1 2013-08-06 Caroline Tice <cmtice@google.com>
3 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5 * tree-pass.h: Add pass_vtable_verify.
6 * varasm.c (assemble_variable): Add code to properly set the comdat
7 section and name for the .vtable_map_vars section.
8 (assemble_vtyv_preinit_initializer): New function.
9 (default_sectin_type_flags): Make sure .vtable_map_vars section has
11 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12 * vtable-verify.c: New file.
13 * vtable-verify.h: New file.
14 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
16 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
17 * passes.def: Insert pass_vtable_verify.
18 * aclocal.m4: Reorder includes.
19 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
21 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
22 as appropriate, if -fvtable-verify=... is used.
23 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
24 -fvtable-verify=... is used.
25 * Makefile.in (OBJS): Add vtable-verify.o to list.
26 (vtable-verify.o): Add new build rule.
27 (GTFILES): Add vtable-verify.c to list.
28 * common.opt (fvtable-verify=): New flag.
29 (vtv_priority): Values for fvtable-verify= flag.
30 (fvtv-counts): New flag.
31 (fvtv-debug): New flag.
32 * tree.h (save_vtable_map_decl): New extern function decl.
34 2013-08-07 David Malcolm <dmalcolm@redhat.com>
36 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
37 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
38 (pass_data_rl78_devirt): ...new pass_data instance and...
39 (make_pass_rl78_devirt): ...new function.
40 (rl78_asm_file_start): Port pass registration to new C++ API.
42 2013-08-07 David Malcolm <dmalcolm@redhat.com>
44 * coretypes.h (rtl_opt_pass): Add.
46 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
47 (epiphany_init): Port to new C++ pass API.
48 (epiphany_optimize_mode_switching): Likewise.
49 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
50 (pass_manager::get_pass_mode_switching): New.
51 (pass_manager::get_pass_peephole2): New.
52 * mode-switching.c (pass_mode_switching): Add clone method.
53 * recog.c (pass_peephole2): Add clone method.
54 (pass_split_all_insns): Add clone method.
56 2013-08-06 David Malcolm <dmalcolm@redhat.com>
58 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move
60 (mips_option_override): ...here, porting to new C++ API for
63 2013-08-06 Jan Hubicka <jh@suse.cz>
65 * cgraph.c (cgraph_get_body): New function based on lto.c
67 * cgraph.h (cgraph_get_body): Declare.
68 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and LTO paths.
69 * cgraphunit.c (expand_function): Get body prior expanding.
70 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
71 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't really need.
72 * passes.c (do_per_function_toporder): Get body.
73 * tree-inline.c (expand_call_inline): Get body prior inlining it.
74 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
76 2013-08-06 Martin Jambor <mjambor@suse.cz>
79 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
80 is not re-finalized. Rename second parameter to no_collect.
82 2013-08-06 Martin Jambor <mjambor@suse.cz>
85 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
86 MEM_REF has proper alignment information.
88 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
91 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
93 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
94 rtx (*) (rtx, ...) with insn_gen_fn.
95 * genoutput.c (output_insn_data): Cast gen_? function pointers to
96 insn_gen_fn::stored_funcptr. Add initializer braces.
98 2013-08-05 David Malcolm <dmalcolm@redhat.com>
100 Rewrite how instances of passes are cloned to remove assumptions
101 about their sizes (thus allowing pass subclasses to have
102 additional data fields, albeit non-GC-managed ones at this point).
104 * passes.c (make_pass_instance): Now that passes have clone
105 methods, rewrite this function to eliminate XNEW and memcpy
106 calls that used hardcoded sizes. Since this function no longer
107 creates pass instances, rename it to...
108 (add_pass_instance): ...this. Document the old way that passes
109 were numbered and flagged, and rework this function to continue
111 (next_pass_1): Add an initial_pass argument for use by
113 (position_pass): When adding multiple instances of a pass, use
114 the pass's clone method, rather than relying on the XNEW/memcpy
115 within the former make_pass_instance (now add_pass_instance).
116 (pass_manager::pass_manager): When invoking next_pass_1, also
117 supply the initial instance of the current pass within the
120 2013-08-05 David Malcolm <dmalcolm@redhat.com>
122 This is the automated part of the conversion of passes from C
123 structs to C++ classes.
125 Patch autogenerated by refactor_passes.py from
126 https://github.com/davidmalcolm/gcc-refactoring-scripts
127 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
129 * asan.c (pass_asan): Convert from a global struct to a subclass of
130 gimple_opt_pass along with...
131 (pass_data_asan): ...new pass_data instance and...
132 (make_pass_asan): ...new function.
133 (pass_asan_O0): Convert from a global struct to a subclass of
134 gimple_opt_pass along with...
135 (pass_data_asan_O0): ...new pass_data instance and...
136 (make_pass_asan_O0): ...new function.
137 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
138 subclass of rtl_opt_pass along with...
139 (pass_data_inc_dec): ...new pass_data instance and...
140 (make_pass_inc_dec): ...new function.
141 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
142 a subclass of rtl_opt_pass along with...
143 (pass_data_reorder_blocks): ...new pass_data instance and...
144 (make_pass_reorder_blocks): ...new function.
145 (pass_duplicate_computed_gotos): Convert from a global struct to a
146 subclass of rtl_opt_pass along with...
147 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
148 (make_pass_duplicate_computed_gotos): ...new function.
149 (pass_partition_blocks): Convert from a global struct to a subclass of
150 rtl_opt_pass along with...
151 (pass_data_partition_blocks): ...new pass_data instance and...
152 (make_pass_partition_blocks): ...new function.
153 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
154 struct to a subclass of rtl_opt_pass along with...
155 (pass_data_branch_target_load_optimize1): ...new pass_data instance
157 (make_pass_branch_target_load_optimize1): ...new function.
158 (pass_branch_target_load_optimize2): Convert from a global struct to a
159 subclass of rtl_opt_pass along with...
160 (pass_data_branch_target_load_optimize2): ...new pass_data instance
162 (make_pass_branch_target_load_optimize2): ...new function.
163 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
164 of rtl_opt_pass along with...
165 (pass_data_jump): ...new pass_data instance and...
166 (make_pass_jump): ...new function.
167 (pass_jump2): Convert from a global struct to a subclass of
168 rtl_opt_pass along with...
169 (pass_data_jump2): ...new pass_data instance and...
170 (make_pass_jump2): ...new function.
171 * cfgexpand.c (pass_expand): Convert from a global struct to a
172 subclass of rtl_opt_pass along with...
173 (pass_data_expand): ...new pass_data instance and...
174 (make_pass_expand): ...new function.
175 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
176 of rtl_opt_pass along with...
177 (pass_data_free_cfg): ...new pass_data instance and...
178 (make_pass_free_cfg): ...new function.
179 (pass_into_cfg_layout_mode): Convert from a global struct to a
180 subclass of rtl_opt_pass along with...
181 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
182 (make_pass_into_cfg_layout_mode): ...new function.
183 (pass_outof_cfg_layout_mode): Convert from a global struct to a
184 subclass of rtl_opt_pass along with...
185 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
186 (make_pass_outof_cfg_layout_mode): ...new function.
187 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
188 struct to a subclass of gimple_opt_pass along with...
189 (pass_data_build_cgraph_edges): ...new pass_data instance and...
190 (make_pass_build_cgraph_edges): ...new function.
191 (pass_rebuild_cgraph_edges): Convert from a global struct to a
192 subclass of gimple_opt_pass along with...
193 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
194 (make_pass_rebuild_cgraph_edges): ...new function.
195 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
196 subclass of gimple_opt_pass along with...
197 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
199 (make_pass_remove_cgraph_callee_edges): ...new function.
200 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
201 struct to a subclass of rtl_opt_pass along with...
202 (pass_data_stack_adjustments): ...new pass_data instance and...
203 (make_pass_stack_adjustments): ...new function.
204 * combine.c (pass_combine): Convert from a global struct to a subclass
205 of rtl_opt_pass along with...
206 (pass_data_combine): ...new pass_data instance and...
207 (make_pass_combine): ...new function.
208 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
209 global struct to a subclass of rtl_opt_pass along with...
210 (pass_data_compare_elim_after_reload): ...new pass_data instance
212 (make_pass_compare_elim_after_reload): ...new function.
213 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
214 of rtl_opt_pass along with...
215 (pass_data_rtl_cprop): ...new pass_data instance and...
216 (make_pass_rtl_cprop): ...new function.
217 * cse.c (pass_cse): Convert from a global struct to a subclass of
218 rtl_opt_pass along with...
219 (pass_data_cse): ...new pass_data instance and...
220 (make_pass_cse): ...new function.
221 (pass_cse2): Convert from a global struct to a subclass of
222 rtl_opt_pass along with...
223 (pass_data_cse2): ...new pass_data instance and...
224 (make_pass_cse2): ...new function.
225 (pass_cse_after_global_opts): Convert from a global struct to a
226 subclass of rtl_opt_pass along with...
227 (pass_data_cse_after_global_opts): ...new pass_data instance and...
228 (make_pass_cse_after_global_opts): ...new function.
229 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
230 of rtl_opt_pass along with...
231 (pass_data_ud_rtl_dce): ...new pass_data instance and...
232 (make_pass_ud_rtl_dce): ...new function.
233 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
234 rtl_opt_pass along with...
235 (pass_data_fast_rtl_dce): ...new pass_data instance and...
236 (make_pass_fast_rtl_dce): ...new function.
237 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
238 a subclass of rtl_opt_pass along with...
239 (pass_data_df_initialize_opt): ...new pass_data instance and...
240 (make_pass_df_initialize_opt): ...new function.
241 (pass_df_initialize_no_opt): Convert from a global struct to a
242 subclass of rtl_opt_pass along with...
243 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
244 (make_pass_df_initialize_no_opt): ...new function.
245 (pass_df_finish): Convert from a global struct to a subclass of
246 rtl_opt_pass along with...
247 (pass_data_df_finish): ...new pass_data instance and...
248 (make_pass_df_finish): ...new function.
249 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
250 rtl_opt_pass along with...
251 (pass_data_rtl_dse1): ...new pass_data instance and...
252 (make_pass_rtl_dse1): ...new function.
253 (pass_rtl_dse2): Convert from a global struct to a subclass of
254 rtl_opt_pass along with...
255 (pass_data_rtl_dse2): ...new pass_data instance and...
256 (make_pass_rtl_dse2): ...new function.
257 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
258 subclass of rtl_opt_pass along with...
259 (pass_data_dwarf2_frame): ...new pass_data instance and...
260 (make_pass_dwarf2_frame): ...new function.
261 * except.c (pass_set_nothrow_function_flags): Convert from a global
262 struct to a subclass of rtl_opt_pass along with...
263 (pass_data_set_nothrow_function_flags): ...new pass_data instance
265 (make_pass_set_nothrow_function_flags): ...new function.
266 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
267 subclass of rtl_opt_pass along with...
268 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
270 (make_pass_convert_to_eh_region_ranges): ...new function.
271 * final.c (pass_compute_alignments): Convert from a global struct to a
272 subclass of rtl_opt_pass along with...
273 (pass_data_compute_alignments): ...new pass_data instance and...
274 (make_pass_compute_alignments): ...new function.
275 (pass_final): Convert from a global struct to a subclass of
276 rtl_opt_pass along with...
277 (pass_data_final): ...new pass_data instance and...
278 (make_pass_final): ...new function.
279 (pass_shorten_branches): Convert from a global struct to a subclass of
280 rtl_opt_pass along with...
281 (pass_data_shorten_branches): ...new pass_data instance and...
282 (make_pass_shorten_branches): ...new function.
283 (pass_clean_state): Convert from a global struct to a subclass of
284 rtl_opt_pass along with...
285 (pass_data_clean_state): ...new pass_data instance and...
286 (make_pass_clean_state): ...new function.
287 * function.c (pass_instantiate_virtual_regs): Convert from a global
288 struct to a subclass of rtl_opt_pass along with...
289 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
290 (make_pass_instantiate_virtual_regs): ...new function.
291 (pass_leaf_regs): Convert from a global struct to a subclass of
292 rtl_opt_pass along with...
293 (pass_data_leaf_regs): ...new pass_data instance and...
294 (make_pass_leaf_regs): ...new function.
295 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
296 subclass of rtl_opt_pass along with...
297 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
299 (make_pass_thread_prologue_and_epilogue): ...new function.
300 (pass_match_asm_constraints): Convert from a global struct to a
301 subclass of rtl_opt_pass along with...
302 (pass_data_match_asm_constraints): ...new pass_data instance and...
303 (make_pass_match_asm_constraints): ...new function.
304 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
305 subclass of rtl_opt_pass along with...
306 (pass_data_rtl_fwprop): ...new pass_data instance and...
307 (make_pass_rtl_fwprop): ...new function.
308 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
309 rtl_opt_pass along with...
310 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
311 (make_pass_rtl_fwprop_addr): ...new function.
312 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
313 rtl_opt_pass along with...
314 (pass_data_rtl_pre): ...new pass_data instance and...
315 (make_pass_rtl_pre): ...new function.
316 (pass_rtl_hoist): Convert from a global struct to a subclass of
317 rtl_opt_pass along with...
318 (pass_data_rtl_hoist): ...new pass_data instance and...
319 (make_pass_rtl_hoist): ...new function.
320 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
321 subclass of gimple_opt_pass along with...
322 (pass_data_lower_cf): ...new pass_data instance and...
323 (make_pass_lower_cf): ...new function.
324 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
325 from a global struct to a subclass of gimple_opt_pass along with...
326 (pass_data_strength_reduction): ...new pass_data instance and...
327 (make_pass_strength_reduction): ...new function.
328 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
329 of rtl_opt_pass along with...
330 (pass_data_rtl_ifcvt): ...new pass_data instance and...
331 (make_pass_rtl_ifcvt): ...new function.
332 (pass_if_after_combine): Convert from a global struct to a subclass of
333 rtl_opt_pass along with...
334 (pass_data_if_after_combine): ...new pass_data instance and...
335 (make_pass_if_after_combine): ...new function.
336 (pass_if_after_reload): Convert from a global struct to a subclass of
337 rtl_opt_pass along with...
338 (pass_data_if_after_reload): ...new pass_data instance and...
339 (make_pass_if_after_reload): ...new function.
340 * init-regs.c (pass_initialize_regs): Convert from a global struct to
341 a subclass of rtl_opt_pass along with...
342 (pass_data_initialize_regs): ...new pass_data instance and...
343 (make_pass_initialize_regs): ...new function.
344 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
345 of ipa_opt_pass_d along with...
346 (pass_data_ipa_cp): ...new pass_data instance and...
347 (make_pass_ipa_cp): ...new function.
348 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
349 global struct to a subclass of gimple_opt_pass along with...
350 (pass_data_inline_parameters): ...new pass_data instance and...
351 (make_pass_inline_parameters): ...new function.
352 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
353 subclass of gimple_opt_pass along with...
354 (pass_data_early_inline): ...new pass_data instance and...
355 (make_pass_early_inline): ...new function.
356 (pass_ipa_inline): Convert from a global struct to a subclass of
357 ipa_opt_pass_d along with...
358 (pass_data_ipa_inline): ...new pass_data instance and...
359 (make_pass_ipa_inline): ...new function.
360 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
361 struct to a subclass of gimple_opt_pass along with...
362 (pass_data_local_pure_const): ...new pass_data instance and...
363 (make_pass_local_pure_const): ...new function.
364 (pass_ipa_pure_const): Convert from a global struct to a subclass of
365 ipa_opt_pass_d along with...
366 (pass_data_ipa_pure_const): ...new pass_data instance and...
367 (make_pass_ipa_pure_const): ...new function.
368 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
369 to a subclass of ipa_opt_pass_d along with...
370 (pass_data_ipa_reference): ...new pass_data instance and...
371 (make_pass_ipa_reference): ...new function.
372 * ipa-split.c (pass_split_functions): Convert from a global struct to
373 a subclass of gimple_opt_pass along with...
374 (pass_data_split_functions): ...new pass_data instance and...
375 (make_pass_split_functions): ...new function.
376 (pass_feedback_split_functions): Convert from a global struct to a
377 subclass of gimple_opt_pass along with...
378 (pass_data_feedback_split_functions): ...new pass_data instance and...
379 (make_pass_feedback_split_functions): ...new function.
380 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
381 global struct to a subclass of simple_ipa_opt_pass along with...
382 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
384 (make_pass_ipa_function_and_variable_visibility): ...new function.
385 (pass_ipa_free_inline_summary): Convert from a global struct to a
386 subclass of simple_ipa_opt_pass along with...
387 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
388 (make_pass_ipa_free_inline_summary): ...new function.
389 (pass_ipa_whole_program_visibility): Convert from a global struct to a
390 subclass of ipa_opt_pass_d along with...
391 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
393 (make_pass_ipa_whole_program_visibility): ...new function.
394 (pass_ipa_profile): Convert from a global struct to a subclass of
395 ipa_opt_pass_d along with...
396 (pass_data_ipa_profile): ...new pass_data instance and...
397 (make_pass_ipa_profile): ...new function.
398 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
399 ipa_opt_pass_d along with...
400 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
401 (make_pass_ipa_cdtor_merge): ...new function.
402 * ira.c (pass_ira): Convert from a global struct to a subclass of
403 rtl_opt_pass along with...
404 (pass_data_ira): ...new pass_data instance and...
405 (make_pass_ira): ...new function.
406 (pass_reload): Convert from a global struct to a subclass of
407 rtl_opt_pass along with...
408 (pass_data_reload): ...new pass_data instance and...
409 (make_pass_reload): ...new function.
410 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
411 subclass of rtl_opt_pass along with...
412 (pass_data_cleanup_barriers): ...new pass_data instance and...
413 (make_pass_cleanup_barriers): ...new function.
414 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
415 of rtl_opt_pass along with...
416 (pass_data_loop2): ...new pass_data instance and...
417 (make_pass_loop2): ...new function.
418 (pass_rtl_loop_init): Convert from a global struct to a subclass of
419 rtl_opt_pass along with...
420 (pass_data_rtl_loop_init): ...new pass_data instance and...
421 (make_pass_rtl_loop_init): ...new function.
422 (pass_rtl_loop_done): Convert from a global struct to a subclass of
423 rtl_opt_pass along with...
424 (pass_data_rtl_loop_done): ...new pass_data instance and...
425 (make_pass_rtl_loop_done): ...new function.
426 (pass_rtl_move_loop_invariants): Convert from a global struct to a
427 subclass of rtl_opt_pass along with...
428 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
429 (make_pass_rtl_move_loop_invariants): ...new function.
430 (pass_rtl_unswitch): Convert from a global struct to a subclass of
431 rtl_opt_pass along with...
432 (pass_data_rtl_unswitch): ...new pass_data instance and...
433 (make_pass_rtl_unswitch): ...new function.
434 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
435 subclass of rtl_opt_pass along with...
436 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
438 (make_pass_rtl_unroll_and_peel_loops): ...new function.
439 (pass_rtl_doloop): Convert from a global struct to a subclass of
440 rtl_opt_pass along with...
441 (pass_data_rtl_doloop): ...new pass_data instance and...
442 (make_pass_rtl_doloop): ...new function.
443 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
444 a subclass of rtl_opt_pass along with...
445 (pass_data_lower_subreg): ...new pass_data instance and...
446 (make_pass_lower_subreg): ...new function.
447 (pass_lower_subreg2): Convert from a global struct to a subclass of
448 rtl_opt_pass along with...
449 (pass_data_lower_subreg2): ...new pass_data instance and...
450 (make_pass_lower_subreg2): ...new function.
451 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
452 struct to a subclass of ipa_opt_pass_d along with...
453 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
454 (make_pass_ipa_lto_gimple_out): ...new function.
455 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
456 of ipa_opt_pass_d along with...
457 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
458 (make_pass_ipa_lto_finish_out): ...new function.
459 * mode-switching.c (pass_mode_switching): Convert from a global struct
460 to a subclass of rtl_opt_pass along with...
461 (pass_data_mode_switching): ...new pass_data instance and...
462 (make_pass_mode_switching): ...new function.
463 * modulo-sched.c (pass_sms): Convert from a global struct to a
464 subclass of rtl_opt_pass along with...
465 (pass_data_sms): ...new pass_data instance and...
466 (make_pass_sms): ...new function.
467 * omp-low.c (pass_expand_omp): Convert from a global struct to a
468 subclass of gimple_opt_pass along with...
469 (pass_data_expand_omp): ...new pass_data instance and...
470 (make_pass_expand_omp): ...new function.
471 (pass_lower_omp): Convert from a global struct to a subclass of
472 gimple_opt_pass along with...
473 (pass_data_lower_omp): ...new pass_data instance and...
474 (make_pass_lower_omp): ...new function.
475 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
476 of gimple_opt_pass along with...
477 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
478 (make_pass_diagnose_omp_blocks): ...new function.
479 * passes.c (pass_early_local_passes): Convert from a global struct to
480 a subclass of simple_ipa_opt_pass along with...
481 (pass_data_early_local_passes): ...new pass_data instance and...
482 (make_pass_early_local_passes): ...new function.
483 (pass_all_early_optimizations): Convert from a global struct to a
484 subclass of gimple_opt_pass along with...
485 (pass_data_all_early_optimizations): ...new pass_data instance and...
486 (make_pass_all_early_optimizations): ...new function.
487 (pass_all_optimizations): Convert from a global struct to a subclass
488 of gimple_opt_pass along with...
489 (pass_data_all_optimizations): ...new pass_data instance and...
490 (make_pass_all_optimizations): ...new function.
491 (pass_all_optimizations_g): Convert from a global struct to a subclass
492 of gimple_opt_pass along with...
493 (pass_data_all_optimizations_g): ...new pass_data instance and...
494 (make_pass_all_optimizations_g): ...new function.
495 (pass_rest_of_compilation): Convert from a global struct to a subclass
496 of rtl_opt_pass along with...
497 (pass_data_rest_of_compilation): ...new pass_data instance and...
498 (make_pass_rest_of_compilation): ...new function.
499 (pass_postreload): Convert from a global struct to a subclass of
500 rtl_opt_pass along with...
501 (pass_data_postreload): ...new pass_data instance and...
502 (make_pass_postreload): ...new function.
503 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
504 subclass of rtl_opt_pass along with...
505 (pass_data_gcse2): ...new pass_data instance and...
506 (make_pass_gcse2): ...new function.
507 * postreload.c (pass_postreload_cse): Convert from a global struct to
508 a subclass of rtl_opt_pass along with...
509 (pass_data_postreload_cse): ...new pass_data instance and...
510 (make_pass_postreload_cse): ...new function.
511 * predict.c (pass_profile): Convert from a global struct to a subclass
512 of gimple_opt_pass along with...
513 (pass_data_profile): ...new pass_data instance and...
514 (make_pass_profile): ...new function.
515 (pass_strip_predict_hints): Convert from a global struct to a subclass
516 of gimple_opt_pass along with...
517 (pass_data_strip_predict_hints): ...new pass_data instance and...
518 (make_pass_strip_predict_hints): ...new function.
519 * recog.c (pass_peephole2): Convert from a global struct to a subclass
520 of rtl_opt_pass along with...
521 (pass_data_peephole2): ...new pass_data instance and...
522 (make_pass_peephole2): ...new function.
523 (pass_split_all_insns): Convert from a global struct to a subclass of
524 rtl_opt_pass along with...
525 (pass_data_split_all_insns): ...new pass_data instance and...
526 (make_pass_split_all_insns): ...new function.
527 (pass_split_after_reload): Convert from a global struct to a subclass
528 of rtl_opt_pass along with...
529 (pass_data_split_after_reload): ...new pass_data instance and...
530 (make_pass_split_after_reload): ...new function.
531 (pass_split_before_regstack): Convert from a global struct to a
532 subclass of rtl_opt_pass along with...
533 (pass_data_split_before_regstack): ...new pass_data instance and...
534 (make_pass_split_before_regstack): ...new function.
535 (pass_split_before_sched2): Convert from a global struct to a subclass
536 of rtl_opt_pass along with...
537 (pass_data_split_before_sched2): ...new pass_data instance and...
538 (make_pass_split_before_sched2): ...new function.
539 (pass_split_for_shorten_branches): Convert from a global struct to a
540 subclass of rtl_opt_pass along with...
541 (pass_data_split_for_shorten_branches): ...new pass_data instance
543 (make_pass_split_for_shorten_branches): ...new function.
544 * ree.c (pass_ree): Convert from a global struct to a subclass of
545 rtl_opt_pass along with...
546 (pass_data_ree): ...new pass_data instance and...
547 (make_pass_ree): ...new function.
548 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
549 subclass of rtl_opt_pass along with...
550 (pass_data_stack_regs): ...new pass_data instance and...
551 (make_pass_stack_regs): ...new function.
552 (pass_stack_regs_run): Convert from a global struct to a subclass of
553 rtl_opt_pass along with...
554 (pass_data_stack_regs_run): ...new pass_data instance and...
555 (make_pass_stack_regs_run): ...new function.
556 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
557 subclass of rtl_opt_pass along with...
558 (pass_data_cprop_hardreg): ...new pass_data instance and...
559 (make_pass_cprop_hardreg): ...new function.
560 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
561 subclass of rtl_opt_pass along with...
562 (pass_data_reginfo_init): ...new pass_data instance and...
563 (make_pass_reginfo_init): ...new function.
564 * regmove.c (pass_regmove): Convert from a global struct to a subclass
565 of rtl_opt_pass along with...
566 (pass_data_regmove): ...new pass_data instance and...
567 (make_pass_regmove): ...new function.
568 * regrename.c (pass_regrename): Convert from a global struct to a
569 subclass of rtl_opt_pass along with...
570 (pass_data_regrename): ...new pass_data instance and...
571 (make_pass_regrename): ...new function.
572 * reorg.c (pass_delay_slots): Convert from a global struct to a
573 subclass of rtl_opt_pass along with...
574 (pass_data_delay_slots): ...new pass_data instance and...
575 (make_pass_delay_slots): ...new function.
576 (pass_machine_reorg): Convert from a global struct to a subclass of
577 rtl_opt_pass along with...
578 (pass_data_machine_reorg): ...new pass_data instance and...
579 (make_pass_machine_reorg): ...new function.
580 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
581 of rtl_opt_pass along with...
582 (pass_data_sched): ...new pass_data instance and...
583 (make_pass_sched): ...new function.
584 (pass_sched2): Convert from a global struct to a subclass of
585 rtl_opt_pass along with...
586 (pass_data_sched2): ...new pass_data instance and...
587 (make_pass_sched2): ...new function.
588 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
589 to a subclass of rtl_opt_pass along with...
590 (pass_data_stack_ptr_mod): ...new pass_data instance and...
591 (make_pass_stack_ptr_mod): ...new function.
592 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
593 to a subclass of rtl_opt_pass along with...
594 (pass_data_rtl_store_motion): ...new pass_data instance and...
595 (make_pass_rtl_store_motion): ...new function.
596 * tracer.c (pass_tracer): Convert from a global struct to a subclass
597 of gimple_opt_pass along with...
598 (pass_data_tracer): ...new pass_data instance and...
599 (make_pass_tracer): ...new function.
600 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
601 to a subclass of gimple_opt_pass along with...
602 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
603 (make_pass_diagnose_tm_blocks): ...new function.
604 (pass_lower_tm): Convert from a global struct to a subclass of
605 gimple_opt_pass along with...
606 (pass_data_lower_tm): ...new pass_data instance and...
607 (make_pass_lower_tm): ...new function.
608 (pass_tm_init): Convert from a global struct to a subclass of
609 gimple_opt_pass along with...
610 (pass_data_tm_init): ...new pass_data instance and...
611 (make_pass_tm_init): ...new function.
612 (pass_tm_mark): Convert from a global struct to a subclass of
613 gimple_opt_pass along with...
614 (pass_data_tm_mark): ...new pass_data instance and...
615 (make_pass_tm_mark): ...new function.
616 (pass_tm_edges): Convert from a global struct to a subclass of
617 gimple_opt_pass along with...
618 (pass_data_tm_edges): ...new pass_data instance and...
619 (make_pass_tm_edges): ...new function.
620 (pass_tm_memopt): Convert from a global struct to a subclass of
621 gimple_opt_pass along with...
622 (pass_data_tm_memopt): ...new pass_data instance and...
623 (make_pass_tm_memopt): ...new function.
624 (pass_ipa_tm): Convert from a global struct to a subclass of
625 simple_ipa_opt_pass along with...
626 (pass_data_ipa_tm): ...new pass_data instance and...
627 (make_pass_ipa_tm): ...new function.
628 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
629 subclass of gimple_opt_pass along with...
630 (pass_data_call_cdce): ...new pass_data instance and...
631 (make_pass_call_cdce): ...new function.
632 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
633 subclass of gimple_opt_pass along with...
634 (pass_data_build_cfg): ...new pass_data instance and...
635 (make_pass_build_cfg): ...new function.
636 (pass_split_crit_edges): Convert from a global struct to a subclass of
637 gimple_opt_pass along with...
638 (pass_data_split_crit_edges): ...new pass_data instance and...
639 (make_pass_split_crit_edges): ...new function.
640 (pass_warn_function_return): Convert from a global struct to a
641 subclass of gimple_opt_pass along with...
642 (pass_data_warn_function_return): ...new pass_data instance and...
643 (make_pass_warn_function_return): ...new function.
644 (pass_warn_function_noreturn): Convert from a global struct to a
645 subclass of gimple_opt_pass along with...
646 (pass_data_warn_function_noreturn): ...new pass_data instance and...
647 (make_pass_warn_function_noreturn): ...new function.
648 (pass_warn_unused_result): Convert from a global struct to a subclass
649 of gimple_opt_pass along with...
650 (pass_data_warn_unused_result): ...new pass_data instance and...
651 (make_pass_warn_unused_result): ...new function.
652 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
653 a subclass of gimple_opt_pass along with...
654 (pass_data_merge_phi): ...new pass_data instance and...
655 (make_pass_merge_phi): ...new function.
656 * tree-complex.c (pass_lower_complex): Convert from a global struct to
657 a subclass of gimple_opt_pass along with...
658 (pass_data_lower_complex): ...new pass_data instance and...
659 (make_pass_lower_complex): ...new function.
660 (pass_lower_complex_O0): Convert from a global struct to a subclass of
661 gimple_opt_pass along with...
662 (pass_data_lower_complex_O0): ...new pass_data instance and...
663 (make_pass_lower_complex_O0): ...new function.
664 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
665 subclass of gimple_opt_pass along with...
666 (pass_data_lower_eh): ...new pass_data instance and...
667 (make_pass_lower_eh): ...new function.
668 (pass_refactor_eh): Convert from a global struct to a subclass of
669 gimple_opt_pass along with...
670 (pass_data_refactor_eh): ...new pass_data instance and...
671 (make_pass_refactor_eh): ...new function.
672 (pass_lower_resx): Convert from a global struct to a subclass of
673 gimple_opt_pass along with...
674 (pass_data_lower_resx): ...new pass_data instance and...
675 (make_pass_lower_resx): ...new function.
676 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
677 of gimple_opt_pass along with...
678 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
679 (make_pass_lower_eh_dispatch): ...new function.
680 (pass_cleanup_eh): Convert from a global struct to a subclass of
681 gimple_opt_pass along with...
682 (pass_data_cleanup_eh): ...new pass_data instance and...
683 (make_pass_cleanup_eh): ...new function.
684 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
685 to a subclass of simple_ipa_opt_pass along with...
686 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
687 (make_pass_ipa_lower_emutls): ...new function.
688 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
689 a subclass of gimple_opt_pass along with...
690 (pass_data_if_conversion): ...new pass_data instance and...
691 (make_pass_if_conversion): ...new function.
692 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
693 subclass of gimple_opt_pass along with...
694 (pass_data_build_ssa): ...new pass_data instance and...
695 (make_pass_build_ssa): ...new function.
696 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
697 global struct to a subclass of gimple_opt_pass along with...
698 (pass_data_loop_distribution): ...new pass_data instance and...
699 (make_pass_loop_distribution): ...new function.
700 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
701 subclass of gimple_opt_pass along with...
702 (pass_data_mudflap_1): ...new pass_data instance and...
703 (make_pass_mudflap_1): ...new function.
704 (pass_mudflap_2): Convert from a global struct to a subclass of
705 gimple_opt_pass along with...
706 (pass_data_mudflap_2): ...new pass_data instance and...
707 (make_pass_mudflap_2): ...new function.
708 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
709 subclass of gimple_opt_pass along with...
710 (pass_data_mudflap_1): ...new pass_data instance and...
711 (make_pass_mudflap_1): ...new function.
712 (pass_mudflap_2): Convert from a global struct to a subclass of
713 gimple_opt_pass along with...
714 (pass_data_mudflap_2): ...new pass_data instance and...
715 (make_pass_mudflap_2): ...new function.
716 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
717 gimple_opt_pass along with...
718 (pass_data_nrv): ...new pass_data instance and...
719 (make_pass_nrv): ...new function.
720 (pass_return_slot): Convert from a global struct to a subclass of
721 gimple_opt_pass along with...
722 (pass_data_return_slot): ...new pass_data instance and...
723 (make_pass_return_slot): ...new function.
724 * tree-object-size.c (pass_object_sizes): Convert from a global struct
725 to a subclass of gimple_opt_pass along with...
726 (pass_data_object_sizes): ...new pass_data instance and...
727 (make_pass_object_sizes): ...new function.
728 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
729 global struct to a subclass of gimple_opt_pass along with...
730 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
732 (make_pass_cleanup_cfg_post_optimizing): ...new function.
733 (pass_fixup_cfg): Convert from a global struct to a subclass of
734 gimple_opt_pass along with...
735 (pass_data_fixup_cfg): ...new pass_data instance and...
736 (make_pass_fixup_cfg): ...new function.
737 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
738 (make_pass_mudflap_1): ...new function.
739 (pass_mudflap_2): Replace declaration with that of...
740 (make_pass_mudflap_2): ...new function.
741 (pass_asan): Replace declaration with that of...
742 (make_pass_asan): ...new function.
743 (pass_asan_O0): Replace declaration with that of...
744 (make_pass_asan_O0): ...new function.
745 (pass_tsan): Replace declaration with that of...
746 (make_pass_tsan): ...new function.
747 (pass_tsan_O0): Replace declaration with that of...
748 (make_pass_tsan_O0): ...new function.
749 (pass_lower_cf): Replace declaration with that of...
750 (make_pass_lower_cf): ...new function.
751 (pass_refactor_eh): Replace declaration with that of...
752 (make_pass_refactor_eh): ...new function.
753 (pass_lower_eh): Replace declaration with that of...
754 (make_pass_lower_eh): ...new function.
755 (pass_lower_eh_dispatch): Replace declaration with that of...
756 (make_pass_lower_eh_dispatch): ...new function.
757 (pass_lower_resx): Replace declaration with that of...
758 (make_pass_lower_resx): ...new function.
759 (pass_build_cfg): Replace declaration with that of...
760 (make_pass_build_cfg): ...new function.
761 (pass_early_tree_profile): Replace declaration with that of...
762 (make_pass_early_tree_profile): ...new function.
763 (pass_cleanup_eh): Replace declaration with that of...
764 (make_pass_cleanup_eh): ...new function.
765 (pass_sra): Replace declaration with that of...
766 (make_pass_sra): ...new function.
767 (pass_sra_early): Replace declaration with that of...
768 (make_pass_sra_early): ...new function.
769 (pass_early_ipa_sra): Replace declaration with that of...
770 (make_pass_early_ipa_sra): ...new function.
771 (pass_tail_recursion): Replace declaration with that of...
772 (make_pass_tail_recursion): ...new function.
773 (pass_tail_calls): Replace declaration with that of...
774 (make_pass_tail_calls): ...new function.
775 (pass_tree_loop): Replace declaration with that of...
776 (make_pass_tree_loop): ...new function.
777 (pass_tree_loop_init): Replace declaration with that of...
778 (make_pass_tree_loop_init): ...new function.
779 (pass_lim): Replace declaration with that of...
780 (make_pass_lim): ...new function.
781 (pass_tree_unswitch): Replace declaration with that of...
782 (make_pass_tree_unswitch): ...new function.
783 (pass_predcom): Replace declaration with that of...
784 (make_pass_predcom): ...new function.
785 (pass_iv_canon): Replace declaration with that of...
786 (make_pass_iv_canon): ...new function.
787 (pass_scev_cprop): Replace declaration with that of...
788 (make_pass_scev_cprop): ...new function.
789 (pass_empty_loop): Replace declaration with that of...
790 (make_pass_empty_loop): ...new function.
791 (pass_record_bounds): Replace declaration with that of...
792 (make_pass_record_bounds): ...new function.
793 (pass_graphite): Replace declaration with that of...
794 (make_pass_graphite): ...new function.
795 (pass_graphite_transforms): Replace declaration with that of...
796 (make_pass_graphite_transforms): ...new function.
797 (pass_if_conversion): Replace declaration with that of...
798 (make_pass_if_conversion): ...new function.
799 (pass_loop_distribution): Replace declaration with that of...
800 (make_pass_loop_distribution): ...new function.
801 (pass_vectorize): Replace declaration with that of...
802 (make_pass_vectorize): ...new function.
803 (pass_slp_vectorize): Replace declaration with that of...
804 (make_pass_slp_vectorize): ...new function.
805 (pass_complete_unroll): Replace declaration with that of...
806 (make_pass_complete_unroll): ...new function.
807 (pass_complete_unrolli): Replace declaration with that of...
808 (make_pass_complete_unrolli): ...new function.
809 (pass_parallelize_loops): Replace declaration with that of...
810 (make_pass_parallelize_loops): ...new function.
811 (pass_loop_prefetch): Replace declaration with that of...
812 (make_pass_loop_prefetch): ...new function.
813 (pass_iv_optimize): Replace declaration with that of...
814 (make_pass_iv_optimize): ...new function.
815 (pass_tree_loop_done): Replace declaration with that of...
816 (make_pass_tree_loop_done): ...new function.
817 (pass_ch): Replace declaration with that of...
818 (make_pass_ch): ...new function.
819 (pass_ccp): Replace declaration with that of...
820 (make_pass_ccp): ...new function.
821 (pass_phi_only_cprop): Replace declaration with that of...
822 (make_pass_phi_only_cprop): ...new function.
823 (pass_build_ssa): Replace declaration with that of...
824 (make_pass_build_ssa): ...new function.
825 (pass_build_alias): Replace declaration with that of...
826 (make_pass_build_alias): ...new function.
827 (pass_build_ealias): Replace declaration with that of...
828 (make_pass_build_ealias): ...new function.
829 (pass_dominator): Replace declaration with that of...
830 (make_pass_dominator): ...new function.
831 (pass_dce): Replace declaration with that of...
832 (make_pass_dce): ...new function.
833 (pass_dce_loop): Replace declaration with that of...
834 (make_pass_dce_loop): ...new function.
835 (pass_cd_dce): Replace declaration with that of...
836 (make_pass_cd_dce): ...new function.
837 (pass_call_cdce): Replace declaration with that of...
838 (make_pass_call_cdce): ...new function.
839 (pass_merge_phi): Replace declaration with that of...
840 (make_pass_merge_phi): ...new function.
841 (pass_split_crit_edges): Replace declaration with that of...
842 (make_pass_split_crit_edges): ...new function.
843 (pass_pre): Replace declaration with that of...
844 (make_pass_pre): ...new function.
845 (pass_profile): Replace declaration with that of...
846 (make_pass_profile): ...new function.
847 (pass_strip_predict_hints): Replace declaration with that of...
848 (make_pass_strip_predict_hints): ...new function.
849 (pass_lower_complex_O0): Replace declaration with that of...
850 (make_pass_lower_complex_O0): ...new function.
851 (pass_lower_complex): Replace declaration with that of...
852 (make_pass_lower_complex): ...new function.
853 (pass_lower_vector): Replace declaration with that of...
854 (make_pass_lower_vector): ...new function.
855 (pass_lower_vector_ssa): Replace declaration with that of...
856 (make_pass_lower_vector_ssa): ...new function.
857 (pass_lower_omp): Replace declaration with that of...
858 (make_pass_lower_omp): ...new function.
859 (pass_diagnose_omp_blocks): Replace declaration with that of...
860 (make_pass_diagnose_omp_blocks): ...new function.
861 (pass_expand_omp): Replace declaration with that of...
862 (make_pass_expand_omp): ...new function.
863 (pass_expand_omp_ssa): Replace declaration with that of...
864 (make_pass_expand_omp_ssa): ...new function.
865 (pass_object_sizes): Replace declaration with that of...
866 (make_pass_object_sizes): ...new function.
867 (pass_strlen): Replace declaration with that of...
868 (make_pass_strlen): ...new function.
869 (pass_fold_builtins): Replace declaration with that of...
870 (make_pass_fold_builtins): ...new function.
871 (pass_stdarg): Replace declaration with that of...
872 (make_pass_stdarg): ...new function.
873 (pass_early_warn_uninitialized): Replace declaration with that of...
874 (make_pass_early_warn_uninitialized): ...new function.
875 (pass_late_warn_uninitialized): Replace declaration with that of...
876 (make_pass_late_warn_uninitialized): ...new function.
877 (pass_cse_reciprocals): Replace declaration with that of...
878 (make_pass_cse_reciprocals): ...new function.
879 (pass_cse_sincos): Replace declaration with that of...
880 (make_pass_cse_sincos): ...new function.
881 (pass_optimize_bswap): Replace declaration with that of...
882 (make_pass_optimize_bswap): ...new function.
883 (pass_optimize_widening_mul): Replace declaration with that of...
884 (make_pass_optimize_widening_mul): ...new function.
885 (pass_warn_function_return): Replace declaration with that of...
886 (make_pass_warn_function_return): ...new function.
887 (pass_warn_function_noreturn): Replace declaration with that of...
888 (make_pass_warn_function_noreturn): ...new function.
889 (pass_cselim): Replace declaration with that of...
890 (make_pass_cselim): ...new function.
891 (pass_phiopt): Replace declaration with that of...
892 (make_pass_phiopt): ...new function.
893 (pass_forwprop): Replace declaration with that of...
894 (make_pass_forwprop): ...new function.
895 (pass_phiprop): Replace declaration with that of...
896 (make_pass_phiprop): ...new function.
897 (pass_tree_ifcombine): Replace declaration with that of...
898 (make_pass_tree_ifcombine): ...new function.
899 (pass_dse): Replace declaration with that of...
900 (make_pass_dse): ...new function.
901 (pass_nrv): Replace declaration with that of...
902 (make_pass_nrv): ...new function.
903 (pass_rename_ssa_copies): Replace declaration with that of...
904 (make_pass_rename_ssa_copies): ...new function.
905 (pass_sink_code): Replace declaration with that of...
906 (make_pass_sink_code): ...new function.
907 (pass_fre): Replace declaration with that of...
908 (make_pass_fre): ...new function.
909 (pass_check_data_deps): Replace declaration with that of...
910 (make_pass_check_data_deps): ...new function.
911 (pass_copy_prop): Replace declaration with that of...
912 (make_pass_copy_prop): ...new function.
913 (pass_vrp): Replace declaration with that of...
914 (make_pass_vrp): ...new function.
915 (pass_uncprop): Replace declaration with that of...
916 (make_pass_uncprop): ...new function.
917 (pass_return_slot): Replace declaration with that of...
918 (make_pass_return_slot): ...new function.
919 (pass_reassoc): Replace declaration with that of...
920 (make_pass_reassoc): ...new function.
921 (pass_rebuild_cgraph_edges): Replace declaration with that of...
922 (make_pass_rebuild_cgraph_edges): ...new function.
923 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
924 (make_pass_remove_cgraph_callee_edges): ...new function.
925 (pass_build_cgraph_edges): Replace declaration with that of...
926 (make_pass_build_cgraph_edges): ...new function.
927 (pass_local_pure_const): Replace declaration with that of...
928 (make_pass_local_pure_const): ...new function.
929 (pass_tracer): Replace declaration with that of...
930 (make_pass_tracer): ...new function.
931 (pass_warn_unused_result): Replace declaration with that of...
932 (make_pass_warn_unused_result): ...new function.
933 (pass_diagnose_tm_blocks): Replace declaration with that of...
934 (make_pass_diagnose_tm_blocks): ...new function.
935 (pass_lower_tm): Replace declaration with that of...
936 (make_pass_lower_tm): ...new function.
937 (pass_tm_init): Replace declaration with that of...
938 (make_pass_tm_init): ...new function.
939 (pass_tm_mark): Replace declaration with that of...
940 (make_pass_tm_mark): ...new function.
941 (pass_tm_memopt): Replace declaration with that of...
942 (make_pass_tm_memopt): ...new function.
943 (pass_tm_edges): Replace declaration with that of...
944 (make_pass_tm_edges): ...new function.
945 (pass_split_functions): Replace declaration with that of...
946 (make_pass_split_functions): ...new function.
947 (pass_feedback_split_functions): Replace declaration with that of...
948 (make_pass_feedback_split_functions): ...new function.
949 (pass_strength_reduction): Replace declaration with that of...
950 (make_pass_strength_reduction): ...new function.
951 (pass_ipa_lower_emutls): Replace declaration with that of...
952 (make_pass_ipa_lower_emutls): ...new function.
953 (pass_ipa_function_and_variable_visibility): Replace declaration with
955 (make_pass_ipa_function_and_variable_visibility): ...new function.
956 (pass_ipa_tree_profile): Replace declaration with that of...
957 (make_pass_ipa_tree_profile): ...new function.
958 (pass_early_local_passes): Replace declaration with that of...
959 (make_pass_early_local_passes): ...new function.
960 (pass_ipa_whole_program_visibility): Replace declaration with that
962 (make_pass_ipa_whole_program_visibility): ...new function.
963 (pass_ipa_lto_gimple_out): Replace declaration with that of...
964 (make_pass_ipa_lto_gimple_out): ...new function.
965 (pass_ipa_increase_alignment): Replace declaration with that of...
966 (make_pass_ipa_increase_alignment): ...new function.
967 (pass_ipa_inline): Replace declaration with that of...
968 (make_pass_ipa_inline): ...new function.
969 (pass_ipa_free_lang_data): Replace declaration with that of...
970 (make_pass_ipa_free_lang_data): ...new function.
971 (pass_ipa_free_inline_summary): Replace declaration with that of...
972 (make_pass_ipa_free_inline_summary): ...new function.
973 (pass_ipa_cp): Replace declaration with that of...
974 (make_pass_ipa_cp): ...new function.
975 (pass_ipa_reference): Replace declaration with that of...
976 (make_pass_ipa_reference): ...new function.
977 (pass_ipa_pure_const): Replace declaration with that of...
978 (make_pass_ipa_pure_const): ...new function.
979 (pass_ipa_pta): Replace declaration with that of...
980 (make_pass_ipa_pta): ...new function.
981 (pass_ipa_lto_finish_out): Replace declaration with that of...
982 (make_pass_ipa_lto_finish_out): ...new function.
983 (pass_ipa_tm): Replace declaration with that of...
984 (make_pass_ipa_tm): ...new function.
985 (pass_ipa_profile): Replace declaration with that of...
986 (make_pass_ipa_profile): ...new function.
987 (pass_ipa_cdtor_merge): Replace declaration with that of...
988 (make_pass_ipa_cdtor_merge): ...new function.
989 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
991 (make_pass_cleanup_cfg_post_optimizing): ...new function.
992 (pass_init_datastructures): Replace declaration with that of...
993 (make_pass_init_datastructures): ...new function.
994 (pass_fixup_cfg): Replace declaration with that of...
995 (make_pass_fixup_cfg): ...new function.
996 (pass_expand): Replace declaration with that of...
997 (make_pass_expand): ...new function.
998 (pass_instantiate_virtual_regs): Replace declaration with that of...
999 (make_pass_instantiate_virtual_regs): ...new function.
1000 (pass_rtl_fwprop): Replace declaration with that of...
1001 (make_pass_rtl_fwprop): ...new function.
1002 (pass_rtl_fwprop_addr): Replace declaration with that of...
1003 (make_pass_rtl_fwprop_addr): ...new function.
1004 (pass_jump): Replace declaration with that of...
1005 (make_pass_jump): ...new function.
1006 (pass_jump2): Replace declaration with that of...
1007 (make_pass_jump2): ...new function.
1008 (pass_lower_subreg): Replace declaration with that of...
1009 (make_pass_lower_subreg): ...new function.
1010 (pass_cse): Replace declaration with that of...
1011 (make_pass_cse): ...new function.
1012 (pass_fast_rtl_dce): Replace declaration with that of...
1013 (make_pass_fast_rtl_dce): ...new function.
1014 (pass_ud_rtl_dce): Replace declaration with that of...
1015 (make_pass_ud_rtl_dce): ...new function.
1016 (pass_rtl_dce): Replace declaration with that of...
1017 (make_pass_rtl_dce): ...new function.
1018 (pass_rtl_dse1): Replace declaration with that of...
1019 (make_pass_rtl_dse1): ...new function.
1020 (pass_rtl_dse2): Replace declaration with that of...
1021 (make_pass_rtl_dse2): ...new function.
1022 (pass_rtl_dse3): Replace declaration with that of...
1023 (make_pass_rtl_dse3): ...new function.
1024 (pass_rtl_cprop): Replace declaration with that of...
1025 (make_pass_rtl_cprop): ...new function.
1026 (pass_rtl_pre): Replace declaration with that of...
1027 (make_pass_rtl_pre): ...new function.
1028 (pass_rtl_hoist): Replace declaration with that of...
1029 (make_pass_rtl_hoist): ...new function.
1030 (pass_rtl_store_motion): Replace declaration with that of...
1031 (make_pass_rtl_store_motion): ...new function.
1032 (pass_cse_after_global_opts): Replace declaration with that of...
1033 (make_pass_cse_after_global_opts): ...new function.
1034 (pass_rtl_ifcvt): Replace declaration with that of...
1035 (make_pass_rtl_ifcvt): ...new function.
1036 (pass_into_cfg_layout_mode): Replace declaration with that of...
1037 (make_pass_into_cfg_layout_mode): ...new function.
1038 (pass_outof_cfg_layout_mode): Replace declaration with that of...
1039 (make_pass_outof_cfg_layout_mode): ...new function.
1040 (pass_loop2): Replace declaration with that of...
1041 (make_pass_loop2): ...new function.
1042 (pass_rtl_loop_init): Replace declaration with that of...
1043 (make_pass_rtl_loop_init): ...new function.
1044 (pass_rtl_move_loop_invariants): Replace declaration with that of...
1045 (make_pass_rtl_move_loop_invariants): ...new function.
1046 (pass_rtl_unswitch): Replace declaration with that of...
1047 (make_pass_rtl_unswitch): ...new function.
1048 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
1049 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1050 (pass_rtl_doloop): Replace declaration with that of...
1051 (make_pass_rtl_doloop): ...new function.
1052 (pass_rtl_loop_done): Replace declaration with that of...
1053 (make_pass_rtl_loop_done): ...new function.
1054 (pass_web): Replace declaration with that of...
1055 (make_pass_web): ...new function.
1056 (pass_cse2): Replace declaration with that of...
1057 (make_pass_cse2): ...new function.
1058 (pass_df_initialize_opt): Replace declaration with that of...
1059 (make_pass_df_initialize_opt): ...new function.
1060 (pass_df_initialize_no_opt): Replace declaration with that of...
1061 (make_pass_df_initialize_no_opt): ...new function.
1062 (pass_reginfo_init): Replace declaration with that of...
1063 (make_pass_reginfo_init): ...new function.
1064 (pass_inc_dec): Replace declaration with that of...
1065 (make_pass_inc_dec): ...new function.
1066 (pass_stack_ptr_mod): Replace declaration with that of...
1067 (make_pass_stack_ptr_mod): ...new function.
1068 (pass_initialize_regs): Replace declaration with that of...
1069 (make_pass_initialize_regs): ...new function.
1070 (pass_combine): Replace declaration with that of...
1071 (make_pass_combine): ...new function.
1072 (pass_if_after_combine): Replace declaration with that of...
1073 (make_pass_if_after_combine): ...new function.
1074 (pass_ree): Replace declaration with that of...
1075 (make_pass_ree): ...new function.
1076 (pass_partition_blocks): Replace declaration with that of...
1077 (make_pass_partition_blocks): ...new function.
1078 (pass_match_asm_constraints): Replace declaration with that of...
1079 (make_pass_match_asm_constraints): ...new function.
1080 (pass_regmove): Replace declaration with that of...
1081 (make_pass_regmove): ...new function.
1082 (pass_split_all_insns): Replace declaration with that of...
1083 (make_pass_split_all_insns): ...new function.
1084 (pass_fast_rtl_byte_dce): Replace declaration with that of...
1085 (make_pass_fast_rtl_byte_dce): ...new function.
1086 (pass_lower_subreg2): Replace declaration with that of...
1087 (make_pass_lower_subreg2): ...new function.
1088 (pass_mode_switching): Replace declaration with that of...
1089 (make_pass_mode_switching): ...new function.
1090 (pass_sms): Replace declaration with that of...
1091 (make_pass_sms): ...new function.
1092 (pass_sched): Replace declaration with that of...
1093 (make_pass_sched): ...new function.
1094 (pass_ira): Replace declaration with that of...
1095 (make_pass_ira): ...new function.
1096 (pass_reload): Replace declaration with that of...
1097 (make_pass_reload): ...new function.
1098 (pass_clean_state): Replace declaration with that of...
1099 (make_pass_clean_state): ...new function.
1100 (pass_branch_prob): Replace declaration with that of...
1101 (make_pass_branch_prob): ...new function.
1102 (pass_value_profile_transformations): Replace declaration with that
1104 (make_pass_value_profile_transformations): ...new function.
1105 (pass_postreload_cse): Replace declaration with that of...
1106 (make_pass_postreload_cse): ...new function.
1107 (pass_gcse2): Replace declaration with that of...
1108 (make_pass_gcse2): ...new function.
1109 (pass_split_after_reload): Replace declaration with that of...
1110 (make_pass_split_after_reload): ...new function.
1111 (pass_branch_target_load_optimize1): Replace declaration with that
1113 (make_pass_branch_target_load_optimize1): ...new function.
1114 (pass_thread_prologue_and_epilogue): Replace declaration with that
1116 (make_pass_thread_prologue_and_epilogue): ...new function.
1117 (pass_stack_adjustments): Replace declaration with that of...
1118 (make_pass_stack_adjustments): ...new function.
1119 (pass_peephole2): Replace declaration with that of...
1120 (make_pass_peephole2): ...new function.
1121 (pass_if_after_reload): Replace declaration with that of...
1122 (make_pass_if_after_reload): ...new function.
1123 (pass_regrename): Replace declaration with that of...
1124 (make_pass_regrename): ...new function.
1125 (pass_cprop_hardreg): Replace declaration with that of...
1126 (make_pass_cprop_hardreg): ...new function.
1127 (pass_reorder_blocks): Replace declaration with that of...
1128 (make_pass_reorder_blocks): ...new function.
1129 (pass_branch_target_load_optimize2): Replace declaration with that
1131 (make_pass_branch_target_load_optimize2): ...new function.
1132 (pass_leaf_regs): Replace declaration with that of...
1133 (make_pass_leaf_regs): ...new function.
1134 (pass_split_before_sched2): Replace declaration with that of...
1135 (make_pass_split_before_sched2): ...new function.
1136 (pass_compare_elim_after_reload): Replace declaration with that of...
1137 (make_pass_compare_elim_after_reload): ...new function.
1138 (pass_sched2): Replace declaration with that of...
1139 (make_pass_sched2): ...new function.
1140 (pass_stack_regs): Replace declaration with that of...
1141 (make_pass_stack_regs): ...new function.
1142 (pass_stack_regs_run): Replace declaration with that of...
1143 (make_pass_stack_regs_run): ...new function.
1144 (pass_df_finish): Replace declaration with that of...
1145 (make_pass_df_finish): ...new function.
1146 (pass_compute_alignments): Replace declaration with that of...
1147 (make_pass_compute_alignments): ...new function.
1148 (pass_duplicate_computed_gotos): Replace declaration with that of...
1149 (make_pass_duplicate_computed_gotos): ...new function.
1150 (pass_variable_tracking): Replace declaration with that of...
1151 (make_pass_variable_tracking): ...new function.
1152 (pass_free_cfg): Replace declaration with that of...
1153 (make_pass_free_cfg): ...new function.
1154 (pass_machine_reorg): Replace declaration with that of...
1155 (make_pass_machine_reorg): ...new function.
1156 (pass_cleanup_barriers): Replace declaration with that of...
1157 (make_pass_cleanup_barriers): ...new function.
1158 (pass_delay_slots): Replace declaration with that of...
1159 (make_pass_delay_slots): ...new function.
1160 (pass_split_for_shorten_branches): Replace declaration with that of...
1161 (make_pass_split_for_shorten_branches): ...new function.
1162 (pass_split_before_regstack): Replace declaration with that of...
1163 (make_pass_split_before_regstack): ...new function.
1164 (pass_convert_to_eh_region_ranges): Replace declaration with that
1166 (make_pass_convert_to_eh_region_ranges): ...new function.
1167 (pass_shorten_branches): Replace declaration with that of...
1168 (make_pass_shorten_branches): ...new function.
1169 (pass_set_nothrow_function_flags): Replace declaration with that of...
1170 (make_pass_set_nothrow_function_flags): ...new function.
1171 (pass_dwarf2_frame): Replace declaration with that of...
1172 (make_pass_dwarf2_frame): ...new function.
1173 (pass_final): Replace declaration with that of...
1174 (make_pass_final): ...new function.
1175 (pass_rtl_seqabstr): Replace declaration with that of...
1176 (make_pass_rtl_seqabstr): ...new function.
1177 (pass_release_ssa_names): Replace declaration with that of...
1178 (make_pass_release_ssa_names): ...new function.
1179 (pass_early_inline): Replace declaration with that of...
1180 (make_pass_early_inline): ...new function.
1181 (pass_inline_parameters): Replace declaration with that of...
1182 (make_pass_inline_parameters): ...new function.
1183 (pass_update_address_taken): Replace declaration with that of...
1184 (make_pass_update_address_taken): ...new function.
1185 (pass_convert_switch): Replace declaration with that of...
1186 (make_pass_convert_switch): ...new function.
1187 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1188 to a subclass of simple_ipa_opt_pass along with...
1189 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1190 (make_pass_ipa_tree_profile): ...new function.
1191 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1192 subclass of gimple_opt_pass along with...
1193 (pass_data_sra_early): ...new pass_data instance and...
1194 (make_pass_sra_early): ...new function.
1195 (pass_sra): Convert from a global struct to a subclass of
1196 gimple_opt_pass along with...
1197 (pass_data_sra): ...new pass_data instance and...
1198 (make_pass_sra): ...new function.
1199 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1200 gimple_opt_pass along with...
1201 (pass_data_early_ipa_sra): ...new pass_data instance and...
1202 (make_pass_early_ipa_sra): ...new function.
1203 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1204 subclass of gimple_opt_pass along with...
1205 (pass_data_ccp): ...new pass_data instance and...
1206 (make_pass_ccp): ...new function.
1207 (pass_fold_builtins): Convert from a global struct to a subclass of
1208 gimple_opt_pass along with...
1209 (pass_data_fold_builtins): ...new pass_data instance and...
1210 (make_pass_fold_builtins): ...new function.
1211 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1212 subclass of gimple_opt_pass along with...
1213 (pass_data_copy_prop): ...new pass_data instance and...
1214 (make_pass_copy_prop): ...new function.
1215 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1216 global struct to a subclass of gimple_opt_pass along with...
1217 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1218 (make_pass_rename_ssa_copies): ...new function.
1219 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1220 subclass of gimple_opt_pass along with...
1221 (pass_data_dce): ...new pass_data instance and...
1222 (make_pass_dce): ...new function.
1223 (pass_dce_loop): Convert from a global struct to a subclass of
1224 gimple_opt_pass along with...
1225 (pass_data_dce_loop): ...new pass_data instance and...
1226 (make_pass_dce_loop): ...new function.
1227 (pass_cd_dce): Convert from a global struct to a subclass of
1228 gimple_opt_pass along with...
1229 (pass_data_cd_dce): ...new pass_data instance and...
1230 (make_pass_cd_dce): ...new function.
1231 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1232 subclass of gimple_opt_pass along with...
1233 (pass_data_dominator): ...new pass_data instance and...
1234 (make_pass_dominator): ...new function.
1235 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1236 gimple_opt_pass along with...
1237 (pass_data_phi_only_cprop): ...new pass_data instance and...
1238 (make_pass_phi_only_cprop): ...new function.
1239 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1240 subclass of gimple_opt_pass along with...
1241 (pass_data_dse): ...new pass_data instance and...
1242 (make_pass_dse): ...new function.
1243 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1244 a subclass of gimple_opt_pass along with...
1245 (pass_data_forwprop): ...new pass_data instance and...
1246 (make_pass_forwprop): ...new function.
1247 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1248 struct to a subclass of gimple_opt_pass along with...
1249 (pass_data_tree_ifcombine): ...new pass_data instance and...
1250 (make_pass_tree_ifcombine): ...new function.
1251 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1252 subclass of gimple_opt_pass along with...
1253 (pass_data_ch): ...new pass_data instance and...
1254 (make_pass_ch): ...new function.
1255 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1256 subclass of gimple_opt_pass along with...
1257 (pass_data_tree_loop): ...new pass_data instance and...
1258 (make_pass_tree_loop): ...new function.
1259 (pass_tree_loop_init): Convert from a global struct to a subclass of
1260 gimple_opt_pass along with...
1261 (pass_data_tree_loop_init): ...new pass_data instance and...
1262 (make_pass_tree_loop_init): ...new function.
1263 (pass_lim): Convert from a global struct to a subclass of
1264 gimple_opt_pass along with...
1265 (pass_data_lim): ...new pass_data instance and...
1266 (make_pass_lim): ...new function.
1267 (pass_tree_unswitch): Convert from a global struct to a subclass of
1268 gimple_opt_pass along with...
1269 (pass_data_tree_unswitch): ...new pass_data instance and...
1270 (make_pass_tree_unswitch): ...new function.
1271 (pass_predcom): Convert from a global struct to a subclass of
1272 gimple_opt_pass along with...
1273 (pass_data_predcom): ...new pass_data instance and...
1274 (make_pass_predcom): ...new function.
1275 (pass_vectorize): Convert from a global struct to a subclass of
1276 gimple_opt_pass along with...
1277 (pass_data_vectorize): ...new pass_data instance and...
1278 (make_pass_vectorize): ...new function.
1279 (pass_graphite): Convert from a global struct to a subclass of
1280 gimple_opt_pass along with...
1281 (pass_data_graphite): ...new pass_data instance and...
1282 (make_pass_graphite): ...new function.
1283 (pass_graphite_transforms): Convert from a global struct to a subclass
1284 of gimple_opt_pass along with...
1285 (pass_data_graphite_transforms): ...new pass_data instance and...
1286 (make_pass_graphite_transforms): ...new function.
1287 (pass_check_data_deps): Convert from a global struct to a subclass of
1288 gimple_opt_pass along with...
1289 (pass_data_check_data_deps): ...new pass_data instance and...
1290 (make_pass_check_data_deps): ...new function.
1291 (pass_iv_canon): Convert from a global struct to a subclass of
1292 gimple_opt_pass along with...
1293 (pass_data_iv_canon): ...new pass_data instance and...
1294 (make_pass_iv_canon): ...new function.
1295 (pass_scev_cprop): Convert from a global struct to a subclass of
1296 gimple_opt_pass along with...
1297 (pass_data_scev_cprop): ...new pass_data instance and...
1298 (make_pass_scev_cprop): ...new function.
1299 (pass_record_bounds): Convert from a global struct to a subclass of
1300 gimple_opt_pass along with...
1301 (pass_data_record_bounds): ...new pass_data instance and...
1302 (make_pass_record_bounds): ...new function.
1303 (pass_complete_unroll): Convert from a global struct to a subclass of
1304 gimple_opt_pass along with...
1305 (pass_data_complete_unroll): ...new pass_data instance and...
1306 (make_pass_complete_unroll): ...new function.
1307 (pass_complete_unrolli): Convert from a global struct to a subclass of
1308 gimple_opt_pass along with...
1309 (pass_data_complete_unrolli): ...new pass_data instance and...
1310 (make_pass_complete_unrolli): ...new function.
1311 (pass_parallelize_loops): Convert from a global struct to a subclass
1312 of gimple_opt_pass along with...
1313 (pass_data_parallelize_loops): ...new pass_data instance and...
1314 (make_pass_parallelize_loops): ...new function.
1315 (pass_loop_prefetch): Convert from a global struct to a subclass of
1316 gimple_opt_pass along with...
1317 (pass_data_loop_prefetch): ...new pass_data instance and...
1318 (make_pass_loop_prefetch): ...new function.
1319 (pass_iv_optimize): Convert from a global struct to a subclass of
1320 gimple_opt_pass along with...
1321 (pass_data_iv_optimize): ...new pass_data instance and...
1322 (make_pass_iv_optimize): ...new function.
1323 (pass_tree_loop_done): Convert from a global struct to a subclass of
1324 gimple_opt_pass along with...
1325 (pass_data_tree_loop_done): ...new pass_data instance and...
1326 (make_pass_tree_loop_done): ...new function.
1327 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1328 struct to a subclass of gimple_opt_pass along with...
1329 (pass_data_cse_reciprocals): ...new pass_data instance and...
1330 (make_pass_cse_reciprocals): ...new function.
1331 (pass_cse_sincos): Convert from a global struct to a subclass of
1332 gimple_opt_pass along with...
1333 (pass_data_cse_sincos): ...new pass_data instance and...
1334 (make_pass_cse_sincos): ...new function.
1335 (pass_optimize_bswap): Convert from a global struct to a subclass of
1336 gimple_opt_pass along with...
1337 (pass_data_optimize_bswap): ...new pass_data instance and...
1338 (make_pass_optimize_bswap): ...new function.
1339 (pass_optimize_widening_mul): Convert from a global struct to a
1340 subclass of gimple_opt_pass along with...
1341 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1342 (make_pass_optimize_widening_mul): ...new function.
1343 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1344 subclass of gimple_opt_pass along with...
1345 (pass_data_phiopt): ...new pass_data instance and...
1346 (make_pass_phiopt): ...new function.
1347 (pass_cselim): Convert from a global struct to a subclass of
1348 gimple_opt_pass along with...
1349 (pass_data_cselim): ...new pass_data instance and...
1350 (make_pass_cselim): ...new function.
1351 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1352 subclass of gimple_opt_pass along with...
1353 (pass_data_phiprop): ...new pass_data instance and...
1354 (make_pass_phiprop): ...new function.
1355 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1356 subclass of gimple_opt_pass along with...
1357 (pass_data_pre): ...new pass_data instance and...
1358 (make_pass_pre): ...new function.
1359 (pass_fre): Convert from a global struct to a subclass of
1360 gimple_opt_pass along with...
1361 (pass_data_fre): ...new pass_data instance and...
1362 (make_pass_fre): ...new function.
1363 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1364 subclass of gimple_opt_pass along with...
1365 (pass_data_reassoc): ...new pass_data instance and...
1366 (make_pass_reassoc): ...new function.
1367 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1368 subclass of gimple_opt_pass along with...
1369 (pass_data_sink_code): ...new pass_data instance and...
1370 (make_pass_sink_code): ...new function.
1371 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1372 subclass of gimple_opt_pass along with...
1373 (pass_data_strlen): ...new pass_data instance and...
1374 (make_pass_strlen): ...new function.
1375 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1376 struct to a subclass of gimple_opt_pass along with...
1377 (pass_data_build_alias): ...new pass_data instance and...
1378 (make_pass_build_alias): ...new function.
1379 (pass_build_ealias): Convert from a global struct to a subclass of
1380 gimple_opt_pass along with...
1381 (pass_data_build_ealias): ...new pass_data instance and...
1382 (make_pass_build_ealias): ...new function.
1383 (pass_ipa_pta): Convert from a global struct to a subclass of
1384 simple_ipa_opt_pass along with...
1385 (pass_data_ipa_pta): ...new pass_data instance and...
1386 (make_pass_ipa_pta): ...new function.
1387 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1388 subclass of gimple_opt_pass along with...
1389 (pass_data_uncprop): ...new pass_data instance and...
1390 (make_pass_uncprop): ...new function.
1391 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1392 global struct to a subclass of gimple_opt_pass along with...
1393 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1394 (make_pass_late_warn_uninitialized): ...new function.
1395 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1396 to a subclass of gimple_opt_pass along with...
1397 (pass_data_init_datastructures): ...new pass_data instance and...
1398 (make_pass_init_datastructures): ...new function.
1399 (pass_early_warn_uninitialized): Convert from a global struct to a
1400 subclass of gimple_opt_pass along with...
1401 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1402 (make_pass_early_warn_uninitialized): ...new function.
1403 (pass_update_address_taken): Convert from a global struct to a
1404 subclass of gimple_opt_pass along with...
1405 (pass_data_update_address_taken): ...new pass_data instance and...
1406 (make_pass_update_address_taken): ...new function.
1407 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1408 struct to a subclass of gimple_opt_pass along with...
1409 (pass_data_release_ssa_names): ...new pass_data instance and...
1410 (make_pass_release_ssa_names): ...new function.
1411 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1412 subclass of gimple_opt_pass along with...
1413 (pass_data_stdarg): ...new pass_data instance and...
1414 (make_pass_stdarg): ...new function.
1415 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1416 global struct to a subclass of gimple_opt_pass along with...
1417 (pass_data_convert_switch): ...new pass_data instance and...
1418 (make_pass_convert_switch): ...new function.
1419 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1420 to a subclass of gimple_opt_pass along with...
1421 (pass_data_tail_recursion): ...new pass_data instance and...
1422 (make_pass_tail_recursion): ...new function.
1423 (pass_tail_calls): Convert from a global struct to a subclass of
1424 gimple_opt_pass along with...
1425 (pass_data_tail_calls): ...new pass_data instance and...
1426 (make_pass_tail_calls): ...new function.
1427 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1428 struct to a subclass of gimple_opt_pass along with...
1429 (pass_data_lower_vector): ...new pass_data instance and...
1430 (make_pass_lower_vector): ...new function.
1431 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1432 gimple_opt_pass along with...
1433 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1434 (make_pass_lower_vector_ssa): ...new function.
1435 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
1436 to a subclass of gimple_opt_pass along with...
1437 (pass_data_slp_vectorize): ...new pass_data instance and...
1438 (make_pass_slp_vectorize): ...new function.
1439 (pass_ipa_increase_alignment): Convert from a global struct to a
1440 subclass of simple_ipa_opt_pass along with...
1441 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
1442 (make_pass_ipa_increase_alignment): ...new function.
1443 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
1444 gimple_opt_pass along with...
1445 (pass_data_vrp): ...new pass_data instance and...
1446 (make_pass_vrp): ...new function.
1447 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
1448 subclass of simple_ipa_opt_pass along with...
1449 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
1450 (make_pass_ipa_free_lang_data): ...new function.
1451 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
1452 gimple_opt_pass along with...
1453 (pass_data_tsan): ...new pass_data instance and...
1454 (make_pass_tsan): ...new function.
1455 (pass_tsan_O0): Convert from a global struct to a subclass of
1456 gimple_opt_pass along with...
1457 (pass_data_tsan_O0): ...new pass_data instance and...
1458 (make_pass_tsan_O0): ...new function.
1459 * var-tracking.c (pass_variable_tracking): Convert from a global
1460 struct to a subclass of rtl_opt_pass along with...
1461 (pass_data_variable_tracking): ...new pass_data instance and...
1462 (make_pass_variable_tracking): ...new function.
1463 * web.c (pass_web): Convert from a global struct to a subclass of
1464 rtl_opt_pass along with...
1465 (pass_data_web): ...new pass_data instance and...
1466 (make_pass_web): ...new function.
1467 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
1468 declaration with that of...
1469 (make_pass_mode_switch_use): ...new function.
1470 (pass_resolve_sw_modes): Replace declaration with that of...
1471 (make_pass_resolve_sw_modes): ...new function.
1472 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
1473 from a global struct to a subclass of rtl_opt_pass along with...
1474 (pass_data_mode_switch_use): ...new pass_data instance and...
1475 (make_pass_mode_switch_use): ...new function.
1476 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
1477 from a global struct to a subclass of rtl_opt_pass along with...
1478 (pass_data_resolve_sw_modes): ...new pass_data instance and...
1479 (make_pass_resolve_sw_modes): ...new function.
1480 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
1481 struct to a subclass of rtl_opt_pass along with...
1482 (pass_data_insert_vzeroupper): ...new pass_data instance and...
1483 (make_pass_insert_vzeroupper): ...new function.
1484 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
1485 global struct to a subclass of rtl_opt_pass along with...
1486 (pass_data_work_around_errata): ...new pass_data instance and...
1487 (make_pass_work_around_errata): ...new function.
1488 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
1489 struct to a subclass of rtl_opt_pass along with...
1490 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
1491 (make_pass_mips_machine_reorg2): ...new function.
1493 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1495 * passes.c (pass_manager::operator new): New.
1497 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1499 Handwritten part of conversion of passes to C++ classes.
1501 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
1502 (toplev.o): Add dep on PASS_MANAGER_H.
1503 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
1504 of early local pases to reflect this moving from a global to a
1505 member of gcc::pass_manager.
1506 (cgraph_add_new_function): Likewise.
1507 * lto-cgraph.c (lto_output_node): Update for conversion of
1508 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
1509 * passes.c (opt_pass::clone): New.
1510 (opt_pass::gate): New.
1511 (opt_pass::execute): New.
1512 (opt_pass::opt_pass): New.
1513 (pass_manager::execute_early_local_passes): New.
1514 (pass_manager::execute_pass_mode_switching): new.
1515 (finish_optimization_passes): Convert to...
1516 (pass_manager::finish_optimization_passes): ...this.
1517 (finish_optimization_passes): Update for conversion of passes to
1519 (register_dump_files_1): Use has_gate since we cannot portably
1520 check a vtable entry against NULL.
1521 (dump_one_pass): Likewise.
1522 (ipa_write_summaries_2): Likewise.
1523 (ipa_write_optimization_summaries_1): Likewise.
1524 (ipa_read_summaries_1): Likewise.
1525 (ipa_read_optimization_summaries_1): Likewise.
1526 (execute_ipa_stmt_fixups): Likewise.
1527 (pass_manager::pass_manager): Rewrite pass-creation, invoking
1528 pass-creation functions rather than wiring up globals, and
1529 storing the results in fields of pass_manager generated using
1531 (pass_manager::dump_profile_report): Update for conversion of
1532 passes to C++ classes.
1533 (pass_manager::execute_ipa_summary_passes): Likewise.
1534 (execute_one_ipa_transform_pass): Likewise.
1535 (execute_one_pass): Use has_gate and has_execute since we cannot
1536 portably check a vtable entry against NULL.
1537 * pass_manager.h (pass_manager::finish_optimization_passes): New.
1538 (pass_manager): Use pass-instances.def to add fields for the
1539 various pass instances.
1540 * toplev.c (finalize): Update for move of
1541 finish_optimization_passes to a method of gcc::pass_manager.
1542 * toplev.h (finish_optimization_passes): Move to method of class
1544 * tree-pass.h (struct pass_data): New.
1545 (opt_pass): Convert to C++ class, make it a subclass of
1547 (opt_pass::gate): Convert to virtual function.
1548 (opt_pass::~opt_pass): New.
1549 (opt_pass::clone): New.
1550 (opt_pass::execute): Convert to virtual function.
1551 (opt_pass::opt_pass): New.
1552 (opt_pass::ctxt_): new.
1553 (gimple_opt_pass): Convert to subclass of opt_pass.
1554 (gimple_opt_pass::gimple_opt_pass): New.
1555 (rtl_opt_pass): Convert to subclass of opt_pass.
1556 (rtl_opt_pass::rtl_opt_pass): New.
1557 (ipa_opt_pass_d): Convert to subclass of opt_pass.
1558 (ipa_opt_pass_d::ipa_opt_pass_d): New.
1559 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
1560 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
1561 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
1562 invocation of pass_mode_switching to reflect this moving from a
1563 global to a member of gcc::pass_manager.
1564 (ix86_option_override): Rework how pass_insert_vzeroupper is
1565 added to the pass_manager to reflect autogenerated changes.
1566 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and
1569 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
1571 PR rtl-optimization/57708
1572 * recog.c (peep2_find_free_register): Validate all regs in a
1575 2013-08-05 Jan Hubicka <jh@suse.cz>
1578 * cgraph.c (verify_cgraph_node): Accept local flags from other partitions.
1579 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
1580 (function_and_variable_visibility): Likewise.
1581 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
1583 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1585 * graph.c (init_graph_slim_pretty_print): Remove.
1586 (print_graph_cfg): Do not call it. Use local pretty printer.
1587 (start_graph_dump): Likewise.
1589 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1591 * gimple-pretty-print.c (buffer): Remove.
1592 (initialized): Likewise.
1593 (maybe_init_pretty_print): Likewise.
1594 (print_gimple_stmt): Do not call it. Use non-static local
1595 pretty_printer variable.
1596 (print_gimple_expr): Likewise.
1597 (print_gimple_seq): Likewise.
1598 (gimple_dump_bb): Likewise.
1600 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1602 * asan.c (asan_pp): Remove.
1603 (asan_pp_initialized): Likewise.
1604 (asan_pp_initialize): Likewise.
1605 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
1606 (asan_emit_stack_protection): Tidy. Use local pretty printer.
1607 (asan_add_global): Likewise.
1609 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
1611 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
1612 * diagnostic.h (diagnostic_flush_buffer): Adjust.
1613 * pretty-print.c (pp_formatted_text_data): Likewise.
1614 (pp_indent): Rename from pp_base_indent.
1615 (pp_format): Rename from pp_base_format.
1616 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
1617 (pp_format_verbatim): Rename from pp_base_format_verbatim.
1618 (pp_flush): Rename from pp_base_flush.
1619 (pp_set_line_maximum_length): Rename from
1620 pp_base_set_line_maximum_length.
1621 (pp_clear_output_area): Rename from pp_base_clear_output_area.
1622 (pp_set_prefix): Rename from pp_base_set_prefix.
1623 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
1624 (pp_emit_prefix): Rename from pp_base_emit_prefix.
1625 (pp_append_text): Rename from pp_base_append_text.
1626 (pp_formatted_text): Rename from pp_base_formatted_text.
1627 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
1628 (pp_remaining_character_count_for_line): Rename from
1629 pp_base_remaining_character_count_for_line.
1630 (pp_newline): Rename from pp_base_newline.
1631 (pp_character): Rename from pp_base_character.
1632 (pp_string): Rename from pp_base_string.
1633 (pp_maybe_space): Rename from pp_base_maybe_space.
1634 * asan.c (asan_pp_string): Adjust.
1635 (asan_emit_stack_protection): Likewise.
1636 (asan_add_global): Likewise.
1637 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
1638 * tree-mudflap.c (mf_varname_tree): Likewise.
1639 * tree-pretty-print.c (pp_tree_identifier): Rename from
1640 pp_base_tree_identifier.
1641 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
1642 Declare as function.
1644 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1646 * pretty-print.h (pp_bar_bar): New.
1647 (pp_ampersand_ampersand): Likewise.
1648 (pp_less_equal): Likewise.
1649 (pp_greater_equal): Likewise.
1650 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
1651 printer functions instead of pp_string or operators and punctuators.
1652 (dump_gimple_call): Likewise.
1653 (dump_gimple_omp_for): Likewise.
1654 (dump_gimple_transaction): Likewise.
1655 (dump_gimple_phi): Likewise.
1656 (pp_gimple_stmt_1): Likewise.
1657 * sched-vis.c (print_insn): Likewise.
1658 * tree-mudflap.c (mf_varname_tree): Likewise.
1659 * tree-pretty-print.c (dump_block_node): Likewise.
1660 (dump_generic_node): Likewise.
1662 2013-08-02 Jan Hubicka <jh@suse.cz>
1664 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
1666 * lto-streamer-out.c (tree_is_indexable): Results decls and
1667 parm decls are not indexable.
1668 (DFS_write_tree_body): Do not follow args and results.
1669 (hash_tree): Likewise.
1670 (output_functions): Rearrange so struct function is needed
1671 only when real body is output; be able to also ouptut abstract
1672 functions; output DECL_ARGUMENTS and DECL_RESULT.
1673 (lto_output): When not in WPA, ale store abstract functions.
1674 (write_symbol): Do not care about RESULT_DECL.
1675 (output_symbol_p): Handle correctly sbtract decls.
1676 * lto-streamer-in.c (input_function): Rearrange so struct
1677 function can be NULL at entry; allow streaming of
1678 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
1679 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
1680 sanity check during LTO.
1681 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
1682 RESULT_DECl and DECL_ARGUMENTS.
1683 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
1686 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1688 * pretty-print.h (pp_underscore): New.
1690 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
1691 printer functions instead of pp_character.
1692 (dump_binary_rhs): Likewise.
1693 (dump_ternary_rhs): Likewise.
1694 (dump_gimple_call_args): Likewise.
1695 (pp_points_to_solution): Likewise.
1696 (dump_gimple_call): Likewise.
1697 (dump_gimple_switch): Likewise.
1698 (dump_gimple_cond): Likewise.
1699 (dump_gimple_bind): Likewise.
1700 (dump_gimple_try): Likewise.
1701 (dump_gimple_omp_for): Likewise.
1702 (dump_gimple_omp_continue): Likewise.
1703 (dump_gimple_omp_single): Likewise.
1704 (dump_gimple_omp_sections): Likewise.
1705 (dump_gimple_omp_block): Likewise.
1706 (dump_gimple_omp_critical): Likewise.
1707 (dump_gimple_transaction): Likewise.
1708 (dump_gimple_asm): Likewise.
1709 (dump_gimple_phi): Likewise.
1710 (dump_gimple_omp_parallel): Likewise.
1711 (dump_gimple_omp_task): Likewise.
1712 (dump_gimple_omp_atomic_load): Likewise.
1713 (dump_gimple_omp_atomic_store): Likewise.
1714 (dump_gimple_mem_ops): Likewise.
1715 (pp_gimple_stmt_1): Likewise.
1716 (pp_cfg_jump): Likewise.
1717 (dump_implicit_edges): Likewise.
1718 (gimple_dump_bb_for_graph): Likewise.
1719 * graph.c (draw_cfg_node): Likewise.
1720 * langhooks.c (lhd_print_error_function): Likewise.
1721 * sched-vis.c (print_exp): Likewise.
1722 (print_value): Likewise.
1723 (print_pattern): Likewise.
1724 (print_insn): Likewise.
1725 (rtl_dump_bb_for_graph): Likewise.
1726 * tree-pretty-print.c (dump_function_declaration): Likewise.
1727 (dump_array_domain): Likewise.
1728 (dump_omp_clause): Likewise.
1729 (dump_location): Likewise.
1730 (dump_generic_node): Likewise.
1731 (print_struct_decl): Likewise.
1732 * diagnostic.c (diagnostic_show_locus): Use pp_space.
1734 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1736 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
1737 candidate table when replacing a candidate statement.
1738 (replace_rhs_if_not_dup): Likewise.
1739 (replace_one_candidate): Likewise.
1741 2013-08-02 Jan Hubicka <jh@suse.cz>
1742 Martin Liska <marxin.liska@gmail.com>
1744 * cgraphunit.c (add_new_function): Fix logic when adding from
1746 (assemble_thunk): Rename to ...
1747 (expand_thunk); .. this one; export; get it working with general functions;
1748 make produced gimple valid.
1749 * cgraph.h (expand_thunk): Declare.
1751 2013-08-02 Jan Hubicka <jh@suse.cz>
1753 * ipa-cp.c (gather_context_independent_values): Use ipa_get_param_move_cost.
1754 (get_replacement_map): Remove PARAM; move parameter folding into tree-inline.c
1755 (create_specialized_node): Update.
1756 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
1757 assert that we have gimple body; update move_cost.
1758 (count_formal_params): Assert that we have gimple body.
1759 (ipa_dump_param): New function.
1760 (ipa_alloc_node_params): Break out from ...
1761 (ipa_initialize_node_params): ... here.
1762 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
1763 (ipa_write_node_info): Stream move costs.
1764 (ipa_read_node_info): Read move costs.
1765 (ipa_update_after_lto_read): Do not recompute node params.
1766 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
1767 (ipa_get_param): Check we are not in WPA.
1768 (ipa_get_param_move_cost): New.
1769 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
1770 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
1771 parm numbers to be present.
1773 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1775 PR rtl-optimization/58048
1776 * lra-constraints.c (process_alt_operands): Don't check asm
1777 operand on register.
1779 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
1781 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
1782 the implied StoreLoad barrier for atomic operations if before.
1784 2013-08-02 Jan Hubicka <jh@suse.cz>
1785 Martin Liska <marxin.liska@gmail.com>
1787 * cgraph.c (cgraph_function_body_availability): Do not check cgrpah flags.
1788 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
1789 symtab_node_availability): Declare.
1790 * ipa.c (can_replace_by_local_alias): New.
1791 (function_and_variable_visibility): Use it.
1792 * symtab.c (symtab_for_node_and_aliases, symtab_nonoverwritable_alias_1,
1793 symtab_nonoverwritable_alias): New.
1795 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1797 PR rtl-optimization/57963
1798 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p):
1800 (lra_constraints): Use them.
1802 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
1804 * config/arm/types.md (define_attr "type"): Add "load_acq" and "store_rel".
1805 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
1807 (cortex_a53_store1): Likewise.
1809 2013-08-01 Jan Hubicka <jh@suse.cz>
1811 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other partitions are
1814 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
1816 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
1817 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
1818 MAYBE_NON_Q_CLASS_P where appropriate.
1820 2013-08-01 Jan Hubicka <jh@suse.cz>
1822 * cgraph.h (release_function_body): Declare.
1823 * tree.c (free_lang_data_in_decl): Free, parameters and return values
1824 of unused delcarations.
1826 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1828 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical RTL form
1829 when subtracting a constant.
1831 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1833 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
1834 Generate canonical plus rtx with negated immediate instead of minus
1836 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
1838 2013-08-01 Jan Hubicka <jh@suse.cz>
1840 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
1841 (cgraph_release_function_body): Likewise.
1842 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1843 * cgraph.h (cgrpah_node): Rename abstract_and_needed
1844 to used_as_abstract_origin.
1845 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
1846 symbols used as abstract origins.
1847 * cgraphunit.c (analyze_functions): Update.
1848 * ipa.c (symtab_remove_unreachable_nodes): Recompute
1849 used_as_abstract_origin.
1850 * tree-inline.c (tree_function_versioning): Update
1851 used_as_abstract_origin; be ready for DECL_RESULT and
1852 DECL_ARGUMENTS to be NULL.
1854 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
1855 for abstract functions.
1856 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
1859 2013-08-01 Jan Hubicka <jh@suse.cz>
1861 * profile.c (compute_value_histograms): Fix thinko.
1863 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1865 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
1866 aarch-common-protos.h to extra_headers.
1867 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1868 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
1869 * config/aarch64/t-aarch64 (aarch-common.o): Define.
1871 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1873 * config/aarch64/aarch64.md (define_attr "type"): Delete.
1874 Include "../arm/types.md". Define "type" attribute for all patterns.
1875 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
1878 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
1880 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
1881 to support power8 load fusion.
1882 (fusion_gpr_mem_load): Likewise.
1884 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
1886 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
1887 declarations for power8 load fusion.
1888 (emit_fusion_gpr_load): Likewise.
1890 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
1891 tuning for power8, turn on fusion mode by default. Turn on sign
1892 extending fusion mode if normal fusion mode is on, and we are at
1894 (fusion_gpr_load_p): New function, return true if we can fuse an
1895 addis instruction with a dependent load to a GPR.
1896 (emit_fusion_gpr_load): Emit the instructions for power8 load
1899 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
1900 (VSX load fusion peepholes): New peepholes to fuse together an
1901 addi instruction with a VSX load instruction.
1903 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
1904 peepholes to fuse an addis instruction with a load to a GPR base
1905 register. If we are supporting sign extending fusions, convert
1906 sign extending loads to zero extending loads and add an explicit
1909 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1911 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
1912 aarch-common-protos.h to extra_headers.
1913 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1914 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
1915 (arm_early_store_addr_dep): Likewise.
1916 (arm_no_early_alu_shift_dep): Likewise.
1917 (arm_no_early_alu_shift_value_dep): Likewise.
1918 (arm_no_early_mul_dep): Likewise.
1919 (arm_no_early_store_addr_dep): Likewise.
1920 (arm_mac_accumulator_is_mul_result): Likewise.
1921 (arm_mac_accumulator_is_result): Likewise.
1922 * config/arm/aarch-common.c: ... here. New file.
1923 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
1925 (arm_early_store_addr_dep): Likewise.
1926 (arm_no_early_alu_shift_dep): Likewise.
1927 (arm_no_early_alu_shift_value_dep): Likewise.
1928 (arm_no_early_mul_dep): Likewise.
1929 (arm_no_early_store_addr_dep): Likewise.
1930 (arm_mac_accumulator_is_mul_result): Likewise.
1931 (arm_mac_accumulator_is_result): Likewise.
1932 * config/arm/aarch-common-protos.h: ... here. New file.
1933 * config/arm/t-arm (aarch-common.o): Define.
1935 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1937 * config/arm/arm.md: Include new file "types.md".
1938 (define_attr "type"): Move from here to ...
1939 (define_attr "mul32"): Likewise.
1940 (define_attr "mul64"): Likewise.
1941 * config/arm/types.md: ... here. New file.
1943 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
1945 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
1946 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
1948 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1950 * gen-pass-instances.awk: Fix offset of substr().
1952 2013-07-31 David Malcolm <dmalcolm@redhat.com>
1954 * Makefile.in (pass-instances.def): New.
1955 (passes.o): Replace dependency on passes.def with one on
1958 * gen-pass-instances.awk: New.
1960 * passes.c (pass_manager::pass_manager): Use pass-instances.def
1961 rather than passes.def, updating local definition of NEXT_PASS
1962 macro to add an extra NUM parameter (currently unused).
1964 2013-07-30 David Malcolm <dmalcolm@redhat.com>
1966 * Makefile.in (PASS_MANAGER_H): New.
1967 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
1968 (passes.o): Likewise.
1969 (statistics.o): Likewise.
1970 (cgraphunit.o): Likewise.
1971 (context.o): Depend on PASS_MANAGER_H.
1973 * pass_manager.h: New.
1975 * cgraphunit.c (cgraph_add_new_function): Update for moves
1976 of globals to fields of pass_manager.
1977 (analyze_function): Likewise.
1978 (expand_function): Likewise.
1979 (ipa_passes): Likewise.
1980 (compile): Likewise.
1982 * context.c (context::context): New.
1983 * context.h (context::context): New.
1984 (context::get_passes): New.
1985 (context::passes_): New.
1987 * lto-cgraph.c (input_node): Update for moves of globals to
1988 fields of pass_manager.
1990 * passes.c (all_passes): Remove, in favor of a field of the
1991 same name within the new class pass_manager.
1992 (all_small_ipa_passes): Likewise.
1993 (all_lowering_passes): Likewise.
1994 (all_regular_ipa_passes): Likewise.
1995 (all_late_ipa_passes): Likewise.
1996 (all_lto_gen_passes): Likewise.
1997 (passes_by_id): Likewise.
1998 (passes_by_id_size): Likewise.
1999 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2000 the new class pass_manager.
2001 (set_pass_for_id): Convert to...
2002 (pass_manager::set_pass_for_id): ...method.
2003 (get_pass_for_id): Convert to...
2004 (pass_manager::get_pass_for_id): ...method.
2005 (register_one_dump_file): Move body of implementation into...
2006 (pass_manager::register_one_dump_file): ...here.
2007 (register_dump_files_1): Convert to...
2008 (pass_manager::register_dump_files_1): ...method.
2009 (register_dump_files): Convert to...
2010 (pass_manager::register_dump_files): ...method.
2011 (create_pass_tab): Update for moves of globals to fields of
2013 (dump_passes): Move body of implementation into...
2014 (pass_manager::dump_passes): ...here.
2015 (register_pass): Move body of implementation into...
2016 (pass_manager::register_pass): ...here.
2017 (init_optimization_passes): Convert into...
2018 (pass_manager::pass_manager): ...constructor for new
2019 pass_manager class, and initialize the pass_lists array.
2020 (check_profile_consistency): Update for moves of globals to
2021 fields of pass_manager.
2022 (dump_profile_report): Move body of implementation into...
2023 (pass_manager::dump_profile_report): ...here.
2024 (ipa_write_summaries_1): Update for moves of pass lists from
2025 being globals to fields of pass_manager.
2026 (ipa_write_optimization_summaries): Likewise.
2027 (ipa_read_summaries): Likewise.
2028 (ipa_read_optimization_summaries): Likewise.
2029 (execute_all_ipa_stmt_fixups): Likewise.
2031 * statistics.c (statistics_fini): Update for moves of globals to
2032 fields of pass_manager.
2034 * toplev.c (general_init): Replace call to
2035 init_optimization_passes with construction of the pass_manager
2038 * tree-pass.h (all_passes): Remove, in favor of a field of the
2039 same name within the new class pass_manager.
2040 (all_small_ipa_passes): Likewise.
2041 (all_lowering_passes): Likewise.
2042 (all_regular_ipa_passes): Likewise.
2043 (all_lto_gen_passes): Likewise.
2044 (all_late_ipa_passes): Likewise.
2045 (passes_by_id): Likewise.
2046 (passes_by_id_size): Likewise.
2047 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2048 the new class pass_manager.
2049 (get_pass_for_id): Remove.
2051 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2053 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
2056 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2058 * arm.md (mulhi3): New expand pattern.
2060 2013-07-30 Jan Hubicka <jh@suse.cz>
2061 Martin Liska <marxin.liska@gmail.com>
2063 * profile.c (compute_value_histograms): Do not ICE when
2064 there is mismatch only on some counters.
2066 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2068 PR rtl-optimization/57637
2069 * function.c (move_insn_for_shrink_wrap): Also check the
2070 GEN set of the LIVE problem for the liveness analysis
2071 if it exists, otherwise give up.
2073 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2075 PR tree-optimization/57993
2076 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
2077 replaced statement in the candidate table.
2078 (phi_add_costs): Return infinite cost when the hidden basis does
2079 not dominate all phis on which the candidate is dependent.
2080 (replace_one_candidate): Record replaced statement in the
2083 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
2085 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
2086 (ashlv2si3): New expander.
2087 (*ashlv2si3_i): New define_insn_and_split.
2088 * predicates.md (float_operation): Allow patterns with three
2091 PR rtl-optimization/58021
2092 * mode-switching.c (create_pre_exit): Always split off preceding
2093 insns if we are not at the basic block head.
2095 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2097 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2098 (UCLIBC_DYNAMIC_LINKER): New macro.
2099 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2101 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2102 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2104 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2105 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2106 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2107 for SF and DF modes. Use ieee_quad_format for TF mode.
2108 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2109 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2110 (mips_option_override): Handle `-mnan=legacy'.
2111 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2112 `-mabs=2008' and `-mnan=2008'.
2113 (OPTION_DEFAULT_SPECS): Add "nan" default.
2114 (ASM_SPEC): Handle `-mnan='.
2115 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2116 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2117 comment accordingly.
2118 (neg<mode>2): Likewise.
2119 * config/mips/mips.opt (mabs, mnan): New options.
2120 * doc/install.texi (Configuration): Document `--with-nan=' option.
2121 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2123 (MIPS Options): Document them.
2124 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2125 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2126 * configure: Regenerate.
2127 * config.in: Regenerate.
2129 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2131 * config/i386/i386.md (float post-reload splitters): Do not check
2132 for subregs of SSE registers.
2134 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2135 H.J. Lu <hongjiu.lu@intel.com>
2139 * config/i386/i386.md (post-reload splitter
2140 to avoid partial SSE reg dependency stalls): New pattern.
2142 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2144 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2146 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2148 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2149 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2151 2013-07-26 Cary Coutant <ccoutant@google.com>
2153 * dwarf2out.c (die_checksum_ordered): Don't include template
2154 instantiations in signature.
2155 (is_template_parameter): New function.
2156 (is_template_instantiation): New function.
2157 (generate_skeleton_bottom_up): Don't include template instantiations
2159 (generate_skeleton): Likewise.
2160 (break_out_comdat_types): Move recursive call to break out nested
2162 (prune_unused_types_mark_generic_parms_dies): Call
2163 is_template_parameter.
2165 2013-07-26 Ian Bolton <ian.bolton@arm.com>
2167 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2168 uses vector registers.
2169 * config/aarch64/iterators.md: Add attributes rtn and vas.
2171 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2172 Richard Earnshaw <richard.earnshaw@arm.com>
2174 * combine.c (simplify_comparison): Re-canonicalize operands
2176 * config/arm/arm.md (movcond_addsi): New splitter.
2178 2013-07-25 Sterling Augustine <saugustine@google.com>
2180 * dwarf2out.c (size_of_pubnames): Move code to...
2181 (include_pubname_in_output): ...here. New.
2182 (want_pubnames): Rearrange.
2183 (output_pubnames): Call include_pubname_in_output. Move assertion.
2185 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2187 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2189 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2192 * doc/extend.texi: Remove obsolete builtins. Fix
2193 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2195 2013-07-25 Jan Hubicka <jh@suse.cz>
2197 * cgraph.c (release_function_body): Break out from ...
2198 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2200 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2201 old_tree in the map.
2202 (create_specialized_node): Update.
2203 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2205 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2206 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
2207 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2208 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2211 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2213 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2214 addsi3_carryin_alt2_<optab>): Correct output template.
2216 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2218 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2219 Adjust for arm_restrict_it.
2220 Remove trailing whitespace.
2222 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2224 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2225 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2227 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2229 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2231 PR rtl-optimization/57960
2232 * lra-constraints.c (process_alt_operands): Use the right mode
2233 when checking strict_low.
2235 2013-07-25 Jan Hubicka <jh@suse.cz>
2237 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2238 * cgraph.c (cgraph_remove_node): Do not release function body
2239 when in cgraph streaming.
2240 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
2241 in other partitions are not considered reachable; fix handling of
2244 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2246 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2248 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2253 * config/arm/arm.md ("*sibcall_insn): Replace use of
2254 Ss with US. Adjust output for v5 and v4t.
2255 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
2257 * config/arm/constraints.md ("Ss"): Rename to US.
2259 2013-07-25 Terry Guo <terry.guo@arm.com>
2261 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2262 shift_add/shift_sub0/shift_sub1 RTXs.
2264 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2265 Anton Blanchard <anton@au1.ibm.com>
2267 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2268 (altivec_vpks<VI_char>ss): Likewise.
2269 (altivec_vpks<VI_char>us): Likewise.
2270 (altivec_vpku<VI_char>us): Likewise.
2271 (altivec_vpku<VI_char>um): Likewise.
2273 2013-07-24 David Malcolm <dmalcolm@redhat.com>
2275 Introduce context class.
2277 * Makefile.in (CONTEXT_H): New.
2278 (OBJS): Add context.o.
2279 (toplev.o): Add CONTEXT_H to dependencies.
2282 * toplev.c (general_init): Create the singleton gcc::context instance.
2288 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2290 PR rtl-optimization/57968
2291 * mode-switching.c (create_pre_exit): Allow instructions that
2292 don't set a return register to need a non-exit mode.
2294 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2295 Anton Blanchard <anton@au1.ibm.com>
2297 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2298 operands to vperm for little endian.
2299 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2300 of lvsl to create the control mask for a vperm for little endian.
2302 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2303 Anton Blanchard <anton@au1.ibm.com>
2305 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2306 two operands for little-endian.
2308 2013-07-23 Steve Ellcey <sellcey@mips.com>
2310 * config/mips/mips.c (mips_case_values_threshold): New.
2311 (TARGET_CASE_VALUES_THRESHOLD): Define.
2313 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2314 Anton Blanchard <anton@au1.ibm.com>
2316 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2317 selection of field for vector splat in little endian mode.
2319 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2321 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2322 expanders to rs6000.md.
2323 (ior<mode>3): Likewise.
2324 (and<mode>3): Likewise.
2325 (one_cmpl<mode>2): Likewise.
2326 (nor<mode>3): Likewise.
2327 (andc<mode>3): Likewise.
2328 (eqv<mode>3): Likewise.
2329 (nand<mode>3): Likewise.
2330 (orc<mode>3): Likewise.
2332 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2335 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2336 to split multi-word logical operations.
2337 (rs6000_split_logical_di): Likewise.
2338 (rs6000_split_logical): Likewise.
2340 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2341 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2342 and allow TImode operations in 32-bit.
2343 (vsx_and<mode>3_64bit): Likewise.
2344 (vsx_ior<mode>3_32bit): Likewise.
2345 (vsx_ior<mode>3_64bit): Likewise.
2346 (vsx_xor<mode>3_32bit): Likewise.
2347 (vsx_xor<mode>3_64bit): Likewise.
2348 (vsx_one_cmpl<mode>2_32bit): Likewise.
2349 (vsx_one_cmpl<mode>2_64bit): Likewise.
2350 (vsx_nor<mode>3_32bit): Likewise.
2351 (vsx_nor<mode>3_64bit): Likewise.
2352 (vsx_andc<mode>3_32bit): Likewise.
2353 (vsx_andc<mode>3_64bit): Likewise.
2354 (vsx_eqv<mode>3_32bit): Likewise.
2355 (vsx_eqv<mode>3_64bit): Likewise.
2356 (vsx_nand<mode>3_32bit): Likewise.
2357 (vsx_nand<mode>3_64bit): Likewise.
2358 (vsx_orc<mode>3_32bit): Likewise.
2359 (vsx_orc<mode>3_64bit): Likewise.
2361 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2362 logical types in GPRs.
2364 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2365 logical insns to rs6000.md, and allow TImode operations in
2367 (altivec_ior<mode>3): Likewise.
2368 (altivec_xor<mode>3): Likewise.
2369 (altivec_one_cmpl<mode>2): Likewise.
2370 (altivec_nor<mode>3): Likewise.
2371 (altivec_andc<mode>3): Likewise.
2373 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2374 attributes for moving the 128-bit logical operations into
2376 (BOOL_REGS_OUTPUT): Likewise.
2377 (BOOL_REGS_OP1): Likewise.
2378 (BOOL_REGS_OP2): Likewise.
2379 (BOOL_REGS_UNARY): Likewise.
2380 (BOOL_REGS_AND_CR0): Likewise.
2381 (one_cmpl<mode>2): Add support for DI logical operations on
2382 32-bit, splitting the operations to 32-bit.
2386 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2387 changes to combine the 32/64-bit code, allow logical operations on
2388 TI mode in 32-bit, and to use similar match_operator patterns like
2389 scalar mode uses. Combine the Altivec and VSX code for logical
2390 operations, and move it here.
2391 (ior<mode>3, 128-bit types): Likewise.
2392 (xor<mode>3, 128-bit types): Likewise.
2393 (one_cmpl<mode>3, 128-bit types): Likewise.
2394 (nor<mode>3, 128-bit types): Likewise.
2395 (andc<mode>3, 128-bit types): Likewise.
2396 (eqv<mode>3, 128-bit types): Likewise.
2397 (nand<mode>3, 128-bit types): Likewise.
2398 (orc<mode>3, 128-bit types): Likewise.
2399 (and<mode>3_internal): Likewise.
2400 (bool<mode>3_internal): Likewise.
2401 (boolc<mode>3_internal1): Likewise.
2402 (boolc<mode>3_internal2): Likewise.
2403 (boolcc<mode>3_internal1): Likewise.
2404 (boolcc<mode>3_internal2): Likewise.
2405 (eqv<mode>3_internal1): Likewise.
2406 (eqv<mode>3_internal2): Likewise.
2407 (one_cmpl1<mode>3_internal): Likewise.
2409 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2411 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2412 Rename flag_stack_usage to flag_stack_usage_info.
2414 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2416 * config/microblaze/sync.md: New file.
2417 * config/microblaze/microblaze.md: Include sync.md
2418 * config/microblaze/microblaze.c: Add print_operand 'y'.
2419 * config/microblaze/constraints.md: Add memory_contraint
2420 'Q' which is a single register.
2422 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2424 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2426 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2428 * reload.c (find_reloads): Exit loop once we find this operand
2429 cannot be reloaded somehow for this alternative.
2431 * reload.c (find_reloads): Exit loop once we find a hard register.
2433 * rtlanal.c (computed_jump_p): Exit loop once we find label
2436 * i386.c (ix86_pad_returns): Exit loop after setting replace.
2438 * cfgloopmanip.c (remove_path): Exit loop after setting
2441 * gensupport.c (subst_dup): Avoid loop if code is not
2442 MATCH_DUP nor MATCH_OP_DUP.
2444 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
2446 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
2448 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2450 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
2451 true for SP_REGNUM if mode == ptr_mode.
2452 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
2453 with value R0_REGNUM + 31.
2455 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2457 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
2458 pad pointer-typed argument downward.
2460 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2462 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
2463 and __ILP32__ when the ILP32 model is in use.
2465 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2467 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
2468 (aarch64_load_symref_appropriately): In the case of
2469 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
2470 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
2471 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
2472 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
2473 if 'mode' doesn't equal to 'ptr_mode'.
2474 (aarch64_output_mi_thunk): Add an assertion on the alignment of
2475 'vcall_offset'; change to call aarch64_emit_move differently depending
2476 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
2477 to calculate the upper bound of 'vcall_offset'.
2478 (aarch64_cannot_force_const_mem): Change to also return true if
2480 (aarch64_legitimize_reload_address): In the case of large
2481 displacements, add new local variable 'xmode' and an assertion
2482 based on it; change to use 'xmode' to generate the new rtx and
2484 (aarch64_asm_trampoline_template): Change to generate the template
2485 differently depending on TARGET_ILP32 or not; change to use
2486 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
2487 (aarch64_trampoline_size): Removed.
2488 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
2489 and replace immediate literals with it. Change to use 'ptr_mode'
2490 instead of 'DImode' and call convert_memory_address if the mode
2491 of 'fnaddr' doesn't equal to 'ptr_mode'.
2492 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
2494 (aarch64_elf_asm_destructor): Likewise.
2495 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
2496 on TARGET_ILP32 instead of aarch64_trampoline_size.
2497 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
2498 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
2499 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
2500 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
2501 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
2502 (storewb_pair<GPI:mode>_<P:mode>): ... this.
2503 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
2504 depending on the value of 'mode'.
2505 (add_losym_<mode>): New.
2506 (ldr_got_small_<mode>): New, based on ldr_got_small.
2507 (ldr_got_small): Remove.
2508 (ldr_got_small_sidi): New.
2509 * config/aarch64/iterators.md (P): New.
2510 (PTR): Change to 'ptr_mode' in the condition.
2512 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2514 * config.gcc (aarch64*-*-*): Support --with-abi.
2515 (aarch64*-*-elf): Support --with-multilib-list.
2516 (aarch64*-*-linux*): Likewise.
2517 (supported_defaults): Add abi to aarch64*-*-*.
2518 * configure.ac: Mention AArch64 for --with-multilib-list.
2519 * configure: Re-generated.
2520 * config/aarch64/biarchilp32.h: New file.
2521 * config/aarch64/biarchlp64.h: New file.
2522 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
2524 (MULTILIB_DEFAULTS): Ditto.
2525 (DRIVER_SELF_SPECS): Ditto.
2526 (ASM_SPEC): Update to also substitute -mabi.
2527 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
2528 file whose name depends on -mabi= and -mbig-endian.
2529 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
2531 (POINTER_SIZE): New define.
2532 (POINTERS_EXTEND_UNSIGNED): Ditto.
2533 (enum aarch64_abi_type): New enumeration tag.
2534 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
2535 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
2536 (TARGET_ILP32): New define.
2537 * config/aarch64/aarch64.opt (mabi): New.
2539 (ilp32, lp64): New values for -mabi.
2540 * config/aarch64/t-aarch64 (comma): New define.
2541 (MULTILIB_OPTIONS): Ditto.
2542 (MULTILIB_DIRNAMES): Ditto.
2543 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
2544 * doc/invoke.texi: Document -mabi for AArch64.
2546 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
2548 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
2550 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2551 Anton Blanchard <anton@au1.ibm.com>
2553 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
2554 endianness when selecting field to splat.
2556 2013-07-22 Eric Christopher <echristo@gmail.com>
2558 * dwarf2out.c (die_odr_checksum): New function to use
2559 CHECKSUM_ macros and ULEB128 for DIE tag.
2560 (generate_type_signature): Use.
2562 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
2564 * config.gcc (sparc*-*-*): Accept leon3 processor.
2565 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
2566 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
2567 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
2568 * config/sparc/sparc.opt (enum processor_type): Add leon3.
2569 (mfix-ut699): Adjust comment.
2570 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
2571 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
2572 (CPP_CPU_SPEC): Likewise.
2573 (ASM_CPU_SPEC): Likewise.
2574 * config/sparc/sparc.c (leon3_cost): New constant.
2575 (sparc_option_override): Add leon3 support.
2576 (mem_ref): New function.
2577 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
2578 (sparc_do_work_around_errata): Look into the instruction in the delay
2579 slot and adjust accordingly. Add fix for the data cache nullify issues
2580 of the UT699. Change insertion position for the NOP.
2581 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
2582 (leon3_load): New reservation.
2583 (leon_store): Bump latency to 2.
2584 (grfpu): New automaton.
2585 (grfpu_alu): New unit.
2586 (grfpu_ds): Likewise.
2587 (leon_fp_alu): Adjust.
2588 (leon_fp_mult): Delete.
2589 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
2590 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
2591 * config/sparc/sparc.md (cpu): Add leon3.
2592 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
2594 (atomic_test_and_set): Likewise.
2597 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
2599 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
2600 default for R5900 targets.
2601 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
2602 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
2603 * config/mips/mips.c (mips_option_override): Report an error for
2604 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
2605 for -march=r5900 -mhard-float.
2607 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2609 * df-problems.c (can_move_insns_across): Exit loop once we
2610 find a non-fixed, non-global register.
2612 * ipa-pure-const.c (propagate_nothrow): Exit loop after
2615 * omega.c (omega_eliminate_red): Break after setting red_found.
2616 (omega_problem_has_red_equations): Similarly after setting found.
2617 (omega_query_variable): Similarly after setting coupled.
2619 2013-07-22 Marek Polacek <polacek@redhat.com>
2621 * gimplify.c: Don't include gimple.h twice.
2623 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2625 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
2626 instead of TARGET_THUMB1.
2627 (Pz): New constraint.
2628 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
2630 (compare_negsi_si): Likewise.
2631 (compare_addsi2_op0): Likewise.
2632 (compare_addsi2_op1): Likewise.
2633 (addsi3_carryin_<optab>): Likewise.
2634 (addsi3_carryin_alt2_<optab>): Likewise.
2635 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
2636 for arm_restrict_it.
2637 (subsi3_carryin): Likewise.
2638 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
2639 (minmax_arithsi): Disable for arm_restrict_it.
2640 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
2641 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
2642 (satsi_<SAT:code>_shift): Likewise.
2643 (arm_shiftsi3): Add alternative for 16-bit encoding.
2644 (arm32_movhf): Disable for arm_restrict_it.
2645 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
2646 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
2648 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
2650 * config/arm/arm.md (attribute "insn"): Delete.
2651 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
2652 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
2653 (not_shiftsi): Update for attribute change.
2654 (not_shiftsi_compare0): Likewise.
2655 (not_shiftsi_compare0_scratch): Likewise.
2656 (arm_one_cmplsi2): Likewise.
2657 (thumb1_one_cmplsi2): Likewise.
2658 (notsi_compare0): Likewise.
2659 (notsi_compare0_scratch): Likewise.
2660 (thumb1_movdi_insn): Likewise.
2661 (arm_movsi_insn): Likewise.
2662 (movhi_insn_arch4): Likewise.
2663 (movhi_bytes): Likewise.
2664 (arm_movqi_insn): Likewise.
2665 (thumb1_movqi_insn): Likewise.
2666 (arm32_movhf): Likewise.
2667 (thumb1_movhf): Likewise.
2668 (arm_movsf_soft_insn): Likewise.
2669 (thumb1_movsf_insn): Likewise.
2670 (thumb_movdf_insn): Likewise.
2671 (movsicc_insn): Likewise.
2672 (movsfcc_soft_insn): Likewise.
2673 (and_scc): Likewise.
2674 (cond_move): Likewise.
2675 (if_move_not): Likewise.
2676 (if_not_move): Likewise.
2677 (if_shift_move): Likewise.
2678 (if_move_shift): Likewise.
2679 (if_shift_shift): Likewise.
2680 (if_not_arith): Likewise.
2681 (if_arith_not): Likewise.
2682 (cond_move_not): Likewise.
2683 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
2684 (neon_mov<mode>): Likewise.
2685 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
2686 (thumb2_movsi_vfp): Likewise.
2687 (movsf_vfp): Likewise.
2688 (thumb2_movsf_vfp): Likewise.
2689 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
2691 (cortexa7_older_only): Likewise.
2692 (cortexa7_younger): Likewise.
2693 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
2694 (1020alu_shift_op): Likewise.
2695 (1020alu_shift_reg_op): Likewise.
2696 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
2697 (alu_shift_op): Likewise.
2698 (alu_shift_reg_op): Likewise.
2699 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
2700 (11_alu_shift_op): Likewise.
2701 (11_alu_shift_reg_op): Likewise.
2702 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
2703 (9_alu_shift_reg_op): Likewise.
2704 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2706 (cortex_a15_alu_shift): Likewise.
2707 (cortex_a15_alu_shift_reg): Likewise.
2708 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
2709 (cortex_a5_alu_shift): Likewise.
2710 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
2712 (cortex_a53_alu_shift): Likewise.
2713 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
2715 (cortex_a7_alu_reg): Likewise.
2716 (cortex_a7_alu_shift): Likewise.
2717 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2718 (cortex_a8_alu_shift): Likewise.
2719 (cortex_a8_alu_shift_reg): Likewise.
2720 (cortex_a8_mov): Likewise.
2721 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2722 (cortex_a9_dp_shift): Likewise.
2723 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
2724 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
2725 (cortex_r4_mov): Likewise.
2726 (cortex_r4_alu_shift): Likewise.
2727 (cortex_r4_alu_shift_reg): Likewise.
2728 * config/arm/fa526.md (526_alu_op): Update for attribute change.
2729 (526_alu_shift_op): Likewise.
2730 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
2731 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
2732 (626te_alu_shift_op): Likewise.
2733 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
2734 (726te_alu_op): Likewise.
2735 (726te_alu_shift_op): Likewise.
2736 (726te_alu_shift_reg_op): Likewise.
2737 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
2738 (mp626_alu_shift_op): Likewise.
2739 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
2740 (pj4_alu_e1_conds): Likewise.
2741 (pj4_alu): Likewise.
2742 (pj4_alu_conds): Likewise.
2743 (pj4_shift): Likewise.
2744 (pj4_shift_conds): Likewise.
2745 (pj4_alu_shift): Likewise.
2746 (pj4_alu_shift_conds): Likewise.
2748 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2750 * config/arm/predicates.md (shiftable_operator_strict_it):
2752 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
2753 Disable cond_exec version for arm_restrict_it.
2754 (thumb2_smaxsi3): Convert to generate cond_exec.
2755 (thumb2_sminsi3): Likewise.
2756 (thumb32_umaxsi3): Likewise.
2757 (thumb2_uminsi3): Likewise.
2758 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
2759 (thumb2_neg_abssi2): Likewise.
2760 (thumb2_mov_scc): Add alternative for 16-bit encoding.
2761 (thumb2_movsicc_insn): Adjust alternatives.
2762 (thumb2_mov_negscc): Disable for arm_restrict_it.
2763 (thumb2_mov_negscc_strict_it): New pattern.
2764 (thumb2_mov_notscc_strict_it): New pattern.
2765 (thumb2_mov_notscc): Disable for arm_restrict_it.
2766 (thumb2_ior_scc): Likewise.
2767 (thumb2_ior_scc_strict_it): New pattern.
2768 (thumb2_cond_move): Adjust for arm_restrict_it.
2769 (thumb2_cond_arith): Disable for arm_restrict_it.
2770 (thumb2_cond_arith_strict_it): New pattern.
2771 (thumb2_cond_sub): Adjust for arm_restrict_it.
2772 (thumb2_movcond): Likewise.
2773 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
2774 (thumb2_zero_extendhisi2_v6): Likewise.
2775 (thumb2_zero_extendqisi2_v6): Likewise.
2776 (orsi_notsi_si): Likewise.
2777 (orsi_not_shiftsi_si): Likewise.
2779 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
2781 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
2782 instruction sequence is 1 byte shorter.
2784 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
2786 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
2787 it is not needed after split.
2789 2013-07-20 Iain Sandoe <iain@codesourcery.com>
2792 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
2793 second label for nonlocal goto receivers. Don't output pic base labels
2794 unless we're producing PIC; mark that action unreachable().
2795 (ix86_save_reg): If the function contains a nonlocal label, save the
2797 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
2798 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
2799 (update_pic_label_number_if_needed): New.
2800 (machopic_output_function_base_name): Adjust for nonlocal receiver
2802 (machopic_should_output_picbase_label): New.
2803 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
2804 (nonlocal_goto_receiver): New insn and split.
2806 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
2808 * config/aarch64/aarch64-builtins.c
2809 (aarch64_fold_builtin): Fold abs in all modes.
2810 * config/aarch64/aarch64-simd-builtins.def
2811 (abs): Enable for all modes.
2812 * config/aarch64/arm_neon.h
2813 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
2814 (vabs_f64): Add missing intrinsic.
2816 2013-07-19 Ian Bolton <ian.bolton@arm.com>
2818 * config/aarch64/arm_neon.h (vabs_s64): New function
2820 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
2823 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
2824 * config/avr/avr.md (adjust_len): Add `round'.
2825 * config/avr/avr-protos.h (avr_out_round): New prototype.
2826 (avr_out_plus): Add `out_label' argument.
2827 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
2828 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
2829 Handle the case where `insn' is just a pattern.
2830 (avr_out_bitop): Handle the case where `insn' is just a pattern.
2831 (avr_out_round): New function.
2832 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
2834 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
2836 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2837 Add check for flag_stack_usage to handle -fstack-usage support
2839 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
2841 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
2842 interaction for new Power8 flags and VSX.
2844 2013-07-18 Sriraman Tallam <tmsriram@google.com>
2847 * tree-inline.c (expand_call_inline): Emit errors during
2848 early_inlining only if optimization is not turned on.
2850 2013-07-18 David Malcolm <dmalcolm@redhat.com>
2854 * passes.c (init_optimization_passes): Move the construction of
2855 the pass hierarchy into a new passes.def file.
2857 * Makefile.in (passes.o): Add dependency on passes.def.
2859 2013-07-18 David Malcolm <dmalcolm@redhat.com>
2861 * passes.c (init_optimization_passes): Introduce macros for
2862 constructing the tree of passes (INSERT_PASSES_AFTER,
2863 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
2864 TERMINATE_PASS_LIST).
2866 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
2867 Wei Mi <wmi@google.com>
2869 PR rtl-optimization/57878
2870 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
2872 (reload_pseudo_compare_func): Check nregs first for reload
2875 2013-07-18 David Malcolm <dmalcolm@redhat.com>
2877 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
2879 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
2881 * read-rtl.c (validate_const_int): Once an invalid character is
2882 seen, quit the loop.
2884 * gengtype.c (write_roots): Similarly once we find the "deletable"
2885 or "if_marked" option.
2887 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2889 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
2890 "xtab" and "sat". Move value "clz" from here to ...
2891 (attriubte "type"): ... here.
2892 (satsi_<SAT:code>): Delete "insn" attribute.
2893 (satsi_<SAT:code>_shift): Likewise.
2894 (arm_zero_extendqisi2addsi): Likewise.
2895 (arm_extendqisi2addsi): Likewise.
2896 (clzsi2): Update for attribute changes.
2897 (rbitsi2): Likewise.
2898 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
2900 (arm_usatsihi): Likewise.
2901 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2903 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2905 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
2906 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
2907 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
2908 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
2909 in alphabetical order.
2910 (attribute "core_cycles"): Update for attribute changes.
2911 (arm_addsi3): Likewise.
2912 (addsi3_compare0): Likewise.
2913 (addsi3_compare0_scratch): Likewise.
2914 (addsi3_compare_op1): Likewise.
2915 (addsi3_compare_op2): Likewise.
2916 (compare_addsi2_op0): Likewise.
2917 (compare_addsi2_op1): Likewise.
2918 (addsi3_carryin_shift_<optab>): Likewise.
2919 (subsi3_carryin_shift): Likewise.
2920 (rsbsi3_carryin_shift): Likewise.
2921 (arm_subsi3_insn): Likewise.
2922 (subsi3_compare0): Likewise.
2923 (subsi3_compare): Likewise.
2924 (arm_andsi3_insn): Likewise.
2925 (thumb1_andsi3_insn): Likewise.
2926 (andsi3_compare0): Likewise.
2927 (andsi3_compare0_scratch): Likewise.
2928 (zeroextractsi_compare0_scratch
2929 (andsi_not_shiftsi_si): Likewise.
2930 (iorsi3_insn): Likewise.
2931 (iorsi3_compare0): Likewise.
2932 (iorsi3_compare0_scratch): Likewise.
2933 (arm_xorsi3): Likewise.
2934 (thumb1_xorsi3_insn): Likewise.
2935 (xorsi3_compare0): Likewise.
2936 (xorsi3_compare0_scratch): Likewise.
2937 (satsi_<SAT:code>_shift): Likewise.
2939 (arm_shiftsi3): Likewise.
2940 (shiftsi3_compare0): Likewise.
2941 (not_shiftsi): Likewise.
2942 (not_shiftsi_compare0): Likewise.
2943 (not_shiftsi_compare0_scratch): Likewise.
2944 (arm_one_cmplsi2): Likewise.
2945 (thumb_one_complsi2): Likewise.
2946 (notsi_compare0): Likewise.
2947 (notsi_compare0_scratch): Likewise.
2948 (thumb1_zero_extendhisi2): Likewise.
2949 (arm_zero_extendhisi2): Likewise.
2950 (arm_zero_extendhisi2_v6): Likewise.
2951 (arm_zero_extendhisi2addsi): Likewise.
2952 (thumb1_zero_extendqisi2): Likewise.
2953 (thumb1_zero_extendqisi2_v6): Likewise.
2954 (arm_zero_extendqisi2): Likewise.
2955 (arm_zero_extendqisi2_v6): Likewise.
2956 (arm_zero_extendqisi2addsi): Likewise.
2957 (thumb1_extendhisi2): Likewise.
2958 (arm_extendhisi2): Likewise.
2959 (arm_extendhisi2_v6): Likewise.
2960 (arm_extendqisi): Likewise.
2961 (arm_extendqisi_v6): Likewise.
2962 (arm_extendqisi2addsi): Likewise.
2963 (thumb1_extendqisi2): Likewise.
2964 (thumb1_movdi_insn): Likewise.
2965 (arm_movsi_insn): Likewise.
2966 (movsi_compare0): Likewise.
2967 (movhi_insn_arch4): Likewise.
2968 (movhi_bytes): Likewise.
2969 (arm_movqi_insn): Likewise.
2970 (thumb1_movqi_insn): Likewise.
2971 (arm32_movhf): Likewise.
2972 (thumb1_movhf): Likewise.
2973 (arm_movsf_soft_insn): Likewise.
2974 (thumb1_movsf_insn): Likewise.
2975 (movdf_soft_insn): Likewise.
2976 (thumb_movdf_insn): Likewise.
2977 (arm_cmpsi_insn): Likewise.
2978 (cmpsi_shiftsi): Likewise.
2979 (cmpsi_shiftsi_swp): Likewise.
2980 (arm_cmpsi_negshiftsi_si): Likewise.
2981 (movsicc_insn): Likewise.
2982 (movsfcc_soft_insn): Likewise.
2983 (arith_shiftsi): Likewise.
2984 (arith_shiftsi_compare0
2985 (arith_shiftsi_compare0_scratch
2986 (sub_shiftsi): Likewise.
2987 (sub_shiftsi_compare0
2988 (sub_shiftsi_compare0_scratch
2989 (and_scc): Likewise.
2990 (cond_move): Likewise.
2991 (if_plus_move): Likewise.
2992 (if_move_plus): Likewise.
2993 (if_move_not): Likewise.
2994 (if_not_move): Likewise.
2995 (if_shift_move): Likewise.
2996 (if_move_shift): Likewise.
2997 (if_shift_shift): Likewise.
2998 (if_not_arith): Likewise.
2999 (if_arith_not): Likewise.
3000 (cond_move_not): Likewise.
3001 (thumb1_ashlsi3): Set type attribute.
3002 (thumb1_ashrsi3): Likewise.
3003 (thumb1_lshrsi3): Likewise.
3004 (thumb1_rotrsi3): Likewise.
3005 (shiftsi3_compare0_scratch): Likewise.
3006 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
3007 (neon_mov<mode>): Likewise.
3008 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
3010 (thumb2_movsi_insn): Likewise.
3011 (thumb2_cmpsi_neg_shiftsi): Likewise.
3012 (thumb2_extendqisi_v6): Likewise.
3013 (thumb2_zero_extendhisi2_v6): Likewise.
3014 (thumb2_zero_extendqisi2_v6): Likewise.
3015 (thumb2_shiftsi3_short): Likewise.
3016 (thumb2_addsi3_compare0_scratch): Likewise.
3017 (orsi_not_shiftsi_si): Likewise.
3018 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3019 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
3021 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3022 (1020alu_shift_op): Likewise.
3023 (1020alu_shift_reg_op): Likewise.
3024 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3025 (alu_shift_op): Likewise.
3026 (alu_shift_reg_op): Likewise.
3027 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3028 (11_alu_shift_op): Likewise.
3029 (11_alu_shift_reg_op): Likewise.
3030 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3031 (9_alu_shift_reg_op): Likewise.
3032 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3034 (cortex_a15_alu_shift): Likewise.
3035 (cortex_a15_alu_shift_reg): Likewise.
3036 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
3038 (cortex_a5_alu_shift): Likewise.
3039 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
3041 (cortex_a53_alu_shift): Likewise.
3042 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3044 (cortex_a7_alu_reg): Likewise.
3045 (cortex_a7_alu_shift): Likewise.
3046 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
3048 (cortex_a8_alu_shift): Likewise.
3049 (cortex_a8_alu_shift_reg): Likewise.
3050 (cortex_a8_mov): Likewise.
3051 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3052 (cortex_a9_dp_shift): Likewise.
3053 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
3055 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
3057 (cortex_r4_mov): Likewise.
3058 (cortex_r4_alu_shift): Likewise.
3059 (cortex_r4_alu_shift_reg): Likewise.
3060 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3061 (526_alu_shift_op): Likewise.
3062 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3063 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3064 (626te_alu_shift_op): Likewise.
3065 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
3066 (726te_alu_op): Likewise.
3067 (726te_alu_shift_op): Likewise.
3068 (726te_alu_shift_reg_op): Likewise.
3069 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3070 (mp626_alu_shift_op): Likewise.
3071 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
3072 (pj4_alu_e1_conds): Likewise.
3073 (pj4_alu): Likewise.
3074 (pj4_alu_conds): Likewise.
3075 (pj4_shift): Likewise.
3076 (pj4_shift_conds): Likewise.
3077 (pj4_alu_shift): Likewise.
3078 (pj4_alu_shift_conds): Likewise.
3079 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3081 (cortexa7_older_only): Likewise.
3082 (cortexa7_younger): Likewise.
3084 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3086 * ipa-pure-const.c (generate_summary): Rename to...
3087 (pure_const_generate_summary): ... this.
3089 2013-07-17 Iain Sandoe <iain@codesourcery.com>
3091 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3093 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
3096 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3099 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3101 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3102 enabled without -march=zEC12.
3103 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3106 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3108 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3109 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3110 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3111 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3112 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3113 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3114 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3116 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3117 and ISA_HAS_NMADD3_NMSUB3.
3118 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3119 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3120 (nmsub4<mode>, nmsub3<mode>): Likewise.
3121 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3123 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3125 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3126 TARGET_MIPS5400 checking.
3128 2013-07-16 Jakub Jelinek <jakub@redhat.com>
3129 Peter Bergner <bergner@vnet.ibm.com>
3131 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3132 registers in the comment.
3133 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3134 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3135 rather than FIRST_PSEUDO_REGISTERS.
3137 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
3139 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3140 enable extra ISA flags with TARGET_HTM.
3142 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3144 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3147 2013-07-15 Cong Hou <congh@google.com>
3149 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3150 in compare function for sorting.
3152 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
3154 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3155 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3156 * config/rs6000/rs6000.opt: Add -mhtm option.
3157 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3158 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3159 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3160 __HTM__ if the HTM instructions are available.
3161 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3162 htm_spr_reg_operand): New define_predicates.
3163 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3164 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3166 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3167 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3168 HTM builtin functions.
3169 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3170 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3171 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3172 (rs6000_builtin_mask_calculate): Likewise.
3173 (rs6000_option_override_internal): Likewise.
3174 (bdesc_htm): Add new HTM builtin support.
3175 (htm_spr_num): New function.
3176 (htm_spr_regno): Likewise.
3177 (rs6000_htm_spr_icode): Likewise.
3178 (htm_expand_builtin): Likewise.
3179 (htm_init_builtins): Likewise.
3180 (rs6000_expand_builtin): Add support for HTM builtin functions.
3181 (rs6000_init_builtins): Likewise.
3182 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3184 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3185 (TARGET_HTM, MASK_HTM): Define macros.
3186 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3187 (FIXED_REGISTERS): Likewise.
3188 (CALL_USED_REGISTERS): Likewise.
3189 (CALL_REALLY_USED_REGISTERS): Likewise.
3190 (REG_ALLOC_ORDER): Likewise.
3191 (enum reg_class): Likewise.
3192 (REG_CLASS_NAMES): Likewise.
3193 (REG_CLASS_CONTENTS): Likewise.
3194 (REGISTER_NAMES): Likewise.
3195 (ADDITIONAL_REGISTER_NAMES): Likewise.
3196 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3197 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3198 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3199 * config/rs6000/htm.md: New file.
3200 * config/rs6000/htmintrin.h: New file.
3201 * config/rs6000/htmxlintrin.h: New file.
3203 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3205 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3206 Define SYMBOL_TINY_GOT, update comment.
3207 * config/aarch64/aarch64.c
3208 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3209 (aarch64_expand_mov_immediate): Likewise.
3210 (aarch64_print_operand): Likewise.
3211 (aarch64_classify_symbol): Likewise.
3212 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3213 (ldr_got_tiny): Define.
3215 2013-07-13 Tobias Grosser <tobias@grosser.es>
3217 PR tree-optimization/54094
3218 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3219 scheduling dimension for the parallelism check from the polyhedral
3220 information in the AST.
3221 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3224 2013-07-13 Jason Merrill <jason@redhat.com>
3226 * print-tree.c (debug_vec_tree): Use debug_raw.
3227 (debug_raw (vec<tree, va_gc> &)): New.
3228 (debug_raw (vec<tree, va_gc> *)): New.
3229 * tree.h: Declare them.
3231 2013-07-13 Bin Cheng <bin.cheng@arm.com>
3233 * ifcvt.c (ifcvt_after_combine): New static variable.
3234 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3236 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3237 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3238 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3240 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3242 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3244 2013-07-12 Michael Matz <matz@suse.de>
3247 * convert.c (convert_to_real): Reject non-float inner types.
3249 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3251 * config/aarch64/aarch64-protos.h
3252 (aarch64_simd_immediate_valid_for_move): Remove.
3253 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3254 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3256 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3258 2013-07-11 Steve Ellcey <sellcey@mips.com>
3260 * config/mips/mips.c (mips_conditional_register_usage): Do not
3261 use t[0-7] registers in MIPS16 mode when optimizing for size.
3263 2013-07-11 Sriraman Tallam <tmsriram@google.com>
3265 * config/i386/i386.c (dispatch_function_versions): Fix array
3266 indexing of function_version_info to match actual_versions.
3268 2013-07-11 Teresa Johnson <tejohnson@google.com>
3270 * vec.h (struct va_gc): Move release out-of-line.
3271 (va_gc::release): Call ggc_free on released vec.
3273 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3275 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3276 Require GOT register as additional operand in UNSPEC.
3277 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3278 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3279 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3280 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3281 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3282 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3283 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3285 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
3288 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3289 name seen by assembler/linker rather if available.
3291 2013-07-11 Andreas Schwab <schwab@suse.de>
3293 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3295 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3297 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
3299 2013-07-10 Joseph Myers <joseph@codesourcery.com>
3301 * doc/tm.texi.in: Move hook documentation to ....
3302 * target.def: ... here.
3304 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3306 * doc/tm.texi: Regenerate.
3308 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3311 * doc/invoke.texi: Document -Wconditionally-supported.
3313 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3316 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3319 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3322 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3323 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3324 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3325 Remove duplicate devices.
3326 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3327 * config/avr/t-multilib: Regenerate.
3328 * config/avr/avr-tables.opt: Regenerate.
3329 * doc/avr-mmcu.texi: Regenerate.
3331 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3334 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3336 2013-07-10 Graham Stott <graham.stott@btinternet.com>
3338 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3339 the cost of MULT when optimizing for size.
3341 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3343 * config/cr16/cr16-protos.h: Don't include target.h.
3345 2013-07-09 Joseph Myers <joseph@codesourcery.com>
3347 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3348 adjust register size for TDmode and TFmode for VSX registers.
3350 2013-07-08 Kai Tietz <ktietz@redhat.com>
3353 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3354 hook_bool_const_tree_true.
3356 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3358 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3359 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3360 * config/s390/s390.md: Define FPR*_REGNUM constants.
3361 Fix FPR2_REGNUM constant (18 -> 17).
3362 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3363 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3364 ("trunc<BFP:mode><DFP_ALL:mode>2")
3365 ("trunc<DFP_ALL:mode><BFP:mode>2")
3366 ("*extend<BFP:mode><DFP_ALL:mode>2")
3367 ("*extend<DFP_ALL:mode><BFP:mode>2")
3368 ("extend<BFP:mode><DFP_ALL:mode>2")
3369 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3372 2013-07-08 Graham Stott <graham.stott@btinternet.com>
3374 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
3376 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3378 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3379 and cfun_fpr_bit_p to cfun_fpr_save_p.
3380 (s390_frame_area, s390_register_info, s390_frame_info)
3381 (s390_emit_prologue, s390_emit_epilogue)
3382 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3384 * config/s390/s390.h: Define *_REGNUM macros for floating point
3387 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3389 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3391 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3393 PR rtl-optimization/57786
3394 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3395 and break out of the loop when it is set to false.
3397 2013-07-08 Jakub Jelinek <jakub@redhat.com>
3400 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3401 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3402 (const_int 63)) 0)).
3403 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3404 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3405 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3407 PR rtl-optimization/57829
3408 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3409 mask bits outside of mode are just sign-extension from mode to HWI.
3411 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3413 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3414 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3415 adjust_address instead of change_address to keep info about alignment.
3416 (emit_strmov): Remove.
3417 (emit_memmov): New function.
3418 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3419 (expand_movmem_epilogue): Likewise and return updated rtx for
3421 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3422 destination and source.
3423 (decide_alignment): Refactor, handle vector_loop.
3424 (ix86_expand_movmem): Likewise.
3425 (ix86_expand_setmem): Likewise.
3426 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3428 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
3430 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3431 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
3433 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3435 * config/i386/sse.md (sse_movlhps): Change alternative 3
3436 of operand 2 to "m".
3438 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3441 * config/i386/sse.md (iptr): New mode attribute.
3442 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
3443 (<sse>_vm<plusminus_insn><mode>3): Ditto.
3444 (<sse>_vmmul<mode>3): Ditto.
3445 (<sse>_vmdiv<mode>3): Ditto.
3446 (sse_vmrcpv4sf2): Ditto.
3447 (<sse>_vmsqrt<mode>2): Ditto.
3448 (sse_vmrsqrtv4sf2): Ditto.
3449 (<sse>_vm<code><mode>3): Ditto.
3450 (avx_vmcmp<mode>3): Ditto.
3451 (<sse>_vmmaskcmp<mode>3): Ditto.
3452 (<sse>_comi): Ditto.
3453 (<sse>_ucomi): Ditto.
3454 (*xop_vmfrcz_<mode>): Ditto.
3455 (*fmai_fmadd_<mode>): Ditto.
3456 (*fmai_fmsub_<mode>): Ditto.
3457 (*fmai_fnmadd_<mode>): Ditto.
3458 (*fmai_fnmsub_<mode>): Ditto.
3459 (*fma4i_vmfmadd_<mode>): Ditto.
3460 (*fma4i_vmfmsub_<mode>): Ditto.
3461 (*fma4i_vmfnmadd_<mode>): Ditto.
3462 (*fma4i_vmfnmsub_<mode>): Ditto.
3463 (*xop_vmfrcz_<mode>): Ditto.
3464 (sse_cvtps2pi): Ditto.
3465 (sse_cvttps2pi): Ditto.
3466 (sse_cvtss2si): Ditto.
3467 (sse_cvtss2si_2): Ditto.
3468 (sse_cvtss2siq_2): Ditto.
3469 (sse_cvttss2si): Ditto.
3470 (sse_cvttss2siq): Ditto.
3471 (sse_cvtsd2si): Ditto.
3472 (sse_cvtsd2si_2): Ditto.
3473 (sse_cvtsd2siq_2): Ditto.
3474 (sse_cvttsd2si): Ditto.
3475 (sse_cvttsd2siq): Ditto.
3476 (sse_cvtsd2ss): Ditto.
3477 (sse_cvtss2sd): Ditto.
3478 (avx2_pbroadcast<mode>): Ditto.
3479 (avx2_pbroadcast<mode>_1): Ditto.
3480 (*avx_vperm_broadcast_v4sf): Ditto.
3482 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
3483 (sse_movlhps): Ditto.
3484 (sse_storehps): Ditto.
3485 (sse_loadhps): Ditto.
3486 (sse_storelps): Ditto.
3487 (sse_loadlps): Ditto.
3488 (*vec_concatv4sf): Ditto.
3489 (*vec_interleave_highv2df): Ditto.
3490 (*vec_interleave_lowv2df): Ditto.
3491 (*vec_extractv2df_1_sse): Ditto.
3492 (*vec_extractv2df_0_sse): Ditto.
3493 (sse2_storelpd): Ditto.
3494 (sse2_loadlpd): Ditto.
3495 (sse2_movsd): Ditto.
3496 (*vec_concatv4si): Ditto.
3497 (vec_concatv2di): Ditto.
3499 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
3500 for Intel asm dialect.
3501 (mmx_punpcklwd): Ditto.
3502 (mmx_punpckldq): Ditto.
3504 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
3505 for intel assembler dialect.
3507 2013-07-06 Jakub Jelinek <jakub@redhat.com>
3510 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
3511 for BUILT_IN_C{LZ,LRSB}*.
3512 * tree.h (CASE_INT_FN): Add FN##IMAX case.
3513 * tree-vrp.c (extract_range_basic): Handle
3514 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
3515 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
3516 fall thru to code calling set_value*.
3517 * builtins.c (expand_builtin): Remove *IMAX cases.
3518 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
3519 if width is bigger than 2*HWI.
3521 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
3523 PR rtl-optimization/55342
3524 * lra-int.h (lra_subreg_reload_pseudos): New.
3525 * lra.c: Add undoing optional reloads to the block diagram.
3526 (lra_subreg_reload_pseudos): New.
3527 (lra_optional_reload_pseudos): Change comments.
3528 (lra): Init and clear lra_subreg_reload_pseudos. Clear
3529 lra_optional_reload_pseudos after undo transformations.
3530 * lra-assigns.c (pseudo_prefix_title): New.
3531 (lra_setup_reg_renumber): Use it.
3532 (spill_for): Ditto. Check subreg reload pseudos too.
3533 (assign_by_spills): Consider subreg reload pseudos too.
3534 * lra-constraints.c (simplify_operand_subreg): Use
3535 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
3536 (curr_insn_transform): Recognize and do optional reloads.
3537 (undo_optional_reloads): New.
3538 (lra_undo_inheritance): Call undo_optional_reloads.
3540 2013-07-05 Thomas Quinot <quinot@adacore.com>
3542 * tree-complex.c (expand_complex_operations_1): Fix typo.
3544 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
3546 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
3547 (tune_params): New member 'const vec_costs'.
3548 * config/aarch64/aarch64.c (generic_vector_cost): New.
3549 (generic_tunings): New member 'generic_vector_cost'.
3550 (aarch64_builtin_vectorization_cost): New.
3551 (aarch64_add_stmt_cost): New.
3552 (TARGET_VECTORIZE_ADD_STMT_COST): New.
3553 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
3555 2013-07-03 Jakub Jelinek <jakub@redhat.com>
3558 * config/i386/predicates.md (vsib_address_operand): Disallow
3559 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
3561 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
3564 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
3565 expand_builtin_setjmp_receiver.
3566 (expand_label): Adjust, call expand_builtin_setjmp_receiver
3567 with NULL for the label parameter.
3568 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
3569 the frame-pointer. Adjust comments.
3570 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
3571 only if LABEL is non-NULL.
3573 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
3575 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
3576 (ARM_ABI_AAPCS64): Ditto.
3578 (ARM_DEFAULT_ABI): Ditto.
3580 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
3582 * config/aarch64/aarch64-builtins.c
3583 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
3584 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
3586 * config/aarch64/aarch64-simd.md
3587 (aarch64_ld1<VALL:mode>): New.
3588 (aarch64_st1<VALL:mode>): Likewise.
3589 * config/aarch64/arm_neon.h
3590 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
3592 2013-07-02 Sriraman Tallam <tmsriram@google.com>
3594 * config/i386/i386.c (gate_insert_vzeroupper): Check if
3596 (ix86_option_override_internal):Turn on all -mavx target flags by
3597 default as they are dependent on AVX anyway.
3599 2013-07-02 Cary Coutant <ccoutant@google.com>
3601 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
3603 (loc_checksum_ordered): Likewise.
3604 (hash_loc_operands): Remove inline keyword.
3606 2013-07-02 Jakub Jelinek <jakub@redhat.com>
3608 PR tree-optimization/57741
3609 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
3610 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
3611 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
3612 Allow REAL_CST step_exprs if flag_associative_math.
3613 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
3615 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3617 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
3619 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3621 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
3623 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3625 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
3627 (iorsi3_insn): Likewise.
3628 (arm_xorsi3): Likewise.
3630 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
3632 * arm.md (attribute "wtype"): Delete. Move attribute values from here
3634 (attribute "type"): ... here, and prefix with "wmmx_".
3635 (attribute "core_cycles"): Update for attribute changes.
3636 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
3637 (tbcstv4hi): Likewise.
3638 (tbcstv2si): Likewise.
3639 (iwmmxt_iordi3): Likewise.
3640 (iwmmxt_xordi3): Likewise.
3641 (iwmmxt_anddi3): Likewise.
3642 (iwmmxt_nanddi3): Likewise.
3643 (iwmmxt_arm_movdi): Likewise.
3644 (iwmmxt_movsi_insn): Likewise.
3645 (mov<mode>_internal): Likewise.
3646 (and<mode>3_iwmmxt): Likewise.
3647 (ior<mode>3_iwmmxt): Likewise.
3648 (xor<mode>3_iwmmxt): Likewise.
3649 (add<mode>3_iwmmxt): Likewise.
3650 (ssaddv8qi3): Likewise.
3651 (ssaddv4hi3): Likewise.
3652 (ssaddv2si3): Likewise.
3653 (usaddv8qi3): Likewise.
3654 (usaddv4hi3): Likewise.
3655 (usaddv2si3): Likewise.
3656 (sub<mode>3_iwmmxt): Likewise.
3657 (sssubv8qi3): Likewise.
3658 (sssubv4hi3): Likewise.
3659 (sssubv2si3): Likewise.
3660 (ussubv8qi3): Likewise.
3661 (ussubv4hi3): Likewise.
3662 (ussubv2si3): Likewise.
3663 (mulv4hi3_iwmmxt): Likewise.
3664 (smulv4hi3_highpart): Likewise.
3665 (umulv4hi3_highpart): Likewise.
3666 (iwmmxt_wmacs): Likewise.
3667 (iwmmxt_wmacsz): Likewise.
3668 (iwmmxt_wmacu): Likewise.
3669 (iwmmxt_wmacuz): Likewise.
3670 (iwmmxt_clrdi): Likewise.
3671 (iwmmxt_clrv8qi): Likewise.
3672 (iwmmxt_clr4hi): Likewise.
3673 (iwmmxt_clr2si): Likewise.
3674 (iwmmxt_uavgrndv8qi3): Likewise.
3675 (iwmmxt_uavgrndv4hi3): Likewise.
3676 (iwmmxt_uavgv8qi3): Likewise.
3677 (iwmmxt_uavgv4hi3): Likewise.
3678 (iwmmxt_tinsrb): Likewise.
3679 (iwmmxt_tinsrh): Likewise.
3680 (iwmmxt_tinsrw): Likewise.
3681 (iwmmxt_textrmub): Likewise.
3682 (iwmmxt_textrmsb): Likewise.
3683 (iwmmxt_textrmuh): Likewise.
3684 (iwmmxt_textrmsh): Likewise.
3685 (iwmmxt_textrmw): Likewise.
3686 (iwmxxt_wshufh): Likewise.
3687 (eqv8qi3): Likewise.
3688 (eqv4hi3): Likewise.
3689 (eqv2si3): Likewise.
3690 (gtuv8qi3): Likewise.
3691 (gtuv4hi3): Likewise.
3692 (gtuv2si3): Likewise.
3693 (gtv8qi3): Likewise.
3694 (gtv4hi3): Likewise.
3695 (gtv2si3): Likewise.
3696 (smax<mode>3_iwmmxt): Likewise.
3697 (umax<mode>3_iwmmxt): Likewise.
3698 (smin<mode>3_iwmmxt): Likewise.
3699 (umin<mode>3_iwmmxt): Likewise.
3700 (iwmmxt_wpackhss): Likewise.
3701 (iwmmxt_wpackwss): Likewise.
3702 (iwmmxt_wpackdss): Likewise.
3703 (iwmmxt_wpackhus): Likewise.
3704 (iwmmxt_wpackwus): Likewise.
3705 (iwmmxt_wpackdus): Likewise.
3706 (iwmmxt_wunpckihb): Likewise.
3707 (iwmmxt_wunpckihh): Likewise.
3708 (iwmmxt_wunpckihw): Likewise.
3709 (iwmmxt_wunpckilb): Likewise.
3710 (iwmmxt_wunpckilh): Likewise.
3711 (iwmmxt_wunpckilw): Likewise.
3712 (iwmmxt_wunpckehub): Likewise.
3713 (iwmmxt_wunpckehuh): Likewise.
3714 (iwmmxt_wunpckehuw): Likewise.
3715 (iwmmxt_wunpckehsb): Likewise.
3716 (iwmmxt_wunpckehsh): Likewise.
3717 (iwmmxt_wunpckehsw): Likewise.
3718 (iwmmxt_wunpckelub): Likewise.
3719 (iwmmxt_wunpckeluh): Likewise.
3720 (iwmmxt_wunpckeluw): Likewise.
3721 (iwmmxt_wunpckelsb): Likewise.
3722 (iwmmxt_wunpckelsh): Likewise.
3723 (iwmmxt_wunpckelsw): Likewise.
3724 (ror<mode>3): Likewise.
3725 (ashr<mode>3_iwmmxt): Likewise.
3726 (lshr<mode>3_iwmmxt): Likewise.
3727 (ashl<mode>3_iwmmxt): Likewise.
3728 (ror<mode>3_di): Likewise.
3729 (ashr<mode>3_di): Likewise.
3730 (lshr<mode>3_di): Likewise.
3731 (ashl<mode>3_di): Likewise.
3732 (iwmmxt_wmadds): Likewise.
3733 (iwmmxt_wmaddu): Likewise.
3734 (iwmmxt_tmia): Likewise.
3735 (iwmmxt_tmiaph): Likewise.
3736 (iwmmxt_tmiabb): Likewise.
3737 (iwmmxt_tmiatb): Likewise.
3738 (iwmmxt_tmiabt): Likewise.
3739 (iwmmxt_tmiatt): Likewise.
3740 (iwmmxt_tmovmskb): Likewise.
3741 (iwmmxt_tmovmskh): Likewise.
3742 (iwmmxt_tmovmskw): Likewise.
3743 (iwmmxt_waccb): Likewise.
3744 (iwmmxt_wacch): Likewise.
3745 (iwmmxt_waccw): Likewise.
3746 (iwmmxt_waligni): Likewise.
3747 (iwmmxt_walignr): Likewise.
3748 (iwmmxt_walignr0): Likewise.
3749 (iwmmxt_walignr1): Likewise.
3750 (iwmmxt_walignr2): Likewise.
3751 (iwmmxt_walignr3): Likewise.
3752 (iwmmxt_wsadb): Likewise.
3753 (iwmmxt_wsadh): Likewise.
3754 (iwmmxt_wsadbz): Likewise.
3755 (iwmmxt_wsadhz): Likewise.
3756 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
3757 (iwmmxt_wabsdiffb): Likewise.
3758 (iwmmxt_wabsdiffh): Likewise.
3759 (iwmmxt_wabsdiffw): Likewise.
3760 (iwmmxt_waddsubhx): Likewise
3761 (iwmmxt_wsubaddhx): Likewise.
3762 (addc<mode>3): Likewise.
3763 (iwmmxt_avg4): Likewise.
3764 (iwmmxt_avg4r): Likewise.
3765 (iwmmxt_wmaddsx): Likewise.
3766 (iwmmxt_wmaddux): Likewise.
3767 (iwmmxt_wmaddsn): Likewise.
3768 (iwmmxt_wmaddun): Likewise.
3769 (iwmmxt_wmulwsm): Likewise.
3770 (iwmmxt_wmulwum): Likewise.
3771 (iwmmxt_wmulsmr): Likewise.
3772 (iwmmxt_wmulumr): Likewise.
3773 (iwmmxt_wmulwsmr): Likewise.
3774 (iwmmxt_wmulwumr): Likewise.
3775 (iwmmxt_wmulwl): Likewise.
3776 (iwmmxt_wqmulm): Likewise.
3777 (iwmmxt_wqmulwm): Likewise.
3778 (iwmmxt_wqmulmr): Likewise.
3779 (iwmmxt_wqmulwmr): Likewise.
3780 (iwmmxt_waddbhusm): Likewise.
3781 (iwmmxt_waddbhusl): Likewise.
3782 (iwmmxt_wqmiabb): Likewise.
3783 (iwmmxt_wqmiabt): Likewise.
3784 (iwmmxt_wqmiatb): Likewise.
3785 (iwmmxt_wqmiatt): Likewise.
3786 (iwmmxt_wqmiabbn): Likewise.
3787 (iwmmxt_wqmiabtn): Likewise.
3788 (iwmmxt_wqmiatbn): Likewise.
3789 (iwmmxt_wqmiattn): Likewise.
3790 (iwmmxt_wmiabb): Likewise.
3791 (iwmmxt_wmiabt): Likewise.
3792 (iwmmxt_wmiatb): Likewise.
3793 (iwmmxt_wmiatt): Likewise.
3794 (iwmmxt_wmiabbn): Likewise.
3795 (iwmmxt_wmiabtn): Likewise.
3796 (iwmmxt_wmiatbn): Likewise.
3797 (iwmmxt_wmiattn): Likewise.
3798 (iwmmxt_wmiawbb): Likewise.
3799 (iwmmxt_wmiawbt): Likewise.
3800 (iwmmxt_wmiawtb): Likewise.
3801 (iwmmxt_wmiawtt): Likewise.
3802 (iwmmxt_wmiawbbn): Likewise.
3803 (iwmmxt_wmiawbtn): Likewise.
3804 (iwmmxt_wmiawtbn): Likewise.
3805 (iwmmxt_wmiawttn): Likewise.
3806 (iwmmxt_wmerge): Likewise.
3807 (iwmmxt_tandc<mode>3): Likewise.
3808 (iwmmxt_torc<mode>3): Likewise.
3809 (iwmmxt_torvsc<mode>3): Likewise.
3810 (iwmmxt_textrc<mode>3): Likewise.
3811 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
3812 (wmmxt_pack): Likewise.
3813 (wmmxt_mult_c1): Likewise.
3814 (wmmxt_mult_c2): Likewise.
3815 (wmmxt_alu_c1): Likewise.
3816 (wmmxt_alu_c2): Likewise.
3817 (wmmxt_alu_c3): Likewise.
3818 (wmmxt_transfer_c1): Likewise.
3819 (wmmxt_transfer_c2): Likewise.
3820 (wmmxt_transfer_c3): Likewise.
3821 (marvell_f_iwmmxt_wstr): Likewise.
3822 (marvell_f_iwmmxt_wldr): Likewise.
3824 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
3826 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
3828 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3831 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3832 * lra-constraints.c (need_for_split_p): Check call used hard regs
3833 living through calls.
3835 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
3836 call used regs for call insn.
3838 2013-06-28 Jakub Jelinek <jakub@redhat.com>
3841 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
3842 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
3844 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
3846 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
3848 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3850 * lra-constraints.c (need_for_split_p): Check call used hard regs
3851 living through calls.
3853 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
3856 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
3857 to tie with any other modes. Eliminate Altivec vector mode tests,
3858 since these are a subset of ALTIVEC or VSX vector modes. Simplify
3859 code, to return 0 if testing MODE2 for a condition, if we've
3860 already tested MODE1 for the same condition.
3862 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3864 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
3867 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3869 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3870 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
3872 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3874 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
3876 (aarch64_symbolic_constant_p): Remove.
3877 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
3878 static. Fix line length and white space.
3879 (aarch64_symbolic_constant_p): Remove.
3880 * config/aarch64/predicates.md (aarch64_valid_symref):
3881 Use aarch64_classify_symbol_expression.
3883 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3885 * config/arm/constraints.md (Ts): New constraint.
3886 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
3888 (compare_scc): Use "Ts" constraint for operand 0.
3889 (ior_scc_scc): Likewise.
3890 (and_scc_scc): Likewise.
3891 (and_scc_scc_nodom): Likewise.
3892 (ior_scc_scc_cmp): Likewise for operand 7.
3893 (and_scc_scc_cmp): Likewise.
3894 * config/arm/thumb2.md (thumb2_movsi_insn):
3895 Add alternatives for 16-bit encodings.
3896 (thumb2_movhi_insn): Likewise.
3897 (thumb2_movsicc_insn): Likewise.
3898 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
3899 (thumb2_negscc): Use "Ts" constraint.
3900 Move mvn instruction outside cond_exec block.
3901 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
3902 for 16-bit encodings.
3904 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3906 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
3908 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
3909 (mulsi3subsi): Likewise.
3910 (mulsidi3adddi): Likewise.
3911 (mulsidi3_v6): Likewise.
3912 (umulsidi3_v6): Likewise.
3913 (umulsidi3adddi_v6): Likewise.
3914 (smulsi3_highpart_v6): Likewise.
3915 (umulsi3_highpart_v6): Likewise.
3916 (mulhisi3tb): Likewise.
3917 (mulhisi3bt): Likewise.
3918 (mulhisi3tt): Likewise.
3919 (maddhisi4): Likewise.
3920 (maddhisi4tb): Likewise.
3921 (maddhisi4tt): Likewise.
3922 (maddhidi4): Likewise.
3923 (maddhidi4tb): Likewise.
3924 (maddhidi4tt): Likewise.
3925 (zeroextractsi_compare0_scratch): Likewise.
3926 (insv_zero): Likewise.
3927 (insv_t2): Likewise.
3928 (anddi_notzesidi_di): Likewise.
3929 (anddi_notsesidi_di): Likewise.
3930 (andsi_notsi_si): Likewise.
3931 (iordi_zesidi_di): Likewise.
3932 (xordi_zesidi_di): Likewise.
3933 (andsi_iorsi3_notsi): Likewise.
3935 (smax_m1): Likewise.
3937 (not_shiftsi): Likewise.
3938 (unaligned_loadsi): Likewise.
3939 (unaligned_loadhis): Likewise.
3940 (unaligned_loadhiu): Likewise.
3941 (unaligned_storesi): Likewise.
3942 (unaligned_storehi): Likewise.
3943 (extv_reg): Likewise.
3944 (extzv_t2): Likewise.
3946 (udivsi3): Likewise.
3947 (arm_zero_extendhisi2addsi): Likewise.
3948 (arm_zero_extendqisi2addsi): Likewise.
3949 (compareqi_eq0): Likewise.
3950 (arm_extendhisi2_v6): Likewise.
3951 (arm_extendqisi2addsi): Likewise.
3952 (arm_movt): Likewise.
3953 (thumb2_ldrd): Likewise.
3954 (thumb2_ldrd_base): Likewise.
3955 (thumb2_ldrd_base_neg): Likewise.
3956 (thumb2_strd): Likewise.
3957 (thumb2_strd_base): Likewise.
3958 (thumb2_strd_base_neg): Likewise.
3959 (arm_negsi2): Add alternative for 16-bit encoding.
3960 (arm_one_cmplsi2): Likewise.
3962 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3964 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
3965 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
3966 (movdfcc): Likewise.
3967 * config/arm/vfp.md (*thumb2_movsf_vfp):
3968 Disable predication for arm_restrict_it.
3969 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
3970 (*thumb2_movdfcc_vfp): Likewise.
3971 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
3972 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
3973 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
3974 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
3975 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
3976 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
3977 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
3978 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
3979 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
3980 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
3981 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
3982 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
3983 Disable predication for arm_restrict_it.
3985 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
3987 * config/i386/bmiintrin.h (_bextr_u32): New.
3988 (_bextr_u64): Ditto.
3990 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
3992 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
3993 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
3994 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
3995 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
3996 * config/mips/n32-elf.h: ...this new file.
3998 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
4001 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
4002 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
4004 2013-06-27 Catherine Moore <clm@codesourcery.com>
4006 * config/mips/mips-tables.opt: Regenerate.
4007 * config/mips/mips-cpus.def: Add m14ke and m14kec.
4008 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
4009 * doc/invoke.texi: Add -m14kc.
4011 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4014 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
4015 constraints of operand 1 and 2.
4018 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
4019 to match RTL canonicalization. Swap predicates and
4020 constraints of operand 1 and 2.
4022 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
4024 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
4025 Process OP_INOUT regs for splitting too.
4027 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4029 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
4030 decl before the loop, initialize to NULL.
4031 (vectorizable_load): Initialize ptr_incr to NULL.
4033 2013-06-27 Martin Jambor <mjambor@suse.cz>
4036 * ipa-ref.h (ipa_maybe_record_reference): Declare.
4037 * ipa-ref.c (ipa_maybe_record_reference): New function.
4038 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
4039 * ipa-cp.c (create_specialized_node): Record potential references from
4041 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
4043 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4045 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
4046 parameter 'mode' of type 'enum machine_mode mode'; change to pass
4047 'mode' to force_reg.
4048 (aarch64_add_offset): Update calls to aarch64_force_temporary.
4049 (aarch64_expand_mov_immediate): Likewise.
4051 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4053 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
4054 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
4056 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4058 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
4059 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4060 (struct machine_function): Add tbegin_p.
4061 (s390_canonicalize_comparison): Fold CC mode compares to
4062 conditional jump if possible.
4063 (s390_emit_jump): Return the emitted jump.
4064 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
4065 Handle CCRAWmode compares.
4066 (s390_option_override): Default to -mhtm if available.
4067 (s390_reg_clobbered_rtx): Handle floating point regs as well.
4068 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
4069 FPRs instead of df_regs_ever_live_p.
4070 (s390_optimize_nonescaping_tx): New function.
4071 (s390_init_frame_layout): Extend clobbered_regs array to cover
4073 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
4074 (s390_expand_tbegin): New function.
4075 (enum s390_builtin): New enum definition.
4076 (code_for_builtin): New array definition.
4077 (s390_init_builtins): New function.
4078 (s390_expand_builtin): New function.
4079 (TARGET_INIT_BUILTINS): Define.
4080 (TARGET_EXPAND_BUILTIN): Define.
4081 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
4082 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
4083 (s390_alc_comparison): Likewise.
4084 * config/s390/s390-modes.def: Add CCRAWmode.
4085 * config/s390/s390.h (processor_flags): Add PF_TX.
4086 (TARGET_CPU_HTM): Define macro.
4087 (TARGET_HTM): Define macro.
4088 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4089 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4090 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4091 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
4092 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
4093 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4094 ("*cc_to_int"): Move up.
4095 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4096 constants other than 0.
4097 ("*ccraw_to_int"): New insn and splitter definition.
4098 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4099 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4100 ("tx_assist"): New expander.
4101 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4102 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4103 * config/s390/s390.opt: Add -mhtm option.
4104 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4105 * config/s390/htmxlintrin.h: New file.
4106 * config/s390/htmintrin.h: New file.
4107 * config/s390/s390intrin.h: New file.
4108 * doc/extend.texi: Document htm builtins.
4109 * config.gcc: Add the new header files to extra_headers.
4111 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4113 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
4114 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
4116 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4117 Pat Haugen <pthaugen@us.ibm.com>
4118 Peter Bergner <bergner@vnet.ibm.com>
4120 * config/rs6000/power8.md: New.
4121 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4122 setting for power8 entry.
4123 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4124 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4125 test for Power4/Power5 only.
4126 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4128 (force_new_group): Adjust comment.
4129 * config/rs6000/rs6000.md: Include power8.md.
4131 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4133 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4134 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4136 (tune_params): Update comment.
4137 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4138 (arm_max_conditional_execute): New function.
4139 (thumb2_final_prescan_insn): Use max_insn_skipped and
4140 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4142 2013-06-25 Jakub Jelinek <jakub@redhat.com>
4144 PR tree-optimization/57705
4145 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4146 SSA_NAME step, provided that it is not defined inside the loop.
4147 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
4148 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4150 2013-06-25 Martin Jambor <mjambor@suse.cz>
4153 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4154 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4156 (ipa_note_param_call): Initialize member_ptr flag.
4157 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4158 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4159 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4160 (ipa_read_indirect_edge_info): Likewise.
4162 2013-06-25 Richard Biener <rguenther@suse.de>
4165 * passes.c (init_optimization_passes): Move pass_fold_builtins
4166 and pass_dce earlier with -Og.
4168 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4170 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4171 <BIT_FIELD_REF>: Remove trailing TAB.
4172 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4175 2013-06-24 Martin Jambor <mjambor@suse.cz>
4177 PR tree-optimization/57358
4178 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4179 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4180 (ipa_analyze_params_uses): Generate pessimistic info when true.
4182 2013-06-24 Martin Jambor <mjambor@suse.cz>
4184 PR tree-optimization/57539
4185 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4186 global.inlined_to of the new node to it. All callers changed.
4187 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4188 inlining_into, pass it to cgraph_clone_node.
4189 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4190 ipa_free_edge_args_substructures.
4191 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4192 rdesc linked list. Do not assert rdesc edges have inlined caller.
4193 Assert we have found an rdesc in the rdesc list.
4195 2013-06-24 Richard Biener <rguenther@suse.de>
4197 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
4198 (pointer_set_lookup): Declare.
4199 (class pointer_map): New template class implementing a
4200 generic pointer to T map.
4201 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4202 pointer_map<T>::contains, pointer_map<T>::insert,
4203 pointer_map<T>::traverse): New functions.
4204 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4205 (pointer_set_lookup): New function.
4206 (pointer_set_contains): Use pointer_set_lookup.
4207 (pointer_set_insert): Likewise.
4208 (insert_aux): Remove.
4209 (struct pointer_map_t): Embed a pointer_set_t.
4210 (pointer_map_create): Adjust.
4211 (pointer_map_destroy): Likewise.
4212 (pointer_map_contains): Likewise.
4213 (pointer_map_insert): Likewise.
4214 (pointer_map_traverse): Likewise.
4215 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4216 pointer_map<unsigned> instead of a pointer_map_t.
4217 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4218 (streamer_tree_cache_lookup): Likewise.
4219 (streamer_tree_cache_create): Likewise.
4220 (streamer_tree_cache_delete): Likewise.
4221 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4222 pointer_map<unsigned> instead of a pointer_map_t.
4223 (lto_init_tree_ref_encoder): Adjust.
4224 (lto_destroy_tree_ref_encoder): Likewise.
4225 * lto-section-out.c (lto_output_decl_index): Likewise.
4226 (lto_record_function_out_decl_state): Likewise.
4227 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4229 2013-06-24 Richard Biener <rguenther@suse.de>
4231 PR tree-optimization/57488
4232 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4234 2013-06-24 Alan Modra <amodra@gmail.com>
4236 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4237 (gen_easy_altivec_constant): Likewise.
4238 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4239 easy_vector_constant_msb): Likewise.
4241 2013-06-23 Jakub Jelinek <jakub@redhat.com>
4244 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4245 add missing return true.
4247 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4250 * config/sh/predicates.md (general_extend_operand): Invoke
4251 general_movsrc_operand for memory operands.
4252 (general_movsrc_operand): Allow reg+reg addressing, do not use
4253 general_operand for memory operands.
4255 2013-06-23 Sriraman Tallam <tmsriram@google.com>
4257 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4258 when current target options does not apply.
4259 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4260 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
4261 * config/i386/bmiintrin.h: Pass appropriate target
4262 attributes to header.
4263 * config/i386/mmintrin.h: Ditto.
4264 * config/i386/nmmintrin.h: Ditto.
4265 * config/i386/avx2intrin.h: Ditto.
4266 * config/i386/fxsrintrin.h: Ditto.
4267 * config/i386/tbmintrin.h: Ditto.
4268 * config/i386/xsaveintrin.h: Ditto.
4269 * config/i386/f16cintrin.h: Ditto.
4270 * config/i386/xtestintrin.h: Ditto.
4271 * config/i386/xsaveoptintrin.h: Ditto.
4272 * config/i386/bmi2intrin.h: Ditto.
4273 * config/i386/lzcntintrin.h: Ditto.
4274 * config/i386/smmintrin.h: Ditto.
4275 * config/i386/wmmintrin.h: Ditto.
4276 * config/i386/x86intrin.h: Remove all header include guards.
4277 * config/i386/prfchwintrin.h: Ditto.
4278 * config/i386/pmmintrin.h: Ditto.
4279 * config/i386/tmmintrin.h: Ditto.
4280 * config/i386/xmmintrin.h: Ditto.
4281 * config/i386/popcntintrin.h: Ditto.
4282 * config/i386/rdseedintrin.h: Ditto.
4283 * config/i386/ammintrin.h: Ditto.
4284 * config/i386/emmintrin.h: Ditto.
4285 * config/i386/immintrin.h: Remove all header include guards.
4286 * config/i386/fma4intrin.h: Ditto.
4287 * config/i386/lwpintrin.h: Ditto.
4288 * config/i386/xopintrin.h: Ditto.
4289 * config/i386/ia32intrin.h: Ditto.
4290 * config/i386/avxintrin.h: Ditto.
4291 * config/i386/rtmintrin.h: Ditto.
4292 * config/i386/fmaintrin.h: Ditto.
4293 * config/i386/mm3dnow.h: Ditto.
4295 2013-06-22 Sriraman Tallam <tmsriram@google.com>
4297 * common/config/i386/i386-common.c: Handle LZCNT.
4299 2013-06-22 Andi Kleen <ak@linux.intel.com>
4301 * doc/extend.texi: Use __atomic_store_n instead of
4302 __atomic_store in HLE example.
4304 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4306 * config/sh/sh.c: Remove <cstdlib> workaround.
4308 2013-06-21 Andi Kleen <ak@linux.intel.com>
4310 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
4312 2013-06-21 Andi Kleen <ak@linux.intel.com>
4314 * doc/extend.texi: Document that __atomic_clear and
4315 __atomic_test_and_set should only be used with bool.
4317 2013-06-20 Jan Hubicka <jh@suse.cz>
4319 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4321 * tree.c (decls_same_for_odr): New function.
4322 (same_for_edr): New function.
4323 (types_same_for_odr): New function.
4324 (get_binfo_at_offset): Use it.
4325 * tree.h (types_same_for_odr): Declare.
4327 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
4328 Jason Merrill <jason@redhat.com>
4330 * system.h: Include <cstdlib> as well as <stdlib.h>.
4332 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
4335 * config/i386/i386.c (construct_container): Report error if
4336 long double is used with disabled x87 float returns.
4338 2013-06-20 Jan Hubicka <jh@suse.cz>
4340 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4342 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4344 PR rtl-optimization/57425
4345 PR rtl-optimization/57569
4346 * alias.c (write_dependence_p): Remove parameters mem_mode and
4347 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4348 Changed all callers.
4349 (canon_anti_dependence): Get comments and semantics in sync.
4350 Add parameter mem_canonicalized. Changed all callers.
4351 * rtl.h (canon_anti_dependence): Update prototype.
4353 2013-06-20 Richard Biener <rguenther@suse.de>
4355 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4356 byte case, inline streamer_read_uchar and defer section
4359 2013-06-20 Richard Biener <rguenther@suse.de>
4361 PR tree-optimization/57584
4362 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4363 SSA names into the expanded expression that take part in
4364 abnormal coalescing.
4366 2013-06-19 Sharad Singhai <singhai@google.com>
4368 * gcov.c (print_usage): Handle new option.
4369 (process_args): Ditto.
4370 (get_gcov_intermediate_filename): New function.
4371 (output_intermediate_file): New function.
4372 (output_gcov_file): New function
4373 (generate_results): Handle new option.
4374 (release_function): Relase demangled name.
4375 (read_graph_file): Handle demangled name.
4376 (output_lines): Ditto.
4377 * doc/gcov.texi: Document gcov intermediate format.
4379 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4382 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4383 (lra_emit_add): Use the functions. Add comment about Y as an
4386 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
4389 * collect2.c (collect_atexit): New.
4390 (collect_exit): Delete.
4391 (main): Register collect_atexit with atexit.
4392 (collect_wait): Change collect_exit to exit.
4394 * collect2.h (collect_exit): Delete.
4395 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4397 2013-06-19 Wei Mi <wmi@google.com>
4399 PR rtl-optimization/57518
4400 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4401 if regno is used in paradoxical subreg.
4402 (update_equiv_regs): Check pdx_subregs[regno] before
4403 set a reg to be equivalent with a mem.
4405 2013-06-19 Matthias Klose <doko@ubuntu.com>
4408 * file-find.h (find_a_file): Add a mode parameter.
4409 * file-find.c (find_a_file): Likewise.
4410 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4411 with X_OK for the executables.
4412 * collect2.c (main): Call find_a_file with X_OK.
4414 2013-06-19 Steve Ellcey <sellcey@mips.com>
4417 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4418 Use NEXT_INSN instead of next_real_insn.
4420 2013-06-19 Jan Hubicka <jh@suse.cz>
4422 * cgraph.h (const_value_known_p): Replace by ...
4423 (ctor_for_folding): .. this one.
4424 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4425 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4426 * expr.c (expand_expr_real_1): Likewise.
4427 (string_constant): Likewise.
4428 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4429 * ipa.c (process_references): Likewise.
4430 (symtab_remove_unreachable_nodes): Likewise.
4431 * ipa-inline-analysis.c (param_change_prob): Likewise.
4432 * gimple-fold.c (canonicalize_constructor_val): Likewise.
4433 (get_base_constructor): Likwise.
4434 * varpool.c (varpool_remove_node): Likewise.
4435 (varpool_remove_initializer): LIkewise.
4436 (dump_varpool_node): LIkwise.
4437 (const_value_known_p): Rewrite to ...
4438 (ctor_for_folding): ... this one.
4440 2013-06-19 Jakub Jelinek <jakub@redhat.com>
4443 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
4444 PERSONALITY in $PATH derived prefixes.
4446 2013-06-19 Jeff Law <law@redhat.com>
4448 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
4451 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
4452 (simplify_bitwise_binary): Use it to simpify certain binary ops on
4455 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
4457 * config/arm/vfp.md: Move VFP instruction classification documentation
4459 * config/arm/arm.md: ... here. Update instruction classification
4462 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
4464 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
4465 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
4466 pattern. Use more efficient sequences on ARMv5 and Thumb2.
4468 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
4471 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4472 with NEXT_INSN. Use tablejump_p to check for jump table data
4475 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
4478 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
4479 that now in C++ the value is correct per the C++ standards.
4481 2013-06-19 Richard Biener <rguenther@suse.de>
4483 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
4486 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4489 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4492 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4493 with next_active_insn.
4495 2013-06-18 Sriraman Tallam <tmsriram@google.com>
4497 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
4498 functions are inlined during failures to flag an error.
4499 * tree-inline.c (expand_call_inline): Allow the error to be flagged
4500 in early inline pass.
4502 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
4504 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
4507 2013-06-18 Julian Brown <julian@codesourcery.com>
4509 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
4510 Permit virtual register pre-reload if !strict.
4511 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
4513 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
4515 * config/arm/neon.md (movmisalign<mode>): Use
4516 neon_perm_struct_or_reg_operand instead of
4517 neon_struct_or_register_operand.
4518 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
4519 neon_permissive_struct_operand instead of neon_struct_operand.
4520 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
4521 neon_vector_mem_operand.
4522 * config/arm/predicates.md (neon_struct_operand): Adjust call to
4523 neon_vector_mem_operand.
4524 (neon_permissive_struct_operand): New.
4525 (neon_struct_or_register_operand): Rename to...
4526 (neon_perm_struct_or_reg_operand): This. Adjust call to
4527 neon_vector_mem_operand.
4529 2013-06-18 Richard Biener <rguenther@suse.de>
4531 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
4532 * lto-streamer.h: Include pointer-set.h.
4533 (struct lto_decl_slot): Remove.
4534 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
4535 Remove next_index entry.
4536 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4537 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4538 (lto_init_tree_ref_encoder): Adjust.
4539 (lto_destroy_tree_ref_encoder): Likewise.
4540 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4541 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4542 (lto_output_decl_index): Adjust.
4543 (lto_new_out_decl_state): Likewise.
4544 (lto_record_function_out_decl_state): Likewise.
4545 * lto-streamer-out.c (copy_function): Likewise.
4547 2013-06-18 Richard Biener <rguenther@suse.de>
4549 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
4550 * cgraphunit.c: Include cfgloop.h.
4551 (init_lowered_empty_function): Initialize the loop tree.
4552 (assemble_thunk): Insert new BBs into loops.
4554 2013-06-18 Richard Biener <rguenther@suse.de>
4556 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
4557 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
4558 the map from cache entry to cache index optional.
4559 (streamer_tree_cache_replace_tree): Adjust accordingly.
4560 (streamer_tree_cache_append): Likewise.
4561 (streamer_tree_cache_delete): Likewise.
4562 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
4563 streamer cache map from cache entry to cache index.
4564 * lto-streamer-out.c (create_output_block): Adjust.
4566 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
4568 * config/arm/arm.md (attribute "insn"): Move multiplication and
4569 division attributes to...
4570 (attribute "type"): ... here. Remove mult.
4571 (attribute "mul32"): New attribute.
4572 (attribute "mul64"): Add umaal.
4573 (*arm_mulsi3): Update attributes.
4574 (*arm_mulsi3_v6): Likewise.
4575 (*thumb_mulsi3): Likewise.
4576 (*thumb_mulsi3_v6): Likewise.
4577 (*mulsi3_compare0): Likewise.
4578 (*mulsi3_compare0_v6): Likewise.
4579 (*mulsi_compare0_scratch): Likewise.
4580 (*mulsi_compare0_scratch_v6): Likewise.
4581 (*mulsi3addsi): Likewise.
4582 (*mulsi3addsi_v6): Likewise.
4583 (*mulsi3addsi_compare0): Likewise.
4584 (*mulsi3addsi_compare0_v6): Likewise.
4585 (*mulsi3addsi_compare0_scratch): Likewise.
4586 (*mulsi3addsi_compare0_scratch_v6): Likewise.
4587 (*mulsi3subsi): Likewise.
4588 (*mulsidi3adddi): Likewise.
4589 (*mulsi3addsi_v6): Likewise.
4590 (*mulsidi3adddi_v6): Likewise.
4591 (*mulsidi3_nov6): Likewise.
4592 (*mulsidi3_v6): Likewise.
4593 (*umulsidi3_nov6): Likewise.
4594 (*umulsidi3_v6): Likewise.
4595 (*umulsidi3adddi): Likewise.
4596 (*umulsidi3adddi_v6): Likewise.
4597 (*smulsi3_highpart_nov6): Likewise.
4598 (*smulsi3_highpart_v6): Likewise.
4599 (*umulsi3_highpart_nov6): Likewise.
4600 (*umulsi3_highpart_v6): Likewise.
4601 (mulhisi3): Likewise.
4602 (*mulhisi3tb): Likewise.
4603 (*mulhisi3bt): Likewise.
4604 (*mulhisi3tt): Likewise.
4605 (maddhisi4): Likewise.
4606 (*maddhisi4tb): Likewise.
4607 (*maddhisi4tt): Likewise.
4608 (maddhidi4): Likewise.
4609 (*maddhidi4tb): Likewise.
4610 (*maddhidi4tt): Likewise.
4612 (udivsi3): Likewise.
4613 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
4614 (thumb2_mulsi_short_compare0): Likewise.
4615 (thumb2_mulsi_short_compare0_scratch): Likewise.
4616 * config/arm/arm1020e.md (1020mult1): Update attribute change.
4617 (1020mult2): Likewise.
4618 (1020mult3): Likewise.
4619 (1020mult4): Likewise.
4620 (1020mult5): Likewise.
4621 (1020mult6): Likewise.
4622 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
4624 (cortex_a15_mult64): Likewise.
4625 (cortex_a15_sdiv): Likewise.
4626 (cortex_a15_udiv): Likewise.
4627 * config/arm/arm1026ejs.md (mult1): Update attribute change.
4633 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
4634 (pj4_ir_div): Likewise.
4635 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
4636 (11_mult2): Likewise.
4637 (11_mult3): Likewise.
4638 (11_mult4): Likewise.
4639 (11_mult5): Likewise.
4640 (11_mult6): Likewise.
4641 (11_mult7): Likewise.
4642 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
4643 (cortex_a8_mla): Likewise.
4644 (cortex_a8_mull): Likewise.
4645 (cortex_a8_smulwy): Likewise.
4646 (cortex_a8_smlald): Likewise.
4647 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
4648 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
4649 (cortex_r4_mul_3): Likewise.
4650 (cortex_r4_mla_4): Likewise.
4651 (cortex_r4_mla_3): Likewise.
4652 (cortex_r4_smlald): Likewise.
4653 (cortex_r4_mull): Likewise.
4654 (cortex_r4_sdiv): Likewise.
4655 (cortex_r4_udiv): Likewise.
4656 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
4657 (cortex_a7_idiv): Likewise.
4658 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
4659 (9_mult2): Likewise.
4660 (9_mult3): Likewise.
4661 (9_mult4): Likewise.
4662 (9_mult5): Likewise.
4663 (9_mult6): Likewise.
4664 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
4665 (cortex_a53_sdiv): Likewise.
4666 (cortex_a53_udiv): Likewise.
4667 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
4668 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
4669 (mp626_mult2): Likewise.
4670 (mp626_mult3): Likewise.
4671 (mp626_mult4): Likewise.
4672 * config/arm/fa526.md (526_mult1): Update attribute change.
4673 (526_mult2): Likewise.
4674 * config/arm/arm-generic.md (mult): Update attribute change.
4675 (mult_ldsched_strongarm): Likewise.
4676 (mult_ldsched): Likewise.
4677 (multi_cycle): Likewise.
4678 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
4679 * config/arm/fa606te.md (606te_mult1): Update attribute change.
4680 (606te_mult2): Likewise.
4681 (606te_mult3): Likewise.
4682 (606te_mult4): Likewise.
4683 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
4684 (cortex_a9_mac16): Likewise.
4685 (cortex_a9_multiply): Likewise.
4686 (cortex_a9_mac): Likewise.
4687 (cortex_a9_multiply_long): Likewise.
4688 * config/arm/fa626te.md (626te_mult1): Update attribute change.
4689 (626te_mult2): Likewise.
4690 (626te_mult3): Likewise.
4691 (626te_mult4): Likewise.
4693 2013-06-18 Richard Biener <rguenther@suse.de>
4696 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
4698 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4701 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4702 with next_active_insn.
4704 2013-06-18 Alan Modra <amodra@gmail.com>
4706 * config/rs6000/rs6000.h (enum data_align): New.
4707 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
4708 (DATA_ABI_ALIGNMENT): Define.
4709 (CONSTANT_ALIGNMENT): Correct comment.
4710 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
4711 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
4713 2013-06-17 David Malcolm <dmalcolm@redhat.com>
4715 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
4716 ATTRIBUTE_UNUSED marking.
4718 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
4720 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
4721 alternative and update.
4722 (aarch64_dup_lanedi): Delete.
4723 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
4724 * config/aarch64/aarch64-simd-builtins.def: Update.
4726 2013-06-17 Richard Biener <rguenther@suse.de>
4728 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
4729 (lto_input_scc): Declare.
4730 (lto_input_tree_1): Likewise.
4731 (struct lto_stats_d): Add num_tree_bodies_output and
4732 num_pickle_refs_output.
4733 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
4734 (lto_read_tree_1): Split out from ...
4735 (lto_read_tree): ... this.
4736 (lto_input_scc): New function.
4737 (lto_input_tree_1): Split out from ...
4738 (lto_input_tree): ... this. Handle LTO_tree_scc.
4739 (lto_data_in_create): Create the streamer cache without hashes.
4740 * lto-streamer-out.c (create_output_block): Create the streamer
4741 cache with hashes when not doing WPA.
4742 (lto_write_tree_1): Split out from ...
4743 (lto_write_tree): ... this.
4744 (get_symbol_initial_value): New function.
4745 (lto_output_tree_1): Split out from ...
4746 (lto_output_tree): ... this. Write trees as series of SCCs
4747 using a DFS walk via DFS_write_tree.
4748 (struct sccs, struct scc_entry): New types.
4749 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
4750 (DFS_write_tree_body): New function.
4751 (DFS_write_tree): Likewise.
4752 (hash_tree): Likewise.
4753 (scc_entry_compare): Likewise.
4754 (hash_scc): Likewise.
4755 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
4756 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
4757 TREE_CHAIN as regular reference.
4758 (streamer_read_integer_cst): Remove.
4759 (streamer_get_pickled_tree): Adjust.
4760 * tree-streamer-out.c (streamer_write_chain): Disable streaming
4761 of DECL_EXTERNALs in BLOCK_VARS for now.
4762 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
4764 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
4765 Add hash value argument and record that if hashes are recorded
4767 (streamer_tree_cache_insert_1): Adjust.
4768 (streamer_tree_cache_insert): Likewise.
4769 (streamer_tree_cache_insert_at): Rename to ...
4770 (streamer_tree_cache_replace_tree): ... this and adjust.
4771 (streamer_tree_cache_append): Adjust.
4772 (record_common_node): Likewise.
4773 (streamer_tree_cache_create): Add argument whether to
4774 record hash values together with trees.
4775 (streamer_tree_cache_delete): Adjust.
4776 * tree-streamer.h (struct streamer_tree_cache_d): Add
4778 (streamer_read_integer_cst): Remove.
4779 (streamer_tree_cache_insert): Adjust.
4780 (streamer_tree_cache_append): Likewise.
4781 (streamer_tree_cache_insert_at): Rename to ...
4782 (streamer_tree_cache_replace_tree): ... this and adjust.
4783 (streamer_tree_cache_create): Add argument whether to record hashes.
4784 (streamer_tree_cache_get): Rename to ...
4785 (streamer_tree_cache_get_tree): ... this.
4786 (streamer_tree_cache_get_hash): New function.
4787 * tree.c (cache_integer_cst): New function.
4788 * tree.h (cache_integer_cst): Declare.
4789 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
4790 * lto-symtab.c (lto_varpool_replace_node): Only release
4791 DECL_INITIAL of non-prevailing decls.
4792 * varpool.c (varpool_remove_initializer): Do not release
4793 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
4795 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4797 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
4798 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
4799 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
4800 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
4801 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
4802 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
4803 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
4804 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
4805 instead of TARGET_64BIT.
4806 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
4807 Require ISA_HAS_<D>DIV.
4809 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
4811 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
4812 (mips*-*-linux*): Move default with_llsc setting to where other
4814 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
4816 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
4817 Likewise. Remove default with_tune setting. Move default float
4818 setting to its own block. Handle with_llsc in the same block as above.
4820 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
4822 PR rtl-optimization/57425
4823 PR rtl-optimization/57569
4824 * alias.c (write_dependence_p): Add new parameters mem_mode,
4825 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
4826 Changed all callers.
4827 (canon_anti_dependence): New function.
4828 * cse.c (check_dependence): Use canon_anti_dependence.
4829 * cselib.c (cselib_invalidate_mem): Likewise.
4830 * rtl.h (canon_anti_dependence): Declare.
4832 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4834 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
4835 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
4836 ".set mips3" for 64-bit targets.
4838 2013-06-15 Dehao Chen <dehao@google.com>
4840 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
4841 * gimple-low.c (gimple_check_call_matching_types): Likewise.
4842 (gimple_check_call_args): Likewise.
4843 * value-prof.c (check_ic_target): Likewise.
4844 * ipa-inline.c (early_inliner): Likewise.
4845 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
4846 * cgraph.c (cgraph_create_edge_1): Likewise.
4847 (cgraph_make_edge_direct): Likewise.
4849 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
4852 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
4853 rs6000_output_move_128bit to handle emitting quad memory
4854 operations. Set attribute length to 8 bytes.
4856 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
4858 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
4860 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
4861 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
4862 (aarch64_<su>mlsl<mode>): Likewise.
4864 2013-06-14 Mike Stump <mikestump@comcast.net>
4866 * Makefile.in (TARGET_H): Add insn-codes.h.
4868 2013-06-14 Alan Modra <amodra@gmail.com>
4872 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
4873 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
4874 bitfield expansion when EXPAND_MEMORY.
4875 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
4877 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
4879 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
4880 test for clearing quad memory on 32-bit later.
4882 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
4884 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
4885 (fold_negate_expr): Likewise.
4886 (fold_real_zero_addition_p): Handle vectors.
4887 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
4889 2013-06-14 Alan Modra <amodra@gmail.com>
4891 * varasm.c (force_const_mem): Revert 2013-06-07 change.
4893 2013-06-13 Jan Hubicka <jh@suse.cz>
4895 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
4896 Local comdats are not externally visible.
4897 * symtab.c (dump_symtab_base): Dump externally visible.
4898 (verify_symtab_base): Verify back links in the symtab hash.
4900 2013-06-13 Bin Cheng <bin.cheng@arm.com>
4902 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
4903 CONVERT_EXPR as equal nodes.
4905 2013-06-13 Bin Cheng <bin.cheng@arm.com>
4907 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
4909 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
4911 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
4912 Generalize to complex and vector.
4913 * tree.c (build_all_ones_cst): New function.
4914 * tree.h (build_all_ones_cst): Declare it.
4916 2013-06-13 Alan Modra <amodra@gmail.com>
4918 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
4919 * config/rs6000/rs6000.md (signbittf2): New insn.
4920 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
4921 (abstf2_internal, cmptf_internal2): Likewise.
4922 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
4924 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4925 Pat Haugen <pthaugen@us.ibm.com>
4926 Peter Bergner <bergner@vnet.ibm.com>
4928 * config/rs6000/rs6000.c (emit_load_locked): Add support for
4929 power8 byte, half-word, and quad-word atomic instructions.
4930 (emit_store_conditional): Likewise.
4931 (rs6000_expand_atomic_compare_and_swap): Likewise.
4932 (rs6000_expand_atomic_op): Likewise.
4934 * config/rs6000/sync.md (larx): Add new modes for power8.
4936 (AINT): New mode iterator to include TImode as well as normal
4937 integer modes on power8.
4938 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
4939 that VSX registers are not considered. Use AINT mode iterator
4940 instead of INT1 to allow inclusion of quad word atomic operations
4942 (load_locked<mode>): Likewise.
4943 (store_conditional<mode>): Likewise.
4944 (atomic_compare_and_swap<mode>): Likewise.
4945 (atomic_exchange<mode>): Likewise.
4946 (atomic_nand<mode>): Likewise.
4947 (atomic_fetch_<fetchop_name><mode>): Likewise.
4948 (atomic_nand_fetch<mode>): Likewise.
4949 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
4951 (ATOMIC): On power8, add QImode, HImode modes.
4952 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
4953 modes that promote to SImode.
4954 (load_lockedti): Convert TImode arguments to PTImode, so that we
4955 get a guaranteed even/odd register pair.
4956 (load_lockedpti): Likewise.
4957 (store_conditionalti): Likewise.
4958 (store_conditionalpti): Likewise.
4960 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
4961 atomic load/store instructions.
4964 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
4966 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
4968 (insn_count): New attribute, with most cases extracted from...
4969 (length): ...here. Redefine most cases in terms of insn_count.
4970 (single_insn): Delete.
4971 (can_delay): Use insn_count to check for single instructions.
4972 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
4973 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
4974 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
4975 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
4976 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
4977 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
4978 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
4979 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
4980 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
4981 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
4982 rather than "length".
4983 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
4984 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
4985 Use "insn_count" rather than "length".
4986 * config/mips/mips-dsp.md
4987 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
4988 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
4991 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
4993 PR tree-optimization/57361
4994 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
4996 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
4998 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
5000 (aarch64_simd_combine<mode>): New instruction expansion.
5001 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
5003 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
5004 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
5006 2013-06-12 Jan Hubicka <jh@suse.cz>
5008 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
5009 decl has when in streaming stage.
5010 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
5011 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
5013 2013-06-12 Roland Stigge <stigge@antcom.de>
5016 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
5018 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5020 PR tree-optimization/57537
5021 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
5022 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
5024 2013-06-12 Richard Biener <rguenther@suse.de>
5026 * data-streamer.h (streamer_write_char_stream): CSE
5027 obs->current_pointer.
5028 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
5029 streamer_write_char_stream manually and optimize the resulting loop.
5030 (streamer_write_hwi_stream): Likewise.
5032 2013-06-12 Jan Hubicka <jh@suse.cz>
5034 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
5035 * cgraph.h (varpool_create_empty_node): Declare.
5036 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
5038 * symtab.c (symtab_unregister_node): Be lax about missin entries
5040 (symtab_get_node): Update comment.
5041 * varpool.c (varpool_create_empty_node): Break out from ...
5042 (varpool_node_for_decl): ... here.
5043 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
5045 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
5047 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
5048 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
5049 part. Use straight-line flow at the end.
5050 <COMPONENT_REF>: Remove superfluous else.
5051 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
5053 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5056 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
5057 target hook even for !TREE_PUBLIC decls. If no resolution info
5058 is available, return false for common and external decls.
5060 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5062 * config/rl78/constraints.md (U): New constraint.
5063 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
5066 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5069 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
5070 to allow returning address to AT_PLATFORM name.
5072 2013-06-11 Jan Hubicka <jh@suse.cz>
5074 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
5075 * cgraph.h (symtab_node_base): Add weakref flag.
5076 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
5077 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
5078 (output_weakrefs): Use weakref flag.
5079 * fold-const.c (simple_operand_p): Handle WEAK.
5080 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
5081 * ipa.c (varpool_externally_visible_p): Drop weakref.
5082 (function_and_variable_visibility): Update comment; fix weakref
5083 sanity checks; do not clear DECL_WEAK on them.
5084 * lto-cgraph.c (lto_output_node): update.
5085 (lto_output_varpool_node): Update.
5086 (input_overwrite_node): Update.
5087 (input_node): Update.
5088 (input_varpool_node): Update.
5089 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5090 (lto_symtab_merge_symbols): Add sanity check.
5091 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5092 * passes.c (rest_of_decl_compilation): Set static flag, too.
5093 * symtab.c (dump_symtab_base): Dump weakref.
5094 (verify_symtab_base): Sanity check weakrefs.
5095 (symtab_make_decl_local): Remove duplicated code.
5096 (symtab_alias_ultimate_target): Simplify.
5097 * varpool.c (varpool_create_variable_alias): Set weakref flag.
5099 2013-06-11 DJ Delorie <dj@redhat.com>
5101 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5102 (rl78_unwind_word_mode): New.
5104 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5106 * final.c (debug_prefix_maps): Make static.
5108 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5110 * function.c (initial_trampoline): Remove stray copy.
5112 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5114 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5116 2013-06-11 Martin Jambor <mjambor@suse.cz>
5118 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5119 within bounds at the beginning of the function.
5121 2013-06-11 Alan Modra <amodra@gmail.com>
5123 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5125 (get_named_section): Don't NULL !DECL_P decl.
5127 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5129 * doc/invoke.texi (core-avx2): Document.
5131 (atom): Updated with MOVBE.
5133 2013-06-11 Richard Biener <rguenther@suse.de>
5135 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
5137 2013-06-11 Anton Blanchard <anton@samba.org>
5139 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5140 correct shift value in little-endian mode.
5142 2013-06-11 Jakub Jelinek <jakub@redhat.com>
5145 * varasm.c (get_variable_align): Move #endif to the right place.
5147 2013-06-10 Cary Coutant <ccoutant@google.com>
5149 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5150 for hash so that hash table traversal order is deterministic.
5152 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5153 Pat Haugen <pthaugen@us.ibm.com>
5154 Peter Bergner <bergner@vnet.ibm.com>
5156 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5157 of vectors in GPRS if they are direct moves or quad word load or
5160 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5162 (direct_move_p): Likewise.
5163 (quad_load_store_p): Likewise.
5165 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5166 classes into bins based on the physical register type.
5167 (reg_class_to_reg_type): Likewise.
5168 (IS_STD_REG_TYPE): Likewise.
5169 (IS_FP_VECT_REG_TYPE): Likewise.
5170 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5171 use direct move instructions.
5172 (reload_gpr_vsx): Likewise.
5173 (reload_vsx_gpr): Likewise.
5174 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5175 information that is a simplification of register classes. Also
5176 precalculate direct move reload helpers.
5177 (direct_move_p): New function to return true if the operation can
5178 be done as a direct move instruciton.
5179 (quad_load_store_p): New function to return true if the operation
5180 is a quad memory operation.
5181 (rs6000_legitimize_address): If quad memory, only allow register
5182 indirect for TImode addresses.
5183 (rs6000_legitimate_address_p): Likewise.
5184 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5185 (rs6000_reload_register_type): Likewise.
5186 (register_to_reg_type): Return register type.
5187 (rs6000_secondary_reload_simple_move): New helper function for
5188 secondary reload and secondary memory needed to identify anything
5189 that is a simple move, and does not need reloading.
5190 (rs6000_secondary_reload_direct_move): New helper function for
5191 secondary reload to identify cases that can be done with several
5192 instructions via the direct move instructions.
5193 (rs6000_secondary_reload_move): New helper function for secondary
5194 reload to identify moves between register types that can be done.
5195 (rs6000_secondary_reload): Add support for quad memory operations
5196 and for direct move.
5197 (rs6000_secondary_memory_needed): Likewise.
5198 (rs6000_debug_secondary_memory_needed): Change argument names.
5199 (rs6000_output_move_128bit): New function to return the move to
5200 use for 128-bit moves, including knowing about the various
5201 limitations of quad memory operations.
5203 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5204 memory operations. call rs6000_output_move_128bit for the actual
5205 instruciton(s) to generate.
5206 (vsx_movti_64bit): Likewise.
5208 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5209 (UNSPEC_P8V_MTVSRWZ): Likewise.
5210 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5211 (UNSPEC_P8V_MTVSRD): Likewise.
5212 (UNSPEC_P8V_XXPERMDI): Likewise.
5213 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5214 (UNSPEC_FUSION_GPR): Likewise.
5215 (FMOVE128_GPR): New iterator for direct move.
5216 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
5219 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5220 loads and direct move instructions.
5221 (zero_extendsidi2_lfiwzx): Likewise.
5222 (extendsidi2_lfiwax): Likewise.
5223 (extendsidi2_nocell): Likewise.
5224 (floatsi<mode>2_lfiwax): Likewise.
5226 (floatunssi<mode>2_lfiwzx): Likewise.
5228 (fix_trunc<mode>_stfiwx): Likewise.
5229 (fixuns_trunc<mode>_stfiwx): Likewise.
5230 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5231 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5232 (parity<mode>2_cmpb): Set length/type attr.
5233 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5234 for 'mr.' to fast_compare.
5235 (bpermd_<mode>): Change type attr to popcnt.
5236 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5237 (p8_mtvsrwz_1): Likewise.
5238 (p8_mtvsrwz_2): Likewise.
5239 (reload_fpr_from_gpr<mode>): Likewise.
5240 (p8_mtvsrd_1): Likewise.
5241 (p8_mtvsrd_2): Likewise.
5242 (p8_xxpermdi_<mode>): Likewise.
5243 (reload_vsx_from_gpr<mode>): Likewise.
5244 (reload_vsx_from_gprsf): Likewise.
5245 (p8_mfvsrd_3_<mode>): LIkewise.
5246 (reload_gpr_from_vsx<mode>): Likewise.
5247 (reload_gpr_from_vsxsf): Likewise.
5248 (p8_mfvsrd_4_disf): Likewise.
5249 (multi-word GPR splits): Do not split direct moves or quad memory
5252 2013-06-10 David Malcolm <dmalcolm@redhat.com>
5254 * tree-into-ssa.c (interesting_blocks): Make static.
5256 2013-06-10 Jakub Jelinek <jakub@redhat.com>
5259 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5260 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5261 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5262 (get_variable_align): New function.
5263 (get_variable_section, emit_bss, emit_common,
5264 assemble_variable_contents, place_block_symbol): Use
5265 get_variable_align instead of DECL_ALIGN.
5266 (assemble_noswitch_variable): Add align argument, use it
5267 instead of DECL_ALIGN.
5268 (assemble_variable): Adjust caller. Use get_variable_align
5269 instead of DECL_ALIGN.
5270 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5272 (DATA_ABI_ALIGNMENT): Define.
5273 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5274 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5275 opt is false, only return the psABI mandated alignment increase.
5276 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5277 (DATA_ABI_ALIGNMENT): ... this.
5278 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5279 (DATA_ABI_ALIGNMENT): ... this.
5280 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5281 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5282 (DATA_ABI_ALIGNMENT): ... this.
5283 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5284 * doc/tm.texi: Regenerated.
5286 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
5288 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5289 cmp_code to construct REG_EQUAL note.
5291 2013-06-09 Jakub Jelinek <jakub@redhat.com>
5294 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5295 that operands[2] doesn't overlap with operands[0].
5297 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
5298 Jan Hubicka <jh@suse.cz>
5300 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5301 hack to mark symbols as used.
5303 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5305 PR rtl-optimization/57559
5306 * lra-constraints.c (process_alt_operands): Don't discourage
5307 memory with known offset for offsetable memory constraint.
5308 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5310 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5312 * varasm.c (struct oc_local_state): Reorder fields.
5313 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5314 and adjust accordingly.
5315 (output_constructor): Reorder initialization code and adjust call to
5316 output_constructor_bitfield.
5318 2013-06-07 Jan Hubicka <jh@suse.cz>
5320 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5322 2013-06-07 David Malcolm <dmalcolm@redhat.com>
5324 * tree-object-size.c (unknown): Make const.
5326 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5328 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5329 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5330 for last alternative in the cpu_facility attribute.
5332 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5335 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5336 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5337 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5338 * config/arm/constraints.md (Dg): New constraint.
5339 * config/arm/neon.md (xordi3_neon): Remove.
5340 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5341 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5343 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5345 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5346 Clean up alternatives.
5348 2013-06-07 Alan Modra <amodra@gmail.com>
5350 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5353 2013-06-07 Alan Modra <amodra@gmail.com>
5355 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5357 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5359 * config/arm/constraints.md (Df): New constraint.
5360 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5361 Correct length attribute for last two alternatives.
5363 2013-06-07 Alan Modra <amodra@gmail.com>
5365 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5366 override user -mfp-in-toc.
5367 (offsettable_ok_by_alignment): Consider just the current access
5368 rather than the whole object, unless BLKmode. Handle
5369 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5370 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5371 for -mcmodel=medium.
5372 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5373 override user -mfp-in-toc or -msum-in-toc. Default to
5374 -mno-fp-in-toc for -mcmodel=medium.
5376 2013-06-06 DJ Delorie <dj@redhat.com>
5378 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5379 TARGET_VALID_POINTER_MODE.
5381 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5382 Pat Haugen <pthaugen@us.ibm.com>
5383 Peter Bergner <bergner@vnet.ibm.com>
5385 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5386 Document new power8 builtins.
5388 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5389 condition code register, to allow 128-bit logical operations to be
5390 done in the VSX or GPR registers.
5391 (nor<mode>3): Use the canonical form for nor.
5392 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5393 vclz*, and vpopcnt* vector instructions.
5394 (nand<mode>3): Likewise.
5395 (orc<mode>3): Likewise.
5396 (clz<mode>2): LIkewise.
5397 (popcount<mode>2): Likewise.
5399 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5400 that only the GPRs are recognized.
5402 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5403 support for new power8 builtins.
5405 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5407 (xscvdpspn): Likewise.
5413 (vpopcnt): Likewise.
5414 (vpopcntb): Likewise.
5415 (vpopcnth): Likewise.
5416 (vpopcntw): Likewise.
5417 (vpopcntd): Likewise.
5422 (eqv_v16qi3): Likewise.
5423 (eqv_v8hi3): Likewise.
5424 (eqv_v4si3): Likewise.
5425 (eqv_v2di3): Likewise.
5426 (eqv_v4sf3): Likewise.
5427 (eqv_v2df3): Likewise.
5429 (nand_v16qi3): Likewise.
5430 (nand_v8hi3): Likewise.
5431 (nand_v4si3): Likewise.
5432 (nand_v2di3): Likewise.
5433 (nand_v4sf3): Likewise.
5434 (nand_v2df3): Likewise.
5436 (orc_v16qi3): Likewise.
5437 (orc_v8hi3): Likewise.
5438 (orc_v4si3): Likewise.
5439 (orc_v2di3): Likewise.
5440 (orc_v4sf3): Likewise.
5441 (orc_v2df3): Likewise.
5443 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
5444 allow power8 quad mode in 64-bit.
5445 (rs6000_builtin_vectorized_function): Add support to vectorize
5446 ISA 2.07 count leading zeros, population count builtins.
5447 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
5448 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
5449 (builtin_function_type): Add vgbbd builtin function which takes an
5451 (altivec_expand_vec_perm_const): Add support for new power8 merge
5454 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
5455 that does not include TImdoe for use with 32-bit.
5456 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
5458 (UNSPEC_VSX_CVDPSPN): Likewise.
5459 (vsx_xscvdpspn): Likewise.
5460 (vsx_xscvspdpn): Likewise.
5461 (vsx_xscvdpspn_scalar): Likewise.
5462 (vsx_xscvspdpn_directmove): Likewise.
5463 (vsx_and<mode>3): Split logical operations into 32-bit and
5464 64-bit. Add support to do logical operations on TImode as well as
5465 VSX vector types. Allow logical operations to be done in either
5466 VSX registers or in general purpose registers in 64-bit mode. Add
5467 splitters if GPRs were used. For AND, add clobber of CCmode to
5468 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
5470 (vsx_and<mode>3_32bit): Likewise.
5471 (vsx_and<mode>3_64bit): Likewise.
5472 (vsx_ior<mode>3): Likewise.
5473 (vsx_ior<mode>3_32bit): Likewise.
5474 (vsx_ior<mode>3_64bit): Likewise.
5475 (vsx_xor<mode>3): Likewise.
5476 (vsx_xor<mode>3_32bit): Likewise.
5477 (vsx_xor<mode>3_64bit): Likewise.
5478 (vsx_one_cmpl<mode>2): Likewise.
5479 (vsx_one_cmpl<mode>2_32bit): Likewise.
5480 (vsx_one_cmpl<mode>2_64bit): Likewise.
5481 (vsx_nor<mode>3): Likewise.
5482 (vsx_nor<mode>3_32bit): Likewise.
5483 (vsx_nor<mode>3_64bit): Likewise.
5484 (vsx_andc<mode>3): Likewise.
5485 (vsx_andc<mode>3_32bit): Likewise.
5486 (vsx_andc<mode>3_64bit): Likewise.
5487 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
5488 and xxlorc instructions.
5489 (vsx_eqv<mode>3_64bit): Likewise.
5490 (vsx_nand<mode>3_32bit): Likewise.
5491 (vsx_nand<mode>3_64bit): Likewise.
5492 (vsx_orc<mode>3_32bit): Likewise.
5493 (vsx_orc<mode>3_64bit): Likewise.
5495 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
5497 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
5499 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
5500 (p8_vmrgow): Likewise.
5501 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
5502 GPRs to be split under VSX.
5503 (p8v_clz<mode>2): Add power8 count leading zero support.
5504 (p8v_popcount<mode>2): Add power8 population count support.
5505 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
5508 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
5511 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
5513 (vec_nand): Likewise.
5514 (vec_vclz): Likewise.
5515 (vec_vclzb): Likewise.
5516 (vec_vclzd): Likewise.
5517 (vec_vclzh): Likewise.
5518 (vec_vclzw): Likewise.
5519 (vec_vgbbd): Likewise.
5520 (vec_vmrgew): Likewise.
5521 (vec_vmrgow): Likewise.
5522 (vec_vpopcnt): Likewise.
5523 (vec_vpopcntb): Likewise.
5524 (vec_vpopcntd): Likewise.
5525 (vec_vpopcnth): Likewise.
5526 (vec_vpopcntw): Likewise.
5528 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5530 PR rtl-optimization/57468
5531 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
5534 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5536 PR rtl-optimization/57459
5537 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
5538 type when setting live regs.
5540 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5542 * config/s390/s390.opt (mlra): New option.
5543 * config/s390/s390.c (s390_decompose_address): Check displacement
5544 for all registers for LRA.
5545 (s390_secondary_reload): Don't used secondary reloads for LRA.
5546 (s390_lra_p): New function.
5547 (TARGET_LRA_P): Define.
5548 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
5549 of attribute cpu_facility to zarch for the last alternative.
5550 (*cmpmem_short): Ditto.
5552 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
5554 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
5555 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
5556 (arm_expand_prologue): Likewise.
5558 2013-06-06 Teresa Johnson <tejohnson@google.com>
5561 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
5562 as this is now done by redirect_edge_and_branch_force.
5563 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
5564 barriers, and fix interaction with splitting.
5565 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
5566 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
5567 reflect changes made in the routine.
5568 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
5569 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
5570 since this is called in cfglayout mode, and replace partition fixup
5571 with assert as that is now done by force_nonfallthru_and_redirect.
5572 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
5573 already be marked with region crossing note.
5574 (insert_section_boundary_note): Make non-static, gate on flag
5575 has_bb_partition, rewrite to also check for multiple partitions.
5576 (rest_of_handle_reorder_blocks): Remove call to
5577 insert_section_boundary_note, now done later during free_cfg.
5578 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
5579 * bb-reorder.h (insert_section_boundary_note): Declare.
5580 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
5581 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
5582 invoke insert_section_boundary_note.
5583 (try_redirect_by_replacing_jump): Remove unnecessary
5584 check for region crossing note.
5585 (fixup_partition_crossing): New function.
5586 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
5587 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
5588 in non-cfglayout mode.
5589 (force_nonfallthru_and_redirect): Fixup partition boundaries,
5590 remove old code that tried to do this. Emit barrier correctly
5591 when we are in cfglayout mode.
5592 (last_bb_in_partition): New function.
5593 (rtl_split_edge): Correctly fixup partition boundaries.
5594 (commit_one_edge_insertion): Remove old code that tried to
5595 fixup region crossing edge since this is now handled in
5596 split_block, and set up insertion point correctly since
5597 block may now end in a jump.
5598 (verify_hot_cold_block_grouping): Guard against checking when not in
5599 linearized RTL mode.
5600 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
5602 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
5603 rtl_verify_flow_info, so not called in cfglayout mode.
5604 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
5605 (fixup_reorder_chain): Remove old code that attempted to fixup region
5606 crossing note as this is now handled in force_nonfallthru_and_redirect.
5607 (duplicate_insn_chain): Don't duplicate switch section notes.
5608 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
5610 * basic-block.h (emit_barrier_after_bb): Declare.
5612 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5614 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
5615 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
5616 arm_usatsihi): Adjust alternatives for arm_restrict_it.
5618 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5620 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
5622 * config/arm/ldmstm.md: Regenerate.
5624 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5626 * config/arm/sync.md (atomic_loaddi_1):
5627 Disable predication for arm_restrict_it.
5628 (arm_load_exclusive<mode>): Likewise.
5629 (arm_load_exclusivesi): Likewise.
5630 (arm_load_exclusivedi): Likewise.
5631 (arm_load_acquire_exclusive<mode>): Likewise.
5632 (arm_load_acquire_exclusivesi): Likewise.
5633 (arm_load_acquire_exclusivedi): Likewise.
5634 (arm_store_exclusive<mode>): Likewise.
5635 (arm_store_exclusive<mode>): Likewise.
5636 (arm_store_release_exclusivedi): Likewise.
5637 (arm_store_release_exclusive<mode>): Likewise.
5639 2013-06-06 Richard Biener <rguenther@suse.de>
5641 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
5643 (lto_tag_is_tree_code_p): Adjust.
5644 (lto_tag_is_gimple_code_p): Likewise.
5645 (lto_gimple_code_to_tag): Likewise.
5646 (lto_tag_to_gimple_code): Likewise.
5647 (lto_tree_code_to_tag): Likewise.
5648 (lto_tag_to_tree_code): Likewise.
5649 * data-streamer.h (streamer_write_hwi_in_range): Use
5650 uhwi streaming to stream the normalized range.
5651 (streamer_read_hwi_in_range): Likewise.
5653 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5655 * config/arm/arm.md (enabled_for_depr_it): New attribute.
5656 (predicable_short_it): Likewise.
5657 (predicated): Likewise.
5658 (enabled): Handle above.
5659 (define_cond_exec): Set predicated attribute to yes.
5661 2013-06-05 Mike Stump <mikestump@comcast.net>
5663 * gdbinit.in (__FUNCTION__): Add.
5665 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
5667 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
5668 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
5670 2013-06-05 Jan Hubicka <jh@suse.cz>
5672 * varasm.c (mark_decl_referenced): Revert the removal until targets
5675 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
5677 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
5678 instead of mark_decl_referenced.
5680 2013-06-05 Jan Hubicka <jh@suse.cz>
5682 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
5683 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
5684 and symtab_used_from_object_file_p.
5685 (cgraph_make_node_local_1): Clear forced_by_abi.
5686 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
5687 * cgraph.h (symtab_node_base): Add forced_by_abi.
5688 (decide_is_variable_needed): Remove.
5689 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
5690 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
5691 (decide_is_symbol_needed): ... this one; handle symbols in general;
5692 always analyze virtuals; honnor forced_by_abi.
5693 (cgraph_finalize_function): Update.
5694 (varpool_finalize_decl): Update.
5695 (symbol_defined_and_needed): Remove.
5696 (analyze_functions): Update.
5697 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
5698 output_refs, input_overwrite_node): Handle forced_by_abi.
5699 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
5700 (address_taken_from_non_vtable_p): ... this one.
5701 (comdat_can_be_unshared_p_1): New function.
5702 (cgraph_comdat_can_be_unshared_p): Rename to ...
5703 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
5704 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
5705 (function_and_variable_visibility): Clear forced_by_abi as needed.
5706 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
5707 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
5708 * symtab.c (dump_symtab_base): Dump forced_by_abi.
5709 * varpool.c (decide_is_variable_needed): Remove.
5711 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5713 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
5714 (arm_option_override): Override arm_restrict_it where appropriate.
5715 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
5716 * config/arm/arm.opt (mrestrict-it): New command-line option.
5717 * doc/invoke.texi: Document -mrestrict-it.
5719 2013-06-05 David Malcolm <dmalcolm@redhat.com>
5721 * tsan.c (tsan_atomic_table): Make const.
5723 2013-06-05 Richard Biener <rguenther@suse.de>
5725 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
5726 index associated with the tree we are supposed to replace.
5727 * tree-streamer-out.c (pack_ts_base_value_fields): Output
5728 TREE_ASM_WRITTEN as zero for everything but SSA names.
5730 2013-06-05 David Malcolm <dmalcolm@redhat.com>
5732 * tree-ssa-structalias.c (call_stmt_vars): Make static.
5734 2013-06-04 Jan Hubicka <jh@suse.cz>
5736 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
5737 (input_node, input_varpool_node): Handle correctly external same
5739 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
5740 nodes at ltrans stage.
5742 2013-06-04 Jan Hubicka <jh@suse.cz>
5744 * ipa-inline.c (update_caller_keys): Fix availability test.
5745 (update_callee_keys): Likewise.
5746 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
5747 to follow ELF standard.
5749 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
5751 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
5752 (mips64r5900el-*-elf*): New configurations.
5753 * config/mips/mips-cpus.def (r5900): New processor.
5754 * config/mips/mips-tables.opt: Regenerate.
5755 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
5756 (mips_issue_rate): Handle PROCESSOR_R5900.
5757 (mips_reorg_process_insns): Force reorder mode for the R5900.
5758 * config/mips/mips.h (TARGET_MIPS5900): Define.
5759 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
5761 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
5763 * config/mips/mips.md (processor): Add r5900.
5764 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
5766 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5768 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
5769 into function to generate MOVI instruction.
5770 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
5771 (aarch64_preferred_simd_mode): Turn into wrapper.
5772 (aarch64_output_scalar_simd_mov_immediate): New function.
5773 * config/aarch64/aarch64-protos.h: Add prototype for above.
5775 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5777 * config/aarch64/aarch64.c (simd_immediate_info): Remove
5778 element_char member.
5779 (sizetochar): Return signed char.
5780 (aarch64_simd_valid_immediate): Remove elchar and other
5781 unnecessary variables.
5782 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
5783 Calculate element_char as required.
5784 * config/aarch64/aarch64-protos.h: Update and move prototype
5785 for aarch64_output_simd_mov_immediate.
5786 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
5789 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5791 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
5792 information completed by aarch64_simd_valid_immediate.
5793 (aarch64_legitimate_constant_p): Update arguments.
5794 (aarch64_simd_valid_immediate): Work with struct rather than many
5796 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
5797 (aarch64_simd_make_constant): Update arguments.
5798 (aarch64_output_simd_mov_immediate): Work with struct rather than
5799 many pointers. Output immediate directly rather than as operand.
5800 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5802 * config/aarch64/constraints.md (Dn): Update arguments.
5804 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5806 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
5808 (aarch64_simd_immediate_valid_for_move): Remove.
5809 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
5810 (aarch64_simd_make_constant): Update call.
5811 (aarch64_output_simd_mov_immediate): Update call.
5812 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5814 * config/aarch64/constraints.md (Dn): Update call.
5816 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5818 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
5819 return type to bool for prototype.
5820 (aarch64_legitimate_constant_p): Check for true instead of not -1.
5821 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
5822 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
5824 2013-06-04 Catherine Moore <clm@codesourcery.com>
5826 * config/mips/mips.opt (meva): New.
5827 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
5828 (ASM_SPEC): Handle -meva.
5829 * doc/invoke.texi (meva): Document.
5831 2013-06-04 Alan Modra <amodra@gmail.com>
5833 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
5836 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5838 * rtl.def: Add extra fourth optional field to define_cond_exec.
5839 * gensupport.c (process_one_cond_exec): Process attributes from
5841 * doc/md.texi: Document fourth field in define_cond_exec.
5843 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
5845 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
5846 out the processing order as in store_bit_field_1.
5848 2013-06-04 Jan Hubicka <jh@suse.cz>
5851 * cgraphunit.c (cgraph_process_same_body_aliases): Create
5852 non-VAR_DECL node if it does not exist yet.
5854 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
5856 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
5857 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
5858 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
5859 target_cpu_default setting.
5861 2013-06-03 Teresa Johnson <tejohnson@google.com>
5863 * dumpfile.c (opt_info_switch_p): Change -fopt-info
5864 default to -fopt-info=optimized instead of all.
5865 * doc/invoke.texi: Ditto.
5866 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
5867 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
5868 (execute_vect_slp): Emit BB vectorization success under
5869 MSG_OPTIMIZED_LOCATIONS.
5870 * tree-vect-slp.c (vect_slp_transform_bb): Change
5871 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
5872 * tree-vect-loop.c (vect_transform_loop): Ditto.
5874 2013-06-03 Jason Merrill <jason@redhat.com>
5877 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
5878 Use TARGET_EXPR for C++.
5880 2013-06-03 Jakub Jelinek <jakub@redhat.com>
5882 PR rtl-optimization/57268
5883 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
5884 if DEBUG_INSN_P (insn).
5887 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
5889 PR rtl-optimization/57268
5890 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5891 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5893 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
5895 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
5896 (ix86_avoid_lea_for_addr): Likewise.
5897 (exact_dependency_1): Likewise.
5898 (ix86_adjust_cost): Likewise.
5899 (swap_top_of_ready_list): Fix formatting and !reload_completed check
5901 (do_reorder_for_imul): Fix typo, formatting and
5902 !reload_completed check removed.
5903 (ix86_sched_reorder): Fix typo and formatting.
5904 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
5907 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
5909 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
5911 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
5913 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
5914 <STRING_CST>: Likewise.
5915 <VECTOR_CST>: Likewise.
5917 2013-06-01 Janus Weil <janus@gcc.gnu.org>
5918 Mikael Morin <mikael@gcc.gnu.org>
5920 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
5921 * config.in: Regenerated.
5922 * configure: Regenerated.
5924 2013-06-01 Jan Hubicka <jh@suse.cz>
5927 * cgraphunit.c (compile): When weakref is not supported,
5928 set up transparent aliases before final output pass.
5929 * varasm.c (assemble_alias): Do not try to do it here.
5931 2013-06-01 Jan Hubicka <jh@suse.cz>
5934 * passes.c (for_per_function): Skip unanalyzed functions.
5936 2013-06-01 Jan Hubicka <jh@suse.cz>
5938 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
5939 (lto_symtab_merge_symbols_1): ... this one.
5940 (lto_symtab_merge_cgraph_nodes): Rename to ...
5941 (lto_symtab_merge_symbols): ... this one; simplify.
5942 * cgraph.c (same_body_aliases_done): Rename to ...
5943 (cpp_implicit_aliases_done): ... this one.
5944 (cgraph_create_function_alias): Update.
5945 (cgraph_same_body_alias): Update.
5946 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
5947 (verify_edge_corresponds_to_fndecl): Simplify.
5948 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
5949 (cgraph_node): Remove same_body_alias.
5950 (varpool_node): Remove alias_of and extra_name_alias.
5951 (same_body_aliases_done): Rename to ..
5952 (cpp_implicit_aliases_done): ... this one.
5953 (symtab_alias_ultimate_target): Add default parameter.
5954 (symtab_resolve_alias): New function.
5955 (fixup_same_cpp_alias_visibility): Declare.
5956 (cgraph_function_node): Add default parameter.
5957 (cgraph_node_asm_name): Likewise.
5958 (cgraph_function_or_thunk_node): Add default parameter; do
5959 not ICE when it is NULL.
5960 (varpool_variable_node): Likewise.
5961 * tree-emutls.c (create_emultls_var): Update.
5962 (ipa_lower_emutls): Update.
5963 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
5964 (cgraph_reset_node): Reset alias info.
5965 (cgraph_finalize_function): Update.
5966 (fixup_same_cpp_alias_visibility): Move to symtab.c.
5967 (analyze_function): Simplify.
5968 (cgraph_process_same_body_aliases): Simplify.
5969 (analyze_functions): Fixup same body aliases.
5970 (handle_alias_pairs): Simplify.
5971 (assemble_thunk): Update.
5972 (assemble_thunks_and_aliases): Update.
5973 (output_weakrefs): Rewrite.
5974 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
5975 (lto_output_varpool_node): Likewise.
5976 (compute_ltrans_boundary): Remve assert.
5977 (get_alias_symbol): New functoin.
5978 (input_node): Rewrite alias handling.
5979 (input_varpool_node): Likewise.
5980 * ipa-pure-const.c (propagate_pure_const): Fix formating.
5981 * ipa.c (process_references): Handle weakrefs correctly.
5982 (symtab_remove_unreachable_nodes): Likewise.
5983 * trans-mem.c (get_cg_data): Update.
5984 (ipa_tm_create_version_alias): Update.
5985 (ipa_tm_execute): Update.
5986 * symtab.c (dump_symtab_base): Dump aliases.
5987 (verify_symtab_base): Verify aliases.
5988 (symtab_node_availability): New function.
5989 (symtab_alias_ultimate_target): Simplify.
5990 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
5991 handle all the fixup cases.
5992 (symtab_resolve_alias): New function.
5993 * passes.c (ipa_write_summaries): Handle weakrefs.
5994 * varpool.c (varpool_analyze_node): Simplify.
5995 (assemble_aliases): Update.
5996 (varpool_create_variable_alias): Simplify.
5997 (varpool_extra_name_alias): Simplify.
5998 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
5999 (lto_symtab_merge_symbols): ... this one.
6001 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
6004 PR rtl-optimization/57268
6005 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6006 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6008 2013-06-01 Tobias Burnus <burnus@net-b.de>
6011 2013-05-31 Tobias Burnus <burnus@net-b.de>
6014 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
6017 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6019 PR rtl-optimization/57268
6020 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6021 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6023 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6025 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
6026 unordered comparison operators when -fno-trapping-math is in effect
6028 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
6029 and implement unordered comparison operators properly on the e500.
6031 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6033 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
6034 for constant scalar integers.
6035 (simplify_relational_operation_1): Likewise.
6037 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
6039 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
6040 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
6043 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6044 Igor Zamyatin <igor.zamyatin@intel.com>
6046 Silvermont (SLM) architecture performance tuning.
6047 * config/i386/i386.h (enum ix86_tune_indices): Add
6048 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
6049 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
6051 * config/i386/i386.c (initial_ix86_tune_features)
6052 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
6053 (ix86_lea_outperforms): Handle Silvermont tuning.
6054 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
6056 (ix86_use_lea_for_mov): Likewise.
6057 (ix86_avoid_lea_for_addr): Likewise.
6058 (ix86_lea_for_add_ok): Likewise.
6059 (exact_dependency_1): New function.
6060 (exact_store_load_dependency): Likewise.
6061 (ix86_adjust_cost): Handle Silvermont tuning.
6062 (do_reoder_for_imul): Likewise.
6063 (swap_top_of_ready_list): New function.
6064 (ix86_sched_reorder): Changed to handle Silvermont tuning.
6066 * config/i386/i386.md (peepholes that split memory operand in fp
6069 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6071 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6072 Remove un-necessary braces.
6074 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6076 * config/aarch64/aarch64.c (aarch64_classify_symbol):
6077 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
6079 2013-05-31 Tobias Burnus <burnus@net-b.de>
6082 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
6084 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6087 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6088 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6089 * config/arm/neon.md (iordi3_neon): Remove.
6090 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6091 * config/arm/predicates.md (imm_for_neon_logic_operand):
6092 Move to earlier in the file.
6093 (neon_logic_op2): Likewise.
6094 (arm_iordi_operand_neon): New predicate.
6096 2013-05-31 Richard Biener <rguenther@suse.de>
6098 PR tree-optimization/57478
6099 PR tree-optimization/57453
6100 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6103 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6105 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6106 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6108 2013-05-30 Tobias Burnus <burnus@net-b.de>
6109 Thomas Koenig <tkoenig@gcc.gnu.org>
6112 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6113 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6115 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6117 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
6119 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6121 * target.def (register_usage_leveling_p): New hook.
6122 * targhooks.c (default_register_usage_leveling_p): New.
6123 * targhooks.h (default_register_usage_leveling_p): New prototype.
6124 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6125 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6126 * doc/tm.texi: Update.
6127 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6129 2013-05-30 Ian Bolton <ian.bolton@arm.com>
6131 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6132 (*insv_reg<mode>): New define_insn.
6134 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6136 PR rtl-optimization/57439
6137 * postreload.c (move2add_valid_value_p): Check that we have
6138 a zero subreg_regno_offset when accessing the register in
6141 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6142 Igor Zamyatin <igor.zamyatin@intel.com>
6144 Silvermont (SLM) architecture pipeline model, tuning and
6146 * config.gcc: Add slm config options and target.
6148 * config/i386/slm.md: New.
6150 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6152 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6154 (ix86_target_macros_internal): Likewise.
6156 * gcc/config/i386/i386.c (slm_cost): New cost.
6157 (m_SLM): New macro flag.
6158 (initial_ix86_tune_features): Set m_SLM.
6159 (x86_accumulate_outgoing_args): Likewise.
6160 (x86_arch_always_fancy_math_387): Likewise.
6161 (processor_target_table): Add slm cost.
6162 (cpu_names): Add slm cpu name.
6163 (x86_option_override_internal): Set SLM ISA.
6164 (ix86_issue_rate): New case PROCESSOR_SLM.
6165 (ia32_multipass_dfa_lookahead): Likewise.
6166 (fold_builtin_cpu): Add slm.
6168 * config/i386/i386.h (TARGET_SLM): New target macro.
6169 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6170 (processor_type): Add PROCESSOR_SLM.
6172 * config/i386/i386.md (cpu): Add new value "slm".
6173 (slm.md): Include slm.md.
6175 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6176 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6178 * config/arm/arm-protos.h: Add and update function protos.
6179 * config/arm/arm.c (use_simple_return_p): New added.
6180 (thumb2_expand_return): Check simple_return flag.
6181 * config/arm/arm.md: Add simple_return and conditional simple_return.
6182 * config/arm/iterators.md: Add iterator for return and simple_return.
6184 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6186 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6187 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6188 (arm_emit_vfp_multi_reg_pop): Likewise.
6189 (thumb2_emit_ldrd_pop): Likewise.
6190 (arm_expand_epilogue): Add misc REG_CFA notes.
6191 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6193 2013-05-29 Lawrence Crowl <crowl@google.com>
6195 * config/arm/t-arm: Update for below.
6197 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6198 Change type to hash_table. Update dependent calls and types.
6200 * config/i386/t-cygming: Update for below.
6202 * config/i386/t-interix: Update for below.
6204 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6205 Change type to hash_table. Update dependent calls and types.
6206 (i386_find_on_wrapper_list::wrappers): Likewise.
6208 * config/ia64/t-ia64: Update for below.
6210 * config/ia64/ia64.c (bundle_state_table):
6211 Change type to hash_table. Update dependent calls and types.
6213 * config/mips/mips.c (mips_reorg_process_insns::htab):
6214 Change type to hash_table. Update dependent calls and types.
6216 * config/sol2.c (solaris_comdat_htab):
6217 Change type to hash_table. Update dependent calls and types.
6219 * config/t-sol2: Update for above.
6221 2013-05-29 Teresa Johnson <tejohnson@google.com>
6223 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6224 functions are not yet marked as defined.
6226 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6227 Pat Haugen <pthaugen@us.ibm.com>
6228 Peter Bergner <bergner@vnet.ibm.com>
6230 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6234 (vrotl<mode>3): Likewise.
6235 (vashl<mode>3): Likewise.
6236 (vlshr<mode>3): Likewise.
6237 (vashr<mode>3): Likewise.
6239 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6240 support for power8 V2DI builtins.
6242 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6243 power8 V2DI builtins.
6244 (vupkhsw): Likewise.
6245 (vupklsw): Likewise.
6246 (vaddudm): Likewise.
6251 (vpkudum): Likewise.
6252 (vpksdss): Likewise.
6253 (vpkudus): Likewise.
6254 (vpksdus): Likewise.
6259 (vsubudm): Likewise.
6260 (vcmpequd): Likewise.
6261 (vcmpgtsd): Likewise.
6262 (vcmpgtud): Likewise.
6263 (vcmpequd_p): Likewise.
6264 (vcmpgtsd_p): Likewise.
6265 (vcmpgtud_p): Likewise.
6266 (vupkhsw): Likewise.
6267 (vupklsw): Likewise.
6268 (vaddudm): Likewise.
6273 (vpksdss): Likewise.
6274 (vpksdus): Likewise.
6275 (vpkudum): Likewise.
6276 (vpkudus): Likewise.
6281 (vsubudm): Likewise.
6283 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6284 support for power8 V2DI instructions.
6286 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6287 power8 V2DI instructions. Combine pack and unpack insns to use an
6288 iterator for each mode. Check whether a particular mode supports
6289 Altivec instructions instead of just checking TARGET_ALTIVEC.
6290 (UNSPEC_VPKUWUM): Likewise.
6291 (UNSPEC_VPKSHSS): Likewise.
6292 (UNSPEC_VPKSWSS): Likewise.
6293 (UNSPEC_VPKUHUS): Likewise.
6294 (UNSPEC_VPKSHUS): Likewise.
6295 (UNSPEC_VPKUWUS): Likewise.
6296 (UNSPEC_VPKSWUS): Likewise.
6297 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6298 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6299 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6300 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6301 (UNSPEC_VUPKHSB): Likewise.
6302 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6303 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6304 (UNSPEC_VUPKHSH): Likewise.
6305 (UNSPEC_VUPKLSB): Likewise.
6306 (UNSPEC_VUPKLSH): Likewise.
6308 (VI_char): Likewise.
6309 (VI_scalar): Likewise.
6310 (VI_unit): Likewise.
6312 (VP_small): Likewise.
6313 (VP_small_lc): Likewise.
6314 (VU_char): Likewise.
6315 (add<mode>3): Likewise.
6316 (altivec_vaddcuw): Likewise.
6317 (altivec_vaddu<VI_char>s): Likewise.
6318 (altivec_vadds<VI_char>s): Likewise.
6319 (sub<mode>3): Likewise.
6320 (altivec_vsubcuw): Likewise.
6321 (altivec_vsubu<VI_char>s): Likewise.
6322 (altivec_vsubs<VI_char>s): Likewise.
6323 (altivec_vavgs<VI_char>): Likewise.
6324 (altivec_vcmpbfp): Likewise.
6325 (altivec_eq<mode>): Likewise.
6326 (altivec_gt<mode>): Likewise.
6327 (altivec_gtu<mode>): Likewise.
6328 (umax<mode>3): Likewise.
6329 (smax<mode>3): Likewise.
6330 (umin<mode>3): Likewise.
6331 (smin<mode>3): Likewise.
6332 (altivec_vpkuhum): Likewise.
6333 (altivec_vpkuwum): Likewise.
6334 (altivec_vpkshss): Likewise.
6335 (altivec_vpkswss): Likewise.
6336 (altivec_vpkuhus): Likewise.
6337 (altivec_vpkshus): Likewise.
6338 (altivec_vpkuwus): Likewise.
6339 (altivec_vpkswus): Likewise.
6340 (altivec_vpks<VI_char>ss): Likewise.
6341 (altivec_vpks<VI_char>us): Likewise.
6342 (altivec_vpku<VI_char>us): Likewise.
6343 (altivec_vpku<VI_char>um): Likewise.
6344 (altivec_vrl<VI_char>): Likewise.
6345 (altivec_vsl<VI_char>): Likewise.
6346 (altivec_vsr<VI_char>): Likewise.
6347 (altivec_vsra<VI_char>): Likewise.
6348 (altivec_vsldoi_<mode>): Likewise.
6349 (altivec_vupkhsb): Likewise.
6350 (altivec_vupkhs<VU_char>): Likewise.
6351 (altivec_vupkls<VU_char>): Likewise.
6352 (altivec_vupkhsh): Likewise.
6353 (altivec_vupklsb): Likewise.
6354 (altivec_vupklsh): Likewise.
6355 (altivec_vcmpequ<VI_char>_p): Likewise.
6356 (altivec_vcmpgts<VI_char>_p): Likewise.
6357 (altivec_vcmpgtu<VI_char>_p): Likewise.
6358 (abs<mode>2): Likewise.
6359 (vec_unpacks_hi_v16qi): Likewise.
6360 (vec_unpacks_hi_v8hi): Likewise.
6361 (vec_unpacks_lo_v16qi): Likewise.
6362 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6363 (vec_unpacks_lo_v8hi): Likewise.
6364 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6365 (vec_pack_trunc_v8h): Likewise.
6366 (vec_pack_trunc_v4si): Likewise.
6367 (vec_pack_trunc_<mode>): Likewise.
6369 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6371 (vec_vmaxsd): Likewise.
6372 (vec_vmaxud): Likewise.
6373 (vec_vminsd): Likewise.
6374 (vec_vminud): Likewise.
6375 (vec_vpksdss): Likewise.
6376 (vec_vpksdus): Likewise.
6377 (vec_vpkudum): Likewise.
6378 (vec_vpkudus): Likewise.
6379 (vec_vrld): Likewise.
6380 (vec_vsld): Likewise.
6381 (vec_vsrad): Likewise.
6382 (vec_vsrd): Likewise.
6383 (vec_vsubudm): Likewise.
6384 (vec_vupkhsw): Likewise.
6385 (vec_vupklsw): Likewise.
6387 2013-05-29 Jan Hubicka <jh@suse.cz>
6389 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6390 flags; reorder rest of fields in more consistent way.
6391 (varpool_node): Remove analyzed, finalized and alias.
6392 (cgraph_ndoe): Likewise.
6393 (symtab_alias_ultimate_target): New function.
6394 (cgraph_function_node): Move offline.
6395 (cgraph_reset_node): Declare.
6396 (cgraph_comdat_can_be_unshared_p): Remove.
6397 (varpool_remove_initializer): Declare.
6398 (varpool_first_defined_variable, varpool_next_defined_variable
6399 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6400 (cgraph_function_with_gimple_body_p): Update.
6401 (varpool_all_refs_explicit_p): Update.
6402 (symtab_alias_target): New function.
6403 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6404 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
6405 (cgraph_function_or_thunk_node): Simplify using
6406 symtab_alias_ultimate_target.
6407 (varpool_variable_node): Likewise.
6408 * cgraph.c (cgraph_create_function_alias): Update.
6409 (cgraph_add_thunk): Update.
6410 (cgraph_remove_node): Update.
6411 (dump_cgraph_node): Do not dump removed flags.
6412 (cgraph_function_body_availability): Update.
6413 (cgraph_propagate_frequency): Update.
6414 (verify_cgraph_node): Check sanity of local flag.
6415 (cgraph_function_node): Move here from cgraph.h; revamp for
6416 cgraph_function_or_thunk_node.
6417 * lto-symtab.c (lto_varpool_replace_node): Update.
6418 (lto_symtab_resolve_can_prevail_p): Update.
6419 (lto_symtab_merge_cgraph_nodes): Update.
6420 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6421 propagate_constants_accross_call, devirtualization_time_bonus,
6422 ipcp_propagate_stage): Update.
6423 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
6424 * ipa-inline-transform.c (clone_inlined_nodes,
6425 preserve_function_body_p): Update.
6426 * ipa-reference.c (propagate): Update.
6427 (write_node_summary_p): Update.
6428 * toplev.c (wrapup_global_declaration_2): Update.
6429 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6430 (analyze_function) ... this one.
6431 (cgraph_process_new_functions): Update.
6432 (cgraph_reset_node): Export.
6433 (cgraph_finalize_function): Update.
6434 (cgraph_add_new_function): Update.
6435 (process_function_and_variable_attributes): Update.
6436 (varpool_finalize_decl): Update.
6437 (symbol_finalized): Remove.
6438 (symbol_finalized_and_needed): Rename to ...
6439 (symbol_defined_and_needed): ... update.
6440 (cgraph_analyze_functions): Update.
6441 (handle_alias_pairs): Update.
6442 (mark_functions_to_output): Update.
6443 (assemble_thunk): Update.
6444 (output_in_order): Update.
6445 (output_weakrefs): Update.
6446 (finalize_compilation_unit): Update.
6447 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
6448 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
6449 input_node, input_varpool_node): Update.
6450 * dbxout.c (dbxout_expand_expr): Update.
6451 * cgraphclones.c (cgraph_clone_node): Update.
6452 (cgraph_copy_node_for_versioning): Update.
6453 (cgraph_materialize_clone): Update.
6454 (cgraph_materialize_all_clones): Update.
6455 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
6456 propagate_pure_const, propagate_nothrow): Update.
6457 * lto-streamer-out.c (lto_output, write_symbol): Update.
6458 * ipa-utils.c (ipa_reverse_postorder): Update.
6459 * ipa-inline.c (can_inline_edge_p): Update.
6460 (update_caller_keys, ipa_inline): Update.
6461 * dwarf2out.c (reference_to_unused,
6462 premark_types_used_by_global_vars_helper): Update.
6463 * tree-eh.c (tree_could_trap_p): Update.
6464 * ipa-split.c (consider_split, execute_split_functions): Update.
6465 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
6466 has_addr_references_p): Update; move ahead in file for better
6468 (process_references): Simplify.
6469 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
6471 (cgraph_comdat_can_be_unshared_p): Make static.
6472 (cgraph_externally_visible_p): Update.
6473 (varpool_externally_visible_p): Update.
6474 (function_and_variable_visibility): Update.
6475 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
6476 ipa_tm_mark_force_output_node): Update.
6477 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
6478 estimate_edge_devirt_benefit, inline_generate_summary,
6479 inline_write_summary): Update.
6480 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
6481 * ipa-prop.c (ipa_compute_jump_functions): Update.
6482 (ipa_print_node_params, ipa_prop_read_section,
6483 ipa_update_after_lto_read, read_replacements_section): Update.
6484 * varasm.c (mark_decl_referenced): Update.
6485 (assemble_alias, dump_tm_clone_pairs): Update.
6486 * tree-inline.c (copy_bb): Update.
6487 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
6489 * symtab.c (dump_symtab_base): Print new flags.
6490 (verify_symtab_base): Verify new flags.
6491 (symtab_alias_ultimate_target): New function.
6492 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
6493 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
6495 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
6497 * i386.c (ix86_get_function_versions_dispatcher,
6498 ix86_generate_version_dispatcher_body): Update.
6499 (fold_builtin_cpu): Use varpool_add_new_variable.
6500 * varpool.c (varpool_remove_initializer): Break out from ...
6501 (varpool_remove_node): ... this one.
6502 (dump_varpool_node, varpool_node_for_asm,
6503 cgraph_variable_initializer_availability, varpool_analyze_node,
6504 varpool_assemble_decl, varpool_remove_unreferenced_decls,
6505 varpool_finalize_named_section_flags, varpool_create_variable_alias):
6508 2013-05-29 Jan Hubicka <jh@suse.cz>
6510 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
6512 2013-05-29 Easwaran Raman <eraman@google.com>
6514 PR tree-optimization/57442
6515 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
6516 when control exits the main loop.
6518 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
6520 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
6522 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
6523 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
6524 * rx/t-rx: Add rx100 under multi library matches option for nofpu
6527 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6529 PR tree-optimization/57441
6530 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
6531 Don't limit size of incr_vec to number of candidates.
6533 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
6535 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
6536 and mips16 directories.
6537 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
6538 (MULTILIB_DIRNAMES): Ditto.
6539 (MULTILIB_EXCEPTIONS): Add new exceptions.
6540 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
6541 (MULTILIB_DIRNAMES): Ditto.
6542 (MULTILIB_EXCEPTIONS): Add new exceptions.
6544 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6545 Marcus Shawcroft <marcus.shawcroft@arm.com>
6547 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
6548 SYMBOL_TINY_ABSOLUTE.
6549 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
6550 SYMBOL_TINY_ABSOLUTE.
6551 (aarch64_expand_mov_immediate): Likewise.
6552 (aarch64_classify_symbol): Likewise.
6553 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
6554 Permit SYMBOL_TINY_ABSOLUTE.
6555 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
6557 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6558 Marcus Shawcroft <marcus.shawcroft@arm.com>
6560 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
6561 Refactor if/switch. Replace gcc_assert with if.
6563 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
6565 * config/i386/i386.c (initial_ix86_tune_features): Enable
6566 FP Reassociation for AMD bdver1 and bdver2.
6568 2013-05-29 Martin Jambor <mjambor@suse.cz>
6570 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
6571 and IMAGPART_EXPR do not occur within other handled_components.
6573 2013-05-29 Richard Biener <rguenther@suse.de>
6575 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
6576 access on whether the use is in the BB we currently try to
6578 (vect_bb_vectorization_profitable_p): Pass the BB we currently
6579 vectorize to vect_bb_slp_scalar_cost.
6581 2013-05-29 Richard Biener <rguenther@suse.de>
6583 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
6584 computing scalar cost offsetted by stmts that are kept live
6586 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
6587 for computation of scalar cost.
6589 2013-05-28 Steve Ellcey <sellcey@mips.com>
6591 * config/mips/mips-cpus.def (mips32r2): Change processor type.
6593 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6595 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
6596 array notation built-in reduction functions.
6597 * doc/passes.texi (Passes): Added documentation about changes done
6599 * doc/invoke.texi (C Dialect Options): Added documentation about
6600 the -fcilkplus flag.
6601 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
6602 (BUILTINS_DEF): Depend on cilkplus.def.
6603 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
6604 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
6605 * cilkplus.def: New file.
6607 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
6609 PR rtl-optimization/57439
6610 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
6612 2013-05-28 Easwaran Raman <eraman@google.com>
6614 PR tree-optimization/57337
6615 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
6616 (find_insert_point): Correctly identify the insertion point
6617 when two statements with the same UID is compared.
6619 2013-05-28 Richard Biener <rguenther@suse.de>
6621 PR tree-optimization/56787
6622 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
6623 from the list of data references.
6624 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
6626 (vect_analyze_loop_operations): Likewise.
6627 (vect_transform_loop): Remove clobbers.
6629 2013-05-28 Martin Jambor <mjambor@suse.cz>
6631 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
6632 and REALPART_EXPRs have scalar type.
6634 2013-05-28 Richard Biener <rguenther@suse.de>
6636 PR tree-optimization/57411
6637 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
6639 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
6640 virtual operand propagation.
6642 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6644 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
6645 destination register for bmasksi_vis.
6646 (vector_init_bshuffle): Likewise.
6647 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
6649 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6651 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
6652 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
6653 mode if the instruction isn't available in the original mode.
6654 * config/sparc/sparc.opt (mfix-ut699): New option.
6655 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
6656 (divdf3): Turn into expander.
6657 (divdf3_nofix): New insn.
6658 (divdf3_fix): Likewise.
6659 (divsf3): Disable if -mfix-ut699.
6660 (sqrtdf2): Turn into expander.
6661 (sqrtdf2_nofix): New insn.
6662 (sqrtdf2_fix): Likewise.
6663 (sqrtsf2): Disable if -mfix-ut699.
6665 2013-05-27 Richard Biener <rguenther@suse.de>
6668 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
6669 block for the new loop.
6671 2013-05-27 Richard Biener <rguenther@suse.de>
6673 PR tree-optimization/57343
6674 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
6675 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
6676 (number_of_iterations_cond): Do not build the folded tree.
6678 2013-05-27 Richard Biener <rguenther@suse.de>
6682 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6683 OEP_CONSTANT_ADDRESS_OF retained.
6685 PR tree-optimization/57417
6686 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
6688 (set_ssa_val_to): Compare addresses using
6689 get_addr_base_and_unit_offset.
6691 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
6693 PR rtl-optimization/56833
6694 * postreload.c (move2add_record_mode): New function.
6695 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
6696 (move2add_use_add2_insn): Use move2add_record_sym_value.
6697 (move2add_use_add3_insn): Likewise.
6698 (reload_cse_move2add): Use move2add_valid_value_p and
6699 move2add_record_mode. Invalidate call-clobbered and REG_INC
6700 affected regs by setting reg_mode to VOIDmode.
6701 (move2add_note_store): Don't pretend the inside of a SUBREG is
6702 the actual destination. Invalidate single/leading registers by
6703 setting reg_mode to VOIDmode.
6704 Use move2add_record_sym_value, move2add_valid_value_p and
6705 move2add_record_mode.
6707 2013-05-27 Richard Biener <rguenther@suse.de>
6709 PR tree-optimization/57396
6710 * tree-affine.c (double_int_constant_multiple_p): Properly
6711 return false for val == 0 and div != 0.
6713 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6715 * config/mips/mips.h: Use #elif in preprocessor conditions.
6717 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6720 * config/mips/constraints.md (kl): New constraint.
6721 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
6722 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
6723 constraint for operand 0. Split after CSE for MIPS16. Emit a move
6725 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
6727 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6730 * config/mips/mips.c (mips_can_inline_p): New function.
6731 (TARGET_CAN_INLINE_P): Define.
6733 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
6735 * sched-int.h (ds_t, dw_t): Make unsigned int.
6736 Fix documentation that describes how all the ds_t bits are used.
6737 Reserve the last bit for delayed-branch scheduling.
6738 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
6739 (BITS_PER_DEP_WEAK): Fix definition and documentation.
6740 (gen_dep_weak_1): Remove prototype.
6741 * sched-deps.c (get_dep_weak_1): Make static.
6742 * target.def (speculate_insn, needs_block_p, gen_spec_check,
6743 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
6744 * doc/tm.texi: Regenerate.
6745 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
6747 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
6750 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
6752 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
6753 Sandra Loosemore <sandra@codesourcery.com>
6755 * config.gcc (powerpc-*): Allow native for with-cpu.
6757 2013-05-24 Jeff Law <law@redhat.com>
6759 PR tree-optimization/57124
6760 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
6761 conversion feeding a condition if the range has an overflow
6762 if -fstrict-overflow. Add warnings for when we do make the
6765 2013-05-24 Dehao Chen <dehao@google.com>
6767 * tree-cfg.c (locus_discrim_map): Fix the typo.
6768 (locus_discrim_hasher): Likewise.
6769 (locus_discrim_hasher::hash): Likewise.
6770 (locus_discrim_hasher::equal): Likewise.
6772 2013-05-24 Martin Jambor <mjambor@suse.cz>
6774 PR tree-optimization/57294
6775 * cgraph.h (ipa_record_stmt_references): Declare.
6776 * cgraphbuild.c (ipa_record_stmt_references): New function.
6777 (build_cgraph_edges): Use ipa_record_stmt_references.
6778 (rebuild_cgraph_edges): Likewise.
6779 (cgraph_rebuild_references): Likewise.
6780 * ipa-prop.c (ipa_modify_call_arguments): Discard references
6781 associated with the old statement and build references from the
6782 newly built statements.
6783 * ipa-ref.c (ipa_remove_stmt_references): New function.
6784 * ipa-ref.h (ipa_remove_stmt_references): Declare.
6786 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
6788 * lra-constraints.c (emit_spill_move): Use smaller mode for
6790 (check_and_process_move): Consider mem-reg moves for secondary
6792 (curr_insn_transform): Don't lose insns emitted before for
6793 secondary memory moves.
6794 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
6795 reg set up in the current insn.
6797 2013-05-24 Dehao Chen <dehao@google.com>
6799 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
6801 (locus_descrim_hasher::equal): Likewise.
6802 (build_gimple_cfg): New discriminator assignment algorithm.
6803 (make_edges): Likewise.
6804 (next_discriminator_for_locus): Likewise.
6805 (same_line_p): Likewise.
6806 (assign_discriminators): Likewise.
6807 (make_cond_expr_edges): Likewise.
6808 (make_gimple_switch_edges): Likewise.
6809 (make_goto_expr_edges): Likewise.
6810 (make_gimple_asm_edges): Likewise.
6812 2013-05-24 Ian Bolton <ian.bolton@arm.com>
6814 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
6815 X format specifier to only display bottom 16 bits.
6816 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
6817 immediate to match for operand 2, since it will be masked.
6819 2013-05-24 Richard Biener <rguenther@suse.de>
6821 PR tree-optimization/57287
6822 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
6823 all SSA names that occur in abnormal PHIs.
6825 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
6828 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
6829 that index is not negative.
6831 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
6833 PR rtl-optimization/55177
6834 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
6835 (simplify_byte_swapping_operation): New.
6836 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
6837 (simplify_relational_operation_1): Deal with BSWAP.
6839 2013-05-23 Richard Henderson <rth@redhat.com>
6842 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
6843 (ix86_reorg): Call it.
6845 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
6848 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
6849 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
6850 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
6852 2013-05-23 Christian Bruel <christian.bruel@st.com>
6855 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
6857 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6858 Marcus Shawcroft <marcus.shawcroft@arm.com>
6860 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
6861 * config/aarch64/constraints.md (Usa): Remove.
6862 * doc/md.texi (AArch64 Usa): Remove.
6864 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6865 Marcus Shawcroft <marcus.shawcroft@arm.com>
6867 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
6868 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
6869 * config/aarch64/predicates.md (aarch64_const_address): Remove.
6870 (aarch64_mov_operand): Use aarch64_mov_operand_p.
6872 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
6874 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
6875 instruction (AdvSIMD).
6876 * config/aarch64/aarch64-builtins.c
6877 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
6878 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
6880 2013-05-23 Martin Jambor <mjambor@suse.cz>
6883 * tree.h (contains_bitfld_component_ref_p): Declare.
6884 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
6885 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
6887 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
6888 not access a bit-field. Assert all final offsets are byte-aligned.
6890 2013-05-23 Richard Biener <rguenther@suse.de>
6892 PR tree-optimization/57380
6893 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
6894 least one invariant or re-used load.
6895 * passes.c (init_optimization_passes): Move pass_phiprop before
6898 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
6900 * config/aarch64/aarch64-simd.md
6901 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
6903 2013-05-23 Richard Biener <rguenther@suse.de>
6906 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6907 OEP_CONSTANT_ADDRESS_OF retained.
6909 2013-05-23 Jakub Jelinek <jakub@redhat.com>
6912 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
6913 don't lower unit. Handle unit not being always BITS_PER_WORD.
6915 2013-05-23 Richard Biener <rguenther@suse.de>
6917 PR rtl-optimization/57341
6918 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
6919 instead of true_dependence.
6921 2013-05-22 David Malcolm <dmalcolm@redhat.com>
6923 * bb-reorder.c (branch_threshold): Make const.
6924 (exec_threshold): Ditto.
6926 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6927 Pat Haugen <pthaugen@us.ibm.com>
6928 Peter Bergner <bergner@vnet.ibm.com>
6930 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
6931 documentation for the power8 crypto builtins.
6933 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
6935 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
6936 macros for defining power8 builtin functions.
6937 (BU_P8V_AV_2): Likewise.
6938 (BU_P8V_AV_P): Likewise.
6939 (BU_P8V_VSX_1): Likewise.
6940 (BU_P8V_OVERLOAD_1): Likewise.
6941 (BU_P8V_OVERLOAD_2): Likewise.
6942 (BU_CRYPTO_1): Likewise.
6943 (BU_CRYPTO_2): Likewise.
6944 (BU_CRYPTO_3): Likewise.
6945 (BU_CRYPTO_OVERLOAD_1): Likewise.
6946 (BU_CRYPTO_OVERLOAD_2): Likewise.
6947 (XSCVSPDP): Fix typo, point to the correct instruction.
6948 (VCIPHER): Add power8 crypto builtins.
6949 (VCIPHERLAST): Likewise.
6950 (VNCIPHER): Likewise.
6951 (VNCIPHERLAST): Likewise.
6952 (VPMSUMB): Likewise.
6953 (VPMSUMH): Likewise.
6954 (VPMSUMW): Likewise.
6955 (VPERMXOR_V2DI): Likewise.
6956 (VPERMXOR_V4SI: Likewise.
6957 (VPERMXOR_V8HI: Likewise.
6958 (VPERMXOR_V16QI: Likewise.
6959 (VSHASIGMAW): Likewise.
6960 (VSHASIGMAD): Likewise.
6962 (VPERMXOR): Likewise.
6963 (VSHASIGMA): Likewise.
6965 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6966 __CRYPTO__ if the crypto instructions are available.
6967 (altivec_overloaded_builtins): Add support for overloaded power8
6970 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
6971 support for power8 crypto builtins.
6972 (builtin_function_type): Likewise.
6973 (altivec_init_builtins): Add support for builtins that take vector
6974 long long (V2DI) arguments.
6976 * config/rs6000/crypto.md: New file, define power8 crypto
6979 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6980 Pat Haugen <pthaugen@us.ibm.com>
6981 Peter Bergner <bergner@vnet.ibm.com>
6983 * doc/invoke.texi (Option Summary): Add power8 options.
6984 (RS/6000 and PowerPC Options): Likewise.
6986 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
6987 constraints.md instead of rs6000.h. Reorder w* constraints. Add
6988 wm, wn, wr documentation.
6990 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
6991 registers if direct move instructions are enabled.
6992 (wn): New constraint for no registers.
6993 (wq): New constraint for quad word even GPR registers.
6994 (wr): New constraint if 64-bit instructions are enabled.
6995 (wv): New constraint if power8 vector instructions are enabled.
6996 (wQ): New constraint for quad word memory locations.
6998 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
6999 constraint for 0..15 for crypto instructions.
7000 (gpc_reg_operand): If VSX allow registers in VSX registers as well
7001 as GPR and floating point registers.
7002 (int_reg_operand): New predicate to match only GPR registers.
7003 (base_reg_operand): New predicate to match base registers.
7004 (quad_int_reg_operand): New predicate to match even GPR registers
7005 for quad memory operations.
7006 (vsx_reg_or_cint_operand): New predicate to allow vector logical
7007 operations in both GPR and VSX registers.
7008 (quad_memory_operand): New predicate for quad memory operations.
7009 (reg_or_indexed_operand): New predicate for direct move support.
7011 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
7012 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
7013 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
7014 (POWERPC_MASKS): Add power8 options.
7015 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
7018 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7019 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
7021 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
7022 (-mpower8-fusion): New power8 options.
7023 (-mpower8-fusion-sign): Likewise.
7024 (-mpower8-vector): Likewise.
7025 (-mcrypto): Likewise.
7026 (-mdirect-move): Likewise.
7027 (-mquad-memory): Likewise.
7029 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
7031 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
7033 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
7034 (rs6000_debug_vector_unit): Add p8_vector.
7035 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
7036 definitions. Also print fusion state.
7037 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
7038 (rs6000_builtin_mask_calculate): Add power8 builtin support.
7039 (rs6000_option_override_internal): Add support for power8.
7040 (rs6000_common_init_builtins): Add debugging for skipped builtins
7042 (rs6000_adjust_cost): Add power8 support.
7043 (rs6000_issue_rate): Likewise.
7044 (insn_must_be_first_in_group): Likewise.
7045 (insn_must_be_last_in_group): Likewise.
7046 (force_new_group): Likewise.
7047 (rs6000_register_move_cost): Likewise.
7048 (rs6000_opt_masks): Likewise.
7050 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
7051 power8 capable assembler, default to power7 options.
7052 (TARGET_DIRECT_MOVE): Likewise.
7053 (TARGET_CRYPTO): Likewise.
7054 (TARGET_P8_VECTOR): Likewise.
7055 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
7056 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
7057 (VECTOR_MEM_P8_VECTOR_P): Likewise.
7058 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
7059 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
7060 (TARGET_XSCVDPSPN): Likewise.
7061 (TARGET_XSCVSPDPN): Likewsie.
7062 (TARGET_SYNC_HI_QI): Likewise.
7063 (TARGET_SYNC_TI): Likewise.
7064 (MASK_CRYPTO): Likewise.
7065 (MASK_DIRECT_MOVE): Likewise.
7066 (MASK_P8_FUSION): Likewise.
7067 (MASK_P8_VECTOR): Likewise.
7068 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
7069 temporary used by some of the direct move instructions to get two FP
7070 temporary registers does not force creation of a stack frame.
7071 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
7072 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
7073 that any VSX registers are tieable, even if they are also an
7074 Altivec vector mode.
7075 (r6000_reg_class_enum): Add wm, wr, wv constraints.
7076 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
7077 (RS6000_BTM_CRYPTO): Likewise.
7078 (RS6000_BTM_COMMON): Likewise.
7080 * config/rs6000/rs6000.md (cpu attribute): Add power8.
7081 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
7082 (enum rs6000_vector): Add power8 vector support.
7084 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7088 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7089 (any_sibcall_could_use_r3): this and handle indirect calls.
7090 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7092 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7094 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7096 2013-05-22 Richard Biener <rguenther@suse.de>
7099 * profile.c (branch_prob): Do not split blocks that are
7100 abnormally receiving from ECF_RETURNS_TWICE functions.
7102 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7104 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7105 address mode. Move pointer mode initialization to the same place.
7107 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7109 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7110 while it has any effect.
7112 2013-05-21 Easwaran Raman <eraman@google.com>
7114 PR tree-optimization/57322
7115 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7116 UID of the statement added to the BB to be 1.
7118 2013-05-21 Jakub Jelinek <jakub@redhat.com>
7120 PR tree-optimization/57331
7121 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7122 of conversion from pointer type to integral type with integer.
7124 2013-05-21 Martin Jambor <mjambor@suse.cz>
7127 * ipa-prop.c (ipa_read_node_info): Process param_used and
7128 controlled_uses in the same order as when writing.
7130 2013-05-21 Magnus Granberg <baldrick@free.fr>
7133 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
7135 2013-05-21 Richard Biener <rguenther@suse.de>
7137 PR tree-optimization/57318
7138 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7139 estimate stmts with side-effects as likely eliminated.
7141 2013-05-21 Richard Biener <rguenther@suse.de>
7143 PR tree-optimization/57330
7144 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7145 preserve the call stmts fntype.
7147 2013-05-21 Richard Biener <rguenther@suse.de>
7149 PR tree-optimization/57303
7150 * tree-ssa-sink.c (statement_sink_location): Improve killing
7151 stmt detection and properly handle self-assignments.
7153 2013-05-21 Christian Bruel <christian.bruel@st.com>
7155 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7156 spanning registers. LEAF_REG_REMAP is supported only for contiguous
7157 registers. Set register size out of the PARALLEL loop.
7159 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7162 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7163 (*fmasf4, *fmasf4_media): New insns.
7165 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7167 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7168 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7169 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7170 (mips_idiv_insns): Update the comments to say that the returned
7171 instruction counts are in units of BASE_INSN_LENGTH.
7172 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7173 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7174 using 2 rather than 4 as the length of indirect MIPS16 and
7175 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7176 length of a NOP. Don't divide MIPS16 lengths by 2.
7177 (mips16_split_long_branches): Assume a branch is long if the
7178 length is greater than 4 rather than 8.
7179 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7180 rather than multiplying them by 2. Multiply instruction counts
7181 by BASE_INSN_LENGTH rather than 4.
7182 (*jump_mips16, tls_get_tp_mips16_<mode>)
7183 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7185 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7187 * config/mips/mips.md (extended_mips16): Remove branch case.
7188 (length): Remove duplicated extended_mips16 test.
7190 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7192 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7194 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7196 * recog.h (Recog_data): Rename to...
7197 (recog_data_d): ...this.
7198 (recog_data): Update accordingly.
7199 * recog.c (recog_data): Likewise.
7200 * reload.c (save_recog_data): Likewise.
7201 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7202 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7204 2013-05-17 Julian Brown <julian@codesourcery.com>
7206 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7207 found in a REG_EQUAL note, invalidate it.
7209 2013-05-17 Easwaran Raman <eraman@google.com>
7211 * tree-ssa-reassoc.c (find_insert_point): New function.
7212 (insert_stmt_after): Likewise.
7213 (get_def_stmt): Likewise.
7214 (ensure_ops_are_available): Likewise.
7215 (not_dominated_by): Likewise.
7216 (rewrite_expr_tree): Do not move statements beyond what is
7217 necessary. Remove call to swap_ops_for_binary_stmt...
7218 (reassociate_bb): ... and move it here.
7219 (build_and_add_sum): Assign UIDs for new statements.
7220 (linearize_expr): Likewise.
7221 (do_reassoc): Renumber gimple statement UIDs.
7223 2013-05-17 Jan Hubicka <jh@suse.cz>
7225 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7227 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
7228 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7230 (output_weakrefs): Update.
7232 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7233 Martin Jambor <mjambor@suse.cz>
7236 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7237 value that corresponds to the given aggval is found in values vector.
7239 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
7241 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7242 sse, sse2, sse3, ssse3 and sse4a flags to options.
7244 2013-05-17 David Malcolm <dmalcolm@redhat.com>
7246 * gengtype-state.c: (s_expr_writer): New class, to handle
7247 prettifying of output layout of s-expressions.
7248 (state_writer): New class, to write out gtype.state.
7249 (state_written_type_count): Move this variable into member data of
7251 (s_expr_writer::s_expr_writer): New code: constructor for new class
7252 (state_writer::state_writer(): ditto
7253 (s_expr_writer::write_new_line): New function
7254 (s_expr_writer::write_any_indent): ditto
7255 (s_expr_writer::begin_s_expr): ditto
7256 (s_expr_writer::end_s_expr): ditto
7257 (write_state_fileloc): convert to method of state_writer...
7258 (state_writer:: write_state_fileloc): ...and use methods of
7259 s_expr_writer to write indentation into the gtype.state output file
7260 to visually represent the hierarchical structure of the list
7262 (write_state_fields): ditto, renaming to...
7263 (state_writer::write_state_fields)
7264 (write_state_a_string): ditto, renaming to...
7265 (state_writer::write_state_a_string)
7266 (write_state_string_option): ditto, renaming to...
7267 (state_writer::write_state_string_option)
7268 (write_state_type_option): ditto, renaming to...
7269 (state_writer::write_state_type_option)
7270 (write_state_nested_option): ditto, renaming to...
7271 (state_writer::write_state_nested_option)
7272 (write_state_option): ditto, renaming to...
7273 (state_writer::write_state_option)
7274 (write_state_options): ditto, renaming to...
7275 (state_writer::write_state_options)
7276 (write_state_lang_bitmap): ditto, renaming to...
7277 (state_writer::write_state_lang_bitmap)
7278 (write_state_version): ditto, renaming to...
7279 (state_writer::write_state_version)
7280 (write_state_scalar_type): ditto, renaming to...
7281 (state_writer::write_state_scalar_type)
7282 (write_state_string_type): ditto, renaming to...
7283 (state_writer::write_state_string_type)
7284 (write_state_undefined_type): ditto, renaming to...
7285 (state_writer::write_state_undefined_type)
7286 (write_state_struct_union_type): ditto, renaming to...
7287 (state_writer::write_state_struct_union_type)
7288 (write_state_struct_type): ditto, renaming to...
7289 (state_writer::write_state_struct_type)
7290 (write_state_user_struct_type): ditto, renaming to...
7291 (state_writer::write_state_user_struct_type)
7292 (write_state_lang_struct_type): ditto, renaming to...
7293 (state_writer::write_state_lang_struct_type)
7294 (write_state_param_struct_type): ditto, renaming to...
7295 (state_writer::write_state_param_struct_type)
7296 (write_state_pointer_type): ditto, renaming to...
7297 (state_writer::write_state_pointer_type)
7298 (write_state_array_type): ditto, renaming to...
7299 (state_writer::write_state_array_type)
7300 (write_state_gc_used): ditto, renaming to...
7301 (state_writer::write_state_gc_used)
7302 (write_state_common_type_content): ditto, renaming to...
7303 (state_writer::write_state_common_type_content)
7304 (write_state_type): ditto, renaming to...
7305 (state_writer::write_state_type)
7306 (write_state_pair_list): ditto, renaming to...
7307 (state_writer::write_state_pair_list)
7308 (write_state_pair): ditto, renaming to...
7309 (state_writer::write_state_pair)
7310 (write_state_typedefs): ditto, renaming to...
7311 (state_writer::write_state_typedefs)
7312 (write_state_structures): ditto, renaming to...
7313 (state_writer::write_state_structures)
7314 (write_state_param_structs): ditto, renaming to...
7315 (state_writer::write_state_param_structs)
7316 (write_state_variables): ditto, renaming to...
7317 (state_writer::write_state_variables)
7318 (write_state_srcdir): ditto, renaming to...
7319 (state_writer::write_state_srcdir)
7320 (write_state_files_list): ditto, renaming to...
7321 (state_writer::write_state_files_list)
7322 (write_state_languages): ditto, renaming to...
7323 (state_writer::write_state_languages)
7324 (write_state): create a state_writer instance and use it when
7325 writing out the state file
7327 2013-05-17 Mike Stump <mikestump@comcast.net>
7329 PR rtl-optimization/57304
7330 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7331 accessing DF_REF_REAL_LOC.
7333 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7335 PR rtl-optimization/57281
7336 PR rtl-optimization/57300
7337 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7338 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7339 what the other splitter did if the registers are dead.
7341 2013-05-17 Richard Biener <rguenther@suse.de>
7343 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7346 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7348 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7351 2013-05-17 Marek Polacek <polacek@redhat.com>
7353 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7354 length when doing non-zero store of storing '\0' to '\0'.
7356 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7358 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7359 vect_external_def oprnd1 with loop_vinfo, try to emit
7360 optional cast, negation and and stmts on the loop preheader
7361 edge instead of into the pattern def seq.
7363 PR tree-optimization/57051
7364 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7365 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7367 2013-05-16 Nick Clifton <nickc@redhat.com>
7369 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7370 (rl78_is_naked_func): New function.
7371 (rl78_expand_prologue): Skip prologue generation for naked functions.
7372 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
7373 * doc/extend.texi (naked): Add RL78 to the list of processors
7374 that supports this attribute.
7376 2013-05-16 Jeff Law <law@redhat.com>
7378 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7380 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7382 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7383 cache parameters using detect_caches_amd also for CYRIX,
7384 NSC and TM2 signatures.
7386 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7387 Dzianis Kahanovich <mahatma@eu.by>
7391 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7392 VIA/Centaur processors and determine their cache parameters
7393 using detect_caches_amd.
7395 2013-05-16 Teresa Johnson <tejohnson@google.com>
7397 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7398 (rtl_verify_edges): New function.
7399 (rtl_verify_bb_insns): Ditto.
7400 (rtl_verify_bb_pointers): Ditto.
7401 (rtl_verify_bb_insn_chain): Ditto.
7402 (rtl_verify_fallthru): Ditto.
7403 (rtl_verify_bb_layout): Ditto.
7404 (rtl_verify_flow_info_1): Outline checks into new functions.
7405 (rtl_verify_flow_info): Ditto.
7407 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
7409 * cfghooks.c (copy_bbs): Add update_dominance argument.
7410 * cfghooks.h (copy_bbs): Update prototype.
7411 * tree-cfg.c (gimple_duplicate_sese_region):
7412 Add update_dominance argument.
7413 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7414 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7415 gimple_duplicate_sese_region call.
7416 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7417 Update copy_bbs call.
7418 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7419 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7421 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7423 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7424 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7425 vect_recog_rotate_pattern.
7426 (vect_recog_rotate_pattern): New function.
7428 2013-05-16 Jason Merrill <jason@redhat.com>
7430 * Makefile.in (LLINKER): New variable.
7431 (mostlyclean): Remove link mutex.
7432 * configure.ac: Handle --enable-link-mutex.
7433 * lock-and-run.sh: New script.
7435 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7438 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
7441 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7443 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
7445 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
7447 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
7448 * config/arm/arm.c (next_consecutive_mem): New function.
7449 (gen_movmem_ldrd_strd): Likewise.
7450 * config/arm/arm.md (movmemqi): Update condition and code.
7451 (unaligned_loaddi, unaligned_storedi): New patterns.
7453 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7455 * config.gcc: Obsolete *-*-solaris2.9*.
7456 * doc/install.texi (Specific, *-*-solaris2*): Document it.
7458 2013-05-16 Richard Biener <rguenther@suse.de>
7460 * passes.c (init_optimization_passes): Move pass_parallelize_loops
7461 earlier, after GRAPHITE transforms and IV canonicalization.
7463 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7465 * omp-low.c (extract_omp_for_data): For collapsed loops,
7466 if at least one of the loops is known at compile time to
7467 iterate zero times, set count to 0.
7468 (expand_omp_regimplify_p): New function.
7469 (expand_omp_for_generic): For collapsed loops, if at least
7470 one of the loops isn't known to iterate at least once,
7471 add runtime check with setting count to 0.
7472 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7473 For unsigned types if it isn't known at compile time that
7474 the loop will iterate at least once, add runtime check to bypass
7475 the whole loop if initial condition isn't true.
7477 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
7479 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
7481 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
7484 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
7485 transformations to avoid an infinite loop.
7487 2013-05-16 Marek Polacek <polacek@redhat.com>
7489 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
7491 2013-05-15 Leif Ekblad <leif@rdos.net>
7493 * config/i386/i386.c (ix86_decompose_address): Use
7494 DEFAULT_TLS_SEG_REG to access TLS segment register.
7495 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
7496 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
7497 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
7499 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
7502 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
7503 sibling calls to functions that would normally be lazily bound,
7504 unless $gp is call-clobbered.
7506 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
7508 * config/i386/i386.c (ix86_option_override_internal): Update
7509 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
7510 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
7511 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
7512 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
7514 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
7516 2013-05-15 Andreas Schwab <schwab@suse.de>
7518 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
7519 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
7522 2013-05-15 Teresa Johnson <tejohnson@google.com>
7524 * loop-unroll.c (report_unroll_peel): Check decision before
7525 emitting unroll/peel message.
7527 2013-05-15 Teresa Johnson <tejohnson@google.com>
7529 * function.h (has_bb_partition): New rtl_data flag.
7530 (bb_reorder_complete): Ditto.
7531 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
7532 instead of flag_reorder_blocks_and_partition.
7533 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
7534 with some enhancements.
7535 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
7536 * bb-reorder.c (connect_traces): Check for has_bb_partition
7537 instead of flag_reorder_blocks_and_partition.
7538 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
7539 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
7540 verify_hot_cold_block_grouping.
7541 (partition_hot_cold_basic_blocks): Set has_bb_partition.
7543 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7546 * config/arm/predicates.md (call_insn_operand): New predicate.
7547 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
7548 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
7549 if insn is not a tail call.
7550 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
7552 * config/arm/arm.h (enum reg_class): New caller save register class.
7553 (REG_CLASS_NAMES): Likewise.
7554 (REG_CLASS_CONTENTS): Likewise.
7555 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
7558 2013-05-15 Richard Biener <rguenther@suse.de>
7560 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
7561 of MSG_OPTIMIZED_LOCATIONS.
7562 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
7563 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
7565 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
7566 of MSG_OPTIMIZED_LOCATIONS.
7567 (execute_vect_slp): Likewise.
7568 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
7569 (vect_create_cond_for_alias_checks): Likewise.
7570 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
7571 (vect_recog_widen_mult_pattern): Likewise.
7572 (vect_recog_widen_sum_pattern): Likewise.
7573 (vect_recog_over_widening_pattern): Likewise.
7574 (vect_recog_widen_shift_pattern): Likewise.
7575 (vect_recog_vector_vector_shift_pattern): Likewise.
7576 (vect_recog_divmod_pattern): Likewise.
7577 (vect_recog_mixed_size_cond_pattern): Likewise.
7578 (vect_recog_bool_pattern): Likewise.
7579 (vect_pattern_recog_1): Likewise.
7581 2013-05-15 Martin Jambor <mjambor@suse.cz>
7583 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
7584 non-functions to builtin_unreachable.
7585 * ipa-inline-transform.c (inline_call): Do not assert estimates were
7586 correct when new direct edges were discovered.
7588 2013-05-15 Martin Jambor <mjambor@suse.cz>
7590 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
7591 header, print symbol order instead of node uid, print more information
7592 about indirect edge targets.
7593 (ipa_make_edge_direct_to_target): Print symbol order instead of node
7595 (ipa_make_edge_direct_to_target): Likewise.
7596 (remove_described_reference): Likewise.
7597 (propagate_controlled_uses): Likewise.
7598 (ipa_print_node_params): Also print symbol order.
7599 (ipcp_transform_function): Print symbol order instead of node uids.
7600 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
7601 (cgraph_get_create_real_symbol_node): Likewise.
7602 * ipa-cp.c (print_lattice): Likewise.
7603 (print_all_lattices): Likewise.
7604 (determine_versionability): Likewise.
7605 (initialize_node_lattices): Likewise.
7606 (estimate_local_effects): Likewise.
7607 (update_profiling_info): Likewise.
7608 (create_specialized_node): Likewise.
7609 (perhaps_add_new_callers): Likewise.
7610 (decide_about_value): Likewise.
7611 (decide_whether_version_node): Likewise.
7612 (identify_dead_nodes): Likewise.
7613 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
7614 (dump_inline_summary): Likewise.
7615 (estimate_node_size_and_time): Likewise.
7616 (inline_analyze_function): Likewise.
7617 * ipa-inline.c (report_inline_failed_reason): Likewise.
7618 (want_early_inline_function_p): Likewise.
7619 (edge_badness): Likewise.
7620 (update_edge_key): Likewise.
7621 (inline_small_functions): Likewise. Add dumping of order to two other
7623 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
7624 instead of node uids.
7625 (propagate_pure_const): Likewise.
7626 (propagate_pure_const): Likewise.
7627 * ipa-utils.c (dump_cgraph_node_set): Likewise.
7628 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
7629 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
7631 * tree-pretty-print.c (dump_function_header): Likewise.
7632 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
7633 Print symbol order instead of node uids.
7635 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7637 * config/s390/s390.c (s390_register_move_cost): Don't impose the
7638 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
7640 2013-05-15 Richard Biener <rguenther@suse.de>
7642 PR tree-optimization/57275
7643 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
7644 return value for fail to do runtime alias checks for gather loads.
7646 2013-05-15 Jan Hubicka <jh@suse.cz>
7650 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
7651 weakrefs are not external.
7652 (lto_symtab_merge_decls): Fix thinko when dealing with
7653 non-lto_symtab decls.
7654 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
7655 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
7656 * varpool.c (dump_varpool_node): Dump more flags.
7658 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7660 * config/i386/i386.c (processor_alias_table): Add instruction
7661 FSGSBASE for AMD bdver3 architecture.
7663 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7665 * tree.c (warn_deprecated_use): Print file:line using locus color.
7666 * diagnostic.c (diagnostic_report_current_module): Print file:line
7667 and file:line:column using locus color.
7669 2013-05-14 Mike Stump <mikestump@comcast.net>
7671 * gdbinit.in: Add __null.
7673 2013-05-14 Mike Stump <mikestump@comcast.net>
7675 * recog.h: Rename struct recog_data to Recog_data.
7676 * recog.c: Likewise.
7677 * reload.c (can_reload_into): Likewise.
7678 * config/picochip/picochip.c: Likewise.
7680 2013-05-14 Mike Stump <mikestump@comcast.net>
7682 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
7684 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
7686 * resource.h (struct resources): Remove unch_memory member.
7687 (CLEAR_RESOURCE): Don't clear unch_memory.
7688 * resource.c (mark_referenced_resources): Don't set it.
7689 (mark_set_resources): Likewise.
7690 (mark_target_live_regs): Don't clear it.
7691 (init_resource_info): Likewise.
7692 * reorg.c (resource_conflicts_p): Don't compare it.
7693 (redundant_insn): Don't set it.
7695 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
7697 * emit-rtl.c (next_label): Remove unused function.
7698 (skip_consecutive_labels, link_cc0_insns): Move to ...
7699 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
7700 only place where these functions are used, and make them static.
7702 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
7704 * fold-const.c (fold_negate_expr): Handle vectors.
7705 (fold_truth_not_expr): Make it static.
7706 (fold_invert_truthvalue): New static function.
7707 (invert_truthvalue_loc): Handle vectors. Do not call
7708 fold_truth_not_expr directly.
7709 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
7710 <TRUTH_NOT_EXPR>: Do not cast to boolean.
7711 (fold_comparison): Handle vector constants.
7712 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
7713 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
7714 * tree.h (fold_truth_not_expr): Remove declaration.
7716 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
7718 * config/aarch64/aarch64-simd.md
7719 (aarch64_vcond_internal<mode>): Rename to...
7720 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
7721 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
7722 float modes. Clarify all iterator modes.
7723 (vcond<mode><mode>): Use new name for vcond expanders.
7724 (vcond<v_cmp_result><mode>): Likewise.
7725 (vcondu<mode><mode>: Likewise.
7726 * config/aarch64/iterators.md (VDQF_COND): New.
7728 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
7731 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
7732 variable for the shift amount. Check that we shift by non-negative
7735 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
7738 * config/arm/arm.h (EPILOGUE_USES): Only return true
7739 for LR_REGNUM after epilogue_completed.
7741 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
7743 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
7746 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7749 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
7750 and Solaris 11+/x86 with gld.
7751 * configure: Regenerate.
7753 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7755 * expmed.c (expand_shift_1): Canonicalize rotates by
7756 constant bitsize / 2 to bitsize - 1.
7757 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
7758 case ROTATERT>: Likewise.
7761 2013-05-10 Jakub Jelinek <jakub@redhat.com>
7763 * config/i386/i386.md (rotateinv): New code attr.
7764 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7765 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7766 roll $31, %eax, etc.
7768 2013-05-14 Richard Biener <rguenther@suse.de>
7771 * tree-eh.c (sink_clobbers): Give up for successors with
7772 multiple predecessors and no virtual uses.
7774 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
7776 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
7777 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
7779 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7782 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
7783 the case when both op0 and op1 have VOIDmode.
7785 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7787 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
7788 in multiply-accumulate mode.
7790 2013-05-13 Guozhi Wei <carrot@google.com>
7792 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
7794 2013-05-13 Kai Tietz <ktietz@redhat.com>
7797 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
7798 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
7799 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
7800 * config/i386/i386.c (ix86_option_override_internal): Likewise.
7801 (ix86_expand_prologue): Likewise.
7802 (ix86_expand_split_stack_prologue): Likewise.
7803 (legitimate_pic_address_disp_p): Likewise.
7804 (legitimize_pic_address): Likewise.
7805 (legitimize_tls_address): Likewise.
7806 (legitimize_pe_coff_symbol): Likewise.
7807 (output_pic_addr_const): Likewise.
7808 (construct_plt_address): Likewise.
7809 (ix86_expand_call): Likewise.
7810 (x86_output_mi_thunk): Likewise.
7811 (x86_function_profiler): Likewise.
7813 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
7815 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
7816 similar switch cases.
7817 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
7818 (aarch64_simd_mov_to_<mode>low): Delete.
7819 (aarch64_simd_mov_to_<mode>high): Delete.
7820 (move_lo_quad_<mode>): Add w<-r alternative.
7821 (aarch64_simd_move_hi_quad_<mode>): Likewise.
7822 (aarch64_simd_mov_from_*): Update type attribute.
7823 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
7826 2013-05-13 Jan Hubicka <jh@suse.cz>
7828 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
7829 * config/i386/i386.c (ix86_compute_frame_layout,
7830 ix86_expand_epilogue, emit_i387_cw_initialization,
7831 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
7832 ix86_local_alignment): Fix use of size/speed predicates.
7834 2013-05-13 Jakub Jelinek <jakub@redhat.com>
7836 PR tree-optimization/45216
7837 PR tree-optimization/57157
7838 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
7839 the (-Y) & (B - 1) variant if OP is |.
7840 * expmed.c (expand_shift_1): For rotations by const0_rtx just
7841 return shifted. Use (-op1) & (prec - 1) as other_amount
7842 instead of prec - op1.
7844 2013-05-13 Martin Jambor <mjambor@suse.cz>
7847 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
7848 (ipa_constant_data): New type.
7849 (ipa_jump_func): Use ipa_constant_data to hold information about
7850 constant jump functions.
7851 (ipa_get_jf_constant): Adjust to jump function type changes.
7852 (ipa_get_jf_constant_rdesc): New function.
7853 (ipa_param_descriptor): New field controlled_uses.
7854 (ipa_get_controlled_uses): New function.
7855 (ipa_set_controlled_uses): Likewise.
7856 * ipa-ref.h (ipa_find_reference): Declare.
7857 * ipa-prop.c (ipa_cst_ref_desc): New type.
7858 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
7860 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
7861 New parameter cs. Adjust all callers.
7862 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
7863 (remove_described_reference): New function.
7864 (jfunc_rdesc_usable): Likewise.
7865 (try_make_edge_direct_simple_call): Decrement controlled use count,
7866 attempt to remove reference if it hits zero.
7867 (combine_controlled_uses_counters): New function.
7868 (propagate_controlled_uses): Likewise.
7869 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
7870 (ipa_edge_duplication_hook): Duplicate reference descriptions.
7871 (ipa_print_node_params): Print described use counter.
7872 (ipa_write_jump_function): Adjust to jump function type changes.
7873 (ipa_read_jump_function): New parameter CS, pass it to
7874 ipa_set_jf_constant. Adjust caller.
7875 (ipa_write_node_info): Stream controlled use count
7876 (ipa_read_node_info): Likewise.
7877 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
7879 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
7880 count. Remove cloning-added reference if it reaches zero.
7881 * ipa-ref.c (ipa_find_reference): New function.
7883 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7885 * config/i386/i386.c (processor_target_table): Modified default
7886 alignment values for AMD BD and BT architectures.
7888 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
7890 * tree-vect-generic.c (uniform_vector_p): Move ...
7891 * tree.c (uniform_vector_p): ... here.
7892 * tree.h (uniform_vector_p): Declare it.
7893 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
7896 2013-05-13 Jakub Jelinek <jakub@redhat.com>
7898 PR tree-optimization/57230
7899 * tree-ssa-strlen.c (handle_char_store): Record length for
7900 array store from STRING_CST.
7902 PR tree-optimization/57230
7903 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
7906 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
7908 * config/epiphany/epiphany.c (epiphany_init): Check size of
7909 NUM_MODES_FOR_MODE_SWITCHING.
7910 (epiphany_expand_prologue):
7911 Remove CONFIG_REGNUM initial value handling code.
7912 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
7913 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
7914 (emit_set_fp_mode, epiphany_mode_after): Likewise.
7915 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
7916 Don't return 1 for FP_MODE_NONE.
7917 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
7918 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
7919 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
7920 * config/epiphany/epiphany.md (save_config): New pattern.
7922 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
7924 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
7926 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
7928 * config/i386/i386.md (memory): Handle sseishft1.
7929 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
7930 (*vec_extractv2di_1): Ditto.
7932 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
7934 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
7937 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
7939 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
7940 Add mthumb/march=armv7-a multilib.
7941 Add mthumb/march=armv7-r multilib.
7942 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
7944 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
7946 * config/v850/t-rtems: Add more multilibs.
7948 2013-05-10 Richard Biener <rguenther@suse.de>
7950 PR tree-optimization/57214
7951 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
7952 not propagate from SSA names that occur in abnormal PHI nodes.
7954 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
7956 * stor-layout.c (element_precision): New function.
7957 * machmode.h (element_precision): Declare it.
7958 * tree.c (build_minus_one_cst): New function.
7959 (element_precision): Likewise.
7960 * tree.h (build_minus_one_cst): Declare new function.
7961 (element_precision): Likewise.
7962 * fold-const.c (operand_equal_p): Use element_precision.
7963 (fold_binary_loc): Handle vector types.
7964 * convert.c (convert_to_integer): Use element_precision.
7965 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
7968 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
7970 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
7971 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
7972 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
7973 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7974 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
7975 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
7976 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
7977 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7978 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
7979 (Uuw8): New constraints.
7980 (Usb4): Move into alphabetical order.
7981 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
7982 (sd8_operand, ub8_operand, uw8_operand): New predicates.
7983 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
7984 previously unnamed patterns.
7985 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
7986 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
7987 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
7988 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
7989 of set_attr_alternative/if_then_else. Use extended_mips16 instead
7990 of specific lengths.
7992 2013-05-10 Jakub Jelinek <jakub@redhat.com>
7994 * config/i386/i386.md (rotateinv): New code attr.
7995 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7996 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7997 roll $31, %eax, etc.
7999 PR tree-optimization/45216
8000 PR tree-optimization/57157
8001 * tree-ssa-forwprop.c (simplify_rotate): New function.
8002 (ssa_forward_propagate_and_combine): Call it.
8004 2013-05-10 Richard Biener <rguenther@suse.de>
8006 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
8007 disable peeling when we version for aliasing.
8008 (vector_alignment_reachable_p): Honor explicit user alignment.
8009 (vect_supportable_dr_alignment): Likewise.
8010 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
8011 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
8012 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
8013 then peeling to arrange for the cost-model check to come first.
8015 2013-05-10 Alan Modra <amodra@gmail.com>
8017 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
8018 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
8019 * configure: Regenerate.
8021 2013-05-10 Alan Modra <amodra@gmail.com>
8024 * varasm.c (default_elf_select_section): Move !DECL_P check..
8025 (get_named_section): ..to here before calling get_section_name.
8027 (default_section_type_flags): Add DECL_P check.
8028 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
8029 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
8031 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
8033 * config/epiphany/epiphany.c (epiphany_expand_prologue):
8034 When using gen_stack_adjust_str with a register offset, add a
8035 REG_FRAME_RELATED_EXPR note.
8037 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
8039 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
8040 (*vec_extractv4si_zext_mem): Ditto.
8041 (*vec_extractv2di): Add 0->x and x->x alternatives.
8042 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
8043 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
8045 2013-05-09 Jason Merrill <jason@redhat.com>
8047 N3639 C++1y VLA support
8048 * gimplify.c (gimplify_vla_decl): Don't touch an existing
8051 * tree.c (build_constructor_va): New.
8052 * tree.h: Declare it.
8054 2013-05-09 Martin Jambor <mjambor@suse.cz>
8057 * gimple-fold.c (canonicalize_constructor_val): Call
8058 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
8060 2013-05-09 Jan Hubicka <jh@suse.cz>
8061 Richard Biener <rguenther@suse.de>
8064 * symtab.c (symtab_make_decl_local): Do not add private names.
8066 2013-05-09 Jan Hubicka <jh@suse.cz>
8069 * symtab.c (insert_to_assembler_name_hash): Handle clones.
8070 (unlink_from_assembler_name_hash): Likewise.
8071 (symtab_prevail_in_asm_name_hash, symtab_register_node,
8072 symtab_unregister_node, symtab_initialize_asm_name_hash,
8073 change_decl_assembler_name): Update.
8075 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
8077 * config/aarch64/aarch64.md: New movtf split.
8078 (*movtf_aarch64): Update.
8079 (aarch64_movdi_tilow): Handle TF modes and rename to
8080 aarch64_movdi_<mode>low.
8081 (aarch64_movdi_tihigh): Handle TF modes and rename to
8082 aarch64_movdi_<mode>high
8083 (aarch64_movtihigh_di): Handle TF modes and rename to
8084 aarch64_mov<mode>high_di
8085 (aarch64_movtilow_di): Handle TF modes and rename to
8086 aarch64_mov<mode>low_di
8087 (aarch64_movtilow_tilow): Remove spurious whitespace.
8088 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8090 (aarch64_print_operand): Update.
8092 2013-05-09 Alan Modra <amodra@gmail.com>
8094 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8096 * configure: Regenerate.
8098 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8100 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8101 splitter preparation statements.
8102 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8103 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8106 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8108 * gimple-ssa-strength-reduction.c (count_candidates): Change
8109 return value to int.
8110 (analyze_candidates_and_replace): Change type of length to int.
8112 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8114 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8115 (*vec_extract<mode>): Use VI12_128 mode iterator.
8116 (*vec_extract<mode>_mem): Ditto.
8117 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8120 2013-05-08 Diego Novillo <dnovillo@google.com>
8125 2012-08-17 Diego Novillo <dnovillo@google.com>
8128 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8129 * config.in: Regenerate.
8130 * configure: Regenerate.
8131 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
8133 2013-05-08 Jan Hubicka <jh@suse.cz>
8136 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8137 * cgraph.h (symtab_node_base): Add unique_name.
8138 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8139 input_overwrite_node, input_varpool_node): Stream unique_name.
8140 * cgraphclones.c (cgraph_create_virtual_clone,
8141 cgraph_function_versioning): Set unique_name.
8142 * ipa.c (function_and_variable_visibility): Set unique_name.
8144 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8146 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8147 (alloc_cand_and_find_basis): Restrict conditional candidate
8148 processing to CAND_MULTs.
8150 2013-05-08 Jan Hubicka <jh@suse.cz>
8153 lto-symtab.c (lto_symtab_symbol_p): New function.
8154 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8155 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8156 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8157 Skip static symbols.
8159 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8161 PR tree-optimization/57200
8162 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8163 Only call inform if the preceding warning_at returns true.
8165 2013-05-07 Han Shen <shenhan@google.com>
8167 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8168 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8169 * common.opt (fstack-protector-strong): New option.
8170 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8171 * doc/invoke.texi (Optimization Options): Document
8172 "-fstack-protector-strong".
8173 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8175 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8177 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8179 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8181 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8182 it is not lhs of the insn.
8183 (setup_reg_equiv): Remove insn having equiv memory which it is not
8185 * lra-constraints.c (process_address): Try to improve generation
8186 code for address base + disp.
8187 (lra_constraints): Make correct the code for checking insn setting
8188 up backward equivalence. Remove insn only if it is in the init
8190 * lra-eliminations.c (update_reg_eliminate): Change return value.
8191 (lra_eliminate): Use the result.
8193 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
8195 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8196 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8197 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8198 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8199 register target operands.
8200 (*vec_extractv8hi_sse2): New pattern.
8201 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8202 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8203 (*vec_extract<mode>_mem): New insn and split pattern.
8205 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8207 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8208 (TARGET_ASAN_SHADOW_OFFSET): Define.
8209 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8210 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8212 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8214 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8215 (incr_vec_index): Return -1 if increment not found.
8216 (create_add_on_incoming_edge): Assert if increment not found.
8217 (record_increment): Limit number of increments recorded.
8218 (all_phi_incrs_profitable): Return false if an increment not found.
8219 (replace_profitable_candidates): Don't process increments that were
8221 (analyze_candidates_and_replace): Limit size of incr_vec.
8223 2013-05-07 Richard Biener <rguenther@suse.de>
8225 * calls.c (special_function_p): setjmp-like functions are leaf.
8226 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8227 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8229 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8231 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8232 (aarch64_simd_mov<mode>): New expander.
8233 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8234 (aarch64_simd_mov_to_<mode>high): Likewise.
8235 (aarch64_simd_mov_from_<mode>low): Likewise.
8236 (aarch64_simd_mov_from_<mode>high): Likewise.
8237 (aarch64_dup_lane<mode>): Update.
8238 (aarch64_dup_lanedi): New instruction pattern.
8239 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8240 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8242 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8244 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8245 (replace_mult_candidate): Remove unnecessary argument; remove
8246 unnecessary parameter from call to introduce_cast_before_cand.
8247 (replace_unconditional_candidate): Remove unnecessary parameter
8248 from call to replace_mult_candidate.
8249 (replace_conditional_candidate): Likewise.
8250 (insert_initializers): Use make_temp_ssa_name.
8251 (introduce_cast_before_cand): Remove unnecessary argument; use
8253 (replace_one_candidate): Remove unnecessary argument; remove
8254 unnecessary parameter from calls to introduce_cast_before_cand.
8255 (replace_profitable_candidates): Remove unnecessary parameters
8256 from calls to replace_one_candidate.
8258 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8260 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8261 phi def as possibly hiding a basis for a CAND_ADD whose operands
8262 have been commuted in the analysis.
8263 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8265 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8267 * config/aarch64/aarch64.md
8268 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8269 shift value between 0-4.
8271 2013-05-07 Richard Biener <rguenther@suse.de>
8273 * double-int.h (rshift): New overload.
8274 * double-int.c (rshift): New function.
8275 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8276 (create_reference_ops_from_ref): Remove.
8277 (vn_reference_insert): Use shared ops for constructing the
8278 reference and copy it.
8280 2013-05-07 Richard Biener <rguenther@suse.de>
8283 * tree-eh.c (sink_clobbers): Properly propagate
8284 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8286 2013-05-07 Jakub Jelinek <jakub@redhat.com>
8288 PR tree-optimization/57149
8289 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8290 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8291 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8292 uninit_undefined_value_p instead of ssa_undefined_value_p.
8295 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8296 for modifier == EXPAND_INITIALIZER.
8298 2013-05-07 Anton Blanchard <anton@samba.org>
8300 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8301 for powerpc64 little endian.
8302 * configure: Regenerate.
8304 2013-05-06 Graham Stott <grahams@btinternet.com>
8306 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8307 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8308 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8309 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8311 2013-05-06 Graham Stott <grahams@btinternet.com>
8313 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8314 codes which allow non-lvalues.
8316 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
8318 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8319 components are all 1s.
8320 (integer_minus_onep): New function.
8321 * tree.h (integer_minus_onep): Declare it.
8322 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8323 integer_minus_onep instead of integer_all_onesp.
8325 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8328 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8329 variations of these patterns.
8331 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8333 * config/i386/i386.md (isa): Add x64_sse4 member.
8334 (enabled): Handle x64_sse4.
8335 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8336 instruction for 64bit SSE4_1 targets. Update insn attributes.
8337 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8338 instruction for SSE4_1 targets. Update insn attributes.
8339 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8340 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8341 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8343 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8344 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8345 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8347 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8350 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8352 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8354 * final.c (do_assembler_dialects): Don't handle curly braces and
8355 vertical bar escaped by % as dialect delimiters.
8356 (output_asm_insn): Print curly braces and vertical bar if escaped
8357 by % and ASSEMBLER_DIALECT defined.
8358 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8359 * doc/tm.texi: Regenerated.
8361 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8363 * config/mips/mips.c: Include tree-pass.h.
8364 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8365 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8366 (pass_mips_machine_reorg2): New machine specific pass.
8367 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8368 (mips_option_override): Register the new pass.
8369 * rtl.h (cleanup_barriers): Remove prototype.
8370 (dbr_schedule): Likewise.
8371 * jump.c (cleanup_barriers): Make static.
8372 * reorg.c (dbr_schedule): Likewise.
8374 2013-05-06 Richard Biener <rguenther@suse.de>
8376 PR tree-optimization/57185
8377 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8380 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8383 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8384 between "first_arg" and "insn", not "last" and "insn".
8386 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8388 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8389 (find_candidates_in_block): Re-enable slsr_process_phi.
8390 (create_phi_basis): Fix double counting of candidate adjustment.
8392 2013-05-06 Richard Biener <rguenther@suse.de>
8395 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8396 the edge is also fallthru, preserve it and just clear the
8398 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8399 also complex, preserve that and just clear the fallthru flag.
8400 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8401 update virtual operands.
8403 2013-05-06 Alan Modra <amodra@gmail.com>
8405 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8406 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8407 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8408 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8409 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8410 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8412 2013-05-06 Alan Modra <amodra@gmail.com>
8414 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8415 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8416 (DEFAULT_ASM_ENDIAN): Define.
8417 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8418 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8419 Update -K PIC clause from sysv4.h.
8420 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8421 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8423 2013-05-06 Alan Modra <amodra@gmail.com>
8425 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8426 twice for little-endian.
8427 (ashrdi3_no_power, ashrdi3): Support little-endian.
8429 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8432 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
8433 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
8435 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
8436 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
8437 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
8440 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
8441 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
8443 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
8444 * config/pa/pa.opt: Make mbig-switch a no-op.
8445 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
8446 (CASE_VECTOR_MODE): Always return SImode.
8447 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
8448 for the !TARGET_BIG_SWITCH case.
8449 * config/pa/pa-linux.h: Likewise.
8450 * config/pa/pa-openbsd.h: Likewise.
8451 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
8452 * config/pa/pa.md (short_jump): Remove define_insn.
8453 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
8454 (casesi0): Remove define_insn.
8455 (type): Remove btable_branch.
8456 (pa_combine_type): Likewise.
8457 (in_nullified_branch_delay): Likewise.
8458 (in_call_delay): Likewise.
8459 (define_delay): Likewise.
8460 (define_insn_reservation "Z3"): Likewise.
8461 (define_insn_reservation "Z4"): Likewise.
8462 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
8463 (pa_adjust_insn_length): Remove adjustment for btable branches.
8464 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
8467 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
8469 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8470 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
8471 Add m->r,x alternatives.
8472 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
8473 splitters using SWI48x mode iterator.
8474 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
8475 TARGET_64BIT. Add m->x alternative.
8476 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
8477 Add o->x alternative. Enable for TARGET_SSE.
8478 (sse_storeq): Remove expander.
8479 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
8480 with memory input operand.
8481 (*vec_extractv2di_1 splitter): New.
8482 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
8483 * config/i386/i386.md (ssevecmodelower): New mode attribute.
8485 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8487 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
8488 (INT_LOWPART): Delete.
8489 (extract_MB): Adjust.
8490 (extract_ME): Adjust.
8491 (print_operand): Adjust.
8493 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8495 * config/rs6000/predicates.md (reg_or_add_cint_operand,
8496 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8497 (reg_or_logical_cint_operand, easy_fp_constant,
8498 logical_const_operand): Delete "CONST_DOUBLE" case.
8499 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
8500 "HOST_BITS_PER_WIDE_INT == 64" test.
8501 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
8502 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
8503 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
8504 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
8506 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
8507 CONST_DOUBLE DImode/VOIDmode case.
8508 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
8509 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8510 CONST_DOUBLE VOIDmode case.
8511 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8512 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
8513 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
8514 Delete CONST_DOUBLE case.
8515 (splitters for mov FMOVE64 const_double): Delete
8516 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8517 "HOST_BITS_PER_WIDE_INT >= 64" test.
8518 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
8520 (mov DI const_double): Delete.
8522 2013-05-04 Jakub Jelinek <jakub@redhat.com>
8524 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
8525 on op shows all bits zero in mode of a lowpart subreg, return zero.
8527 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8530 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
8531 to save TFmode registers and DImode to save TImode registers for
8532 caller save operations.
8533 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
8534 mark being partially clobbered since they only use the first
8537 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
8538 and TDmode only use the upper 64-bits of each VSX register.
8540 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8542 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
8543 (find_candidates_in_block): Disable slsr_process_phi.
8545 2013-05-03 Guozhi Wei <carrot@google.com>
8547 * coverage.c (coverage_obj_init): Move the construction of gcov
8549 (build_init_ctor): ... here.
8551 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8553 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
8554 (slsr_cand_d): Redefine def_phi.
8555 (stride_status, phi_adjust_status, count_phis_status): New enums.
8556 (find_phi_def): New.
8557 (find_basis_for_base_expr): New.
8558 (find_basis_for_candidate): Handle hidden bases.
8559 (alloc_cand_and_find_basis): Handle phi candidates.
8560 (slsr_process_phi): New.
8561 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
8562 (create_mul_imm_cand): Likewise.
8563 (create_add_ssa_cand): Exclude phi base candidates.
8564 (create_add_imm_cand): Likewise.
8565 (slsr_process_cast): Likewise.
8566 (slsr_process_copy): Likewise.
8567 (find_candidates_in_block): Handle phi candidates.
8568 (dump_candidate): Likewise.
8569 (unconditional_cands): Delete.
8570 (unconditional_cands_with_known_stride_p): Delete.
8571 (phi_dependent_cand_p): New.
8572 (cand_increment): Handle phi-dependent candidates.
8573 (replace_dependent): Delete.
8574 (replace_mult_candidate): New.
8575 (replace_unconditional_candidate): New.
8576 (incr_vec_index): Move to avoid forward reference.
8577 (create_add_on_incoming_edge): New.
8578 (create_phi_basis): New.
8579 (replace_dependents): Delete.
8580 (replace_conditional_candidate): New.
8581 (phi_add_costs): New.
8582 (replace_uncond_cands_and_profitable_phis): New.
8583 (record_increment): Handle phi adjustments.
8584 (record_phi_increments): New.
8585 (record_increments): Handle phi adjustments.
8586 (phi_incr_cost): New.
8587 (lowest_cost_path): Handle phis.
8588 (total_savings): Likewise.
8589 (analyze_increments): Likewise.
8590 (ncd_with_phi): New.
8591 (ncd_of_cand_and_phis): New.
8592 (nearest_common_dominator_for_cands): Handle phi increments.
8593 (all_phi_incrs_profitable): New.
8594 (replace_profitable_candidates): Handle phi-dependent candidates.
8595 (analyze_candidates_and_replace): Likewise.
8597 2013-05-03 Teresa Johnson <tejohnson@google.com>
8600 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
8601 do not exceed REG_BR_PROB_BASE.
8603 2013-05-03 Jeff Law <law@redhat.com>
8605 PR tree-optimization/57144
8606 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
8607 operand of the condition will bit into the new type when eliminating
8608 a cast feeding a condition.
8610 2013-05-03 Jakub Jelinek <jakub@redhat.com>
8612 PR rtl-optimization/57130
8613 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
8614 of COMPARE as in_code to the recursive call if needed.
8616 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
8618 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
8619 (enabled): Handle new members.
8620 * config/i386/sse.md (*vec_concatv2si): Merge from
8621 *vec_concatv2si_sse2 and vec_concatv2si_sse.
8622 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
8624 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
8626 PR tree-optimization/57027
8627 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
8628 for fnms opportunity, check we got the prerequisite kind
8629 of tree / gimple before using accessor functions.
8631 2013-05-03 Richard Biener <rguenther@suse.de>
8633 * double-int.h (lshift): New overload without precision
8635 (operator *=, operator +=, operator -=): Move ...
8636 * double-int.c (operator *=, operator +=, operator -=): ... here
8637 and implement more efficiently.
8638 (mul_double_with_sign): Remove.
8639 (lshift_double): Adjust to take unsinged shift argument, push
8640 dispatching code to callers.
8641 (mul_double_wide_with_sign): Add early out for callers that
8642 are not interested in high parts or overflow.
8643 (lshift): New function.
8644 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
8646 (lrotate, rrotate): Use logical shifts.
8647 * expr.c (get_inner_reference): Use lshift.
8648 * fixed-value.c (do_fixed_divide): Likewise.
8649 * tree-dfa.c (get_ref_base_and_extent): Likewise.
8650 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
8651 (indirect_refs_may_alias_p): Likewise.
8652 (stmt_kills_ref_p_1): Likewise.
8654 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8656 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
8658 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8660 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
8661 scalar form of FABD instruction.
8663 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
8665 * lra-constraints.c (process_alt_operands): Add checking alt
8666 number to choose the best alternative.
8668 2013-05-02 Richard Biener <rguenther@suse.de>
8670 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
8671 bitmap and its handling.
8672 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
8674 2013-05-02 Richard Biener <rguenther@suse.de>
8677 * tree-inline.c (copy_loops): Properly handle removed loops.
8678 (copy_cfg_body): Mark destination loops for fixup if source
8681 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
8684 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
8685 generating simple_return for naked functions.
8687 2013-05-02 Martin Jambor <mjambor@suse.cz>
8690 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
8691 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
8693 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
8694 ipa_agg_replacement_value structures.
8695 (known_aggs_to_agg_replacement_list): Likewise.
8696 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
8697 (read_agg_replacement_chain): Likewise.
8698 (ipcp_transform_function): Also check that by_ref flags match.
8700 2013-05-02 Richard Biener <rguenther@suse.de>
8702 * graphds.h (struct graph): Add obstack member.
8703 * graphds.c (new_graph): Initialize obstack and allocate
8705 (add_edge): Allocate edge from the obstack.
8706 (free_graph): Free the obstack instead of all edges and vertices.
8708 2013-05-02 Teresa Johnson <tejohnson@google.com>
8710 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
8712 * cfg.c (update_bb_profile_for_threading): Ditto.
8713 * tree-inline.c (copy_bb): Ditto.
8714 (copy_edges_for_bb): Ditto.
8715 (initialize_cfun): Ditto.
8716 (copy_cfg_body): Ditto.
8717 (expand_call_inline): Ditto.
8718 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
8719 (estimate_node_size_and_time): Ditto.
8720 (inline_merge_summary): Ditto.
8721 * cgraphclones.c (cgraph_clone_edge): Ditto.
8722 (cgraph_clone_node): Ditto.
8723 * sched-rgn.c (compute_dom_prob_ps): Ditto.
8724 (compute_trg_info): Ditto.
8726 2013-05-02 Ian Bolton <ian.bolton@arm.com>
8728 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
8729 S reg when fp attribute set.
8730 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
8732 2013-05-02 Ian Bolton <ian.bolton@arm.com>
8734 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
8736 (*and_one_cmplsi3_compare0_uxtw): Likewise.
8737 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
8738 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
8740 2013-05-02 Richard Biener <rguenther@suse.de>
8742 * tree-scalar-evolution.c (scev_info_hasher): Remove.
8743 (struct instantiate_cache_entry): New type.
8744 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
8745 (struct instantiate_cache_type): New type.
8746 (set_instantiated_value, get_instantiated_value): Remove.
8747 (get_instantiated_value_entry): New function.
8748 (instantiate_scev_name): Use the new cache and adjust.
8749 (instantiate_scev_poly): Adjust.
8750 (instantiate_scev_binary): Likewise.
8751 (instantiate_array_ref): Likewise.
8752 (instantiate_scev_convert): Likewise.
8753 (instantiate_scev_not): Likewise.
8754 (instantiate_scev_3): Likewise.
8755 (instantiate_scev_2): Likewise.
8756 (instantiate_scev_r): Likewise.
8757 (instantiate_scev): Likewise.
8758 (resolve_mixers): Likewise.
8760 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
8763 * lra-constraints.c (best_small_class_operands_num): Remove.
8764 (process_alt_operands): Remove small_class_operands_num. Take
8765 small classes operands into losers and only if the operand is not
8766 matched. Modify debugging output.
8767 (curr_insn_transform): Remove best_small_class_operands_num.
8770 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8772 * config/aarch64/aarch64-builtins.c
8773 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
8774 * config/aarch64/aarch64-simd-builtins.def
8775 (reduc_splus_): Add new modes.
8776 (reduc_uplus_): New.
8777 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
8778 (reduc_uplus_v4sf): Likewise.
8779 (reduc_splus_v4sf): Likewise.
8780 (aarch64_addv<mode>): Likewise.
8781 (reduc_uplus_<mode>): Likewise.
8782 (reduc_splus_<mode>): Likewise.
8783 (aarch64_addvv2di): Likewise.
8784 (reduc_uplus_v2di): Likewise.
8785 (reduc_splus_v2di): Likewise.
8786 (aarch64_addvv2si): Likewise.
8787 (reduc_uplus_v2si): Likewise.
8788 (reduc_splus_v2si): Likewise.
8789 (reduc_<sur>plus_<mode>): New.
8790 (reduc_<sur>plus_v2di): Likewise.
8791 (reduc_<sur>plus_v2si): Likewise.
8792 (reduc_<sur>plus_v4sf): Likewise.
8793 (aarch64_addpv4sf): Likewise.
8794 * config/aarch64/arm_neon.h
8795 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
8796 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
8797 add UNSPEC_SADDV, UNSPEC_UADDV.
8799 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
8801 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8803 * config/aarch64/arm_neon.h
8804 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
8806 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8808 * config/aarch64/aarch64-builtins
8809 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
8811 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8813 * config/aarch64/aarch64-simd-builtins.def
8815 (reduc_smin_): Likewise.
8816 (reduc_umax_): Likewise.
8817 (reduc_umin_): Likewise.
8818 (reduc_smax_nan_): Likewise.
8819 (reduc_smin_nan_): Likewise.
8822 (smax): Update for V2SF, V4SF and V2DF modes.
8825 (smin_nan): Likewise.
8826 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
8827 (<su><maxmin><mode>3): ...This, refactor.
8828 (s<maxmin><mode>3): New.
8829 (<maxmin_uns><mode>3): Likewise.
8830 (reduc_<maxmin_uns>_<mode>): Refactor.
8831 (reduc_<maxmin_uns>_v4sf): Likewise.
8832 (reduc_<maxmin_uns>_v2si): Likewise.
8833 (aarch64_<fmaxmin><mode>: Remove.
8834 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
8836 (vmin<q>_f<32,64>): Likewise.
8837 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8839 (su): Add mappings for smax, smin, umax, umin.
8841 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8842 (FMAXMIN): Rename as...
8843 (FMAXMIN_UNS): ...This.
8845 (fmaxminv): Likewise.
8846 (fmaxmin): Likewise.
8848 (maxmin_uns_op): Likewise.
8850 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8852 * config/aarch64/arm_neon.h
8853 (vac<ge, gt><sd>_f<32, 64>): Rename to...
8854 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
8855 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
8857 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8859 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
8860 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
8862 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8864 * config/aarch64/aarch64-simd.md
8865 (vcond<mode>_internal): Handle special cases for constant masks.
8866 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
8867 (vcondu<mode><mode>): Likewise.
8868 (vcond<v_cmp_result><mode>): New.
8870 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8872 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
8873 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
8874 * config/aarch64/aarch64-simd-builtins.def
8875 (cmeq): Update to BUILTIN_VALLDI.
8880 * config/aarch64/arm_neon.h
8881 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
8882 to builtins or C as appropriate.
8884 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8886 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
8888 (cmhi): Rename to...
8890 * config/aarch64/aarch64-simd.md
8891 (simd_mode): Add SF.
8892 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
8893 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
8894 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
8895 (cstore<mode>_neg): ...This.
8896 * config/aarch64/iterators.md
8898 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
8900 (UCOMPARISONS): Likewise.
8901 (optab): Add missing comparisons.
8909 (V_cmp_result): Add DF, SF modes.
8910 (v_cmp_result): Likewise.
8913 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
8915 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
8917 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
8918 define_insn to define_insn_and_split.
8919 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
8920 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
8921 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
8922 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
8923 (thumb2_negscc): Likewise.
8925 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8927 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
8929 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8931 * config/arm/thumb2.md: Remove trailing whitespaces.
8933 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8935 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
8936 Use gen_int_mode rather than GEN_INT.
8938 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
8940 * value-prof.c (stream_in_histogram_value): Remove the strayed
8943 2013-04-30 Richard Biener <rguenther@suse.de>
8946 * cfghooks.c (split_edge): Properly check for the loop latch edge.
8948 2013-04-30 Richard Biener <rguenther@suse.de>
8951 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
8953 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
8955 PR rtl-optimization/56957
8956 PR rtl-optimization/57105
8957 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
8958 variable. Use just INSN_UID for determining whether an insn
8959 should be only disconnected from the insn stream.
8960 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
8962 2013-04-30 Jakub Jelinek <jakub@redhat.com>
8964 PR tree-optimization/57104
8965 * tsan.c (instrument_expr): Don't instrument accesses to
8966 DECL_HARD_REGISTER VAR_DECLs.
8968 2013-04-30 Richard Biener <rguenther@suse.de>
8970 * function.h (loops_for_fn): New inline function.
8971 (set_loops_for_fn): Likewise.
8972 * cfgloop.h (place_new_loop): Add struct function parameter.
8973 (get_loop): Likewise.
8974 (get_loops): Likewise.
8975 (number_of_loops): Likewise.
8977 (fel_init): Likewise.
8978 * cfg.c (get_loop_copy): Adjust.
8979 * cfgloop.c (flow_loops_dump): Likewise.
8980 (record_loop_exits): Likewise.
8981 (verify_loop_structure): Likewise.
8982 * cfgloopanal.c (mark_irreducible_loops): Likewise.
8983 (estimate_reg_pressure_cost): Likewise.
8984 (mark_loop_exit_edges): Likewise.
8985 * cfgloopmanip.c (place_new_loop): Likewise.
8986 (add_loop): Likewise.
8987 (duplicate_loop): Likewise.
8988 * graph.c (draw_cfg_nodes): Likewise.
8989 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
8990 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
8991 (extract_affine_chrec): Likewise.
8992 (build_scop_iteration_domain): Likewise.
8993 * graphite.c (graphite_initialize): Likewise.
8994 * ira-build.c (create_loop_tree_nodes): Likewise.
8995 (more_one_region_p): Likewise.
8996 (rebuild_regno_allocno_maps): Likewise.
8997 (mark_loops_for_removal): Likewise.
8998 (mark_all_loops_for_removal): Likewise.
8999 (remove_unnecessary_regions): Likewise.
9000 (ira_build): Likewise.
9001 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
9002 * loop-init.c (fix_loop_structure): Likewise.
9003 (gate_rtl_move_loop_invariants): Likewise.
9004 (gate_rtl_unswitch): Likewise.
9005 (gate_rtl_unroll_and_peel_loops): Likewise.
9006 (rtl_doloop): Likewise.
9007 * lto-streamer-in.c (input_cfg): Likewise.
9008 * lto-streamer-out.c (output_cfg): Likewise.
9009 * modulo-sched.c (sms_schedule): Likewise.
9010 * predict.c (tree_estimate_probability): Likewise.
9011 (tree_estimate_probability_driver): Likewise.
9012 (estimate_loops): Likewise.
9013 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
9014 (move_sese_region_to_fn): Likewise.
9015 (debug_loop_num): Likewise.
9016 * tree-chrec.c (chrec_evaluate): Likewise.
9017 (hide_evolution_in_other_loops_than_loop): Likewise.
9018 (chrec_component_in_loop_num): Likewise.
9019 (reset_evolution_in_loop): Likewise.
9020 (evolution_function_is_invariant_rec_p): Likewise.
9021 * tree-if-conv.c (main_tree_if_conversion): Likewise.
9022 * tree-inline.c (copy_loops): Likewise.
9023 (copy_cfg_body): Likewise.
9024 (tree_function_versioning): Likewise.
9025 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
9026 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
9028 (add_to_evolution_1): Likewise.
9029 (scev_const_prop): Likewise.
9030 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
9031 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
9032 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
9033 (tree_ssa_lim_initialize): Likewise.
9034 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
9035 (verify_loop_closed_ssa): Likewise.
9036 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
9037 (tree_ssa_loop_im): Likewise.
9038 (tree_ssa_loop_unswitch): Likewise.
9039 (tree_vectorize): Likewise.
9040 (check_data_deps): Likewise.
9041 (tree_ssa_loop_ivcanon): Likewise.
9042 (tree_ssa_loop_bounds): Likewise.
9043 (tree_complete_unroll): Likewise.
9044 (tree_complete_unroll_inner): Likewise.
9045 (tree_parallelize_loops): Likewise.
9046 (tree_ssa_loop_prefetch): Likewise.
9047 (tree_ssa_loop_ivopts): Likewise.
9048 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
9049 * tree-vectorizer.c (vectorize_loops): Likewise.
9051 2013-04-29 Mike Frysinger <vapier@gentoo.org>
9053 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
9054 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
9055 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
9056 with EABI_LINK_SPEC.
9058 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9061 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
9064 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
9067 * lra-constraints.c (process_alt_operands): Discourage a bit more
9068 using memory for pseudos. Print cost dump for alternatives.
9069 Modify cost values for conflicts with early clobbers.
9070 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
9072 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9075 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
9077 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9079 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
9081 (movdi_aarch64): Support LDR/STR from/to D register.
9083 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9085 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
9086 or higher by default.
9088 2013-04-29 Richard Biener <rguenther@suse.de>
9091 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9092 even if not adding abnormal edges for calls that can make
9095 2013-04-29 Richard Biener <rguenther@suse.de>
9098 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9100 (move_stmt_r): Remove redundant checking.
9102 2013-04-29 Teresa Johnson <tejohnson@google.com>
9105 * basic-block.h (apply_scale): New function.
9106 (apply_probability): Use apply_scale.
9107 * gimple-streamer-in.c (input_bb): Ditto.
9108 * lto-streamer-in.c (input_cfg): Ditto.
9109 * lto-cgraph.c (merge_profile_summaries): Ditto.
9110 * tree-optimize.c (execute_fixup_cfg): Ditto.
9111 * tree-inline.c (copy_bb): Update comment to use apply_scale.
9112 (copy_edges_for_bb): Ditto.
9113 (copy_cfg_body): Ditto.
9115 2013-04-29 Tom de Vries <tom@codesourcery.com>
9117 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9118 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9119 (tail_merge_optimize): Handle current_loops == NULL.
9121 2013-04-26 Jeff Law <law@redhat.com>
9123 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9124 (simplify_cond_using_ranges): Generalize code to simplify
9125 COND_EXPRs where one argument is a constant and the other
9126 is an SSA_NAME created by an integral type conversion.
9128 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9130 * config/arm/arm.md (store_minmaxsi): Use only when
9131 optimize_insn_for_size_p.
9133 2013-04-29 Christian Bruel <christian.bruel@st.com>
9136 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9138 2013-04-29 Richard Biener <rguenther@suse.de>
9141 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9142 loop tree make sure to schedule a fixup for the child as well.
9143 (expand_omp_for_generic): Properly add loops.
9144 (expand_omp_for_static_nochunk): Likewise.
9145 (expand_omp_for_static_chunk): Likewise.
9146 (expand_omp_for): For the degenerate case fixup loops.
9147 (expand_omp_sections): Fix default bb placement in loops.
9148 (expand_omp_atomic_pipeline): Properly add loops.
9150 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9152 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9154 2013-04-29 Tom de Vries <tom@codesourcery.com>
9156 * tree-ssa-tail-merge.c: Update header comment.
9158 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9160 * config/aarch64/arm_neon.h
9161 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9162 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9163 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9164 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9165 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9166 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9167 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9168 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9170 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9172 * config/aarch64/aarch64-simd.md
9173 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9174 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9175 fix_trunc, fixuns_trunc.
9176 (ftrunc<VDQF:mode>2): New.
9177 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9178 (fix_trunc_optab): New.
9180 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9182 * config/aarch64/aarch64-builtins.c
9183 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9184 iceilf, lround, iroundf.
9186 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9189 * config/i386/i386.h (enum ix86_tune_indices)
9190 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9191 New, split from X86_TUNE_INTER_UNIT_MOVES.
9192 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9193 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9194 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9195 (TARGET_INTER_UNIT_MOVES): Remove.
9196 * config/i386/i386.c (initial_ix86_tune_features): Update.
9197 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9198 (ix86_expand_convert_uns_didf_sse): Use
9199 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9200 (ix86_expand_vector_init_one_nonzero): Ditto.
9201 (ix86_expand_vector_init_interleave): Ditto.
9202 (inline_secondary_memory_needed): Return true for moves from SSE class
9203 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9204 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9205 * config/i386/constraints.md (Yi, Ym): Depend on
9206 TARGET_INTER_UNIT_MOVES_TO_VEC.
9207 (Yj, Yn): New constraints.
9208 * config/i386/i386.md (*movdi_internal): Change constraints of
9209 operand 1 from Yi to Yj and from Ym to Yn.
9210 (*movsi_internal): Ditto.
9211 (*movdf_internal): Ditto.
9212 (*movsf_internal): Ditto.
9213 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9214 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9215 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9216 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9217 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9218 * config/i386/sse.md (movdi_to_sse): Ditto.
9219 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9220 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9221 TARGET_INTER_UNIT_MOVES.
9222 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9223 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9224 instead of TARGET_INTER_UNIT_MOVES.
9225 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9226 operand 1 from Yi to Yj and from Ym to Yn.
9228 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9230 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9231 (float_truncate_hi_): Likewise.
9232 (float_extend_lo_): Likewise.
9233 (float_truncate_lo_): Likewise.
9234 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9235 (aarch64_float_extend_lo_v2df): Likewise.
9236 (vec_unpacks_hi_v4sf): Likewise.
9237 (aarch64_float_truncate_lo_v2sf): Likewise.
9238 (aarch64_float_truncate_hi_v4sf): Likewise.
9239 (vec_pack_trunc_v2df): Likewise.
9240 (vec_pack_trunc_df): Likewise.
9242 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9244 * config/aarch64/aarch64-builtins.c
9245 (aarch64_fold_builtin): Fold float conversions.
9246 * config/aarch64/aarch64-simd-builtins.def
9247 (floatv2si, floatv4si, floatv2di): New.
9248 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9249 * config/aarch64/aarch64-simd.md
9250 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9251 * config/aarch64/iterators.md (FLOATUORS): New.
9252 (optab): Add float, floatuns.
9253 (su_optab): Likewise.
9255 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9257 * config/aarch64/aarch64-builtins.c
9258 (aarch64_builtin_vectorized_function): Use new names for
9260 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9261 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9262 (fcvtzu): Split as...
9263 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9264 (fcvtas): Split as...
9265 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9266 (fcvtau): Split as...
9267 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9268 (fcvtps): Split as...
9269 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9270 (fcvtpu): Split as...
9271 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9272 (fcvtms): Split as...
9273 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9274 (fcvtmu): Split as...
9275 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9276 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9277 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9278 (lfrintnusf, lfrintnudf): Likewise.
9279 * config/aarch64/aarch64-simd.md
9280 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9282 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9283 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9284 (fcvt_pattern): Likewise.
9286 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9288 * config/aarch64/aarch64-simd.md
9289 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9290 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9292 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9294 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9295 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9296 (vrnd<a,m,n,p>_f32): Implement using builtins.
9297 (vrnd<i,x><q>_f<32, 64>): New.
9299 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9301 * config/aarch64/aarch64-builtins.c
9302 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9303 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9304 (frintz): Rename to...
9306 (frintp): Rename to...
9308 (frintm): Rename to...
9310 (frinti): Rename to...
9311 (nearbyint): ...this.
9312 (frintx): Rename to...
9314 (frinta): Rename to...
9316 * config/aarch64/aarch64-simd.md
9317 (aarch64_frint<frint_suffix><mode>): Delete.
9318 (<frint_pattern><mode>2): Convert to insn.
9319 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9320 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9321 (frint_pattern): Likewise.
9322 (frint_suffix): Likewise.
9324 2013-04-29 Richard Biener <rguenther@suse.de>
9326 PR tree-optimization/57081
9327 * loop-init.c: Include tree-flow.h.
9328 (loop_optimizer_finalize): Free number of iteration estimates.
9329 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9331 2013-04-29 Jakub Jelinek <jakub@redhat.com>
9333 PR tree-optimization/57083
9334 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9335 non-singleton shift count range, zero extend low_bound for uns case.
9337 * config/i386/predicates.md (general_vector_operand): New predicate.
9338 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9339 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9340 if they aren't nonimmediate operands. If their original values
9341 satisfy const_vector_equal_evenodd_p, don't shift them.
9342 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9343 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9344 if not nonimmediate_operand.
9345 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9346 instead of register_operand.
9347 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9349 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
9351 * stor-layout.c (finalize_size_functions): Allocate a structure and
9352 reset cfun before dumping the functions.
9354 2013-04-27 Jakub Jelinek <jakub@redhat.com>
9356 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9359 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9360 use xop_pmacsdqh if uns_p.
9361 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9362 the immediate rotate count.
9364 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9366 * rtl.h (struct rtx_def): Add comment for field jump.
9367 (LRA_SUBREG_P): New macro.
9368 * recog.c (register_operand): Check LRA_SUBREG_P.
9369 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9371 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9372 (lra_final_code_change): Skip subreg change for operators.
9373 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9374 if there are no operand changes.
9375 * lra-constraints.c (curr_insn_set): New.
9376 (match_reload): Set LRA_SUBREG_P.
9377 (emit_spill_move): Ditto.
9378 (check_and_process_move): Use curr_insn_set. Process only single
9379 set insns. Don't initialize sec_mem_p and change_p.
9380 (simplify_operand_subreg): Use LRA_SUBREG_P.
9381 (reg_in_class_p): New function.
9382 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9383 of #ifdef. Add code to remove cycling.
9384 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9385 non-null disp. Reload inner instead of disp when base and index
9386 are null. Try to put lo_sum into register.
9387 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9388 (check_and_process_move): Move code for move cost check to
9389 simple_move_p. Remove equiv_substitution.
9390 (simple_move_p): New function.
9391 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9392 curr_insn_set. Call check_and_process_move only for single set
9393 insns. Use the new function. Move call of check_and_process_move
9394 after operand equiv substitution and address process.
9396 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9399 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9400 with nonlocal goto receivers or returns twice calls, ignore
9401 unininitialized values from abnormal edges to nl goto receiver
9402 or returns twice call.
9404 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9406 PR tree-optimization/57051
9407 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9408 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9411 2013-04-26 Richard Biener <rguenther@suse.de>
9413 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9414 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9415 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9416 (fixup_loop_arrays_after_move): New function.
9417 (move_sese_region_to_fn): Properly outline the loop tree parts
9420 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
9422 * config/i386/i386.md (type, unit): Fix long lines.
9424 2013-04-26 Richard Biener <rguenther@suse.de>
9426 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9427 (lto-streamer-out.o): Likewise.
9428 * cfgloop.c (init_loops_structure): Export, add struct function
9429 argument and adjust.
9430 (flow_loops_find): Adjust.
9431 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9432 (init_loops_structure): Declare.
9433 * lto-streamer-in.c: Include cfgloop.h.
9434 (input_cfg): Input the loop tree.
9435 * lto-streamer-out.c: Include cfgloop.h.
9436 (output_cfg): Output the loop tree.
9437 (output_struct_function_base): Do not drop PROP_loops.
9439 2013-03-26 Richard Biener <rguenther@suse.de>
9441 * tree-cfg.c (execute_build_cfg): Build the loop tree.
9442 (pass_build_cfg): Provide PROP_loops.
9443 (move_sese_region_to_fn): Remove loops that are outlined into fn
9445 * tree-inline.c: Include cfgloop.h.
9446 (initialize_cfun): Do not drop PROP_loops.
9447 (copy_loops): New function.
9448 (copy_cfg_body): Copy loop structure.
9449 (tree_function_versioning): Initialize destination loop tree.
9450 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
9451 (pass_parallelize_loops): Do IL verification.
9452 * loop-init.c (loop_optimizer_init): Fixup loops if required.
9453 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
9454 the CFG make sure we fixup loops as well.
9455 * tree-ssa-tail-merge.c: Include cfgloop.h.
9456 (replace_block_by): When merging loop latches mark loops for fixup.
9457 * lto-streamer-out.c (output_struct_function_base): Drop
9459 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
9460 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
9461 * ipa-split.c: Include cfgloop.h.
9462 (split_function): Add the new return block to the loop tree root.
9463 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
9464 whether we have removed the forwarder block.
9465 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
9466 * cfgloop.h (place_new_loop): Declare.
9467 * cfgloopmanip.c (place_new_loop): Export.
9468 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
9469 (tree-switch-conversion.o): Likewise.
9470 (tree-complex.o): Likewise.
9471 (tree-inline.o): Likewise.
9472 (tree-ssa-tailmerge.o): Likewise.
9473 (ipa-split.o): Likewise.
9474 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
9475 (tree-ssa-copy.o): Likewise.
9476 * tree-switch-conversion.c: Include cfgloop.h
9477 (process_switch): If we emit a bit-test cascade, schedule loops
9479 * tree-complex.c: Include cfgloop.h.
9480 (expand_complex_div_wide): Properly add new basic-blocks to loops.
9481 * asan.c: Include cfgloop.h.
9482 (create_cond_insert_point): Properly add new basic-blocks to
9483 loops, schedule loop fixup.
9484 * cfgloop.c (verify_loop_structure): Check that looks are not
9486 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
9488 (expand_omp_for_generic): Likewise.
9489 (expand_omp_sections): Likewise.
9490 (expand_omp_atomic_pipeline): Schedule loops for fixup.
9491 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
9492 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
9493 is initialized, not when loops are present.
9494 * tree-parloops.c (parallelize_loops): Remove checking here.
9495 * passes.c (init_optimization_passes): Schedule a copy-propagation
9496 pass before complete unrolling of inner loops.
9498 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9500 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
9501 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
9502 (colorize_init): Add argument to _WIN32 version.
9503 * toplev.c: Include diagnostic-color.h.
9504 (process_options): Default to -fdiagnostics-color=auto if
9505 GCC_COLORS env var is in the environment.
9506 * common.opt (fdiagnostics-color=): Add Var and Init.
9507 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
9508 env var is in the environment, the default is auto rather than never.
9510 * diagnostic.h (file_name_as_prefix): Add context argument.
9511 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
9512 the string as locus.
9513 * langhooks.c (lhd_print_error_function): Adjust caller.
9515 2013-04-25 Lawrence Crowl <crowl@google.com>
9517 * var-tracking.c (shared_hash_def::htab):
9518 Change type to hash_table. Update dependent calls and types.
9520 2013-04-25 Lawrence Crowl <crowl@google.com>
9522 * Makefile.in: Update as needed below.
9524 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
9525 Move declaration to after the type's method definitons.
9527 * attribs.c (htab_t scoped_attributes::attribute_hash):
9528 Change type to hash_table. Update dependent calls and types.
9530 * bitmap.c (htab_t bitmap_desc_hash):
9531 Change type to hash_table. Update dependent calls and types.
9533 * cselib.c (htab_t cselib_hash_table):
9534 Change type to hash_table. Update dependent calls and types.
9536 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
9537 (hash_string_slot_node): Move implementation into lto-streamer.h
9538 struct string_slot_hasher.
9539 (eq_string_slot_node): Likewise.
9541 * data-streamer-out.c: Update output_block::string_hash_table
9542 dependent calls and types.
9544 * dwarf2cfi.c (htab_t trace_index):
9545 Change type to hash_table. Update dependent calls and types.
9547 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
9548 Change type to hash_table. Update dependent calls and types.
9549 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
9550 (htab_t optimize_external_refs::map): Likewise.
9551 (htab_t output_comp_unit::extern_map): Likewise.
9552 (htab_t output_comdat_type_unit::extern_map): Likewise.
9553 (htab_t output_macinfo::macinfo_htab): Likewise.
9554 (htab_t optimize_location_lists::htab): Likewise.
9555 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
9557 * except.c (htab_t ehspec_hash_type):
9558 Change type to hash_table. Update dependent calls and types.
9559 (assign_filter_values::ttypes): Likewise.
9560 (assign_filter_values::ehspec): Likewise.
9561 (sjlj_assign_call_site_values::ar_hash): Likewise.
9562 (convert_to_eh_region_ranges::ar_hash): Likewise.
9564 * gcse.c (htab_t pre_ldst_table):
9565 Change type to hash_table. Update dependent calls and types.
9567 * ggc-common.c (htab_t saving_htab):
9568 Change type to hash_table. Update dependent calls and types.
9569 (htab_t loc_hash): Likewise.
9570 (htab_t ptr_hash): Likewise.
9571 (call_count): Rename ggc_call_count.
9572 (call_alloc): Rename ggc_call_alloc.
9573 (loc_descriptor): Rename make_loc_descriptor.
9574 (add_statistics): Rename ggc_add_statistics.
9576 * ggc-common.c (saving_htab):
9577 Change type to hash_table. Update dependent calls and types.
9579 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
9580 (push_gimplify_context): Likewise.
9581 (pop_gimplify_context): Likewise.
9582 (struct gimple_temp_hash_elt): Added.
9583 (struct gimplify_hasher): Likewise.
9584 (struct gimplify_ctx.temp_htab):
9585 Change type to hash_table. Update dependent calls and types.
9587 * gimple-fold.c: Include gimplify-ctx.h.
9589 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
9590 Change type to hash_table. Update dependent calls and types.
9591 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
9592 avoid potential global name collision.
9594 * gimplify.c: Include gimplify-ctx.h.
9595 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
9596 (htab_t gimplify_ctx::temp_htab):
9597 Update dependent calls and types for new type hash_table.
9598 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
9599 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
9601 * gimplify-ctx.h: New.
9602 (struct gimple_temp_hash_elt): Move from gimplify.c.
9603 (class gimplify_hasher): New.
9604 (struct gimplify_ctx): Move from gimple.h.
9605 (htab_t gimplify_ctx::temp_htab):
9606 Change type to hash_table. Update dependent calls and types.
9608 * graphite-clast-to-gimple.c: Include graphite-htab.h.
9609 (htab_t ivs_params::newivs_index):
9610 Change type to hash_table. Update dependent calls and types.
9611 (htab_t ivs_params::params_index): Likewise.
9612 (htab_t print_generated_program::params_index): Likewise.
9613 (htab_t gloog::newivs_index): Likewise.
9614 (htab_t gloog::params_index): Likewise.
9616 * graphite.c: Include graphite-htab.h.
9617 4htab_t graphite_transform_loops::bb_pbb_mapping):
9618 Change type to hash_table. Update dependent calls and types.
9620 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
9621 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
9622 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
9624 * graphite-dependences.c: Include graphite-htab.h.
9625 (loop_is_parallel_p): Change hash table type of parameter.
9627 * graphite-htab.h: New.
9628 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
9629 (extern find_pbb_via_hash): Move from graphite-poly.h.
9630 (extern loop_is_parallel_p): Move from graphite-poly.h.
9631 (extern get_loop_body_pbbs): Move from graphite-poly.h.
9633 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
9634 (extern loop_is_parallel_p): Move to graphite-htab.h.
9635 (extern get_loop_body_pbbs): Move to graphite-htab.h.
9637 * haifa-sched.c (htab_t delay_htab):
9638 Change type to hash_table. Update dependent calls and types.
9639 (htab_t delay_htab_i2): Likewise.
9641 * ira-color.c (htab_t allocno_hard_regs_htab):
9642 Change type to hash_table. Update dependent calls and types.
9644 * ira-costs.c (htab_t cost_classes_htab):
9645 Change type to hash_table. Update dependent calls and types.
9647 * loop-invariant.c (htab_t merge_identical_invariants::eq):
9648 Change type to hash_table. Update dependent calls and types.
9650 * loop-iv.c (htab_t bivs):
9651 Change type to hash_table. Update dependent calls and types.
9653 * loop-unroll.c (htab_t opt_info::insns_to_split):
9654 Change type to hash_table. Update dependent calls and types.
9655 (htab_t opt_info::insns_with_var_to_expand): Likewise.
9657 * lto-streamer.h (struct string_slot): Move from data-streamer.h
9658 (struct string_slot_hasher): New.
9659 (htab_t output_block::string_hash_table):
9660 Change type to hash_table. Update dependent calls and types.
9662 * lto-streamer-in.c (freeing_string_slot_hasher): New.
9663 (htab_t file_name_hash_table):
9664 Change type to hash_table. Update dependent calls and types.
9666 * lto-streamer-out.c: Update output_block::string_hash_table dependent
9669 * lto-streamer.c (htab_t tree_htab):
9670 Change type to hash_table. Update dependent calls and types.
9672 * omp-low.c: Include gimplify-ctx.h.
9674 * passes.c (htab_t name_to_pass_map):
9675 Change type to hash_table. Update dependent calls and types.
9676 (pass_traverse): Rename to passes_pass_traverse.
9678 * plugin.c (htab_t event_tab):
9679 Change type to hash_table. Update dependent calls and types.
9681 * postreload-gcse.c (htab_t expr_table):
9682 Change type to hash_table. Update dependent calls and types.
9683 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
9685 * sese.c (debug_rename_map_1): Make extern.
9686 (htab_t copy_bb_and_scalar_dependences::rename_map):
9687 Change type to hash_table. Update dependent calls and types.
9689 * sese.h (extern debug_rename_map): Move to .c file.
9691 * store-motion.c (htab_t store_motion_mems_table):
9692 Change type to hash_table. Update dependent calls and types.
9694 * trans-mem.c (htab_t tm_new_mem_hash):
9695 Change type to hash_table. Update dependent calls and types.
9697 * tree-browser.c (htab_t TB_up_ht):
9698 Change type to hash_table. Update dependent calls and types.
9700 * tree-cfg.c (htab_t discriminator_per_locus):
9701 Change type to hash_table. Update dependent calls and types.
9703 * tree-complex.c: Include tree-hasher.h
9704 (htab_t complex_variable_components):
9705 Change type to hash_table. Update dependent calls and types.
9707 * tree-eh.c (htab_t finally_tree):
9708 Change type to hash_table. Update dependent calls and types.
9710 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
9711 struct int_tree_hasher.
9712 (extern int_tree_map_eq): Likewise.
9713 (uid_decl_map_hash): Removed.
9714 (extern decl_tree_map_eq): Likewise.
9716 * tree-hasher.h: New.
9717 (struct int_tree_hasher): New.
9718 (typedef int_tree_htab_type): New.
9720 * tree-inline.c: Include gimplify-ctx.h.
9722 * tree-mudflap.c: Include gimplify-ctx.h.
9724 * tree-parloops.c: Include tree-hasher.h.
9725 (htab_t eliminate_local_variables_stmt::decl_address):
9726 Change type to hash_table. Update dependent calls and types.
9727 (htab_t separate_decls_in_region::decl_copies): Likewise.
9729 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
9730 Change type to hash_table. Update dependent calls and types.
9732 * tree-sra.c (candidates):
9733 Change type to hash_table. Update dependent calls and types.
9735 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
9737 (int_tree_map_hash): Likewise.
9739 * tree-ssa-dom.c (htab_t avail_exprs):
9740 Change type to hash_table. Update dependent calls and types.
9742 * tree-ssa-live.c (var_map_base_init::tree_to_index):
9743 Change type to hash_table. Update dependent calls and types.
9745 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
9746 Change type to hash_table. Update dependent calls and types.
9748 * tree-ssa-phiopt.c (seen_ssa_names):
9749 Change type to hash_table. Update dependent calls and types.
9751 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
9752 Change type to hash_table. Update dependent calls and types.
9754 * tree-ssa-uncprop.c (equiv):
9755 Change type to hash_table. Update dependent calls and types.
9757 2013-04-25 Jakub Jelinek <jakub@redhat.com>
9759 PR rtl-optimization/57003
9760 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
9761 call note_stores with kill_clobbered_value callback again after
9762 killing regs_invalidated_by_call.
9764 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9766 * config/aarch64/aarch64-simd.md
9767 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
9768 (aarch64_simd_bsl<mode>): Likewise.
9769 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
9771 2013-04-25 Marek Polacek <polacek@redhat.com>
9773 PR tree-optimization/57066
9774 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
9776 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
9778 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
9780 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9782 * config/aarch64/aarch64-builtins.c
9783 (aarch64_fold_builtin): New.
9784 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
9785 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
9786 * config/aarch64/aarch64-simd-builtins.def (abs): New.
9787 * config/aarch64/arm_neon.h
9788 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
9790 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9791 Tejas Belagod <tejas.belagod@arm.com>
9793 * config/aarch64/aarch64-builtins.c
9794 (aarch64_gimple_fold_builtin): New.
9795 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
9796 * config/aarch64/aarch64-simd-builtins.def (addv): New.
9797 * config/aarch64/aarch64-simd.md (addpv4sf): New.
9799 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
9801 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9803 * config/aarch64/aarch64.md
9804 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
9806 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9808 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
9809 (*ngcsi_uxtw): New pattern.
9811 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9812 Julian Brown <julian@codesourcery.com>
9814 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
9815 (TB_DREG): Add T_V4HF.
9816 (v4hf_UP): New macro.
9817 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
9818 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
9819 Handle initialisation of V4HF. Adjust initialisation of reinterpret
9821 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
9822 (arm_vector_mode_supported_p): Handle V4HF.
9823 (arm_mangle_map): Handle V4HFmode.
9824 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
9825 * config/arm/arm_neon_builtins.def: Add entries for
9826 vcvtv4hfv4sf, vcvtv4sfv4hf.
9827 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
9828 (neon_vcvtv4hfv4sf): Likewise.
9829 * config/arm/neon-gen.ml: Handle half-precision floating point
9831 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
9832 * config/arm/arm_neon.h: Regenerate.
9833 * config/arm/neon.ml (type elts): Add F16.
9834 (type vectype): Add T_float16x4, T_floatHF.
9835 (type vecmode): Add V4HF.
9836 (type features): Add Requires_FP_bit feature.
9837 (elt_width): Handle F16.
9838 (elt_class): Likewise.
9839 (elt_of_class_width): Likewise.
9840 (mode_of_elt): Refactor.
9841 (type_for_elt): Handle F16, fix error messages.
9842 (vectype_size): Handle T_float16x4.
9843 (vcvt_sh): New function.
9844 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
9845 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
9846 (string_of_mode): Handle V4HF.
9847 * doc/arm-neon-intrinsics.texi: Regenerate.
9849 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9851 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
9852 format specifier in 'X' case.
9854 2013-04-25 Alan Modra <amodra@gmail.com>
9857 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
9858 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
9859 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
9860 Repeat for many other rotate/shift and mask patterns using subregs.
9861 Name lshiftrt insns.
9862 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
9863 on WORDS_BIG_ENDIAN.
9865 2013-04-25 Alan Modra <amodra@gmail.com>
9867 * config.gcc: Support little-endian powerpc-linux targets.
9868 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
9869 (LINK_OS_LINUX_SPEC): Define.
9870 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
9871 Preserve MASK_LITTLE_ENDIAN.
9872 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
9873 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
9874 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
9875 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
9876 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
9877 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
9878 Correct fp word order for little-endian. Don't shift toc entries
9879 smaller than a word for little-endian.
9880 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
9881 (bswapdi2 splits): Correct low-part subreg for little-endian.
9882 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
9883 low/high where such is correct only for be.
9884 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
9885 little-endian for -mcall-aixdesc.
9887 2013-04-25 Alan Modra <amodra@gmail.com>
9889 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9890 replace_equiv_address_nv.
9892 2013-04-25 Alan Modra <amodra@gmail.com>
9894 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
9896 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9899 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9900 * rtl.h (struct rtx_def): ...
9902 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9904 PR rtl-optimizations/57046
9905 * lra-constraints (split_reg): Set up lra_risky_transformations_p
9906 for multi-reg splits.
9908 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
9910 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
9912 2013-04-24 Sterling Augustine <saugustine@google.com>
9914 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
9915 (comp_dir_string, debug_str_dwo_section): New.
9916 (DEBUG_STR_DWO_SECTION): Rename to ...
9917 (DEBUG_DWO_STR_SECTION): ... this.
9918 (DEBUG_NORM_STR_SECTION): Delete.
9919 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
9920 (DEBUG_STR_DWO_SECTION_FLAGS): New.
9921 (find_AT_string): Move most logic to ...
9922 (find_AT_string_in_table): ... here. New.
9923 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
9924 add_skeleton_AT_string. Delete logic.
9925 (output_skeleton_debug_sections): Remove call to
9926 add_top_level_skeleton_die_attrs.
9927 (add_comp_dir_attribute): Move logic to comp_dir_string.
9928 (dwarf2out_init): Initialize debug_str_dwo_section.
9929 (output_indirect_string): Call find_string_form.
9930 (output_indirect_strings): Rewrite.
9931 (prune_unused_types): Empty skeleton_debug_str_hash.
9932 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
9933 (dwarf2out_finish): Call output_indirect_strings.
9935 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9937 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
9939 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9941 * rtl.h (struct rtx_def): Add comment for field jump.
9942 (LRA_SUBREG_P): New macro.
9943 * recog.c (register_operand): Check LRA_SUBREG_P.
9944 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9946 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9947 (lra_final_code_change): Skip subreg change for operators.
9948 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9949 if there are no operand changes.
9950 * lra-constraints.c (curr_insn_set): New.
9951 (match_reload): Set LRA_SUBREG_P.
9952 (emit_spill_move): Ditto.
9953 (check_and_process_move): Use curr_insn_set. Process only single
9954 set insns. Don't initialize sec_mem_p and change_p.
9955 (simplify_operand_subreg): Use LRA_SUBREG_P.
9956 (reg_in_class_p): New function.
9957 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9958 of #ifdef. Add code to remove cycling.
9959 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9960 non-null disp. Reload inner instead of disp when base and index
9961 are null. Try to put lo_sum into register.
9962 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9963 (check_and_process_move): Move code for move cost check to
9964 simple_move_p. Remove equiv_substitution.
9965 (simple_move_p): New function.
9966 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9967 curr_insn_set. Call check_and_process_move only for single set
9968 insns. Use the new function. Move call of check_and_process_move
9969 after operand equiv substitution and address process.
9971 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
9973 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
9974 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
9975 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
9977 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9979 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
9981 2013-04-24 Marek Polacek <polacek@redhat.com>
9983 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
9984 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
9985 (select_loops_exit_conditions): Likewise.
9986 (number_of_iterations_for_all_loops): Likewise.
9987 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
9988 (scev_analysis): Likewise.
9990 2013-04-02 Catherine Moore <clm@codesourcery.com>
9991 Chao-ying Fu <fu@mips.com>
9993 * config/mips/micromips.md (jraddiusp): New pattern.
9994 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
9995 instruction if possible.
9997 2013-04-24 Alan Modra <amodra@gmail.com>
9999 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
10001 2013-04-24 Julian Brown <julian@codesourcery.com>
10002 Chung-Lin Tang <cltang@codesourcery.com>
10004 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
10005 dependency behavior in enumeration type DIE generation. Add TODO note
10006 to comments about future DW_FORM_sdata/udata re-work of related code.
10008 2013-04-23 Lawrence Crowl <crowl@google.com>
10010 * Makefile.in: Update as needed below.
10012 * hash-table.h (class hash_table):
10013 Correct many methods with parameter types compare_type to the correct
10014 value_type. (Correct code was unlikely to notice the change.)
10015 (hash_table::elements_with_deleted) New.
10016 (class hashtable::iterator): New.
10017 (hashtable::begin()): New.
10018 (hashtable::end()): New.
10019 (FOR_EACH_HASH_TABLE_ELEMENT): New.
10021 * statistics.c (statistics_hashes):
10022 Change type to hash_table. Update dependent calls and types.
10024 * tree-into-ssa.c (var_infos):
10025 Change type to hash_table. Update dependent calls and types.
10027 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
10028 Change type to hash_table. Update dependent calls and types.
10030 * tree-ssa-loop-im.c (struct mem_ref.refs):
10031 Change type to hash_table. Update dependent calls and types.
10033 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
10034 Change type to hash_table. Update dependent calls and types.
10036 * tree-ssa-sccvn.c (vn_tables_s::nary):
10037 Change type to hash_table. Update dependent calls and types.
10038 (vn_tables_s::phis): Likewise.
10039 (vn_tables_s::references): Likewise.
10041 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
10042 (vn_reference_eq): Update parameter and return types.
10044 * tree-ssa-structalias.c (pointer_equiv_class_table):
10045 Change type to hash_table. Update dependent calls and types.
10046 (location_equiv_class_table): Likewise.
10048 * tree-vect-data-refs.c: Consequential changes for making
10049 peeling a hash_table.
10051 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
10052 (destroy_loop_vec_info): Dependent hash_table update.
10054 * tree-vectorizer.h (peeling_htab):
10055 Change type to hash_table. Update dependent calls and types.
10057 2013-04-23 Shiva Chen <shiva0217@gmail.com>
10059 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
10060 to check the register content is equal or not.
10061 * lra-constraints.c (match_reload): Use lra_assign_reg_val
10062 to assign register content record.
10063 * lra-eliminations.c (update_reg_eliminate): Use
10064 lra_update_reg_val_offset to update register content offset.
10065 * lra-int.h (struct lra_reg): Add offset member.
10066 (lra_reg_val_equal_p): New static inline function.
10067 (lra_update_reg_val_offset): New static inline function.
10068 (lra_assign_reg_val): New static inline function.
10069 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
10070 to assign register content record.
10071 (initialize_lra_reg_info_element): Initial offset to zero.
10073 2013-04-23 Catherine Moore <clm@codesourcery.com>
10075 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
10076 operands. Record compression.
10078 2013-04-23 Xinliang David Li <davidxl@google.com>
10080 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
10082 2013-04-23 Richard Biener <rguenther@suse.de>
10084 PR middle-end/57036
10085 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
10086 parameter, only add abnormal goto edges from the copied body
10087 if the call could perform abnormal gotos.
10088 (copy_cfg_body): Adjust.
10090 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10092 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10094 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10096 * coretypes.h (gimple_stmt_iterator): Add struct to make
10099 2013-04-23 Richard Biener <rguenther@suse.de>
10101 PR tree-optimization/57026
10102 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10103 from SSA names occuring in abnormal PHI nodes.
10105 2013-04-22 Andi Kleen <ak@linux.intel.com>
10107 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10109 2013-04-22 Andi Kleen <ak@linux.intel.com>
10111 * lto/lto.c (print_lto_report_1): Declare early.
10112 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10114 2013-04-22 Andi Kleen <ak@linux.intel.com>
10116 * common.opt (-flto-report-wpa): Add.
10117 * doc/invoke.texi (-flto-report-wpa): Add.
10118 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10121 2013-04-22 Xinliang David Li <davidxl@google.com>
10123 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10124 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10125 * Makefile.in: New dependency
10127 David Daney <ddaney.cavm@gmail.com>
10129 * configure.ac (gcc_cv_as_micromips_support): Use the
10130 --fatal-warnings option.
10131 * configure: Regenerate.
10133 2013-04-22 Marek Polacek <polacek@redhat.com>
10136 * tsan.c (instrument_expr): Don't instrument expression
10137 in case its size is zero.
10139 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
10143 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10145 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10147 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10149 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10150 (gimple_stmt_iterator): New typedef.
10151 * gimple.h (gimple_stmt_iterator): Rename to...
10152 (gimple_stmt_iterator_d): ... This.
10153 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10154 trees be valid for GIMPLE and GENERIC.
10155 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10156 * gimple-fold.c (gimple_fold_call): Call target hook
10157 gimple_fold_builtin.
10158 * hooks.c (hook_bool_gsiptr_false): New.
10159 * hooks.h (hook_bool_gsiptr_false): New.
10160 * target.def (fold_stmt): New.
10161 * doc/tm.texi: Regenerate.
10163 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10166 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10167 a set sp if no stack realignment.
10169 2013-04-22 Nick Clifton <nickc@redhat.com>
10171 * config.gcc (tilegx-linux): Extend extra_objs rather than
10173 (tilepro-linux): Likewise.
10175 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10177 * config/aarch64/aarch64-builtins.c
10179 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10180 (VAR<1-12>): Add MAP parameter.
10181 (BUILTIN_*): Likewise.
10182 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10183 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10184 (aarch64_ushl_n<mode>): Likewise.
10185 (aarch64_sshr_n<mode>): Likewise.
10186 (aarch64_ushr_n<mode>): Likewise.
10187 (aarch64_<maxmin><mode>): Likewise.
10188 (aarch64_sqrt<mode>): Likewise.
10189 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10190 (vshr<q>_n_*): Likewise.
10192 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10194 * config/aarch64/aarch64-builtins.c
10195 (aarch64_simd_builtin_type_mode): Handle SF types.
10197 (BUILTIN_GPF): Define.
10198 (aarch64_init_simd_builtins): Handle SF types.
10199 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10200 (frecps): Likewise.
10201 (frecpx): Likewise.
10202 * config/aarch64/aarch64-simd.md
10203 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10204 (aarch64_frecpe<mode>): New.
10205 (aarch64_frecps<mode>): Likewise.
10206 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10207 (v8type): Add frecp<esx>.
10208 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10209 (aarch64_frecps<mode>): Likewise.
10210 * config/aarch64/iterators.md (FRECP): New.
10211 (frecp_suffix): Likewise.
10212 * config/aarch64/arm_neon.h
10213 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10215 2013-04-22 Christian Bruel <christian.bruel@st.com>
10218 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10219 (REG_CLASS_NAMES): Idem.
10220 (REG_CLASS_CONTENTS): Idem.
10221 (REGCLASS_HAS_FP_REG): Idem.
10222 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10223 (sh_conditional_register_usage): Idem.
10225 2013-04-21 Jeff Law <law@redhat.com>
10227 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10228 (ssa_forward_propagate_and_combine): Use it.
10230 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10232 * lra.c: Update the flow chart diagram.
10234 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10236 PR rtl-optimization/56847
10237 * lra-constraints.c (process_alt_operands): Discourage alternative
10238 with non-matche doffsettable memory constraint fro memory with
10241 2013-04-19 Richard Biener <rguenther@suse.de>
10243 PR tree-optimization/56982
10244 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10246 * gimplify.c (gimplify_call_expr): Notice special calls.
10247 (gimplify_modify_expr): Likewise.
10248 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10249 abnormal control flow receivers.
10250 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10251 in the same way as cfun->has_nonlocal_labels.
10252 (gimple_purge_dead_abnormal_call_edges): Likewise.
10253 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10254 receivers start a basic-block.
10256 2013-04-19 Richard Biener <rguenther@suse.de>
10258 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10260 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10261 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10262 (SLP_TREE_LOAD_PERMUTATION): Add.
10263 (vect_transform_slp_perm_load): Adjust prototype.
10264 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10265 (vect_free_slp_instance): Likewise.
10266 (vect_create_new_slp_node): Likewise.
10267 (vect_supported_slp_permutation_p): Remove.
10268 (vect_slp_rearrange_stmts): Adjust.
10269 (vect_supported_load_permutation_p): Likewise. Inline
10270 vect_supported_slp_permutation_p here.
10271 (vect_analyze_slp_instance): Compute load permutations per
10272 slp node instead of per instance.
10273 (vect_get_slp_defs): Adjust.
10274 (vect_transform_slp_perm_load): Likewise.
10275 (vect_schedule_slp_instance): Remove redundant code.
10276 (vect_schedule_slp): Remove hack for PR56270, add it ...
10277 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10278 CSE loads for SLP. Adjust.
10280 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10282 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10283 spelling in two comments.
10285 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10288 * config/arm/arm.c (load_multiple_sequence): Require SP
10289 as base register for loads if SP is in the register list.
10291 2013-04-19 Martin Jambor <mjambor@suse.cz>
10293 PR tree-optimization/56718
10294 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10295 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10296 and made public. Adjusted all callers.
10297 (ipa_intraprocedural_devirtualization): New function.
10298 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10299 (ipa_intraprocedural_devirtualization): Likewise.
10300 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10302 2013-04-19 Richard Biener <rguenther@suse.de>
10304 PR tree-optimization/57000
10305 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10307 2013-04-19 Terry Guo <terry.guo@arm.com>
10309 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10311 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10312 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10313 (cortex_m4_fmacs): Use new reservations.
10314 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10316 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10318 PR rtl-optimization/56999
10319 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10321 (lra_coalesce): Remove split_origin_bitmap and related code.
10322 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10323 ranges if necessary.
10325 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
10327 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10329 (ix86_expand_call): Remove clobbered_registers array and use
10330 x86_64_ms_sysv_extra_clobbered_registers instead.
10331 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10333 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10335 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10336 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10337 (*call_value_rex64_ms_sysv): Ditto.
10339 2013-04-18 Cary Coutant <ccoutant@google.com>
10341 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10342 parent instead of die_mark.
10344 2013-04-18 Diego Novillo <dnovillo@google.com>
10346 * gimple.c (create_gimple_tmp): New.
10347 (get_expr_type): New.
10348 (build_assign): New.
10349 (build_type_cast): New.
10350 * gimple.h (enum ssa_mode): Define.
10351 (gimple_seq_set_location): New.
10352 * asan.c (build_check_stmt): Change some gimple_build_* calls
10353 to use build_assign and build_type_cast.
10355 2013-04-18 Richard Biener <rguenther@suse.de>
10357 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10358 handle negative step. Remove redundant checks.
10359 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10360 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10361 for negative step and grouped loads fail to vectorize.
10363 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10365 * emit-rtl.c (reset_insn_used_flags): New function.
10366 (reset_all_used_flags): Use it.
10367 (verify_insn_sharing): New function.
10368 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10370 2013-04-18 Jakub Jelinek <jakub@redhat.com>
10372 PR tree-optimization/56984
10373 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10374 and (x >> M) >= N don't register any assertion if N << M is the
10377 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10379 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10380 do not delete_insn insns that have not yet been emitted, only
10381 unlink them with remove_insn.
10382 * df-scan.c (df_insn_delete): Revert r197492.
10384 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10386 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10387 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10389 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10391 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10392 define_insn_and_split.
10393 (and_scc,ior_scc,negscc): Likewise.
10394 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10396 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10398 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10399 can benefit from using a sequence of LDRD instructions in epilogue
10400 instead of a single LDM instruction.
10402 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10405 * doc/extend.texi: Fix typo.
10407 2013-04-17 Richard Biener <rguenther@suse.de>
10409 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10410 (vect_build_slp_tree): ... here.
10411 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10412 match. Remove special-casing of mismatched complex loads.
10413 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10414 re-try the match with swapped commutative operands.
10415 (vect_supported_load_permutation_p): Remove special-casing of
10416 mismatched complex loads.
10417 (vect_analyze_slp_instance): Adjust.
10419 2013-04-17 Richard Biener <rguenther@suse.de>
10421 PR rtl-optimization/56921
10422 * cfgloop.h (struct loop): Add simple_loop_desc member.
10423 (struct niter_desc): Mark with GTY(()).
10424 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10425 * loop-iv.c (get_simple_loop_desc): Likewise.
10426 (free_simple_loop_desc): Likewise.
10429 2013-04-16 Richard Biener <rguenther@suse.de>
10431 PR rtl-optimization/56921
10432 * loop-init.c (pass_rtl_move_loop_invariants): Add
10433 TODO_do_not_ggc_collect to todo_flags_finish.
10434 (pass_rtl_unswitch): Same.
10435 (pass_rtl_unroll_and_peel_loops): Same.
10436 (pass_rtl_doloop): Same.
10438 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
10440 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
10441 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
10442 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
10444 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
10445 * tree-streamer.c (record_common_node): Adjust reference in comment.
10447 2013-04-17 Terry Guo <terry.guo@arm.com>
10449 * config/arm/cortex-m4.md: Add a new bypass.
10451 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10453 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
10455 (*subs_<optab><mode>_multp2): New pattern.
10456 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10457 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10459 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10461 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
10462 (*subs_mul_imm_<mode>): New pattern.
10464 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
10467 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
10468 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
10469 (vsx_movti_32bit): Change j->wa to O->wa.
10471 2013-04-16 Richard Biener <rguenther@suse.de>
10473 PR rtl-optimization/56921
10474 * loop-init.c (pass_rtl_move_loop_invariants): Add
10475 TODO_do_not_ggc_collect to todo_flags_finish.
10476 (pass_rtl_unswitch): Same.
10477 (pass_rtl_unroll_and_peel_loops): Same.
10478 (pass_rtl_doloop): Same.
10480 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
10482 * config/arm/arm.c (emit_multi_reg_push): New declaration
10483 for an existing function.
10484 (arm_emit_strd_push): New function.
10485 (arm_expand_prologue): Used here.
10486 (arm_emit_ldrd_pop): New function.
10487 (arm_expand_epilogue): Used here.
10488 (arm_get_frame_offsets): Update condition.
10489 (arm_emit_multi_reg_pop): Add a special case for load of a single
10490 register with writeback.
10492 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
10494 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
10497 2013-04-16 Richard Biener <rguenther@suse.de>
10499 PR tree-optimization/56756
10500 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
10501 (first_mem_ref_loc): New.
10502 (execute_sm): Place the load temporarily before a previous
10503 access instead of in the latch edge to ensure its SSA dependencies
10504 are defined at points dominating the load.
10506 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
10508 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
10509 correct fix by moving header and footer insn to the footer of
10510 the merged basic block. Clear BB_END of the merged-away block.
10512 PR middle-end/43631
10513 * emit-rtl.c (make_note_raw): New function.
10514 (link_insn_into_chain): New static inline function.
10515 (add_insn): Use it.
10516 (add_insn_before, add_insn_after): Factor insn chain linking code...
10517 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
10518 using link_insn_into_chain.
10519 (note_outside_basic_block_p): New helper function for emit_note_after
10520 and emit_note_before.
10521 (emit_note_after): Use nobb variant of add_insn_after if the note
10522 should not be contained in a basic block.
10523 (emit_note_before): Use nobb variant of add_insn_before if the note
10524 should not be contained in a basic block.
10525 (emit_note_copy): Use make_note_raw.
10526 (emit_note): Likewise.
10527 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
10528 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
10529 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
10530 the moved barrier the tail of the basic block it follows.
10531 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
10533 2013-04-15 Jakub Jelinek <jakub@redhat.com>
10535 PR tree-optimization/56962
10536 * gimple-ssa-strength-reduction.c (record_increment): Only set
10537 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
10538 either rhs1 or rhs2 is equal to c->base_expr.
10540 2013-04-15 Richard Biener <rguenther@suse.de>
10542 PR tree-optimization/56933
10543 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
10545 (GROUP_READ_WRITE_DEPENDENCE): Remove.
10546 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
10547 * tree-vect-data-refs.c (vect_analyze_group_access): Move
10548 dependence check ...
10549 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
10551 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
10552 GROUP_READ_WRITE_DEPENDENCE.
10554 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10556 * emit-rtl.c (reset_all_used_flags): New function.
10557 (verify_rtl_sharing): Call reset_all_used_flags before and after
10558 performing the checks.
10560 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10562 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
10563 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
10564 * config/arm/constraints.md (De): New constraint.
10565 * config/arm/neon.md (anddi3_neon): Delete.
10566 (neon_vand<mode>): Expand to standard anddi3 pattern.
10567 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
10568 Move earlier in the file.
10569 (neon_inv_logic_op2): Likewise.
10570 (arm_anddi_operand_neon): New predicate.
10572 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10574 * configure.ac (gcc_cv_ld_as_needed): Set
10575 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
10576 Use -z ignore, -z record on *-*-solaris2*.
10577 (HAVE_LD_AS_NEEDED): Update comment.
10578 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
10579 * configure: Regenerate.
10580 * config.in: Regenerate.
10581 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
10582 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
10583 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
10584 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
10585 equivalents. Fix markup.
10586 * doc/tm.texi: Regenerate.
10588 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
10590 * config/i386/i386.opt: New option mstack-protector-guard=.
10591 * config/i386/i386-opts.h: Add enum stack_protector_guard.
10592 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
10593 TARGET_SSP_TLS_GUARD.
10594 * config/i386/i386.c (ix86_option_override_internal): Set
10595 ix86_stack_protector_guard.
10596 * config/i386/i386.md (stack_protect_set): Enable for
10597 TARGET_SSP_TLS_GUARD only.
10598 (stack_protect_set_<mode>): Ditto.
10599 (stack_protect_test): Ditto.
10600 (stack_protect_test_<mode>): Ditto.
10601 * doc/invoke.texi (i386 Option): Document.
10603 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
10606 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
10607 (S_MODES): Set H_MODE bit.
10608 (SF_MODES): Set only S_MODE and SF_MODE bits.
10609 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
10610 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
10611 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
10612 <MODE_FLOAT>: Likewise.
10614 2013-04-15 Joey Ye <joey.ye@arm.com>
10616 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
10618 2013-04-15 Joey Ye <joey.ye@arm.com>
10620 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
10622 (thumb_far_jump_used_p): Count instruction size and set
10625 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
10627 * reorg.c (fill_simple_delay_slots): Reindent block of code.
10628 * resource.c (mark_target_live_regs): Reformat conditional block.
10630 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10632 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
10633 notes, they are emitted only just before final.
10634 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
10636 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10638 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
10639 * cfgrtl.c (delete_insn): Call it here instead.
10640 * lra-spills.c (lra_final_code_change): Use delete_insn.
10641 * haifa-sched.c (sched_remove_insn): Likewise.
10642 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
10643 returning to the nop pool.
10644 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
10645 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
10647 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10649 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
10650 * doc/tm.texi: Regenerated.
10652 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
10654 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
10657 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10659 * df-core.c (df_find_def): Compare register numbers.
10660 (df_find_use): Likewise.
10662 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
10665 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
10666 lra_in_progress for return.
10668 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10670 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
10671 define_insn into define_insn_and_split and emit movsicc patterns.
10673 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10675 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
10677 2013-04-12 Richard Biener <rguenther@suse.de>
10679 * tree-pass.h (TODO_do_not_ggc_collect): New.
10680 * passes.c (execute_one_ipa_transform_pass): Honor
10681 TODO_do_not_ggc_collect.
10682 (execute_one_pass): Likewise.
10685 2013-04-10 Richard Biener <rguenther@suse.de>
10687 * passes.c (init_optimization_passes): Remove reload pass.
10688 * ira.c (do_reload): Merge into ...
10690 (rest_of_handle_reload): Remove.
10691 (pass_reload): Likewise.
10692 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10693 of reload for vzeroupper pass placement.
10695 2013-04-12 Jakub Jelinek <jakub@redhat.com>
10697 PR tree-optimization/56918
10698 PR tree-optimization/56920
10699 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
10700 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
10701 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
10702 use wide_mul_with_sign method.
10704 2013-04-12 Richard Biener <rguenther@suse.de>
10706 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
10707 not be considered a gimple constant.
10709 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
10711 * fold-const.c (const_binop): Handle vector shifts by a scalar.
10712 (fold_binary_loc): Call const_binop also for mixed vector-scalar
10715 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
10716 Jakub Jelinek <jakub@redhat.com>
10718 * opts.c: Include diagnostic-color.h.
10719 (common_handle_option): Handle OPT_fdiagnostics_color_.
10720 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
10721 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
10722 (diagnostic-color.o): New.
10723 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
10724 (diagnostic_color_rule): New enum.
10725 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
10726 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
10727 the location string.
10728 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
10729 either NULL, or color kind.
10730 * diagnostic-color.c: New file.
10731 * diagnostic-color.h: New file.
10732 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
10734 * doc/invoke.texi (-fdiagnostics-color): Document.
10735 * pretty-print.h (pp_show_color): Define.
10736 (struct pretty_print_info): Add show_color field.
10737 * diagnostic.c: Include diagnostic-color.h.
10738 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
10739 macros. Colorize error:, warning: etc. strings and also the location
10741 (diagnostic_show_locus): Colorize the caret line.
10742 * pretty-print.c: Include diagnostic-color.h.
10743 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
10744 inside of %< %> quotes or quoted through q format modifier.
10746 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10748 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
10750 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10752 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
10753 code in CC_NZ mode.
10754 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
10757 2013-04-11 Marek Polacek <polacek@redhat.com>
10759 PR tree-optimization/48184
10760 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
10762 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
10764 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
10765 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
10766 (skip_simple_arithmetic): Tidy up.
10767 * tree.h (skip_simple_constant_arithmetic): Declare.
10769 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10771 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
10773 2013-04-11 Richard Biener <rguenther@suse.de>
10775 * tree-vect-loop.c (get_initial_def_for_induction): Properly
10776 generate vector constants.
10778 2013-04-11 Richard Biener <rguenther@suse.de>
10780 PR tree-optimization/56878
10781 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
10782 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
10784 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
10785 Prefer to align the DR with the most invariant base address.
10787 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10789 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
10792 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
10794 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
10795 floating-point vector comparisons against 0.
10797 2013-04-11 Jakub Jelinek <jakub@redhat.com>
10799 PR tree-optimization/56899
10800 * fold-const.c (extract_muldiv_1): Apply distributive law
10801 only if TYPE_OVERFLOW_WRAPS (ctype).
10803 2013-04-11 Bin Cheng <bin.cheng@arm.com>
10806 * ira-costs.c (scan_one_insn): Check whether the source rtx of
10807 loading has side effect.
10809 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
10811 * config/sparc/sparc.c: Include tree-pass.h.
10812 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
10813 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
10814 head of file. Change return type. Split off gate function.
10815 (sparc_gate_work_around_errata): New function.
10816 (pass_work_around_errata): New pass definition.
10817 (insert_pass_work_around_errata) New pass insert definition to
10818 insert pass_work_around_errata just after delayed-branch scheduling.
10819 (sparc_option_override): Insert the pass.
10820 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
10822 2013-04-10 David S. Miller <davem@davemloft.net>
10824 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
10825 or -mcpu=hypersparc.
10827 * target.def (cstore_mode): New hook.
10828 * target.h: Include insn-codes.h
10829 * targhooks.c: Likewise.
10830 (default_cstore_mode): New function.
10831 * targhooks.h: Declare it.
10832 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
10833 * doc/tm.texi: Rebuild.
10834 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
10835 target hook, rather than inspecting the insn_data.
10836 * config/sparc/sparc.c (sparc_cstore_mode): New function.
10837 (TARGET_CSTORE_MODE): Redefine.
10838 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
10840 * config/sparc/predicates.md (cstore_result_operand): New special
10842 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
10843 Use it for operand 0.
10844 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
10845 (*snesi_special): Likewise.
10846 (*snesi_zero): Likewise.
10847 (*seqsi_zero): Likewise.
10848 (*sltu_insn): Likewise.
10849 (*sgeu_insn): Likewise.
10850 (*seqdi_special): Make operand 0 and comparison operation be of
10852 (*snedi_special): Likewise.
10853 (*snedi_special_vis3): Likewise.
10854 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
10855 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
10856 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
10857 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
10858 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
10859 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
10860 (*sltu_extend_sp64): Likewise.
10861 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
10862 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
10863 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
10864 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
10865 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
10867 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
10869 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
10870 (aarch64_start_file): Use the new function.
10872 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10873 Jason Merrill <jason@redhat.com>
10875 * common.opt: Add -gdwarf.
10876 * opts.c (common_handle_option): Handle it.
10877 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
10879 2013-04-10 Richard Biener <rguenther@suse.de>
10881 * passes.c (execute_todo): Do not call ggc_collect conditional here.
10882 (execute_one_ipa_transform_pass): But unconditionally here.
10883 (execute_one_pass): And here.
10884 (init_optimization_passes): Remove reload pass.
10885 * tree-pass.h (TODO_ggc_collect): Remove.
10886 (pass_reload): Likewise.
10887 * ira.c (do_reload): Merge into ...
10889 (rest_of_handle_reload): Remove.
10890 (pass_reload): Likewise.
10891 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10892 of reload for vzeroupper pass placement.
10893 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
10894 and todo_flags_finish of all passes.
10896 2013-04-10 Richard Biener <rguenther@suse.de>
10898 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
10899 first_const_oprnd field, rename first_def_type to first_op_type.
10900 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
10901 (vect_get_and_check_slp_defs): Always use the type of the
10902 operand. Allow mixed vect_external_def, vect_constant_def types.
10903 (vect_get_constant_vectors): Handle mixed vect_external_def,
10904 vect_constant_def types.
10906 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
10908 PR tree-optimization/55524
10909 * tree-ssa-math-opts.c
10910 (convert_mult_to_fma): Don't use an fms construct
10911 when we don't have an fms operation, but fnma, and it looks
10912 likely that we'll be able to use the latter.
10914 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
10916 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
10918 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
10919 inline fail caused by overwritable functions.
10921 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
10923 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
10924 unnecessary bits in the constant power of two case.
10926 2013-04-10 Richard Biener <rguenther@suse.de>
10928 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
10929 broken code swapping operands.
10930 (vect_build_slp_tree): Do not compute load permutations here.
10931 (vect_analyze_slp_instance): Compute load permutations here,
10932 after building the SLP tree.
10934 2013-04-09 Christian Bruel <christian.bruel@st.com>
10936 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
10937 of next/prev_real_insn.
10939 2013-04-09 Jan Hubicka <jh@suse.cz>
10941 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10942 Drop aliased parameter.
10943 (function_and_variable_visibility): Do not handle alias pairs.
10944 * cgraph.c (varpool_externally_visible_p): Update prototype.
10945 * varpool.c (varpool_add_new_variable): Update.
10947 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10949 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
10951 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
10953 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
10955 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
10957 2013-04-09 Marek Polacek <polacek@redhat.com>
10959 PR tree-optimization/48762
10960 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
10962 2013-04-09 Richard Biener <rguenther@suse.de>
10964 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
10966 (vect_build_slp_tree): Likewise.
10967 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
10968 calculating the cost of a SLP instance.
10969 (vect_analyze_slp_instance): Use it from here, after building
10972 2013-04-09 Jakub Jelinek <jakub@redhat.com>
10974 PR middle-end/56883
10975 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
10976 expand_omp_for_static_chunk): Use simple_p = true in
10977 force_gimple_operand_gsi calls when assigning to addressable decls.
10979 2013-04-09 Jeff Law <law@redhat.com>
10981 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
10982 when the boolean was created by converting a wider object which
10983 had a boolean range.
10985 2013-04-09 Richard Biener <rguenther@suse.de>
10987 * tree-vectorizer.h (slp_void_p): Remove.
10988 (slp_tree): Typedef before _slp_tree declaration.
10989 (struct _slp_tree): Use a vector of slp_tree as children.
10990 (vect_get_place_in_interleaving_chain): Remove.
10991 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
10993 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
10995 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
10996 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
10997 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
10998 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
10999 Use slp_node instead of slp_void_p and adjust.
11001 2013-04-09 Richard Biener <rguenther@suse.de>
11003 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
11004 work that is not necessary.
11006 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11008 PR tree-optimization/56854
11009 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
11010 forward into clobber stmts if it would change MEM_REF lhs into
11013 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
11015 * tree.c (type_hash_lookup, type_hash_add): Make static.
11016 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
11018 2013-04-09 Richard Biener <rguenther@suse.de>
11020 * tree.h (unsave_expr_now): Remove.
11021 * tree-inline.c (mark_local_for_remap_r): Remove.
11022 (unsave_expr_1): Likewise.
11023 (unsave_r): Likewise.
11024 (unsave_expr_now): Likewise.
11025 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
11026 (propagate_tree_value): Likewise.
11028 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
11030 * doc/rtl.texi (sequence): Rewrite documentation to match the
11031 current use of SEQUENCE rtl objects.
11032 * rtl.def (SEQUENCE): Likewise.
11034 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
11035 Update documentation.
11036 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
11037 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
11039 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
11041 2013-04-08 Teresa Johnson <tejohnson@google.com>
11043 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
11044 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
11046 (estimate_edge_size_and_time): Add comment to suggest using rounding
11048 (estimate_node_size_and_time): Ditto.
11049 (remap_edge_change_prob): Use helper rounding divide methods.
11050 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
11051 (gimple_mod_pow2_value_transform): Ditto.
11052 (gimple_mod_subtract_transform): Ditto.
11053 (gimple_ic_transform): Ditto.
11054 (gimple_stringops_transform): Ditto.
11055 * stmt.c (conditional_probability): Ditto.
11056 (emit_case_dispatch_table): Ditto.
11057 * lto-cgraph.c (merge_profile_summaries): Ditto.
11058 * tree-optimize.c (execute_fixup_cfg): Ditto.
11059 * cfgcleanup.c (try_forward_edges): Ditto.
11060 * cfgloopmanip.c (scale_loop_profile): Ditto.
11062 (duplicate_loop_to_header_edge): Ditto.
11063 (lv_adjust_loop_entry_edge): Ditto.
11064 * tree-vect-loop.c (vect_transform_loop): Ditto.
11065 * profile.c (compute_branch_probabilities): Ditto.
11066 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
11067 * lto-streamer-in.c (input_cfg): Ditto.
11068 * gimple-streamer-in.c (input_bb): Ditto.
11069 * ipa-cp.c (update_profiling_info): Ditto.
11070 (update_specialized_profile): Ditto.
11071 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
11072 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
11074 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11075 (compute_trg_info): Ditto.
11076 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
11077 (purge_dead_edges): Ditto.
11078 * loop-unswitch.c (unswitch_loop): Ditto.
11079 * cgraphclones.c (cgraph_clone_edge): Ditto.
11080 (cgraph_clone_node): Ditto.
11081 * tree-inline.c (copy_bb): Ditto.
11082 (copy_edges_for_bb): Ditto.
11083 (initialize_cfun): Ditto.
11084 (copy_cfg_body): Ditto.
11085 (expand_call_inline): Ditto.
11087 2013-04-08 Kai Tietz <ktietz@redhat.com>
11089 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11090 TARGET_CYGWIN64 by TARGET_64BIT.
11092 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11094 * config/epiphany/epiphany.md (GPR_1): New constant.
11095 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11096 * config/epiphany/epiphany.c (gen_compare_reg):
11097 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11098 is already in place.
11099 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11100 Don't require being called during rtl expansion; If y operlaps r0,
11102 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11103 (epiphany_expand_epilogue): Likewise.
11105 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11106 Don't use CC_FPmode for ORDERED / UNORDERED.
11107 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
11109 * config/epiphany/constraints.md (CnL): New constraint.
11110 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11111 * config/epiphany/predicates.md (add_operand): Allow 1024.
11113 * config/epiphany/epiphany.md (logical_op): New code iterator.
11114 (op_mnc): New code attribute.
11115 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11116 (mov_f+1, mov_f+2): New peephole2 patterns.
11118 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11119 (cstoresi4): Also allow re-use of zero result when doing a NE
11120 comparison to a non-zero operand.
11121 Use (clobber (scratch)) for first insn if the gpr output is not needed.
11123 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11124 Use gen_addsi3_i / gen_subsi3_i.
11126 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11130 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11131 contain anything but clobbers, at most one __builtin_stack_restore,
11132 optionally debug stmts and final resx, and if it has at least one
11133 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11134 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11135 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11136 which isn't defaut definition, remove them.
11137 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11138 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11139 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11140 with MEM_REF LHS with SSA_NAME address.
11142 2013-04-08 Jeff Law <law@redhat.com>
11144 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
11146 2013-04-08 Richard Biener <rguenther@suse.de>
11148 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11150 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11151 determined vector type.
11152 (vect_analyze_data_refs): Likewise.
11153 (vect_get_new_vect_var): Adjust.
11154 (vect_create_destination_var): Preserve SSA name versions.
11155 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11156 not dump anything here.
11158 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11160 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11161 Add member lr_slot_known.
11162 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11164 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11165 Remove code that sets lr_slot_offset according to what a previous
11166 version of epiphany_emit_save_restore used to do.
11167 (epiphany_emit_save_restore): When doing an lr save or restore,
11168 set/verify lr_slot_known and lr_slot_offset.
11170 2013-04-08 Xinyu Qi <xyqi@marvell.com>
11173 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
11176 2013-04-08 Richard Biener <rguenther@suse.de>
11178 * alias.c (find_base_term): Fix thinko in previous change.
11180 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11182 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11183 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11184 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11185 if possible to compute val.
11186 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11187 For QImode integers don't require anything about precision. Use
11188 const_with_all_bytes_same to find out if the constant doesn't have
11189 repeated bytes in it.
11191 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11193 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11196 2013-04-08 Marek Polacek <polacek@redhat.com>
11198 PR rtl-optimization/48182
11199 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11202 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11205 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11206 nuses, make sure we have a label.
11208 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11211 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11212 (rs6000_emit_swdiv_low_precision): Remove.
11213 (rs6000_emit_swdiv): Rewrite to handle between one and four
11214 iterations of Newton-Raphson generally; modify required number of
11215 iterations for some cases.
11216 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11218 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11220 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11221 set-but-unused variable.
11223 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11224 basic blocks of released function bodies garbage-collectable.
11226 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11227 (struct rtl_opt_pass): Add TODO_df_finish.
11229 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11231 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11233 * config/arm/constraints.md (q): New constraint.
11234 * config/arm/ldrdstrd.md: New file.
11235 * config/arm/arm.md (ldrdstrd.md) New include.
11236 (arm_movdi): Use "q" instead of "r" constraint
11237 for double-word memory access.
11238 (movdf_soft_insn): Likewise.
11239 * config/arm/vfp.md (movdi_vfp): Likewise.
11240 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
11241 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
11242 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11243 (mem_ok_for_ldrd_strd): Likewise.
11244 (output_move_double): Update assertion.
11246 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11248 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11250 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11252 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11253 define_insn_and_split.
11254 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11256 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11258 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11259 define_insn_and_split.
11260 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
11261 (shiftsi3_compare): New pattern.
11262 (rrx): New pattern.
11263 * config/arm/unspecs.md (UNSPEC_RRX): New.
11265 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11267 * config/arm/arm.md (negdi_extendsidi): New pattern.
11268 (negdi_zero_extendsidi): Likewise.
11270 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11272 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11273 define_insn_and_split.
11274 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11275 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11277 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11279 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11280 define_insn_and_split.
11281 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11282 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11284 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11286 * config/arm/arm.md (subsi3_carryin): New pattern.
11287 (subsi3_carryin_const): Likewise.
11288 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11289 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11291 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11293 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11295 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11297 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
11298 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
11300 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11302 * config/arm/arm.c (arm_expand_builtin): Change fcode
11303 type to unsigned int.
11305 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11307 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11309 2013-04-04 Ian Lance Taylor <iant@google.com>
11311 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11314 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11316 PR middle-end/56729
11317 * df-scan.c (df_insn_delete): Disable failing assert.
11319 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11321 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11322 New function prototype.
11323 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11324 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11325 (arm_builtin_vectorized_function): New function.
11327 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11329 * config/arm/arm_neon_builtins.def: New file.
11330 * config/arm/arm.c (neon_builtin_data): Move contents to
11331 arm_neon_builtins.def.
11332 (enum arm_builtins): Include neon builtin definitions.
11333 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
11334 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
11336 2013-04-04 Marek Polacek <polacek@redhat.com>
11338 PR tree-optimization/48186
11339 * predict.c (maybe_hot_frequency_p): Return false if
11340 HOT_BB_FREQUENCY_FRACTION is 0.
11341 (cgraph_maybe_hot_edge_p): Likewise.
11343 2013-04-04 Richard Biener <rguenther@suse.de>
11345 PR tree-optimization/56826
11346 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11349 2013-04-04 Richard Biener <rguenther@suse.de>
11351 PR tree-optimization/56213
11352 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
11353 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
11355 2013-04-04 Richard Biener <rguenther@suse.de>
11357 PR tree-optimization/56837
11358 * tree-loop-distribution.c (classify_partition): For non-zero
11359 values require that the value has the same precision as its
11360 mode to be useful as memset value.
11362 2013-04-03 Nick Clifton <nickc@redhat.com>
11364 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
11365 (fmssf4): Use fmsf.s on E3V5 architectures.
11366 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11367 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11369 2013-04-03 Jeff Law <law@redhat.com>
11371 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11372 (lra-eliminations.o): Likewise.
11374 2013-04-03 Teresa Johnson <tejohnson@google.com>
11376 * gcov-io.c (compute_working_sets): Moved most of body of old
11377 compute_working_sets here from profile.c.
11378 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
11379 (gcov_working_set_t): Moved typedef here from basic-block.h
11380 (compute_working_set): Declare.
11381 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11382 (get_working_sets): Renamed from compute_working_set,
11383 replace most of body with call to new compute_working_sets.
11384 (get_exec_counts): Replace call to compute_working_sets
11385 to get_working_sets.
11386 * profile.h (get_working_sets): Renamed from compute_working_set.
11387 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
11388 to get_working_sets.
11389 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11390 * gcov-dump.c (dump_working_sets): New function.
11392 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11394 * hwint.c (sext_hwi, zext_hwi): New functions.
11395 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11396 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11397 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11398 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11399 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11400 (sext_hwi, zext_hwi): New functions.
11402 2013-04-03 Jeff Law <law@redhat.com>
11404 PR tree-optimization/56799
11405 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11406 back test for widening conversion erroneously dropped in prior change.
11408 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11411 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11412 instead of next_real_insn.
11414 2013-04-03 Marek Polacek <polacek@redhat.com>
11417 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
11419 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11422 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11424 (thumb1_output_casesi): Likewise.
11425 (thumb2_output_casesi): Likewise.
11427 2013-04-03 Richard Biener <rguenther@suse.de>
11429 PR tree-optimization/56817
11430 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11432 (tree_unroll_loops_completely_1): ... new function to manually
11433 walk the loop tree, properly defering outer loops of unrolled
11434 loops to later iterations.
11436 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11438 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
11439 (vectorizable_load): Likewise.
11440 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
11441 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
11443 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11445 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
11448 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11450 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
11452 2013-04-03 Bin Cheng <bin.cheng@arm.com>
11454 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
11456 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11458 PR tree-optimization/56790
11459 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
11462 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11464 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
11466 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
11467 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
11470 2013-04-03 Jakub Jelinek <jakub@redhat.com>
11473 * tree.h (force_folding_builtin_constant_p): New decl.
11474 * builtins.c (force_folding_builtin_constant_p): New variable.
11475 (fold_builtin_constant_p): Fold immediately also if
11476 force_folding_builtin_constant_p.
11478 2013-04-03 Richard Biener <rguenther@suse.de>
11480 PR tree-optimization/56812
11481 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
11482 DRs of the same interleaving chain are independent.
11484 2013-04-02 Jason Merrill <jason@redhat.com>
11486 * gdbinit.in (pbb): Use debug fn.
11488 2013-04-02 Lawrence Crowl <crowl@google.com>
11490 * sese.h (struct ivtype_map_elt_s): Remove unused.
11491 (extern debug_ivtype_map): Remove unused.
11492 (extern eq_ivtype_map_elts): Remove unused.
11493 * sese.c (debug_ivtype_map): Removed unused.
11494 (debug_ivtype_map_1): Removed unused.
11495 (debug_ivtype_elt): Remove unused.
11496 (eq_ivtype_map_elts): Remove unused.
11499 2013-04-02 Kai Tietz <ktietz@redhat.com>
11502 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
11503 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
11504 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
11506 (legitimize_pe_coff_symbol): Likewise.
11507 (is_imported_p): New helper-function.
11508 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
11509 for Windows x64 targets.
11510 (ix86_expand_prologue): Optimize for pe-coff targets.
11511 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
11512 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
11513 medium/large code-model.
11514 (legitimize_pic_address): Likewise.
11515 (legitimize_tls_address): Likewise.
11516 (ix86_expand_call): Likewise.
11517 (x86_output_mi_thunk): Likewise.
11518 (get_dllimport_decl): Add new beimport argument.
11519 (construct_plt_address): Don't assert for x64 pe-coff targets.
11520 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
11522 (SYMBOL_FLAG_STUBVAR): New macro.
11523 (SYMBOL_REF_STUBVAR_P): Likewise.
11524 * config/i386/winnt.c (stub_list): New structure.
11525 (stub_head): New local variable.
11526 (i386_pe_record_stub): New function.
11527 (i386_pe_file_end): Emit refptr-stubs.
11529 2013-04-02 Jakub Jelinek <jakub@redhat.com>
11531 PR rtl-optimization/56745
11532 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
11533 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
11536 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
11537 and both of them are MEM_REFs, just compare first argument for
11538 equality and attempt to deal even with differing offsets.
11541 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
11542 of gimple_clobber_p to be MEM_REF.
11543 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
11544 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
11545 after gimplification.
11546 * asan.c (get_mem_ref_of_assignment): Don't instrument
11547 gimple_clobber_p stmts.
11548 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
11549 gimple_clobber_p stmt if they have MEM_REF lhs and
11550 are dead because of another gimple_clobber_p stmt.
11551 * tree-ssa-live.c (clear_unused_block_pointer): Treat
11552 gimple_clobber_p stmts like debug stmts.
11553 (remove_unused_locals): Remove clobbers with MEM_REF lhs
11554 that refer to unused VAR_DECLs or uninitialized values.
11555 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
11556 gimple_clobber_p stmts if they refer to removed parameters.
11557 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
11560 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
11562 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
11563 using SWI48 mode attribute.
11565 2013-04-02 Wei Mi <wmi@google.com>
11567 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
11568 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
11569 *<rotate_insn><mode>3_mask in i386.md.
11571 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
11573 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
11575 2013-04-02 Richard Biener <rguenther@suse.de>
11577 PR tree-optimization/56778
11578 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
11579 Runtime alias tests are not supported for gather loads.
11580 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
11581 stmts referenced from SSA operands before updating SSA form.
11583 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
11584 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11586 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
11587 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
11588 * config/arm/cortex-a53.md: New file.
11589 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
11590 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
11591 * config/arm/arm.c (arm_issue_rate): Likewise.
11592 * config/arm/arm-tune.md: Regenerate
11593 * config/arm/arm-tables.opt: Regenerate.
11594 * config/arm/arm-cores.def: Add cortex-a53.
11596 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
11598 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
11601 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11603 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
11604 scalar load/store operations using B/H registers.
11605 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
11607 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11609 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
11611 * config/aarch64/aarch64.c
11612 (aarch64_simd_scalar_immediate_valid_for_move): New.
11613 * config/aarch64/aarch64-protos.h
11614 (aarch64_simd_scalar_immediate_valid_for_move): New.
11615 * config/aarch64/constraints.md (Dh, Dq): New.
11616 * config/aarch64/iterators.md (hq): New.
11618 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
11620 * reorg.c (get_branch_condition): Deal with conditional returns.
11621 (fill_simple_delay_slots): Remove dead code dealing with jumps.
11623 2013-04-01 Wei Mi <wmi@google.com>
11625 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
11626 Truncate operand 2 using %b asm operand modifier.
11627 (*<shift_insn><mode>3_mask): Ditto.
11628 (*<rotate_insn><mode>3_mask): Ditto.
11630 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
11632 PR middle-end/56798
11633 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
11635 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
11637 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
11639 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
11641 2013-03-30 Lawrence Crowl <crowl@google.com>
11643 * dse.c (clear_alias_sets): Remove never set.
11644 (disqualified_clear_alias_sets): Remove never set.
11645 (clear_alias_mode_pool): Remove never set.
11646 (dse_step0): Remove condition that is never true.
11647 (canon_address): Remove condition that is never true.
11648 (dse_step7): Remove condition that is never true.
11649 (rest_of_handle_dse): Remove condition that is never true.
11650 (rest_of_handle_dse::did_global): Remove never read from above.
11651 (dse_step2_spill): Remove never called from above.
11652 (dse_step5_spill): Remove never called from above.
11654 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
11656 * doc/md.texi (Standard Names) <casesi>: Update documentation for
11657 JUMP_TABLE_DATA changes.
11658 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
11659 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
11660 (Insns) <jump_table_data>: New entry.
11661 * doc/tm.texi: Regenerate.
11663 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
11665 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
11666 for table jump at the end of a basic block using tablejump_p.
11667 * targhooks.c (default_invalid_within_doloop): Likewise.
11668 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
11669 target hook implementation that is identical to the default hook.
11670 (rs6000_invalid_within_doloop): Remove.
11672 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
11673 unused variable from tablejump_p call.
11675 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
11676 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
11677 (INSN_DELETED_P): Likewise.
11678 (emit_jump_table_data): New prototype.
11679 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
11680 after 4th as unused.
11681 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
11682 * sched-vis.c (print_insn): Likewise.
11683 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
11684 insn for compatibility with back ends that use next_active_insn to
11685 identify jump table data.
11686 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
11687 (remove_insn): Likewise.
11688 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
11690 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
11691 (emit_jump_table_data): New function.
11693 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
11694 basic block, a JUMP_TABLE_DATA never is.
11695 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
11696 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
11697 off from code handling real insns.
11698 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
11699 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
11700 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
11701 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
11702 is not a NONDEBUG_INSN_P.
11703 * ira-costs.c (scan_one_insn): Likewise.
11704 * jump.c (mark_all_labels): Likewise.
11705 (mark_jump_label_1): Likewise.
11706 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11707 * lra.c (get_insn_freq): Expect all insns reaching here to be in
11709 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
11710 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
11711 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
11712 JUMP_TABLE_DATA_P insns.
11713 (calculate_elim_costs_all_insns): Likewise.
11714 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
11715 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
11716 (delete_output_reload): Code style fixups.
11717 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
11718 insn flags on this non-insn.
11719 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
11720 as scheduling barriers, for pre-change compatibility.
11721 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
11722 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
11724 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
11725 redundant JUMP_TABLE_DATA_P test.
11726 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
11727 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11728 (frv_for_each_packet): Likewise.
11729 * config/i386/i386.c (min_insn_size): Likewise.
11730 (ix86_avoid_jump_mispredicts): Likewise.
11731 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11732 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
11733 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11734 (mips16_insn_length): Robustify.
11735 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
11736 (mips16_split_long_branches): Likewise.
11737 * config/pa/pa.c (pa_combine_instructions): Likewise.
11738 * config/rs6000/rs6000.c (get_next_active_insn): Treat
11739 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
11740 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
11741 as contributing to pool range lengths.
11742 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
11743 Remove redundant JUMP_TABLE_DATA_P test.
11744 (sh_loop_align): Likewise.
11745 (split_branches): Likewise.
11746 (sh_insn_length_adjustment): Likewise.
11747 * config/spu/spu.c (get_branch_target): Likewise.
11749 2013-03-29 Jan Hubicka <jh@suse.cz>
11751 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
11752 gcov streaming; stream hot bb threshold to ltrans.
11753 * predict.c (get_hot_bb_threshold): Break out from ....
11754 (maybe_hot_count_p): ... here.
11755 (set_hot_bb_threshold): New function.
11756 * lto-section-in.c (lto_section_name): Add profile.
11757 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
11758 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
11759 and data-streamer.h
11760 (histogram_entry): New structure.
11761 (histogram, histogram_pool): New global vars.
11762 (histogram_hash): New structure.
11763 (histogram_hash::hash): New method.
11764 (histogram_hash::equal): Likewise.
11765 (account_time_size): New function.
11766 (cmp_counts): New function.
11767 (dump_histogram): New function.
11768 (ipa_profile_generate_summary): New function.
11769 (ipa_profile_write_summary): New function.
11770 (ipa_profile_read_summary): New function.
11771 (ipa_profile): Decide on threshold.
11772 (pass_ipa_profile): Add ipa_profile_write_summary and
11773 ipa_profile_read_summary.
11774 * Makefile.in (ipa.o): Update dependencies.
11775 * lto-streamer.h (LTO_section_ipa_profile): New section.
11777 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
11779 * tree.h (VAR_P): New.
11781 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
11784 * doc/invoke.texi ([-fwhole-program]): Fix typo.
11786 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
11788 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
11789 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11790 (control_flow_insn_p): Likewise.
11791 * cfgrtl.c (duplicate_insn_chain): Likewise.
11792 * final.c (get_attr_length_1): Likewise.
11793 (shorten_branches): Likewise.
11794 (final_scan_insn): Likewise.
11795 * function.c (instantiate_virtual_regs): Likewise.
11796 * gcse.c (insert_insn_end_basic_block): Likewise.
11797 * ira-costs.c (scan_one_insn): Likewise.
11798 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11799 * lra.c (check_rtl): Likewise.
11800 * reload1.c (elimination_costs_in_insn): Likewise.
11801 * reorg.c (follow_jumps): Likewise.
11803 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
11804 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11805 (thumb_far_jump_used_p): Likewise.
11806 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
11807 (workaround_speculation): Likewise.
11808 (add_sched_insns_for_speculation): Likewise.
11809 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
11810 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11811 (frv_for_each_packet): Likewise.
11812 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
11813 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
11814 (final_emit_insn_group_barriers): Likewise.
11815 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11816 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11817 (mips16_insn_length): Likewise.
11818 * config/pa/pa.c (pa_reorg): Likewise.
11819 (pa_combine_instructions): Likewise.
11820 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
11821 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
11822 (sh_reorg): Likewise.
11823 (split_branches): Likewise.
11824 * config/spu/spu.c (get_branch_target): Likewise.
11826 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
11829 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
11831 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
11832 Fix declaration name.
11834 2013-03-28 Lawrence Crowl <crowl@google.com>
11836 * graphds.h (struct graph.indicies): Remove unused.
11837 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
11838 (SCOP_ORIGINAL_PDDRS): Remove unused.
11839 * sese.h (extern insert_loop_close_phis): Removed unused.
11840 (extern insert_guard_phis): Removed unused.
11841 (extern ivtype_map_elt_info): Removed unused.
11842 (new_ivtype_map_elt): Removed unused.
11843 * sese.c (ivtype_map_elt_info): Removed unused.
11845 2013-03-28 Lawrence Crowl <crowl@google.com>
11847 * Makefile.in: Add several missing include dependences.
11849 (test-dump.o): New. This object is not added to any executable,
11850 but is present for ad-hoc testing.
11852 (debug (const bitmap_head_def &)): New.
11853 (debug (const bitmap_head_def *)): New.
11855 (extern debug (const bitmap_head_def &)): New.
11856 (extern debug (const bitmap_head_def *)): New.
11858 (debug (edge_def &)): New.
11859 (debug (edge_def *)): New.
11861 (debug (basic_block_def &)): New.
11862 (debug (basic_block_def *)): New.
11864 (dump_node (const_tree, int, FILE *)): Correct source file.
11866 (debug (die_struct &)): New.
11867 (debug (die_struct *)): New.
11869 (extern debug (die_struct &)): New.
11870 (extern debug (die_struct *)): New.
11871 * gimple-pretty-print.c
11872 (debug (gimple_statement_d &)): New.
11873 (debug (gimple_statement_d *)): New.
11874 * gimple-pretty-print.h
11875 (extern debug (gimple_statement_d &)): New.
11876 (extern debug (gimple_statement_d *)): New.
11878 (debug (ira_allocno_copy &)): New.
11879 (debug (ira_allocno_copy *)): New.
11880 (debug (ira_allocno &)): New.
11881 (debug (ira_allocno *)): New.
11883 (extern debug (ira_allocno_copy &)): New.
11884 (extern debug (ira_allocno_copy *)): New.
11885 (extern debug (ira_allocno &)): New.
11886 (extern debug (ira_allocno *)): New.
11888 (debug (live_range &)): New.
11889 (debug (live_range *)): New.
11891 (debug (lra_live_range &)): New.
11892 (debug (lra_live_range *)): New.
11894 (debug (lra_live_range &)): New.
11895 (debug (lra_live_range *)): New.
11897 (debug (omega_pb_d &)): New.
11898 (debug (omega_pb_d *)): New.
11900 (extern debug (omega_pb_d &)): New.
11901 (extern debug (omega_pb_d *)): New.
11903 (debug (const rtx_def &)): New.
11904 (debug (const rtx_def *)): New.
11906 (debug_tree (tree): Move within file.
11907 (debug_raw (const tree_node &)): New.
11908 (debug_raw (const tree_node *)): New.
11909 (dump_tree_via_hooks (const tree_node *, int)): New.
11910 (debug (const tree_node &)): New.
11911 (debug (const tree_node *)): New.
11912 (debug_verbose (const tree_node &)): New.
11913 (debug_verbose (const tree_node *)): New.
11914 (debug_head (const tree_node &)): New.
11915 (debug_head (const tree_node *)): New.
11916 (debug_body (const tree_node &)): New.
11917 (debug_body (const tree_node *)): New.
11918 (debug_vec_tree (tree): Move and reimplement in terms of dump.
11919 (debug (vec<tree, va_gc> &)): New.
11920 (debug (vec<tree, va_gc> *)): New.
11922 (extern debug (const rtx_def &)): New.
11923 (extern debug (const rtx_def *)): New.
11925 (debug_raw (simple_bitmap_def &)): New.
11926 (debug_raw (simple_bitmap_def *)): New.
11927 (debug (simple_bitmap_def &)): New.
11928 (debug (simple_bitmap_def *)): New.
11930 (extern debug (simple_bitmap_def &)): New.
11931 (extern debug (simple_bitmap_def *)): New.
11932 (extern debug_raw (simple_bitmap_def &)): New.
11933 (extern debug_raw (simple_bitmap_def *)): New.
11935 (debug (vinsn_def &)): New.
11936 (debug (vinsn_def *)): New.
11937 (debug_verbose (vinsn_def &)): New.
11938 (debug_verbose (vinsn_def *)): New.
11939 (debug (expr_def &)): New.
11940 (debug (expr_def *)): New.
11941 (debug_verbose (expr_def &)): New.
11942 (debug_verbose (expr_def *)): New.
11943 (debug (vec<rtx> &)): New.
11944 (debug (vec<rtx> *)): New.
11946 (extern debug (vinsn_def &)): New.
11947 (extern debug (vinsn_def *)): New.
11948 (extern debug_verbose (vinsn_def &)): New.
11949 (extern debug_verbose (vinsn_def *)): New.
11950 (extern debug (expr_def &)): New.
11951 (extern debug (expr_def *)): New.
11952 (extern debug_verbose (expr_def &)): New.
11953 (extern debug_verbose (expr_def *)): New.
11954 (extern debug (vec<rtx> &)): New.
11955 (extern debug (vec<rtx> *)): New.
11957 (_list_iter_cond_expr): Make inline instead of static.
11959 (debug (sreal &)): New.
11960 (debug (sreal *)): New.
11962 (extern debug (sreal &)): New.
11963 (extern debug (sreal *)): New.
11965 (extern debug_raw (const tree_node &)): New.
11966 (extern debug_raw (const tree_node *)): New.
11967 (extern debug (const tree_node &)): New.
11968 (extern debug (const tree_node *)): New.
11969 (extern debug_verbose (const tree_node &)): New.
11970 (extern debug_verbose (const tree_node *)): New.
11971 (extern debug_head (const tree_node &)): New.
11972 (extern debug_head (const tree_node *)): New.
11973 (extern debug_body (const tree_node &)): New.
11974 (extern debug_body (const tree_node *)): New.
11975 (extern debug (vec<tree, va_gc> &)): New.
11976 (extern debug (vec<tree, va_gc> *)): New.
11978 (debug (struct loop &)): New.
11979 (debug (struct loop *)): New.
11980 (debug_verbose (struct loop &)): New.
11981 (debug_verbose (struct loop *)): New.
11982 * tree-dump.c: Add header dependence.
11984 (extern debug (struct loop &)): New.
11985 (extern debug (struct loop *)): New.
11986 (extern debug_verbose (struct loop &)): New.
11987 (extern debug_verbose (struct loop *)): New.
11989 (debug (data_reference &)): New.
11990 (debug (data_reference *)): New.
11991 (debug (vec<data_reference_p> &)): New.
11992 (debug (vec<data_reference_p> *)): New.
11993 (debug (vec<ddr_p> &)): New.
11994 (debug (vec<ddr_p> *)): New.
11996 (extern debug (data_reference &)): New.
11997 (extern debug (data_reference *)): New.
11998 (extern debug (vec<data_reference_p> &)): New.
11999 (extern debug (vec<data_reference_p> *)): New.
12000 (extern debug (vec<ddr_p> &)): New.
12001 (extern debug (vec<ddr_p> *)): New.
12003 (debug (pt_solution &)): New.
12004 (debug (pt_solution *)): New.
12006 (extern debug (pt_solution &)): New.
12007 (extern debug (pt_solution *)): New.
12009 (debug (_var_map &)): New.
12010 (debug (_var_map *)): New.
12011 (debug (tree_live_info_d &)): New.
12012 (debug (tree_live_info_d *)): New.
12014 (extern debug (_var_map &)): New.
12015 (extern debug (_var_map *)): New.
12016 (extern debug (tree_live_info_d &)): New.
12017 (extern debug (tree_live_info_d *)): New.
12019 2013-03-28 Jan Hubicka <jh@suse.cz>
12021 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
12023 2013-03-28 Ian Bolton <ian.bolton@arm.com>
12025 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
12026 record only when desired or required.
12028 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
12030 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
12031 *vec_extractv2di_1_rex64. Use x64 isa attribute.
12033 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12035 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
12036 (*andsi3_compare0_uxtw): New pattern.
12037 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
12038 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
12040 2013-03-28 Jan Hubicka <jh@suse.cz>
12042 * data-streamer-in.c (streamer_read_gcov_count): New function.
12043 * gimple-streamer-out.c: Include value-prof.h.
12044 (output_gimple_stmt): Output histogram.
12045 (output_bb): Use streamer_write_gcov_count.
12046 * value-prof.c: Include data-streamer.h
12047 (dump_histogram_value): Add HIST_TYPE_MAX.
12048 (stream_out_histogram_value): New function.
12049 (stream_in_histogram_value): New function.
12050 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
12051 (stream_out_histogram_value, stream_in_histogram_value): Declare.
12052 * data-streamer-out.c (streamer_write_gcov_count): New function.
12053 (streamer_write_gcov_count_stream): New function.
12054 * lto-cgraph.c (lto_output_edge): Update counter streaming.
12055 (lto_output_node): Likewise.
12056 (input_node, input_edge): Likewise.
12057 * lto-streamer-out.c (output_cfg): Update streaming.
12058 * lto-streamer-in.c (input_cfg): Likewise.
12059 * data-streamer.h (streamer_write_gcov_count,
12060 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
12061 * gimple-streamer-in.c: Include value-prof.h
12062 (input_gimple_stmt): Input histograms.
12063 (input_bb): Update profile streaming.
12065 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12067 * genmodes.c (emit_max_int): New function.
12068 (emit_insn_modes_h): Added call to emit_max_function.
12069 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
12071 * machmode.def: Fixed comment.
12073 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12075 * combine.c (try_combine): Removed useless assert.
12076 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
12078 2013-03-28 Marek Polacek <polacek@redhat.com>
12079 Richard Biener <rguenther@suse.de>
12081 PR tree-optimization/56695
12082 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
12083 build signed result of a vector comparison.
12084 * tree-cfg.c (verify_gimple_comparison): Check that a result
12085 of a vector comparison has signed type.
12087 2013-03-28 Richard Biener <rguenther@suse.de>
12089 PR tree-optimization/37021
12090 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12091 do not restrict gaps between groups.
12092 * tree-vect-stmts.c (vectorizable_load): Properly account for
12093 a gap between groups.
12095 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12097 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12098 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12101 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12103 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12104 * doc/extend.texi (Named Address Spaces): Ditto.
12105 (Variable Attributes): Ditto.
12107 2013-03-27 Kai Tietz <ktietz@redhat.com>
12109 * config.build: Add support for cygwin x64 target.
12110 * config.gcc: Likewise.
12111 * config.host: Likewise.
12112 * configure.ac: Likewise
12113 * configure: Regenerated.
12115 2013-03-27 Kai Tietz <ktietz@redhat.com>
12117 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12118 * config/i386/t-cygwin-w64: New file.
12119 * config/i386/cygwin-w64.h: New file.
12120 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12121 and add support for x64-cygwin target.
12122 (CPP_SPEC): Likewise.
12123 (CXX_WRAP_SPEC_LIST): Undefine before define.
12124 (LIBGCJ_SONAME): Use 15 as version.
12126 2013-03-27 Richard Biener <rguenther@suse.de>
12128 PR tree-optimization/56716
12129 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12130 dumping for ref nodes.
12132 2013-03-27 Martin Jambor <mjambor@suse.cz>
12134 PR tree-optimization/55334
12135 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12136 restricted pointers to arrays.
12138 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12140 * Makefile.in (.SUFFIXES): Add .cc.
12141 (.c.o): Apply same recipe for implicit rule .cc.o.
12143 2013-03-27 Richard Biener <rguenther@suse.de>
12145 PR tree-optimization/37021
12146 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12147 REALPART/IMAGPART_EXPRs around the supported refs.
12148 * tree-ssa-structalias.c (find_func_aliases): Assume that
12149 floating-point values are not used to transfer pointers.
12151 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12153 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12154 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12155 * doc/tm.texi: Regenerate.
12156 * targhooks.h (default_has_ifunc_p): New.
12157 * targhooks.c (default_has_ifunc_p): Ditto.
12158 * config/linux-protos.h: New file.
12159 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
12160 this hook for linux which disables support of indirect functions in
12162 * config/linux-android.c: New file.
12163 * config/t-linux-android.c: Ditto.
12164 * config.gcc: Added new object file linux-android.o.
12165 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
12166 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
12167 * varasm.c (do_assemble_alias): Likewise.
12168 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12169 doesn't support indirect functions.
12170 * configure: Regenerate.
12172 2013-03-27 Bin Cheng <bin.cheng@arm.com>
12175 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12176 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12179 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12181 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12183 2013-03-27 Terry Guo <terry.guo@arm.com>
12185 * config/arm/arm-cores.def: Added core cortex-r7.
12186 * config/arm/arm-tune.md: Regenerated.
12187 * config/arm/arm-tables.opt: Regenerated.
12188 * doc/invoke.texi: Added entry for core cortex-r7.
12190 2013-03-27 Walter Lee <walt@tilera.com>
12192 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12193 double-decrement of next_scratch_regno.
12195 2013-03-27 Walter Lee <walt@tilera.com>
12197 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
12199 (insn_v1mulus): Ditto.
12200 (insn_v2muls): Ditto.
12202 2013-03-27 Walter Lee <walt@tilera.com>
12204 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
12205 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12207 2013-03-27 Walter Lee <walt@tilera.com>
12209 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
12210 (*sibcall_value): Ditto.
12212 2013-03-27 Walter Lee <walt@tilera.com>
12214 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12215 (insn_mnz_v8qi): ... this ...
12216 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12218 (insn_v<n>mnz): Replaced by ...
12219 (insn_v1mnz): ... this ...
12220 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12222 (insn_mz_<mode>): Replaced by ...
12223 (insn_mz_v8qi): ... this ...
12224 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12226 (insn_v<n>mz): Replaced by ...
12227 (insn_v1mz): ... this ...
12228 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12231 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12233 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12235 2013-03-26 Roland McGrath <mcgrathr@google.com>
12237 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12238 than fprintf with a non-constant, non-format string.
12240 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12242 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12243 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12244 operand 0 predicate.
12245 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12246 attribute. Use general_x64nomem_operand as operand 1 predicate.
12247 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12248 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12249 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12250 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12251 (mov<mode>_insv_1): Remove expander. Merge insn with
12252 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12253 Use general_x64nomem_operand as operand 1 predicate.
12254 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12255 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12256 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12257 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12258 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12259 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12260 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12261 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12262 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12263 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12264 (general_x64nomem_operand): Ditto.
12266 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
12268 * config/rtems.opt: Add -pthread option.
12270 2013-03-26 Richard Biener <rguenther@suse.de>
12272 * alias.c (find_base_term): Avoid redundant and not used recursion.
12273 (base_alias_check): Get the initial base term from the caller.
12274 (true_dependence_1): Compute and pass base terms to base_alias_check.
12275 (write_dependence_p): Likewise.
12276 (may_alias_p): Likewise.
12278 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12280 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12281 PC-relative load in SI modes and above only.
12283 2013-03-26 Xinyu Qi <xyqi@marvell.com>
12285 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
12286 * config/arm/iwmmxt.md (WCGR0): Update.
12287 (WCGR1, WCGR2, WCGR3): Likewise.
12289 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12291 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12292 Use x64 and nox64 isa attributes.
12294 2013-03-26 Richard Biener <rguenther@suse.de>
12296 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12297 alignment computations and rely on get_object_alignment_1
12298 for the !TYPE_P case.
12299 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12301 2013-03-26 Walter Lee <walt@tilera.com>
12303 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12304 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12306 2013-03-25 Jeff Law <law@redhat.com>
12308 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12309 check for INTEGRAL_TYPE_P that was missing due to checking in
12310 wrong version of prior patch.
12312 2013-03-25 Walter Lee <walt@tilera.com>
12314 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12315 TILEGX_INSN_SHUFFLEBYTES1.
12316 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12318 (tilegx_builtins): Ditto.
12319 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12321 2013-03-25 Walter Lee <walt@tilera.com>
12323 * config/tilegx/tilegx.md (floatsisf2): New pattern.
12324 (floatunssisf2): New pattern.
12325 (floatsidf2): New pattern.
12326 (floatunssidf2): New pattern.
12328 2013-03-25 Walter Lee <walt@tilera.com>
12330 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12331 tests for constraint J, K, N, P.
12333 2013-03-25 Walter Lee <walt@tilera.com>
12335 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12336 Use indirect/pcrel encoding.
12337 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12340 2013-03-25 Steve Ellcey <sellcey@mips.com>
12342 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12343 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12344 * config/mips/mips.c (mips_option_override): Set IMADD default.
12345 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12346 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12347 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12348 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12349 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12351 2013-03-25 Jeff Law <law@redhat.com>
12353 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12354 slightly to avoid creating and folding useless trees. Simplify
12355 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12357 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
12359 * config/i386/i386.md (*zero_extendsidi2): Merge with
12360 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12361 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12362 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12363 targets. Clarify comment.
12365 2013-03-25 Martin Jambor <mjambor@suse.cz>
12367 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12368 pass-through jump functions differently.
12369 (ipa_read_jump_function): Likewise. Also use setter functions to set
12372 2013-03-25 Martin Jambor <mjambor@suse.cz>
12374 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12375 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12377 (ipa_get_indirect_edge_target): New function.
12378 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12379 ipa_get_indirect_edge_target. Update all callers.
12380 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12381 ipa_get_indirect_edge_target_1 instead of calling
12382 ipa_get_indirect_edge_target.
12383 (create_specialized_node): Pass aggvlas to
12384 ipcp_discover_new_direct_edges.
12386 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12388 * config/arm/arm.md (f_sels, f_seld): New types.
12389 (*cmov<mode>): New pattern.
12390 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12393 2013-03-25 Kai Tietz <ktietz@redhat.com>
12395 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12396 POSIX-printf for mingw-hosted builds.
12398 2013-03-25 Richard Biener <rguenther@suse.de>
12400 PR middle-end/56694
12401 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12402 must-not-throw stmt location.
12404 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12406 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12407 Emit load-acquire versions when acq is true.
12408 (arm_emit_store_exclusive): Add rel parameter.
12409 Emit store-release versions when rel is true.
12410 (arm_split_compare_and_swap): Use acquire-release instructions
12412 of barriers when appropriate.
12413 (arm_split_atomic_op): Likewise.
12414 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12415 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12416 (VUNSPEC_SLX): Likewise.
12417 (VUNSPEC_LDA): Likewise.
12418 (VUNSPEC_STL): Likewise.
12419 * config/arm/sync.md (atomic_load<mode>): New pattern.
12420 (atomic_store<mode>): Likewise.
12421 (arm_load_acquire_exclusive<mode>): Likewise.
12422 (arm_load_acquire_exclusivesi): Likewise.
12423 (arm_load_acquire_exclusivedi): Likewise.
12424 (arm_store_release_exclusive<mode>): Likewise.
12426 2013-03-25 Catherine Moore <clm@codesourcery.com>
12428 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12429 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12430 * config/mip/predicates.md (lwsp_swsp_operand,
12431 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12432 sb16_operand, db4_operand, db7_operand, ib3_operand,
12433 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
12434 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
12435 andi16_operand): New predicates.
12436 * config/mips/mips.md (compression): New attribute.
12437 (enabled): New attribute.
12438 (length): Consider compression in computing length.
12439 (shift_compression): New code attribute.
12440 (*add<mode>3): New operands. Record compression.
12441 (sub<mode>3): Likewise.
12442 (one_cmpl<mode>2): Likewise.
12443 (*and<mode>3): Likewise.
12444 (*ior<mode>3): Likewise.
12445 (unnamed pattern for xor): Likewise.
12446 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
12447 (*<optab><mode>3): Likewise.
12448 (*mov<mode>_internal: Likewise.
12449 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
12450 (mips_unsigned_immediate_p): New.
12451 (umips_lwsp_swsp_address_p): New.
12452 (m16_based_address_p): New.
12453 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
12454 (mips_unsigned_immediate_p): New prototype.
12455 (lwsp_swsp_address_p): New prototype.
12456 (m16_based_address_p): New prototype.
12457 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
12458 (mips_signed_immediate_p): New function.
12459 (m16_based_address_p): New function.
12460 (lwsp_swsp_address_p): New function.
12461 (mips_print_operand_punctuation): Recognize short delay slot insns
12462 for microMIPS.add<mode>3"
12464 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12467 * config/arm/iterators.md (v_cmp_result): New mode attribute.
12468 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
12470 2013-03-25 Richard Biener <rguenther@suse.de>
12472 PR tree-optimization/56689
12473 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
12476 2013-03-25 Richard Biener <rguenther@suse.de>
12478 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
12480 (memory_references): Likewise.
12481 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
12482 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
12483 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
12484 (gather_mem_refs_in_loops): Fold into ...
12485 (analyze_memory_references): ... this. Move initialization
12486 to tree_ssa_lim_initialize.
12487 (fill_always_executed_in): Rename to ...
12488 (fill_always_executed_in_1): ... this.
12489 (fill_always_executed_in): Move contains_call computation to
12490 this new function from ...
12491 (tree_ssa_lim_initialize): ... here.
12492 (tree_ssa_lim): Call fill_always_executed_in.
12494 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12496 * postreload.c (reload_combine): Fix code detecting returns.
12498 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12500 * function.c (emit_use_return_register_into_block): On cc0 targets,
12501 do not emit the sequence between cc0 setter and user.
12503 2013-03-25 Kai Tietz <ktietz@redhat.com>
12505 * config/i386/predicates.md (local_symbolic_operand): Interpret
12506 dll-imported symbols as none-local.
12508 2013-03-25 Richard Biener <rguenther@suse.de>
12510 * tree-ssa-loop-im.c (struct depend): Remove.
12511 (struct lim_aux_data): Make depends a vec of gimples.
12512 (free_lim_aux_data): Adjust.
12513 (add_dependency): Likewise.
12514 (set_level): Likewise.
12516 2013-03-25 Richard Biener <rguenther@suse.de>
12518 PR middle-end/56434
12519 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
12520 the pointer returned by calls with ECF_MALLOC set.
12522 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12524 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
12526 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12528 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
12529 using MMXMODE mode iterator.
12530 (*move<mode>_internal): Merge with *movv2sf_internal and
12531 *movv2sf_internal_rex64 using MMXMODE mode iterator.
12533 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12535 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
12536 (record_last_mem_set_info): Likewise.
12538 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
12539 of XNEWVEC followed by memset.
12540 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
12542 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12544 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
12545 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
12546 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
12547 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
12548 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
12549 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
12550 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
12551 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
12552 BARRIER_P instead of GET_CODE.
12554 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12556 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
12557 inaccuracy in the probing code.
12559 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
12560 (ctrapdi4): Likewise.
12562 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12564 * calls.c (expand_call): Add missing guard to code handling return
12565 of non-BLKmode structures in MSB.
12566 * function.c (expand_function_end): Likewise.
12568 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12570 * combine.c (try_combine): Adjust comment. Do not add the set of
12571 insn #0 if the destination indirectly is set or dies in insn #2.
12572 Tidy up code to distribute a new note.
12574 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12576 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
12577 also for alternatives 16 and 17.
12579 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12581 * config/i386/sse.md (*mov<mode>_internal): Merge with
12582 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
12583 Emit insn template depending on type attribute. Use
12584 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12585 movd instead of movq mnemonic for interunit moves. Rewrite mode
12586 attribute calculation. Remove unit attribute calculation.
12587 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
12588 Set prefix_data16 attribute for DImode ssemov types.
12589 Use Ym instead of y for SSE-MMX conversion alternatives.
12590 Reorder operand constraints.
12592 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
12594 * df.h (df_insn_delete): Adjust prototype.
12595 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
12596 and let it decide whether mark the basic block dirty.
12597 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
12598 * df-scan.c (df_insn_info_delete): New helper function, split
12599 off from df_insn_delete.
12600 (df_scan_free_bb_info): Use it.
12601 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
12603 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
12604 that the insn is actually an insn and it has a non-NULL basic block.
12605 Do not mark basic block dirty if only deleting a DEBUG_INSN.
12607 2013-03-22 Richard Biener <rguenther@suse.de>
12609 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
12611 (mem_ref_alloc): Do not allocate them.
12612 (refs_independent_p): Do not query or maintain a cache.
12614 2013-03-22 Richard Biener <rguenther@suse.de>
12616 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
12617 (gather_mem_refs_in_loops): Do not compute it.
12618 (analyze_memory_references): Do not allocate it.
12619 (tree_ssa_lim_finalize): Do not free it.
12620 (for_all_locs_in_loop): Do not query all_refs_in_loop.
12622 2013-03-22 Richard Biener <rguenther@suse.de>
12624 * is-a.h (as_a): Use gcc_checking_assert.
12626 2013-03-22 Ian Bolton <ian.bolton@arm.com>
12628 * config/aarch64/aarch64.c (aarch64_print_operand): New
12629 format specifier for printing a constant in hex.
12630 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
12631 format specifier for printing second operand.
12633 2013-03-22 Richard Biener <rguenther@suse.de>
12635 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
12637 (gather_mem_refs_in_loops): Perform store accumulation here.
12638 (create_vop_ref_mapping_loop): Remove.
12639 (create_vop_ref_mapping): Likewise.
12640 (analyze_memory_references): Initialize refs_stored_in_loop.
12641 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
12642 (record_indep_loop): Remove.
12643 (record_dep_loop): New function.
12644 (ref_indep_loop_p_1): Adjust to only walk over references
12645 in the loop, not its subloops.
12646 (ref_indep_loop_p): Rename to ...
12647 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
12648 maintaining a more fine-grained cache.
12649 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
12650 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
12652 2013-03-22 Richard Biener <rguenther@suse.de>
12654 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
12655 (struct mem_ref): Make accesses_in_loop a vec of a vec of
12656 aggregate mem_ref_loc.
12657 (free_mem_ref_locs): Inline into ...
12658 (memref_free): ... this and adjust.
12659 (mem_ref_alloc): Adjust.
12660 (mem_ref_locs_alloc): Remove.
12661 (record_mem_ref_loc): Adjust.
12662 (get_all_locs_in_loop): Rewrite into ...
12663 (for_all_locs_in_loop): ... this iterator.
12664 (rewrite_mem_ref_loc): New functor.
12665 (rewrite_mem_refs): Use for_all_locs_in_loop.
12666 (sm_set_flag_if_changed): New functor.
12667 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
12668 (ref_always_accessed): New functor.
12669 (ref_always_accessed_p): Use for_all_locs_in_loop.
12671 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12673 * tree-pass.h (PROP_gimple_lvec): New.
12674 * passes.c (dump_properties): Handle PROP_gimple_lvec.
12675 (init_optimization_passes): Move pass_lower_vector.
12676 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
12678 (pass_lower_vector): Provide PROP_gimple_lvec.
12679 (pass_lower_vector_ssa): Likewise.
12680 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
12682 2013-03-21 Mark Wielaard <mjw@redhat.com>
12684 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
12686 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
12688 * config/i386/i386.md (*movdi_internal): Disparage slightly
12689 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
12690 conversion alternatives.
12692 2013-03-21 Jakub Jelinek <jakub@redhat.com>
12694 PR middle-end/48087
12695 * diagnostic.def (DK_WERROR): New kind.
12696 * diagnostic.h (werrorcount): Define.
12697 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
12698 promoted to DK_ERROR, increment DK_WERROR counter instead of
12700 * toplev.c (toplev_main): Call print_ignored_options even if
12701 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
12702 even if just werrorcount is non-zero.
12705 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
12707 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
12708 (string_cst_pool_decl): New function.
12709 (optimize_one_addr_into_implicit_ptr): New function.
12710 (resolve_addr_in_expr): Optimize DWARF location expression
12711 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
12712 which doesn't live in memory, but has DW_AT_location or
12713 DW_AT_const_value, or refers to a string literal, into
12714 DW_OP_GNU_implicit_pointer.
12715 (optimize_location_into_implicit_ptr): New function.
12716 (resolve_addr): If removing DW_AT_location of a variable because
12717 it was DW_OP_addr of address of the variable, but the variable doesn't
12718 live in memory, try to emit const value attribute for the initializer.
12720 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12722 * tree.h (VECTOR_TYPE_P): New macro.
12723 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
12724 TYPE_MODE): Use it.
12725 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
12726 VEC_COND_EXPR cannot be lvalues.
12727 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
12729 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12731 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
12732 Restrict the transformation to equal modes.
12734 2013-03-21 Richard Biener <rguenther@suse.de>
12736 PR tree-optimization/39326
12737 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
12738 (MEM_ANALYZABLE): Adjust.
12739 (record_mem_ref_loc): Move bitmap ops ...
12740 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
12741 unanalyzable refs, do not record locations for it.
12742 (analyze_memory_references): Allocate ref zero as shared
12744 (refs_independent_p): Do not test for unanalyzed mems here.
12745 (ref_indep_loop_p_1): Special-case disambiguation against
12746 the unanalyzed ref.
12747 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
12749 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
12751 * config/arm/arm-protos.h (tune_params): Add
12752 prefer_neon_for_64bits field.
12753 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
12754 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
12755 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
12756 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
12757 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
12758 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
12759 (arm_option_override): Handle -mneon-for-64bits new option.
12760 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
12761 (prefer_neon_for_64bits): Declare new variable.
12762 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
12763 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
12764 (arch_enabled): Handle new arch types. Remove support for onlya8
12766 (one_cmpldi2): Use new arch names.
12767 (zero_extend<mode>di2, extend<mode>di2): Ditto.
12768 * config/arm/arm.opt (mneon-for-64bits): Add option.
12769 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
12770 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
12771 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
12773 * doc/invoke.texi (-mneon-for-64bits): Document.
12775 2013-03-21 Richard Biener <rguenther@suse.de>
12777 PR tree-optimization/39326
12778 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
12779 (sort_bbs_in_loop_postorder_cmp): New function.
12780 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
12782 2013-03-21 Richard Biener <rguenther@suse.de>
12784 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
12785 (vect_insert_into_interleaving_chain): Likewise.
12786 (vect_drs_dependent_in_basic_block): Inline ...
12787 (vect_slp_analyze_data_ref_dependence): ... here. New function,
12789 (vect_analyze_data_ref_dependence): ... here. Simplify.
12790 (vect_check_interleaving): Simplify.
12791 (vect_analyze_data_ref_dependences): Likewise. Split out ...
12792 (vect_slp_analyze_data_ref_dependences): ... this new function.
12793 (dr_group_sort_cmp): New function.
12794 (vect_analyze_data_ref_accesses): Compute data-reference groups
12795 here instead of in vect_analyze_data_ref_dependence. Use
12796 a more efficient algorithm.
12797 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
12798 vect_slp_analyze_data_ref_dependences. Call
12799 vect_analyze_data_ref_accesses earlier.
12800 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
12801 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
12802 (vect_slp_analyze_data_ref_dependences): New prototype.
12804 2013-03-21 Richard Biener <rguenther@suse.de>
12806 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
12807 ref is stored in the loop.
12808 (find_refs_for_sm): Walk only over all stores.
12809 (store_motion_loop): Allocate from lim_bitmap_obstack.
12810 (store_motion): Likewise.
12812 2013-03-21 Richard Biener <rguenther@suse.de>
12814 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
12815 Update virtual SSA form.
12817 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12819 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
12820 * configure: Regenerate.
12821 * config.in: Regenerate.
12822 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
12823 if !HAVE_LD_EH_FRAME_CIEV3.
12825 2013-03-21 Richard Biener <rguenther@suse.de>
12827 * tree-cfg.c (verify_expr_no_block): New function.
12828 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
12829 nor DECL_VALUE_EXPR have locations with associated blocks.
12830 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
12831 (clear_unused_block_pointer): Remove code dealing with
12832 blocks in DECL_DEBUG_EXPR locations.
12834 2013-03-21 Richard Biener <rguenther@suse.de>
12836 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
12837 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
12838 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
12839 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
12840 instead of DECL_DEBUG_EXPR_IS_FROM.
12841 * gimplify.c (gimplify_modify_expr): Likewise.
12842 * tree-cfg.c (verify_expr_location_1): Likewise.
12843 * tree-complex.c (create_one_component_var): Likewise.
12844 * tree-sra.c (create_access_replacement): Likewise.
12845 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
12846 (clear_unused_block_pointer): Likewise.
12847 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
12848 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
12849 * var-tracking.c (var_debug_decl): Likewise.
12850 (track_expr_p): Likewise.
12851 * tree-inline.c (add_local_variables): Likewise. Set
12852 DECL_HAS_DEBUG_EXPR_P after copying it.
12853 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
12854 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
12856 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
12859 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
12860 * configure: Regenerate.
12861 * config.in: Regenerate.
12862 * config/i386/i386.md (*movdf_internal): Use
12863 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12864 movd instead of movq mnemonic for interunit moves.
12865 (*movdi_internal): Ditto.
12867 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12869 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
12870 (abd<mode>_3): New pattern.
12871 (aba<mode>_3): New pattern.
12872 (fabd<mode>_3): New pattern.
12874 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12876 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
12877 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
12878 occurrence of REGISTER_PREFIX as its empty string.
12880 2013-03-20 Jeff Law <law@redhat.com>
12882 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
12883 addititional equivalences for equality comparisons between an SSA_NAME
12884 and a constant where the SSA_NAME was set from a widening conversion.
12886 2013-03-20 Walter Lee <walt@tilera.com>
12888 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
12890 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
12892 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
12893 depending on type attribute.
12894 (*movti_internal): Ditto.
12895 (*movtf_internal): Ditto.
12896 (*movxf_internal): Ditto.
12897 (*movdf_internal): Ditto.
12898 (*movsf_internal): Ditto.
12900 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
12902 * config/i386/i386.md (*movti_internal): Set prefix attribute to
12903 maybe_vex for sselog1 and ssemov types.
12904 (*movdi_internal): Reorder operand constraints.
12905 (*movsi_internal): Ditto. Set prefix attribute to
12906 maybe_vex for sselog1 and ssemov types.
12907 (*movtf_internal): Set prefix attribute to maybe_vex
12908 for sselog1 and ssemov types.
12909 (*movdf_internal): Ditto. Set prefix_data16 attribute for
12910 DImode ssemov types. Reorder operand constraints.
12911 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
12912 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
12913 attribute for SImode ssemov types. Reorder operand constraints.
12915 2013-03-20 Martin Jambor <mjambor@suse.cz>
12917 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
12918 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
12920 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
12922 * config/rs6000/predicates.md (indexed_address, update_address_mem
12923 update_indexed_address_mem): New predicates.
12924 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
12925 attribute for load/store instructions.
12926 * config/rs6000/dfp.md (movsd_store): Likewise.
12927 (movsd_load): Likewise.
12928 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
12929 (unnamed HI->DI extend define_insn): Likewise.
12930 (unnamed SI->DI extend define_insn): Likewise.
12931 (unnamed QI->SI extend define_insn): Likewise.
12932 (unnamed QI->HI extend define_insn): Likewise.
12933 (unnamed HI->SI extend define_insn): Likewise.
12934 (unnamed HI->SI extend define_insn): Likewise.
12935 (extendsfdf2_fpr): Likewise.
12936 (movsi_internal1): Likewise.
12937 (movsi_internal1_single): Likewise.
12938 (movhi_internal): Likewise.
12939 (movqi_internal): Likewise.
12940 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
12941 attribute for load/store instructions.
12942 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
12944 (mov<mode>_softfloat): Likewise.
12945 (mov<mode>_hardfloat32): Likewise.
12946 (mov<mode>_hardfloat64): Likewise.
12947 (mov<mode>_softfloat64): Likewise.
12948 (movdi_internal32): Likewise.
12949 (movdi_internal64): Likewise.
12950 (probe_stack_<mode>): Likewise.
12952 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
12954 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
12955 floating point, and decimal floating point to reload iterator.
12957 * config/rs6000/constraints.md (wl constraint): New constraints to
12958 return FLOAT_REGS if certain options are used to reduce the number
12959 of separate patterns that exist in the file.
12960 (wx constraint): Likewise.
12961 (wz constraint): Likewise.
12963 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
12964 -mdebug=reg, print wg, wl, wx, and wz constraints.
12965 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
12966 Initialize the reload functions for 64-bit binary/decimal floating
12968 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
12969 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
12970 create the buffer on the stack to overcome not having a 32-bit
12972 (rs6000_emit_move): Likewise.
12973 (rs6000_secondary_memory_needed_rtx): Likewise.
12974 (rs6000_alloc_sdmode_stack_slot): Likewise.
12975 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
12976 via xxlxor, just like DFmode 0.0.
12978 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
12979 define as 1 if we are running on a power7 or newer.
12980 (enum r6000_reg_class_enum): Add new constraints.
12982 * config/rs6000/dfp.md (movsd): Delete, combine with binary
12983 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
12984 with other moves by using conditional constraits (wg). Use LFIWZX
12985 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
12986 (movsd splitter): Likewise.
12987 (movsd_hardfloat): Likewise.
12988 (movsd_softfloat): Likewise.
12990 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
12991 binary and decimal floating point moves.
12992 (fmove_ok): New attributes to combine binary and decimal floating
12993 point moves, and to combine power6x (mfpgpr) moves along normal
12995 (real_value_to_target): Likewise.
12996 (f32_lr): Likewise.
12997 (f32_lm): Likewise.
12998 (f32_li): Likewise.
12999 (f32_sr): Likewise.
13000 (f32_sm): Likewise.
13001 (f32_si): Likewise.
13002 (movsf): Combine binary and decimal floating point moves. Combine
13003 power6x (mfpgpr) moves with other moves by using conditional
13004 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
13005 (mov<mode> for SFmode/SDmode); Likewise.
13006 (SFmode/SDmode splitters): Likewise.
13007 (movsf_hardfloat): Likewise.
13008 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
13009 (movsf_softfloat): Likewise.
13010 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
13012 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
13013 wx and wz constraints.
13015 * config/rs6000/constraints.md (wg constraint): New constraint to
13016 return FLOAT_REGS if -mmfpgpr (power6x) was used.
13018 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
13021 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13022 -mdebug=reg, print wg, wl, wx, and wz constraints.
13023 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13024 Initialize the reload functions for 64-bit binary/decimal floating
13026 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13027 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13028 create the buffer on the stack to overcome not having a 32-bit
13030 (rs6000_emit_move): Likewise.
13031 (rs6000_secondary_memory_needed_rtx): Likewise.
13032 (rs6000_alloc_sdmode_stack_slot): Likewise.
13033 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13034 via xxlxor, just like DFmode 0.0.
13036 * config/rs6000/dfp.md (movdd): Delete, combine with binary
13037 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13038 with other moves by using conditional constraits (wg). Use LFIWZX
13039 and STFIWX for loading SDmode on power7.
13040 (movdd splitters): Likewise.
13041 (movdd_hardfloat32): Likewise.
13042 (movdd_softfloat32): Likewise.
13043 (movdd_hardfloat64_mfpgpr): Likewise.
13044 (movdd_hardfloat64): Likewise.
13045 (movdd_softfloat64): Likewise.
13047 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
13048 64-bit binary and decimal floating point moves.
13049 (FMOVE64X): Likewise.
13050 (movdf): Combine 64-bit binary and decimal floating point moves.
13051 Combine power6x (mfpgpr) moves with other moves by using
13052 conditional constraits (wg).
13053 (mov<mode> for DFmode/DDmode): Likewise.
13054 (DFmode/DDmode splitters): Likewise.
13055 (movdf_hardfloat32): Likewise.
13056 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
13057 (movdf_softfloat32): Likewise.
13058 (movdf_hardfloat64_mfpgpr): Likewise.
13059 (movdf_hardfloat64): Likewise.
13060 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
13061 (movdf_softfloat64): Likewise.
13062 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
13063 (reload_<mode>_load): Move to later in the file so they aren't in
13064 the middle of the floating point move insns.
13065 (reload_<mode>_store): Likewise.
13067 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
13070 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
13071 constraint if -mdebug=reg.
13072 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
13073 Enable using dd reload support if needed.
13075 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
13076 binary and decimal floating point moves in rs6000.md.
13077 (movtd_internal): Likewise.
13079 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
13080 decimal floating point moves.
13082 (movtf_internal): Likewise.
13083 (mov<mode>_internal, TDmode/TFmode): Likewise.
13084 (movtf_softfloat): Likewise.
13085 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
13087 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13088 movdi_internal64, using wg constraint for move direct operations.
13089 (movdi_internal64): Likewise.
13091 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13092 MODES_TIEABLE_P for selected modes. Print the numerical value of
13093 the various virtual registers. Use GPR/FPR first/last values,
13094 instead of hard coding the register numbers. Print which modes
13095 have reload functions registered.
13096 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13097 settings before/after setting cpu, target and subtarget settings.
13098 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13099 and for secondary reload failures in rs6000_secondary_reload_inner.
13100 (rs6000_secondary_reload_fail): Likewise.
13101 (rs6000_secondary_reload_inner): Likewise.
13103 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13104 macros for first/last GPR and FPR registers.
13105 (LAST_GPR_REGNO): Likewise.
13106 (FIRST_FPR_REGNO): Likewise.
13107 (LAST_FPR_REGNO): Likewise.
13109 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13110 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13111 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13112 (vcond<mode><mode>): Likewise.
13113 (vcondu<mode><mode>): Likewise.
13114 (vector_gtu<mode>): Likewise.
13115 (vector_gte<mode>): Likewise.
13116 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
13117 to prevent the compiler from converting DImode operations to TImode.
13118 (ior<mode>3): Likewise.
13119 (and<mode>3): Likewise.
13120 (one_cmpl<mode>2): Likewise.
13121 (nor<mode>3): Likewise.
13122 (andc<mode>3): Likewise.
13124 * config/rs6000/constraints.md (wt constraint): New constraint
13125 that returns VSX_REGS if TImode is allowed in VSX registers.
13127 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13128 constant under VSX.
13130 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13131 similar to TImode, but it is restricted to being in the GPRs.
13133 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13134 TImode to occupy a single VSX register.
13136 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13137 -mvsx-timode for power7/power8.
13138 (power7 cpu): Likewise.
13139 (power8 cpu): Likewise.
13141 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13142 sure that TFmode/TDmode take up two registers if they are ever
13143 allowed in the upper VSX registers.
13144 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13146 (rs6000_init_hard_regno_mode_ok): Likewise.
13147 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13148 constraint. Print if LRA is turned on.
13149 (rs6000_option_override_internal): Give an error if -mvsx-timode
13150 and VSX is not enabled.
13151 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13152 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13153 to reg+offset addressing. Use PTImode when checking offset
13154 addresses for validity.
13155 (reg_offset_addressing_ok_p): Likewise.
13156 (rs6000_legitimate_offset_address_p): Likewise.
13157 (rs6000_legitimize_address): Likewise.
13158 (rs6000_legitimize_reload_address): Likewise.
13159 (rs6000_legitimate_address_p): Likewise.
13160 (rs6000_eliminate_indexed_memrefs): Likewise.
13161 (rs6000_emit_move): Likewise.
13162 (rs6000_secondary_reload): Likewise.
13163 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13164 reloads to fpr registers to continue to use reg+offset addressing,
13165 but 64-bit reloads to altivec registers need reg+reg addressing.
13166 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13167 it. Treat LO_SUM like a PLUS operation.
13168 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
13169 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
13170 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13171 registers to share a register with a smaller sized type, since VSX
13172 puts scalars in the upper 64-bits.
13173 (print_operand): Add support for PTImode.
13174 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13175 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13176 registers, but don't have arithmetic support.
13177 (rs6000_memory_move_cost): Add test for VSX.
13178 (rs6000_opt_masks): Add -mvsx-timode.
13180 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13183 (VSr): Use wt constraint for TImode.
13184 (VSv): Drop TImode support.
13185 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13186 (vsx_movti_64bit): Likewise.
13187 (vsx_movti_32bit): Likewise.
13188 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13189 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13190 one '?' on the appropriate output constraint. Do not allow TImode
13191 logical operations on 32-bit systems.
13192 (vsx_ior<mode>3): Likewise.
13193 (vsx_xor<mode>3): Likewise.
13194 (vsx_one_cmpl<mode>2): Likewise.
13195 (vsx_nor<mode>3): Likewise.
13196 (vsx_andc<mode>3): Likewise.
13197 (vsx_concat_<mode>): Likewise.
13198 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13200 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13201 OPTION_MASK_VSX_TIMODE.
13202 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13203 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13205 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13206 (TI2 iterator): New iterator for TImode, PTImode.
13207 (wd mode attribute): Add values for vector types.
13208 (movti_string): Replace TI move operations with operations for TImode
13209 and PTImode. Add support for TImode being allowed in VSX registers.
13210 (mov<mode>_string, TImode/PTImode): Likewise.
13211 (movti_ppc64): Likewise.
13212 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13213 (TI mode splitters): Likewise.
13215 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13218 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
13220 PR tree-optimization/56355
13221 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13222 Also handle integers with undefined overflow.
13224 2013-03-20 Catherine Moore <clm@codesourcery.com>
13225 Maciej W. Rozycki <macro@codesourcery.com>
13226 Tom de Vries <tom@codesourcery.com>
13227 Nathan Sidwell <nathan@codesourcery.com>
13228 Iain Sandoe <iain@codesourcery.com>
13229 Nathan Froyd <froydnj@codesourcery.com>
13230 Chao-ying Fu <fu@mips.com>
13232 * doc/extend.texi: (micromips, nomicromips, nocompression):
13233 Document new function attributes.
13234 * doc/invoke.texi (minterlink-compressed, mmicromips,
13235 m14k, m14ke, m14kec): Document new options.
13236 (minterlink-mips16): Update documentation.
13237 * doc/md.texi (ZC, ZD): Document new constraints.
13238 * configure.ac (gcc_cv_as_micromips): Check if linker
13239 supports the .set micromips directive.
13240 * configure: Regenerate.
13241 * config.in: Regenerate.
13242 * config/mips/mips-tables.opt: Regenerate.
13243 * config/mips/micromips.md: New file.
13244 * constraints.md (ZC, ZD): New constraints.
13245 * config/mips/predicates.md (movep_src_register): New predicate.
13246 (movep_src_operand): New predicate.
13247 (non_volatile_mem_operand): New predicate.
13248 * config/mips/mips.md (multimem): New type.
13249 (length): Differentiate between 17-bit and 18-bit branch offsets.
13250 (MOVEP1, MOVEP2): New mode iterator.
13251 (mov_<load>l): Use ZC constraint.
13252 (mov_<load>r): Likewise.
13253 (mov_<store>l): Likewise.
13254 (mov_<store>r): Likewise.
13255 (*branch_equality<mode>_inverted): Add microMIPS support.
13256 (*branch_equality<mode>): Likewise.
13257 (*jump_absolute): Likewise.
13258 (indirect_jump_<mode>): Likewise.
13259 (tablejump_<mode>): Likewise.
13260 (<optab>_internal): Likewise.
13261 (sibcall_internal): Likewise.
13262 (sibcall_value_internal): Likewise.
13263 (prefetch): Use constraint ZD.
13264 * config/mips/mips.opt (minterlink-compressed): New option.
13265 (minterlink-mips16): Now an alias for minterlink-compressed.
13266 (mmicromips): New option.
13267 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13268 (compare_and_swap_12): Likewise.
13269 (sync_add<mode>): Likewise.
13270 (sync_<optab>_12): Likewise.
13271 (sync_old_<optab>_12): Likewise.
13272 (sync_new_<optab>_12): Likewise.
13273 (sync_nand_12): Likewise.
13274 (sync_old_nand_12): Likewise.
13275 (sync_new_nand_12): Likewise.
13276 (sync_sub<mode>): Likewise.
13277 (sync_old_add<mode>): Likewise.
13278 (sync_old_sub<mode>): Likewise.
13279 (sync_new_add<mode>): Likewise.
13280 (sync_new_sub<mode>): Likewise.
13281 (sync_<optab><mode>): Likewise.
13282 (sync_old_<optab><mode>): Likewise.
13283 (sync_new_<optab><mode>): Likewise.
13284 (sync_nand<mode>): Likewise.
13285 (sync_old_nand<mode>): Likewise.
13286 (sync_new_nand<mode>): Likewise.
13287 (sync_lock_test_and_set<mode>): Likewise.
13288 (test_and_set_12): Likewise.
13289 (atomic_compare_and_swap<mode>): Likewise.
13290 (atomic_exchange<mode>_llsc): Likewise.
13291 (atomic_fetch_add<mode>_llsc): Likewise.
13292 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13293 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13294 (umips_save_restore_pattern_p): Likewise.
13295 (umips_load_store_pair_p): Likewise.
13296 (umips_output_load_store_pair): Likewise.
13297 (umips_movep_target_p): Likewise.
13298 (umips_12bit_offset_address_p): Likewise.
13299 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13300 (mips_base_mips16): Rename this...
13301 (mips_base_compression_flags): ...to this. Update all uses.
13302 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13303 (mips_mips16_decl_p): Delete.
13304 (mips_nomips16_decl_p): Delete.
13305 (mips_get_compress_on_flags): New function.
13306 (mips_get_compress_off_flags): New function.
13307 (mips_get_compress_mode): New function.
13308 (mips_get_compress_on_name): New function.
13309 (mips_get_compress_off_name): New function.
13310 (mips_insert_attributes): Support multiple compression types.
13311 (mips_merge_decl_attributes): Likewise.
13312 (umips_12bit_offset_address_p): New function.
13313 (mips_start_function_definition): Emit .set micromips directive.
13314 (mips_call_may_need_jalx_p): New function.
13315 (mips_function_ok_for_sibcall): Add microMIPS support.
13316 (mips_print_operand_punctuation): Support short delay slots and
13318 (umips_swm_mask, umips_swm_encoding): New.
13319 (umips_build_save_restore): New function.
13320 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13321 (was_mips16_p): Remove.
13322 (old_compression_mode): New.
13323 (mips_set_compression_mode): New function.
13324 (mips_set_current_function): Add microMIPS support.
13325 (mips_option_override): Likewise.
13326 (umips_save_restore_pattern_p): New function.
13327 (umips_output_save_restore): New function.
13328 (umips_load_store_pair_p_1): New function.
13329 (umips_load_store_pair_p): New function.
13330 (umips_output_load_store_pair_1): New function.
13331 (umips_output_load_store_pair): New function.
13332 (umips_movep_target_p) New function.
13333 (mips_prepare_pch_save): Add microMIPS support.
13334 * config/mips/mips.h (TARGET_COMPRESSION): New.
13335 (TARGET_CPU_CPP_BUILTINS): Update macro
13336 to use new compression flags and to support microMIPS.
13337 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13338 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13339 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13340 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13341 (ASM_SPEC): Support mmicromips and mno-micromips.
13342 (M16STORE_REG_P): New macro.
13343 (MIPS_CALL): Support TARGET_MICROMIPS.
13344 (MICROMIPS_J): New macro.
13345 (mips_base_mips16): Rename this...
13346 (mips_base_compression_flags): ...to this.
13347 (UMIPS_12BIT_OFFSET_P): New macro.
13348 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13349 (MULTILIB_DIRNAMES): Likewise.
13350 2013-03-20 Richard Biener <rguenther@suse.de>
13352 PR tree-optimization/56661
13353 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13354 the result does not have to be distinct.
13356 2013-03-20 Richard Biener <rguenther@suse.de>
13358 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13361 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13362 Steven Bosscher <steven@gcc.gnu.org>
13364 PR rtl-optimization/56605
13365 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13367 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13370 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13371 that require movd instead of movq.
13373 2013-03-20 Richard Biener <rguenther@suse.de>
13375 * tree-ssa-structalias.c (struct variable_info): Add pointer
13376 to the first field of an aggregate with sub-vars. Make
13377 this and the pointer to the next subfield its ID.
13378 (vi_next): New function.
13379 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13380 storedanything_id, integer_id): Increment by one.
13381 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13382 get_call_clobber_vi): Adjust.
13383 (solution_set_expand): Simplify and speedup.
13384 (solution_set_add): Inline into ...
13385 (set_union_with_increment): ... this. Adjust accordingly.
13386 (do_sd_constraint): Likewise.
13387 (do_ds_constraint): Likewise.
13388 (do_complex_constraint): Simplify.
13389 (build_pred_graph): Adjust.
13390 (solve_graph): Likewise. Simplify and speedup.
13391 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13392 get_constraint_for_component_ref, get_constraint_for_1,
13393 first_vi_for_offset, first_or_preceding_vi_for_offset,
13394 create_function_info_for, create_variable_info_for_1,
13395 create_variable_info_for, intra_create_variable_infos): Adjust.
13396 (init_base_vars): Push NULL for ID zero.
13397 (compute_points_to_sets): Adjust.
13399 2013-03-20 Richard Biener <rguenther@suse.de>
13401 * cfgloop.c (verify_loop_structure): Streamline and avoid
13402 ICEing on corrupt loop tree.
13403 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13406 2013-03-20 Richard Biener <rguenther@suse.de>
13408 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13409 check whether an SSA update is needed.
13411 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13413 * config/mips/constraints.md (T): Rename to...
13417 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13418 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13420 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13422 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13423 (*subsi3_carryin_uxtw): Likewise.
13425 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13427 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13428 (*rorsi3_insn_uxtw): Likewise.
13430 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13432 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
13433 (*extrsi5_insn_uxtw): Likewise.
13435 2013-03-19 Richard Biener <rguenther@suse.de>
13437 PR tree-optimization/56273
13438 * passes.c (init_optimization_passes): Move second VRP after DOM.
13440 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
13442 * config/i386/i386.md (*movti_internal): Merge from
13443 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
13444 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
13445 nox64 isa attributes.
13447 2013-03-18 Richard Biener <rguenther@suse.de>
13449 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
13451 (merge_node_constraints): Likewise.
13452 (build_succ_graph): Likewise.
13453 (valid_graph_edge): Inline into single caller.
13454 (unify_nodes): Likewise. Use bitmap_set_bit return value
13456 (scc_visit): Fix formatting and variable use.
13457 (do_sd_constraint): Use gcc_checking_assert.
13458 (do_ds_constraint): Likewise.
13459 (do_complex_constraint): Likewise.
13460 (condense_visit): Likewise. Cleanup.
13461 (dump_pred_graph): New function.
13462 (perform_var_substitution): Dump the pred-graph before
13463 variable substitution.
13464 (find_equivalent_node): Use gcc_checking_assert.
13465 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
13467 2013-03-18 Richard Biener <rguenther@suse.de>
13469 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13470 Remove cond_expr_stmt_list argument and do not gimplify the
13472 (vect_loop_versioning): Adjust.
13473 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
13474 Cleanup to use less temporaries.
13475 (vect_create_data_ref_ptr): Cleanup.
13477 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13479 PR tree-optimization/56635
13480 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
13481 require types_compatible_p types.
13483 2013-03-18 Nick Clifton <nickc@redhat.com>
13485 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
13486 spurious backslash.
13488 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
13489 Add missing line to comment describing function.
13491 2013-03-18 Richard Biener <rguenther@suse.de>
13493 PR tree-optimization/56210
13494 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
13495 Handle string / character search functions.
13496 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
13498 2013-03-18 Richard Biener <rguenther@suse.de>
13500 PR middle-end/56483
13501 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
13502 and implement properly.
13503 * gimple.h (gimple_cond_single_var_p): Remove.
13505 2013-03-18 Richard Biener <rguenther@suse.de>
13507 * tree-data-ref.h (find_data_references_in_loop): Declare.
13508 * tree-data-ref.c (get_references_in_stmt): Use a stack
13509 vector pre-allocated in the callers.
13510 (find_data_references_in_stmt): Adjust.
13511 (graphite_find_data_references_in_stmt): Likewise.
13512 (create_rdg_vertices): Likewise.
13513 (find_data_references_in_loop): Export.
13514 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
13515 Compute dependences here...
13516 (vect_analyze_data_refs): ...not here. When we encounter
13517 a non-vectorizable data reference in basic-block vectorization
13518 truncate the data reference vector. Do not bother to
13519 fixup data-dependence information for gather loads.
13520 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
13521 of data references, as reported.
13523 2013-03-18 Richard Biener <rguenther@suse.de>
13525 PR tree-optimization/3713
13526 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
13527 has_constants and expr.
13528 (stmt_has_constants): Properly valueize SSA names when deciding
13529 whether the stmt has constants.
13531 2013-03-18 Richard Biener <rguenther@suse.de>
13533 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
13534 whole function when there is nothing to do.
13535 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
13536 * tree-vectorizer.c (vectorize_loops): Update virtual and
13537 loop-closed SSA once.
13538 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
13540 2013-03-18 Richard Biener <rguenther@suse.de>
13542 PR middle-end/56113
13543 * domwalk.c (bb_postorder): New global static.
13544 (cmp_bb_postorder): New function.
13545 (walk_dominator_tree): Replace scheme imposing an order for
13546 visiting dominator sons by one sorting them at the time they
13547 are pushed on the stack.
13549 2013-03-18 Richard Biener <rguenther@suse.de>
13551 PR tree-optimization/39326
13552 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
13553 (struct mem_ref): Replace mem member with ao_ref typed member.
13554 (MEM_ANALYZABLE): Adjust.
13555 (memref_eq): Likewise.
13556 (mem_ref_alloc): Likewise.
13557 (gather_mem_refs_stmt): Likewise.
13558 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
13559 (execute_sm_if_changed_flag_set): Adjust.
13560 (execute_sm): Likewise.
13561 (ref_always_accessed_p): Likewise.
13562 (refs_independent_p): Likewise.
13563 (can_sm_ref_p): Likewise.
13565 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13568 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
13569 return 1 even for !unsignedp.
13571 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13573 * config/i386/i386.md (isa): Add x64 and nox64.
13574 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
13575 (*pushtf): Enable *roF alternative for x64 isa only.
13576 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
13577 mode attribute of integer alternatives to DImode for TARGET_64BIT.
13578 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
13579 (*movtf_internal): Merge from *movtf_internal_rex64 and
13580 *movtf_internal_sse. Use x64 and nox64 isa attributes.
13581 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
13582 nox64 isa attributes.
13583 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
13584 nox64 isa attributes.
13585 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
13587 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13589 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13591 2013-03-17 Jakub Jelinek <jakub@redhat.com>
13594 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
13595 class names. Remove trailing comma after "ALL_REGS".
13597 2013-03-16 Jan Hubicka <jh@suse.cz>
13599 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
13600 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
13601 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
13602 of cgraph_get_create_node.
13603 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
13605 2013-03-16 Jason Merrill <jason@redhat.com>
13608 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
13609 with DW_AT_default_value.
13611 2013-03-16 Jakub Jelinek <jakub@redhat.com>
13613 * BASE-VER: Set to 4.9.0.
13615 2013-03-14 Andi Kleen <ak@linux.intel.com>
13618 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
13619 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
13620 Document _x* TSX intrinsics.
13622 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
13623 David Holsgrove <david.holsgrove@xilinx.com>
13625 * configure.ac: Add MicroBlaze TLS support detection.
13626 * configure: Regenerate.
13627 * config/microblaze/microblaze-protos.h
13628 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
13629 symbol_mentioned_p, label_mentioned_p): Add prototypes.
13630 * config/microblaze/microblaze.c (microblaze_address_type): Add
13631 ADDRESS_TLS and tls_reloc address types.
13632 (microblaze_address_info): Add tls_reloc.
13633 (TARGET_HAVE_TLS): Define.
13634 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
13635 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
13636 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
13637 load_tls_operand, microblaze_call_tls_get_addr,
13638 microblaze_legitimize_tls_address): New functions.
13639 (microblaze_classify_unspec): Handle UNSPEC_TLS.
13640 (get_base_reg): Use microblaze_tls_symbol_p.
13641 (microblaze_classify_address): Handle TLS.
13642 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
13643 label_mentioned_p and microblaze_tls_referenced_p.
13644 (microblaze_legitimize_address): Handle TLS.
13645 (microblaze_address_insns): Handle ADDRESS_TLS.
13646 (pic_address_needs_scratch): Handle TLS.
13647 (print_operand_address): Handle TLS.
13648 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
13649 (microblaze_expand_move): Handle TLS.
13650 (microblaze_legitimate_constant_p): Check
13651 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
13652 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
13653 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
13654 (PIC_OFFSET_TABLE_REGNUM): Set.
13655 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
13656 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
13657 (addsi3, movsi_internal2, movdf_internal): Update constraints
13658 * config/microblaze/predicates.md (arith_plus_operand): Define
13659 (move_operand): Redefine as move_src_operand,
13660 check microblaze_tls_referenced_p.
13662 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13664 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
13665 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
13667 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13669 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
13672 2013-03-14 Jakub Jelinek <jakub@redhat.com>
13674 PR tree-optimization/53265
13675 * common.opt (Waggressive-loop-optimizations): New option.
13676 * tree-ssa-loop-niter.c: Include tree-pass.h.
13677 (do_warn_aggressive_loop_optimizations): New function.
13678 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
13679 if number_of_latch_executions returned constant.
13680 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
13681 early. If number_of_latch_executions returned constant, set
13682 nb_iterations_upper_bound back to it.
13683 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
13685 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
13686 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
13688 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
13689 (MULTILIB_OSDIRNAMES): Set.
13690 * genmultilib: If defaultosdirname doesn't start with :: , set
13691 defaultosdirname2 instead, clear it and emit two . multilib_raw
13692 entries instead of just one.
13694 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
13696 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13697 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
13698 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13699 (SUBTARGET_OVERRIDE_OPTIONS): New.
13701 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
13704 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
13705 (musermode): Convert to Var(TARGET_USERMODE).
13706 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
13707 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
13708 * config/sh/sh.c (sh_option_override): Use
13709 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
13710 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
13712 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
13714 (udivsi3_i4_single, divsi3_i4_single): Use
13715 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
13717 2013-03-13 Dave Korn <dave.korn.cygwin@....>
13719 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
13722 2013-03-13 Richard Biener <rguenther@suse.de>
13724 PR tree-optimization/56608
13725 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
13726 calls when vectorizing basic-blocks.
13728 2013-03-13 Jakub Jelinek <jakub@redhat.com>
13731 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
13734 2013-03-12 Jakub Jelinek <jakub@redhat.com>
13736 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
13738 2013-03-11 Jan Hubicka <jh@suse.cz>
13741 * lto-streamer-out.c (output_symbol_p): Skip references from
13742 constructors of external variables.
13744 2013-03-11 Jan Hubicka <jh@suse.cz>
13746 PR middle-end/56571
13747 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
13749 * emit-rtl.c (verify_rtx_sharing): Likewise.
13750 (copy_insn_1): Likewise.
13751 * rtl.c (copy_rtx): Likewise.
13753 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
13756 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
13757 output_operand_lossage message.
13759 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
13762 * arm.c (shift_op): Validate RTL pattern on the fly.
13763 (arm_print_operand, case 'S'): Don't use shift_operator to validate
13766 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13769 * config/pa/pa.md (call_value): Check for calls to powf and direct to
13770 new call patterns that clobber %fr12.
13771 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
13772 split and postreload patterns.
13773 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
13774 registers %fr12 and %fr12R as call used.
13776 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
13778 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
13779 (canon_address, record_store, replace_read, check_mem_read_rtx,
13780 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
13781 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
13782 rest_of_handle_dse): Likewise.
13784 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
13786 PR middle-end/56524
13787 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
13789 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
13790 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
13791 (save_optabs_if_changed): Replace with...
13792 (init_tree_optimization_optabs): ...this.
13793 * optabs.c (save_optabs_if_changed): Rename to...
13794 (init_tree_optimization_optabs): ...this. Take the optimization node
13795 as argument. Do nothing if the base optabs are already correct.
13796 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
13797 to recompute optabs.
13798 * function.h (function): Remove optabs field.
13799 * function.c (invoke_set_current_function_hook): Call
13800 init_tree_optimization_optabs. Use the result to initialize
13803 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
13805 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
13806 if GTMA_HAS_NO_INSTRUMENTATION.
13807 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
13808 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
13809 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
13810 * gimple-pretty-print.c (dump_gimple_transaction): Handle
13811 GTMA_HAS_NO_INSTRUMENTATION.
13813 2013-03-08 Jakub Jelinek <jakub@redhat.com>
13815 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
13818 2013-03-08 Marek Polacek <polacek@redhat.com>
13819 Jakub Jelinek <jakub@redhat.com>
13821 PR tree-optimization/56478
13822 * predict.c (is_comparison_with_loop_invariant_p): Change the
13823 type of loop_step to tree.
13824 (predict_loops): Adjust.
13825 (predict_iv_comparison): Perform the computations on double_ints.
13827 2013-03-08 Richard Biener <rguenther@suse.de>
13829 PR tree-optimization/56570
13830 * tree-cfg.c (verify_expr_location_1): Verify locations for
13832 * tree-sra.c (create_access_replacement): Strip locations
13833 from DECL_DEBUG_EXPRs.
13835 2013-03-08 Richard Biener <rguenther@suse.de>
13837 * tree-inline.c (expand_call_inline): Do not associate
13838 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
13839 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
13841 2013-03-08 Richard Biener <rguenther@suse.de>
13843 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
13844 or block changes with -Og. Fix for location / block encoding
13845 changes and PHI arguments with locations.
13847 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
13849 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
13851 (struct output_info): Likewise.
13852 (register_overhead): Remove bad gcc_assert.
13853 (bitmap_find_bit): If there is only a single bitmap element, do not
13854 count a miss as a search.
13855 (print_statistics): Update for counter type changes.
13856 (dump_bitmap_statistics): Likewise. Print headers such that they
13857 are properly lined up with the printed counters.
13859 2013-03-07 Jakub Jelinek <jakub@redhat.com>
13861 PR tree-optimization/56559
13862 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
13863 check that it has only a single use.
13865 2013-03-07 Richard Biener <rguenther@suse.de>
13867 * doc/invoke.texi (fwhole-program): Discourage use in combination
13870 2013-03-06 Jakub Jelinek <jakub@redhat.com>
13872 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
13874 PR tree-optimization/56539
13875 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
13876 instead of GSI_CONTINUE_LINKING as last argument to
13877 force_gimple_operand_gsi. Adjust function comment.
13879 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
13882 PR middle-end/56548
13883 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
13884 promoted mode, convert the result back to the original mode.
13886 2013-03-06 Richard Biener <rguenther@suse.de>
13888 PR middle-end/56294
13889 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
13890 (insert_updated_phi_nodes_compare_uids): New function.
13891 (update_ssa): Sort symbols_to_rename after UID before
13892 traversing it to insert PHI nodes.
13894 2013-03-06 Richard Biener <rguenther@suse.de>
13896 PR middle-end/50494
13897 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
13898 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
13901 2013-02-13 Richard Biener <rguenther@suse.de>
13904 * varasm.c (output_constant_def_1): Get the decl representing
13905 the constant as argument.
13906 (output_constant_def): Wrap output_constant_def_1.
13907 (make_decl_rtl): Use output_constant_def_1 with the decl
13908 representing the constant.
13909 (build_constant_desc): Optionally re-use a decl already
13910 representing the constant.
13911 (tree_output_constant_def): Adjust.
13913 2013-03-06 Joey Ye <joey.ye@arm.com>
13916 * gcc.c (convert_white_space): New function.
13917 (main): Handles white space in function name.
13919 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
13922 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
13923 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
13924 to SH_DIV_CALL_TABLE for TARGET_SH2.
13925 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
13927 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
13928 call-table options.
13930 2013-03-05 Sterling Augustine <saugustine@google.com>
13931 Cary Coutant <ccoutant@google.com>
13934 * dwarf2out.c (resolve_addr): Don't call
13935 remove_loc_list_addr_table_entries a second time for the same
13938 2013-03-05 Jakub Jelinek <jakub@redhat.com>
13941 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
13942 (avoid_complex_debug_insns): New function.
13943 (expand_debug_locations): Call it.
13945 PR rtl-optimization/56484
13946 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
13947 lifetimes of hard registers on small register class machines.
13949 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
13951 * config/microblaze/microblaze-protos.h: Rename
13952 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
13953 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
13955 (microblaze_fast_interrupt_function_p): New function.
13956 (microblaze_is_interrupt_handler): Rename to
13957 microblaze_is_interrupt_variant and add fast_interrupt check.
13958 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
13959 (save_restore_insns): Likewise.
13960 (compute_frame_size): Likewise.
13961 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
13962 (microblaze_globalize_label): Likewise.
13963 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
13964 * config/microblaze/microblaze.md: Use wrapper
13965 microblaze_is_interrupt_variant.
13967 2013-03-05 Kai Tietz <ktietz@redhat.com>
13969 * sdbout.c (sdbout_one_type): Switch to current function's section
13970 supporting cold/hot.
13972 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
13974 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
13977 2013-03-05 Jakub Jelinek <jakub@redhat.com>
13979 PR middle-end/56461
13980 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
13981 if VALGRIND_GET_VBITS is defined, temporarily make object
13982 memory all defined, and restore previous valgrind addressability
13983 and definability afterwards. Free this_object at the end.
13985 PR middle-end/56461
13986 * lra.c (lra): Call lra_clear_live_ranges if live_p,
13987 right before calling lra_create_live_ranges, also call it
13988 when clearing live_p. Only call lra_clear_live_ranges
13989 at the end if live_p.
13991 PR middle-end/56461
13992 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
13994 2013-03-05 Richard Biener <rguenther@suse.de>
13996 PR tree-optimization/56521
13997 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
14000 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
14003 * except.h (remove_unreachable_eh_regions): New prototype.
14004 * except.c (remove_eh_handler_splicer): New function, split out
14005 of remove_eh_handler.
14006 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
14007 warning about running it on many EH regions one at a time.
14008 (remove_unreachable_eh_regions_worker): New function, walk the
14009 EH tree in depth-first order and remove non-marked regions.
14010 (remove_unreachable_eh_regions): New function.
14011 * tree-eh.c (mark_reachable_handlers): New function, split out
14012 from remove_unreachable_handlers.
14013 (remove_unreachable_handlers): Use mark_reachable_handlers and
14014 remove_unreachable_eh_regions.
14015 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
14016 and remove_unreachable_eh_regions.
14018 2013-03-05 Richard Biener <rguenther@suse.de>
14020 PR middle-end/56525
14021 * loop-init.c (fix_loop_structure): Remove loops in two stages,
14022 not freeing them until the end.
14024 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14026 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
14028 2013-03-05 Richard Biener <rguenther@suse.de>
14030 PR tree-optimization/56270
14031 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
14032 of loads after scheduling an SLP instance.
14034 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14036 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
14038 (check_gcc_parallelize): Run guality.exp as a separate job from
14039 vect.exp with unsorted.exp and $(dg_target_exps) separately from
14040 struct-layout-1.exp with stackalign.exp.
14042 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
14044 PR middle-end/56461
14045 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
14046 load_index sbitmap even if some bit in it isn't set.
14048 PR middle-end/56461
14049 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
14050 (discover_iteration_bound_by_body_walk): Change queues to
14051 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
14052 spelling in comment. Call safe_push on queues[bound_index] directly.
14053 Release queues[queue_index] in every iteration unconditionally.
14054 Release bounds vector.
14056 PR middle-end/56461
14057 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
14058 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
14059 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
14062 2013-03-05 Richard Biener <rguenther@suse.de>
14065 * tree-inline.c (remap_blocks_to_null): New function.
14066 (expand_call_inline): When expanding a call stmt without
14067 an associated block inline remap all callee blocks to NULL.
14069 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14071 PR rtl-optimization/56494
14072 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
14073 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
14074 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
14076 PR middle-end/56461
14077 * sel-sched-ir.c (free_sched_pools): Release
14078 succs_info_pool.stack[succs_info_pool.max_top] vectors too
14079 if succs_info_pool.max_top isn't -1.
14082 * opts.c (opts_obstack, opts_concat): Moved to...
14083 * opts-common.c (opts_obstack, opts_concat): ... here.
14085 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14087 PR middle-end/56461
14088 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14090 2013-03-04 Martin Jambor <mjambor@suse.cz>
14092 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14093 all appropriate places.
14095 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14097 PR tree-optimization/56424
14098 * ipa-split.c (split_function): Do not set the RSO flag if result is
14099 not by reference and its type is a register type.
14101 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
14103 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
14104 (microblaze_legitimate_pic_operand): Likewise
14105 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
14106 new function microblaze_legitimate_pic_operand
14107 * config/microblaze/microblaze-protos.h
14108 (microblaze_legitimate_pic_operand): Declare.
14110 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14112 * config/microblaze/predicates.md (call_insn_simple_operand):
14113 New predicate for supported rtx code types.
14114 * config/microblaze/microblaze.md (call_internal1): Use
14115 call_insn_simple_operand predicate.
14117 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14119 PR middle-end/56461
14120 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14121 partitions.ordered_remove.
14123 PR middle-end/56461
14124 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14125 vec_oprnds0.create (1) for modifier == NONE.
14127 PR middle-end/56461
14128 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14129 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14130 vec_oprnds1 right before pushing anything to it for
14133 PR middle-end/56461
14134 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14135 set nbbs to 0 instead of having separate code path.
14136 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14137 instead of false as last argument if returning NULL.
14139 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14141 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14142 the attribute is now called "target" instead of "option".
14143 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14144 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14145 attribute/pragma name for TARGET_OPTION_VALID_P and
14146 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14147 * doc/tm.texi: Regenerated.
14149 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
14151 * config/microblaze/microblaze.c:
14152 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
14153 * config/microblaze/microblaze.h: Add -mxl-reorder to
14155 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14156 instructions emitted if TARGET_REORDER.
14157 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14158 or 0 for -m/-mno case, but initialises as 2 to detect default use case
14161 2013-03-01 Xinliang David Li <davidxl@google.com>
14163 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14166 2013-03-01 Jakub Jelinek <jakub@redhat.com>
14168 PR middle-end/56461
14169 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14170 vector even when returning true. Fix up function comment formatting.
14172 PR middle-end/56461
14173 * ira-build.c (ira_loop_nodes_count): New variable.
14174 (create_loop_tree_nodes): Initialize it.
14175 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14177 PR middle-end/56461
14178 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14179 method on dr_chain and result_chain.
14180 * tree-vect-stmts.c (vectorizable_store): Only call
14181 result_chain.create if j == 0.
14183 PR middle-end/56461
14184 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14185 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14186 before overwriting it.
14188 2013-03-01 Tobias Burnus <burnus@net-b.de>
14190 * doc/extended.texi (C Extensions): Change order in @menu
14192 (Other MIPS Built-in Functions): Move last MIPS entry before
14193 "picoChip Built-in Functions".
14194 (SH Built-in Functions): Move after RX Built-in Functions.
14195 * doc/gcc.texi (Introduction): Change order in @menu
14197 * doc/md.texi (Constraints): Ditto.
14198 * gty.texi (Type Information): Ditto.
14199 (User-provided marking routines for template types): Make
14201 * doc/invoke.texi (AArch64 Options): Move before
14202 "Adapteva Epiphany Options".
14204 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14205 Jakub Jelinek <jakub@redhat.com>
14208 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14209 no_address_safety_analysis attribute.
14210 * doc/extend.texi (no_address_safety_attribute): Rename to
14211 no_sanitize_address attribute, mention no_address_safety_analysis
14212 attribute as deprecated alias.
14214 2013-02-28 Jakub Jelinek <jakub@redhat.com>
14216 PR middle-end/56461
14217 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14218 type to vec<vec<tree> > *.
14219 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14220 to be vec<tree> instead of vec<tree> *, set vec_defs
14221 to vNULL and call vec_defs.create (number_of_vects), adjust other
14223 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14224 vectorizable_condition): Adjust vect_get_slp_defs callers.
14226 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14228 * config/aarch64/aarch64.c
14229 (aarch64_float_const_representable): Remove unused variable.
14231 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14233 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14235 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14237 * config/aarch64/aarch64-builtins.c
14238 (aarch64_init_simd_builtins): Make static.
14240 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14242 * config/aarch64/aarch64.c
14243 (aarch64_simd_make_constant): Make static.
14245 2013-02-28 Martin Jambor <mjambor@suse.cz>
14247 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14248 with no initialization to the RHS of debug statements.
14250 2013-02-28 Martin Jambor <mjambor@suse.cz>
14252 PR tree-optimization/56294
14253 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14255 (get_access_replacement): Do not call create_access_replacement.
14256 Assert a replacement exists.
14257 (get_repl_default_def_ssa_name): Create the replacement declaration
14260 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14262 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14263 final_end_function.
14265 2013-02-28 Marek Polacek <polacek@redhat.com>
14267 PR rtl-optimization/56466
14268 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14269 if we're changing a loop.
14270 (peel_loops_completely): Likewise.
14272 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14275 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14277 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14280 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14281 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14282 INTX_FTYPE_FX, FX_FTYPE_INTX.
14283 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14285 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14287 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14288 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14289 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14290 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14291 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14292 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14293 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14294 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14295 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14296 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14297 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14298 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14299 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14300 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14301 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14302 (avrxmega6): Increase max flash segments from 5 to 6.
14303 * config/avr/t-multilib: Regenerate.
14304 * config/avr/avr-tables.opt: Regenerate.
14305 * doc/avr-mmcu.texi: Regenerate.
14307 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14309 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14310 (avr_device_to_arch): Rename to avr_device_to_ld.
14311 (avr_device_to_as): New prototype.
14312 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14313 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14314 * config/avr/driver-avr.c (avr_device_to_as): New.
14315 (avr_device_to_arch): Rename to avr_device_to_ld.
14317 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14319 PR middle-end/56461
14320 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14321 method on dr_chain and result_chain.
14323 PR middle-end/56461
14324 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14325 pointer_set_destroy on not_executed_last_iteration.
14327 PR middle-end/56461
14328 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
14330 PR middle-end/56461
14331 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14332 FOR_EACH_DEFINED_FUNCTION when freeing state.
14334 PR middle-end/56461
14335 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14337 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14340 PR middle-end/56461
14341 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14342 known_aggs[i].items and release known_aggs vector.
14344 PR middle-end/56461
14345 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14347 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14349 * config/microblaze/microblaze.c (microblaze_emit_compare):
14350 Use xor for EQ/NE comparisions.
14351 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14352 (cbranchsf4): Adjust operator to comparison_operator.
14354 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14356 PR middle-end/56461
14357 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14359 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14360 vec_safe_push, always update *slot.
14361 (redirect_edge_var_map_clear): Use vec_free.
14362 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14363 (free_var_map_entry): Use vec_free.
14364 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14365 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14367 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
14369 PR middle-end/45472
14370 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14371 when the may_trap_p bit of the exprs being merged differs.
14372 Reorder tests for speculativeness in the logical and operator.
14374 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14376 * incpath.c (add_standard_paths): Use reconcat instead of concat
14377 where appropriate and avoid leaking memory.
14379 * opts.h: Include obstack.h.
14380 (opts_concat): New prototype.
14381 (opts_obstack): New declaration.
14382 * opts.c (opts_concat): New function.
14383 (opts_obstack): New variable.
14384 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14385 (finish_options): Use opts_concat instead of concat
14386 and XOBNEWVEC instead of XNEWVEC.
14387 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14388 generate_option): Likewise.
14389 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14390 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14393 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14394 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14396 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14398 PR middle-end/56461
14399 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14401 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14403 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14404 (arm_block_move_unaligned_straight): Likewise.
14405 (arm_adjust_block_mem): Likewise.
14407 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14410 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14411 temp, cond and label.
14412 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14415 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14416 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14419 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14420 of prologue/epilogue functions.
14423 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14424 Remove unused variable cfa_offset.
14425 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14428 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14429 type promotion to unsigned.
14432 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
14433 for HOST_WIDE_INT of 32 bit / same size as int.
14434 (arm_block_move_unaligned_straight): Likewise.
14435 (arm_adjust_block_mem): Likewise.
14438 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
14441 2013-02-26 Marek Polacek <polacek@redhat.com>
14443 PR tree-optimization/56426
14444 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
14446 2013-02-26 Richard Biener <rguenther@suse.de>
14449 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
14450 unused variable loops.
14452 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14454 PR tree-optimization/56448
14455 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
14456 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
14457 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
14458 later operands of the references, or even first operand for
14459 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
14461 PR tree-optimization/56443
14462 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
14463 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
14464 to type_for_mode langhook.
14466 2013-02-25 Matt Turner <mattst88@gmail.com>
14468 * doc/invoke.texi: Document r4700.
14470 2013-02-25 Richard Biener <rguenther@suse.de>
14472 PR tree-optimization/56175
14473 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
14475 (simplify_bitwise_binary): ... here. Also guard the conversion
14476 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
14478 2013-02-25 Catherine Moore <clm@codesourcery.com>
14481 2013-02-24 Catherine Moore <clm@codesourcery.com>
14482 Maciej W. Rozycki <macro@codesourcery.com>
14483 Tom de Vries <tom@codesourcery.com>
14484 Nathan Sidwell <nathan@codesourcery.com>
14485 Iain Sandoe <iain@codesourcery.com>
14486 Nathan Froyd <froydnj@codesourcery.com>
14487 Chao-ying Fu <fu@mips.com>
14489 * doc/extend.texi: (micromips, nomicromips, nocompression):
14490 Document new function attributes.
14491 * doc/invoke.texi (minterlink-compressed, mmicromips,
14492 m14k, m14ke, m14kec): Document new options.
14493 (minterlink-mips16): Update documentation.
14494 * doc/md.texi (ZC, ZD): Document new constraints.
14495 * configure.ac (gcc_cv_as_micromips): Check if linker
14496 supports the .set micromips directive.
14497 * configure: Regenerate.
14498 * config.in: Regenerate.
14499 * config/mips/mips-tables.opt: Regenerate.
14500 * config/mips/micromips.md: New file.
14501 * constraints.md (ZC, AD): New constraints.
14502 * config/mips/predicates.md (movep_src_register): New predicate.
14503 (movep_src_operand): New predicate.
14504 (non_volatile_mem_operand): New predicate.
14505 * config/mips/mips.md (multimem): New type.
14506 (length): Differentiate between 17-bit and 18-bit branch offsets.
14507 (MOVEP1, MOVEP2): New mode iterator.
14508 (mov_<load>l): Use ZC constraint.
14509 (mov_<load>r): Likewise.
14510 (mov_<store>l): Likewise.
14511 (mov_<store>r): Likewise.
14512 (*branch_equality<mode>_inverted): Add microMIPS support.
14513 (*branch_equality<mode>): Likewise.
14514 (*jump_absolute): Likewise.
14515 (indirect_jump_<mode>): Likewise.
14516 (tablejump_<mode>): Likewise.
14517 (<optab>_internal): Likewise.
14518 (sibcall_internal): Likewise.
14519 (sibcall_value_internal): Likewise.
14520 (prefetch): Use constraint ZD.
14521 * config/mips/mips.opt (minterlink-compressed): New option.
14522 (minterlink-mips16): Now an alias for minterlink-compressed.
14523 (mmicromips): New option.
14524 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14525 (compare_and_swap_12): Likewise.
14526 (sync_add<mode>): Likewise.
14527 (sync_<optab>_12): Likewise.
14528 (sync_old_<optab>_12): Likewise.
14529 (sync_new_<optab>_12): Likewise.
14530 (sync_nand_12): Likewise.
14531 (sync_old_nand_12): Likewise.
14532 (sync_new_nand_12): Likewise.
14533 (sync_sub<mode>): Likewise.
14534 (sync_old_add<mode>): Likewise.
14535 (sync_old_sub<mode>): Likewise.
14536 (sync_new_add<mode>): Likewise.
14537 (sync_new_sub<mode>): Likewise.
14538 (sync_<optab><mode>): Likewise.
14539 (sync_old_<optab><mode>): Likewise.
14540 (sync_new_<optab><mode>): Likewise.
14541 (sync_nand<mode>): Likewise.
14542 (sync_old_nand<mode>): Likewise.
14543 (sync_new_nand<mode>): Likewise.
14544 (sync_lock_test_and_set<mode>): Likewise.
14545 (test_and_set_12): Likewise.
14546 (atomic_compare_and_swap<mode>): Likewise.
14547 (atomic_exchange<mode>_llsc): Likewise.
14548 (atomic_fetch_add<mode>_llsc): Likewise.
14549 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14550 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14551 (umips_save_restore_pattern_p): Likewise.
14552 (umips_load_store_pair_p): Likewise.
14553 (umips_output_load_store_pair): Likewise.
14554 (umips_movep_target_p): Likewise.
14555 (umips_12bit_offset_address_p): Likewise.
14556 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14557 (mips_base_mips16): Rename this...
14558 (mips_base_compression_flags): ...to this. Update all uses.
14559 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14560 (mips_mips16_decl_p): Delete.
14561 (mips_nomips16_decl_p): Delete.
14562 (mips_get_compress_on_flags): New function.
14563 (mips_get_compress_off_flags): New function.
14564 (mips_get_compress_mode): New function.
14565 (mips_get_compress_on_name): New function.
14566 (mips_get_compress_off_name): New function.
14567 (mips_insert_attributes): Support multiple compression types.
14568 (mips_merge_decl_attributes): Likewise.
14569 (umips_12bit_offset_address_p): New function.
14570 (mips_start_function_definition): Emit .set micromips directive.
14571 (mips_call_may_need_jalx_p): New function.
14572 (mips_function_ok_for_sibcall): Add microMIPS support.
14573 (mips_print_operand_punctuation): Support short delay slots and
14575 (umips_swm_mask, umips_swm_encoding): New.
14576 (umips_build_save_restore): New function.
14577 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14578 (was_mips16_p): Remove.
14579 (old_compression_mode): New.
14580 (mips_set_compression_mode): New function.
14581 (mips_set_current_function): Add microMIPS support.
14582 (mips_option_override): Likewise.
14583 (umips_save_restore_pattern_p): New function.
14584 (umips_output_save_restore): New function.
14585 (umips_load_store_pair_p_1): New function.
14586 (umips_load_store_pair_p): New function.
14587 (umips_output_load_store_pair_1): New function.
14588 (umips_output_load_store_pair): New function.
14589 (umips_movep_target_p) New function.
14590 (mips_prepare_pch_save): Add microMIPS support.
14591 * config/mips/mips.h (TARGET_COMPRESSION): New.
14592 (TARGET_CPU_CPP_BUILTINS): Update macro
14593 to use new compression flags and to support microMIPS.
14594 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14595 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14596 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14597 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14598 (ASM_SPEC): Support mmicromips and mno-micromips.
14599 (M16STORE_REG_P): New macro.
14600 (MIPS_CALL): Support TARGET_MICROMIPS.
14601 (MICROMIPS_J): New macro.
14602 (mips_base_mips16): Rename this...
14603 (mips_base_compression_flags): ...to this.
14604 (UMIPS_12BIT_OFFSET_P): New macro.
14605 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14606 (MULTILIB_DIRNAMES): Likewise.
14608 2013-02-25 Tom de Vries <tom@codesourcery.com>
14610 PR rtl-optimization/56131
14611 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
14612 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
14613 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
14615 2013-02-25 Tobias Burnus <burnus@net-b.de>
14617 * doc/invoke.texi (-fsanitize=): Move from optimization
14618 to debugging options.
14620 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14622 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
14624 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14625 Alexander Monakov <amonakov@ispras.ru>
14627 PR middle-end/56077
14628 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
14629 flush pending lists also on non-jumps. Adjust comment.
14631 2013-02-24 Catherine Moore <clm@codesourcery.com>
14632 Maciej W. Rozycki <macro@codesourcery.com>
14633 Tom de Vries <tom@codesourcery.com>
14634 Nathan Sidwell <nathan@codesourcery.com>
14635 Iain Sandoe <iain@codesourcery.com>
14636 Nathan Froyd <froydnj@codesourcery.com>
14637 Chao-ying Fu <fu@mips.com>
14639 * doc/extend.texi: (micromips, nomicromips, nocompression):
14640 Document new function attributes.
14641 * doc/invoke.texi (minterlink-compressed, mmicromips,
14642 m14k, m14ke, m14kec): Document new options.
14643 (minterlink-mips16): Update documentation.
14644 * doc/md.texi (ZC, ZD): Document new constraints.
14645 * configure.ac (gcc_cv_as_micromips): Check if linker
14646 supports the .set micromips directive.
14647 * configure: Regenerate.
14648 * config.in: Regenerate.
14649 * config/mips/mips-tables.opt: Regenerate.
14650 * config/mips/micromips.md: New file.
14651 * constraints.md (ZC, AD): New constraints.
14652 * config/mips/predicates.md (movep_src_register): New predicate.
14653 (movep_src_operand): New predicate.
14654 (non_volatile_mem_operand): New predicate.
14655 * config/mips/mips.md (multimem): New type.
14656 (length): Differentiate between 17-bit and 18-bit branch offsets.
14657 (MOVEP1, MOVEP2): New mode iterator.
14658 (mov_<load>l): Use ZC constraint.
14659 (mov_<load>r): Likewise.
14660 (mov_<store>l): Likewise.
14661 (mov_<store>r): Likewise.
14662 (*branch_equality<mode>_inverted): Add microMIPS support.
14663 (*branch_equality<mode>): Likewise.
14664 (*jump_absolute): Likewise.
14665 (indirect_jump_<mode>): Likewise.
14666 (tablejump_<mode>): Likewise.
14667 (<optab>_internal): Likewise.
14668 (sibcall_internal): Likewise.
14669 (sibcall_value_internal): Likewise.
14670 (prefetch): Use constraint ZD.
14671 * config/mips/mips.opt (minterlink-compressed): New option.
14672 (minterlink-mips16): Now an alias for minterlink-compressed.
14673 (mmicromips): New option.
14674 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14675 (compare_and_swap_12): Likewise.
14676 (sync_add<mode>): Likewise.
14677 (sync_<optab>_12): Likewise.
14678 (sync_old_<optab>_12): Likewise.
14679 (sync_new_<optab>_12): Likewise.
14680 (sync_nand_12): Likewise.
14681 (sync_old_nand_12): Likewise.
14682 (sync_new_nand_12): Likewise.
14683 (sync_sub<mode>): Likewise.
14684 (sync_old_add<mode>): Likewise.
14685 (sync_old_sub<mode>): Likewise.
14686 (sync_new_add<mode>): Likewise.
14687 (sync_new_sub<mode>): Likewise.
14688 (sync_<optab><mode>): Likewise.
14689 (sync_old_<optab><mode>): Likewise.
14690 (sync_new_<optab><mode>): Likewise.
14691 (sync_nand<mode>): Likewise.
14692 (sync_old_nand<mode>): Likewise.
14693 (sync_new_nand<mode>): Likewise.
14694 (sync_lock_test_and_set<mode>): Likewise.
14695 (test_and_set_12): Likewise.
14696 (atomic_compare_and_swap<mode>): Likewise.
14697 (atomic_exchange<mode>_llsc): Likewise.
14698 (atomic_fetch_add<mode>_llsc): Likewise.
14699 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14700 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14701 (umips_save_restore_pattern_p): Likewise.
14702 (umips_load_store_pair_p): Likewise.
14703 (umips_output_load_store_pair): Likewise.
14704 (umips_movep_target_p): Likewise.
14705 (umips_12bit_offset_address_p): Likewise.
14706 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14707 (mips_base_mips16): Rename this...
14708 (mips_base_compression_flags): ...to this. Update all uses.
14709 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14710 (mips_mips16_decl_p): Delete.
14711 (mips_nomips16_decl_p): Delete.
14712 (mips_get_compress_on_flags): New function.
14713 (mips_get_compress_off_flags): New function.
14714 (mips_get_compress_mode): New function.
14715 (mips_get_compress_on_name): New function.
14716 (mips_get_compress_off_name): New function.
14717 (mips_insert_attributes): Support multiple compression types.
14718 (mips_merge_decl_attributes): Likewise.
14719 (umips_12bit_offset_address_p): New function.
14720 (mips_start_function_definition): Emit .set micromips directive.
14721 (mips_call_may_need_jalx_p): New function.
14722 (mips_function_ok_for_sibcall): Add microMIPS support.
14723 (mips_print_operand_punctuation): Support short delay slots and
14725 (umips_swm_mask, umips_swm_encoding): New.
14726 (umips_build_save_restore): New function.
14727 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14728 (was_mips16_p): Remove.
14729 (old_compression_mode): New.
14730 (mips_set_compression_mode): New function.
14731 (mips_set_current_function): Add microMIPS support.
14732 (mips_option_override): Likewise.
14733 (umips_save_restore_pattern_p): New function.
14734 (umips_output_save_restore): New function.
14735 (umips_load_store_pair_p_1): New function.
14736 (umips_load_store_pair_p): New function.
14737 (umips_output_load_store_pair_1): New function.
14738 (umips_output_load_store_pair): New function.
14739 (umips_movep_target_p) New function.
14740 (mips_prepare_pch_save): Add microMIPS support.
14741 * config/mips/mips.h (TARGET_COMPRESSION): New.
14742 (TARGET_CPU_CPP_BUILTINS): Update macro
14743 to use new compression flags and to support microMIPS.
14744 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14745 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14746 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14747 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14748 (ASM_SPEC): Support mmicromips and mno-micromips.
14749 (M16STORE_REG_P): New macro.
14750 (MIPS_CALL): Support TARGET_MICROMIPS.
14751 (MICROMIPS_J): New macro.
14752 (mips_base_mips16): Rename this...
14753 (mips_base_compression_flags): ...to this.
14754 (UMIPS_12BIT_OFFSET_P): New macro.
14755 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14756 (MULTILIB_DIRNAMES): Likewise.
14758 2013-02-24 Jakub Jelinek <jakub@redhat.com>
14761 * target-globals.c (save_target_globals): For init_reg_sets and
14762 target_reinit remporarily set this_fn_optabs to this_target_optabs.
14764 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
14766 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
14767 * config/aarch64/t-aarch64
14768 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
14770 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
14772 PR inline-asm/56148
14773 * lra-constraints.c (process_alt_operands): Reload operand
14774 conflicting with earlier clobber only if no more other conflicting
14777 2013-02-22 Jakub Jelinek <jakub@redhat.com>
14780 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
14781 if not linking a shared library.
14783 2013-02-22 Seth LaForge <sethml@google.com>
14785 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
14787 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
14789 * config/arm/arm.md (split for extendsidi): Update condition.
14790 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
14791 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
14792 (qhs_zextenddi_cstr): Likewise.
14794 2013-02-21 Jakub Jelinek <jakub@redhat.com>
14796 PR middle-end/56420
14797 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
14798 avoid signed wrapping.
14799 (expand_mult): Handle properly multiplication by
14800 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
14801 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
14802 in the compiler if coeff is HOST_WIDE_INT_MIN.
14803 (expand_divmod): Don't make ext_op1 static, change it's type to
14804 uhwi. Avoid undefined behavior in -INTVAL (op1).
14806 PR rtl-optimization/50339
14807 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
14809 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
14810 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
14811 into splitting_ashiftrt field.
14812 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
14814 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
14817 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
14819 PR middle-end/56108
14820 * trans-mem.c (execute_tm_mark): Do not expand transactions that
14821 are sure to go irrevocable.
14823 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
14825 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
14826 scalars are valid operands.
14828 2013-02-21 Martin Jambor <mjambor@suse.cz>
14830 PR tree-optimization/56310
14831 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
14832 only matching indices and non-negative final offsets.
14833 (intersect_aggregates_with_edge): Pass src_idx to
14834 agg_replacements_to_vector. Pass src_idx insstead of index to
14835 intersect_with_agg_replacements.
14837 2013-02-21 Martin Jambor <mjambor@suse.cz>
14839 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
14840 instead of hard-wired defaults.
14842 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
14844 * doc/invoke.texi (MIPS Options): Update documentation of the
14845 floating-point multiply-accumulate instruction restrictions.
14847 2013-02-21 Kostya Serebryany <kcc@google.com>
14849 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
14850 asan_shadow_offset on x86_64 linux.
14852 2013-02-21 Richard Biener <rguenther@suse.de>
14854 PR tree-optimization/56415
14856 2013-02-11 Richard Biener <rguenther@suse.de>
14858 PR tree-optimization/56273
14859 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
14862 2013-02-21 Jakub Jelinek <jakub@redhat.com>
14865 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
14868 PR inline-asm/56405
14869 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
14870 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
14872 2013-02-20 Jan Hubicka <jh@suse.cz>
14874 PR tree-optimization/56265
14875 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
14876 when target is referenced for first time.
14878 2013-02-20 Richard Biener <rguenther@suse.de>
14880 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
14881 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
14882 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
14883 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
14884 not return anything.
14885 (rename_ssa_copies): Do not remove unused locals.
14886 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
14887 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
14888 * passes.c (execute_function_todo): Do not schedule unused locals
14889 removal if cleanup_tree_cfg did something.
14890 * tree-ssa-live.c (remove_unused_locals): Dump statistics
14891 about the number of removed locals.
14893 2013-02-20 Richard Biener <rguenther@suse.de>
14895 PR tree-optimization/56398
14896 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
14898 2013-02-20 Martin Jambor <mjambor@suse.cz>
14900 PR tree-optimization/55334
14901 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
14902 restricted pointers to arrays.
14904 2013-02-20 Richard Biener <rguenther@suse.de>
14905 Jakub Jelinek <jakub@redhat.com>
14907 PR tree-optimization/56396
14908 * tree-ssa-ccp.c (n_const_val): New static variable.
14909 (get_value): Return NULL for SSA names we don't have a lattice
14911 (ccp_initialize): Initialize n_const_val.
14912 * tree-ssa-copy.c (n_copy_of): New static variable.
14913 (init_copy_prop): Initialize n_copy_of.
14914 (get_value): Return NULL_TREE for SSA names we don't have a
14917 2013-02-20 Martin Jambor <mjambor@suse.cz>
14919 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
14921 2013-02-20 Richard Biener <rguenther@suse.de>
14923 * genpreds.c (write_lookup_constraint): Do not compare first
14924 letter of the constraint again.
14926 2013-02-20 Richard Biener <rguenther@suse.de>
14928 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
14930 (get_use_iv_cost): Terminate hashtable walk when coming across
14933 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
14935 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
14936 reassociation for avx2 targets.
14938 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14940 * config/microblaze/microblaze.c: microblaze_has_clz = 0
14941 Add version check for v8.10.a to enable microblaze_has_clz
14942 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
14943 version and TARGET_PATTERN_COMPARE check
14944 * config/microblaze/microblaze.md: New clzsi2 instruction
14946 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14948 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
14949 function before branching.
14951 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
14953 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
14955 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
14957 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
14959 PR middle-end/55889
14960 * sel-sched.c: Include ira.h.
14961 (implicit_clobber_conflict_p): New function.
14962 (moveup_expr): Use it.
14963 * Makefile.in (sel-sched.o): Depend on ira.h.
14965 2013-02-19 Richard Biener <rguenther@suse.de>
14967 PR tree-optimization/56384
14968 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
14969 (vn_hash_type): Split out from ...
14970 (vn_hash_constant_with_type): ... here.
14971 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
14972 (vn_phi_eq): Compare types from vn_phi_s structure.
14973 (vn_phi_lookup): Populate vn_phi_s type.
14974 (vn_phi_insert): Likewise.
14976 2013-02-19 Jakub Jelinek <jakub@redhat.com>
14978 PR tree-optimization/56350
14979 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
14980 if haven't found reduction or nested cycle operand, rather than
14981 asserting we must find it.
14983 PR tree-optimization/56381
14984 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
14987 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
14988 Jakub Jelinek <jakub@redhat.com>
14991 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
14992 (swap_optab_enable): Same.
14993 (init_all_optabs): Use argument instead of global.
14994 * tree.h (struct tree_optimization_option): New field target_optabs.
14995 * expr.h (init_all_optabs): Add argument to prototype.
14996 (TREE_OPTIMIZATION_OPTABS): New.
14997 (save_optabs_if_changed): Protoize.
14998 * optabs.h: Declare this_fn_optabs.
14999 * optabs.c (save_optabs_if_changed): New.
15000 Declare this_fn_optabs.
15001 (init_optabs): Add argument to init_all_optabs() call.
15002 * function.c (invoke_set_current_function_hook): Handle per
15004 * function.h (struct function): New field optabs.
15005 * config/mips/mips.c (mips_set_mips16_mode): Handle when
15006 optimization_current_node has changed.
15007 * target-globals.h (save_target_globals_default_opts): Protoize.
15008 * target-globals.c (save_target_globals_default_opts): New.
15010 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15013 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
15014 registers %fr12 and %fr12R as call used.
15017 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
15018 and HImode, require all displacements to be an integer multiple of
15020 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
15021 only allow QImode and HImode when reload is in progress and strict is
15022 true. Likewise for symbolic addresses. Use base14_operand to check
15023 displacements in REG+BASE addresses.
15025 2013-02-18 Richard Biener <rguenther@suse.de>
15027 PR tree-optimization/56366
15028 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15029 handle sign-conversion of outer-loop initial induction value.
15031 2013-02-18 Richard Biener <rguenther@suse.de>
15033 PR middle-end/56349
15034 * cfghooks.c (merge_blocks): If we merge a latch into another
15035 block adjust references to it.
15036 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
15037 (verify_loop_structure): Verify that a recorded latch is in fact
15040 2013-02-18 Richard Biener <rguenther@suse.de>
15042 PR tree-optimization/56321
15043 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
15044 order SSA name release and virtual operand unlinking.
15046 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15048 * config/microblaze/microblaze.md (save_stack_block): Define.
15049 (restore_stack_block): Likewise.
15051 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15053 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
15054 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
15055 * config/microblaze/microblaze.c (microblaze_option_override):
15056 Bail out early for PIC modes when target does not support PIC.
15058 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15060 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
15061 Replace with a microblaze version.
15062 (microblaze_trampoline_init): Adapt for microblaze.
15063 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
15066 2013-02-16 Jakub Jelinek <jakub@redhat.com>
15067 Dodji Seketeli <dodji@redhat.com>
15070 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
15071 (instrument_mem_region_access): Do not forget to always put
15072 instrumentation of the of 'base' and 'base + len' in a "if (len !=
15073 0) statement, even for cases where either 'base' or 'base + len'
15074 are not instrumented -- because they have been previously
15075 instrumented. Simplify the logic by putting all the statements
15076 instrument 'base + len' inside a sequence, and then insert that
15077 sequence right before the current insertion point. Then, to
15078 instrument 'base + len', just get an iterator on that statement.
15079 And do not forget to update the pointer to iterator the function
15080 received as argument.
15082 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
15084 PR rtl-optimization/56348
15085 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
15087 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15089 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15090 (clean_graph_dump_file): Pass base to start_graph_dump.
15092 2013-02-14 Richard Henderson <rth@redhat.com>
15095 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15097 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15099 * collect2-aix.h: Define F_LOADONLY.
15101 2013-02-14 Richard Biener <rguenther@suse.de>
15104 * varasm.c (output_constant_def_1): Get the decl representing
15105 the constant as argument.
15106 (output_constant_def): Wrap output_constant_def_1.
15107 (make_decl_rtl): Use output_constant_def_1 with the decl
15108 representing the constant.
15109 (build_constant_desc): Optionally re-use a decl already
15110 representing the constant.
15111 (tree_output_constant_def): Adjust.
15113 2013-02-14 Dodji Seketeli <dodji@redhat.com>
15116 * asan.c (instrument_builtin_call): Really put the length of the
15117 second source argument into src1_len.
15119 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15121 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15122 argument. If it is false, don't create edge from then_bb to
15124 (insert_if_then_before_iter): Pass true to it.
15125 (build_check_stmt): Pass false to it.
15126 (transform_statements): Flush hash table only on extended basic
15127 block boundaries, rather than at the beginning of every bb.
15128 Don't flush hash table on nonfreeing_call_p calls.
15129 * tree-flow.h (nonfreeing_call_p): New prototype.
15130 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15132 2013-02-13 David S. Miller <davem@davemloft.net>
15134 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15136 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15139 * ira.c (max_regno_before_ira): Move from ...
15141 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15144 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15146 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15148 2013-02-13 Richard Biener <rguenther@suse.de>
15151 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15152 globals in MEM_REFs.
15154 2013-02-13 Richard Biener <rguenther@suse.de>
15156 * loop-init.c (loop_optimizer_init): Clear loop state when
15157 re-initializing preserved loops.
15158 * loop-unswitch.c (unswitch_single_loop): Return whether
15159 we unswitched the loop. Do not verify loop state here.
15160 (unswitch_loops): When we unswitched a loop discover new loops.
15162 2013-02-13 Kostya Serebryany <kcc@google.com>
15164 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15166 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15168 2013-02-12 Dodji Seketeli <dodji@redhat.com>
15170 Avoid instrumenting duplicated memory access in the same basic block
15171 * Makefile.in (asan.o): Add new dependency on hash-table.h
15172 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15173 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15174 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15175 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15176 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15177 (get_mem_ref_of_assignment): New functions.
15178 (get_mem_refs_of_builtin_call): Extract from
15179 instrument_builtin_call and tweak a little bit to make it fit with
15181 (instrument_builtin_call): Use the new
15182 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15183 of is_gimple_builtin_call.
15184 (instrument_derefs, instrument_mem_region_access): Insert the
15185 instrumented memory reference into the hash table.
15186 (maybe_instrument_assignment): Renamed instrument_assignment into
15187 this, and change it to advance the iterator when instrumentation
15188 actually happened and return true in that case. This makes it
15189 homogeneous with maybe_instrument_assignment, and thus give a
15190 chance to callers to be more 'regular'.
15191 (transform_statements): Clear the memory reference hash table
15192 whenever we enter a new BB, when we cross a function call, or when
15193 we are done transforming statements. Use
15194 maybe_instrument_assignment instead of instrumentation. No more
15195 need to special case maybe_instrument_assignment and advance the
15196 iterator after calling it; it's now handled just like
15197 maybe_instrument_call. Update comment.
15199 2013-02-13 Richard Biener <rguenther@suse.de>
15201 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15202 Fix loop discovery code.
15204 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15206 PR inline-asm/56148
15207 * lra-constraints.c (process_alt_operands): Match early clobber
15208 operand with itself. Check conflicts with earlyclobber only if
15209 the operand is not reloaded. Prefer to reload conflicting operand
15210 if earlyclobber and matching operands are the same.
15212 2013-02-12 Richard Biener <rguenther@suse.de>
15215 * lto-streamer-out.c (write_symbol): Do not output symbols
15216 for hard register variables.
15218 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
15221 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15222 (umulsidi3_insn, mulsidi3_insn): New insns.
15224 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
15226 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15227 (struct tune_params): Add vec_costs field.
15228 * config/arm/arm.c (arm_builtin_vectorization_cost)
15229 (arm_add_stmt_cost): New functions.
15230 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15231 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15232 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15233 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15234 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15235 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15236 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15238 2013-02-12 Richard Biener <rguenther@suse.de>
15241 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15242 decls again if possible.
15244 2013-02-12 Richard Biener <rguenther@suse.de>
15246 PR middle-end/56288
15247 * tree-ssa.c (verify_ssa_name): Fix check, move
15248 SSA_NAME_IN_FREE_LIST check up.
15250 2013-02-12 Jakub Jelinek <jakub@redhat.com>
15251 Steven Bosscher <steven@gcc.gnu.org>
15253 PR rtl-optimization/56151
15254 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15255 equal to op0 or op1, and last_insn pattern is CODE operation
15256 with MEM dest and one of the operands matches that MEM.
15258 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15260 * doc/extend.texi: Document Function Multiversioning and "default"
15261 parameter string to target attribute.
15262 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15263 target attribute parameter is "default".
15264 (ix86_compare_version_priority): Remove checks for target attribute.
15265 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15266 Remove check for target attribute equal to NULL. Add assert.
15267 (ix86_generate_version_dispatcher_body): Change error to sorry.
15269 2013-02-11 Iain Sandoe <iain@codesourcery.com>
15270 Jack Howarth <howarth@bromo.med.uc.edu>
15271 Patrick Marlier <patrick.marlier@gmail.com>
15274 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15275 define ENDFILE_SPEC as TM_DESTRUCTOR.
15276 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15278 2013-02-11 Alexander Potapenko <glider@google.com>
15279 Jack Howarth <howarth@bromo.med.uc.edu>
15280 Jakub Jelinek <jakub@redhat.com>
15283 * config/darwin.c (cdtor_record): Rename ctor_record.
15284 (sort_cdtor_records): Rename sort_ctor_records.
15285 (finalize_dtors): New routine to sort destructors by
15286 priority before use in assemble_integer.
15287 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15289 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
15291 PR rtl-optimization/56275
15292 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15293 offset is non-negative and less than cmode size before
15294 calling simplify_subreg.
15296 2013-02-11 Richard Biener <rguenther@suse.de>
15298 PR tree-optimization/56264
15299 * cfgloop.h (fix_loop_structure): Adjust prototype.
15300 * loop-init.c (fix_loop_structure): Return the number of
15301 newly discovered loops.
15302 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15303 are discovered, do a full loop-closed SSA rewrite.
15305 2013-02-11 Richard Biener <rguenther@suse.de>
15307 PR tree-optimization/56273
15308 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15310 (check_array_ref): Fix missing newline in dumps.
15311 (search_for_addr_array): Likewise.
15313 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
15315 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15317 2013-02-09 Jakub Jelinek <jakub@redhat.com>
15320 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15322 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15324 PR rtl-optimization/56246
15325 * lra-constraints.c (simplify_operand_subreg): Try to reuse
15327 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15328 constraints are satisfied.
15330 2013-02-08 Jeff Law <law@redhat.com>
15333 * emit-rtl.c (reg_is_parm_p): New function.
15334 * regs.h (reg_is_parm_p): New prototype.
15335 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15336 callee-clobbered registers.
15338 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15341 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15342 If there is no implicit builtin declaration, just return NULL.
15344 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15346 * config/i386/sse.md (FMAMODEM): New mode iterator.
15347 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15348 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15350 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15352 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15353 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15354 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15356 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15358 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15359 (microblaze*-*-elf): Likewise.
15360 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15362 * config/microblaze/microblaze-c.c: Add builtin defines for
15363 _LITTLE_ENDIAN and _BIG_ENDIAN.
15364 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15365 add to TARGET_DEFAULT flags.
15366 Expand ASM_SPEC and LINK_SPEC.
15367 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15368 * config/microblaze/microblaze.md: Update extendsidi2 and
15369 movdi_internal instructions to use low-order / high-order reg
15371 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15372 options and inversemask / mask of LITTLE_ENDIAN.
15373 * config/microblaze/t-microblaze: Expand multilib options to
15374 include mlittle-endian (le) and update exceptions patterns.
15376 2013-02-08 Jakub Jelinek <jakub@redhat.com>
15378 PR rtl-optimization/56195
15379 * lra-constraints.c (get_reload_reg): Don't reuse regs
15380 if they have smaller mode than requested, if they have
15381 wider mode than requested, try to return a SUBREG.
15383 PR tree-optimization/56250
15384 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15385 if type is unsigned and code isn't MULT_EXPR.
15387 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15389 PR tree-optimization/56064
15390 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15391 bits according to mode.
15392 * fixed-value.h (fixed_from_double_int)
15393 (const_fixed_from_double_int): Adjust comments.
15395 2013-02-08 Richard Biener <rguenther@suse.de>
15398 * lto-streamer.h (struct data_in): Remove current_file, current_line
15399 and current_col members.
15400 * lto-streamer-out.c (lto_output_location): Stream changed bits
15401 en-block for efficiency.
15402 * lto-streamer-in.c (clear_line_info): Remove.
15403 (lto_input_location): Cache current file, line and column
15404 globally via local statics. Read changed bits en-block.
15405 (input_function): Do not call clear_line_info.
15406 (lto_read_body): Likewise.
15407 (lto_input_toplevel_asms): Likewise.
15409 2013-02-08 Michael Matz <matz@suse.de>
15411 PR tree-optimization/52448
15412 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15413 (nt_call_phase): New static.
15414 (add_or_mark_expr): Only mark accesses with newer phase than any
15416 (nonfreeing_call_p): New.
15417 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15418 (nt_fini_block): Keep blocks marked as visited.
15419 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15421 2013-02-08 Richard Biener <rguenther@suse.de>
15423 * ira.c (ira): Free broken dominator information.
15425 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15427 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15429 2013-02-08 Marek Polacek <polacek@redhat.com>
15431 * cfgloop.c (verify_loop_structure): Add more checking of headers.
15433 2013-02-08 Richard Biener <rguenther@suse.de>
15435 PR middle-end/56181
15436 * cfgloop.h (flow_loops_find): Adjust.
15437 (bb_loop_header_p): Declare.
15438 * cfgloop.c (bb_loop_header_p): New function split out from ...
15439 (flow_loops_find): ... here. Adjust function signature,
15440 support incremental loop structure update.
15441 (verify_loop_structure): Cleanup. Verify a loop is a loop.
15442 * cfgloopmanip.c (fix_loop_structure): Move ...
15443 * loop-init.c (fix_loop_structure): ... here.
15444 (apply_loop_flags): Split out from ...
15445 (loop_optimizer_init): ... here.
15446 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
15447 in incremental mode, only remove dead loops here.
15449 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15452 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
15453 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
15454 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
15455 (*round<mode>3.libgcc): New insns for fixed-modes.
15456 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
15457 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
15458 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
15459 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
15460 implementations. Define to __builtin_avr_absFX,
15461 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
15462 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
15463 __builtin_avr_countlsFX, respectively.
15464 * config/avr/avr-c.c (target.h): Include it.
15465 (enum avr_builtin_id): New enum.
15466 (avr_resolve_overloaded_builtin): New static function.
15467 (avr_register_target_pragmas): Use it to set
15468 targetm.resolve_overloaded_builtin.
15469 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
15470 tree nodes used by DEF_BUILTIN.
15471 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
15472 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
15473 <AVR_BUILTIN_xxBITS>: Same.
15475 2013-02-08 Richard Biener <rguenther@suse.de>
15477 * cfgloop.c (verify_loop_structure): Properly handle
15478 a loop exiting to another loop header.
15479 * ira-int.h (ira_loops): Remove.
15480 * ira.c (ira_loops): Remove.
15481 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
15482 (do_reload): Use loop_optimizer_finalize.
15483 * ira-build.c (create_loop_tree_nodes): Use get_loops and
15484 number_of_loops to access the loop tree.
15485 (more_one_region_p): Likewise.
15486 (finish_loop_tree_nodes): Likewise.
15487 (rebuild_regno_allocno_maps): Likewise.
15488 (mark_loops_for_removal): Likewise.
15489 (mark_all_loops_for_removal): Likewise.
15490 (remove_unnecessary_regions): Likewise.
15491 (ira_build): Likewise.
15492 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15494 2013-02-08 Richard Biener <rguenther@suse.de>
15496 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
15497 * ipa-pure-const.c (analyze_function): Avoid calling
15498 mark_irreducible_loops twice.
15499 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
15501 2013-02-07 David S. Miller <davem@davemloft.net>
15503 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
15505 * var-tracking.c (vt_add_function_parameter): Test the presence of
15506 HAVE_window_save properly and do not remap argument registers when
15507 we have a leaf function.
15509 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
15512 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
15514 * config/i386/i386.c (ix86_print_operand): Ditto.
15516 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15518 * lra-constraints.c (process_alt_operands): Fix recently added comment.
15520 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15522 PR rtl-optimization/56225
15523 * lra-constraints.c (process_alt_operands): Check that reload hard
15524 reg can hold value for strict_low_part.
15526 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15529 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
15530 dwarf2out_end_function.
15531 (in_first_function_p, maybe_at_text_label_p,
15532 first_loclabel_num_not_at_text_label): New variables.
15533 (dwarf2out_var_location): In the first function find out
15534 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
15535 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
15538 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
15540 PR rtl-optimization/56178
15541 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
15542 SUBREG of a register. Tidy up related block of code.
15543 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
15544 note if the source is a register or a SUBREG of a register.
15546 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15549 * config/rs6000/rs6000.md (ptrm): New mode attr.
15550 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
15551 call_value_indirect_aix<pttrsize>,
15552 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
15555 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
15557 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
15558 if -bnortl. Convert to strcmp and strncmp.
15560 2013-02-07 Alan Modra <amodra@gmail.com>
15563 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
15564 addresses won't wrap when offsetting.
15565 (rs6000_secondary_reload): Provide secondary reloads needed for
15566 wrapping LO_SUM addresses.
15568 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
15570 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
15571 MACH, just __MACH__.
15573 2013-02-06 Richard Biener <rguenther@suse.de>
15575 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
15576 instead of calling fix_loop_structure.
15578 2013-02-06 Jakub Jelinek <jakub@redhat.com>
15580 PR middle-end/56217
15581 * omp-low.c (use_pointer_for_field): Return false if
15582 lower_send_shared_vars doesn't generate any copy-out code.
15584 2013-02-06 Tom de Vries <tom@codesourcery.com>
15586 PR rtl-optimization/56131
15587 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
15588 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
15589 of the label is NULL. Add comment.
15591 2013-02-05 Jakub Jelinek <jakub@redhat.com>
15593 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
15596 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
15597 (STATIC_LIBTSAN_LIBS): Likewise.
15598 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
15599 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
15600 is defined, don't add anything else beyond that.
15601 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
15602 (LINK_COMMAND_SPEC): Use them.
15604 PR tree-optimization/56205
15605 * tree-stdarg.c (check_all_va_list_escapes): Return true if
15606 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
15607 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
15609 2013-02-05 Richard Biener <rguenther@suse.de>
15611 PR tree-optimization/53342
15612 PR tree-optimization/53185
15613 * tree-vectorizer.h (vect_check_strided_load): Remove.
15614 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
15615 not disallow peeling for vectorized strided loads.
15616 (vect_check_strided_load): Make static and simplify.
15617 (vect_analyze_data_refs): Adjust.
15618 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
15619 correctly when vectorizing strided loads.
15621 2013-02-05 Richard Biener <rguenther@suse.de>
15623 * doc/install.texi: Refer to ISL, not PPL.
15625 2013-02-05 Jan Hubicka <jh@suse.cz>
15627 PR tree-optimization/55789
15628 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
15630 2013-02-05 Jan Hubicka <jh@suse.cz>
15632 PR tree-optimization/55789
15633 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
15634 the dead call anyway.
15636 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
15639 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
15641 2013-02-04 Alexander Potapenko <glider@google.com>
15642 Jack Howarth <howarth@bromo.med.uc.edu>
15643 Jakub Jelinek <jakub@redhat.com>
15646 * config/darwin.c (sort_ctor_records): Stabilized qsort
15647 on constructor priority by using original position.
15648 (finalize_ctors): New routine to sort constructors by
15649 priority before use in assemble_integer.
15650 (machopic_asm_out_constructor): Use finalize_ctors if needed.
15652 2013-02-04 Jakub Jelinek <jakub@redhat.com>
15655 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
15656 about visibility on artificial decls.
15657 * config/sol2.c (solaris_assemble_visibility): Likewise.
15659 2013-02-04 Kai Tietz <ktietz@redhat.com>
15662 * config/i386/i386.c (function_value_ms_64): Add additional valtype
15663 argument and improve checking of return-argument types for 16-byte
15665 (ix86_function_value_1): Add additional valtype argument on call
15666 of function_value_64.
15667 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
15668 handling infunction_value_64 function.
15670 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
15672 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
15674 2013-02-04 Richard Biener <rguenther@suse.de>
15676 PR tree-optimization/56188
15677 * tree-ssa-structalias.c (label_visit): Consider case with
15678 initially non-empty points-to set.
15679 (perform_var_substitution): Dump node mapping and clean up.
15681 2013-02-04 Richard Guenther <rguenther@suse.de>
15684 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
15685 node prevail as last resort.
15686 (lto_symtab_merge_decls): Remove guard on LTRANS here.
15687 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
15689 2013-02-04 Richard Biener <rguenther@suse.de>
15691 PR tree-optimization/56113
15692 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
15694 (equiv_class_lookup_or_add): ... this.
15695 (label_visit): Adjust and fix error in previous patch.
15696 (perform_var_substitution): Adjust.
15698 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
15700 * config/sh/divtab.c: Fix formatting and comments throughout the file.
15701 * config/sh/sh4-300.md: Likewise.
15702 * config/sh/sh4a.md: Likewise.
15703 * config/sh/constraints.md: Likewise.
15704 * config/sh/sh.md: Likewise.
15705 * config/sh/netbsd-elf.h: Likewise.
15706 * config/sh/predicates.md: Likewise.
15707 * config/sh/sh-protos.h: Likewise.
15708 * config/sh/ushmedia.h: Likewise.
15709 * config/sh/linux.h: Likewise.
15710 * config/sh/sh.c: Likewise.
15711 * config/sh/superh.h: Likewise.
15712 * config/sh/elf.h: Likewise.
15713 * config/sh/sh4.md: Likewise.
15714 * config/sh/sh.h: Likewise.
15716 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15718 * config/pa/constraints.md: Adjust unused letters. Change "T"
15719 constraint to match_test floating_point_store_memory_operand().
15720 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
15721 (base14_operand): New.
15722 (floating_point_store_memory_operand): New.
15723 (integer_store_memory_operand): Revise to use base14_operand and
15724 reg_plus_base_memory_operand.
15725 (move_dest_operand): Allow symbolic_memory_operands.
15726 (symbolic_memory_operand): Check for LO_SOM.
15727 (symbolic_operand): Change default case to break.
15728 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
15729 CONST_DOUBLE values to be reloaded by putting them into memory when
15730 the destination is a floating point register.
15731 (movdf): Remove code to handle CONST_DOUBLE.
15733 (reload_indf_r1): New.
15734 (reload_insf_r1): New.
15735 Consistently use "Q" and "T" constraints with integer and floating
15736 point move instructions, respectively.
15737 (movdi): Remove FAIL.
15738 Change predicate for source operand unamed DImode move from
15739 general_operand to move_src_operand.
15740 (umulsidi3): Change predicate for destination operand to
15742 Likewise for similar unamed patterns.
15743 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
15744 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
15745 (hppa_legitimize_address): Simplify mask calculation.
15746 (pa_emit_move_sequence): Revised handling of secondary reloads from
15747 REG+D addresses for floating point loads and stores. Directly handle
15748 loading CONST0_RTX (mode) to a floating point register.
15749 (pa_secondary_reload): Handle reloading DF and SFmode constant values
15750 to floating point registers. Don't restrict secondary reloads to
15751 floating point registers to integer modes. Revise some comments and
15753 (TARGET_LEGITIMATE_ADDRESS_P): Define.
15754 (pa_legitimate_address_p): New.
15755 (pa_legitimize_reload_address): New.
15756 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
15757 (STRICT_REG_OK_FOR_BASE_P): New.
15758 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
15759 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
15761 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
15762 Andrew Dixie <andrewd@gentrack.com>
15764 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
15767 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
15769 * expmed.c (extract_bit_field_1): Pass the full width of the
15770 structure to get_best_reg_extraction_insn.
15772 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
15775 * configure.ac (use_cxa_atexit): Add AIX.
15776 * configure: Regenerate.
15778 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
15780 2013-02-01 Jakub Jelinek <jakub@redhat.com>
15783 * final.c (need_profile_function): New variable.
15784 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
15785 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
15786 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
15787 notes, targetm.asm_out.function_prologue doesn't emit anything,
15788 HAVE_prologue and profiler should be emitted before prologue,
15789 set need_profile_function instead of emitting it.
15790 (final_scan_insn): If need_profile_function, emit
15791 profile_function on the first NOTE_INSN_BASIC_BLOCK or
15792 NOTE_INSN_FUNCTION_BEG note.
15794 2013-02-01 Richard Henderson <rth@redhat.com>
15796 * config/rs6000/rs6000.md (smulditi3): New.
15799 * config/alpha/alpha.md (umulditi3): New.
15801 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
15803 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
15804 (ASM_OUTPUT_ALIGNED_LOCAL): New.
15806 2013-02-01 Richard Biener <rguenther@suse.de>
15808 PR tree-optimization/56113
15809 * tree-ssa-structalias.c (label_visit): Reduce work for
15810 single-predecessor nodes.
15812 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
15814 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
15815 range isn't testing for zero.
15817 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
15819 PR middle-end/56113
15820 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
15822 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
15823 Nick Clifton <nickc@redhat.com>
15825 * config/v850/constraints.md (Q): Define as a memory constraint.
15826 * config/v850/predicates.md (label_ref_operand): New predicate.
15827 (e3v5_shift_operand): New predicate.
15828 (ior_operator): New predicate.
15829 * config/v850/t-v850: Add e3v5 multilib.
15830 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
15831 (v850_gen_movdi): Prototype.
15832 * config/v850/v850.c: Add support for e3v5 architecture.
15833 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
15835 (construct_save_jarl): Add e3v5 long JARL support.
15836 (v850_adjust_insn_length): New function. Adjust length of call
15837 insns when using e3v5 instructions.
15838 (v850_gen_movdi): New function: Generate instructions to move a
15840 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
15841 (CPP_SPEC): Define __v850e3v5__ as appropriate.
15842 (TARGET_USE_FPU): Enable for e3v5.
15843 (CONST_OK_FOR_W): New macro.
15844 (ADJUST_INSN_LENGTH): Define.
15845 * config/v850/v850.md (UNSPEC_LOOP): Define.
15846 (attr cpu): Add v850e3v5.
15847 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
15848 (movdi): New pattern.
15849 (movdi_internal): New pattern.
15850 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
15851 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
15852 (cstoresf4): Likewise.
15853 (cstoredf4): Likewise.
15854 (insv): New pattern.
15855 (rotlso3_a): New pattern.
15856 (rotlsi3_b): New pattern
15857 (rotlsi3_v850e3v5): New pattern.
15858 (doloop_begin): New pattern.
15859 (fix_loop_counter): New pattern.
15860 (doloop_end): New pattern.
15861 (branch_normal): Add e3v5 long branch support.
15862 (branch_invert): Likewise.
15863 (branch_z_normal): Likewise.
15864 (branch_z_invert): Likewise.
15865 (branch_nz_normal): Likewise.
15866 (branch_nz_invert): Likewise.
15867 (call_internal_short): Add e3v5 register-indirect JARL support.
15868 (call_internal_long): Likewise.
15869 (call_value_internal_short): Likewise.
15870 (call_value_internal_long): Likewise.
15871 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
15872 (mloop): New option.
15873 * config.gcc: Add support for configuring v840e3v5 target.
15874 * doc/invoke.texi: Document new v850 specific command line options.
15876 2013-01-31 Paul Koning <ni1d@arrl.net>
15880 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
15881 children if parent is a class.
15882 (prune_unused_types_prune): Don't add DW_AT_declaration.
15884 2013-01-31 Richard Biener <rguenther@suse.de>
15886 PR tree-optimization/56157
15887 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
15888 match up operand with SLP child.
15890 2013-01-31 Jason Merrill <jason@redhat.com>
15893 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
15894 parameters the first time.
15895 (gen_scheduled_generic_parms_dies): Check completeness here.
15897 2013-01-31 Richard Biener <rguenther@suse.de>
15899 PR middle-end/53073
15900 * common.opt (faggressive-loop-optimizations): New flag,
15901 enabled by default.
15902 * doc/invoke.texi (faggressive-loop-optimizations): Document.
15903 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
15904 infer_loop_bounds_from_undefined by it.
15906 2013-01-31 Richard Biener <rguenther@suse.de>
15908 PR tree-optimization/56150
15909 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
15910 visit virtual operands.
15911 (find_uses_to_rename_bb): Likewise.
15913 2013-01-31 Richard Biener <rguenther@suse.de>
15915 PR tree-optimization/56150
15916 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
15917 mixed store non-store stmts.
15919 2013-01-30 Jakub Jelinek <jakub@redhat.com>
15922 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
15923 LIBASAN_EARLY_SPEC is defined.
15924 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
15925 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
15927 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
15930 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
15931 invalid args instead of ICEing on it.
15932 (ix86_valid_target_attribute_tree): Return error_mark_node if
15933 ix86_valid_target_attribute_inner_p failed.
15934 (ix86_valid_target_attribute_p): Return false only if
15935 ix86_valid_target_attribute_tree returned error_mark_node. Allow
15936 target("default") attribute.
15937 (sorted_attr_string): Change argument from const char * to tree,
15938 merge in all target attribute arguments rather than just one.
15939 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
15940 instead of free. Avoid using strcat.
15941 (ix86_mangle_function_version_assembler_name): Mangle
15942 target("default") as if no target attribute is present. Adjust
15943 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
15944 instead of xmalloc and XDELETEVEC instead of free.
15945 (ix86_function_versions): Don't return true if one of the decls
15946 doesn't have target attribute. If they don't and one of the decls
15947 is DECL_FUNCTION_VERSIONED, report an error. Adjust
15948 sorted_attr_string caller. Use XDELETEVEC instead of free.
15949 (ix86_supports_function_versions): Remove.
15950 (make_name): Fix up formatting.
15951 (make_dispatcher_decl): Remove resolver_name and its initialization.
15952 Avoid leaking memory.
15953 (is_function_default_version): Return true if there is
15954 target("default") attribute rather than no target attribute at all.
15955 (make_resolver_func): Avoid leaking memory.
15956 (ix86_generate_version_dispatcher_body): Likewise.
15957 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
15958 * target.def (supports_function_versions): Remove.
15959 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
15960 * doc/tm.texi: Regenerated.
15962 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
15964 PR rtl-optimization/56144
15965 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
15966 for values with side effects.
15968 2013-01-30 Richard Biener <rguenther@suse.de>
15970 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
15971 (sparseset_pop): Likewise.
15972 * cfganal.c (compute_idf): Likewise. Increase work-stack size
15973 to be able to use quick_push in the worker loop.
15975 2013-01-30 Marek Polacek <polacek@redhat.com>
15977 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
15979 2013-01-30 Richard Biener <rguenther@suse.de>
15982 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
15984 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
15986 PR tree-optimization/56064
15987 * fixed-value.c (fixed_from_double_int): New function.
15988 * fixed-value.h (fixed_from_double_int): New prototype.
15989 (const_fixed_from_double_int): New static inline function.
15990 * fold-const.c (native_interpret_fixed): New static function.
15991 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
15992 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
15993 (native_encode_fixed): New static function.
15994 (native_encode_expr) <FIXED_CST>: Use it.
15995 (native_interpret_int): Move double_int worker code to...
15996 * double-int.c (double_int::from_buffer): ...this new static method.
15997 * double-int.h (double_int::from_buffer): Prototype it.
15999 2013-01-30 Richard Biener <rguenther@suse.de>
16001 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
16002 New pointer-map and obstack.
16003 (init_alias_vars): Allocate pointer-map and obstack.
16004 (delete_points_to_sets): Free them.
16005 (find_what_var_points_to): Cache result.
16006 (find_what_p_points_to): Adjust for changed interface of
16007 find_what_var_points_to.
16008 (compute_points_to_sets): Likewise.
16009 (ipa_pta_execute): Likewise.
16011 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16013 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
16014 * configure: Regenerate.
16015 * config.in: Regenerate.
16016 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
16017 #nobits/#progbits if supported.
16019 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
16022 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
16023 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
16024 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
16026 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16028 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
16029 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
16031 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16033 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
16034 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
16036 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16038 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
16040 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
16041 * config/arm/cortex-a7.md: New bypasses using
16042 arm_mac_accumulator_is_result.
16044 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16046 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
16047 (cortex_a7_neon_mla): Likewise.
16048 (cortex_a7_fpfmad): New reservation.
16049 (cortex_a7_fpmacs): Use ffmas and update required units.
16050 (cortex_a7_fpmuld): Update required units and latency.
16051 (cortex_a7_fpmacd): Likewise.
16052 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
16053 (cortex_a7_neon). Likewise.
16054 (bypass) Update participating units.
16056 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16058 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
16059 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
16061 * config/arm/vfp11.md (vfp_farith): Use ffmas.
16062 (vfp_fmul): Use ffmad.
16063 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
16064 (cortex_r4_fmacd): Use ffmad.
16065 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
16066 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
16067 (cortex_a9_fmacd): Use ffmad.
16068 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
16069 (cortex_a8_vfp_macd): Use ffmad.
16070 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
16071 (cortex_a5_fpmacd): Use ffmad.
16072 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
16073 (cortex_a15_vfp_macd): Use ffmad.
16074 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
16076 2013-01-29 Jason Merrill <jason@redhat.com>
16079 * varasm.c (default_assemble_visibility): Don't warn about
16080 visibility on artificial decls.
16082 2013-01-29 Richard Biener <rguenther@suse.de>
16084 PR tree-optimization/56113
16085 * tree-ssa-structalias.c (equiv_class_lookup): Also return
16087 (label_visit): Free duplicate bitmaps and record the leader instead.
16088 (perform_var_substitution): Adjust.
16090 2013-01-29 Richard Biener <rguenther@suse.de>
16092 PR tree-optimization/55270
16093 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16094 the CFG, schedule loops for fixup.
16096 2013-01-29 Nick Clifton <nickc@redhat.com>
16098 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16101 2013-01-28 Leif Ekblad <leif@rdos.net>
16103 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16104 * config/i386/i386.h (TARGET_RDOS): New macro.
16105 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16106 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16107 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16108 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16109 DEFAULT_LARGE_SECTION_THRESHOLD.
16110 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16111 * config/i386/rdos.h: New file.
16112 * config/i386/rdos64.h: New file.
16114 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16117 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16120 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16122 PR rtl-optimization/56117
16123 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16124 call cselib_lookup_from_insn on the MEM before calling
16125 add_insn_mem_dependence.
16127 2013-01-28 Richard Biener <rguenther@suse.de>
16129 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16130 to a stmt that didn't have one.
16131 (copy_phis_for_bb): Likewise for PHI arguments.
16132 (copy_debug_stmt): Likewise for debug stmts.
16134 2013-01-28 Richard Biener <rguenther@suse.de>
16136 PR tree-optimization/56034
16137 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
16138 (partition_builtin_p): Adjust.
16139 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16140 it is the last partition.
16141 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16142 up the vertex for the definition.
16143 (classify_partition): Classify whether a partition is a
16144 PKIND_REDUCTION, thus has uses outside of the loop.
16145 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16146 Merge all PKIND_REDUCTION partitions into the last partition.
16147 (tree_loop_distribution): Seed partitions from reductions as well.
16149 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16151 PR tree-optimization/56125
16152 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16153 pow(x,c) into sqrt(x) * powi(x, n/2) or
16154 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16155 optimizing for size.
16156 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16157 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16160 PR tree-optimization/56094
16161 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16162 to UNKNOWN_LOCATION while gimplifying expr.
16164 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
16167 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16168 operand 0 in movabs insn template for -masm=intel asm alternative.
16169 (*movabs<mode>_2): Ditto for operand 1.
16171 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
16174 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16177 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16179 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16182 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16184 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16185 Undef to avoid warning.
16187 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16189 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16190 * configure: Regenerate.
16192 2013-01-25 Jakub Jelinek <jakub@redhat.com>
16194 PR tree-optimization/56098
16195 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16196 for stmts with volatile ops.
16197 (cond_store_replacement): Don't optimize if assign has volatile ops.
16198 (cond_if_else_store_replacement_1): Don't optimize if either
16199 then_assign or else_assign have volatile ops.
16200 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16203 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16205 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16207 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16209 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16210 missing ':' in asm example.
16212 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16214 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16215 entries into lane and laneq entries.
16216 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16217 Remove AdvSIMD scalar modes.
16218 (aarch64_sq<r>dmulh_laneq<mode>): New.
16219 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16221 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16222 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16223 * config/aarch64/iterators.md (VCOND): New.
16226 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16229 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16230 Add NULL LIBNAME argument to existing definitions.
16231 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16232 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16233 * config/avr/avr.c (DEF_BUILTIN): Same.
16234 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16235 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16236 implementation is available (DECL_ASSEMBLER_NAME is set).
16237 (avr_fold_absfx): New static function.
16238 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16239 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16240 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16241 AVR_BUILTIN_ABSLLK.
16242 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16243 (abshk, absk, abslk, absllk): Provide as static inline functions.
16245 2013-01-25 Marek Polacek <polacek@redhat.com>
16247 PR tree-optimization/56035
16248 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16250 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
16252 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16253 (*movtf_internal_rex64): Add (!o,C) alternative
16254 (*movxf_internal_rex64): Ditto.
16255 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16257 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
16259 * doc/invoke.texi: fix typo.
16260 * doc/objc.texi: fix typo.
16262 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16264 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16265 for the first two alternatives.
16267 2013-01-24 Diego Novillo <dnovillo@google.com>
16269 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
16270 (ggc-zone.o): Remove.
16271 * configure.ac: Remove option --with-gc.
16272 * configure: Re-generate.
16273 * doc/install.texi: Remove documentation for --with-gc.
16274 * gengtype.c (write_enum_defn): Remove. Update all users.
16275 (write_Types_process_field): Remove generation of gt_e_* argument.
16276 (output_type_enum): Remove. Update all users.
16277 (write_enum_defn): Remove. Update all users.
16278 (enum alloc_zone): Remove. Update all users.
16279 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
16280 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16281 Update all callers.
16282 (struct ptr_data): Remove field TYPE. Update all users.
16283 (gt_pch_note_object): Remove argument TYPE. Update all users.
16284 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16286 * ggc-none.c (ggc_alloc_typed_stat): Remove.
16287 (struct alloc_zone): Remove.
16288 (ggc_internal_alloc_zone_stat): Remove.
16289 (ggc_internal_cleared_alloc_zone_stat): Remove.
16290 * ggc-page.c (ggc_alloc_typed_stat): Remove.
16291 (ggc_pch_count_object): Remove last argument. Update all users.
16292 (ggc_pch_alloc_object): Remove last argument. Update all users.
16293 (struct alloc_zone): Remove.
16294 * ggc-zone.c: Remove.
16295 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
16296 (struct alloc_zone): Remove.
16297 (ggc_alloc_typed_stat): Remove.
16298 (ggc_alloc_typed): Remove.
16299 (ggc_splay_alloc): Remove first argument.
16300 (rtl_zone): Remove. Update all users.
16301 (tree_zone): Remove. Update all users.
16302 (tree_id_zone): Remove. Update all users.
16303 (ggc_internal_zone_alloc_stat): Remove. Update all users.
16304 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
16305 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
16306 * tree-ssanames.c: Remove references to zone allocator in comments.
16308 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
16310 * config/avr/avr.c (avr_out_fract): Make register numbers that
16311 might be outside of source operand signed.
16313 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
16315 * config/i386/constraints.md (Yf): New constraint.
16316 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16317 of f constraint to conditionaly disable x87 register preferences.
16318 (*movdf_internal): Ditto.
16319 (*movsf_internal): Ditto.
16321 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16323 PR inline-asm/55934
16324 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16325 that have operands with impossible constraints.
16326 Add a FIXME for a speed-up opportunity.
16327 * lra-constraints.c (process_alt_operands): Verify that a class
16328 selected from constraints on asms is valid for the operand mode.
16329 (curr_insn_transform): Remove incorrect comment.
16331 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
16333 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16334 TOC operand is a valid symbol ref in the constant pool.
16336 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16338 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
16340 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
16343 * config/avr/stdfix.h: New file.
16344 * t-avr (stdfix-gcc.h): New rule to build it.
16345 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16347 2013-01-23 Kostya Serebryany <kcc@google.com>
16349 * config/darwin.h: remove dependency on
16350 CoreFoundation (asan on Mac OS).
16352 2013-01-23 Jakub Jelinek <jakub@redhat.com>
16355 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16356 instead of cmpdi_operand for first comparison operand.
16357 Don't assert that comparison operands aren't both constants.
16359 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16361 * doc/install.texi (Downloading the Source): Update references to
16362 downloading separate components.
16364 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16366 * doc/extend.texi (__int128): Improve grammar.
16368 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
16371 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16372 alternative to (o,r).
16373 (*movdi_internal_rex64): Remove (!o,n) alternative.
16374 (DImode immediate->memory splitter): Remove.
16375 (DImode immediate->memory peephole2): Remove.
16376 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16377 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16378 alternative to (!o,*r).
16379 (*movtf_internal_sse): New pattern.
16380 (*movxf_internal_rex64): New pattern.
16381 (*movxf_internal): Disable for TARGET_64BIT.
16382 (*movdf_internal_rex64): Remove (!o,F) alternative.
16384 2013-01-22 Jakub Jelinek <jakub@redhat.com>
16386 PR middle-end/56074
16387 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16388 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16389 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16390 stmt locations where LOCATION_LOCUS of the stmt location is
16391 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16394 * config/i386/i386.md (UNSPEC_STOS): New.
16395 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16396 *strsetqi_1): Add UNSPEC_STOS.
16398 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16401 * doc/invoke.texi: Remove left over -Wsynth example.
16403 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16405 PR tree-optimization/56051
16406 * fold-const.c (fold_binary_loc): Don't fold
16407 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16408 a narrowing conversion, or widening conversion from signed
16411 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16413 PR rtl-optimization/56023
16414 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16415 dependent on debug instruction.
16417 2013-01-21 Martin Jambor <mjambor@suse.cz>
16419 PR middle-end/56022
16420 * function.c (allocate_struct_function): Call
16421 invoke_set_current_function_hook earlier.
16423 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16425 * reload1.c (init_reload): Only initialize reload_obstack
16426 during the first call.
16428 2013-01-21 Marek Polacek <polacek@redhat.com>
16430 * cfgloop.c (verify_loop_structure): Fix up grammar.
16432 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
16434 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
16435 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
16437 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16440 * config/arm/marvell-pj4.md: Update copyright year.
16441 Fix up use of alu to alu_reg and simple_alu_imm.
16443 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16445 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
16447 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
16450 * lra-constraints.c (curr_insn_transform): Don't reuse original
16451 insn for secondary memory move when memory mode should be different.
16453 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16455 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
16456 atomic_storedi_1): New patterns.
16458 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16460 btver2 pipeline descriptions.
16461 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
16463 * config/i386/i386.md (btver2_decode): New type attributes.
16464 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
16466 * config/i386/btver2.md: New file describing btver2 pipelines.
16468 2013-01-19 Andrew Pinski <apinski@cavium.com>
16470 PR tree-optimization/52631
16471 * tree-ssa-sccvn (visit_use): Before looking up the original
16472 statement, try looking up the simplified expression.
16474 2013-01-19 Anthony Green <green@moxielogic.com>
16476 * config/moxie/moxie.c (moxie_expand_prologue): Set
16477 current_function_static_stack_size.
16479 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16481 PR tree-optimization/56029
16482 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
16483 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
16485 2013-01-18 Sharad Singhai <singhai@google.com>
16487 PR tree-optimization/55995
16488 * dumpfile.c (dump_loc): Print location only if available.
16489 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
16491 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
16494 * lra-constraints.c (curr_insn_transform): Reuse original insn for
16495 secondary memory move.
16496 (inherit_reload_reg): Use rclass instead of cl for
16497 check_secondary_memory_needed_p.
16499 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16501 PR middle-end/56015
16502 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
16503 the case where writing real complex part of target modifies op1.
16505 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
16507 * config/aarch64/aarch64-simd.md
16508 (aarch64_vcond_internal<mode>): Handle unordered cases.
16509 * config/aarch64/iterators.md (v_cmp_result): New.
16511 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
16512 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16514 * config/arm/marvell-pj4.md: New file.
16515 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
16516 * config/arm/arm.md (generic_sched): Add marvell_pj4.
16517 (generic_vfp): Likewise.
16518 * config/arm/arm-cores.def: Add marvell-pj4.
16519 * config/arm/arm-tune.md: Regenerate.
16520 * config/arm/arm-tables.opt: Regenerate.
16521 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
16522 * doc/invoke.texi: Document marvell-pj4.
16524 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
16526 * config/aarch64/arm_neon.h: Map scalar types to standard types.
16528 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16533 * var-tracking.c (negative_power_of_two_p): New.
16534 (global_get_addr_cache, local_get_addr_cache): New.
16535 (get_addr_from_global_cache, get_addr_from_local_cache): New.
16536 (vt_canonicalize_addr): Rewrite using the above. Adjust the
16538 (vt_stack_offset_p): Remove.
16539 (vt_canon_true_dep): Always canonicalize loc's address.
16540 (clobber_overlapping_mems): Make sure we have a MEM.
16541 (local_get_addr_clear_given_value): New.
16542 (val_reset): Clear local cached entries.
16543 (compute_bb_dataflow): Create and release the local cache.
16544 Disable duplicate MEMs clobbering.
16545 (emit_notes_in_bb): Clobber MEMs likewise.
16546 (vt_emit_notes): Create and release the local cache.
16547 (vt_initialize, vt_finalize): Create and release the global
16548 cache, respectively.
16549 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
16551 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16553 PR libmudflap/53359
16554 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
16555 not found in the symtab.
16557 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16560 PR rtl-optimization/55547
16561 PR rtl-optimization/53827
16564 * alias.c (offset_overlap_p): New, factored out of...
16565 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
16566 the conservative special case for symbolic constants. Don't
16567 adjust zero sizes on alignment.
16569 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
16571 PR rtl-optimization/52573
16572 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
16573 REG_UNUSED for the same register.
16575 2013-01-17 Richard Biener <rguenther@suse.de>
16576 Marek Polacek <polacek@redhat.com>
16578 PR rtl-optimization/55833
16579 * loop-unswitch.c (unswitch_loops): Move loop verification...
16580 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
16581 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
16582 Set it to true when we're removing a loop from hierarchy tree in
16583 an irreducible region.
16584 (fix_bb_placements): Adjust caller.
16585 (fix_loop_placements): Likewise.
16587 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
16589 * config/avr/builtins.def (DEF_BUILTIN): Factor out
16590 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
16591 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
16592 Remove ID. Adjust comments.
16593 * config/avr/avr-c.c (avr_builtin_name): Remove.
16594 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
16595 * config/avr/avr.c (avr_tolower): New static function.
16596 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
16597 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
16598 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
16601 2013-01-17 Jan Hubicka <jh@suse.cz>
16603 PR tree-optimization/55273
16604 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
16606 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
16609 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
16610 store through atomic_store<mode>_1.
16611 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
16613 2013-01-17 Martin Jambor <mjambor@suse.cz>
16615 PR tree-optimizations/55264
16616 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
16617 for virtual methods.
16618 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
16619 virtual methods before inlining is over.
16620 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
16622 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
16625 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
16627 PR rtl-optimization/56005
16628 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
16629 pending reads for prefetch.
16631 2013-01-16 Ian Bolton <ian.bolton@arm.com>
16633 * config/aarch64/aarch64.md
16634 (*cstoresi_neg_uxtw): New pattern.
16635 (*cmovsi_insn_uxtw): New pattern.
16636 (*<optab>si3_uxtw): New pattern.
16637 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
16638 (*<optab>si3_insn_uxtw): New pattern.
16639 (*bswapsi2_uxtw): New pattern.
16641 2013-01-16 Richard Biener <rguenther@suse.de>
16643 * tree-inline.c (tree_function_versioning): Remove set but
16644 never used variable.
16646 2013-01-16 Richard Biener <rguenther@suse.de>
16648 PR tree-optimization/55964
16649 * tree-flow.h (rename_variables_in_loop): Remove.
16650 (rename_variables_in_bb): Likewise.
16651 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
16652 (copy_loop_before): Adjust and delete update-ssa status.
16653 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
16654 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
16655 (rename_variables_in_loop): Remove.
16656 (slpeel_update_phis_for_duplicate_loop): Likewise.
16657 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
16658 use available cfg machinery instead of duplicating it.
16659 Update PHI nodes and perform poor-mans SSA update here.
16660 (slpeel_tree_peel_loop_to_edge): Adjust.
16662 2013-01-16 Richard Biener <rguenther@suse.de>
16664 PR tree-optimization/54767
16665 PR tree-optimization/53465
16666 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
16667 (vrp_visit_phi_node): For PHI arguments coming via backedges
16668 drop all symbolical range information.
16669 (execute_vrp): Compute backedges.
16671 2013-01-16 Richard Biener <rguenther@suse.de>
16673 * doc/install.texi: Update CLooG and ISL requirements to
16676 2013-01-16 Christian Bruel <christian.bruel@st.com>
16679 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
16680 (broken_move): Handle UNSPECV_SP_SWITCH_B.
16681 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
16683 2013-01-16 DJ Delorie <dj@redhat.com>
16685 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
16686 (UNSPECV_SP_SWITCH_E): New.
16687 (sp_switch_1): Change to an unspec.
16688 (sp_switch_2): Change to an unspec. Don't use post-inc when we
16691 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
16693 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
16694 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
16695 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
16696 (expand_mem_thread_fence): Ditto.
16697 (expand_mem_signal_fence): Ditto.
16698 (expand_atomic_load): Ditto.
16699 (expand_atomic_store): Ditto.
16701 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
16703 PR rtl-optimization/55547
16704 PR rtl-optimization/53827
16707 * alias.c (memrefs_conflict_p): Set sizes to negative after
16710 2013-01-15 Jakub Jelinek <jakub@redhat.com>
16713 * function.c (thread_prologue_and_epilogue_insns): Always
16714 add crtl->drap_reg to set_up_by_prologue.set, even if
16715 stack_realign_drap is false.
16717 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16719 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
16720 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
16721 *call): Fix indention.
16723 2013-01-15 Tom de Vries <tom@codesourcery.com>
16726 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
16729 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
16731 PR rtl-optimization/55153
16732 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
16734 2013-01-15 Martin Jambor <mjambor@suse.cz>
16736 PR tree-optimization/55920
16737 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
16738 accesses as grp_to_be_debug_replaced.
16740 2013-01-15 Jakub Jelinek <jakub@redhat.com>
16742 PR tree-optimization/55920
16743 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
16744 there is non-useless type conversion needed from debug rhs to lhs,
16745 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
16747 2013-01-15 Joseph Myers <joseph@codesourcery.com>
16748 Mikael Pettersson <mikpe@it.uu.se>
16751 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
16753 (ASM_OUTPUT_CASE_LABEL): Remove.
16754 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
16755 * final.c (shorten_branches): Update alignment of labels before
16756 jump tables if CASE_VECTOR_SHORTEN_MODE.
16758 2013-01-15 Richard Biener <rguenther@suse.de>
16761 * system.h: Do not include gmp.h for building host tools.
16763 2013-01-15 Richard Biener <rguenther@suse.de>
16765 PR middle-end/55882
16766 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
16767 account for bitpos when computing alignment.
16769 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
16771 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
16772 (ix86_target_macros_internal): Likewise.
16774 * config/i386/i386.c (m_CORE2I7): Removed.
16775 (m_CORE_HASWELL): New macro.
16776 (m_CORE_ALL): Likewise.
16777 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
16778 (initial_ix86_arch_features): Likewise.
16779 (processor_target_table): Initializations for Core avx2.
16780 (cpu_names): New names "core-avx2".
16781 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
16782 PROCESSOR_CORE_HASWELL.
16783 (ix86_issue_rate): New case.
16784 (ia32_multipass_dfa_lookahead): Likewise.
16785 (ix86_sched_init_global): Likewise.
16787 * config/i386/i386.h (TARGET_HASWELL): New macro.
16788 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
16789 (processor_type): New PROCESSOR_HASWELL.
16791 2013-01-15 Jakub Jelinek <jakub@redhat.com>
16793 PR tree-optimization/55955
16794 * tree-vect-loop.c (vectorizable_reduction): Give up early on
16795 *SHIFT_EXPR and *ROTATE_EXPR codes.
16797 PR tree-optimization/48766
16798 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
16799 -ftrapv disable -fwrapv.
16801 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16804 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
16805 etc. to 1 and not to __flash.
16806 Use LL suffix for __INT24_MAX__ with -mint8.
16807 Use ULL suffix for __UINT24_MAX__ with -mint8.
16809 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16811 * config/avr/avr-arch.h
16812 (struct base_arch_s): Use typedef avr_arch_t instead.
16813 (struct arch_info_s): Use typedef avr_arch_info_t instead.
16814 (struct mcu_type_s): Use typedef avr_mcu_t instead.
16815 * config/avr/avr.c: Same.
16816 * config/avr/avr-devices.c: Same.
16817 * config/avr/driver-avr.c: Same.
16818 * config/avr/gen-avr-mmcu-texi.c: Same.
16819 * config/avr/avr-mcus.def: Adjust comment.
16821 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
16823 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
16824 * config/aarch64/iterators.md (VALLDI): New.
16826 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
16827 Andi Kleen <ak@linux.intel.com>
16830 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
16831 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
16834 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16836 * config/avr/avr-stdint.h: Remove trailing blanks.
16837 * config/avr/avr-log.h: Same.
16838 * config/avr/avr-arch.h: Same.
16839 * config/avr/avr-devices.c: Same.
16840 * config/avr/avr-dimode.md: Same.
16841 * config/avr/predicates.md: Same.
16842 * config/avr/avr-c.c: Same. And fix typo.
16844 * config/avr/avr-protos.h: Same. And:
16845 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
16846 (init_cumulative_args): Rename to avr_init_cumulative_args.
16847 (expand_prologue): Rename to avr_expand_prologue.
16848 (expand_epilogue): Rename to avr_expand_epilogue.
16849 (adjust_insn_length): Rename to avr_adjust_insn_length.
16850 (notice_update_cc): Rename to avr_notice_update_cc.
16851 (final_prescan_insn): Rename to avr_final_prescan_insn.
16852 * config/avr/avr.c: Same.
16853 * config/avr/avr.h: Same.
16854 * config/avr/avr.md: Remove trailing blanks.
16855 (prologue): Use avr_expand_prologue.
16856 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
16858 2013-01-14 Richard Biener <rguenther@suse.de>
16860 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
16861 verify_location, collect_subblocks): New functions.
16862 (verify_gimple_in_cfg): Verify that locations only reference
16863 BLOCKs in the functions BLOCK tree.
16865 2013-01-14 Richard Biener <rguenther@suse.de>
16867 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
16869 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
16871 (insert_out_of_ssa_copy_on_edge): Likewise.
16872 (rewrite_close_phi_out_of_ssa): Likewise.
16873 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
16875 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
16876 propagated constants.
16877 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
16880 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16882 * config/avr/avr-modes.def: Add GPL copyright notice.
16884 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
16886 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
16887 MEMMODEL_MASK to determine memory model.
16888 (atomic_store<mode>): Ditto from operands[2].
16889 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
16891 2013-01-13 Jakub Jelinek <jakub@redhat.com>
16894 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
16895 (fold_gimple_assign): Don't call unshare_expr here.
16896 (fold_ctor_reference): Call unshare_expr.
16898 2013-01-13 Terry Guo <terry.guo@arm.com>
16900 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
16901 * doc/fragments.texi: Document MULTILIB_REUSE.
16902 * gcc.c (multilib_reuse): New internal spec.
16903 (set_multilib_dir): Also search multilib from multilib_reuse.
16904 * genmultilib (tmpmultilib3): Refactor code.
16905 (tmpmultilib4): Ditto.
16906 (multilib_reuse): New multilib argument.
16908 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
16910 * Makefile.in: Update copyright.
16912 2013-01-12 Tom de Vries <tom@codesourcery.com>
16914 PR middle-end/55890
16915 * calls.c (expand_call): Check if arg_nr is valid.
16917 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
16919 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
16920 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
16921 documentation. Add missing '__' in front of
16922 __builtin_ia32_packssdw256.
16924 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16927 * config/s390/s390.c (s390_preferred_reload_class): Do not return
16928 NO_REGS for larl operands.
16929 (s390_reload_larl_operand): Use s390_load_address instead of
16932 2013-01-11 Richard Biener <rguenther@suse.de>
16934 * tree-cfg.c (verify_node_sharing_1): Split out from ...
16935 (verify_node_sharing): ... here.
16936 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
16938 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
16940 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
16941 Substitute TREECHECKING.
16942 * configure: Regenerate.
16943 * Makefile.in (TREECHECKING): New.
16945 2013-01-11 Richard Guenther <rguenther@suse.de>
16947 PR tree-optimization/44061
16948 * tree-vrp.c (extract_range_basic): Compute zero as
16949 value-range for __builtin_constant_p of function parameters.
16951 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
16953 Update copyright years.
16955 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
16957 PR rtl-optimization/55672
16958 * lra-eliminations.c (mark_not_eliminable): Permit addition with
16959 const to be eliminable.
16961 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
16963 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
16964 * configure: Regenerate.
16966 2013-01-10 Richard Biener <rguenther@suse.de>
16968 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
16970 2013-01-10 Richard Biener <rguenther@suse.de>
16973 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
16974 locations for virtual PHI arguments.
16975 (rewrite_update_phi_arguments): Likewise.
16977 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
16979 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
16982 2013-01-10 Jakub Jelinek <jakub@redhat.com>
16984 PR tree-optimization/55921
16985 * tree-complex.c (expand_complex_asm): New function.
16986 (expand_complex_operations_1): Call it for GIMPLE_ASM.
16988 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16991 * config/s390/s390.c (s390_symref_operand_p)
16992 (s390_loadrelative_operand_p): Merge the two functions.
16993 (s390_check_qrst_address, print_operand_address): Add parameters
16994 to s390_loadrelative_operand_p invokation.
16995 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
16996 (s390_reload_larl_operand, s390_secondary_reload): Use
16997 s390_loadrelative_operand_p instead of s390_symref_operand_p.
16998 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
17000 2013-01-09 Mike Stump <mikestump@comcast.net>
17002 * dse.c (record_store): Remove unnecessary assert.
17004 2013-01-09 Jan Hubicka <jh@suse.cz>
17006 PR tree-optimization/55569
17007 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
17008 * cfgloop.h (scale_loop_profile): Likewise.
17010 2013-01-09 Jan Hubicka <jh@suse.cz>
17013 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
17015 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
17017 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
17019 PR middle-end/55114
17020 * expr.h (maybe_emit_group_store): Declare.
17021 * expr.c (maybe_emit_group_store): New function.
17022 * builtins.c (expand_builtin_int_roundingfn): Call it.
17023 (expand_builtin_int_roundingfn_2): Likewise.
17025 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
17027 PR rtl-optimization/55829
17028 * lra-constraints.c (match_reload): Add code for absent output.
17029 (curr_insn_transform): Add code for reloads of matched inputs
17032 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
17034 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
17035 attribute of movddup insn to DF.
17036 (*vec_interleave_lowv2df): Ditto.
17037 (vec_dupv2df): Ditto.
17039 2013-01-09 Jan Hubicka <jh@suse.cz>
17041 PR tree-optimiation/55875
17042 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
17043 EVERY_ITERATION parameter.
17044 (number_of_iterations_exit): Check if exit is executed every iteration.
17045 (idx_infer_loop_bounds): Similarly here.
17046 (n_of_executions_at_most): Simplify
17047 to only test for cases where statement is dominated by the
17048 particular bound; handle correctly the "postdominance" test.
17049 (scev_probably_wraps_p): Use max loop iterations info
17050 as a global bound first.
17052 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
17053 Nick Clifton <nickc@redhat.com>
17055 * config/v850/v850.md (cbranchsf4): New pattern.
17056 (cstoresf4): New pattern.
17057 (cbranchdf4): New pattern.
17058 (cstoredf4): New pattern.
17059 (movsicc): Disallow floating point comparisons.
17060 (cmpsf_le_insn): Fix order of operators.
17061 (cmpsf_lt_insn): Likewise.
17062 (cmpsf_eq_insn): Likewise.
17063 (cmpdf_le_insn): Likewise.
17064 (cmpdf_lt_insn): Likewise.
17065 (cmpdf_eq_insn): Likewise.
17066 (cmpsf_ge_insn): Use LE comparison.
17067 (cmpdf_ge_insn): Likewise.
17068 (cmpsf_gt_insn): Use LT comparison.
17069 (cmpdf_gt_insn): Likewise.
17070 (cmpsf_ne_insn): Delete pattern.
17071 (cmpdf_ne_insn): Delete pattern.
17072 * config/v850/v850.c (v850_gen_float_compare): Use
17073 gen_cmpdf_eq_insn for NE comparison.
17074 (v850_float_z_comparison_operator)
17075 (v850_float_nz_comparison_operator): Move from here ...
17076 * config/v850/predicates.md: ... to here. Move GT and GE
17077 comparisons into v850_float_z_comparison_operator.
17078 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
17080 (v850_float_nz_comparison_operator): Likewise.
17082 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17084 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
17085 with calls to gen_insvsi/gen_insvdi.
17087 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17089 * config/i386/i386.c (initial_ix86_tune_features): Set up
17090 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17092 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17093 Jakub Jelinek <jakub@redhat.com>
17095 PR tree-optimization/48189
17096 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17097 If nitercst is 0, don't predict the exit edge.
17099 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17101 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17102 in asm_fprintf with reg_names.
17103 (aarch64_print_operand_address): Likewise.
17104 (aarch64_return_addr): Likewise.
17105 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17107 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17109 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17110 (INT_U6_BITS): Likewise.
17111 * config/pa/predicates.md (uint6_operand): New predicate.
17112 (shift5_operand, shift6_operand): Likewise.
17113 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17115 (lshrdi3): Use shift6_operand.
17116 (shrpsi4, shrpdi4): New insn patterns.
17117 (extzv): Delete expander.
17118 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17119 predicates in unamed zero extract patterns. Tighten common constraint.
17120 (extv): Delete expander.
17121 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17122 predicates in unamed sign extract patterns. Tighten common constraint.
17123 (insv): Delete expander.
17124 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17125 predicates in unamed insert patterns. Tighten common constraint.
17126 Change uint32_operand predicate to uint6_operand predicate in unamed
17127 DImode pattern to insert constant values of type 1...1xxxx.
17129 2013-01-04 Jan Hubicka <jh@suse.cz>
17131 PR tree-optimization/55823
17132 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17135 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17136 Uros Bizjak <ubizjak@gmail.com>
17138 PR rtl-optimization/55845
17139 * df-problems.c (can_move_insns_across): Stop scanning at
17140 volatile_insn_p source instruction or give up if
17141 across_from .. across_to range contains any volatile_insn_p
17144 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17146 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17147 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17149 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
17150 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
17152 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17155 * asan.c (asan_clear_shadow): New function.
17156 (asan_emit_stack_protection): Use it.
17158 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17160 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17161 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17162 with tab instead of space.
17164 2013-01-08 Nick Clifton <nickc@redhat.com>
17166 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17167 register bank 0 at the start of an interrupt handler.
17168 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17171 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17173 * config/aarch64/aarch64-simd.md
17174 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17175 (aarch64_simd_bsl): Likewise.
17176 (aarch64_vcond_internal<mode>): Likewise.
17177 (vcond<mode><mode>): Likewise.
17178 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17179 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17181 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17183 * config/aarch64/aarch64-builtins.c
17184 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17186 2013-01-08 Martin Jambor <mjambor@suse.cz>
17189 * tree-sra.c (analyze_access_subtree): Return true also after
17190 potentially creating a debug-only replacement.
17192 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17194 PR middle-end/55890
17195 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17197 PR tree-optimization/54120
17198 * tree-vrp.c (range_fits_type_p): Don't allow
17199 src_precision < precision from signed vr to unsigned_p
17200 if vr->min or vr->max is negative.
17201 (simplify_float_conversion_using_ranges): Test can_float_p
17202 against CODE_FOR_nothing.
17204 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17205 Richard Biener <rguenther@suse.de>
17207 PR middle-end/55851
17208 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17209 types instead of just INTEGER_TYPE types.
17211 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
17213 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17216 2013-01-07 Steve Ellcey <sellcey@mips.com>
17219 * config/mips/mips.opt: Change mad to mmad to match documentation.
17221 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17224 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17225 .progmemx.data now.
17227 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17230 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17231 (avr_addrspace_t): Add .section_name field.
17232 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17234 (avr_addrspace): Same. Initialize .section_name. Remove last
17235 NULL entry. Put __memx into .progmemx.data.
17236 (progmem_section_prefix): Remove.
17237 (avr_asm_init_sections): No need to initialize progmem_section.
17238 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17239 section name prefix.
17240 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17241 retrieve the progmem section.
17242 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17243 boundary to run over avr_addrspace[].
17244 (avr_register_target_pragmas): Ditto.
17246 2013-01-06 Jakub Jelinek <jakub@redhat.com>
17248 * varasm.c (output_constant_def_contents): For asan_protect_global
17249 protected strings, adjust DECL_ALIGN if needed, before testing for
17251 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17252 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17254 (output_object_block): For asan protected decls, emit asan padding
17255 after their contents.
17256 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17257 (asan_finish_file): Test it here instead.
17259 2013-01-07 Nick Clifton <nickc@redhat.com>
17260 Matthias Klose <doko@debian.org>
17261 Doug Kwan <dougkwan@google.com>
17262 H.J. Lu <hongjiu.lu@intel.com>
17265 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17267 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17269 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17271 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
17273 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17275 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17278 * doc/install.texi (Cross-Compiler-Specific Options): Document
17281 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17283 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17284 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17285 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17286 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17287 vqmovun_high_s64): Fix source operand number and update copyright.
17289 2013-01-07 Richard Biener <rguenther@suse.de>
17291 PR middle-end/55890
17292 * gimple.h (gimple_call_builtin_p): New overload.
17293 * gimple.c (validate_call): New function.
17294 (gimple_call_builtin_p): Likewise.
17295 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17296 Use gimple_call_builtin_p.
17297 (find_func_clobbers): Likewise.
17298 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17299 (strlen_optimize_stmt): Likewise.
17301 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17303 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17304 (vld1q_dup_*): Likewise.
17305 (vld1_*): Likewise.
17306 (vld1q_*): Likewise.
17307 (vld1_lane_*): Likewise.
17308 (vld1q_lane_*): Likewise.
17310 2013-01-07 Richard Biener <rguenther@suse.de>
17312 * lto-streamer.h (LTO_minor_version): Bump to 2.
17314 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17316 * config/aarch64/aarch64-protos.h
17317 (aarch64_const_double_zero_rtx_p): Rename to...
17318 (aarch64_float_const_zero_rtx_p): ...this.
17319 (aarch64_float_const_representable_p): New.
17320 (aarch64_output_simd_mov_immediate): Likewise.
17321 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17322 move immediate case.
17323 * config/aarch64/aarch64.c
17324 (aarch64_const_double_zero_rtx_p): Rename to...
17325 (aarch64_float_const_zero_rtx_p): ...this.
17326 (aarch64_print_operand): Allow printing of new constants.
17327 (aarch64_valid_floating_const): New.
17328 (aarch64_legitimate_constant_p): Check for valid floating-point
17330 (aarch64_simd_valid_immediate): Likewise.
17331 (aarch64_vect_float_const_representable_p): New.
17332 (aarch64_float_const_representable_p): Likewise.
17333 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17334 (aarch64_output_simd_mov_immediate): New.
17335 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17336 (*movdf_aarch64): Likewise.
17337 * config/aarch64/constraints.md (Ufc): New.
17338 (Y): call aarch64_float_const_zero_rtx.
17339 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17341 2013-01-07 Richard Biener <rguenther@suse.de>
17343 PR tree-optimization/55888
17344 PR tree-optimization/55862
17345 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17346 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17347 not if it is contained therein.
17349 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17351 * config/avr/t-avr: Typo.
17353 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17356 * config/avr/t-avr: Don't automatically rebuild
17357 $(srcdir)/config/avr/t-multilib
17358 $(srcdir)/config/avr/avr-tables.opt
17359 $(srcdir)/doc/avr-mmcu.texi
17360 (avr-mcus): New phony target to build them on request.
17361 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17362 * avr/avr-mcus.def: Adjust comments.
17364 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
17366 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17368 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17370 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17372 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17374 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17376 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
17378 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17379 to generate profiling.
17380 * config/rs6000/aix64.h (LIB_SPEC): Same.
17382 2013-01-04 Andrew Pinski <apinski@cavium.com>
17384 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17386 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17388 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
17390 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17392 (ix86_expand_move): Ditto.
17393 (ix86_zero_extend_to_Pmode): Ditto.
17394 (ix86_expand_call): Ditto.
17395 (ix86_expand_special_args_builtin): Ditto.
17396 (ix86_expand_builtin): Ditto.
17398 2013-01-04 Richard Biener <rguenther@suse.de>
17400 PR tree-optimization/55862
17401 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17402 translating them through PHI nodes.
17404 2013-01-04 Martin Jambor <mjambor@suse.cz>
17406 PR tree-optimization/55755
17407 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17408 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17410 2013-01-04 Marek Polacek <polacek@redhat.com>
17412 PR middle-end/55859
17413 * opts.c (default_options_optimization): Clarify error message.
17415 2013-01-04 Richard Biener <rguenther@suse.de>
17417 PR middle-end/55863
17418 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17421 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17424 * config/pa/pa.md (movsi): Revert previous change.
17425 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17428 2013-01-03 Richard Henderson <rth@redhat.com>
17430 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17431 after eliminating TLS symbols.
17433 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
17436 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
17437 * graphite-poly.c (debug_gmp_value): Likewise.
17439 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
17442 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
17443 selected code model, define __code_mode_small__, __code_model_medium__,
17444 __code_model_large__, __code_model_32__ or __code_model_kernel__.
17445 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
17446 xchg temporary register with %k. Declare temporary register as
17448 [__x86_64__]: For medium and large code models, preserve %rbx register.
17450 2013-01-03 Richard Biener <rguenther@suse.de>
17452 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
17453 (dump_subscript): Adjust.
17454 (finalize_ddr_dependent): Do not dump redundant info.
17455 (analyze_siv_subscript): Adjust.
17456 (subscript_dependence_tester): Likewise.
17457 (compute_affine_dependence): Likewise.
17459 2013-01-03 Richard Biener <rguenther@suse.de>
17462 2013-01-03 Richard Biener <rguenther@suse.de>
17464 PR tree-optimization/55857
17465 * tree-vect-stmts.c (vectorizable_load): Do not setup
17466 re-alignment for invariant loads.
17468 2013-01-02 Richard Biener <rguenther@suse.de>
17470 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17471 invariant load do not generate a vector load from the scalar location.
17473 2013-01-03 Richard Biener <rguenther@suse.de>
17475 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
17476 for not vectorizing.
17477 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
17478 not build INDIRECT_REFs, call get_name once only.
17479 (vect_create_data_ref_ptr): Likewise. Dump base object kind
17480 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
17482 2013-01-03 Richard Biener <rguenther@suse.de>
17484 PR tree-optimization/55857
17485 * tree-vect-stmts.c (vectorizable_load): Do not setup
17486 re-alignment for invariant loads.
17488 2013-01-03 Richard Biener <rguenther@suse.de>
17491 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
17492 prefer a built-in decl.
17494 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17496 * gcc.c (process_command): Update copyright notice dates.
17497 * gcov.c (print_version): Likewise.
17498 * gcov-dump.c (print_version): Likewise.
17500 PR rtl-optimization/55838
17501 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
17502 iv0.step, iv1.step and step.
17504 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17505 Marc Glisse <marc.glisse@inria.fr>
17507 PR tree-optimization/55832
17508 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
17509 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
17510 integer_{one,zero}_node.
17512 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17515 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
17516 * var-tracking.c (reverse_op): Don't add reverse ops to
17517 VALUEs that have already
17518 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
17520 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
17522 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
17524 2013-01-02 Teresa Johnson <tejohnson@google.com>
17526 * dumpfile.c (dump_loc): Print filename with location.
17527 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
17528 new location_t parameter to emit complete unroll message with
17529 new dump framework.
17530 (canonicalize_loop_induction_variables): Compute loops location
17531 and pass to try_unroll_loop_completely.
17532 * loop-unroll.c (report_unroll_peel): New function.
17533 (peel_loops_completely): Use new dump format with location
17534 for main dumpfile message, and invoke report_unroll_peel on success.
17535 (decide_unrolling_and_peeling): Ditto.
17536 (decide_peel_once_rolling): Remove old dumpfile message subsumed
17537 by report_unroll_peel.
17538 (decide_peel_completely): Ditto.
17539 (decide_unroll_constant_iterations): Ditto.
17540 (decide_unroll_runtime_iterations): Ditto.
17541 (decide_peel_simple): Ditto.
17542 (decide_unroll_stupid): Ditto.
17543 * cfgloop.c (get_loop_location): New function.
17544 * cfgloop.h (get_loop_location): Declare.
17546 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17548 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
17551 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17553 PR middle-end/55198
17554 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
17555 BLKmode objects when EXPAND_MEMORY is specified.
17557 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17559 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
17561 (fold_builtin_cpu): Do not share cpu model decls across statements.
17563 2013-01-02 Jason Merrill <jason@redhat.com>
17566 * tree.c (build_array_type_1): Revert earlier change.
17568 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
17570 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
17572 * config/aarch64/aarch64-tune.md: Re-generate.
17574 2013-01-02 Richard Biener <rguenther@suse.de>
17576 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17577 invariant load do not generate a vector load from the scalar location.
17579 2013-01-02 Richard Biener <rguenther@suse.de>
17582 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
17583 * configure: Regenerate.
17585 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17587 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
17588 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
17589 (expand_builtin_int_roundingfn_2): Keep the original target around
17590 for the fallback case.
17592 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17594 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
17595 to be clear for sign changes.
17597 2013-01-01 Jan Hubicka <jh@suse.cz>
17599 * ipa-inline-analysis.c: Fix formatting.
17601 2013-01-01 Jakub Jelinek <jakub@redhat.com>
17603 PR tree-optimization/55831
17604 * tree-vect-loop.c (get_initial_def_for_induction): Use
17605 gsi_after_labels instead of gsi_start_bb.
17607 Copyright (C) 2013 Free Software Foundation, Inc.
17609 Copying and distribution of this file, with or without modification,
17610 are permitted in any medium without royalty provided the copyright
17611 notice and this notice are preserved.