1 2013-08-06 Jan Hubicka <jh@suse.cz>
3 * cgraph.c (cgraph_get_body): New function based on lto.c
5 * cgraph.h (cgraph_get_body): Declare.
6 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and LTO paths.
7 * cgraphunit.c (expand_function): Get body prior expanding.
8 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't really need.
10 * passes.c (do_per_function_toporder): Get body.
11 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
14 2013-08-06 Martin Jambor <mjambor@suse.cz>
17 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
18 is not re-finalized. Rename second parameter to no_collect.
20 2013-08-06 Martin Jambor <mjambor@suse.cz>
23 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
24 MEM_REF has proper alignment information.
26 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
29 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
31 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
32 rtx (*) (rtx, ...) with insn_gen_fn.
33 * genoutput.c (output_insn_data): Cast gen_? function pointers to
34 insn_gen_fn::stored_funcptr. Add initializer braces.
36 2013-08-05 David Malcolm <dmalcolm@redhat.com>
38 Rewrite how instances of passes are cloned to remove assumptions
39 about their sizes (thus allowing pass subclasses to have
40 additional data fields, albeit non-GC-managed ones at this point).
42 * passes.c (make_pass_instance): Now that passes have clone
43 methods, rewrite this function to eliminate XNEW and memcpy
44 calls that used hardcoded sizes. Since this function no longer
45 creates pass instances, rename it to...
46 (add_pass_instance): ...this. Document the old way that passes
47 were numbered and flagged, and rework this function to continue
49 (next_pass_1): Add an initial_pass argument for use by
51 (position_pass): When adding multiple instances of a pass, use
52 the pass's clone method, rather than relying on the XNEW/memcpy
53 within the former make_pass_instance (now add_pass_instance).
54 (pass_manager::pass_manager): When invoking next_pass_1, also
55 supply the initial instance of the current pass within the
58 2013-08-05 David Malcolm <dmalcolm@redhat.com>
60 This is the automated part of the conversion of passes from C
61 structs to C++ classes.
63 Patch autogenerated by refactor_passes.py from
64 https://github.com/davidmalcolm/gcc-refactoring-scripts
65 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
67 * asan.c (pass_asan): Convert from a global struct to a subclass of
68 gimple_opt_pass along with...
69 (pass_data_asan): ...new pass_data instance and...
70 (make_pass_asan): ...new function.
71 (pass_asan_O0): Convert from a global struct to a subclass of
72 gimple_opt_pass along with...
73 (pass_data_asan_O0): ...new pass_data instance and...
74 (make_pass_asan_O0): ...new function.
75 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
76 subclass of rtl_opt_pass along with...
77 (pass_data_inc_dec): ...new pass_data instance and...
78 (make_pass_inc_dec): ...new function.
79 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
80 a subclass of rtl_opt_pass along with...
81 (pass_data_reorder_blocks): ...new pass_data instance and...
82 (make_pass_reorder_blocks): ...new function.
83 (pass_duplicate_computed_gotos): Convert from a global struct to a
84 subclass of rtl_opt_pass along with...
85 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
86 (make_pass_duplicate_computed_gotos): ...new function.
87 (pass_partition_blocks): Convert from a global struct to a subclass of
88 rtl_opt_pass along with...
89 (pass_data_partition_blocks): ...new pass_data instance and...
90 (make_pass_partition_blocks): ...new function.
91 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
92 struct to a subclass of rtl_opt_pass along with...
93 (pass_data_branch_target_load_optimize1): ...new pass_data instance
95 (make_pass_branch_target_load_optimize1): ...new function.
96 (pass_branch_target_load_optimize2): Convert from a global struct to a
97 subclass of rtl_opt_pass along with...
98 (pass_data_branch_target_load_optimize2): ...new pass_data instance
100 (make_pass_branch_target_load_optimize2): ...new function.
101 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
102 of rtl_opt_pass along with...
103 (pass_data_jump): ...new pass_data instance and...
104 (make_pass_jump): ...new function.
105 (pass_jump2): Convert from a global struct to a subclass of
106 rtl_opt_pass along with...
107 (pass_data_jump2): ...new pass_data instance and...
108 (make_pass_jump2): ...new function.
109 * cfgexpand.c (pass_expand): Convert from a global struct to a
110 subclass of rtl_opt_pass along with...
111 (pass_data_expand): ...new pass_data instance and...
112 (make_pass_expand): ...new function.
113 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
114 of rtl_opt_pass along with...
115 (pass_data_free_cfg): ...new pass_data instance and...
116 (make_pass_free_cfg): ...new function.
117 (pass_into_cfg_layout_mode): Convert from a global struct to a
118 subclass of rtl_opt_pass along with...
119 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
120 (make_pass_into_cfg_layout_mode): ...new function.
121 (pass_outof_cfg_layout_mode): Convert from a global struct to a
122 subclass of rtl_opt_pass along with...
123 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
124 (make_pass_outof_cfg_layout_mode): ...new function.
125 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
126 struct to a subclass of gimple_opt_pass along with...
127 (pass_data_build_cgraph_edges): ...new pass_data instance and...
128 (make_pass_build_cgraph_edges): ...new function.
129 (pass_rebuild_cgraph_edges): Convert from a global struct to a
130 subclass of gimple_opt_pass along with...
131 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
132 (make_pass_rebuild_cgraph_edges): ...new function.
133 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
134 subclass of gimple_opt_pass along with...
135 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
137 (make_pass_remove_cgraph_callee_edges): ...new function.
138 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
139 struct to a subclass of rtl_opt_pass along with...
140 (pass_data_stack_adjustments): ...new pass_data instance and...
141 (make_pass_stack_adjustments): ...new function.
142 * combine.c (pass_combine): Convert from a global struct to a subclass
143 of rtl_opt_pass along with...
144 (pass_data_combine): ...new pass_data instance and...
145 (make_pass_combine): ...new function.
146 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
147 global struct to a subclass of rtl_opt_pass along with...
148 (pass_data_compare_elim_after_reload): ...new pass_data instance
150 (make_pass_compare_elim_after_reload): ...new function.
151 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
152 of rtl_opt_pass along with...
153 (pass_data_rtl_cprop): ...new pass_data instance and...
154 (make_pass_rtl_cprop): ...new function.
155 * cse.c (pass_cse): Convert from a global struct to a subclass of
156 rtl_opt_pass along with...
157 (pass_data_cse): ...new pass_data instance and...
158 (make_pass_cse): ...new function.
159 (pass_cse2): Convert from a global struct to a subclass of
160 rtl_opt_pass along with...
161 (pass_data_cse2): ...new pass_data instance and...
162 (make_pass_cse2): ...new function.
163 (pass_cse_after_global_opts): Convert from a global struct to a
164 subclass of rtl_opt_pass along with...
165 (pass_data_cse_after_global_opts): ...new pass_data instance and...
166 (make_pass_cse_after_global_opts): ...new function.
167 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
168 of rtl_opt_pass along with...
169 (pass_data_ud_rtl_dce): ...new pass_data instance and...
170 (make_pass_ud_rtl_dce): ...new function.
171 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
172 rtl_opt_pass along with...
173 (pass_data_fast_rtl_dce): ...new pass_data instance and...
174 (make_pass_fast_rtl_dce): ...new function.
175 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
176 a subclass of rtl_opt_pass along with...
177 (pass_data_df_initialize_opt): ...new pass_data instance and...
178 (make_pass_df_initialize_opt): ...new function.
179 (pass_df_initialize_no_opt): Convert from a global struct to a
180 subclass of rtl_opt_pass along with...
181 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
182 (make_pass_df_initialize_no_opt): ...new function.
183 (pass_df_finish): Convert from a global struct to a subclass of
184 rtl_opt_pass along with...
185 (pass_data_df_finish): ...new pass_data instance and...
186 (make_pass_df_finish): ...new function.
187 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
188 rtl_opt_pass along with...
189 (pass_data_rtl_dse1): ...new pass_data instance and...
190 (make_pass_rtl_dse1): ...new function.
191 (pass_rtl_dse2): Convert from a global struct to a subclass of
192 rtl_opt_pass along with...
193 (pass_data_rtl_dse2): ...new pass_data instance and...
194 (make_pass_rtl_dse2): ...new function.
195 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
196 subclass of rtl_opt_pass along with...
197 (pass_data_dwarf2_frame): ...new pass_data instance and...
198 (make_pass_dwarf2_frame): ...new function.
199 * except.c (pass_set_nothrow_function_flags): Convert from a global
200 struct to a subclass of rtl_opt_pass along with...
201 (pass_data_set_nothrow_function_flags): ...new pass_data instance
203 (make_pass_set_nothrow_function_flags): ...new function.
204 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
205 subclass of rtl_opt_pass along with...
206 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
208 (make_pass_convert_to_eh_region_ranges): ...new function.
209 * final.c (pass_compute_alignments): Convert from a global struct to a
210 subclass of rtl_opt_pass along with...
211 (pass_data_compute_alignments): ...new pass_data instance and...
212 (make_pass_compute_alignments): ...new function.
213 (pass_final): Convert from a global struct to a subclass of
214 rtl_opt_pass along with...
215 (pass_data_final): ...new pass_data instance and...
216 (make_pass_final): ...new function.
217 (pass_shorten_branches): Convert from a global struct to a subclass of
218 rtl_opt_pass along with...
219 (pass_data_shorten_branches): ...new pass_data instance and...
220 (make_pass_shorten_branches): ...new function.
221 (pass_clean_state): Convert from a global struct to a subclass of
222 rtl_opt_pass along with...
223 (pass_data_clean_state): ...new pass_data instance and...
224 (make_pass_clean_state): ...new function.
225 * function.c (pass_instantiate_virtual_regs): Convert from a global
226 struct to a subclass of rtl_opt_pass along with...
227 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
228 (make_pass_instantiate_virtual_regs): ...new function.
229 (pass_leaf_regs): Convert from a global struct to a subclass of
230 rtl_opt_pass along with...
231 (pass_data_leaf_regs): ...new pass_data instance and...
232 (make_pass_leaf_regs): ...new function.
233 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
234 subclass of rtl_opt_pass along with...
235 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
237 (make_pass_thread_prologue_and_epilogue): ...new function.
238 (pass_match_asm_constraints): Convert from a global struct to a
239 subclass of rtl_opt_pass along with...
240 (pass_data_match_asm_constraints): ...new pass_data instance and...
241 (make_pass_match_asm_constraints): ...new function.
242 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
243 subclass of rtl_opt_pass along with...
244 (pass_data_rtl_fwprop): ...new pass_data instance and...
245 (make_pass_rtl_fwprop): ...new function.
246 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
247 rtl_opt_pass along with...
248 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
249 (make_pass_rtl_fwprop_addr): ...new function.
250 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
251 rtl_opt_pass along with...
252 (pass_data_rtl_pre): ...new pass_data instance and...
253 (make_pass_rtl_pre): ...new function.
254 (pass_rtl_hoist): Convert from a global struct to a subclass of
255 rtl_opt_pass along with...
256 (pass_data_rtl_hoist): ...new pass_data instance and...
257 (make_pass_rtl_hoist): ...new function.
258 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
259 subclass of gimple_opt_pass along with...
260 (pass_data_lower_cf): ...new pass_data instance and...
261 (make_pass_lower_cf): ...new function.
262 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
263 from a global struct to a subclass of gimple_opt_pass along with...
264 (pass_data_strength_reduction): ...new pass_data instance and...
265 (make_pass_strength_reduction): ...new function.
266 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
267 of rtl_opt_pass along with...
268 (pass_data_rtl_ifcvt): ...new pass_data instance and...
269 (make_pass_rtl_ifcvt): ...new function.
270 (pass_if_after_combine): Convert from a global struct to a subclass of
271 rtl_opt_pass along with...
272 (pass_data_if_after_combine): ...new pass_data instance and...
273 (make_pass_if_after_combine): ...new function.
274 (pass_if_after_reload): Convert from a global struct to a subclass of
275 rtl_opt_pass along with...
276 (pass_data_if_after_reload): ...new pass_data instance and...
277 (make_pass_if_after_reload): ...new function.
278 * init-regs.c (pass_initialize_regs): Convert from a global struct to
279 a subclass of rtl_opt_pass along with...
280 (pass_data_initialize_regs): ...new pass_data instance and...
281 (make_pass_initialize_regs): ...new function.
282 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
283 of ipa_opt_pass_d along with...
284 (pass_data_ipa_cp): ...new pass_data instance and...
285 (make_pass_ipa_cp): ...new function.
286 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
287 global struct to a subclass of gimple_opt_pass along with...
288 (pass_data_inline_parameters): ...new pass_data instance and...
289 (make_pass_inline_parameters): ...new function.
290 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
291 subclass of gimple_opt_pass along with...
292 (pass_data_early_inline): ...new pass_data instance and...
293 (make_pass_early_inline): ...new function.
294 (pass_ipa_inline): Convert from a global struct to a subclass of
295 ipa_opt_pass_d along with...
296 (pass_data_ipa_inline): ...new pass_data instance and...
297 (make_pass_ipa_inline): ...new function.
298 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
299 struct to a subclass of gimple_opt_pass along with...
300 (pass_data_local_pure_const): ...new pass_data instance and...
301 (make_pass_local_pure_const): ...new function.
302 (pass_ipa_pure_const): Convert from a global struct to a subclass of
303 ipa_opt_pass_d along with...
304 (pass_data_ipa_pure_const): ...new pass_data instance and...
305 (make_pass_ipa_pure_const): ...new function.
306 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
307 to a subclass of ipa_opt_pass_d along with...
308 (pass_data_ipa_reference): ...new pass_data instance and...
309 (make_pass_ipa_reference): ...new function.
310 * ipa-split.c (pass_split_functions): Convert from a global struct to
311 a subclass of gimple_opt_pass along with...
312 (pass_data_split_functions): ...new pass_data instance and...
313 (make_pass_split_functions): ...new function.
314 (pass_feedback_split_functions): Convert from a global struct to a
315 subclass of gimple_opt_pass along with...
316 (pass_data_feedback_split_functions): ...new pass_data instance and...
317 (make_pass_feedback_split_functions): ...new function.
318 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
319 global struct to a subclass of simple_ipa_opt_pass along with...
320 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
322 (make_pass_ipa_function_and_variable_visibility): ...new function.
323 (pass_ipa_free_inline_summary): Convert from a global struct to a
324 subclass of simple_ipa_opt_pass along with...
325 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
326 (make_pass_ipa_free_inline_summary): ...new function.
327 (pass_ipa_whole_program_visibility): Convert from a global struct to a
328 subclass of ipa_opt_pass_d along with...
329 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
331 (make_pass_ipa_whole_program_visibility): ...new function.
332 (pass_ipa_profile): Convert from a global struct to a subclass of
333 ipa_opt_pass_d along with...
334 (pass_data_ipa_profile): ...new pass_data instance and...
335 (make_pass_ipa_profile): ...new function.
336 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
337 ipa_opt_pass_d along with...
338 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
339 (make_pass_ipa_cdtor_merge): ...new function.
340 * ira.c (pass_ira): Convert from a global struct to a subclass of
341 rtl_opt_pass along with...
342 (pass_data_ira): ...new pass_data instance and...
343 (make_pass_ira): ...new function.
344 (pass_reload): Convert from a global struct to a subclass of
345 rtl_opt_pass along with...
346 (pass_data_reload): ...new pass_data instance and...
347 (make_pass_reload): ...new function.
348 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
349 subclass of rtl_opt_pass along with...
350 (pass_data_cleanup_barriers): ...new pass_data instance and...
351 (make_pass_cleanup_barriers): ...new function.
352 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
353 of rtl_opt_pass along with...
354 (pass_data_loop2): ...new pass_data instance and...
355 (make_pass_loop2): ...new function.
356 (pass_rtl_loop_init): Convert from a global struct to a subclass of
357 rtl_opt_pass along with...
358 (pass_data_rtl_loop_init): ...new pass_data instance and...
359 (make_pass_rtl_loop_init): ...new function.
360 (pass_rtl_loop_done): Convert from a global struct to a subclass of
361 rtl_opt_pass along with...
362 (pass_data_rtl_loop_done): ...new pass_data instance and...
363 (make_pass_rtl_loop_done): ...new function.
364 (pass_rtl_move_loop_invariants): Convert from a global struct to a
365 subclass of rtl_opt_pass along with...
366 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
367 (make_pass_rtl_move_loop_invariants): ...new function.
368 (pass_rtl_unswitch): Convert from a global struct to a subclass of
369 rtl_opt_pass along with...
370 (pass_data_rtl_unswitch): ...new pass_data instance and...
371 (make_pass_rtl_unswitch): ...new function.
372 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
373 subclass of rtl_opt_pass along with...
374 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
376 (make_pass_rtl_unroll_and_peel_loops): ...new function.
377 (pass_rtl_doloop): Convert from a global struct to a subclass of
378 rtl_opt_pass along with...
379 (pass_data_rtl_doloop): ...new pass_data instance and...
380 (make_pass_rtl_doloop): ...new function.
381 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
382 a subclass of rtl_opt_pass along with...
383 (pass_data_lower_subreg): ...new pass_data instance and...
384 (make_pass_lower_subreg): ...new function.
385 (pass_lower_subreg2): Convert from a global struct to a subclass of
386 rtl_opt_pass along with...
387 (pass_data_lower_subreg2): ...new pass_data instance and...
388 (make_pass_lower_subreg2): ...new function.
389 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
390 struct to a subclass of ipa_opt_pass_d along with...
391 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
392 (make_pass_ipa_lto_gimple_out): ...new function.
393 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
394 of ipa_opt_pass_d along with...
395 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
396 (make_pass_ipa_lto_finish_out): ...new function.
397 * mode-switching.c (pass_mode_switching): Convert from a global struct
398 to a subclass of rtl_opt_pass along with...
399 (pass_data_mode_switching): ...new pass_data instance and...
400 (make_pass_mode_switching): ...new function.
401 * modulo-sched.c (pass_sms): Convert from a global struct to a
402 subclass of rtl_opt_pass along with...
403 (pass_data_sms): ...new pass_data instance and...
404 (make_pass_sms): ...new function.
405 * omp-low.c (pass_expand_omp): Convert from a global struct to a
406 subclass of gimple_opt_pass along with...
407 (pass_data_expand_omp): ...new pass_data instance and...
408 (make_pass_expand_omp): ...new function.
409 (pass_lower_omp): Convert from a global struct to a subclass of
410 gimple_opt_pass along with...
411 (pass_data_lower_omp): ...new pass_data instance and...
412 (make_pass_lower_omp): ...new function.
413 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
414 of gimple_opt_pass along with...
415 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
416 (make_pass_diagnose_omp_blocks): ...new function.
417 * passes.c (pass_early_local_passes): Convert from a global struct to
418 a subclass of simple_ipa_opt_pass along with...
419 (pass_data_early_local_passes): ...new pass_data instance and...
420 (make_pass_early_local_passes): ...new function.
421 (pass_all_early_optimizations): Convert from a global struct to a
422 subclass of gimple_opt_pass along with...
423 (pass_data_all_early_optimizations): ...new pass_data instance and...
424 (make_pass_all_early_optimizations): ...new function.
425 (pass_all_optimizations): Convert from a global struct to a subclass
426 of gimple_opt_pass along with...
427 (pass_data_all_optimizations): ...new pass_data instance and...
428 (make_pass_all_optimizations): ...new function.
429 (pass_all_optimizations_g): Convert from a global struct to a subclass
430 of gimple_opt_pass along with...
431 (pass_data_all_optimizations_g): ...new pass_data instance and...
432 (make_pass_all_optimizations_g): ...new function.
433 (pass_rest_of_compilation): Convert from a global struct to a subclass
434 of rtl_opt_pass along with...
435 (pass_data_rest_of_compilation): ...new pass_data instance and...
436 (make_pass_rest_of_compilation): ...new function.
437 (pass_postreload): Convert from a global struct to a subclass of
438 rtl_opt_pass along with...
439 (pass_data_postreload): ...new pass_data instance and...
440 (make_pass_postreload): ...new function.
441 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
442 subclass of rtl_opt_pass along with...
443 (pass_data_gcse2): ...new pass_data instance and...
444 (make_pass_gcse2): ...new function.
445 * postreload.c (pass_postreload_cse): Convert from a global struct to
446 a subclass of rtl_opt_pass along with...
447 (pass_data_postreload_cse): ...new pass_data instance and...
448 (make_pass_postreload_cse): ...new function.
449 * predict.c (pass_profile): Convert from a global struct to a subclass
450 of gimple_opt_pass along with...
451 (pass_data_profile): ...new pass_data instance and...
452 (make_pass_profile): ...new function.
453 (pass_strip_predict_hints): Convert from a global struct to a subclass
454 of gimple_opt_pass along with...
455 (pass_data_strip_predict_hints): ...new pass_data instance and...
456 (make_pass_strip_predict_hints): ...new function.
457 * recog.c (pass_peephole2): Convert from a global struct to a subclass
458 of rtl_opt_pass along with...
459 (pass_data_peephole2): ...new pass_data instance and...
460 (make_pass_peephole2): ...new function.
461 (pass_split_all_insns): Convert from a global struct to a subclass of
462 rtl_opt_pass along with...
463 (pass_data_split_all_insns): ...new pass_data instance and...
464 (make_pass_split_all_insns): ...new function.
465 (pass_split_after_reload): Convert from a global struct to a subclass
466 of rtl_opt_pass along with...
467 (pass_data_split_after_reload): ...new pass_data instance and...
468 (make_pass_split_after_reload): ...new function.
469 (pass_split_before_regstack): Convert from a global struct to a
470 subclass of rtl_opt_pass along with...
471 (pass_data_split_before_regstack): ...new pass_data instance and...
472 (make_pass_split_before_regstack): ...new function.
473 (pass_split_before_sched2): Convert from a global struct to a subclass
474 of rtl_opt_pass along with...
475 (pass_data_split_before_sched2): ...new pass_data instance and...
476 (make_pass_split_before_sched2): ...new function.
477 (pass_split_for_shorten_branches): Convert from a global struct to a
478 subclass of rtl_opt_pass along with...
479 (pass_data_split_for_shorten_branches): ...new pass_data instance
481 (make_pass_split_for_shorten_branches): ...new function.
482 * ree.c (pass_ree): Convert from a global struct to a subclass of
483 rtl_opt_pass along with...
484 (pass_data_ree): ...new pass_data instance and...
485 (make_pass_ree): ...new function.
486 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
487 subclass of rtl_opt_pass along with...
488 (pass_data_stack_regs): ...new pass_data instance and...
489 (make_pass_stack_regs): ...new function.
490 (pass_stack_regs_run): Convert from a global struct to a subclass of
491 rtl_opt_pass along with...
492 (pass_data_stack_regs_run): ...new pass_data instance and...
493 (make_pass_stack_regs_run): ...new function.
494 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
495 subclass of rtl_opt_pass along with...
496 (pass_data_cprop_hardreg): ...new pass_data instance and...
497 (make_pass_cprop_hardreg): ...new function.
498 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
499 subclass of rtl_opt_pass along with...
500 (pass_data_reginfo_init): ...new pass_data instance and...
501 (make_pass_reginfo_init): ...new function.
502 * regmove.c (pass_regmove): Convert from a global struct to a subclass
503 of rtl_opt_pass along with...
504 (pass_data_regmove): ...new pass_data instance and...
505 (make_pass_regmove): ...new function.
506 * regrename.c (pass_regrename): Convert from a global struct to a
507 subclass of rtl_opt_pass along with...
508 (pass_data_regrename): ...new pass_data instance and...
509 (make_pass_regrename): ...new function.
510 * reorg.c (pass_delay_slots): Convert from a global struct to a
511 subclass of rtl_opt_pass along with...
512 (pass_data_delay_slots): ...new pass_data instance and...
513 (make_pass_delay_slots): ...new function.
514 (pass_machine_reorg): Convert from a global struct to a subclass of
515 rtl_opt_pass along with...
516 (pass_data_machine_reorg): ...new pass_data instance and...
517 (make_pass_machine_reorg): ...new function.
518 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
519 of rtl_opt_pass along with...
520 (pass_data_sched): ...new pass_data instance and...
521 (make_pass_sched): ...new function.
522 (pass_sched2): Convert from a global struct to a subclass of
523 rtl_opt_pass along with...
524 (pass_data_sched2): ...new pass_data instance and...
525 (make_pass_sched2): ...new function.
526 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
527 to a subclass of rtl_opt_pass along with...
528 (pass_data_stack_ptr_mod): ...new pass_data instance and...
529 (make_pass_stack_ptr_mod): ...new function.
530 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
531 to a subclass of rtl_opt_pass along with...
532 (pass_data_rtl_store_motion): ...new pass_data instance and...
533 (make_pass_rtl_store_motion): ...new function.
534 * tracer.c (pass_tracer): Convert from a global struct to a subclass
535 of gimple_opt_pass along with...
536 (pass_data_tracer): ...new pass_data instance and...
537 (make_pass_tracer): ...new function.
538 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
539 to a subclass of gimple_opt_pass along with...
540 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
541 (make_pass_diagnose_tm_blocks): ...new function.
542 (pass_lower_tm): Convert from a global struct to a subclass of
543 gimple_opt_pass along with...
544 (pass_data_lower_tm): ...new pass_data instance and...
545 (make_pass_lower_tm): ...new function.
546 (pass_tm_init): Convert from a global struct to a subclass of
547 gimple_opt_pass along with...
548 (pass_data_tm_init): ...new pass_data instance and...
549 (make_pass_tm_init): ...new function.
550 (pass_tm_mark): Convert from a global struct to a subclass of
551 gimple_opt_pass along with...
552 (pass_data_tm_mark): ...new pass_data instance and...
553 (make_pass_tm_mark): ...new function.
554 (pass_tm_edges): Convert from a global struct to a subclass of
555 gimple_opt_pass along with...
556 (pass_data_tm_edges): ...new pass_data instance and...
557 (make_pass_tm_edges): ...new function.
558 (pass_tm_memopt): Convert from a global struct to a subclass of
559 gimple_opt_pass along with...
560 (pass_data_tm_memopt): ...new pass_data instance and...
561 (make_pass_tm_memopt): ...new function.
562 (pass_ipa_tm): Convert from a global struct to a subclass of
563 simple_ipa_opt_pass along with...
564 (pass_data_ipa_tm): ...new pass_data instance and...
565 (make_pass_ipa_tm): ...new function.
566 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
567 subclass of gimple_opt_pass along with...
568 (pass_data_call_cdce): ...new pass_data instance and...
569 (make_pass_call_cdce): ...new function.
570 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
571 subclass of gimple_opt_pass along with...
572 (pass_data_build_cfg): ...new pass_data instance and...
573 (make_pass_build_cfg): ...new function.
574 (pass_split_crit_edges): Convert from a global struct to a subclass of
575 gimple_opt_pass along with...
576 (pass_data_split_crit_edges): ...new pass_data instance and...
577 (make_pass_split_crit_edges): ...new function.
578 (pass_warn_function_return): Convert from a global struct to a
579 subclass of gimple_opt_pass along with...
580 (pass_data_warn_function_return): ...new pass_data instance and...
581 (make_pass_warn_function_return): ...new function.
582 (pass_warn_function_noreturn): Convert from a global struct to a
583 subclass of gimple_opt_pass along with...
584 (pass_data_warn_function_noreturn): ...new pass_data instance and...
585 (make_pass_warn_function_noreturn): ...new function.
586 (pass_warn_unused_result): Convert from a global struct to a subclass
587 of gimple_opt_pass along with...
588 (pass_data_warn_unused_result): ...new pass_data instance and...
589 (make_pass_warn_unused_result): ...new function.
590 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
591 a subclass of gimple_opt_pass along with...
592 (pass_data_merge_phi): ...new pass_data instance and...
593 (make_pass_merge_phi): ...new function.
594 * tree-complex.c (pass_lower_complex): Convert from a global struct to
595 a subclass of gimple_opt_pass along with...
596 (pass_data_lower_complex): ...new pass_data instance and...
597 (make_pass_lower_complex): ...new function.
598 (pass_lower_complex_O0): Convert from a global struct to a subclass of
599 gimple_opt_pass along with...
600 (pass_data_lower_complex_O0): ...new pass_data instance and...
601 (make_pass_lower_complex_O0): ...new function.
602 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
603 subclass of gimple_opt_pass along with...
604 (pass_data_lower_eh): ...new pass_data instance and...
605 (make_pass_lower_eh): ...new function.
606 (pass_refactor_eh): Convert from a global struct to a subclass of
607 gimple_opt_pass along with...
608 (pass_data_refactor_eh): ...new pass_data instance and...
609 (make_pass_refactor_eh): ...new function.
610 (pass_lower_resx): Convert from a global struct to a subclass of
611 gimple_opt_pass along with...
612 (pass_data_lower_resx): ...new pass_data instance and...
613 (make_pass_lower_resx): ...new function.
614 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
615 of gimple_opt_pass along with...
616 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
617 (make_pass_lower_eh_dispatch): ...new function.
618 (pass_cleanup_eh): Convert from a global struct to a subclass of
619 gimple_opt_pass along with...
620 (pass_data_cleanup_eh): ...new pass_data instance and...
621 (make_pass_cleanup_eh): ...new function.
622 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
623 to a subclass of simple_ipa_opt_pass along with...
624 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
625 (make_pass_ipa_lower_emutls): ...new function.
626 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
627 a subclass of gimple_opt_pass along with...
628 (pass_data_if_conversion): ...new pass_data instance and...
629 (make_pass_if_conversion): ...new function.
630 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
631 subclass of gimple_opt_pass along with...
632 (pass_data_build_ssa): ...new pass_data instance and...
633 (make_pass_build_ssa): ...new function.
634 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
635 global struct to a subclass of gimple_opt_pass along with...
636 (pass_data_loop_distribution): ...new pass_data instance and...
637 (make_pass_loop_distribution): ...new function.
638 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
639 subclass of gimple_opt_pass along with...
640 (pass_data_mudflap_1): ...new pass_data instance and...
641 (make_pass_mudflap_1): ...new function.
642 (pass_mudflap_2): Convert from a global struct to a subclass of
643 gimple_opt_pass along with...
644 (pass_data_mudflap_2): ...new pass_data instance and...
645 (make_pass_mudflap_2): ...new function.
646 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
647 subclass of gimple_opt_pass along with...
648 (pass_data_mudflap_1): ...new pass_data instance and...
649 (make_pass_mudflap_1): ...new function.
650 (pass_mudflap_2): Convert from a global struct to a subclass of
651 gimple_opt_pass along with...
652 (pass_data_mudflap_2): ...new pass_data instance and...
653 (make_pass_mudflap_2): ...new function.
654 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
655 gimple_opt_pass along with...
656 (pass_data_nrv): ...new pass_data instance and...
657 (make_pass_nrv): ...new function.
658 (pass_return_slot): Convert from a global struct to a subclass of
659 gimple_opt_pass along with...
660 (pass_data_return_slot): ...new pass_data instance and...
661 (make_pass_return_slot): ...new function.
662 * tree-object-size.c (pass_object_sizes): Convert from a global struct
663 to a subclass of gimple_opt_pass along with...
664 (pass_data_object_sizes): ...new pass_data instance and...
665 (make_pass_object_sizes): ...new function.
666 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
667 global struct to a subclass of gimple_opt_pass along with...
668 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
670 (make_pass_cleanup_cfg_post_optimizing): ...new function.
671 (pass_fixup_cfg): Convert from a global struct to a subclass of
672 gimple_opt_pass along with...
673 (pass_data_fixup_cfg): ...new pass_data instance and...
674 (make_pass_fixup_cfg): ...new function.
675 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
676 (make_pass_mudflap_1): ...new function.
677 (pass_mudflap_2): Replace declaration with that of...
678 (make_pass_mudflap_2): ...new function.
679 (pass_asan): Replace declaration with that of...
680 (make_pass_asan): ...new function.
681 (pass_asan_O0): Replace declaration with that of...
682 (make_pass_asan_O0): ...new function.
683 (pass_tsan): Replace declaration with that of...
684 (make_pass_tsan): ...new function.
685 (pass_tsan_O0): Replace declaration with that of...
686 (make_pass_tsan_O0): ...new function.
687 (pass_lower_cf): Replace declaration with that of...
688 (make_pass_lower_cf): ...new function.
689 (pass_refactor_eh): Replace declaration with that of...
690 (make_pass_refactor_eh): ...new function.
691 (pass_lower_eh): Replace declaration with that of...
692 (make_pass_lower_eh): ...new function.
693 (pass_lower_eh_dispatch): Replace declaration with that of...
694 (make_pass_lower_eh_dispatch): ...new function.
695 (pass_lower_resx): Replace declaration with that of...
696 (make_pass_lower_resx): ...new function.
697 (pass_build_cfg): Replace declaration with that of...
698 (make_pass_build_cfg): ...new function.
699 (pass_early_tree_profile): Replace declaration with that of...
700 (make_pass_early_tree_profile): ...new function.
701 (pass_cleanup_eh): Replace declaration with that of...
702 (make_pass_cleanup_eh): ...new function.
703 (pass_sra): Replace declaration with that of...
704 (make_pass_sra): ...new function.
705 (pass_sra_early): Replace declaration with that of...
706 (make_pass_sra_early): ...new function.
707 (pass_early_ipa_sra): Replace declaration with that of...
708 (make_pass_early_ipa_sra): ...new function.
709 (pass_tail_recursion): Replace declaration with that of...
710 (make_pass_tail_recursion): ...new function.
711 (pass_tail_calls): Replace declaration with that of...
712 (make_pass_tail_calls): ...new function.
713 (pass_tree_loop): Replace declaration with that of...
714 (make_pass_tree_loop): ...new function.
715 (pass_tree_loop_init): Replace declaration with that of...
716 (make_pass_tree_loop_init): ...new function.
717 (pass_lim): Replace declaration with that of...
718 (make_pass_lim): ...new function.
719 (pass_tree_unswitch): Replace declaration with that of...
720 (make_pass_tree_unswitch): ...new function.
721 (pass_predcom): Replace declaration with that of...
722 (make_pass_predcom): ...new function.
723 (pass_iv_canon): Replace declaration with that of...
724 (make_pass_iv_canon): ...new function.
725 (pass_scev_cprop): Replace declaration with that of...
726 (make_pass_scev_cprop): ...new function.
727 (pass_empty_loop): Replace declaration with that of...
728 (make_pass_empty_loop): ...new function.
729 (pass_record_bounds): Replace declaration with that of...
730 (make_pass_record_bounds): ...new function.
731 (pass_graphite): Replace declaration with that of...
732 (make_pass_graphite): ...new function.
733 (pass_graphite_transforms): Replace declaration with that of...
734 (make_pass_graphite_transforms): ...new function.
735 (pass_if_conversion): Replace declaration with that of...
736 (make_pass_if_conversion): ...new function.
737 (pass_loop_distribution): Replace declaration with that of...
738 (make_pass_loop_distribution): ...new function.
739 (pass_vectorize): Replace declaration with that of...
740 (make_pass_vectorize): ...new function.
741 (pass_slp_vectorize): Replace declaration with that of...
742 (make_pass_slp_vectorize): ...new function.
743 (pass_complete_unroll): Replace declaration with that of...
744 (make_pass_complete_unroll): ...new function.
745 (pass_complete_unrolli): Replace declaration with that of...
746 (make_pass_complete_unrolli): ...new function.
747 (pass_parallelize_loops): Replace declaration with that of...
748 (make_pass_parallelize_loops): ...new function.
749 (pass_loop_prefetch): Replace declaration with that of...
750 (make_pass_loop_prefetch): ...new function.
751 (pass_iv_optimize): Replace declaration with that of...
752 (make_pass_iv_optimize): ...new function.
753 (pass_tree_loop_done): Replace declaration with that of...
754 (make_pass_tree_loop_done): ...new function.
755 (pass_ch): Replace declaration with that of...
756 (make_pass_ch): ...new function.
757 (pass_ccp): Replace declaration with that of...
758 (make_pass_ccp): ...new function.
759 (pass_phi_only_cprop): Replace declaration with that of...
760 (make_pass_phi_only_cprop): ...new function.
761 (pass_build_ssa): Replace declaration with that of...
762 (make_pass_build_ssa): ...new function.
763 (pass_build_alias): Replace declaration with that of...
764 (make_pass_build_alias): ...new function.
765 (pass_build_ealias): Replace declaration with that of...
766 (make_pass_build_ealias): ...new function.
767 (pass_dominator): Replace declaration with that of...
768 (make_pass_dominator): ...new function.
769 (pass_dce): Replace declaration with that of...
770 (make_pass_dce): ...new function.
771 (pass_dce_loop): Replace declaration with that of...
772 (make_pass_dce_loop): ...new function.
773 (pass_cd_dce): Replace declaration with that of...
774 (make_pass_cd_dce): ...new function.
775 (pass_call_cdce): Replace declaration with that of...
776 (make_pass_call_cdce): ...new function.
777 (pass_merge_phi): Replace declaration with that of...
778 (make_pass_merge_phi): ...new function.
779 (pass_split_crit_edges): Replace declaration with that of...
780 (make_pass_split_crit_edges): ...new function.
781 (pass_pre): Replace declaration with that of...
782 (make_pass_pre): ...new function.
783 (pass_profile): Replace declaration with that of...
784 (make_pass_profile): ...new function.
785 (pass_strip_predict_hints): Replace declaration with that of...
786 (make_pass_strip_predict_hints): ...new function.
787 (pass_lower_complex_O0): Replace declaration with that of...
788 (make_pass_lower_complex_O0): ...new function.
789 (pass_lower_complex): Replace declaration with that of...
790 (make_pass_lower_complex): ...new function.
791 (pass_lower_vector): Replace declaration with that of...
792 (make_pass_lower_vector): ...new function.
793 (pass_lower_vector_ssa): Replace declaration with that of...
794 (make_pass_lower_vector_ssa): ...new function.
795 (pass_lower_omp): Replace declaration with that of...
796 (make_pass_lower_omp): ...new function.
797 (pass_diagnose_omp_blocks): Replace declaration with that of...
798 (make_pass_diagnose_omp_blocks): ...new function.
799 (pass_expand_omp): Replace declaration with that of...
800 (make_pass_expand_omp): ...new function.
801 (pass_expand_omp_ssa): Replace declaration with that of...
802 (make_pass_expand_omp_ssa): ...new function.
803 (pass_object_sizes): Replace declaration with that of...
804 (make_pass_object_sizes): ...new function.
805 (pass_strlen): Replace declaration with that of...
806 (make_pass_strlen): ...new function.
807 (pass_fold_builtins): Replace declaration with that of...
808 (make_pass_fold_builtins): ...new function.
809 (pass_stdarg): Replace declaration with that of...
810 (make_pass_stdarg): ...new function.
811 (pass_early_warn_uninitialized): Replace declaration with that of...
812 (make_pass_early_warn_uninitialized): ...new function.
813 (pass_late_warn_uninitialized): Replace declaration with that of...
814 (make_pass_late_warn_uninitialized): ...new function.
815 (pass_cse_reciprocals): Replace declaration with that of...
816 (make_pass_cse_reciprocals): ...new function.
817 (pass_cse_sincos): Replace declaration with that of...
818 (make_pass_cse_sincos): ...new function.
819 (pass_optimize_bswap): Replace declaration with that of...
820 (make_pass_optimize_bswap): ...new function.
821 (pass_optimize_widening_mul): Replace declaration with that of...
822 (make_pass_optimize_widening_mul): ...new function.
823 (pass_warn_function_return): Replace declaration with that of...
824 (make_pass_warn_function_return): ...new function.
825 (pass_warn_function_noreturn): Replace declaration with that of...
826 (make_pass_warn_function_noreturn): ...new function.
827 (pass_cselim): Replace declaration with that of...
828 (make_pass_cselim): ...new function.
829 (pass_phiopt): Replace declaration with that of...
830 (make_pass_phiopt): ...new function.
831 (pass_forwprop): Replace declaration with that of...
832 (make_pass_forwprop): ...new function.
833 (pass_phiprop): Replace declaration with that of...
834 (make_pass_phiprop): ...new function.
835 (pass_tree_ifcombine): Replace declaration with that of...
836 (make_pass_tree_ifcombine): ...new function.
837 (pass_dse): Replace declaration with that of...
838 (make_pass_dse): ...new function.
839 (pass_nrv): Replace declaration with that of...
840 (make_pass_nrv): ...new function.
841 (pass_rename_ssa_copies): Replace declaration with that of...
842 (make_pass_rename_ssa_copies): ...new function.
843 (pass_sink_code): Replace declaration with that of...
844 (make_pass_sink_code): ...new function.
845 (pass_fre): Replace declaration with that of...
846 (make_pass_fre): ...new function.
847 (pass_check_data_deps): Replace declaration with that of...
848 (make_pass_check_data_deps): ...new function.
849 (pass_copy_prop): Replace declaration with that of...
850 (make_pass_copy_prop): ...new function.
851 (pass_vrp): Replace declaration with that of...
852 (make_pass_vrp): ...new function.
853 (pass_uncprop): Replace declaration with that of...
854 (make_pass_uncprop): ...new function.
855 (pass_return_slot): Replace declaration with that of...
856 (make_pass_return_slot): ...new function.
857 (pass_reassoc): Replace declaration with that of...
858 (make_pass_reassoc): ...new function.
859 (pass_rebuild_cgraph_edges): Replace declaration with that of...
860 (make_pass_rebuild_cgraph_edges): ...new function.
861 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
862 (make_pass_remove_cgraph_callee_edges): ...new function.
863 (pass_build_cgraph_edges): Replace declaration with that of...
864 (make_pass_build_cgraph_edges): ...new function.
865 (pass_local_pure_const): Replace declaration with that of...
866 (make_pass_local_pure_const): ...new function.
867 (pass_tracer): Replace declaration with that of...
868 (make_pass_tracer): ...new function.
869 (pass_warn_unused_result): Replace declaration with that of...
870 (make_pass_warn_unused_result): ...new function.
871 (pass_diagnose_tm_blocks): Replace declaration with that of...
872 (make_pass_diagnose_tm_blocks): ...new function.
873 (pass_lower_tm): Replace declaration with that of...
874 (make_pass_lower_tm): ...new function.
875 (pass_tm_init): Replace declaration with that of...
876 (make_pass_tm_init): ...new function.
877 (pass_tm_mark): Replace declaration with that of...
878 (make_pass_tm_mark): ...new function.
879 (pass_tm_memopt): Replace declaration with that of...
880 (make_pass_tm_memopt): ...new function.
881 (pass_tm_edges): Replace declaration with that of...
882 (make_pass_tm_edges): ...new function.
883 (pass_split_functions): Replace declaration with that of...
884 (make_pass_split_functions): ...new function.
885 (pass_feedback_split_functions): Replace declaration with that of...
886 (make_pass_feedback_split_functions): ...new function.
887 (pass_strength_reduction): Replace declaration with that of...
888 (make_pass_strength_reduction): ...new function.
889 (pass_ipa_lower_emutls): Replace declaration with that of...
890 (make_pass_ipa_lower_emutls): ...new function.
891 (pass_ipa_function_and_variable_visibility): Replace declaration with
893 (make_pass_ipa_function_and_variable_visibility): ...new function.
894 (pass_ipa_tree_profile): Replace declaration with that of...
895 (make_pass_ipa_tree_profile): ...new function.
896 (pass_early_local_passes): Replace declaration with that of...
897 (make_pass_early_local_passes): ...new function.
898 (pass_ipa_whole_program_visibility): Replace declaration with that
900 (make_pass_ipa_whole_program_visibility): ...new function.
901 (pass_ipa_lto_gimple_out): Replace declaration with that of...
902 (make_pass_ipa_lto_gimple_out): ...new function.
903 (pass_ipa_increase_alignment): Replace declaration with that of...
904 (make_pass_ipa_increase_alignment): ...new function.
905 (pass_ipa_inline): Replace declaration with that of...
906 (make_pass_ipa_inline): ...new function.
907 (pass_ipa_free_lang_data): Replace declaration with that of...
908 (make_pass_ipa_free_lang_data): ...new function.
909 (pass_ipa_free_inline_summary): Replace declaration with that of...
910 (make_pass_ipa_free_inline_summary): ...new function.
911 (pass_ipa_cp): Replace declaration with that of...
912 (make_pass_ipa_cp): ...new function.
913 (pass_ipa_reference): Replace declaration with that of...
914 (make_pass_ipa_reference): ...new function.
915 (pass_ipa_pure_const): Replace declaration with that of...
916 (make_pass_ipa_pure_const): ...new function.
917 (pass_ipa_pta): Replace declaration with that of...
918 (make_pass_ipa_pta): ...new function.
919 (pass_ipa_lto_finish_out): Replace declaration with that of...
920 (make_pass_ipa_lto_finish_out): ...new function.
921 (pass_ipa_tm): Replace declaration with that of...
922 (make_pass_ipa_tm): ...new function.
923 (pass_ipa_profile): Replace declaration with that of...
924 (make_pass_ipa_profile): ...new function.
925 (pass_ipa_cdtor_merge): Replace declaration with that of...
926 (make_pass_ipa_cdtor_merge): ...new function.
927 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
929 (make_pass_cleanup_cfg_post_optimizing): ...new function.
930 (pass_init_datastructures): Replace declaration with that of...
931 (make_pass_init_datastructures): ...new function.
932 (pass_fixup_cfg): Replace declaration with that of...
933 (make_pass_fixup_cfg): ...new function.
934 (pass_expand): Replace declaration with that of...
935 (make_pass_expand): ...new function.
936 (pass_instantiate_virtual_regs): Replace declaration with that of...
937 (make_pass_instantiate_virtual_regs): ...new function.
938 (pass_rtl_fwprop): Replace declaration with that of...
939 (make_pass_rtl_fwprop): ...new function.
940 (pass_rtl_fwprop_addr): Replace declaration with that of...
941 (make_pass_rtl_fwprop_addr): ...new function.
942 (pass_jump): Replace declaration with that of...
943 (make_pass_jump): ...new function.
944 (pass_jump2): Replace declaration with that of...
945 (make_pass_jump2): ...new function.
946 (pass_lower_subreg): Replace declaration with that of...
947 (make_pass_lower_subreg): ...new function.
948 (pass_cse): Replace declaration with that of...
949 (make_pass_cse): ...new function.
950 (pass_fast_rtl_dce): Replace declaration with that of...
951 (make_pass_fast_rtl_dce): ...new function.
952 (pass_ud_rtl_dce): Replace declaration with that of...
953 (make_pass_ud_rtl_dce): ...new function.
954 (pass_rtl_dce): Replace declaration with that of...
955 (make_pass_rtl_dce): ...new function.
956 (pass_rtl_dse1): Replace declaration with that of...
957 (make_pass_rtl_dse1): ...new function.
958 (pass_rtl_dse2): Replace declaration with that of...
959 (make_pass_rtl_dse2): ...new function.
960 (pass_rtl_dse3): Replace declaration with that of...
961 (make_pass_rtl_dse3): ...new function.
962 (pass_rtl_cprop): Replace declaration with that of...
963 (make_pass_rtl_cprop): ...new function.
964 (pass_rtl_pre): Replace declaration with that of...
965 (make_pass_rtl_pre): ...new function.
966 (pass_rtl_hoist): Replace declaration with that of...
967 (make_pass_rtl_hoist): ...new function.
968 (pass_rtl_store_motion): Replace declaration with that of...
969 (make_pass_rtl_store_motion): ...new function.
970 (pass_cse_after_global_opts): Replace declaration with that of...
971 (make_pass_cse_after_global_opts): ...new function.
972 (pass_rtl_ifcvt): Replace declaration with that of...
973 (make_pass_rtl_ifcvt): ...new function.
974 (pass_into_cfg_layout_mode): Replace declaration with that of...
975 (make_pass_into_cfg_layout_mode): ...new function.
976 (pass_outof_cfg_layout_mode): Replace declaration with that of...
977 (make_pass_outof_cfg_layout_mode): ...new function.
978 (pass_loop2): Replace declaration with that of...
979 (make_pass_loop2): ...new function.
980 (pass_rtl_loop_init): Replace declaration with that of...
981 (make_pass_rtl_loop_init): ...new function.
982 (pass_rtl_move_loop_invariants): Replace declaration with that of...
983 (make_pass_rtl_move_loop_invariants): ...new function.
984 (pass_rtl_unswitch): Replace declaration with that of...
985 (make_pass_rtl_unswitch): ...new function.
986 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
987 (make_pass_rtl_unroll_and_peel_loops): ...new function.
988 (pass_rtl_doloop): Replace declaration with that of...
989 (make_pass_rtl_doloop): ...new function.
990 (pass_rtl_loop_done): Replace declaration with that of...
991 (make_pass_rtl_loop_done): ...new function.
992 (pass_web): Replace declaration with that of...
993 (make_pass_web): ...new function.
994 (pass_cse2): Replace declaration with that of...
995 (make_pass_cse2): ...new function.
996 (pass_df_initialize_opt): Replace declaration with that of...
997 (make_pass_df_initialize_opt): ...new function.
998 (pass_df_initialize_no_opt): Replace declaration with that of...
999 (make_pass_df_initialize_no_opt): ...new function.
1000 (pass_reginfo_init): Replace declaration with that of...
1001 (make_pass_reginfo_init): ...new function.
1002 (pass_inc_dec): Replace declaration with that of...
1003 (make_pass_inc_dec): ...new function.
1004 (pass_stack_ptr_mod): Replace declaration with that of...
1005 (make_pass_stack_ptr_mod): ...new function.
1006 (pass_initialize_regs): Replace declaration with that of...
1007 (make_pass_initialize_regs): ...new function.
1008 (pass_combine): Replace declaration with that of...
1009 (make_pass_combine): ...new function.
1010 (pass_if_after_combine): Replace declaration with that of...
1011 (make_pass_if_after_combine): ...new function.
1012 (pass_ree): Replace declaration with that of...
1013 (make_pass_ree): ...new function.
1014 (pass_partition_blocks): Replace declaration with that of...
1015 (make_pass_partition_blocks): ...new function.
1016 (pass_match_asm_constraints): Replace declaration with that of...
1017 (make_pass_match_asm_constraints): ...new function.
1018 (pass_regmove): Replace declaration with that of...
1019 (make_pass_regmove): ...new function.
1020 (pass_split_all_insns): Replace declaration with that of...
1021 (make_pass_split_all_insns): ...new function.
1022 (pass_fast_rtl_byte_dce): Replace declaration with that of...
1023 (make_pass_fast_rtl_byte_dce): ...new function.
1024 (pass_lower_subreg2): Replace declaration with that of...
1025 (make_pass_lower_subreg2): ...new function.
1026 (pass_mode_switching): Replace declaration with that of...
1027 (make_pass_mode_switching): ...new function.
1028 (pass_sms): Replace declaration with that of...
1029 (make_pass_sms): ...new function.
1030 (pass_sched): Replace declaration with that of...
1031 (make_pass_sched): ...new function.
1032 (pass_ira): Replace declaration with that of...
1033 (make_pass_ira): ...new function.
1034 (pass_reload): Replace declaration with that of...
1035 (make_pass_reload): ...new function.
1036 (pass_clean_state): Replace declaration with that of...
1037 (make_pass_clean_state): ...new function.
1038 (pass_branch_prob): Replace declaration with that of...
1039 (make_pass_branch_prob): ...new function.
1040 (pass_value_profile_transformations): Replace declaration with that
1042 (make_pass_value_profile_transformations): ...new function.
1043 (pass_postreload_cse): Replace declaration with that of...
1044 (make_pass_postreload_cse): ...new function.
1045 (pass_gcse2): Replace declaration with that of...
1046 (make_pass_gcse2): ...new function.
1047 (pass_split_after_reload): Replace declaration with that of...
1048 (make_pass_split_after_reload): ...new function.
1049 (pass_branch_target_load_optimize1): Replace declaration with that
1051 (make_pass_branch_target_load_optimize1): ...new function.
1052 (pass_thread_prologue_and_epilogue): Replace declaration with that
1054 (make_pass_thread_prologue_and_epilogue): ...new function.
1055 (pass_stack_adjustments): Replace declaration with that of...
1056 (make_pass_stack_adjustments): ...new function.
1057 (pass_peephole2): Replace declaration with that of...
1058 (make_pass_peephole2): ...new function.
1059 (pass_if_after_reload): Replace declaration with that of...
1060 (make_pass_if_after_reload): ...new function.
1061 (pass_regrename): Replace declaration with that of...
1062 (make_pass_regrename): ...new function.
1063 (pass_cprop_hardreg): Replace declaration with that of...
1064 (make_pass_cprop_hardreg): ...new function.
1065 (pass_reorder_blocks): Replace declaration with that of...
1066 (make_pass_reorder_blocks): ...new function.
1067 (pass_branch_target_load_optimize2): Replace declaration with that
1069 (make_pass_branch_target_load_optimize2): ...new function.
1070 (pass_leaf_regs): Replace declaration with that of...
1071 (make_pass_leaf_regs): ...new function.
1072 (pass_split_before_sched2): Replace declaration with that of...
1073 (make_pass_split_before_sched2): ...new function.
1074 (pass_compare_elim_after_reload): Replace declaration with that of...
1075 (make_pass_compare_elim_after_reload): ...new function.
1076 (pass_sched2): Replace declaration with that of...
1077 (make_pass_sched2): ...new function.
1078 (pass_stack_regs): Replace declaration with that of...
1079 (make_pass_stack_regs): ...new function.
1080 (pass_stack_regs_run): Replace declaration with that of...
1081 (make_pass_stack_regs_run): ...new function.
1082 (pass_df_finish): Replace declaration with that of...
1083 (make_pass_df_finish): ...new function.
1084 (pass_compute_alignments): Replace declaration with that of...
1085 (make_pass_compute_alignments): ...new function.
1086 (pass_duplicate_computed_gotos): Replace declaration with that of...
1087 (make_pass_duplicate_computed_gotos): ...new function.
1088 (pass_variable_tracking): Replace declaration with that of...
1089 (make_pass_variable_tracking): ...new function.
1090 (pass_free_cfg): Replace declaration with that of...
1091 (make_pass_free_cfg): ...new function.
1092 (pass_machine_reorg): Replace declaration with that of...
1093 (make_pass_machine_reorg): ...new function.
1094 (pass_cleanup_barriers): Replace declaration with that of...
1095 (make_pass_cleanup_barriers): ...new function.
1096 (pass_delay_slots): Replace declaration with that of...
1097 (make_pass_delay_slots): ...new function.
1098 (pass_split_for_shorten_branches): Replace declaration with that of...
1099 (make_pass_split_for_shorten_branches): ...new function.
1100 (pass_split_before_regstack): Replace declaration with that of...
1101 (make_pass_split_before_regstack): ...new function.
1102 (pass_convert_to_eh_region_ranges): Replace declaration with that
1104 (make_pass_convert_to_eh_region_ranges): ...new function.
1105 (pass_shorten_branches): Replace declaration with that of...
1106 (make_pass_shorten_branches): ...new function.
1107 (pass_set_nothrow_function_flags): Replace declaration with that of...
1108 (make_pass_set_nothrow_function_flags): ...new function.
1109 (pass_dwarf2_frame): Replace declaration with that of...
1110 (make_pass_dwarf2_frame): ...new function.
1111 (pass_final): Replace declaration with that of...
1112 (make_pass_final): ...new function.
1113 (pass_rtl_seqabstr): Replace declaration with that of...
1114 (make_pass_rtl_seqabstr): ...new function.
1115 (pass_release_ssa_names): Replace declaration with that of...
1116 (make_pass_release_ssa_names): ...new function.
1117 (pass_early_inline): Replace declaration with that of...
1118 (make_pass_early_inline): ...new function.
1119 (pass_inline_parameters): Replace declaration with that of...
1120 (make_pass_inline_parameters): ...new function.
1121 (pass_update_address_taken): Replace declaration with that of...
1122 (make_pass_update_address_taken): ...new function.
1123 (pass_convert_switch): Replace declaration with that of...
1124 (make_pass_convert_switch): ...new function.
1125 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1126 to a subclass of simple_ipa_opt_pass along with...
1127 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1128 (make_pass_ipa_tree_profile): ...new function.
1129 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1130 subclass of gimple_opt_pass along with...
1131 (pass_data_sra_early): ...new pass_data instance and...
1132 (make_pass_sra_early): ...new function.
1133 (pass_sra): Convert from a global struct to a subclass of
1134 gimple_opt_pass along with...
1135 (pass_data_sra): ...new pass_data instance and...
1136 (make_pass_sra): ...new function.
1137 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1138 gimple_opt_pass along with...
1139 (pass_data_early_ipa_sra): ...new pass_data instance and...
1140 (make_pass_early_ipa_sra): ...new function.
1141 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1142 subclass of gimple_opt_pass along with...
1143 (pass_data_ccp): ...new pass_data instance and...
1144 (make_pass_ccp): ...new function.
1145 (pass_fold_builtins): Convert from a global struct to a subclass of
1146 gimple_opt_pass along with...
1147 (pass_data_fold_builtins): ...new pass_data instance and...
1148 (make_pass_fold_builtins): ...new function.
1149 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1150 subclass of gimple_opt_pass along with...
1151 (pass_data_copy_prop): ...new pass_data instance and...
1152 (make_pass_copy_prop): ...new function.
1153 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1154 global struct to a subclass of gimple_opt_pass along with...
1155 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1156 (make_pass_rename_ssa_copies): ...new function.
1157 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1158 subclass of gimple_opt_pass along with...
1159 (pass_data_dce): ...new pass_data instance and...
1160 (make_pass_dce): ...new function.
1161 (pass_dce_loop): Convert from a global struct to a subclass of
1162 gimple_opt_pass along with...
1163 (pass_data_dce_loop): ...new pass_data instance and...
1164 (make_pass_dce_loop): ...new function.
1165 (pass_cd_dce): Convert from a global struct to a subclass of
1166 gimple_opt_pass along with...
1167 (pass_data_cd_dce): ...new pass_data instance and...
1168 (make_pass_cd_dce): ...new function.
1169 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1170 subclass of gimple_opt_pass along with...
1171 (pass_data_dominator): ...new pass_data instance and...
1172 (make_pass_dominator): ...new function.
1173 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1174 gimple_opt_pass along with...
1175 (pass_data_phi_only_cprop): ...new pass_data instance and...
1176 (make_pass_phi_only_cprop): ...new function.
1177 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1178 subclass of gimple_opt_pass along with...
1179 (pass_data_dse): ...new pass_data instance and...
1180 (make_pass_dse): ...new function.
1181 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1182 a subclass of gimple_opt_pass along with...
1183 (pass_data_forwprop): ...new pass_data instance and...
1184 (make_pass_forwprop): ...new function.
1185 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1186 struct to a subclass of gimple_opt_pass along with...
1187 (pass_data_tree_ifcombine): ...new pass_data instance and...
1188 (make_pass_tree_ifcombine): ...new function.
1189 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1190 subclass of gimple_opt_pass along with...
1191 (pass_data_ch): ...new pass_data instance and...
1192 (make_pass_ch): ...new function.
1193 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1194 subclass of gimple_opt_pass along with...
1195 (pass_data_tree_loop): ...new pass_data instance and...
1196 (make_pass_tree_loop): ...new function.
1197 (pass_tree_loop_init): Convert from a global struct to a subclass of
1198 gimple_opt_pass along with...
1199 (pass_data_tree_loop_init): ...new pass_data instance and...
1200 (make_pass_tree_loop_init): ...new function.
1201 (pass_lim): Convert from a global struct to a subclass of
1202 gimple_opt_pass along with...
1203 (pass_data_lim): ...new pass_data instance and...
1204 (make_pass_lim): ...new function.
1205 (pass_tree_unswitch): Convert from a global struct to a subclass of
1206 gimple_opt_pass along with...
1207 (pass_data_tree_unswitch): ...new pass_data instance and...
1208 (make_pass_tree_unswitch): ...new function.
1209 (pass_predcom): Convert from a global struct to a subclass of
1210 gimple_opt_pass along with...
1211 (pass_data_predcom): ...new pass_data instance and...
1212 (make_pass_predcom): ...new function.
1213 (pass_vectorize): Convert from a global struct to a subclass of
1214 gimple_opt_pass along with...
1215 (pass_data_vectorize): ...new pass_data instance and...
1216 (make_pass_vectorize): ...new function.
1217 (pass_graphite): Convert from a global struct to a subclass of
1218 gimple_opt_pass along with...
1219 (pass_data_graphite): ...new pass_data instance and...
1220 (make_pass_graphite): ...new function.
1221 (pass_graphite_transforms): Convert from a global struct to a subclass
1222 of gimple_opt_pass along with...
1223 (pass_data_graphite_transforms): ...new pass_data instance and...
1224 (make_pass_graphite_transforms): ...new function.
1225 (pass_check_data_deps): Convert from a global struct to a subclass of
1226 gimple_opt_pass along with...
1227 (pass_data_check_data_deps): ...new pass_data instance and...
1228 (make_pass_check_data_deps): ...new function.
1229 (pass_iv_canon): Convert from a global struct to a subclass of
1230 gimple_opt_pass along with...
1231 (pass_data_iv_canon): ...new pass_data instance and...
1232 (make_pass_iv_canon): ...new function.
1233 (pass_scev_cprop): Convert from a global struct to a subclass of
1234 gimple_opt_pass along with...
1235 (pass_data_scev_cprop): ...new pass_data instance and...
1236 (make_pass_scev_cprop): ...new function.
1237 (pass_record_bounds): Convert from a global struct to a subclass of
1238 gimple_opt_pass along with...
1239 (pass_data_record_bounds): ...new pass_data instance and...
1240 (make_pass_record_bounds): ...new function.
1241 (pass_complete_unroll): Convert from a global struct to a subclass of
1242 gimple_opt_pass along with...
1243 (pass_data_complete_unroll): ...new pass_data instance and...
1244 (make_pass_complete_unroll): ...new function.
1245 (pass_complete_unrolli): Convert from a global struct to a subclass of
1246 gimple_opt_pass along with...
1247 (pass_data_complete_unrolli): ...new pass_data instance and...
1248 (make_pass_complete_unrolli): ...new function.
1249 (pass_parallelize_loops): Convert from a global struct to a subclass
1250 of gimple_opt_pass along with...
1251 (pass_data_parallelize_loops): ...new pass_data instance and...
1252 (make_pass_parallelize_loops): ...new function.
1253 (pass_loop_prefetch): Convert from a global struct to a subclass of
1254 gimple_opt_pass along with...
1255 (pass_data_loop_prefetch): ...new pass_data instance and...
1256 (make_pass_loop_prefetch): ...new function.
1257 (pass_iv_optimize): Convert from a global struct to a subclass of
1258 gimple_opt_pass along with...
1259 (pass_data_iv_optimize): ...new pass_data instance and...
1260 (make_pass_iv_optimize): ...new function.
1261 (pass_tree_loop_done): Convert from a global struct to a subclass of
1262 gimple_opt_pass along with...
1263 (pass_data_tree_loop_done): ...new pass_data instance and...
1264 (make_pass_tree_loop_done): ...new function.
1265 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1266 struct to a subclass of gimple_opt_pass along with...
1267 (pass_data_cse_reciprocals): ...new pass_data instance and...
1268 (make_pass_cse_reciprocals): ...new function.
1269 (pass_cse_sincos): Convert from a global struct to a subclass of
1270 gimple_opt_pass along with...
1271 (pass_data_cse_sincos): ...new pass_data instance and...
1272 (make_pass_cse_sincos): ...new function.
1273 (pass_optimize_bswap): Convert from a global struct to a subclass of
1274 gimple_opt_pass along with...
1275 (pass_data_optimize_bswap): ...new pass_data instance and...
1276 (make_pass_optimize_bswap): ...new function.
1277 (pass_optimize_widening_mul): Convert from a global struct to a
1278 subclass of gimple_opt_pass along with...
1279 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1280 (make_pass_optimize_widening_mul): ...new function.
1281 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1282 subclass of gimple_opt_pass along with...
1283 (pass_data_phiopt): ...new pass_data instance and...
1284 (make_pass_phiopt): ...new function.
1285 (pass_cselim): Convert from a global struct to a subclass of
1286 gimple_opt_pass along with...
1287 (pass_data_cselim): ...new pass_data instance and...
1288 (make_pass_cselim): ...new function.
1289 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1290 subclass of gimple_opt_pass along with...
1291 (pass_data_phiprop): ...new pass_data instance and...
1292 (make_pass_phiprop): ...new function.
1293 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1294 subclass of gimple_opt_pass along with...
1295 (pass_data_pre): ...new pass_data instance and...
1296 (make_pass_pre): ...new function.
1297 (pass_fre): Convert from a global struct to a subclass of
1298 gimple_opt_pass along with...
1299 (pass_data_fre): ...new pass_data instance and...
1300 (make_pass_fre): ...new function.
1301 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1302 subclass of gimple_opt_pass along with...
1303 (pass_data_reassoc): ...new pass_data instance and...
1304 (make_pass_reassoc): ...new function.
1305 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1306 subclass of gimple_opt_pass along with...
1307 (pass_data_sink_code): ...new pass_data instance and...
1308 (make_pass_sink_code): ...new function.
1309 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1310 subclass of gimple_opt_pass along with...
1311 (pass_data_strlen): ...new pass_data instance and...
1312 (make_pass_strlen): ...new function.
1313 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1314 struct to a subclass of gimple_opt_pass along with...
1315 (pass_data_build_alias): ...new pass_data instance and...
1316 (make_pass_build_alias): ...new function.
1317 (pass_build_ealias): Convert from a global struct to a subclass of
1318 gimple_opt_pass along with...
1319 (pass_data_build_ealias): ...new pass_data instance and...
1320 (make_pass_build_ealias): ...new function.
1321 (pass_ipa_pta): Convert from a global struct to a subclass of
1322 simple_ipa_opt_pass along with...
1323 (pass_data_ipa_pta): ...new pass_data instance and...
1324 (make_pass_ipa_pta): ...new function.
1325 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1326 subclass of gimple_opt_pass along with...
1327 (pass_data_uncprop): ...new pass_data instance and...
1328 (make_pass_uncprop): ...new function.
1329 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1330 global struct to a subclass of gimple_opt_pass along with...
1331 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1332 (make_pass_late_warn_uninitialized): ...new function.
1333 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1334 to a subclass of gimple_opt_pass along with...
1335 (pass_data_init_datastructures): ...new pass_data instance and...
1336 (make_pass_init_datastructures): ...new function.
1337 (pass_early_warn_uninitialized): Convert from a global struct to a
1338 subclass of gimple_opt_pass along with...
1339 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1340 (make_pass_early_warn_uninitialized): ...new function.
1341 (pass_update_address_taken): Convert from a global struct to a
1342 subclass of gimple_opt_pass along with...
1343 (pass_data_update_address_taken): ...new pass_data instance and...
1344 (make_pass_update_address_taken): ...new function.
1345 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1346 struct to a subclass of gimple_opt_pass along with...
1347 (pass_data_release_ssa_names): ...new pass_data instance and...
1348 (make_pass_release_ssa_names): ...new function.
1349 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1350 subclass of gimple_opt_pass along with...
1351 (pass_data_stdarg): ...new pass_data instance and...
1352 (make_pass_stdarg): ...new function.
1353 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1354 global struct to a subclass of gimple_opt_pass along with...
1355 (pass_data_convert_switch): ...new pass_data instance and...
1356 (make_pass_convert_switch): ...new function.
1357 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1358 to a subclass of gimple_opt_pass along with...
1359 (pass_data_tail_recursion): ...new pass_data instance and...
1360 (make_pass_tail_recursion): ...new function.
1361 (pass_tail_calls): Convert from a global struct to a subclass of
1362 gimple_opt_pass along with...
1363 (pass_data_tail_calls): ...new pass_data instance and...
1364 (make_pass_tail_calls): ...new function.
1365 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1366 struct to a subclass of gimple_opt_pass along with...
1367 (pass_data_lower_vector): ...new pass_data instance and...
1368 (make_pass_lower_vector): ...new function.
1369 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1370 gimple_opt_pass along with...
1371 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1372 (make_pass_lower_vector_ssa): ...new function.
1373 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
1374 to a subclass of gimple_opt_pass along with...
1375 (pass_data_slp_vectorize): ...new pass_data instance and...
1376 (make_pass_slp_vectorize): ...new function.
1377 (pass_ipa_increase_alignment): Convert from a global struct to a
1378 subclass of simple_ipa_opt_pass along with...
1379 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
1380 (make_pass_ipa_increase_alignment): ...new function.
1381 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
1382 gimple_opt_pass along with...
1383 (pass_data_vrp): ...new pass_data instance and...
1384 (make_pass_vrp): ...new function.
1385 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
1386 subclass of simple_ipa_opt_pass along with...
1387 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
1388 (make_pass_ipa_free_lang_data): ...new function.
1389 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
1390 gimple_opt_pass along with...
1391 (pass_data_tsan): ...new pass_data instance and...
1392 (make_pass_tsan): ...new function.
1393 (pass_tsan_O0): Convert from a global struct to a subclass of
1394 gimple_opt_pass along with...
1395 (pass_data_tsan_O0): ...new pass_data instance and...
1396 (make_pass_tsan_O0): ...new function.
1397 * var-tracking.c (pass_variable_tracking): Convert from a global
1398 struct to a subclass of rtl_opt_pass along with...
1399 (pass_data_variable_tracking): ...new pass_data instance and...
1400 (make_pass_variable_tracking): ...new function.
1401 * web.c (pass_web): Convert from a global struct to a subclass of
1402 rtl_opt_pass along with...
1403 (pass_data_web): ...new pass_data instance and...
1404 (make_pass_web): ...new function.
1405 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
1406 declaration with that of...
1407 (make_pass_mode_switch_use): ...new function.
1408 (pass_resolve_sw_modes): Replace declaration with that of...
1409 (make_pass_resolve_sw_modes): ...new function.
1410 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
1411 from a global struct to a subclass of rtl_opt_pass along with...
1412 (pass_data_mode_switch_use): ...new pass_data instance and...
1413 (make_pass_mode_switch_use): ...new function.
1414 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
1415 from a global struct to a subclass of rtl_opt_pass along with...
1416 (pass_data_resolve_sw_modes): ...new pass_data instance and...
1417 (make_pass_resolve_sw_modes): ...new function.
1418 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
1419 struct to a subclass of rtl_opt_pass along with...
1420 (pass_data_insert_vzeroupper): ...new pass_data instance and...
1421 (make_pass_insert_vzeroupper): ...new function.
1422 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
1423 global struct to a subclass of rtl_opt_pass along with...
1424 (pass_data_work_around_errata): ...new pass_data instance and...
1425 (make_pass_work_around_errata): ...new function.
1426 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
1427 struct to a subclass of rtl_opt_pass along with...
1428 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
1429 (make_pass_mips_machine_reorg2): ...new function.
1431 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1433 * passes.c (pass_manager::operator new): New.
1435 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1437 Handwritten part of conversion of passes to C++ classes.
1439 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
1440 (toplev.o): Add dep on PASS_MANAGER_H.
1441 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
1442 of early local pases to reflect this moving from a global to a
1443 member of gcc::pass_manager.
1444 (cgraph_add_new_function): Likewise.
1445 * lto-cgraph.c (lto_output_node): Update for conversion of
1446 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
1447 * passes.c (opt_pass::clone): New.
1448 (opt_pass::gate): New.
1449 (opt_pass::execute): New.
1450 (opt_pass::opt_pass): New.
1451 (pass_manager::execute_early_local_passes): New.
1452 (pass_manager::execute_pass_mode_switching): new.
1453 (finish_optimization_passes): Convert to...
1454 (pass_manager::finish_optimization_passes): ...this.
1455 (finish_optimization_passes): Update for conversion of passes to
1457 (register_dump_files_1): Use has_gate since we cannot portably
1458 check a vtable entry against NULL.
1459 (dump_one_pass): Likewise.
1460 (ipa_write_summaries_2): Likewise.
1461 (ipa_write_optimization_summaries_1): Likewise.
1462 (ipa_read_summaries_1): Likewise.
1463 (ipa_read_optimization_summaries_1): Likewise.
1464 (execute_ipa_stmt_fixups): Likewise.
1465 (pass_manager::pass_manager): Rewrite pass-creation, invoking
1466 pass-creation functions rather than wiring up globals, and
1467 storing the results in fields of pass_manager generated using
1469 (pass_manager::dump_profile_report): Update for conversion of
1470 passes to C++ classes.
1471 (pass_manager::execute_ipa_summary_passes): Likewise.
1472 (execute_one_ipa_transform_pass): Likewise.
1473 (execute_one_pass): Use has_gate and has_execute since we cannot
1474 portably check a vtable entry against NULL.
1475 * pass_manager.h (pass_manager::finish_optimization_passes): New.
1476 (pass_manager): Use pass-instances.def to add fields for the
1477 various pass instances.
1478 * toplev.c (finalize): Update for move of
1479 finish_optimization_passes to a method of gcc::pass_manager.
1480 * toplev.h (finish_optimization_passes): Move to method of class
1482 * tree-pass.h (struct pass_data): New.
1483 (opt_pass): Convert to C++ class, make it a subclass of
1485 (opt_pass::gate): Convert to virtual function.
1486 (opt_pass::~opt_pass): New.
1487 (opt_pass::clone): New.
1488 (opt_pass::execute): Convert to virtual function.
1489 (opt_pass::opt_pass): New.
1490 (opt_pass::ctxt_): new.
1491 (gimple_opt_pass): Convert to subclass of opt_pass.
1492 (gimple_opt_pass::gimple_opt_pass): New.
1493 (rtl_opt_pass): Convert to subclass of opt_pass.
1494 (rtl_opt_pass::rtl_opt_pass): New.
1495 (ipa_opt_pass_d): Convert to subclass of opt_pass.
1496 (ipa_opt_pass_d::ipa_opt_pass_d): New.
1497 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
1498 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
1499 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
1500 invocation of pass_mode_switching to reflect this moving from a
1501 global to a member of gcc::pass_manager.
1502 (ix86_option_override): Rework how pass_insert_vzeroupper is
1503 added to the pass_manager to reflect autogenerated changes.
1504 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and
1507 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
1509 PR rtl-optimization/57708
1510 * recog.c (peep2_find_free_register): Validate all regs in a
1513 2013-08-05 Jan Hubicka <jh@suse.cz>
1516 * cgraph.c (verify_cgraph_node): Accept local flags from other partitions.
1517 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
1518 (function_and_variable_visibility): Likewise.
1519 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
1521 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1523 * graph.c (init_graph_slim_pretty_print): Remove.
1524 (print_graph_cfg): Do not call it. Use local pretty printer.
1525 (start_graph_dump): Likewise.
1527 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1529 * gimple-pretty-print.c (buffer): Remove.
1530 (initialized): Likewise.
1531 (maybe_init_pretty_print): Likewise.
1532 (print_gimple_stmt): Do not call it. Use non-static local
1533 pretty_printer variable.
1534 (print_gimple_expr): Likewise.
1535 (print_gimple_seq): Likewise.
1536 (gimple_dump_bb): Likewise.
1538 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1540 * asan.c (asan_pp): Remove.
1541 (asan_pp_initialized): Likewise.
1542 (asan_pp_initialize): Likewise.
1543 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
1544 (asan_emit_stack_protection): Tidy. Use local pretty printer.
1545 (asan_add_global): Likewise.
1547 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
1549 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
1550 * diagnostic.h (diagnostic_flush_buffer): Adjust.
1551 * pretty-print.c (pp_formatted_text_data): Likewise.
1552 (pp_indent): Rename from pp_base_indent.
1553 (pp_format): Rename from pp_base_format.
1554 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
1555 (pp_format_verbatim): Rename from pp_base_format_verbatim.
1556 (pp_flush): Rename from pp_base_flush.
1557 (pp_set_line_maximum_length): Rename from
1558 pp_base_set_line_maximum_length.
1559 (pp_clear_output_area): Rename from pp_base_clear_output_area.
1560 (pp_set_prefix): Rename from pp_base_set_prefix.
1561 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
1562 (pp_emit_prefix): Rename from pp_base_emit_prefix.
1563 (pp_append_text): Rename from pp_base_append_text.
1564 (pp_formatted_text): Rename from pp_base_formatted_text.
1565 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
1566 (pp_remaining_character_count_for_line): Rename from
1567 pp_base_remaining_character_count_for_line.
1568 (pp_newline): Rename from pp_base_newline.
1569 (pp_character): Rename from pp_base_character.
1570 (pp_string): Rename from pp_base_string.
1571 (pp_maybe_space): Rename from pp_base_maybe_space.
1572 * asan.c (asan_pp_string): Adjust.
1573 (asan_emit_stack_protection): Likewise.
1574 (asan_add_global): Likewise.
1575 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
1576 * tree-mudflap.c (mf_varname_tree): Likewise.
1577 * tree-pretty-print.c (pp_tree_identifier): Rename from
1578 pp_base_tree_identifier.
1579 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
1580 Declare as function.
1582 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1584 * pretty-print.h (pp_bar_bar): New.
1585 (pp_ampersand_ampersand): Likewise.
1586 (pp_less_equal): Likewise.
1587 (pp_greater_equal): Likewise.
1588 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
1589 printer functions instead of pp_string or operators and punctuators.
1590 (dump_gimple_call): Likewise.
1591 (dump_gimple_omp_for): Likewise.
1592 (dump_gimple_transaction): Likewise.
1593 (dump_gimple_phi): Likewise.
1594 (pp_gimple_stmt_1): Likewise.
1595 * sched-vis.c (print_insn): Likewise.
1596 * tree-mudflap.c (mf_varname_tree): Likewise.
1597 * tree-pretty-print.c (dump_block_node): Likewise.
1598 (dump_generic_node): Likewise.
1600 2013-08-02 Jan Hubicka <jh@suse.cz>
1602 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
1604 * lto-streamer-out.c (tree_is_indexable): Results decls and
1605 parm decls are not indexable.
1606 (DFS_write_tree_body): Do not follow args and results.
1607 (hash_tree): Likewise.
1608 (output_functions): Rearrange so struct function is needed
1609 only when real body is output; be able to also ouptut abstract
1610 functions; output DECL_ARGUMENTS and DECL_RESULT.
1611 (lto_output): When not in WPA, ale store abstract functions.
1612 (write_symbol): Do not care about RESULT_DECL.
1613 (output_symbol_p): Handle correctly sbtract decls.
1614 * lto-streamer-in.c (input_function): Rearrange so struct
1615 function can be NULL at entry; allow streaming of
1616 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
1617 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
1618 sanity check during LTO.
1619 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
1620 RESULT_DECl and DECL_ARGUMENTS.
1621 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
1624 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1626 * pretty-print.h (pp_underscore): New.
1628 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
1629 printer functions instead of pp_character.
1630 (dump_binary_rhs): Likewise.
1631 (dump_ternary_rhs): Likewise.
1632 (dump_gimple_call_args): Likewise.
1633 (pp_points_to_solution): Likewise.
1634 (dump_gimple_call): Likewise.
1635 (dump_gimple_switch): Likewise.
1636 (dump_gimple_cond): Likewise.
1637 (dump_gimple_bind): Likewise.
1638 (dump_gimple_try): Likewise.
1639 (dump_gimple_omp_for): Likewise.
1640 (dump_gimple_omp_continue): Likewise.
1641 (dump_gimple_omp_single): Likewise.
1642 (dump_gimple_omp_sections): Likewise.
1643 (dump_gimple_omp_block): Likewise.
1644 (dump_gimple_omp_critical): Likewise.
1645 (dump_gimple_transaction): Likewise.
1646 (dump_gimple_asm): Likewise.
1647 (dump_gimple_phi): Likewise.
1648 (dump_gimple_omp_parallel): Likewise.
1649 (dump_gimple_omp_task): Likewise.
1650 (dump_gimple_omp_atomic_load): Likewise.
1651 (dump_gimple_omp_atomic_store): Likewise.
1652 (dump_gimple_mem_ops): Likewise.
1653 (pp_gimple_stmt_1): Likewise.
1654 (pp_cfg_jump): Likewise.
1655 (dump_implicit_edges): Likewise.
1656 (gimple_dump_bb_for_graph): Likewise.
1657 * graph.c (draw_cfg_node): Likewise.
1658 * langhooks.c (lhd_print_error_function): Likewise.
1659 * sched-vis.c (print_exp): Likewise.
1660 (print_value): Likewise.
1661 (print_pattern): Likewise.
1662 (print_insn): Likewise.
1663 (rtl_dump_bb_for_graph): Likewise.
1664 * tree-pretty-print.c (dump_function_declaration): Likewise.
1665 (dump_array_domain): Likewise.
1666 (dump_omp_clause): Likewise.
1667 (dump_location): Likewise.
1668 (dump_generic_node): Likewise.
1669 (print_struct_decl): Likewise.
1670 * diagnostic.c (diagnostic_show_locus): Use pp_space.
1672 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1674 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
1675 candidate table when replacing a candidate statement.
1676 (replace_rhs_if_not_dup): Likewise.
1677 (replace_one_candidate): Likewise.
1679 2013-08-02 Jan Hubicka <jh@suse.cz>
1680 Martin Liska <marxin.liska@gmail.com>
1682 * cgraphunit.c (add_new_function): Fix logic when adding from
1684 (assemble_thunk): Rename to ...
1685 (expand_thunk); .. this one; export; get it working with general functions;
1686 make produced gimple valid.
1687 * cgraph.h (expand_thunk): Declare.
1689 2013-08-02 Jan Hubicka <jh@suse.cz>
1691 * ipa-cp.c (gather_context_independent_values): Use ipa_get_param_move_cost.
1692 (get_replacement_map): Remove PARAM; move parameter folding into tree-inline.c
1693 (create_specialized_node): Update.
1694 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
1695 assert that we have gimple body; update move_cost.
1696 (count_formal_params): Assert that we have gimple body.
1697 (ipa_dump_param): New function.
1698 (ipa_alloc_node_params): Break out from ...
1699 (ipa_initialize_node_params): ... here.
1700 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
1701 (ipa_write_node_info): Stream move costs.
1702 (ipa_read_node_info): Read move costs.
1703 (ipa_update_after_lto_read): Do not recompute node params.
1704 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
1705 (ipa_get_param): Check we are not in WPA.
1706 (ipa_get_param_move_cost): New.
1707 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
1708 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
1709 parm numbers to be present.
1711 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1713 PR rtl-optimization/58048
1714 * lra-constraints.c (process_alt_operands): Don't check asm
1715 operand on register.
1717 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
1719 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
1720 the implied StoreLoad barrier for atomic operations if before.
1722 2013-08-02 Jan Hubicka <jh@suse.cz>
1723 Martin Liska <marxin.liska@gmail.com>
1725 * cgraph.c (cgraph_function_body_availability): Do not check cgrpah flags.
1726 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
1727 symtab_node_availability): Declare.
1728 * ipa.c (can_replace_by_local_alias): New.
1729 (function_and_variable_visibility): Use it.
1730 * symtab.c (symtab_for_node_and_aliases, symtab_nonoverwritable_alias_1,
1731 symtab_nonoverwritable_alias): New.
1733 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1735 PR rtl-optimization/57963
1736 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p):
1738 (lra_constraints): Use them.
1740 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
1742 * config/arm/types.md (define_attr "type"): Add "load_acq" and "store_rel".
1743 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
1745 (cortex_a53_store1): Likewise.
1747 2013-08-01 Jan Hubicka <jh@suse.cz>
1749 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other partitions are
1752 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
1754 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
1755 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
1756 MAYBE_NON_Q_CLASS_P where appropriate.
1758 2013-08-01 Jan Hubicka <jh@suse.cz>
1760 * cgraph.h (release_function_body): Declare.
1761 * tree.c (free_lang_data_in_decl): Free, parameters and return values
1762 of unused delcarations.
1764 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1766 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical RTL form
1767 when subtracting a constant.
1769 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1771 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
1772 Generate canonical plus rtx with negated immediate instead of minus
1774 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
1776 2013-08-01 Jan Hubicka <jh@suse.cz>
1778 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
1779 (cgraph_release_function_body): Likewise.
1780 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1781 * cgraph.h (cgrpah_node): Rename abstract_and_needed
1782 to used_as_abstract_origin.
1783 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
1784 symbols used as abstract origins.
1785 * cgraphunit.c (analyze_functions): Update.
1786 * ipa.c (symtab_remove_unreachable_nodes): Recompute
1787 used_as_abstract_origin.
1788 * tree-inline.c (tree_function_versioning): Update
1789 used_as_abstract_origin; be ready for DECL_RESULT and
1790 DECL_ARGUMENTS to be NULL.
1792 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
1793 for abstract functions.
1794 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
1797 2013-08-01 Jan Hubicka <jh@suse.cz>
1799 * profile.c (compute_value_histograms): Fix thinko.
1801 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1803 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
1804 aarch-common-protos.h to extra_headers.
1805 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1806 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
1807 * config/aarch64/t-aarch64 (aarch-common.o): Define.
1809 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1811 * config/aarch64/aarch64.md (define_attr "type"): Delete.
1812 Include "../arm/types.md". Define "type" attribute for all patterns.
1813 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
1816 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
1818 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
1819 to support power8 load fusion.
1820 (fusion_gpr_mem_load): Likewise.
1822 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
1824 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
1825 declarations for power8 load fusion.
1826 (emit_fusion_gpr_load): Likewise.
1828 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
1829 tuning for power8, turn on fusion mode by default. Turn on sign
1830 extending fusion mode if normal fusion mode is on, and we are at
1832 (fusion_gpr_load_p): New function, return true if we can fuse an
1833 addis instruction with a dependent load to a GPR.
1834 (emit_fusion_gpr_load): Emit the instructions for power8 load
1837 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
1838 (VSX load fusion peepholes): New peepholes to fuse together an
1839 addi instruction with a VSX load instruction.
1841 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
1842 peepholes to fuse an addis instruction with a load to a GPR base
1843 register. If we are supporting sign extending fusions, convert
1844 sign extending loads to zero extending loads and add an explicit
1847 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1849 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
1850 aarch-common-protos.h to extra_headers.
1851 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1852 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
1853 (arm_early_store_addr_dep): Likewise.
1854 (arm_no_early_alu_shift_dep): Likewise.
1855 (arm_no_early_alu_shift_value_dep): Likewise.
1856 (arm_no_early_mul_dep): Likewise.
1857 (arm_no_early_store_addr_dep): Likewise.
1858 (arm_mac_accumulator_is_mul_result): Likewise.
1859 (arm_mac_accumulator_is_result): Likewise.
1860 * config/arm/aarch-common.c: ... here. New file.
1861 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
1863 (arm_early_store_addr_dep): Likewise.
1864 (arm_no_early_alu_shift_dep): Likewise.
1865 (arm_no_early_alu_shift_value_dep): Likewise.
1866 (arm_no_early_mul_dep): Likewise.
1867 (arm_no_early_store_addr_dep): Likewise.
1868 (arm_mac_accumulator_is_mul_result): Likewise.
1869 (arm_mac_accumulator_is_result): Likewise.
1870 * config/arm/aarch-common-protos.h: ... here. New file.
1871 * config/arm/t-arm (aarch-common.o): Define.
1873 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1875 * config/arm/arm.md: Include new file "types.md".
1876 (define_attr "type"): Move from here to ...
1877 (define_attr "mul32"): Likewise.
1878 (define_attr "mul64"): Likewise.
1879 * config/arm/types.md: ... here. New file.
1881 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
1883 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
1884 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
1886 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1888 * gen-pass-instances.awk: Fix offset of substr().
1890 2013-07-31 David Malcolm <dmalcolm@redhat.com>
1892 * Makefile.in (pass-instances.def): New.
1893 (passes.o): Replace dependency on passes.def with one on
1896 * gen-pass-instances.awk: New.
1898 * passes.c (pass_manager::pass_manager): Use pass-instances.def
1899 rather than passes.def, updating local definition of NEXT_PASS
1900 macro to add an extra NUM parameter (currently unused).
1902 2013-07-30 David Malcolm <dmalcolm@redhat.com>
1904 * Makefile.in (PASS_MANAGER_H): New.
1905 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
1906 (passes.o): Likewise.
1907 (statistics.o): Likewise.
1908 (cgraphunit.o): Likewise.
1909 (context.o): Depend on PASS_MANAGER_H.
1911 * pass_manager.h: New.
1913 * cgraphunit.c (cgraph_add_new_function): Update for moves
1914 of globals to fields of pass_manager.
1915 (analyze_function): Likewise.
1916 (expand_function): Likewise.
1917 (ipa_passes): Likewise.
1918 (compile): Likewise.
1920 * context.c (context::context): New.
1921 * context.h (context::context): New.
1922 (context::get_passes): New.
1923 (context::passes_): New.
1925 * lto-cgraph.c (input_node): Update for moves of globals to
1926 fields of pass_manager.
1928 * passes.c (all_passes): Remove, in favor of a field of the
1929 same name within the new class pass_manager.
1930 (all_small_ipa_passes): Likewise.
1931 (all_lowering_passes): Likewise.
1932 (all_regular_ipa_passes): Likewise.
1933 (all_late_ipa_passes): Likewise.
1934 (all_lto_gen_passes): Likewise.
1935 (passes_by_id): Likewise.
1936 (passes_by_id_size): Likewise.
1937 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
1938 the new class pass_manager.
1939 (set_pass_for_id): Convert to...
1940 (pass_manager::set_pass_for_id): ...method.
1941 (get_pass_for_id): Convert to...
1942 (pass_manager::get_pass_for_id): ...method.
1943 (register_one_dump_file): Move body of implementation into...
1944 (pass_manager::register_one_dump_file): ...here.
1945 (register_dump_files_1): Convert to...
1946 (pass_manager::register_dump_files_1): ...method.
1947 (register_dump_files): Convert to...
1948 (pass_manager::register_dump_files): ...method.
1949 (create_pass_tab): Update for moves of globals to fields of
1951 (dump_passes): Move body of implementation into...
1952 (pass_manager::dump_passes): ...here.
1953 (register_pass): Move body of implementation into...
1954 (pass_manager::register_pass): ...here.
1955 (init_optimization_passes): Convert into...
1956 (pass_manager::pass_manager): ...constructor for new
1957 pass_manager class, and initialize the pass_lists array.
1958 (check_profile_consistency): Update for moves of globals to
1959 fields of pass_manager.
1960 (dump_profile_report): Move body of implementation into...
1961 (pass_manager::dump_profile_report): ...here.
1962 (ipa_write_summaries_1): Update for moves of pass lists from
1963 being globals to fields of pass_manager.
1964 (ipa_write_optimization_summaries): Likewise.
1965 (ipa_read_summaries): Likewise.
1966 (ipa_read_optimization_summaries): Likewise.
1967 (execute_all_ipa_stmt_fixups): Likewise.
1969 * statistics.c (statistics_fini): Update for moves of globals to
1970 fields of pass_manager.
1972 * toplev.c (general_init): Replace call to
1973 init_optimization_passes with construction of the pass_manager
1976 * tree-pass.h (all_passes): Remove, in favor of a field of the
1977 same name within the new class pass_manager.
1978 (all_small_ipa_passes): Likewise.
1979 (all_lowering_passes): Likewise.
1980 (all_regular_ipa_passes): Likewise.
1981 (all_lto_gen_passes): Likewise.
1982 (all_late_ipa_passes): Likewise.
1983 (passes_by_id): Likewise.
1984 (passes_by_id_size): Likewise.
1985 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
1986 the new class pass_manager.
1987 (get_pass_for_id): Remove.
1989 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
1991 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
1994 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
1996 * arm.md (mulhi3): New expand pattern.
1998 2013-07-30 Jan Hubicka <jh@suse.cz>
1999 Martin Liska <marxin.liska@gmail.com>
2001 * profile.c (compute_value_histograms): Do not ICE when
2002 there is mismatch only on some counters.
2004 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2006 PR rtl-optimization/57637
2007 * function.c (move_insn_for_shrink_wrap): Also check the
2008 GEN set of the LIVE problem for the liveness analysis
2009 if it exists, otherwise give up.
2011 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2013 PR tree-optimization/57993
2014 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
2015 replaced statement in the candidate table.
2016 (phi_add_costs): Return infinite cost when the hidden basis does
2017 not dominate all phis on which the candidate is dependent.
2018 (replace_one_candidate): Record replaced statement in the
2021 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
2023 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
2024 (ashlv2si3): New expander.
2025 (*ashlv2si3_i): New define_insn_and_split.
2026 * predicates.md (float_operation): Allow patterns with three
2029 PR rtl-optimization/58021
2030 * mode-switching.c (create_pre_exit): Always split off preceding
2031 insns if we are not at the basic block head.
2033 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2035 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2036 (UCLIBC_DYNAMIC_LINKER): New macro.
2037 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2039 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2040 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2042 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2043 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2044 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2045 for SF and DF modes. Use ieee_quad_format for TF mode.
2046 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2047 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2048 (mips_option_override): Handle `-mnan=legacy'.
2049 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2050 `-mabs=2008' and `-mnan=2008'.
2051 (OPTION_DEFAULT_SPECS): Add "nan" default.
2052 (ASM_SPEC): Handle `-mnan='.
2053 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2054 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2055 comment accordingly.
2056 (neg<mode>2): Likewise.
2057 * config/mips/mips.opt (mabs, mnan): New options.
2058 * doc/install.texi (Configuration): Document `--with-nan=' option.
2059 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2061 (MIPS Options): Document them.
2062 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2063 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2064 * configure: Regenerate.
2065 * config.in: Regenerate.
2067 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2069 * config/i386/i386.md (float post-reload splitters): Do not check
2070 for subregs of SSE registers.
2072 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2073 H.J. Lu <hongjiu.lu@intel.com>
2077 * config/i386/i386.md (post-reload splitter
2078 to avoid partial SSE reg dependency stalls): New pattern.
2080 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2082 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2084 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2086 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2087 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2089 2013-07-26 Cary Coutant <ccoutant@google.com>
2091 * dwarf2out.c (die_checksum_ordered): Don't include template
2092 instantiations in signature.
2093 (is_template_parameter): New function.
2094 (is_template_instantiation): New function.
2095 (generate_skeleton_bottom_up): Don't include template instantiations
2097 (generate_skeleton): Likewise.
2098 (break_out_comdat_types): Move recursive call to break out nested
2100 (prune_unused_types_mark_generic_parms_dies): Call
2101 is_template_parameter.
2103 2013-07-26 Ian Bolton <ian.bolton@arm.com>
2105 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2106 uses vector registers.
2107 * config/aarch64/iterators.md: Add attributes rtn and vas.
2109 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2110 Richard Earnshaw <richard.earnshaw@arm.com>
2112 * combine.c (simplify_comparison): Re-canonicalize operands
2114 * config/arm/arm.md (movcond_addsi): New splitter.
2116 2013-07-25 Sterling Augustine <saugustine@google.com>
2118 * dwarf2out.c (size_of_pubnames): Move code to...
2119 (include_pubname_in_output): ...here. New.
2120 (want_pubnames): Rearrange.
2121 (output_pubnames): Call include_pubname_in_output. Move assertion.
2123 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2125 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2127 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2130 * doc/extend.texi: Remove obsolete builtins. Fix
2131 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2133 2013-07-25 Jan Hubicka <jh@suse.cz>
2135 * cgraph.c (release_function_body): Break out from ...
2136 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2138 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2139 old_tree in the map.
2140 (create_specialized_node): Update.
2141 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2143 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2144 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
2145 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2146 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2149 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2151 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2152 addsi3_carryin_alt2_<optab>): Correct output template.
2154 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2156 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2157 Adjust for arm_restrict_it.
2158 Remove trailing whitespace.
2160 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2162 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2163 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2165 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2167 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2169 PR rtl-optimization/57960
2170 * lra-constraints.c (process_alt_operands): Use the right mode
2171 when checking strict_low.
2173 2013-07-25 Jan Hubicka <jh@suse.cz>
2175 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2176 * cgraph.c (cgraph_remove_node): Do not release function body
2177 when in cgraph streaming.
2178 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
2179 in other partitions are not considered reachable; fix handling of
2182 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2184 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2186 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2191 * config/arm/arm.md ("*sibcall_insn): Replace use of
2192 Ss with US. Adjust output for v5 and v4t.
2193 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
2195 * config/arm/constraints.md ("Ss"): Rename to US.
2197 2013-07-25 Terry Guo <terry.guo@arm.com>
2199 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2200 shift_add/shift_sub0/shift_sub1 RTXs.
2202 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2203 Anton Blanchard <anton@au1.ibm.com>
2205 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2206 (altivec_vpks<VI_char>ss): Likewise.
2207 (altivec_vpks<VI_char>us): Likewise.
2208 (altivec_vpku<VI_char>us): Likewise.
2209 (altivec_vpku<VI_char>um): Likewise.
2211 2013-07-24 David Malcolm <dmalcolm@redhat.com>
2213 Introduce context class.
2215 * Makefile.in (CONTEXT_H): New.
2216 (OBJS): Add context.o.
2217 (toplev.o): Add CONTEXT_H to dependencies.
2220 * toplev.c (general_init): Create the singleton gcc::context instance.
2226 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2228 PR rtl-optimization/57968
2229 * mode-switching.c (create_pre_exit): Allow instructions that
2230 don't set a return register to need a non-exit mode.
2232 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2233 Anton Blanchard <anton@au1.ibm.com>
2235 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2236 operands to vperm for little endian.
2237 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2238 of lvsl to create the control mask for a vperm for little endian.
2240 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2241 Anton Blanchard <anton@au1.ibm.com>
2243 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2244 two operands for little-endian.
2246 2013-07-23 Steve Ellcey <sellcey@mips.com>
2248 * config/mips/mips.c (mips_case_values_threshold): New.
2249 (TARGET_CASE_VALUES_THRESHOLD): Define.
2251 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2252 Anton Blanchard <anton@au1.ibm.com>
2254 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2255 selection of field for vector splat in little endian mode.
2257 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2259 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2260 expanders to rs6000.md.
2261 (ior<mode>3): Likewise.
2262 (and<mode>3): Likewise.
2263 (one_cmpl<mode>2): Likewise.
2264 (nor<mode>3): Likewise.
2265 (andc<mode>3): Likewise.
2266 (eqv<mode>3): Likewise.
2267 (nand<mode>3): Likewise.
2268 (orc<mode>3): Likewise.
2270 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2273 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2274 to split multi-word logical operations.
2275 (rs6000_split_logical_di): Likewise.
2276 (rs6000_split_logical): Likewise.
2278 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2279 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2280 and allow TImode operations in 32-bit.
2281 (vsx_and<mode>3_64bit): Likewise.
2282 (vsx_ior<mode>3_32bit): Likewise.
2283 (vsx_ior<mode>3_64bit): Likewise.
2284 (vsx_xor<mode>3_32bit): Likewise.
2285 (vsx_xor<mode>3_64bit): Likewise.
2286 (vsx_one_cmpl<mode>2_32bit): Likewise.
2287 (vsx_one_cmpl<mode>2_64bit): Likewise.
2288 (vsx_nor<mode>3_32bit): Likewise.
2289 (vsx_nor<mode>3_64bit): Likewise.
2290 (vsx_andc<mode>3_32bit): Likewise.
2291 (vsx_andc<mode>3_64bit): Likewise.
2292 (vsx_eqv<mode>3_32bit): Likewise.
2293 (vsx_eqv<mode>3_64bit): Likewise.
2294 (vsx_nand<mode>3_32bit): Likewise.
2295 (vsx_nand<mode>3_64bit): Likewise.
2296 (vsx_orc<mode>3_32bit): Likewise.
2297 (vsx_orc<mode>3_64bit): Likewise.
2299 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2300 logical types in GPRs.
2302 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2303 logical insns to rs6000.md, and allow TImode operations in
2305 (altivec_ior<mode>3): Likewise.
2306 (altivec_xor<mode>3): Likewise.
2307 (altivec_one_cmpl<mode>2): Likewise.
2308 (altivec_nor<mode>3): Likewise.
2309 (altivec_andc<mode>3): Likewise.
2311 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2312 attributes for moving the 128-bit logical operations into
2314 (BOOL_REGS_OUTPUT): Likewise.
2315 (BOOL_REGS_OP1): Likewise.
2316 (BOOL_REGS_OP2): Likewise.
2317 (BOOL_REGS_UNARY): Likewise.
2318 (BOOL_REGS_AND_CR0): Likewise.
2319 (one_cmpl<mode>2): Add support for DI logical operations on
2320 32-bit, splitting the operations to 32-bit.
2324 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2325 changes to combine the 32/64-bit code, allow logical operations on
2326 TI mode in 32-bit, and to use similar match_operator patterns like
2327 scalar mode uses. Combine the Altivec and VSX code for logical
2328 operations, and move it here.
2329 (ior<mode>3, 128-bit types): Likewise.
2330 (xor<mode>3, 128-bit types): Likewise.
2331 (one_cmpl<mode>3, 128-bit types): Likewise.
2332 (nor<mode>3, 128-bit types): Likewise.
2333 (andc<mode>3, 128-bit types): Likewise.
2334 (eqv<mode>3, 128-bit types): Likewise.
2335 (nand<mode>3, 128-bit types): Likewise.
2336 (orc<mode>3, 128-bit types): Likewise.
2337 (and<mode>3_internal): Likewise.
2338 (bool<mode>3_internal): Likewise.
2339 (boolc<mode>3_internal1): Likewise.
2340 (boolc<mode>3_internal2): Likewise.
2341 (boolcc<mode>3_internal1): Likewise.
2342 (boolcc<mode>3_internal2): Likewise.
2343 (eqv<mode>3_internal1): Likewise.
2344 (eqv<mode>3_internal2): Likewise.
2345 (one_cmpl1<mode>3_internal): Likewise.
2347 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2349 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2350 Rename flag_stack_usage to flag_stack_usage_info.
2352 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2354 * config/microblaze/sync.md: New file.
2355 * config/microblaze/microblaze.md: Include sync.md
2356 * config/microblaze/microblaze.c: Add print_operand 'y'.
2357 * config/microblaze/constraints.md: Add memory_contraint
2358 'Q' which is a single register.
2360 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2362 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2364 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2366 * reload.c (find_reloads): Exit loop once we find this operand
2367 cannot be reloaded somehow for this alternative.
2369 * reload.c (find_reloads): Exit loop once we find a hard register.
2371 * rtlanal.c (computed_jump_p): Exit loop once we find label
2374 * i386.c (ix86_pad_returns): Exit loop after setting replace.
2376 * cfgloopmanip.c (remove_path): Exit loop after setting
2379 * gensupport.c (subst_dup): Avoid loop if code is not
2380 MATCH_DUP nor MATCH_OP_DUP.
2382 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
2384 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
2386 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2388 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
2389 true for SP_REGNUM if mode == ptr_mode.
2390 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
2391 with value R0_REGNUM + 31.
2393 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2395 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
2396 pad pointer-typed argument downward.
2398 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2400 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
2401 and __ILP32__ when the ILP32 model is in use.
2403 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2405 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
2406 (aarch64_load_symref_appropriately): In the case of
2407 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
2408 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
2409 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
2410 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
2411 if 'mode' doesn't equal to 'ptr_mode'.
2412 (aarch64_output_mi_thunk): Add an assertion on the alignment of
2413 'vcall_offset'; change to call aarch64_emit_move differently depending
2414 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
2415 to calculate the upper bound of 'vcall_offset'.
2416 (aarch64_cannot_force_const_mem): Change to also return true if
2418 (aarch64_legitimize_reload_address): In the case of large
2419 displacements, add new local variable 'xmode' and an assertion
2420 based on it; change to use 'xmode' to generate the new rtx and
2422 (aarch64_asm_trampoline_template): Change to generate the template
2423 differently depending on TARGET_ILP32 or not; change to use
2424 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
2425 (aarch64_trampoline_size): Removed.
2426 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
2427 and replace immediate literals with it. Change to use 'ptr_mode'
2428 instead of 'DImode' and call convert_memory_address if the mode
2429 of 'fnaddr' doesn't equal to 'ptr_mode'.
2430 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
2432 (aarch64_elf_asm_destructor): Likewise.
2433 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
2434 on TARGET_ILP32 instead of aarch64_trampoline_size.
2435 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
2436 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
2437 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
2438 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
2439 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
2440 (storewb_pair<GPI:mode>_<P:mode>): ... this.
2441 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
2442 depending on the value of 'mode'.
2443 (add_losym_<mode>): New.
2444 (ldr_got_small_<mode>): New, based on ldr_got_small.
2445 (ldr_got_small): Remove.
2446 (ldr_got_small_sidi): New.
2447 * config/aarch64/iterators.md (P): New.
2448 (PTR): Change to 'ptr_mode' in the condition.
2450 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2452 * config.gcc (aarch64*-*-*): Support --with-abi.
2453 (aarch64*-*-elf): Support --with-multilib-list.
2454 (aarch64*-*-linux*): Likewise.
2455 (supported_defaults): Add abi to aarch64*-*-*.
2456 * configure.ac: Mention AArch64 for --with-multilib-list.
2457 * configure: Re-generated.
2458 * config/aarch64/biarchilp32.h: New file.
2459 * config/aarch64/biarchlp64.h: New file.
2460 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
2462 (MULTILIB_DEFAULTS): Ditto.
2463 (DRIVER_SELF_SPECS): Ditto.
2464 (ASM_SPEC): Update to also substitute -mabi.
2465 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
2466 file whose name depends on -mabi= and -mbig-endian.
2467 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
2469 (POINTER_SIZE): New define.
2470 (POINTERS_EXTEND_UNSIGNED): Ditto.
2471 (enum aarch64_abi_type): New enumeration tag.
2472 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
2473 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
2474 (TARGET_ILP32): New define.
2475 * config/aarch64/aarch64.opt (mabi): New.
2477 (ilp32, lp64): New values for -mabi.
2478 * config/aarch64/t-aarch64 (comma): New define.
2479 (MULTILIB_OPTIONS): Ditto.
2480 (MULTILIB_DIRNAMES): Ditto.
2481 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
2482 * doc/invoke.texi: Document -mabi for AArch64.
2484 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
2486 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
2488 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2489 Anton Blanchard <anton@au1.ibm.com>
2491 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
2492 endianness when selecting field to splat.
2494 2013-07-22 Eric Christopher <echristo@gmail.com>
2496 * dwarf2out.c (die_odr_checksum): New function to use
2497 CHECKSUM_ macros and ULEB128 for DIE tag.
2498 (generate_type_signature): Use.
2500 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
2502 * config.gcc (sparc*-*-*): Accept leon3 processor.
2503 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
2504 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
2505 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
2506 * config/sparc/sparc.opt (enum processor_type): Add leon3.
2507 (mfix-ut699): Adjust comment.
2508 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
2509 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
2510 (CPP_CPU_SPEC): Likewise.
2511 (ASM_CPU_SPEC): Likewise.
2512 * config/sparc/sparc.c (leon3_cost): New constant.
2513 (sparc_option_override): Add leon3 support.
2514 (mem_ref): New function.
2515 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
2516 (sparc_do_work_around_errata): Look into the instruction in the delay
2517 slot and adjust accordingly. Add fix for the data cache nullify issues
2518 of the UT699. Change insertion position for the NOP.
2519 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
2520 (leon3_load): New reservation.
2521 (leon_store): Bump latency to 2.
2522 (grfpu): New automaton.
2523 (grfpu_alu): New unit.
2524 (grfpu_ds): Likewise.
2525 (leon_fp_alu): Adjust.
2526 (leon_fp_mult): Delete.
2527 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
2528 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
2529 * config/sparc/sparc.md (cpu): Add leon3.
2530 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
2532 (atomic_test_and_set): Likewise.
2535 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
2537 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
2538 default for R5900 targets.
2539 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
2540 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
2541 * config/mips/mips.c (mips_option_override): Report an error for
2542 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
2543 for -march=r5900 -mhard-float.
2545 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2547 * df-problems.c (can_move_insns_across): Exit loop once we
2548 find a non-fixed, non-global register.
2550 * ipa-pure-const.c (propagate_nothrow): Exit loop after
2553 * omega.c (omega_eliminate_red): Break after setting red_found.
2554 (omega_problem_has_red_equations): Similarly after setting found.
2555 (omega_query_variable): Similarly after setting coupled.
2557 2013-07-22 Marek Polacek <polacek@redhat.com>
2559 * gimplify.c: Don't include gimple.h twice.
2561 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2563 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
2564 instead of TARGET_THUMB1.
2565 (Pz): New constraint.
2566 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
2568 (compare_negsi_si): Likewise.
2569 (compare_addsi2_op0): Likewise.
2570 (compare_addsi2_op1): Likewise.
2571 (addsi3_carryin_<optab>): Likewise.
2572 (addsi3_carryin_alt2_<optab>): Likewise.
2573 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
2574 for arm_restrict_it.
2575 (subsi3_carryin): Likewise.
2576 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
2577 (minmax_arithsi): Disable for arm_restrict_it.
2578 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
2579 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
2580 (satsi_<SAT:code>_shift): Likewise.
2581 (arm_shiftsi3): Add alternative for 16-bit encoding.
2582 (arm32_movhf): Disable for arm_restrict_it.
2583 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
2584 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
2586 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
2588 * config/arm/arm.md (attribute "insn"): Delete.
2589 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
2590 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
2591 (not_shiftsi): Update for attribute change.
2592 (not_shiftsi_compare0): Likewise.
2593 (not_shiftsi_compare0_scratch): Likewise.
2594 (arm_one_cmplsi2): Likewise.
2595 (thumb1_one_cmplsi2): Likewise.
2596 (notsi_compare0): Likewise.
2597 (notsi_compare0_scratch): Likewise.
2598 (thumb1_movdi_insn): Likewise.
2599 (arm_movsi_insn): Likewise.
2600 (movhi_insn_arch4): Likewise.
2601 (movhi_bytes): Likewise.
2602 (arm_movqi_insn): Likewise.
2603 (thumb1_movqi_insn): Likewise.
2604 (arm32_movhf): Likewise.
2605 (thumb1_movhf): Likewise.
2606 (arm_movsf_soft_insn): Likewise.
2607 (thumb1_movsf_insn): Likewise.
2608 (thumb_movdf_insn): Likewise.
2609 (movsicc_insn): Likewise.
2610 (movsfcc_soft_insn): Likewise.
2611 (and_scc): Likewise.
2612 (cond_move): Likewise.
2613 (if_move_not): Likewise.
2614 (if_not_move): Likewise.
2615 (if_shift_move): Likewise.
2616 (if_move_shift): Likewise.
2617 (if_shift_shift): Likewise.
2618 (if_not_arith): Likewise.
2619 (if_arith_not): Likewise.
2620 (cond_move_not): Likewise.
2621 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
2622 (neon_mov<mode>): Likewise.
2623 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
2624 (thumb2_movsi_vfp): Likewise.
2625 (movsf_vfp): Likewise.
2626 (thumb2_movsf_vfp): Likewise.
2627 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
2629 (cortexa7_older_only): Likewise.
2630 (cortexa7_younger): Likewise.
2631 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
2632 (1020alu_shift_op): Likewise.
2633 (1020alu_shift_reg_op): Likewise.
2634 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
2635 (alu_shift_op): Likewise.
2636 (alu_shift_reg_op): Likewise.
2637 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
2638 (11_alu_shift_op): Likewise.
2639 (11_alu_shift_reg_op): Likewise.
2640 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
2641 (9_alu_shift_reg_op): Likewise.
2642 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2644 (cortex_a15_alu_shift): Likewise.
2645 (cortex_a15_alu_shift_reg): Likewise.
2646 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
2647 (cortex_a5_alu_shift): Likewise.
2648 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
2650 (cortex_a53_alu_shift): Likewise.
2651 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
2653 (cortex_a7_alu_reg): Likewise.
2654 (cortex_a7_alu_shift): Likewise.
2655 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2656 (cortex_a8_alu_shift): Likewise.
2657 (cortex_a8_alu_shift_reg): Likewise.
2658 (cortex_a8_mov): Likewise.
2659 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2660 (cortex_a9_dp_shift): Likewise.
2661 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
2662 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
2663 (cortex_r4_mov): Likewise.
2664 (cortex_r4_alu_shift): Likewise.
2665 (cortex_r4_alu_shift_reg): Likewise.
2666 * config/arm/fa526.md (526_alu_op): Update for attribute change.
2667 (526_alu_shift_op): Likewise.
2668 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
2669 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
2670 (626te_alu_shift_op): Likewise.
2671 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
2672 (726te_alu_op): Likewise.
2673 (726te_alu_shift_op): Likewise.
2674 (726te_alu_shift_reg_op): Likewise.
2675 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
2676 (mp626_alu_shift_op): Likewise.
2677 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
2678 (pj4_alu_e1_conds): Likewise.
2679 (pj4_alu): Likewise.
2680 (pj4_alu_conds): Likewise.
2681 (pj4_shift): Likewise.
2682 (pj4_shift_conds): Likewise.
2683 (pj4_alu_shift): Likewise.
2684 (pj4_alu_shift_conds): Likewise.
2686 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2688 * config/arm/predicates.md (shiftable_operator_strict_it):
2690 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
2691 Disable cond_exec version for arm_restrict_it.
2692 (thumb2_smaxsi3): Convert to generate cond_exec.
2693 (thumb2_sminsi3): Likewise.
2694 (thumb32_umaxsi3): Likewise.
2695 (thumb2_uminsi3): Likewise.
2696 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
2697 (thumb2_neg_abssi2): Likewise.
2698 (thumb2_mov_scc): Add alternative for 16-bit encoding.
2699 (thumb2_movsicc_insn): Adjust alternatives.
2700 (thumb2_mov_negscc): Disable for arm_restrict_it.
2701 (thumb2_mov_negscc_strict_it): New pattern.
2702 (thumb2_mov_notscc_strict_it): New pattern.
2703 (thumb2_mov_notscc): Disable for arm_restrict_it.
2704 (thumb2_ior_scc): Likewise.
2705 (thumb2_ior_scc_strict_it): New pattern.
2706 (thumb2_cond_move): Adjust for arm_restrict_it.
2707 (thumb2_cond_arith): Disable for arm_restrict_it.
2708 (thumb2_cond_arith_strict_it): New pattern.
2709 (thumb2_cond_sub): Adjust for arm_restrict_it.
2710 (thumb2_movcond): Likewise.
2711 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
2712 (thumb2_zero_extendhisi2_v6): Likewise.
2713 (thumb2_zero_extendqisi2_v6): Likewise.
2714 (orsi_notsi_si): Likewise.
2715 (orsi_not_shiftsi_si): Likewise.
2717 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
2719 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
2720 instruction sequence is 1 byte shorter.
2722 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
2724 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
2725 it is not needed after split.
2727 2013-07-20 Iain Sandoe <iain@codesourcery.com>
2730 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
2731 second label for nonlocal goto receivers. Don't output pic base labels
2732 unless we're producing PIC; mark that action unreachable().
2733 (ix86_save_reg): If the function contains a nonlocal label, save the
2735 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
2736 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
2737 (update_pic_label_number_if_needed): New.
2738 (machopic_output_function_base_name): Adjust for nonlocal receiver
2740 (machopic_should_output_picbase_label): New.
2741 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
2742 (nonlocal_goto_receiver): New insn and split.
2744 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
2746 * config/aarch64/aarch64-builtins.c
2747 (aarch64_fold_builtin): Fold abs in all modes.
2748 * config/aarch64/aarch64-simd-builtins.def
2749 (abs): Enable for all modes.
2750 * config/aarch64/arm_neon.h
2751 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
2752 (vabs_f64): Add missing intrinsic.
2754 2013-07-19 Ian Bolton <ian.bolton@arm.com>
2756 * config/aarch64/arm_neon.h (vabs_s64): New function
2758 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
2761 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
2762 * config/avr/avr.md (adjust_len): Add `round'.
2763 * config/avr/avr-protos.h (avr_out_round): New prototype.
2764 (avr_out_plus): Add `out_label' argument.
2765 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
2766 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
2767 Handle the case where `insn' is just a pattern.
2768 (avr_out_bitop): Handle the case where `insn' is just a pattern.
2769 (avr_out_round): New function.
2770 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
2772 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
2774 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2775 Add check for flag_stack_usage to handle -fstack-usage support
2777 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
2779 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
2780 interaction for new Power8 flags and VSX.
2782 2013-07-18 Sriraman Tallam <tmsriram@google.com>
2785 * tree-inline.c (expand_call_inline): Emit errors during
2786 early_inlining only if optimization is not turned on.
2788 2013-07-18 David Malcolm <dmalcolm@redhat.com>
2792 * passes.c (init_optimization_passes): Move the construction of
2793 the pass hierarchy into a new passes.def file.
2795 * Makefile.in (passes.o): Add dependency on passes.def.
2797 2013-07-18 David Malcolm <dmalcolm@redhat.com>
2799 * passes.c (init_optimization_passes): Introduce macros for
2800 constructing the tree of passes (INSERT_PASSES_AFTER,
2801 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
2802 TERMINATE_PASS_LIST).
2804 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
2805 Wei Mi <wmi@google.com>
2807 PR rtl-optimization/57878
2808 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
2810 (reload_pseudo_compare_func): Check nregs first for reload
2813 2013-07-18 David Malcolm <dmalcolm@redhat.com>
2815 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
2817 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
2819 * read-rtl.c (validate_const_int): Once an invalid character is
2820 seen, quit the loop.
2822 * gengtype.c (write_roots): Similarly once we find the "deletable"
2823 or "if_marked" option.
2825 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2827 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
2828 "xtab" and "sat". Move value "clz" from here to ...
2829 (attriubte "type"): ... here.
2830 (satsi_<SAT:code>): Delete "insn" attribute.
2831 (satsi_<SAT:code>_shift): Likewise.
2832 (arm_zero_extendqisi2addsi): Likewise.
2833 (arm_extendqisi2addsi): Likewise.
2834 (clzsi2): Update for attribute changes.
2835 (rbitsi2): Likewise.
2836 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
2838 (arm_usatsihi): Likewise.
2839 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2841 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2843 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
2844 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
2845 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
2846 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
2847 in alphabetical order.
2848 (attribute "core_cycles"): Update for attribute changes.
2849 (arm_addsi3): Likewise.
2850 (addsi3_compare0): Likewise.
2851 (addsi3_compare0_scratch): Likewise.
2852 (addsi3_compare_op1): Likewise.
2853 (addsi3_compare_op2): Likewise.
2854 (compare_addsi2_op0): Likewise.
2855 (compare_addsi2_op1): Likewise.
2856 (addsi3_carryin_shift_<optab>): Likewise.
2857 (subsi3_carryin_shift): Likewise.
2858 (rsbsi3_carryin_shift): Likewise.
2859 (arm_subsi3_insn): Likewise.
2860 (subsi3_compare0): Likewise.
2861 (subsi3_compare): Likewise.
2862 (arm_andsi3_insn): Likewise.
2863 (thumb1_andsi3_insn): Likewise.
2864 (andsi3_compare0): Likewise.
2865 (andsi3_compare0_scratch): Likewise.
2866 (zeroextractsi_compare0_scratch
2867 (andsi_not_shiftsi_si): Likewise.
2868 (iorsi3_insn): Likewise.
2869 (iorsi3_compare0): Likewise.
2870 (iorsi3_compare0_scratch): Likewise.
2871 (arm_xorsi3): Likewise.
2872 (thumb1_xorsi3_insn): Likewise.
2873 (xorsi3_compare0): Likewise.
2874 (xorsi3_compare0_scratch): Likewise.
2875 (satsi_<SAT:code>_shift): Likewise.
2877 (arm_shiftsi3): Likewise.
2878 (shiftsi3_compare0): Likewise.
2879 (not_shiftsi): Likewise.
2880 (not_shiftsi_compare0): Likewise.
2881 (not_shiftsi_compare0_scratch): Likewise.
2882 (arm_one_cmplsi2): Likewise.
2883 (thumb_one_complsi2): Likewise.
2884 (notsi_compare0): Likewise.
2885 (notsi_compare0_scratch): Likewise.
2886 (thumb1_zero_extendhisi2): Likewise.
2887 (arm_zero_extendhisi2): Likewise.
2888 (arm_zero_extendhisi2_v6): Likewise.
2889 (arm_zero_extendhisi2addsi): Likewise.
2890 (thumb1_zero_extendqisi2): Likewise.
2891 (thumb1_zero_extendqisi2_v6): Likewise.
2892 (arm_zero_extendqisi2): Likewise.
2893 (arm_zero_extendqisi2_v6): Likewise.
2894 (arm_zero_extendqisi2addsi): Likewise.
2895 (thumb1_extendhisi2): Likewise.
2896 (arm_extendhisi2): Likewise.
2897 (arm_extendhisi2_v6): Likewise.
2898 (arm_extendqisi): Likewise.
2899 (arm_extendqisi_v6): Likewise.
2900 (arm_extendqisi2addsi): Likewise.
2901 (thumb1_extendqisi2): Likewise.
2902 (thumb1_movdi_insn): Likewise.
2903 (arm_movsi_insn): Likewise.
2904 (movsi_compare0): Likewise.
2905 (movhi_insn_arch4): Likewise.
2906 (movhi_bytes): Likewise.
2907 (arm_movqi_insn): Likewise.
2908 (thumb1_movqi_insn): Likewise.
2909 (arm32_movhf): Likewise.
2910 (thumb1_movhf): Likewise.
2911 (arm_movsf_soft_insn): Likewise.
2912 (thumb1_movsf_insn): Likewise.
2913 (movdf_soft_insn): Likewise.
2914 (thumb_movdf_insn): Likewise.
2915 (arm_cmpsi_insn): Likewise.
2916 (cmpsi_shiftsi): Likewise.
2917 (cmpsi_shiftsi_swp): Likewise.
2918 (arm_cmpsi_negshiftsi_si): Likewise.
2919 (movsicc_insn): Likewise.
2920 (movsfcc_soft_insn): Likewise.
2921 (arith_shiftsi): Likewise.
2922 (arith_shiftsi_compare0
2923 (arith_shiftsi_compare0_scratch
2924 (sub_shiftsi): Likewise.
2925 (sub_shiftsi_compare0
2926 (sub_shiftsi_compare0_scratch
2927 (and_scc): Likewise.
2928 (cond_move): Likewise.
2929 (if_plus_move): Likewise.
2930 (if_move_plus): Likewise.
2931 (if_move_not): Likewise.
2932 (if_not_move): Likewise.
2933 (if_shift_move): Likewise.
2934 (if_move_shift): Likewise.
2935 (if_shift_shift): Likewise.
2936 (if_not_arith): Likewise.
2937 (if_arith_not): Likewise.
2938 (cond_move_not): Likewise.
2939 (thumb1_ashlsi3): Set type attribute.
2940 (thumb1_ashrsi3): Likewise.
2941 (thumb1_lshrsi3): Likewise.
2942 (thumb1_rotrsi3): Likewise.
2943 (shiftsi3_compare0_scratch): Likewise.
2944 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
2945 (neon_mov<mode>): Likewise.
2946 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
2948 (thumb2_movsi_insn): Likewise.
2949 (thumb2_cmpsi_neg_shiftsi): Likewise.
2950 (thumb2_extendqisi_v6): Likewise.
2951 (thumb2_zero_extendhisi2_v6): Likewise.
2952 (thumb2_zero_extendqisi2_v6): Likewise.
2953 (thumb2_shiftsi3_short): Likewise.
2954 (thumb2_addsi3_compare0_scratch): Likewise.
2955 (orsi_not_shiftsi_si): Likewise.
2956 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
2957 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
2959 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2960 (1020alu_shift_op): Likewise.
2961 (1020alu_shift_reg_op): Likewise.
2962 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2963 (alu_shift_op): Likewise.
2964 (alu_shift_reg_op): Likewise.
2965 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2966 (11_alu_shift_op): Likewise.
2967 (11_alu_shift_reg_op): Likewise.
2968 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2969 (9_alu_shift_reg_op): Likewise.
2970 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2972 (cortex_a15_alu_shift): Likewise.
2973 (cortex_a15_alu_shift_reg): Likewise.
2974 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
2976 (cortex_a5_alu_shift): Likewise.
2977 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
2979 (cortex_a53_alu_shift): Likewise.
2980 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
2982 (cortex_a7_alu_reg): Likewise.
2983 (cortex_a7_alu_shift): Likewise.
2984 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
2986 (cortex_a8_alu_shift): Likewise.
2987 (cortex_a8_alu_shift_reg): Likewise.
2988 (cortex_a8_mov): Likewise.
2989 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2990 (cortex_a9_dp_shift): Likewise.
2991 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
2993 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
2995 (cortex_r4_mov): Likewise.
2996 (cortex_r4_alu_shift): Likewise.
2997 (cortex_r4_alu_shift_reg): Likewise.
2998 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2999 (526_alu_shift_op): Likewise.
3000 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3001 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3002 (626te_alu_shift_op): Likewise.
3003 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
3004 (726te_alu_op): Likewise.
3005 (726te_alu_shift_op): Likewise.
3006 (726te_alu_shift_reg_op): Likewise.
3007 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3008 (mp626_alu_shift_op): Likewise.
3009 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
3010 (pj4_alu_e1_conds): Likewise.
3011 (pj4_alu): Likewise.
3012 (pj4_alu_conds): Likewise.
3013 (pj4_shift): Likewise.
3014 (pj4_shift_conds): Likewise.
3015 (pj4_alu_shift): Likewise.
3016 (pj4_alu_shift_conds): Likewise.
3017 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3019 (cortexa7_older_only): Likewise.
3020 (cortexa7_younger): Likewise.
3022 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3024 * ipa-pure-const.c (generate_summary): Rename to...
3025 (pure_const_generate_summary): ... this.
3027 2013-07-17 Iain Sandoe <iain@codesourcery.com>
3029 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3031 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
3034 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3037 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3039 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3040 enabled without -march=zEC12.
3041 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3044 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3046 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3047 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3048 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3049 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3050 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3051 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3052 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3054 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3055 and ISA_HAS_NMADD3_NMSUB3.
3056 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3057 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3058 (nmsub4<mode>, nmsub3<mode>): Likewise.
3059 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3061 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3063 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3064 TARGET_MIPS5400 checking.
3066 2013-07-16 Jakub Jelinek <jakub@redhat.com>
3067 Peter Bergner <bergner@vnet.ibm.com>
3069 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3070 registers in the comment.
3071 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3072 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3073 rather than FIRST_PSEUDO_REGISTERS.
3075 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
3077 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3078 enable extra ISA flags with TARGET_HTM.
3080 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3082 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3085 2013-07-15 Cong Hou <congh@google.com>
3087 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3088 in compare function for sorting.
3090 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
3092 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3093 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3094 * config/rs6000/rs6000.opt: Add -mhtm option.
3095 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3096 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3097 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3098 __HTM__ if the HTM instructions are available.
3099 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3100 htm_spr_reg_operand): New define_predicates.
3101 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3102 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3104 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3105 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3106 HTM builtin functions.
3107 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3108 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3109 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3110 (rs6000_builtin_mask_calculate): Likewise.
3111 (rs6000_option_override_internal): Likewise.
3112 (bdesc_htm): Add new HTM builtin support.
3113 (htm_spr_num): New function.
3114 (htm_spr_regno): Likewise.
3115 (rs6000_htm_spr_icode): Likewise.
3116 (htm_expand_builtin): Likewise.
3117 (htm_init_builtins): Likewise.
3118 (rs6000_expand_builtin): Add support for HTM builtin functions.
3119 (rs6000_init_builtins): Likewise.
3120 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3122 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3123 (TARGET_HTM, MASK_HTM): Define macros.
3124 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3125 (FIXED_REGISTERS): Likewise.
3126 (CALL_USED_REGISTERS): Likewise.
3127 (CALL_REALLY_USED_REGISTERS): Likewise.
3128 (REG_ALLOC_ORDER): Likewise.
3129 (enum reg_class): Likewise.
3130 (REG_CLASS_NAMES): Likewise.
3131 (REG_CLASS_CONTENTS): Likewise.
3132 (REGISTER_NAMES): Likewise.
3133 (ADDITIONAL_REGISTER_NAMES): Likewise.
3134 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3135 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3136 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3137 * config/rs6000/htm.md: New file.
3138 * config/rs6000/htmintrin.h: New file.
3139 * config/rs6000/htmxlintrin.h: New file.
3141 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3143 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3144 Define SYMBOL_TINY_GOT, update comment.
3145 * config/aarch64/aarch64.c
3146 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3147 (aarch64_expand_mov_immediate): Likewise.
3148 (aarch64_print_operand): Likewise.
3149 (aarch64_classify_symbol): Likewise.
3150 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3151 (ldr_got_tiny): Define.
3153 2013-07-13 Tobias Grosser <tobias@grosser.es>
3155 PR tree-optimization/54094
3156 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3157 scheduling dimension for the parallelism check from the polyhedral
3158 information in the AST.
3159 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3162 2013-07-13 Jason Merrill <jason@redhat.com>
3164 * print-tree.c (debug_vec_tree): Use debug_raw.
3165 (debug_raw (vec<tree, va_gc> &)): New.
3166 (debug_raw (vec<tree, va_gc> *)): New.
3167 * tree.h: Declare them.
3169 2013-07-13 Bin Cheng <bin.cheng@arm.com>
3171 * ifcvt.c (ifcvt_after_combine): New static variable.
3172 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3174 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3175 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3176 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3178 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3180 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3182 2013-07-12 Michael Matz <matz@suse.de>
3185 * convert.c (convert_to_real): Reject non-float inner types.
3187 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3189 * config/aarch64/aarch64-protos.h
3190 (aarch64_simd_immediate_valid_for_move): Remove.
3191 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3192 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3194 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3196 2013-07-11 Steve Ellcey <sellcey@mips.com>
3198 * config/mips/mips.c (mips_conditional_register_usage): Do not
3199 use t[0-7] registers in MIPS16 mode when optimizing for size.
3201 2013-07-11 Sriraman Tallam <tmsriram@google.com>
3203 * config/i386/i386.c (dispatch_function_versions): Fix array
3204 indexing of function_version_info to match actual_versions.
3206 2013-07-11 Teresa Johnson <tejohnson@google.com>
3208 * vec.h (struct va_gc): Move release out-of-line.
3209 (va_gc::release): Call ggc_free on released vec.
3211 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3213 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3214 Require GOT register as additional operand in UNSPEC.
3215 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3216 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3217 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3218 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3219 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3220 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3221 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3223 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
3226 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3227 name seen by assembler/linker rather if available.
3229 2013-07-11 Andreas Schwab <schwab@suse.de>
3231 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3233 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3235 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
3237 2013-07-10 Joseph Myers <joseph@codesourcery.com>
3239 * doc/tm.texi.in: Move hook documentation to ....
3240 * target.def: ... here.
3242 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3244 * doc/tm.texi: Regenerate.
3246 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3249 * doc/invoke.texi: Document -Wconditionally-supported.
3251 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3254 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3257 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3260 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3261 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3262 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3263 Remove duplicate devices.
3264 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3265 * config/avr/t-multilib: Regenerate.
3266 * config/avr/avr-tables.opt: Regenerate.
3267 * doc/avr-mmcu.texi: Regenerate.
3269 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3272 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3274 2013-07-10 Graham Stott <graham.stott@btinternet.com>
3276 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3277 the cost of MULT when optimizing for size.
3279 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3281 * config/cr16/cr16-protos.h: Don't include target.h.
3283 2013-07-09 Joseph Myers <joseph@codesourcery.com>
3285 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3286 adjust register size for TDmode and TFmode for VSX registers.
3288 2013-07-08 Kai Tietz <ktietz@redhat.com>
3291 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3292 hook_bool_const_tree_true.
3294 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3296 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3297 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3298 * config/s390/s390.md: Define FPR*_REGNUM constants.
3299 Fix FPR2_REGNUM constant (18 -> 17).
3300 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3301 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3302 ("trunc<BFP:mode><DFP_ALL:mode>2")
3303 ("trunc<DFP_ALL:mode><BFP:mode>2")
3304 ("*extend<BFP:mode><DFP_ALL:mode>2")
3305 ("*extend<DFP_ALL:mode><BFP:mode>2")
3306 ("extend<BFP:mode><DFP_ALL:mode>2")
3307 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3310 2013-07-08 Graham Stott <graham.stott@btinternet.com>
3312 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
3314 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3316 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3317 and cfun_fpr_bit_p to cfun_fpr_save_p.
3318 (s390_frame_area, s390_register_info, s390_frame_info)
3319 (s390_emit_prologue, s390_emit_epilogue)
3320 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3322 * config/s390/s390.h: Define *_REGNUM macros for floating point
3325 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3327 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3329 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3331 PR rtl-optimization/57786
3332 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3333 and break out of the loop when it is set to false.
3335 2013-07-08 Jakub Jelinek <jakub@redhat.com>
3338 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3339 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3340 (const_int 63)) 0)).
3341 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3342 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3343 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3345 PR rtl-optimization/57829
3346 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3347 mask bits outside of mode are just sign-extension from mode to HWI.
3349 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3351 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3352 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3353 adjust_address instead of change_address to keep info about alignment.
3354 (emit_strmov): Remove.
3355 (emit_memmov): New function.
3356 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3357 (expand_movmem_epilogue): Likewise and return updated rtx for
3359 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3360 destination and source.
3361 (decide_alignment): Refactor, handle vector_loop.
3362 (ix86_expand_movmem): Likewise.
3363 (ix86_expand_setmem): Likewise.
3364 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3366 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
3368 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3369 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
3371 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3373 * config/i386/sse.md (sse_movlhps): Change alternative 3
3374 of operand 2 to "m".
3376 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3379 * config/i386/sse.md (iptr): New mode attribute.
3380 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
3381 (<sse>_vm<plusminus_insn><mode>3): Ditto.
3382 (<sse>_vmmul<mode>3): Ditto.
3383 (<sse>_vmdiv<mode>3): Ditto.
3384 (sse_vmrcpv4sf2): Ditto.
3385 (<sse>_vmsqrt<mode>2): Ditto.
3386 (sse_vmrsqrtv4sf2): Ditto.
3387 (<sse>_vm<code><mode>3): Ditto.
3388 (avx_vmcmp<mode>3): Ditto.
3389 (<sse>_vmmaskcmp<mode>3): Ditto.
3390 (<sse>_comi): Ditto.
3391 (<sse>_ucomi): Ditto.
3392 (*xop_vmfrcz_<mode>): Ditto.
3393 (*fmai_fmadd_<mode>): Ditto.
3394 (*fmai_fmsub_<mode>): Ditto.
3395 (*fmai_fnmadd_<mode>): Ditto.
3396 (*fmai_fnmsub_<mode>): Ditto.
3397 (*fma4i_vmfmadd_<mode>): Ditto.
3398 (*fma4i_vmfmsub_<mode>): Ditto.
3399 (*fma4i_vmfnmadd_<mode>): Ditto.
3400 (*fma4i_vmfnmsub_<mode>): Ditto.
3401 (*xop_vmfrcz_<mode>): Ditto.
3402 (sse_cvtps2pi): Ditto.
3403 (sse_cvttps2pi): Ditto.
3404 (sse_cvtss2si): Ditto.
3405 (sse_cvtss2si_2): Ditto.
3406 (sse_cvtss2siq_2): Ditto.
3407 (sse_cvttss2si): Ditto.
3408 (sse_cvttss2siq): Ditto.
3409 (sse_cvtsd2si): Ditto.
3410 (sse_cvtsd2si_2): Ditto.
3411 (sse_cvtsd2siq_2): Ditto.
3412 (sse_cvttsd2si): Ditto.
3413 (sse_cvttsd2siq): Ditto.
3414 (sse_cvtsd2ss): Ditto.
3415 (sse_cvtss2sd): Ditto.
3416 (avx2_pbroadcast<mode>): Ditto.
3417 (avx2_pbroadcast<mode>_1): Ditto.
3418 (*avx_vperm_broadcast_v4sf): Ditto.
3420 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
3421 (sse_movlhps): Ditto.
3422 (sse_storehps): Ditto.
3423 (sse_loadhps): Ditto.
3424 (sse_storelps): Ditto.
3425 (sse_loadlps): Ditto.
3426 (*vec_concatv4sf): Ditto.
3427 (*vec_interleave_highv2df): Ditto.
3428 (*vec_interleave_lowv2df): Ditto.
3429 (*vec_extractv2df_1_sse): Ditto.
3430 (*vec_extractv2df_0_sse): Ditto.
3431 (sse2_storelpd): Ditto.
3432 (sse2_loadlpd): Ditto.
3433 (sse2_movsd): Ditto.
3434 (*vec_concatv4si): Ditto.
3435 (vec_concatv2di): Ditto.
3437 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
3438 for Intel asm dialect.
3439 (mmx_punpcklwd): Ditto.
3440 (mmx_punpckldq): Ditto.
3442 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
3443 for intel assembler dialect.
3445 2013-07-06 Jakub Jelinek <jakub@redhat.com>
3448 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
3449 for BUILT_IN_C{LZ,LRSB}*.
3450 * tree.h (CASE_INT_FN): Add FN##IMAX case.
3451 * tree-vrp.c (extract_range_basic): Handle
3452 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
3453 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
3454 fall thru to code calling set_value*.
3455 * builtins.c (expand_builtin): Remove *IMAX cases.
3456 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
3457 if width is bigger than 2*HWI.
3459 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
3461 PR rtl-optimization/55342
3462 * lra-int.h (lra_subreg_reload_pseudos): New.
3463 * lra.c: Add undoing optional reloads to the block diagram.
3464 (lra_subreg_reload_pseudos): New.
3465 (lra_optional_reload_pseudos): Change comments.
3466 (lra): Init and clear lra_subreg_reload_pseudos. Clear
3467 lra_optional_reload_pseudos after undo transformations.
3468 * lra-assigns.c (pseudo_prefix_title): New.
3469 (lra_setup_reg_renumber): Use it.
3470 (spill_for): Ditto. Check subreg reload pseudos too.
3471 (assign_by_spills): Consider subreg reload pseudos too.
3472 * lra-constraints.c (simplify_operand_subreg): Use
3473 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
3474 (curr_insn_transform): Recognize and do optional reloads.
3475 (undo_optional_reloads): New.
3476 (lra_undo_inheritance): Call undo_optional_reloads.
3478 2013-07-05 Thomas Quinot <quinot@adacore.com>
3480 * tree-complex.c (expand_complex_operations_1): Fix typo.
3482 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
3484 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
3485 (tune_params): New member 'const vec_costs'.
3486 * config/aarch64/aarch64.c (generic_vector_cost): New.
3487 (generic_tunings): New member 'generic_vector_cost'.
3488 (aarch64_builtin_vectorization_cost): New.
3489 (aarch64_add_stmt_cost): New.
3490 (TARGET_VECTORIZE_ADD_STMT_COST): New.
3491 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
3493 2013-07-03 Jakub Jelinek <jakub@redhat.com>
3496 * config/i386/predicates.md (vsib_address_operand): Disallow
3497 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
3499 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
3502 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
3503 expand_builtin_setjmp_receiver.
3504 (expand_label): Adjust, call expand_builtin_setjmp_receiver
3505 with NULL for the label parameter.
3506 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
3507 the frame-pointer. Adjust comments.
3508 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
3509 only if LABEL is non-NULL.
3511 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
3513 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
3514 (ARM_ABI_AAPCS64): Ditto.
3516 (ARM_DEFAULT_ABI): Ditto.
3518 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
3520 * config/aarch64/aarch64-builtins.c
3521 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
3522 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
3524 * config/aarch64/aarch64-simd.md
3525 (aarch64_ld1<VALL:mode>): New.
3526 (aarch64_st1<VALL:mode>): Likewise.
3527 * config/aarch64/arm_neon.h
3528 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
3530 2013-07-02 Sriraman Tallam <tmsriram@google.com>
3532 * config/i386/i386.c (gate_insert_vzeroupper): Check if
3534 (ix86_option_override_internal):Turn on all -mavx target flags by
3535 default as they are dependent on AVX anyway.
3537 2013-07-02 Cary Coutant <ccoutant@google.com>
3539 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
3541 (loc_checksum_ordered): Likewise.
3542 (hash_loc_operands): Remove inline keyword.
3544 2013-07-02 Jakub Jelinek <jakub@redhat.com>
3546 PR tree-optimization/57741
3547 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
3548 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
3549 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
3550 Allow REAL_CST step_exprs if flag_associative_math.
3551 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
3553 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3555 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
3557 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3559 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
3561 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3563 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
3565 (iorsi3_insn): Likewise.
3566 (arm_xorsi3): Likewise.
3568 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
3570 * arm.md (attribute "wtype"): Delete. Move attribute values from here
3572 (attribute "type"): ... here, and prefix with "wmmx_".
3573 (attribute "core_cycles"): Update for attribute changes.
3574 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
3575 (tbcstv4hi): Likewise.
3576 (tbcstv2si): Likewise.
3577 (iwmmxt_iordi3): Likewise.
3578 (iwmmxt_xordi3): Likewise.
3579 (iwmmxt_anddi3): Likewise.
3580 (iwmmxt_nanddi3): Likewise.
3581 (iwmmxt_arm_movdi): Likewise.
3582 (iwmmxt_movsi_insn): Likewise.
3583 (mov<mode>_internal): Likewise.
3584 (and<mode>3_iwmmxt): Likewise.
3585 (ior<mode>3_iwmmxt): Likewise.
3586 (xor<mode>3_iwmmxt): Likewise.
3587 (add<mode>3_iwmmxt): Likewise.
3588 (ssaddv8qi3): Likewise.
3589 (ssaddv4hi3): Likewise.
3590 (ssaddv2si3): Likewise.
3591 (usaddv8qi3): Likewise.
3592 (usaddv4hi3): Likewise.
3593 (usaddv2si3): Likewise.
3594 (sub<mode>3_iwmmxt): Likewise.
3595 (sssubv8qi3): Likewise.
3596 (sssubv4hi3): Likewise.
3597 (sssubv2si3): Likewise.
3598 (ussubv8qi3): Likewise.
3599 (ussubv4hi3): Likewise.
3600 (ussubv2si3): Likewise.
3601 (mulv4hi3_iwmmxt): Likewise.
3602 (smulv4hi3_highpart): Likewise.
3603 (umulv4hi3_highpart): Likewise.
3604 (iwmmxt_wmacs): Likewise.
3605 (iwmmxt_wmacsz): Likewise.
3606 (iwmmxt_wmacu): Likewise.
3607 (iwmmxt_wmacuz): Likewise.
3608 (iwmmxt_clrdi): Likewise.
3609 (iwmmxt_clrv8qi): Likewise.
3610 (iwmmxt_clr4hi): Likewise.
3611 (iwmmxt_clr2si): Likewise.
3612 (iwmmxt_uavgrndv8qi3): Likewise.
3613 (iwmmxt_uavgrndv4hi3): Likewise.
3614 (iwmmxt_uavgv8qi3): Likewise.
3615 (iwmmxt_uavgv4hi3): Likewise.
3616 (iwmmxt_tinsrb): Likewise.
3617 (iwmmxt_tinsrh): Likewise.
3618 (iwmmxt_tinsrw): Likewise.
3619 (iwmmxt_textrmub): Likewise.
3620 (iwmmxt_textrmsb): Likewise.
3621 (iwmmxt_textrmuh): Likewise.
3622 (iwmmxt_textrmsh): Likewise.
3623 (iwmmxt_textrmw): Likewise.
3624 (iwmxxt_wshufh): Likewise.
3625 (eqv8qi3): Likewise.
3626 (eqv4hi3): Likewise.
3627 (eqv2si3): Likewise.
3628 (gtuv8qi3): Likewise.
3629 (gtuv4hi3): Likewise.
3630 (gtuv2si3): Likewise.
3631 (gtv8qi3): Likewise.
3632 (gtv4hi3): Likewise.
3633 (gtv2si3): Likewise.
3634 (smax<mode>3_iwmmxt): Likewise.
3635 (umax<mode>3_iwmmxt): Likewise.
3636 (smin<mode>3_iwmmxt): Likewise.
3637 (umin<mode>3_iwmmxt): Likewise.
3638 (iwmmxt_wpackhss): Likewise.
3639 (iwmmxt_wpackwss): Likewise.
3640 (iwmmxt_wpackdss): Likewise.
3641 (iwmmxt_wpackhus): Likewise.
3642 (iwmmxt_wpackwus): Likewise.
3643 (iwmmxt_wpackdus): Likewise.
3644 (iwmmxt_wunpckihb): Likewise.
3645 (iwmmxt_wunpckihh): Likewise.
3646 (iwmmxt_wunpckihw): Likewise.
3647 (iwmmxt_wunpckilb): Likewise.
3648 (iwmmxt_wunpckilh): Likewise.
3649 (iwmmxt_wunpckilw): Likewise.
3650 (iwmmxt_wunpckehub): Likewise.
3651 (iwmmxt_wunpckehuh): Likewise.
3652 (iwmmxt_wunpckehuw): Likewise.
3653 (iwmmxt_wunpckehsb): Likewise.
3654 (iwmmxt_wunpckehsh): Likewise.
3655 (iwmmxt_wunpckehsw): Likewise.
3656 (iwmmxt_wunpckelub): Likewise.
3657 (iwmmxt_wunpckeluh): Likewise.
3658 (iwmmxt_wunpckeluw): Likewise.
3659 (iwmmxt_wunpckelsb): Likewise.
3660 (iwmmxt_wunpckelsh): Likewise.
3661 (iwmmxt_wunpckelsw): Likewise.
3662 (ror<mode>3): Likewise.
3663 (ashr<mode>3_iwmmxt): Likewise.
3664 (lshr<mode>3_iwmmxt): Likewise.
3665 (ashl<mode>3_iwmmxt): Likewise.
3666 (ror<mode>3_di): Likewise.
3667 (ashr<mode>3_di): Likewise.
3668 (lshr<mode>3_di): Likewise.
3669 (ashl<mode>3_di): Likewise.
3670 (iwmmxt_wmadds): Likewise.
3671 (iwmmxt_wmaddu): Likewise.
3672 (iwmmxt_tmia): Likewise.
3673 (iwmmxt_tmiaph): Likewise.
3674 (iwmmxt_tmiabb): Likewise.
3675 (iwmmxt_tmiatb): Likewise.
3676 (iwmmxt_tmiabt): Likewise.
3677 (iwmmxt_tmiatt): Likewise.
3678 (iwmmxt_tmovmskb): Likewise.
3679 (iwmmxt_tmovmskh): Likewise.
3680 (iwmmxt_tmovmskw): Likewise.
3681 (iwmmxt_waccb): Likewise.
3682 (iwmmxt_wacch): Likewise.
3683 (iwmmxt_waccw): Likewise.
3684 (iwmmxt_waligni): Likewise.
3685 (iwmmxt_walignr): Likewise.
3686 (iwmmxt_walignr0): Likewise.
3687 (iwmmxt_walignr1): Likewise.
3688 (iwmmxt_walignr2): Likewise.
3689 (iwmmxt_walignr3): Likewise.
3690 (iwmmxt_wsadb): Likewise.
3691 (iwmmxt_wsadh): Likewise.
3692 (iwmmxt_wsadbz): Likewise.
3693 (iwmmxt_wsadhz): Likewise.
3694 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
3695 (iwmmxt_wabsdiffb): Likewise.
3696 (iwmmxt_wabsdiffh): Likewise.
3697 (iwmmxt_wabsdiffw): Likewise.
3698 (iwmmxt_waddsubhx): Likewise
3699 (iwmmxt_wsubaddhx): Likewise.
3700 (addc<mode>3): Likewise.
3701 (iwmmxt_avg4): Likewise.
3702 (iwmmxt_avg4r): Likewise.
3703 (iwmmxt_wmaddsx): Likewise.
3704 (iwmmxt_wmaddux): Likewise.
3705 (iwmmxt_wmaddsn): Likewise.
3706 (iwmmxt_wmaddun): Likewise.
3707 (iwmmxt_wmulwsm): Likewise.
3708 (iwmmxt_wmulwum): Likewise.
3709 (iwmmxt_wmulsmr): Likewise.
3710 (iwmmxt_wmulumr): Likewise.
3711 (iwmmxt_wmulwsmr): Likewise.
3712 (iwmmxt_wmulwumr): Likewise.
3713 (iwmmxt_wmulwl): Likewise.
3714 (iwmmxt_wqmulm): Likewise.
3715 (iwmmxt_wqmulwm): Likewise.
3716 (iwmmxt_wqmulmr): Likewise.
3717 (iwmmxt_wqmulwmr): Likewise.
3718 (iwmmxt_waddbhusm): Likewise.
3719 (iwmmxt_waddbhusl): Likewise.
3720 (iwmmxt_wqmiabb): Likewise.
3721 (iwmmxt_wqmiabt): Likewise.
3722 (iwmmxt_wqmiatb): Likewise.
3723 (iwmmxt_wqmiatt): Likewise.
3724 (iwmmxt_wqmiabbn): Likewise.
3725 (iwmmxt_wqmiabtn): Likewise.
3726 (iwmmxt_wqmiatbn): Likewise.
3727 (iwmmxt_wqmiattn): Likewise.
3728 (iwmmxt_wmiabb): Likewise.
3729 (iwmmxt_wmiabt): Likewise.
3730 (iwmmxt_wmiatb): Likewise.
3731 (iwmmxt_wmiatt): Likewise.
3732 (iwmmxt_wmiabbn): Likewise.
3733 (iwmmxt_wmiabtn): Likewise.
3734 (iwmmxt_wmiatbn): Likewise.
3735 (iwmmxt_wmiattn): Likewise.
3736 (iwmmxt_wmiawbb): Likewise.
3737 (iwmmxt_wmiawbt): Likewise.
3738 (iwmmxt_wmiawtb): Likewise.
3739 (iwmmxt_wmiawtt): Likewise.
3740 (iwmmxt_wmiawbbn): Likewise.
3741 (iwmmxt_wmiawbtn): Likewise.
3742 (iwmmxt_wmiawtbn): Likewise.
3743 (iwmmxt_wmiawttn): Likewise.
3744 (iwmmxt_wmerge): Likewise.
3745 (iwmmxt_tandc<mode>3): Likewise.
3746 (iwmmxt_torc<mode>3): Likewise.
3747 (iwmmxt_torvsc<mode>3): Likewise.
3748 (iwmmxt_textrc<mode>3): Likewise.
3749 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
3750 (wmmxt_pack): Likewise.
3751 (wmmxt_mult_c1): Likewise.
3752 (wmmxt_mult_c2): Likewise.
3753 (wmmxt_alu_c1): Likewise.
3754 (wmmxt_alu_c2): Likewise.
3755 (wmmxt_alu_c3): Likewise.
3756 (wmmxt_transfer_c1): Likewise.
3757 (wmmxt_transfer_c2): Likewise.
3758 (wmmxt_transfer_c3): Likewise.
3759 (marvell_f_iwmmxt_wstr): Likewise.
3760 (marvell_f_iwmmxt_wldr): Likewise.
3762 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
3764 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
3766 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3769 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3770 * lra-constraints.c (need_for_split_p): Check call used hard regs
3771 living through calls.
3773 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
3774 call used regs for call insn.
3776 2013-06-28 Jakub Jelinek <jakub@redhat.com>
3779 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
3780 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
3782 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
3784 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
3786 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3788 * lra-constraints.c (need_for_split_p): Check call used hard regs
3789 living through calls.
3791 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
3794 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
3795 to tie with any other modes. Eliminate Altivec vector mode tests,
3796 since these are a subset of ALTIVEC or VSX vector modes. Simplify
3797 code, to return 0 if testing MODE2 for a condition, if we've
3798 already tested MODE1 for the same condition.
3800 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3802 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
3805 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3807 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3808 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
3810 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3812 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
3814 (aarch64_symbolic_constant_p): Remove.
3815 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
3816 static. Fix line length and white space.
3817 (aarch64_symbolic_constant_p): Remove.
3818 * config/aarch64/predicates.md (aarch64_valid_symref):
3819 Use aarch64_classify_symbol_expression.
3821 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3823 * config/arm/constraints.md (Ts): New constraint.
3824 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
3826 (compare_scc): Use "Ts" constraint for operand 0.
3827 (ior_scc_scc): Likewise.
3828 (and_scc_scc): Likewise.
3829 (and_scc_scc_nodom): Likewise.
3830 (ior_scc_scc_cmp): Likewise for operand 7.
3831 (and_scc_scc_cmp): Likewise.
3832 * config/arm/thumb2.md (thumb2_movsi_insn):
3833 Add alternatives for 16-bit encodings.
3834 (thumb2_movhi_insn): Likewise.
3835 (thumb2_movsicc_insn): Likewise.
3836 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
3837 (thumb2_negscc): Use "Ts" constraint.
3838 Move mvn instruction outside cond_exec block.
3839 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
3840 for 16-bit encodings.
3842 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3844 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
3846 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
3847 (mulsi3subsi): Likewise.
3848 (mulsidi3adddi): Likewise.
3849 (mulsidi3_v6): Likewise.
3850 (umulsidi3_v6): Likewise.
3851 (umulsidi3adddi_v6): Likewise.
3852 (smulsi3_highpart_v6): Likewise.
3853 (umulsi3_highpart_v6): Likewise.
3854 (mulhisi3tb): Likewise.
3855 (mulhisi3bt): Likewise.
3856 (mulhisi3tt): Likewise.
3857 (maddhisi4): Likewise.
3858 (maddhisi4tb): Likewise.
3859 (maddhisi4tt): Likewise.
3860 (maddhidi4): Likewise.
3861 (maddhidi4tb): Likewise.
3862 (maddhidi4tt): Likewise.
3863 (zeroextractsi_compare0_scratch): Likewise.
3864 (insv_zero): Likewise.
3865 (insv_t2): Likewise.
3866 (anddi_notzesidi_di): Likewise.
3867 (anddi_notsesidi_di): Likewise.
3868 (andsi_notsi_si): Likewise.
3869 (iordi_zesidi_di): Likewise.
3870 (xordi_zesidi_di): Likewise.
3871 (andsi_iorsi3_notsi): Likewise.
3873 (smax_m1): Likewise.
3875 (not_shiftsi): Likewise.
3876 (unaligned_loadsi): Likewise.
3877 (unaligned_loadhis): Likewise.
3878 (unaligned_loadhiu): Likewise.
3879 (unaligned_storesi): Likewise.
3880 (unaligned_storehi): Likewise.
3881 (extv_reg): Likewise.
3882 (extzv_t2): Likewise.
3884 (udivsi3): Likewise.
3885 (arm_zero_extendhisi2addsi): Likewise.
3886 (arm_zero_extendqisi2addsi): Likewise.
3887 (compareqi_eq0): Likewise.
3888 (arm_extendhisi2_v6): Likewise.
3889 (arm_extendqisi2addsi): Likewise.
3890 (arm_movt): Likewise.
3891 (thumb2_ldrd): Likewise.
3892 (thumb2_ldrd_base): Likewise.
3893 (thumb2_ldrd_base_neg): Likewise.
3894 (thumb2_strd): Likewise.
3895 (thumb2_strd_base): Likewise.
3896 (thumb2_strd_base_neg): Likewise.
3897 (arm_negsi2): Add alternative for 16-bit encoding.
3898 (arm_one_cmplsi2): Likewise.
3900 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3902 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
3903 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
3904 (movdfcc): Likewise.
3905 * config/arm/vfp.md (*thumb2_movsf_vfp):
3906 Disable predication for arm_restrict_it.
3907 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
3908 (*thumb2_movdfcc_vfp): Likewise.
3909 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
3910 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
3911 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
3912 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
3913 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
3914 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
3915 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
3916 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
3917 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
3918 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
3919 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
3920 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
3921 Disable predication for arm_restrict_it.
3923 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
3925 * config/i386/bmiintrin.h (_bextr_u32): New.
3926 (_bextr_u64): Ditto.
3928 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
3930 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
3931 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
3932 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
3933 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
3934 * config/mips/n32-elf.h: ...this new file.
3936 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
3939 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
3940 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
3942 2013-06-27 Catherine Moore <clm@codesourcery.com>
3944 * config/mips/mips-tables.opt: Regenerate.
3945 * config/mips/mips-cpus.def: Add m14ke and m14kec.
3946 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
3947 * doc/invoke.texi: Add -m14kc.
3949 2013-06-27 Jakub Jelinek <jakub@redhat.com>
3952 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
3953 constraints of operand 1 and 2.
3956 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
3957 to match RTL canonicalization. Swap predicates and
3958 constraints of operand 1 and 2.
3960 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
3962 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
3963 Process OP_INOUT regs for splitting too.
3965 2013-06-27 Jakub Jelinek <jakub@redhat.com>
3967 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
3968 decl before the loop, initialize to NULL.
3969 (vectorizable_load): Initialize ptr_incr to NULL.
3971 2013-06-27 Martin Jambor <mjambor@suse.cz>
3974 * ipa-ref.h (ipa_maybe_record_reference): Declare.
3975 * ipa-ref.c (ipa_maybe_record_reference): New function.
3976 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
3977 * ipa-cp.c (create_specialized_node): Record potential references from
3979 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
3981 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
3983 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
3984 parameter 'mode' of type 'enum machine_mode mode'; change to pass
3985 'mode' to force_reg.
3986 (aarch64_add_offset): Update calls to aarch64_force_temporary.
3987 (aarch64_expand_mov_immediate): Likewise.
3989 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
3991 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
3992 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
3994 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3996 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
3997 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
3998 (struct machine_function): Add tbegin_p.
3999 (s390_canonicalize_comparison): Fold CC mode compares to
4000 conditional jump if possible.
4001 (s390_emit_jump): Return the emitted jump.
4002 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
4003 Handle CCRAWmode compares.
4004 (s390_option_override): Default to -mhtm if available.
4005 (s390_reg_clobbered_rtx): Handle floating point regs as well.
4006 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
4007 FPRs instead of df_regs_ever_live_p.
4008 (s390_optimize_nonescaping_tx): New function.
4009 (s390_init_frame_layout): Extend clobbered_regs array to cover
4011 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
4012 (s390_expand_tbegin): New function.
4013 (enum s390_builtin): New enum definition.
4014 (code_for_builtin): New array definition.
4015 (s390_init_builtins): New function.
4016 (s390_expand_builtin): New function.
4017 (TARGET_INIT_BUILTINS): Define.
4018 (TARGET_EXPAND_BUILTIN): Define.
4019 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
4020 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
4021 (s390_alc_comparison): Likewise.
4022 * config/s390/s390-modes.def: Add CCRAWmode.
4023 * config/s390/s390.h (processor_flags): Add PF_TX.
4024 (TARGET_CPU_HTM): Define macro.
4025 (TARGET_HTM): Define macro.
4026 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4027 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4028 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4029 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
4030 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
4031 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4032 ("*cc_to_int"): Move up.
4033 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4034 constants other than 0.
4035 ("*ccraw_to_int"): New insn and splitter definition.
4036 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4037 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4038 ("tx_assist"): New expander.
4039 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4040 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4041 * config/s390/s390.opt: Add -mhtm option.
4042 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4043 * config/s390/htmxlintrin.h: New file.
4044 * config/s390/htmintrin.h: New file.
4045 * config/s390/s390intrin.h: New file.
4046 * doc/extend.texi: Document htm builtins.
4047 * config.gcc: Add the new header files to extra_headers.
4049 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4051 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
4052 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
4054 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4055 Pat Haugen <pthaugen@us.ibm.com>
4056 Peter Bergner <bergner@vnet.ibm.com>
4058 * config/rs6000/power8.md: New.
4059 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4060 setting for power8 entry.
4061 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4062 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4063 test for Power4/Power5 only.
4064 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4066 (force_new_group): Adjust comment.
4067 * config/rs6000/rs6000.md: Include power8.md.
4069 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4071 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4072 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4074 (tune_params): Update comment.
4075 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4076 (arm_max_conditional_execute): New function.
4077 (thumb2_final_prescan_insn): Use max_insn_skipped and
4078 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4080 2013-06-25 Jakub Jelinek <jakub@redhat.com>
4082 PR tree-optimization/57705
4083 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4084 SSA_NAME step, provided that it is not defined inside the loop.
4085 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
4086 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4088 2013-06-25 Martin Jambor <mjambor@suse.cz>
4091 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4092 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4094 (ipa_note_param_call): Initialize member_ptr flag.
4095 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4096 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4097 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4098 (ipa_read_indirect_edge_info): Likewise.
4100 2013-06-25 Richard Biener <rguenther@suse.de>
4103 * passes.c (init_optimization_passes): Move pass_fold_builtins
4104 and pass_dce earlier with -Og.
4106 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4108 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4109 <BIT_FIELD_REF>: Remove trailing TAB.
4110 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4113 2013-06-24 Martin Jambor <mjambor@suse.cz>
4115 PR tree-optimization/57358
4116 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4117 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4118 (ipa_analyze_params_uses): Generate pessimistic info when true.
4120 2013-06-24 Martin Jambor <mjambor@suse.cz>
4122 PR tree-optimization/57539
4123 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4124 global.inlined_to of the new node to it. All callers changed.
4125 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4126 inlining_into, pass it to cgraph_clone_node.
4127 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4128 ipa_free_edge_args_substructures.
4129 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4130 rdesc linked list. Do not assert rdesc edges have inlined caller.
4131 Assert we have found an rdesc in the rdesc list.
4133 2013-06-24 Richard Biener <rguenther@suse.de>
4135 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
4136 (pointer_set_lookup): Declare.
4137 (class pointer_map): New template class implementing a
4138 generic pointer to T map.
4139 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4140 pointer_map<T>::contains, pointer_map<T>::insert,
4141 pointer_map<T>::traverse): New functions.
4142 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4143 (pointer_set_lookup): New function.
4144 (pointer_set_contains): Use pointer_set_lookup.
4145 (pointer_set_insert): Likewise.
4146 (insert_aux): Remove.
4147 (struct pointer_map_t): Embed a pointer_set_t.
4148 (pointer_map_create): Adjust.
4149 (pointer_map_destroy): Likewise.
4150 (pointer_map_contains): Likewise.
4151 (pointer_map_insert): Likewise.
4152 (pointer_map_traverse): Likewise.
4153 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4154 pointer_map<unsigned> instead of a pointer_map_t.
4155 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4156 (streamer_tree_cache_lookup): Likewise.
4157 (streamer_tree_cache_create): Likewise.
4158 (streamer_tree_cache_delete): Likewise.
4159 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4160 pointer_map<unsigned> instead of a pointer_map_t.
4161 (lto_init_tree_ref_encoder): Adjust.
4162 (lto_destroy_tree_ref_encoder): Likewise.
4163 * lto-section-out.c (lto_output_decl_index): Likewise.
4164 (lto_record_function_out_decl_state): Likewise.
4165 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4167 2013-06-24 Richard Biener <rguenther@suse.de>
4169 PR tree-optimization/57488
4170 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4172 2013-06-24 Alan Modra <amodra@gmail.com>
4174 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4175 (gen_easy_altivec_constant): Likewise.
4176 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4177 easy_vector_constant_msb): Likewise.
4179 2013-06-23 Jakub Jelinek <jakub@redhat.com>
4182 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4183 add missing return true.
4185 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4188 * config/sh/predicates.md (general_extend_operand): Invoke
4189 general_movsrc_operand for memory operands.
4190 (general_movsrc_operand): Allow reg+reg addressing, do not use
4191 general_operand for memory operands.
4193 2013-06-23 Sriraman Tallam <tmsriram@google.com>
4195 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4196 when current target options does not apply.
4197 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4198 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
4199 * config/i386/bmiintrin.h: Pass appropriate target
4200 attributes to header.
4201 * config/i386/mmintrin.h: Ditto.
4202 * config/i386/nmmintrin.h: Ditto.
4203 * config/i386/avx2intrin.h: Ditto.
4204 * config/i386/fxsrintrin.h: Ditto.
4205 * config/i386/tbmintrin.h: Ditto.
4206 * config/i386/xsaveintrin.h: Ditto.
4207 * config/i386/f16cintrin.h: Ditto.
4208 * config/i386/xtestintrin.h: Ditto.
4209 * config/i386/xsaveoptintrin.h: Ditto.
4210 * config/i386/bmi2intrin.h: Ditto.
4211 * config/i386/lzcntintrin.h: Ditto.
4212 * config/i386/smmintrin.h: Ditto.
4213 * config/i386/wmmintrin.h: Ditto.
4214 * config/i386/x86intrin.h: Remove all header include guards.
4215 * config/i386/prfchwintrin.h: Ditto.
4216 * config/i386/pmmintrin.h: Ditto.
4217 * config/i386/tmmintrin.h: Ditto.
4218 * config/i386/xmmintrin.h: Ditto.
4219 * config/i386/popcntintrin.h: Ditto.
4220 * config/i386/rdseedintrin.h: Ditto.
4221 * config/i386/ammintrin.h: Ditto.
4222 * config/i386/emmintrin.h: Ditto.
4223 * config/i386/immintrin.h: Remove all header include guards.
4224 * config/i386/fma4intrin.h: Ditto.
4225 * config/i386/lwpintrin.h: Ditto.
4226 * config/i386/xopintrin.h: Ditto.
4227 * config/i386/ia32intrin.h: Ditto.
4228 * config/i386/avxintrin.h: Ditto.
4229 * config/i386/rtmintrin.h: Ditto.
4230 * config/i386/fmaintrin.h: Ditto.
4231 * config/i386/mm3dnow.h: Ditto.
4233 2013-06-22 Sriraman Tallam <tmsriram@google.com>
4235 * common/config/i386/i386-common.c: Handle LZCNT.
4237 2013-06-22 Andi Kleen <ak@linux.intel.com>
4239 * doc/extend.texi: Use __atomic_store_n instead of
4240 __atomic_store in HLE example.
4242 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4244 * config/sh/sh.c: Remove <cstdlib> workaround.
4246 2013-06-21 Andi Kleen <ak@linux.intel.com>
4248 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
4250 2013-06-21 Andi Kleen <ak@linux.intel.com>
4252 * doc/extend.texi: Document that __atomic_clear and
4253 __atomic_test_and_set should only be used with bool.
4255 2013-06-20 Jan Hubicka <jh@suse.cz>
4257 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4259 * tree.c (decls_same_for_odr): New function.
4260 (same_for_edr): New function.
4261 (types_same_for_odr): New function.
4262 (get_binfo_at_offset): Use it.
4263 * tree.h (types_same_for_odr): Declare.
4265 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
4266 Jason Merrill <jason@redhat.com>
4268 * system.h: Include <cstdlib> as well as <stdlib.h>.
4270 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
4273 * config/i386/i386.c (construct_container): Report error if
4274 long double is used with disabled x87 float returns.
4276 2013-06-20 Jan Hubicka <jh@suse.cz>
4278 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4280 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4282 PR rtl-optimization/57425
4283 PR rtl-optimization/57569
4284 * alias.c (write_dependence_p): Remove parameters mem_mode and
4285 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4286 Changed all callers.
4287 (canon_anti_dependence): Get comments and semantics in sync.
4288 Add parameter mem_canonicalized. Changed all callers.
4289 * rtl.h (canon_anti_dependence): Update prototype.
4291 2013-06-20 Richard Biener <rguenther@suse.de>
4293 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4294 byte case, inline streamer_read_uchar and defer section
4297 2013-06-20 Richard Biener <rguenther@suse.de>
4299 PR tree-optimization/57584
4300 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4301 SSA names into the expanded expression that take part in
4302 abnormal coalescing.
4304 2013-06-19 Sharad Singhai <singhai@google.com>
4306 * gcov.c (print_usage): Handle new option.
4307 (process_args): Ditto.
4308 (get_gcov_intermediate_filename): New function.
4309 (output_intermediate_file): New function.
4310 (output_gcov_file): New function
4311 (generate_results): Handle new option.
4312 (release_function): Relase demangled name.
4313 (read_graph_file): Handle demangled name.
4314 (output_lines): Ditto.
4315 * doc/gcov.texi: Document gcov intermediate format.
4317 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4320 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4321 (lra_emit_add): Use the functions. Add comment about Y as an
4324 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
4327 * collect2.c (collect_atexit): New.
4328 (collect_exit): Delete.
4329 (main): Register collect_atexit with atexit.
4330 (collect_wait): Change collect_exit to exit.
4332 * collect2.h (collect_exit): Delete.
4333 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4335 2013-06-19 Wei Mi <wmi@google.com>
4337 PR rtl-optimization/57518
4338 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4339 if regno is used in paradoxical subreg.
4340 (update_equiv_regs): Check pdx_subregs[regno] before
4341 set a reg to be equivalent with a mem.
4343 2013-06-19 Matthias Klose <doko@ubuntu.com>
4346 * file-find.h (find_a_file): Add a mode parameter.
4347 * file-find.c (find_a_file): Likewise.
4348 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4349 with X_OK for the executables.
4350 * collect2.c (main): Call find_a_file with X_OK.
4352 2013-06-19 Steve Ellcey <sellcey@mips.com>
4355 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4356 Use NEXT_INSN instead of next_real_insn.
4358 2013-06-19 Jan Hubicka <jh@suse.cz>
4360 * cgraph.h (const_value_known_p): Replace by ...
4361 (ctor_for_folding): .. this one.
4362 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4363 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4364 * expr.c (expand_expr_real_1): Likewise.
4365 (string_constant): Likewise.
4366 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4367 * ipa.c (process_references): Likewise.
4368 (symtab_remove_unreachable_nodes): Likewise.
4369 * ipa-inline-analysis.c (param_change_prob): Likewise.
4370 * gimple-fold.c (canonicalize_constructor_val): Likewise.
4371 (get_base_constructor): Likwise.
4372 * varpool.c (varpool_remove_node): Likewise.
4373 (varpool_remove_initializer): LIkewise.
4374 (dump_varpool_node): LIkwise.
4375 (const_value_known_p): Rewrite to ...
4376 (ctor_for_folding): ... this one.
4378 2013-06-19 Jakub Jelinek <jakub@redhat.com>
4381 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
4382 PERSONALITY in $PATH derived prefixes.
4384 2013-06-19 Jeff Law <law@redhat.com>
4386 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
4389 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
4390 (simplify_bitwise_binary): Use it to simpify certain binary ops on
4393 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
4395 * config/arm/vfp.md: Move VFP instruction classification documentation
4397 * config/arm/arm.md: ... here. Update instruction classification
4400 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
4402 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
4403 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
4404 pattern. Use more efficient sequences on ARMv5 and Thumb2.
4406 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
4409 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4410 with NEXT_INSN. Use tablejump_p to check for jump table data
4413 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
4416 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
4417 that now in C++ the value is correct per the C++ standards.
4419 2013-06-19 Richard Biener <rguenther@suse.de>
4421 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
4424 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4427 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4430 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4431 with next_active_insn.
4433 2013-06-18 Sriraman Tallam <tmsriram@google.com>
4435 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
4436 functions are inlined during failures to flag an error.
4437 * tree-inline.c (expand_call_inline): Allow the error to be flagged
4438 in early inline pass.
4440 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
4442 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
4445 2013-06-18 Julian Brown <julian@codesourcery.com>
4447 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
4448 Permit virtual register pre-reload if !strict.
4449 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
4451 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
4453 * config/arm/neon.md (movmisalign<mode>): Use
4454 neon_perm_struct_or_reg_operand instead of
4455 neon_struct_or_register_operand.
4456 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
4457 neon_permissive_struct_operand instead of neon_struct_operand.
4458 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
4459 neon_vector_mem_operand.
4460 * config/arm/predicates.md (neon_struct_operand): Adjust call to
4461 neon_vector_mem_operand.
4462 (neon_permissive_struct_operand): New.
4463 (neon_struct_or_register_operand): Rename to...
4464 (neon_perm_struct_or_reg_operand): This. Adjust call to
4465 neon_vector_mem_operand.
4467 2013-06-18 Richard Biener <rguenther@suse.de>
4469 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
4470 * lto-streamer.h: Include pointer-set.h.
4471 (struct lto_decl_slot): Remove.
4472 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
4473 Remove next_index entry.
4474 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4475 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4476 (lto_init_tree_ref_encoder): Adjust.
4477 (lto_destroy_tree_ref_encoder): Likewise.
4478 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4479 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4480 (lto_output_decl_index): Adjust.
4481 (lto_new_out_decl_state): Likewise.
4482 (lto_record_function_out_decl_state): Likewise.
4483 * lto-streamer-out.c (copy_function): Likewise.
4485 2013-06-18 Richard Biener <rguenther@suse.de>
4487 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
4488 * cgraphunit.c: Include cfgloop.h.
4489 (init_lowered_empty_function): Initialize the loop tree.
4490 (assemble_thunk): Insert new BBs into loops.
4492 2013-06-18 Richard Biener <rguenther@suse.de>
4494 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
4495 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
4496 the map from cache entry to cache index optional.
4497 (streamer_tree_cache_replace_tree): Adjust accordingly.
4498 (streamer_tree_cache_append): Likewise.
4499 (streamer_tree_cache_delete): Likewise.
4500 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
4501 streamer cache map from cache entry to cache index.
4502 * lto-streamer-out.c (create_output_block): Adjust.
4504 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
4506 * config/arm/arm.md (attribute "insn"): Move multiplication and
4507 division attributes to...
4508 (attribute "type"): ... here. Remove mult.
4509 (attribute "mul32"): New attribute.
4510 (attribute "mul64"): Add umaal.
4511 (*arm_mulsi3): Update attributes.
4512 (*arm_mulsi3_v6): Likewise.
4513 (*thumb_mulsi3): Likewise.
4514 (*thumb_mulsi3_v6): Likewise.
4515 (*mulsi3_compare0): Likewise.
4516 (*mulsi3_compare0_v6): Likewise.
4517 (*mulsi_compare0_scratch): Likewise.
4518 (*mulsi_compare0_scratch_v6): Likewise.
4519 (*mulsi3addsi): Likewise.
4520 (*mulsi3addsi_v6): Likewise.
4521 (*mulsi3addsi_compare0): Likewise.
4522 (*mulsi3addsi_compare0_v6): Likewise.
4523 (*mulsi3addsi_compare0_scratch): Likewise.
4524 (*mulsi3addsi_compare0_scratch_v6): Likewise.
4525 (*mulsi3subsi): Likewise.
4526 (*mulsidi3adddi): Likewise.
4527 (*mulsi3addsi_v6): Likewise.
4528 (*mulsidi3adddi_v6): Likewise.
4529 (*mulsidi3_nov6): Likewise.
4530 (*mulsidi3_v6): Likewise.
4531 (*umulsidi3_nov6): Likewise.
4532 (*umulsidi3_v6): Likewise.
4533 (*umulsidi3adddi): Likewise.
4534 (*umulsidi3adddi_v6): Likewise.
4535 (*smulsi3_highpart_nov6): Likewise.
4536 (*smulsi3_highpart_v6): Likewise.
4537 (*umulsi3_highpart_nov6): Likewise.
4538 (*umulsi3_highpart_v6): Likewise.
4539 (mulhisi3): Likewise.
4540 (*mulhisi3tb): Likewise.
4541 (*mulhisi3bt): Likewise.
4542 (*mulhisi3tt): Likewise.
4543 (maddhisi4): Likewise.
4544 (*maddhisi4tb): Likewise.
4545 (*maddhisi4tt): Likewise.
4546 (maddhidi4): Likewise.
4547 (*maddhidi4tb): Likewise.
4548 (*maddhidi4tt): Likewise.
4550 (udivsi3): Likewise.
4551 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
4552 (thumb2_mulsi_short_compare0): Likewise.
4553 (thumb2_mulsi_short_compare0_scratch): Likewise.
4554 * config/arm/arm1020e.md (1020mult1): Update attribute change.
4555 (1020mult2): Likewise.
4556 (1020mult3): Likewise.
4557 (1020mult4): Likewise.
4558 (1020mult5): Likewise.
4559 (1020mult6): Likewise.
4560 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
4562 (cortex_a15_mult64): Likewise.
4563 (cortex_a15_sdiv): Likewise.
4564 (cortex_a15_udiv): Likewise.
4565 * config/arm/arm1026ejs.md (mult1): Update attribute change.
4571 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
4572 (pj4_ir_div): Likewise.
4573 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
4574 (11_mult2): Likewise.
4575 (11_mult3): Likewise.
4576 (11_mult4): Likewise.
4577 (11_mult5): Likewise.
4578 (11_mult6): Likewise.
4579 (11_mult7): Likewise.
4580 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
4581 (cortex_a8_mla): Likewise.
4582 (cortex_a8_mull): Likewise.
4583 (cortex_a8_smulwy): Likewise.
4584 (cortex_a8_smlald): Likewise.
4585 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
4586 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
4587 (cortex_r4_mul_3): Likewise.
4588 (cortex_r4_mla_4): Likewise.
4589 (cortex_r4_mla_3): Likewise.
4590 (cortex_r4_smlald): Likewise.
4591 (cortex_r4_mull): Likewise.
4592 (cortex_r4_sdiv): Likewise.
4593 (cortex_r4_udiv): Likewise.
4594 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
4595 (cortex_a7_idiv): Likewise.
4596 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
4597 (9_mult2): Likewise.
4598 (9_mult3): Likewise.
4599 (9_mult4): Likewise.
4600 (9_mult5): Likewise.
4601 (9_mult6): Likewise.
4602 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
4603 (cortex_a53_sdiv): Likewise.
4604 (cortex_a53_udiv): Likewise.
4605 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
4606 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
4607 (mp626_mult2): Likewise.
4608 (mp626_mult3): Likewise.
4609 (mp626_mult4): Likewise.
4610 * config/arm/fa526.md (526_mult1): Update attribute change.
4611 (526_mult2): Likewise.
4612 * config/arm/arm-generic.md (mult): Update attribute change.
4613 (mult_ldsched_strongarm): Likewise.
4614 (mult_ldsched): Likewise.
4615 (multi_cycle): Likewise.
4616 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
4617 * config/arm/fa606te.md (606te_mult1): Update attribute change.
4618 (606te_mult2): Likewise.
4619 (606te_mult3): Likewise.
4620 (606te_mult4): Likewise.
4621 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
4622 (cortex_a9_mac16): Likewise.
4623 (cortex_a9_multiply): Likewise.
4624 (cortex_a9_mac): Likewise.
4625 (cortex_a9_multiply_long): Likewise.
4626 * config/arm/fa626te.md (626te_mult1): Update attribute change.
4627 (626te_mult2): Likewise.
4628 (626te_mult3): Likewise.
4629 (626te_mult4): Likewise.
4631 2013-06-18 Richard Biener <rguenther@suse.de>
4634 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
4636 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4639 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4640 with next_active_insn.
4642 2013-06-18 Alan Modra <amodra@gmail.com>
4644 * config/rs6000/rs6000.h (enum data_align): New.
4645 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
4646 (DATA_ABI_ALIGNMENT): Define.
4647 (CONSTANT_ALIGNMENT): Correct comment.
4648 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
4649 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
4651 2013-06-17 David Malcolm <dmalcolm@redhat.com>
4653 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
4654 ATTRIBUTE_UNUSED marking.
4656 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
4658 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
4659 alternative and update.
4660 (aarch64_dup_lanedi): Delete.
4661 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
4662 * config/aarch64/aarch64-simd-builtins.def: Update.
4664 2013-06-17 Richard Biener <rguenther@suse.de>
4666 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
4667 (lto_input_scc): Declare.
4668 (lto_input_tree_1): Likewise.
4669 (struct lto_stats_d): Add num_tree_bodies_output and
4670 num_pickle_refs_output.
4671 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
4672 (lto_read_tree_1): Split out from ...
4673 (lto_read_tree): ... this.
4674 (lto_input_scc): New function.
4675 (lto_input_tree_1): Split out from ...
4676 (lto_input_tree): ... this. Handle LTO_tree_scc.
4677 (lto_data_in_create): Create the streamer cache without hashes.
4678 * lto-streamer-out.c (create_output_block): Create the streamer
4679 cache with hashes when not doing WPA.
4680 (lto_write_tree_1): Split out from ...
4681 (lto_write_tree): ... this.
4682 (get_symbol_initial_value): New function.
4683 (lto_output_tree_1): Split out from ...
4684 (lto_output_tree): ... this. Write trees as series of SCCs
4685 using a DFS walk via DFS_write_tree.
4686 (struct sccs, struct scc_entry): New types.
4687 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
4688 (DFS_write_tree_body): New function.
4689 (DFS_write_tree): Likewise.
4690 (hash_tree): Likewise.
4691 (scc_entry_compare): Likewise.
4692 (hash_scc): Likewise.
4693 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
4694 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
4695 TREE_CHAIN as regular reference.
4696 (streamer_read_integer_cst): Remove.
4697 (streamer_get_pickled_tree): Adjust.
4698 * tree-streamer-out.c (streamer_write_chain): Disable streaming
4699 of DECL_EXTERNALs in BLOCK_VARS for now.
4700 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
4702 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
4703 Add hash value argument and record that if hashes are recorded
4705 (streamer_tree_cache_insert_1): Adjust.
4706 (streamer_tree_cache_insert): Likewise.
4707 (streamer_tree_cache_insert_at): Rename to ...
4708 (streamer_tree_cache_replace_tree): ... this and adjust.
4709 (streamer_tree_cache_append): Adjust.
4710 (record_common_node): Likewise.
4711 (streamer_tree_cache_create): Add argument whether to
4712 record hash values together with trees.
4713 (streamer_tree_cache_delete): Adjust.
4714 * tree-streamer.h (struct streamer_tree_cache_d): Add
4716 (streamer_read_integer_cst): Remove.
4717 (streamer_tree_cache_insert): Adjust.
4718 (streamer_tree_cache_append): Likewise.
4719 (streamer_tree_cache_insert_at): Rename to ...
4720 (streamer_tree_cache_replace_tree): ... this and adjust.
4721 (streamer_tree_cache_create): Add argument whether to record hashes.
4722 (streamer_tree_cache_get): Rename to ...
4723 (streamer_tree_cache_get_tree): ... this.
4724 (streamer_tree_cache_get_hash): New function.
4725 * tree.c (cache_integer_cst): New function.
4726 * tree.h (cache_integer_cst): Declare.
4727 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
4728 * lto-symtab.c (lto_varpool_replace_node): Only release
4729 DECL_INITIAL of non-prevailing decls.
4730 * varpool.c (varpool_remove_initializer): Do not release
4731 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
4733 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4735 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
4736 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
4737 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
4738 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
4739 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
4740 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
4741 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
4742 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
4743 instead of TARGET_64BIT.
4744 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
4745 Require ISA_HAS_<D>DIV.
4747 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
4749 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
4750 (mips*-*-linux*): Move default with_llsc setting to where other
4752 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
4754 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
4755 Likewise. Remove default with_tune setting. Move default float
4756 setting to its own block. Handle with_llsc in the same block as above.
4758 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
4760 PR rtl-optimization/57425
4761 PR rtl-optimization/57569
4762 * alias.c (write_dependence_p): Add new parameters mem_mode,
4763 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
4764 Changed all callers.
4765 (canon_anti_dependence): New function.
4766 * cse.c (check_dependence): Use canon_anti_dependence.
4767 * cselib.c (cselib_invalidate_mem): Likewise.
4768 * rtl.h (canon_anti_dependence): Declare.
4770 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4772 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
4773 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
4774 ".set mips3" for 64-bit targets.
4776 2013-06-15 Dehao Chen <dehao@google.com>
4778 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
4779 * gimple-low.c (gimple_check_call_matching_types): Likewise.
4780 (gimple_check_call_args): Likewise.
4781 * value-prof.c (check_ic_target): Likewise.
4782 * ipa-inline.c (early_inliner): Likewise.
4783 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
4784 * cgraph.c (cgraph_create_edge_1): Likewise.
4785 (cgraph_make_edge_direct): Likewise.
4787 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
4790 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
4791 rs6000_output_move_128bit to handle emitting quad memory
4792 operations. Set attribute length to 8 bytes.
4794 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
4796 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
4798 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
4799 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
4800 (aarch64_<su>mlsl<mode>): Likewise.
4802 2013-06-14 Mike Stump <mikestump@comcast.net>
4804 * Makefile.in (TARGET_H): Add insn-codes.h.
4806 2013-06-14 Alan Modra <amodra@gmail.com>
4810 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
4811 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
4812 bitfield expansion when EXPAND_MEMORY.
4813 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
4815 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
4817 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
4818 test for clearing quad memory on 32-bit later.
4820 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
4822 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
4823 (fold_negate_expr): Likewise.
4824 (fold_real_zero_addition_p): Handle vectors.
4825 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
4827 2013-06-14 Alan Modra <amodra@gmail.com>
4829 * varasm.c (force_const_mem): Revert 2013-06-07 change.
4831 2013-06-13 Jan Hubicka <jh@suse.cz>
4833 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
4834 Local comdats are not externally visible.
4835 * symtab.c (dump_symtab_base): Dump externally visible.
4836 (verify_symtab_base): Verify back links in the symtab hash.
4838 2013-06-13 Bin Cheng <bin.cheng@arm.com>
4840 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
4841 CONVERT_EXPR as equal nodes.
4843 2013-06-13 Bin Cheng <bin.cheng@arm.com>
4845 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
4847 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
4849 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
4850 Generalize to complex and vector.
4851 * tree.c (build_all_ones_cst): New function.
4852 * tree.h (build_all_ones_cst): Declare it.
4854 2013-06-13 Alan Modra <amodra@gmail.com>
4856 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
4857 * config/rs6000/rs6000.md (signbittf2): New insn.
4858 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
4859 (abstf2_internal, cmptf_internal2): Likewise.
4860 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
4862 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4863 Pat Haugen <pthaugen@us.ibm.com>
4864 Peter Bergner <bergner@vnet.ibm.com>
4866 * config/rs6000/rs6000.c (emit_load_locked): Add support for
4867 power8 byte, half-word, and quad-word atomic instructions.
4868 (emit_store_conditional): Likewise.
4869 (rs6000_expand_atomic_compare_and_swap): Likewise.
4870 (rs6000_expand_atomic_op): Likewise.
4872 * config/rs6000/sync.md (larx): Add new modes for power8.
4874 (AINT): New mode iterator to include TImode as well as normal
4875 integer modes on power8.
4876 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
4877 that VSX registers are not considered. Use AINT mode iterator
4878 instead of INT1 to allow inclusion of quad word atomic operations
4880 (load_locked<mode>): Likewise.
4881 (store_conditional<mode>): Likewise.
4882 (atomic_compare_and_swap<mode>): Likewise.
4883 (atomic_exchange<mode>): Likewise.
4884 (atomic_nand<mode>): Likewise.
4885 (atomic_fetch_<fetchop_name><mode>): Likewise.
4886 (atomic_nand_fetch<mode>): Likewise.
4887 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
4889 (ATOMIC): On power8, add QImode, HImode modes.
4890 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
4891 modes that promote to SImode.
4892 (load_lockedti): Convert TImode arguments to PTImode, so that we
4893 get a guaranteed even/odd register pair.
4894 (load_lockedpti): Likewise.
4895 (store_conditionalti): Likewise.
4896 (store_conditionalpti): Likewise.
4898 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
4899 atomic load/store instructions.
4902 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
4904 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
4906 (insn_count): New attribute, with most cases extracted from...
4907 (length): ...here. Redefine most cases in terms of insn_count.
4908 (single_insn): Delete.
4909 (can_delay): Use insn_count to check for single instructions.
4910 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
4911 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
4912 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
4913 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
4914 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
4915 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
4916 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
4917 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
4918 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
4919 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
4920 rather than "length".
4921 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
4922 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
4923 Use "insn_count" rather than "length".
4924 * config/mips/mips-dsp.md
4925 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
4926 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
4929 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
4931 PR tree-optimization/57361
4932 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
4934 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
4936 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
4938 (aarch64_simd_combine<mode>): New instruction expansion.
4939 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
4941 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
4942 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
4944 2013-06-12 Jan Hubicka <jh@suse.cz>
4946 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
4947 decl has when in streaming stage.
4948 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
4949 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
4951 2013-06-12 Roland Stigge <stigge@antcom.de>
4954 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
4956 2013-06-12 Jakub Jelinek <jakub@redhat.com>
4958 PR tree-optimization/57537
4959 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
4960 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
4962 2013-06-12 Richard Biener <rguenther@suse.de>
4964 * data-streamer.h (streamer_write_char_stream): CSE
4965 obs->current_pointer.
4966 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
4967 streamer_write_char_stream manually and optimize the resulting loop.
4968 (streamer_write_hwi_stream): Likewise.
4970 2013-06-12 Jan Hubicka <jh@suse.cz>
4972 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
4973 * cgraph.h (varpool_create_empty_node): Declare.
4974 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
4976 * symtab.c (symtab_unregister_node): Be lax about missin entries
4978 (symtab_get_node): Update comment.
4979 * varpool.c (varpool_create_empty_node): Break out from ...
4980 (varpool_node_for_decl): ... here.
4981 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
4983 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
4985 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
4986 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
4987 part. Use straight-line flow at the end.
4988 <COMPONENT_REF>: Remove superfluous else.
4989 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
4991 2013-06-12 Jakub Jelinek <jakub@redhat.com>
4994 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
4995 target hook even for !TREE_PUBLIC decls. If no resolution info
4996 is available, return false for common and external decls.
4998 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5000 * config/rl78/constraints.md (U): New constraint.
5001 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
5004 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5007 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
5008 to allow returning address to AT_PLATFORM name.
5010 2013-06-11 Jan Hubicka <jh@suse.cz>
5012 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
5013 * cgraph.h (symtab_node_base): Add weakref flag.
5014 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
5015 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
5016 (output_weakrefs): Use weakref flag.
5017 * fold-const.c (simple_operand_p): Handle WEAK.
5018 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
5019 * ipa.c (varpool_externally_visible_p): Drop weakref.
5020 (function_and_variable_visibility): Update comment; fix weakref
5021 sanity checks; do not clear DECL_WEAK on them.
5022 * lto-cgraph.c (lto_output_node): update.
5023 (lto_output_varpool_node): Update.
5024 (input_overwrite_node): Update.
5025 (input_node): Update.
5026 (input_varpool_node): Update.
5027 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5028 (lto_symtab_merge_symbols): Add sanity check.
5029 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5030 * passes.c (rest_of_decl_compilation): Set static flag, too.
5031 * symtab.c (dump_symtab_base): Dump weakref.
5032 (verify_symtab_base): Sanity check weakrefs.
5033 (symtab_make_decl_local): Remove duplicated code.
5034 (symtab_alias_ultimate_target): Simplify.
5035 * varpool.c (varpool_create_variable_alias): Set weakref flag.
5037 2013-06-11 DJ Delorie <dj@redhat.com>
5039 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5040 (rl78_unwind_word_mode): New.
5042 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5044 * final.c (debug_prefix_maps): Make static.
5046 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5048 * function.c (initial_trampoline): Remove stray copy.
5050 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5052 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5054 2013-06-11 Martin Jambor <mjambor@suse.cz>
5056 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5057 within bounds at the beginning of the function.
5059 2013-06-11 Alan Modra <amodra@gmail.com>
5061 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5063 (get_named_section): Don't NULL !DECL_P decl.
5065 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5067 * doc/invoke.texi (core-avx2): Document.
5069 (atom): Updated with MOVBE.
5071 2013-06-11 Richard Biener <rguenther@suse.de>
5073 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
5075 2013-06-11 Anton Blanchard <anton@samba.org>
5077 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5078 correct shift value in little-endian mode.
5080 2013-06-11 Jakub Jelinek <jakub@redhat.com>
5083 * varasm.c (get_variable_align): Move #endif to the right place.
5085 2013-06-10 Cary Coutant <ccoutant@google.com>
5087 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5088 for hash so that hash table traversal order is deterministic.
5090 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5091 Pat Haugen <pthaugen@us.ibm.com>
5092 Peter Bergner <bergner@vnet.ibm.com>
5094 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5095 of vectors in GPRS if they are direct moves or quad word load or
5098 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5100 (direct_move_p): Likewise.
5101 (quad_load_store_p): Likewise.
5103 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5104 classes into bins based on the physical register type.
5105 (reg_class_to_reg_type): Likewise.
5106 (IS_STD_REG_TYPE): Likewise.
5107 (IS_FP_VECT_REG_TYPE): Likewise.
5108 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5109 use direct move instructions.
5110 (reload_gpr_vsx): Likewise.
5111 (reload_vsx_gpr): Likewise.
5112 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5113 information that is a simplification of register classes. Also
5114 precalculate direct move reload helpers.
5115 (direct_move_p): New function to return true if the operation can
5116 be done as a direct move instruciton.
5117 (quad_load_store_p): New function to return true if the operation
5118 is a quad memory operation.
5119 (rs6000_legitimize_address): If quad memory, only allow register
5120 indirect for TImode addresses.
5121 (rs6000_legitimate_address_p): Likewise.
5122 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5123 (rs6000_reload_register_type): Likewise.
5124 (register_to_reg_type): Return register type.
5125 (rs6000_secondary_reload_simple_move): New helper function for
5126 secondary reload and secondary memory needed to identify anything
5127 that is a simple move, and does not need reloading.
5128 (rs6000_secondary_reload_direct_move): New helper function for
5129 secondary reload to identify cases that can be done with several
5130 instructions via the direct move instructions.
5131 (rs6000_secondary_reload_move): New helper function for secondary
5132 reload to identify moves between register types that can be done.
5133 (rs6000_secondary_reload): Add support for quad memory operations
5134 and for direct move.
5135 (rs6000_secondary_memory_needed): Likewise.
5136 (rs6000_debug_secondary_memory_needed): Change argument names.
5137 (rs6000_output_move_128bit): New function to return the move to
5138 use for 128-bit moves, including knowing about the various
5139 limitations of quad memory operations.
5141 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5142 memory operations. call rs6000_output_move_128bit for the actual
5143 instruciton(s) to generate.
5144 (vsx_movti_64bit): Likewise.
5146 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5147 (UNSPEC_P8V_MTVSRWZ): Likewise.
5148 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5149 (UNSPEC_P8V_MTVSRD): Likewise.
5150 (UNSPEC_P8V_XXPERMDI): Likewise.
5151 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5152 (UNSPEC_FUSION_GPR): Likewise.
5153 (FMOVE128_GPR): New iterator for direct move.
5154 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
5157 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5158 loads and direct move instructions.
5159 (zero_extendsidi2_lfiwzx): Likewise.
5160 (extendsidi2_lfiwax): Likewise.
5161 (extendsidi2_nocell): Likewise.
5162 (floatsi<mode>2_lfiwax): Likewise.
5164 (floatunssi<mode>2_lfiwzx): Likewise.
5166 (fix_trunc<mode>_stfiwx): Likewise.
5167 (fixuns_trunc<mode>_stfiwx): Likewise.
5168 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5169 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5170 (parity<mode>2_cmpb): Set length/type attr.
5171 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5172 for 'mr.' to fast_compare.
5173 (bpermd_<mode>): Change type attr to popcnt.
5174 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5175 (p8_mtvsrwz_1): Likewise.
5176 (p8_mtvsrwz_2): Likewise.
5177 (reload_fpr_from_gpr<mode>): Likewise.
5178 (p8_mtvsrd_1): Likewise.
5179 (p8_mtvsrd_2): Likewise.
5180 (p8_xxpermdi_<mode>): Likewise.
5181 (reload_vsx_from_gpr<mode>): Likewise.
5182 (reload_vsx_from_gprsf): Likewise.
5183 (p8_mfvsrd_3_<mode>): LIkewise.
5184 (reload_gpr_from_vsx<mode>): Likewise.
5185 (reload_gpr_from_vsxsf): Likewise.
5186 (p8_mfvsrd_4_disf): Likewise.
5187 (multi-word GPR splits): Do not split direct moves or quad memory
5190 2013-06-10 David Malcolm <dmalcolm@redhat.com>
5192 * tree-into-ssa.c (interesting_blocks): Make static.
5194 2013-06-10 Jakub Jelinek <jakub@redhat.com>
5197 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5198 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5199 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5200 (get_variable_align): New function.
5201 (get_variable_section, emit_bss, emit_common,
5202 assemble_variable_contents, place_block_symbol): Use
5203 get_variable_align instead of DECL_ALIGN.
5204 (assemble_noswitch_variable): Add align argument, use it
5205 instead of DECL_ALIGN.
5206 (assemble_variable): Adjust caller. Use get_variable_align
5207 instead of DECL_ALIGN.
5208 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5210 (DATA_ABI_ALIGNMENT): Define.
5211 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5212 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5213 opt is false, only return the psABI mandated alignment increase.
5214 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5215 (DATA_ABI_ALIGNMENT): ... this.
5216 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5217 (DATA_ABI_ALIGNMENT): ... this.
5218 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5219 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5220 (DATA_ABI_ALIGNMENT): ... this.
5221 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5222 * doc/tm.texi: Regenerated.
5224 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
5226 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5227 cmp_code to construct REG_EQUAL note.
5229 2013-06-09 Jakub Jelinek <jakub@redhat.com>
5232 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5233 that operands[2] doesn't overlap with operands[0].
5235 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
5236 Jan Hubicka <jh@suse.cz>
5238 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5239 hack to mark symbols as used.
5241 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5243 PR rtl-optimization/57559
5244 * lra-constraints.c (process_alt_operands): Don't discourage
5245 memory with known offset for offsetable memory constraint.
5246 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5248 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5250 * varasm.c (struct oc_local_state): Reorder fields.
5251 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5252 and adjust accordingly.
5253 (output_constructor): Reorder initialization code and adjust call to
5254 output_constructor_bitfield.
5256 2013-06-07 Jan Hubicka <jh@suse.cz>
5258 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5260 2013-06-07 David Malcolm <dmalcolm@redhat.com>
5262 * tree-object-size.c (unknown): Make const.
5264 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5266 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5267 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5268 for last alternative in the cpu_facility attribute.
5270 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5273 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5274 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5275 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5276 * config/arm/constraints.md (Dg): New constraint.
5277 * config/arm/neon.md (xordi3_neon): Remove.
5278 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5279 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5281 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5283 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5284 Clean up alternatives.
5286 2013-06-07 Alan Modra <amodra@gmail.com>
5288 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5291 2013-06-07 Alan Modra <amodra@gmail.com>
5293 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5295 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5297 * config/arm/constraints.md (Df): New constraint.
5298 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5299 Correct length attribute for last two alternatives.
5301 2013-06-07 Alan Modra <amodra@gmail.com>
5303 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5304 override user -mfp-in-toc.
5305 (offsettable_ok_by_alignment): Consider just the current access
5306 rather than the whole object, unless BLKmode. Handle
5307 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5308 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5309 for -mcmodel=medium.
5310 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5311 override user -mfp-in-toc or -msum-in-toc. Default to
5312 -mno-fp-in-toc for -mcmodel=medium.
5314 2013-06-06 DJ Delorie <dj@redhat.com>
5316 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5317 TARGET_VALID_POINTER_MODE.
5319 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5320 Pat Haugen <pthaugen@us.ibm.com>
5321 Peter Bergner <bergner@vnet.ibm.com>
5323 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5324 Document new power8 builtins.
5326 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5327 condition code register, to allow 128-bit logical operations to be
5328 done in the VSX or GPR registers.
5329 (nor<mode>3): Use the canonical form for nor.
5330 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5331 vclz*, and vpopcnt* vector instructions.
5332 (nand<mode>3): Likewise.
5333 (orc<mode>3): Likewise.
5334 (clz<mode>2): LIkewise.
5335 (popcount<mode>2): Likewise.
5337 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5338 that only the GPRs are recognized.
5340 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5341 support for new power8 builtins.
5343 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5345 (xscvdpspn): Likewise.
5351 (vpopcnt): Likewise.
5352 (vpopcntb): Likewise.
5353 (vpopcnth): Likewise.
5354 (vpopcntw): Likewise.
5355 (vpopcntd): Likewise.
5360 (eqv_v16qi3): Likewise.
5361 (eqv_v8hi3): Likewise.
5362 (eqv_v4si3): Likewise.
5363 (eqv_v2di3): Likewise.
5364 (eqv_v4sf3): Likewise.
5365 (eqv_v2df3): Likewise.
5367 (nand_v16qi3): Likewise.
5368 (nand_v8hi3): Likewise.
5369 (nand_v4si3): Likewise.
5370 (nand_v2di3): Likewise.
5371 (nand_v4sf3): Likewise.
5372 (nand_v2df3): Likewise.
5374 (orc_v16qi3): Likewise.
5375 (orc_v8hi3): Likewise.
5376 (orc_v4si3): Likewise.
5377 (orc_v2di3): Likewise.
5378 (orc_v4sf3): Likewise.
5379 (orc_v2df3): Likewise.
5381 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
5382 allow power8 quad mode in 64-bit.
5383 (rs6000_builtin_vectorized_function): Add support to vectorize
5384 ISA 2.07 count leading zeros, population count builtins.
5385 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
5386 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
5387 (builtin_function_type): Add vgbbd builtin function which takes an
5389 (altivec_expand_vec_perm_const): Add support for new power8 merge
5392 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
5393 that does not include TImdoe for use with 32-bit.
5394 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
5396 (UNSPEC_VSX_CVDPSPN): Likewise.
5397 (vsx_xscvdpspn): Likewise.
5398 (vsx_xscvspdpn): Likewise.
5399 (vsx_xscvdpspn_scalar): Likewise.
5400 (vsx_xscvspdpn_directmove): Likewise.
5401 (vsx_and<mode>3): Split logical operations into 32-bit and
5402 64-bit. Add support to do logical operations on TImode as well as
5403 VSX vector types. Allow logical operations to be done in either
5404 VSX registers or in general purpose registers in 64-bit mode. Add
5405 splitters if GPRs were used. For AND, add clobber of CCmode to
5406 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
5408 (vsx_and<mode>3_32bit): Likewise.
5409 (vsx_and<mode>3_64bit): Likewise.
5410 (vsx_ior<mode>3): Likewise.
5411 (vsx_ior<mode>3_32bit): Likewise.
5412 (vsx_ior<mode>3_64bit): Likewise.
5413 (vsx_xor<mode>3): Likewise.
5414 (vsx_xor<mode>3_32bit): Likewise.
5415 (vsx_xor<mode>3_64bit): Likewise.
5416 (vsx_one_cmpl<mode>2): Likewise.
5417 (vsx_one_cmpl<mode>2_32bit): Likewise.
5418 (vsx_one_cmpl<mode>2_64bit): Likewise.
5419 (vsx_nor<mode>3): Likewise.
5420 (vsx_nor<mode>3_32bit): Likewise.
5421 (vsx_nor<mode>3_64bit): Likewise.
5422 (vsx_andc<mode>3): Likewise.
5423 (vsx_andc<mode>3_32bit): Likewise.
5424 (vsx_andc<mode>3_64bit): Likewise.
5425 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
5426 and xxlorc instructions.
5427 (vsx_eqv<mode>3_64bit): Likewise.
5428 (vsx_nand<mode>3_32bit): Likewise.
5429 (vsx_nand<mode>3_64bit): Likewise.
5430 (vsx_orc<mode>3_32bit): Likewise.
5431 (vsx_orc<mode>3_64bit): Likewise.
5433 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
5435 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
5437 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
5438 (p8_vmrgow): Likewise.
5439 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
5440 GPRs to be split under VSX.
5441 (p8v_clz<mode>2): Add power8 count leading zero support.
5442 (p8v_popcount<mode>2): Add power8 population count support.
5443 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
5446 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
5449 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
5451 (vec_nand): Likewise.
5452 (vec_vclz): Likewise.
5453 (vec_vclzb): Likewise.
5454 (vec_vclzd): Likewise.
5455 (vec_vclzh): Likewise.
5456 (vec_vclzw): Likewise.
5457 (vec_vgbbd): Likewise.
5458 (vec_vmrgew): Likewise.
5459 (vec_vmrgow): Likewise.
5460 (vec_vpopcnt): Likewise.
5461 (vec_vpopcntb): Likewise.
5462 (vec_vpopcntd): Likewise.
5463 (vec_vpopcnth): Likewise.
5464 (vec_vpopcntw): Likewise.
5466 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5468 PR rtl-optimization/57468
5469 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
5472 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5474 PR rtl-optimization/57459
5475 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
5476 type when setting live regs.
5478 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5480 * config/s390/s390.opt (mlra): New option.
5481 * config/s390/s390.c (s390_decompose_address): Check displacement
5482 for all registers for LRA.
5483 (s390_secondary_reload): Don't used secondary reloads for LRA.
5484 (s390_lra_p): New function.
5485 (TARGET_LRA_P): Define.
5486 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
5487 of attribute cpu_facility to zarch for the last alternative.
5488 (*cmpmem_short): Ditto.
5490 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
5492 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
5493 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
5494 (arm_expand_prologue): Likewise.
5496 2013-06-06 Teresa Johnson <tejohnson@google.com>
5499 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
5500 as this is now done by redirect_edge_and_branch_force.
5501 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
5502 barriers, and fix interaction with splitting.
5503 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
5504 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
5505 reflect changes made in the routine.
5506 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
5507 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
5508 since this is called in cfglayout mode, and replace partition fixup
5509 with assert as that is now done by force_nonfallthru_and_redirect.
5510 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
5511 already be marked with region crossing note.
5512 (insert_section_boundary_note): Make non-static, gate on flag
5513 has_bb_partition, rewrite to also check for multiple partitions.
5514 (rest_of_handle_reorder_blocks): Remove call to
5515 insert_section_boundary_note, now done later during free_cfg.
5516 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
5517 * bb-reorder.h (insert_section_boundary_note): Declare.
5518 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
5519 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
5520 invoke insert_section_boundary_note.
5521 (try_redirect_by_replacing_jump): Remove unnecessary
5522 check for region crossing note.
5523 (fixup_partition_crossing): New function.
5524 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
5525 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
5526 in non-cfglayout mode.
5527 (force_nonfallthru_and_redirect): Fixup partition boundaries,
5528 remove old code that tried to do this. Emit barrier correctly
5529 when we are in cfglayout mode.
5530 (last_bb_in_partition): New function.
5531 (rtl_split_edge): Correctly fixup partition boundaries.
5532 (commit_one_edge_insertion): Remove old code that tried to
5533 fixup region crossing edge since this is now handled in
5534 split_block, and set up insertion point correctly since
5535 block may now end in a jump.
5536 (verify_hot_cold_block_grouping): Guard against checking when not in
5537 linearized RTL mode.
5538 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
5540 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
5541 rtl_verify_flow_info, so not called in cfglayout mode.
5542 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
5543 (fixup_reorder_chain): Remove old code that attempted to fixup region
5544 crossing note as this is now handled in force_nonfallthru_and_redirect.
5545 (duplicate_insn_chain): Don't duplicate switch section notes.
5546 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
5548 * basic-block.h (emit_barrier_after_bb): Declare.
5550 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5552 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
5553 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
5554 arm_usatsihi): Adjust alternatives for arm_restrict_it.
5556 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5558 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
5560 * config/arm/ldmstm.md: Regenerate.
5562 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5564 * config/arm/sync.md (atomic_loaddi_1):
5565 Disable predication for arm_restrict_it.
5566 (arm_load_exclusive<mode>): Likewise.
5567 (arm_load_exclusivesi): Likewise.
5568 (arm_load_exclusivedi): Likewise.
5569 (arm_load_acquire_exclusive<mode>): Likewise.
5570 (arm_load_acquire_exclusivesi): Likewise.
5571 (arm_load_acquire_exclusivedi): Likewise.
5572 (arm_store_exclusive<mode>): Likewise.
5573 (arm_store_exclusive<mode>): Likewise.
5574 (arm_store_release_exclusivedi): Likewise.
5575 (arm_store_release_exclusive<mode>): Likewise.
5577 2013-06-06 Richard Biener <rguenther@suse.de>
5579 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
5581 (lto_tag_is_tree_code_p): Adjust.
5582 (lto_tag_is_gimple_code_p): Likewise.
5583 (lto_gimple_code_to_tag): Likewise.
5584 (lto_tag_to_gimple_code): Likewise.
5585 (lto_tree_code_to_tag): Likewise.
5586 (lto_tag_to_tree_code): Likewise.
5587 * data-streamer.h (streamer_write_hwi_in_range): Use
5588 uhwi streaming to stream the normalized range.
5589 (streamer_read_hwi_in_range): Likewise.
5591 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5593 * config/arm/arm.md (enabled_for_depr_it): New attribute.
5594 (predicable_short_it): Likewise.
5595 (predicated): Likewise.
5596 (enabled): Handle above.
5597 (define_cond_exec): Set predicated attribute to yes.
5599 2013-06-05 Mike Stump <mikestump@comcast.net>
5601 * gdbinit.in (__FUNCTION__): Add.
5603 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
5605 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
5606 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
5608 2013-06-05 Jan Hubicka <jh@suse.cz>
5610 * varasm.c (mark_decl_referenced): Revert the removal until targets
5613 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
5615 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
5616 instead of mark_decl_referenced.
5618 2013-06-05 Jan Hubicka <jh@suse.cz>
5620 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
5621 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
5622 and symtab_used_from_object_file_p.
5623 (cgraph_make_node_local_1): Clear forced_by_abi.
5624 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
5625 * cgraph.h (symtab_node_base): Add forced_by_abi.
5626 (decide_is_variable_needed): Remove.
5627 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
5628 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
5629 (decide_is_symbol_needed): ... this one; handle symbols in general;
5630 always analyze virtuals; honnor forced_by_abi.
5631 (cgraph_finalize_function): Update.
5632 (varpool_finalize_decl): Update.
5633 (symbol_defined_and_needed): Remove.
5634 (analyze_functions): Update.
5635 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
5636 output_refs, input_overwrite_node): Handle forced_by_abi.
5637 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
5638 (address_taken_from_non_vtable_p): ... this one.
5639 (comdat_can_be_unshared_p_1): New function.
5640 (cgraph_comdat_can_be_unshared_p): Rename to ...
5641 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
5642 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
5643 (function_and_variable_visibility): Clear forced_by_abi as needed.
5644 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
5645 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
5646 * symtab.c (dump_symtab_base): Dump forced_by_abi.
5647 * varpool.c (decide_is_variable_needed): Remove.
5649 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5651 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
5652 (arm_option_override): Override arm_restrict_it where appropriate.
5653 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
5654 * config/arm/arm.opt (mrestrict-it): New command-line option.
5655 * doc/invoke.texi: Document -mrestrict-it.
5657 2013-06-05 David Malcolm <dmalcolm@redhat.com>
5659 * tsan.c (tsan_atomic_table): Make const.
5661 2013-06-05 Richard Biener <rguenther@suse.de>
5663 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
5664 index associated with the tree we are supposed to replace.
5665 * tree-streamer-out.c (pack_ts_base_value_fields): Output
5666 TREE_ASM_WRITTEN as zero for everything but SSA names.
5668 2013-06-05 David Malcolm <dmalcolm@redhat.com>
5670 * tree-ssa-structalias.c (call_stmt_vars): Make static.
5672 2013-06-04 Jan Hubicka <jh@suse.cz>
5674 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
5675 (input_node, input_varpool_node): Handle correctly external same
5677 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
5678 nodes at ltrans stage.
5680 2013-06-04 Jan Hubicka <jh@suse.cz>
5682 * ipa-inline.c (update_caller_keys): Fix availability test.
5683 (update_callee_keys): Likewise.
5684 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
5685 to follow ELF standard.
5687 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
5689 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
5690 (mips64r5900el-*-elf*): New configurations.
5691 * config/mips/mips-cpus.def (r5900): New processor.
5692 * config/mips/mips-tables.opt: Regenerate.
5693 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
5694 (mips_issue_rate): Handle PROCESSOR_R5900.
5695 (mips_reorg_process_insns): Force reorder mode for the R5900.
5696 * config/mips/mips.h (TARGET_MIPS5900): Define.
5697 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
5699 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
5701 * config/mips/mips.md (processor): Add r5900.
5702 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
5704 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5706 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
5707 into function to generate MOVI instruction.
5708 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
5709 (aarch64_preferred_simd_mode): Turn into wrapper.
5710 (aarch64_output_scalar_simd_mov_immediate): New function.
5711 * config/aarch64/aarch64-protos.h: Add prototype for above.
5713 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5715 * config/aarch64/aarch64.c (simd_immediate_info): Remove
5716 element_char member.
5717 (sizetochar): Return signed char.
5718 (aarch64_simd_valid_immediate): Remove elchar and other
5719 unnecessary variables.
5720 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
5721 Calculate element_char as required.
5722 * config/aarch64/aarch64-protos.h: Update and move prototype
5723 for aarch64_output_simd_mov_immediate.
5724 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
5727 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5729 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
5730 information completed by aarch64_simd_valid_immediate.
5731 (aarch64_legitimate_constant_p): Update arguments.
5732 (aarch64_simd_valid_immediate): Work with struct rather than many
5734 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
5735 (aarch64_simd_make_constant): Update arguments.
5736 (aarch64_output_simd_mov_immediate): Work with struct rather than
5737 many pointers. Output immediate directly rather than as operand.
5738 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5740 * config/aarch64/constraints.md (Dn): Update arguments.
5742 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5744 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
5746 (aarch64_simd_immediate_valid_for_move): Remove.
5747 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
5748 (aarch64_simd_make_constant): Update call.
5749 (aarch64_output_simd_mov_immediate): Update call.
5750 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5752 * config/aarch64/constraints.md (Dn): Update call.
5754 2013-06-04 Ian Bolton <ian.bolton@arm.com>
5756 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
5757 return type to bool for prototype.
5758 (aarch64_legitimate_constant_p): Check for true instead of not -1.
5759 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
5760 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
5762 2013-06-04 Catherine Moore <clm@codesourcery.com>
5764 * config/mips/mips.opt (meva): New.
5765 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
5766 (ASM_SPEC): Handle -meva.
5767 * doc/invoke.texi (meva): Document.
5769 2013-06-04 Alan Modra <amodra@gmail.com>
5771 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
5774 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5776 * rtl.def: Add extra fourth optional field to define_cond_exec.
5777 * gensupport.c (process_one_cond_exec): Process attributes from
5779 * doc/md.texi: Document fourth field in define_cond_exec.
5781 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
5783 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
5784 out the processing order as in store_bit_field_1.
5786 2013-06-04 Jan Hubicka <jh@suse.cz>
5789 * cgraphunit.c (cgraph_process_same_body_aliases): Create
5790 non-VAR_DECL node if it does not exist yet.
5792 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
5794 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
5795 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
5796 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
5797 target_cpu_default setting.
5799 2013-06-03 Teresa Johnson <tejohnson@google.com>
5801 * dumpfile.c (opt_info_switch_p): Change -fopt-info
5802 default to -fopt-info=optimized instead of all.
5803 * doc/invoke.texi: Ditto.
5804 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
5805 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
5806 (execute_vect_slp): Emit BB vectorization success under
5807 MSG_OPTIMIZED_LOCATIONS.
5808 * tree-vect-slp.c (vect_slp_transform_bb): Change
5809 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
5810 * tree-vect-loop.c (vect_transform_loop): Ditto.
5812 2013-06-03 Jason Merrill <jason@redhat.com>
5815 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
5816 Use TARGET_EXPR for C++.
5818 2013-06-03 Jakub Jelinek <jakub@redhat.com>
5820 PR rtl-optimization/57268
5821 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
5822 if DEBUG_INSN_P (insn).
5825 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
5827 PR rtl-optimization/57268
5828 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5829 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5831 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
5833 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
5834 (ix86_avoid_lea_for_addr): Likewise.
5835 (exact_dependency_1): Likewise.
5836 (ix86_adjust_cost): Likewise.
5837 (swap_top_of_ready_list): Fix formatting and !reload_completed check
5839 (do_reorder_for_imul): Fix typo, formatting and
5840 !reload_completed check removed.
5841 (ix86_sched_reorder): Fix typo and formatting.
5842 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
5845 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
5847 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
5849 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
5851 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
5852 <STRING_CST>: Likewise.
5853 <VECTOR_CST>: Likewise.
5855 2013-06-01 Janus Weil <janus@gcc.gnu.org>
5856 Mikael Morin <mikael@gcc.gnu.org>
5858 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
5859 * config.in: Regenerated.
5860 * configure: Regenerated.
5862 2013-06-01 Jan Hubicka <jh@suse.cz>
5865 * cgraphunit.c (compile): When weakref is not supported,
5866 set up transparent aliases before final output pass.
5867 * varasm.c (assemble_alias): Do not try to do it here.
5869 2013-06-01 Jan Hubicka <jh@suse.cz>
5872 * passes.c (for_per_function): Skip unanalyzed functions.
5874 2013-06-01 Jan Hubicka <jh@suse.cz>
5876 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
5877 (lto_symtab_merge_symbols_1): ... this one.
5878 (lto_symtab_merge_cgraph_nodes): Rename to ...
5879 (lto_symtab_merge_symbols): ... this one; simplify.
5880 * cgraph.c (same_body_aliases_done): Rename to ...
5881 (cpp_implicit_aliases_done): ... this one.
5882 (cgraph_create_function_alias): Update.
5883 (cgraph_same_body_alias): Update.
5884 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
5885 (verify_edge_corresponds_to_fndecl): Simplify.
5886 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
5887 (cgraph_node): Remove same_body_alias.
5888 (varpool_node): Remove alias_of and extra_name_alias.
5889 (same_body_aliases_done): Rename to ..
5890 (cpp_implicit_aliases_done): ... this one.
5891 (symtab_alias_ultimate_target): Add default parameter.
5892 (symtab_resolve_alias): New function.
5893 (fixup_same_cpp_alias_visibility): Declare.
5894 (cgraph_function_node): Add default parameter.
5895 (cgraph_node_asm_name): Likewise.
5896 (cgraph_function_or_thunk_node): Add default parameter; do
5897 not ICE when it is NULL.
5898 (varpool_variable_node): Likewise.
5899 * tree-emutls.c (create_emultls_var): Update.
5900 (ipa_lower_emutls): Update.
5901 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
5902 (cgraph_reset_node): Reset alias info.
5903 (cgraph_finalize_function): Update.
5904 (fixup_same_cpp_alias_visibility): Move to symtab.c.
5905 (analyze_function): Simplify.
5906 (cgraph_process_same_body_aliases): Simplify.
5907 (analyze_functions): Fixup same body aliases.
5908 (handle_alias_pairs): Simplify.
5909 (assemble_thunk): Update.
5910 (assemble_thunks_and_aliases): Update.
5911 (output_weakrefs): Rewrite.
5912 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
5913 (lto_output_varpool_node): Likewise.
5914 (compute_ltrans_boundary): Remve assert.
5915 (get_alias_symbol): New functoin.
5916 (input_node): Rewrite alias handling.
5917 (input_varpool_node): Likewise.
5918 * ipa-pure-const.c (propagate_pure_const): Fix formating.
5919 * ipa.c (process_references): Handle weakrefs correctly.
5920 (symtab_remove_unreachable_nodes): Likewise.
5921 * trans-mem.c (get_cg_data): Update.
5922 (ipa_tm_create_version_alias): Update.
5923 (ipa_tm_execute): Update.
5924 * symtab.c (dump_symtab_base): Dump aliases.
5925 (verify_symtab_base): Verify aliases.
5926 (symtab_node_availability): New function.
5927 (symtab_alias_ultimate_target): Simplify.
5928 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
5929 handle all the fixup cases.
5930 (symtab_resolve_alias): New function.
5931 * passes.c (ipa_write_summaries): Handle weakrefs.
5932 * varpool.c (varpool_analyze_node): Simplify.
5933 (assemble_aliases): Update.
5934 (varpool_create_variable_alias): Simplify.
5935 (varpool_extra_name_alias): Simplify.
5936 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
5937 (lto_symtab_merge_symbols): ... this one.
5939 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
5942 PR rtl-optimization/57268
5943 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5944 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5946 2013-06-01 Tobias Burnus <burnus@net-b.de>
5949 2013-05-31 Tobias Burnus <burnus@net-b.de>
5952 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
5955 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
5957 PR rtl-optimization/57268
5958 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5959 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5961 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
5963 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
5964 unordered comparison operators when -fno-trapping-math is in effect
5966 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
5967 and implement unordered comparison operators properly on the e500.
5969 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
5971 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
5972 for constant scalar integers.
5973 (simplify_relational_operation_1): Likewise.
5975 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
5977 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
5978 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
5981 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
5982 Igor Zamyatin <igor.zamyatin@intel.com>
5984 Silvermont (SLM) architecture performance tuning.
5985 * config/i386/i386.h (enum ix86_tune_indices): Add
5986 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
5987 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
5989 * config/i386/i386.c (initial_ix86_tune_features)
5990 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
5991 (ix86_lea_outperforms): Handle Silvermont tuning.
5992 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
5994 (ix86_use_lea_for_mov): Likewise.
5995 (ix86_avoid_lea_for_addr): Likewise.
5996 (ix86_lea_for_add_ok): Likewise.
5997 (exact_dependency_1): New function.
5998 (exact_store_load_dependency): Likewise.
5999 (ix86_adjust_cost): Handle Silvermont tuning.
6000 (do_reoder_for_imul): Likewise.
6001 (swap_top_of_ready_list): New function.
6002 (ix86_sched_reorder): Changed to handle Silvermont tuning.
6004 * config/i386/i386.md (peepholes that split memory operand in fp
6007 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6009 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6010 Remove un-necessary braces.
6012 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6014 * config/aarch64/aarch64.c (aarch64_classify_symbol):
6015 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
6017 2013-05-31 Tobias Burnus <burnus@net-b.de>
6020 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
6022 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6025 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6026 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6027 * config/arm/neon.md (iordi3_neon): Remove.
6028 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6029 * config/arm/predicates.md (imm_for_neon_logic_operand):
6030 Move to earlier in the file.
6031 (neon_logic_op2): Likewise.
6032 (arm_iordi_operand_neon): New predicate.
6034 2013-05-31 Richard Biener <rguenther@suse.de>
6036 PR tree-optimization/57478
6037 PR tree-optimization/57453
6038 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6041 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6043 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6044 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6046 2013-05-30 Tobias Burnus <burnus@net-b.de>
6047 Thomas Koenig <tkoenig@gcc.gnu.org>
6050 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6051 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6053 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6055 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
6057 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6059 * target.def (register_usage_leveling_p): New hook.
6060 * targhooks.c (default_register_usage_leveling_p): New.
6061 * targhooks.h (default_register_usage_leveling_p): New prototype.
6062 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6063 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6064 * doc/tm.texi: Update.
6065 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6067 2013-05-30 Ian Bolton <ian.bolton@arm.com>
6069 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6070 (*insv_reg<mode>): New define_insn.
6072 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6074 PR rtl-optimization/57439
6075 * postreload.c (move2add_valid_value_p): Check that we have
6076 a zero subreg_regno_offset when accessing the register in
6079 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6080 Igor Zamyatin <igor.zamyatin@intel.com>
6082 Silvermont (SLM) architecture pipeline model, tuning and
6084 * config.gcc: Add slm config options and target.
6086 * config/i386/slm.md: New.
6088 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6090 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6092 (ix86_target_macros_internal): Likewise.
6094 * gcc/config/i386/i386.c (slm_cost): New cost.
6095 (m_SLM): New macro flag.
6096 (initial_ix86_tune_features): Set m_SLM.
6097 (x86_accumulate_outgoing_args): Likewise.
6098 (x86_arch_always_fancy_math_387): Likewise.
6099 (processor_target_table): Add slm cost.
6100 (cpu_names): Add slm cpu name.
6101 (x86_option_override_internal): Set SLM ISA.
6102 (ix86_issue_rate): New case PROCESSOR_SLM.
6103 (ia32_multipass_dfa_lookahead): Likewise.
6104 (fold_builtin_cpu): Add slm.
6106 * config/i386/i386.h (TARGET_SLM): New target macro.
6107 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6108 (processor_type): Add PROCESSOR_SLM.
6110 * config/i386/i386.md (cpu): Add new value "slm".
6111 (slm.md): Include slm.md.
6113 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6114 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6116 * config/arm/arm-protos.h: Add and update function protos.
6117 * config/arm/arm.c (use_simple_return_p): New added.
6118 (thumb2_expand_return): Check simple_return flag.
6119 * config/arm/arm.md: Add simple_return and conditional simple_return.
6120 * config/arm/iterators.md: Add iterator for return and simple_return.
6122 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6124 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6125 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6126 (arm_emit_vfp_multi_reg_pop): Likewise.
6127 (thumb2_emit_ldrd_pop): Likewise.
6128 (arm_expand_epilogue): Add misc REG_CFA notes.
6129 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6131 2013-05-29 Lawrence Crowl <crowl@google.com>
6133 * config/arm/t-arm: Update for below.
6135 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6136 Change type to hash_table. Update dependent calls and types.
6138 * config/i386/t-cygming: Update for below.
6140 * config/i386/t-interix: Update for below.
6142 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6143 Change type to hash_table. Update dependent calls and types.
6144 (i386_find_on_wrapper_list::wrappers): Likewise.
6146 * config/ia64/t-ia64: Update for below.
6148 * config/ia64/ia64.c (bundle_state_table):
6149 Change type to hash_table. Update dependent calls and types.
6151 * config/mips/mips.c (mips_reorg_process_insns::htab):
6152 Change type to hash_table. Update dependent calls and types.
6154 * config/sol2.c (solaris_comdat_htab):
6155 Change type to hash_table. Update dependent calls and types.
6157 * config/t-sol2: Update for above.
6159 2013-05-29 Teresa Johnson <tejohnson@google.com>
6161 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6162 functions are not yet marked as defined.
6164 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6165 Pat Haugen <pthaugen@us.ibm.com>
6166 Peter Bergner <bergner@vnet.ibm.com>
6168 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6172 (vrotl<mode>3): Likewise.
6173 (vashl<mode>3): Likewise.
6174 (vlshr<mode>3): Likewise.
6175 (vashr<mode>3): Likewise.
6177 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6178 support for power8 V2DI builtins.
6180 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6181 power8 V2DI builtins.
6182 (vupkhsw): Likewise.
6183 (vupklsw): Likewise.
6184 (vaddudm): Likewise.
6189 (vpkudum): Likewise.
6190 (vpksdss): Likewise.
6191 (vpkudus): Likewise.
6192 (vpksdus): Likewise.
6197 (vsubudm): Likewise.
6198 (vcmpequd): Likewise.
6199 (vcmpgtsd): Likewise.
6200 (vcmpgtud): Likewise.
6201 (vcmpequd_p): Likewise.
6202 (vcmpgtsd_p): Likewise.
6203 (vcmpgtud_p): Likewise.
6204 (vupkhsw): Likewise.
6205 (vupklsw): Likewise.
6206 (vaddudm): Likewise.
6211 (vpksdss): Likewise.
6212 (vpksdus): Likewise.
6213 (vpkudum): Likewise.
6214 (vpkudus): Likewise.
6219 (vsubudm): Likewise.
6221 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6222 support for power8 V2DI instructions.
6224 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6225 power8 V2DI instructions. Combine pack and unpack insns to use an
6226 iterator for each mode. Check whether a particular mode supports
6227 Altivec instructions instead of just checking TARGET_ALTIVEC.
6228 (UNSPEC_VPKUWUM): Likewise.
6229 (UNSPEC_VPKSHSS): Likewise.
6230 (UNSPEC_VPKSWSS): Likewise.
6231 (UNSPEC_VPKUHUS): Likewise.
6232 (UNSPEC_VPKSHUS): Likewise.
6233 (UNSPEC_VPKUWUS): Likewise.
6234 (UNSPEC_VPKSWUS): Likewise.
6235 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6236 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6237 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6238 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6239 (UNSPEC_VUPKHSB): Likewise.
6240 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6241 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6242 (UNSPEC_VUPKHSH): Likewise.
6243 (UNSPEC_VUPKLSB): Likewise.
6244 (UNSPEC_VUPKLSH): Likewise.
6246 (VI_char): Likewise.
6247 (VI_scalar): Likewise.
6248 (VI_unit): Likewise.
6250 (VP_small): Likewise.
6251 (VP_small_lc): Likewise.
6252 (VU_char): Likewise.
6253 (add<mode>3): Likewise.
6254 (altivec_vaddcuw): Likewise.
6255 (altivec_vaddu<VI_char>s): Likewise.
6256 (altivec_vadds<VI_char>s): Likewise.
6257 (sub<mode>3): Likewise.
6258 (altivec_vsubcuw): Likewise.
6259 (altivec_vsubu<VI_char>s): Likewise.
6260 (altivec_vsubs<VI_char>s): Likewise.
6261 (altivec_vavgs<VI_char>): Likewise.
6262 (altivec_vcmpbfp): Likewise.
6263 (altivec_eq<mode>): Likewise.
6264 (altivec_gt<mode>): Likewise.
6265 (altivec_gtu<mode>): Likewise.
6266 (umax<mode>3): Likewise.
6267 (smax<mode>3): Likewise.
6268 (umin<mode>3): Likewise.
6269 (smin<mode>3): Likewise.
6270 (altivec_vpkuhum): Likewise.
6271 (altivec_vpkuwum): Likewise.
6272 (altivec_vpkshss): Likewise.
6273 (altivec_vpkswss): Likewise.
6274 (altivec_vpkuhus): Likewise.
6275 (altivec_vpkshus): Likewise.
6276 (altivec_vpkuwus): Likewise.
6277 (altivec_vpkswus): Likewise.
6278 (altivec_vpks<VI_char>ss): Likewise.
6279 (altivec_vpks<VI_char>us): Likewise.
6280 (altivec_vpku<VI_char>us): Likewise.
6281 (altivec_vpku<VI_char>um): Likewise.
6282 (altivec_vrl<VI_char>): Likewise.
6283 (altivec_vsl<VI_char>): Likewise.
6284 (altivec_vsr<VI_char>): Likewise.
6285 (altivec_vsra<VI_char>): Likewise.
6286 (altivec_vsldoi_<mode>): Likewise.
6287 (altivec_vupkhsb): Likewise.
6288 (altivec_vupkhs<VU_char>): Likewise.
6289 (altivec_vupkls<VU_char>): Likewise.
6290 (altivec_vupkhsh): Likewise.
6291 (altivec_vupklsb): Likewise.
6292 (altivec_vupklsh): Likewise.
6293 (altivec_vcmpequ<VI_char>_p): Likewise.
6294 (altivec_vcmpgts<VI_char>_p): Likewise.
6295 (altivec_vcmpgtu<VI_char>_p): Likewise.
6296 (abs<mode>2): Likewise.
6297 (vec_unpacks_hi_v16qi): Likewise.
6298 (vec_unpacks_hi_v8hi): Likewise.
6299 (vec_unpacks_lo_v16qi): Likewise.
6300 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6301 (vec_unpacks_lo_v8hi): Likewise.
6302 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6303 (vec_pack_trunc_v8h): Likewise.
6304 (vec_pack_trunc_v4si): Likewise.
6305 (vec_pack_trunc_<mode>): Likewise.
6307 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6309 (vec_vmaxsd): Likewise.
6310 (vec_vmaxud): Likewise.
6311 (vec_vminsd): Likewise.
6312 (vec_vminud): Likewise.
6313 (vec_vpksdss): Likewise.
6314 (vec_vpksdus): Likewise.
6315 (vec_vpkudum): Likewise.
6316 (vec_vpkudus): Likewise.
6317 (vec_vrld): Likewise.
6318 (vec_vsld): Likewise.
6319 (vec_vsrad): Likewise.
6320 (vec_vsrd): Likewise.
6321 (vec_vsubudm): Likewise.
6322 (vec_vupkhsw): Likewise.
6323 (vec_vupklsw): Likewise.
6325 2013-05-29 Jan Hubicka <jh@suse.cz>
6327 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6328 flags; reorder rest of fields in more consistent way.
6329 (varpool_node): Remove analyzed, finalized and alias.
6330 (cgraph_ndoe): Likewise.
6331 (symtab_alias_ultimate_target): New function.
6332 (cgraph_function_node): Move offline.
6333 (cgraph_reset_node): Declare.
6334 (cgraph_comdat_can_be_unshared_p): Remove.
6335 (varpool_remove_initializer): Declare.
6336 (varpool_first_defined_variable, varpool_next_defined_variable
6337 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6338 (cgraph_function_with_gimple_body_p): Update.
6339 (varpool_all_refs_explicit_p): Update.
6340 (symtab_alias_target): New function.
6341 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6342 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
6343 (cgraph_function_or_thunk_node): Simplify using
6344 symtab_alias_ultimate_target.
6345 (varpool_variable_node): Likewise.
6346 * cgraph.c (cgraph_create_function_alias): Update.
6347 (cgraph_add_thunk): Update.
6348 (cgraph_remove_node): Update.
6349 (dump_cgraph_node): Do not dump removed flags.
6350 (cgraph_function_body_availability): Update.
6351 (cgraph_propagate_frequency): Update.
6352 (verify_cgraph_node): Check sanity of local flag.
6353 (cgraph_function_node): Move here from cgraph.h; revamp for
6354 cgraph_function_or_thunk_node.
6355 * lto-symtab.c (lto_varpool_replace_node): Update.
6356 (lto_symtab_resolve_can_prevail_p): Update.
6357 (lto_symtab_merge_cgraph_nodes): Update.
6358 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6359 propagate_constants_accross_call, devirtualization_time_bonus,
6360 ipcp_propagate_stage): Update.
6361 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
6362 * ipa-inline-transform.c (clone_inlined_nodes,
6363 preserve_function_body_p): Update.
6364 * ipa-reference.c (propagate): Update.
6365 (write_node_summary_p): Update.
6366 * toplev.c (wrapup_global_declaration_2): Update.
6367 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6368 (analyze_function) ... this one.
6369 (cgraph_process_new_functions): Update.
6370 (cgraph_reset_node): Export.
6371 (cgraph_finalize_function): Update.
6372 (cgraph_add_new_function): Update.
6373 (process_function_and_variable_attributes): Update.
6374 (varpool_finalize_decl): Update.
6375 (symbol_finalized): Remove.
6376 (symbol_finalized_and_needed): Rename to ...
6377 (symbol_defined_and_needed): ... update.
6378 (cgraph_analyze_functions): Update.
6379 (handle_alias_pairs): Update.
6380 (mark_functions_to_output): Update.
6381 (assemble_thunk): Update.
6382 (output_in_order): Update.
6383 (output_weakrefs): Update.
6384 (finalize_compilation_unit): Update.
6385 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
6386 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
6387 input_node, input_varpool_node): Update.
6388 * dbxout.c (dbxout_expand_expr): Update.
6389 * cgraphclones.c (cgraph_clone_node): Update.
6390 (cgraph_copy_node_for_versioning): Update.
6391 (cgraph_materialize_clone): Update.
6392 (cgraph_materialize_all_clones): Update.
6393 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
6394 propagate_pure_const, propagate_nothrow): Update.
6395 * lto-streamer-out.c (lto_output, write_symbol): Update.
6396 * ipa-utils.c (ipa_reverse_postorder): Update.
6397 * ipa-inline.c (can_inline_edge_p): Update.
6398 (update_caller_keys, ipa_inline): Update.
6399 * dwarf2out.c (reference_to_unused,
6400 premark_types_used_by_global_vars_helper): Update.
6401 * tree-eh.c (tree_could_trap_p): Update.
6402 * ipa-split.c (consider_split, execute_split_functions): Update.
6403 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
6404 has_addr_references_p): Update; move ahead in file for better
6406 (process_references): Simplify.
6407 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
6409 (cgraph_comdat_can_be_unshared_p): Make static.
6410 (cgraph_externally_visible_p): Update.
6411 (varpool_externally_visible_p): Update.
6412 (function_and_variable_visibility): Update.
6413 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
6414 ipa_tm_mark_force_output_node): Update.
6415 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
6416 estimate_edge_devirt_benefit, inline_generate_summary,
6417 inline_write_summary): Update.
6418 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
6419 * ipa-prop.c (ipa_compute_jump_functions): Update.
6420 (ipa_print_node_params, ipa_prop_read_section,
6421 ipa_update_after_lto_read, read_replacements_section): Update.
6422 * varasm.c (mark_decl_referenced): Update.
6423 (assemble_alias, dump_tm_clone_pairs): Update.
6424 * tree-inline.c (copy_bb): Update.
6425 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
6427 * symtab.c (dump_symtab_base): Print new flags.
6428 (verify_symtab_base): Verify new flags.
6429 (symtab_alias_ultimate_target): New function.
6430 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
6431 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
6433 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
6435 * i386.c (ix86_get_function_versions_dispatcher,
6436 ix86_generate_version_dispatcher_body): Update.
6437 (fold_builtin_cpu): Use varpool_add_new_variable.
6438 * varpool.c (varpool_remove_initializer): Break out from ...
6439 (varpool_remove_node): ... this one.
6440 (dump_varpool_node, varpool_node_for_asm,
6441 cgraph_variable_initializer_availability, varpool_analyze_node,
6442 varpool_assemble_decl, varpool_remove_unreferenced_decls,
6443 varpool_finalize_named_section_flags, varpool_create_variable_alias):
6446 2013-05-29 Jan Hubicka <jh@suse.cz>
6448 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
6450 2013-05-29 Easwaran Raman <eraman@google.com>
6452 PR tree-optimization/57442
6453 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
6454 when control exits the main loop.
6456 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
6458 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
6460 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
6461 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
6462 * rx/t-rx: Add rx100 under multi library matches option for nofpu
6465 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6467 PR tree-optimization/57441
6468 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
6469 Don't limit size of incr_vec to number of candidates.
6471 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
6473 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
6474 and mips16 directories.
6475 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
6476 (MULTILIB_DIRNAMES): Ditto.
6477 (MULTILIB_EXCEPTIONS): Add new exceptions.
6478 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
6479 (MULTILIB_DIRNAMES): Ditto.
6480 (MULTILIB_EXCEPTIONS): Add new exceptions.
6482 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6483 Marcus Shawcroft <marcus.shawcroft@arm.com>
6485 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
6486 SYMBOL_TINY_ABSOLUTE.
6487 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
6488 SYMBOL_TINY_ABSOLUTE.
6489 (aarch64_expand_mov_immediate): Likewise.
6490 (aarch64_classify_symbol): Likewise.
6491 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
6492 Permit SYMBOL_TINY_ABSOLUTE.
6493 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
6495 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6496 Marcus Shawcroft <marcus.shawcroft@arm.com>
6498 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
6499 Refactor if/switch. Replace gcc_assert with if.
6501 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
6503 * config/i386/i386.c (initial_ix86_tune_features): Enable
6504 FP Reassociation for AMD bdver1 and bdver2.
6506 2013-05-29 Martin Jambor <mjambor@suse.cz>
6508 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
6509 and IMAGPART_EXPR do not occur within other handled_components.
6511 2013-05-29 Richard Biener <rguenther@suse.de>
6513 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
6514 access on whether the use is in the BB we currently try to
6516 (vect_bb_vectorization_profitable_p): Pass the BB we currently
6517 vectorize to vect_bb_slp_scalar_cost.
6519 2013-05-29 Richard Biener <rguenther@suse.de>
6521 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
6522 computing scalar cost offsetted by stmts that are kept live
6524 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
6525 for computation of scalar cost.
6527 2013-05-28 Steve Ellcey <sellcey@mips.com>
6529 * config/mips/mips-cpus.def (mips32r2): Change processor type.
6531 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6533 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
6534 array notation built-in reduction functions.
6535 * doc/passes.texi (Passes): Added documentation about changes done
6537 * doc/invoke.texi (C Dialect Options): Added documentation about
6538 the -fcilkplus flag.
6539 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
6540 (BUILTINS_DEF): Depend on cilkplus.def.
6541 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
6542 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
6543 * cilkplus.def: New file.
6545 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
6547 PR rtl-optimization/57439
6548 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
6550 2013-05-28 Easwaran Raman <eraman@google.com>
6552 PR tree-optimization/57337
6553 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
6554 (find_insert_point): Correctly identify the insertion point
6555 when two statements with the same UID is compared.
6557 2013-05-28 Richard Biener <rguenther@suse.de>
6559 PR tree-optimization/56787
6560 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
6561 from the list of data references.
6562 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
6564 (vect_analyze_loop_operations): Likewise.
6565 (vect_transform_loop): Remove clobbers.
6567 2013-05-28 Martin Jambor <mjambor@suse.cz>
6569 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
6570 and REALPART_EXPRs have scalar type.
6572 2013-05-28 Richard Biener <rguenther@suse.de>
6574 PR tree-optimization/57411
6575 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
6577 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
6578 virtual operand propagation.
6580 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6582 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
6583 destination register for bmasksi_vis.
6584 (vector_init_bshuffle): Likewise.
6585 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
6587 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6589 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
6590 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
6591 mode if the instruction isn't available in the original mode.
6592 * config/sparc/sparc.opt (mfix-ut699): New option.
6593 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
6594 (divdf3): Turn into expander.
6595 (divdf3_nofix): New insn.
6596 (divdf3_fix): Likewise.
6597 (divsf3): Disable if -mfix-ut699.
6598 (sqrtdf2): Turn into expander.
6599 (sqrtdf2_nofix): New insn.
6600 (sqrtdf2_fix): Likewise.
6601 (sqrtsf2): Disable if -mfix-ut699.
6603 2013-05-27 Richard Biener <rguenther@suse.de>
6606 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
6607 block for the new loop.
6609 2013-05-27 Richard Biener <rguenther@suse.de>
6611 PR tree-optimization/57343
6612 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
6613 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
6614 (number_of_iterations_cond): Do not build the folded tree.
6616 2013-05-27 Richard Biener <rguenther@suse.de>
6620 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6621 OEP_CONSTANT_ADDRESS_OF retained.
6623 PR tree-optimization/57417
6624 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
6626 (set_ssa_val_to): Compare addresses using
6627 get_addr_base_and_unit_offset.
6629 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
6631 PR rtl-optimization/56833
6632 * postreload.c (move2add_record_mode): New function.
6633 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
6634 (move2add_use_add2_insn): Use move2add_record_sym_value.
6635 (move2add_use_add3_insn): Likewise.
6636 (reload_cse_move2add): Use move2add_valid_value_p and
6637 move2add_record_mode. Invalidate call-clobbered and REG_INC
6638 affected regs by setting reg_mode to VOIDmode.
6639 (move2add_note_store): Don't pretend the inside of a SUBREG is
6640 the actual destination. Invalidate single/leading registers by
6641 setting reg_mode to VOIDmode.
6642 Use move2add_record_sym_value, move2add_valid_value_p and
6643 move2add_record_mode.
6645 2013-05-27 Richard Biener <rguenther@suse.de>
6647 PR tree-optimization/57396
6648 * tree-affine.c (double_int_constant_multiple_p): Properly
6649 return false for val == 0 and div != 0.
6651 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6653 * config/mips/mips.h: Use #elif in preprocessor conditions.
6655 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6658 * config/mips/constraints.md (kl): New constraint.
6659 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
6660 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
6661 constraint for operand 0. Split after CSE for MIPS16. Emit a move
6663 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
6665 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6668 * config/mips/mips.c (mips_can_inline_p): New function.
6669 (TARGET_CAN_INLINE_P): Define.
6671 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
6673 * sched-int.h (ds_t, dw_t): Make unsigned int.
6674 Fix documentation that describes how all the ds_t bits are used.
6675 Reserve the last bit for delayed-branch scheduling.
6676 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
6677 (BITS_PER_DEP_WEAK): Fix definition and documentation.
6678 (gen_dep_weak_1): Remove prototype.
6679 * sched-deps.c (get_dep_weak_1): Make static.
6680 * target.def (speculate_insn, needs_block_p, gen_spec_check,
6681 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
6682 * doc/tm.texi: Regenerate.
6683 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
6685 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
6688 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
6690 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
6691 Sandra Loosemore <sandra@codesourcery.com>
6693 * config.gcc (powerpc-*): Allow native for with-cpu.
6695 2013-05-24 Jeff Law <law@redhat.com>
6697 PR tree-optimization/57124
6698 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
6699 conversion feeding a condition if the range has an overflow
6700 if -fstrict-overflow. Add warnings for when we do make the
6703 2013-05-24 Dehao Chen <dehao@google.com>
6705 * tree-cfg.c (locus_discrim_map): Fix the typo.
6706 (locus_discrim_hasher): Likewise.
6707 (locus_discrim_hasher::hash): Likewise.
6708 (locus_discrim_hasher::equal): Likewise.
6710 2013-05-24 Martin Jambor <mjambor@suse.cz>
6712 PR tree-optimization/57294
6713 * cgraph.h (ipa_record_stmt_references): Declare.
6714 * cgraphbuild.c (ipa_record_stmt_references): New function.
6715 (build_cgraph_edges): Use ipa_record_stmt_references.
6716 (rebuild_cgraph_edges): Likewise.
6717 (cgraph_rebuild_references): Likewise.
6718 * ipa-prop.c (ipa_modify_call_arguments): Discard references
6719 associated with the old statement and build references from the
6720 newly built statements.
6721 * ipa-ref.c (ipa_remove_stmt_references): New function.
6722 * ipa-ref.h (ipa_remove_stmt_references): Declare.
6724 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
6726 * lra-constraints.c (emit_spill_move): Use smaller mode for
6728 (check_and_process_move): Consider mem-reg moves for secondary
6730 (curr_insn_transform): Don't lose insns emitted before for
6731 secondary memory moves.
6732 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
6733 reg set up in the current insn.
6735 2013-05-24 Dehao Chen <dehao@google.com>
6737 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
6739 (locus_descrim_hasher::equal): Likewise.
6740 (build_gimple_cfg): New discriminator assignment algorithm.
6741 (make_edges): Likewise.
6742 (next_discriminator_for_locus): Likewise.
6743 (same_line_p): Likewise.
6744 (assign_discriminators): Likewise.
6745 (make_cond_expr_edges): Likewise.
6746 (make_gimple_switch_edges): Likewise.
6747 (make_goto_expr_edges): Likewise.
6748 (make_gimple_asm_edges): Likewise.
6750 2013-05-24 Ian Bolton <ian.bolton@arm.com>
6752 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
6753 X format specifier to only display bottom 16 bits.
6754 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
6755 immediate to match for operand 2, since it will be masked.
6757 2013-05-24 Richard Biener <rguenther@suse.de>
6759 PR tree-optimization/57287
6760 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
6761 all SSA names that occur in abnormal PHIs.
6763 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
6766 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
6767 that index is not negative.
6769 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
6771 PR rtl-optimization/55177
6772 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
6773 (simplify_byte_swapping_operation): New.
6774 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
6775 (simplify_relational_operation_1): Deal with BSWAP.
6777 2013-05-23 Richard Henderson <rth@redhat.com>
6780 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
6781 (ix86_reorg): Call it.
6783 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
6786 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
6787 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
6788 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
6790 2013-05-23 Christian Bruel <christian.bruel@st.com>
6793 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
6795 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6796 Marcus Shawcroft <marcus.shawcroft@arm.com>
6798 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
6799 * config/aarch64/constraints.md (Usa): Remove.
6800 * doc/md.texi (AArch64 Usa): Remove.
6802 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6803 Marcus Shawcroft <marcus.shawcroft@arm.com>
6805 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
6806 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
6807 * config/aarch64/predicates.md (aarch64_const_address): Remove.
6808 (aarch64_mov_operand): Use aarch64_mov_operand_p.
6810 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
6812 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
6813 instruction (AdvSIMD).
6814 * config/aarch64/aarch64-builtins.c
6815 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
6816 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
6818 2013-05-23 Martin Jambor <mjambor@suse.cz>
6821 * tree.h (contains_bitfld_component_ref_p): Declare.
6822 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
6823 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
6825 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
6826 not access a bit-field. Assert all final offsets are byte-aligned.
6828 2013-05-23 Richard Biener <rguenther@suse.de>
6830 PR tree-optimization/57380
6831 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
6832 least one invariant or re-used load.
6833 * passes.c (init_optimization_passes): Move pass_phiprop before
6836 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
6838 * config/aarch64/aarch64-simd.md
6839 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
6841 2013-05-23 Richard Biener <rguenther@suse.de>
6844 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6845 OEP_CONSTANT_ADDRESS_OF retained.
6847 2013-05-23 Jakub Jelinek <jakub@redhat.com>
6850 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
6851 don't lower unit. Handle unit not being always BITS_PER_WORD.
6853 2013-05-23 Richard Biener <rguenther@suse.de>
6855 PR rtl-optimization/57341
6856 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
6857 instead of true_dependence.
6859 2013-05-22 David Malcolm <dmalcolm@redhat.com>
6861 * bb-reorder.c (branch_threshold): Make const.
6862 (exec_threshold): Ditto.
6864 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6865 Pat Haugen <pthaugen@us.ibm.com>
6866 Peter Bergner <bergner@vnet.ibm.com>
6868 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
6869 documentation for the power8 crypto builtins.
6871 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
6873 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
6874 macros for defining power8 builtin functions.
6875 (BU_P8V_AV_2): Likewise.
6876 (BU_P8V_AV_P): Likewise.
6877 (BU_P8V_VSX_1): Likewise.
6878 (BU_P8V_OVERLOAD_1): Likewise.
6879 (BU_P8V_OVERLOAD_2): Likewise.
6880 (BU_CRYPTO_1): Likewise.
6881 (BU_CRYPTO_2): Likewise.
6882 (BU_CRYPTO_3): Likewise.
6883 (BU_CRYPTO_OVERLOAD_1): Likewise.
6884 (BU_CRYPTO_OVERLOAD_2): Likewise.
6885 (XSCVSPDP): Fix typo, point to the correct instruction.
6886 (VCIPHER): Add power8 crypto builtins.
6887 (VCIPHERLAST): Likewise.
6888 (VNCIPHER): Likewise.
6889 (VNCIPHERLAST): Likewise.
6890 (VPMSUMB): Likewise.
6891 (VPMSUMH): Likewise.
6892 (VPMSUMW): Likewise.
6893 (VPERMXOR_V2DI): Likewise.
6894 (VPERMXOR_V4SI: Likewise.
6895 (VPERMXOR_V8HI: Likewise.
6896 (VPERMXOR_V16QI: Likewise.
6897 (VSHASIGMAW): Likewise.
6898 (VSHASIGMAD): Likewise.
6900 (VPERMXOR): Likewise.
6901 (VSHASIGMA): Likewise.
6903 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6904 __CRYPTO__ if the crypto instructions are available.
6905 (altivec_overloaded_builtins): Add support for overloaded power8
6908 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
6909 support for power8 crypto builtins.
6910 (builtin_function_type): Likewise.
6911 (altivec_init_builtins): Add support for builtins that take vector
6912 long long (V2DI) arguments.
6914 * config/rs6000/crypto.md: New file, define power8 crypto
6917 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6918 Pat Haugen <pthaugen@us.ibm.com>
6919 Peter Bergner <bergner@vnet.ibm.com>
6921 * doc/invoke.texi (Option Summary): Add power8 options.
6922 (RS/6000 and PowerPC Options): Likewise.
6924 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
6925 constraints.md instead of rs6000.h. Reorder w* constraints. Add
6926 wm, wn, wr documentation.
6928 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
6929 registers if direct move instructions are enabled.
6930 (wn): New constraint for no registers.
6931 (wq): New constraint for quad word even GPR registers.
6932 (wr): New constraint if 64-bit instructions are enabled.
6933 (wv): New constraint if power8 vector instructions are enabled.
6934 (wQ): New constraint for quad word memory locations.
6936 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
6937 constraint for 0..15 for crypto instructions.
6938 (gpc_reg_operand): If VSX allow registers in VSX registers as well
6939 as GPR and floating point registers.
6940 (int_reg_operand): New predicate to match only GPR registers.
6941 (base_reg_operand): New predicate to match base registers.
6942 (quad_int_reg_operand): New predicate to match even GPR registers
6943 for quad memory operations.
6944 (vsx_reg_or_cint_operand): New predicate to allow vector logical
6945 operations in both GPR and VSX registers.
6946 (quad_memory_operand): New predicate for quad memory operations.
6947 (reg_or_indexed_operand): New predicate for direct move support.
6949 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
6950 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
6951 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
6952 (POWERPC_MASKS): Add power8 options.
6953 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
6956 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6957 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
6959 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
6960 (-mpower8-fusion): New power8 options.
6961 (-mpower8-fusion-sign): Likewise.
6962 (-mpower8-vector): Likewise.
6963 (-mcrypto): Likewise.
6964 (-mdirect-move): Likewise.
6965 (-mquad-memory): Likewise.
6967 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
6969 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
6971 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
6972 (rs6000_debug_vector_unit): Add p8_vector.
6973 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
6974 definitions. Also print fusion state.
6975 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
6976 (rs6000_builtin_mask_calculate): Add power8 builtin support.
6977 (rs6000_option_override_internal): Add support for power8.
6978 (rs6000_common_init_builtins): Add debugging for skipped builtins
6980 (rs6000_adjust_cost): Add power8 support.
6981 (rs6000_issue_rate): Likewise.
6982 (insn_must_be_first_in_group): Likewise.
6983 (insn_must_be_last_in_group): Likewise.
6984 (force_new_group): Likewise.
6985 (rs6000_register_move_cost): Likewise.
6986 (rs6000_opt_masks): Likewise.
6988 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
6989 power8 capable assembler, default to power7 options.
6990 (TARGET_DIRECT_MOVE): Likewise.
6991 (TARGET_CRYPTO): Likewise.
6992 (TARGET_P8_VECTOR): Likewise.
6993 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
6994 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
6995 (VECTOR_MEM_P8_VECTOR_P): Likewise.
6996 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
6997 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
6998 (TARGET_XSCVDPSPN): Likewise.
6999 (TARGET_XSCVSPDPN): Likewsie.
7000 (TARGET_SYNC_HI_QI): Likewise.
7001 (TARGET_SYNC_TI): Likewise.
7002 (MASK_CRYPTO): Likewise.
7003 (MASK_DIRECT_MOVE): Likewise.
7004 (MASK_P8_FUSION): Likewise.
7005 (MASK_P8_VECTOR): Likewise.
7006 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
7007 temporary used by some of the direct move instructions to get two FP
7008 temporary registers does not force creation of a stack frame.
7009 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
7010 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
7011 that any VSX registers are tieable, even if they are also an
7012 Altivec vector mode.
7013 (r6000_reg_class_enum): Add wm, wr, wv constraints.
7014 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
7015 (RS6000_BTM_CRYPTO): Likewise.
7016 (RS6000_BTM_COMMON): Likewise.
7018 * config/rs6000/rs6000.md (cpu attribute): Add power8.
7019 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
7020 (enum rs6000_vector): Add power8 vector support.
7022 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7026 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7027 (any_sibcall_could_use_r3): this and handle indirect calls.
7028 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7030 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7032 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7034 2013-05-22 Richard Biener <rguenther@suse.de>
7037 * profile.c (branch_prob): Do not split blocks that are
7038 abnormally receiving from ECF_RETURNS_TWICE functions.
7040 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7042 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7043 address mode. Move pointer mode initialization to the same place.
7045 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7047 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7048 while it has any effect.
7050 2013-05-21 Easwaran Raman <eraman@google.com>
7052 PR tree-optimization/57322
7053 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7054 UID of the statement added to the BB to be 1.
7056 2013-05-21 Jakub Jelinek <jakub@redhat.com>
7058 PR tree-optimization/57331
7059 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7060 of conversion from pointer type to integral type with integer.
7062 2013-05-21 Martin Jambor <mjambor@suse.cz>
7065 * ipa-prop.c (ipa_read_node_info): Process param_used and
7066 controlled_uses in the same order as when writing.
7068 2013-05-21 Magnus Granberg <baldrick@free.fr>
7071 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
7073 2013-05-21 Richard Biener <rguenther@suse.de>
7075 PR tree-optimization/57318
7076 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7077 estimate stmts with side-effects as likely eliminated.
7079 2013-05-21 Richard Biener <rguenther@suse.de>
7081 PR tree-optimization/57330
7082 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7083 preserve the call stmts fntype.
7085 2013-05-21 Richard Biener <rguenther@suse.de>
7087 PR tree-optimization/57303
7088 * tree-ssa-sink.c (statement_sink_location): Improve killing
7089 stmt detection and properly handle self-assignments.
7091 2013-05-21 Christian Bruel <christian.bruel@st.com>
7093 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7094 spanning registers. LEAF_REG_REMAP is supported only for contiguous
7095 registers. Set register size out of the PARALLEL loop.
7097 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7100 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7101 (*fmasf4, *fmasf4_media): New insns.
7103 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7105 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7106 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7107 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7108 (mips_idiv_insns): Update the comments to say that the returned
7109 instruction counts are in units of BASE_INSN_LENGTH.
7110 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7111 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7112 using 2 rather than 4 as the length of indirect MIPS16 and
7113 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7114 length of a NOP. Don't divide MIPS16 lengths by 2.
7115 (mips16_split_long_branches): Assume a branch is long if the
7116 length is greater than 4 rather than 8.
7117 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7118 rather than multiplying them by 2. Multiply instruction counts
7119 by BASE_INSN_LENGTH rather than 4.
7120 (*jump_mips16, tls_get_tp_mips16_<mode>)
7121 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7123 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7125 * config/mips/mips.md (extended_mips16): Remove branch case.
7126 (length): Remove duplicated extended_mips16 test.
7128 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7130 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7132 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7134 * recog.h (Recog_data): Rename to...
7135 (recog_data_d): ...this.
7136 (recog_data): Update accordingly.
7137 * recog.c (recog_data): Likewise.
7138 * reload.c (save_recog_data): Likewise.
7139 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7140 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7142 2013-05-17 Julian Brown <julian@codesourcery.com>
7144 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7145 found in a REG_EQUAL note, invalidate it.
7147 2013-05-17 Easwaran Raman <eraman@google.com>
7149 * tree-ssa-reassoc.c (find_insert_point): New function.
7150 (insert_stmt_after): Likewise.
7151 (get_def_stmt): Likewise.
7152 (ensure_ops_are_available): Likewise.
7153 (not_dominated_by): Likewise.
7154 (rewrite_expr_tree): Do not move statements beyond what is
7155 necessary. Remove call to swap_ops_for_binary_stmt...
7156 (reassociate_bb): ... and move it here.
7157 (build_and_add_sum): Assign UIDs for new statements.
7158 (linearize_expr): Likewise.
7159 (do_reassoc): Renumber gimple statement UIDs.
7161 2013-05-17 Jan Hubicka <jh@suse.cz>
7163 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7165 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
7166 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7168 (output_weakrefs): Update.
7170 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7171 Martin Jambor <mjambor@suse.cz>
7174 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7175 value that corresponds to the given aggval is found in values vector.
7177 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
7179 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7180 sse, sse2, sse3, ssse3 and sse4a flags to options.
7182 2013-05-17 David Malcolm <dmalcolm@redhat.com>
7184 * gengtype-state.c: (s_expr_writer): New class, to handle
7185 prettifying of output layout of s-expressions.
7186 (state_writer): New class, to write out gtype.state.
7187 (state_written_type_count): Move this variable into member data of
7189 (s_expr_writer::s_expr_writer): New code: constructor for new class
7190 (state_writer::state_writer(): ditto
7191 (s_expr_writer::write_new_line): New function
7192 (s_expr_writer::write_any_indent): ditto
7193 (s_expr_writer::begin_s_expr): ditto
7194 (s_expr_writer::end_s_expr): ditto
7195 (write_state_fileloc): convert to method of state_writer...
7196 (state_writer:: write_state_fileloc): ...and use methods of
7197 s_expr_writer to write indentation into the gtype.state output file
7198 to visually represent the hierarchical structure of the list
7200 (write_state_fields): ditto, renaming to...
7201 (state_writer::write_state_fields)
7202 (write_state_a_string): ditto, renaming to...
7203 (state_writer::write_state_a_string)
7204 (write_state_string_option): ditto, renaming to...
7205 (state_writer::write_state_string_option)
7206 (write_state_type_option): ditto, renaming to...
7207 (state_writer::write_state_type_option)
7208 (write_state_nested_option): ditto, renaming to...
7209 (state_writer::write_state_nested_option)
7210 (write_state_option): ditto, renaming to...
7211 (state_writer::write_state_option)
7212 (write_state_options): ditto, renaming to...
7213 (state_writer::write_state_options)
7214 (write_state_lang_bitmap): ditto, renaming to...
7215 (state_writer::write_state_lang_bitmap)
7216 (write_state_version): ditto, renaming to...
7217 (state_writer::write_state_version)
7218 (write_state_scalar_type): ditto, renaming to...
7219 (state_writer::write_state_scalar_type)
7220 (write_state_string_type): ditto, renaming to...
7221 (state_writer::write_state_string_type)
7222 (write_state_undefined_type): ditto, renaming to...
7223 (state_writer::write_state_undefined_type)
7224 (write_state_struct_union_type): ditto, renaming to...
7225 (state_writer::write_state_struct_union_type)
7226 (write_state_struct_type): ditto, renaming to...
7227 (state_writer::write_state_struct_type)
7228 (write_state_user_struct_type): ditto, renaming to...
7229 (state_writer::write_state_user_struct_type)
7230 (write_state_lang_struct_type): ditto, renaming to...
7231 (state_writer::write_state_lang_struct_type)
7232 (write_state_param_struct_type): ditto, renaming to...
7233 (state_writer::write_state_param_struct_type)
7234 (write_state_pointer_type): ditto, renaming to...
7235 (state_writer::write_state_pointer_type)
7236 (write_state_array_type): ditto, renaming to...
7237 (state_writer::write_state_array_type)
7238 (write_state_gc_used): ditto, renaming to...
7239 (state_writer::write_state_gc_used)
7240 (write_state_common_type_content): ditto, renaming to...
7241 (state_writer::write_state_common_type_content)
7242 (write_state_type): ditto, renaming to...
7243 (state_writer::write_state_type)
7244 (write_state_pair_list): ditto, renaming to...
7245 (state_writer::write_state_pair_list)
7246 (write_state_pair): ditto, renaming to...
7247 (state_writer::write_state_pair)
7248 (write_state_typedefs): ditto, renaming to...
7249 (state_writer::write_state_typedefs)
7250 (write_state_structures): ditto, renaming to...
7251 (state_writer::write_state_structures)
7252 (write_state_param_structs): ditto, renaming to...
7253 (state_writer::write_state_param_structs)
7254 (write_state_variables): ditto, renaming to...
7255 (state_writer::write_state_variables)
7256 (write_state_srcdir): ditto, renaming to...
7257 (state_writer::write_state_srcdir)
7258 (write_state_files_list): ditto, renaming to...
7259 (state_writer::write_state_files_list)
7260 (write_state_languages): ditto, renaming to...
7261 (state_writer::write_state_languages)
7262 (write_state): create a state_writer instance and use it when
7263 writing out the state file
7265 2013-05-17 Mike Stump <mikestump@comcast.net>
7267 PR rtl-optimization/57304
7268 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7269 accessing DF_REF_REAL_LOC.
7271 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7273 PR rtl-optimization/57281
7274 PR rtl-optimization/57300
7275 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7276 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7277 what the other splitter did if the registers are dead.
7279 2013-05-17 Richard Biener <rguenther@suse.de>
7281 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7284 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7286 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7289 2013-05-17 Marek Polacek <polacek@redhat.com>
7291 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7292 length when doing non-zero store of storing '\0' to '\0'.
7294 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7296 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7297 vect_external_def oprnd1 with loop_vinfo, try to emit
7298 optional cast, negation and and stmts on the loop preheader
7299 edge instead of into the pattern def seq.
7301 PR tree-optimization/57051
7302 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7303 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7305 2013-05-16 Nick Clifton <nickc@redhat.com>
7307 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7308 (rl78_is_naked_func): New function.
7309 (rl78_expand_prologue): Skip prologue generation for naked functions.
7310 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
7311 * doc/extend.texi (naked): Add RL78 to the list of processors
7312 that supports this attribute.
7314 2013-05-16 Jeff Law <law@redhat.com>
7316 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7318 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7320 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7321 cache parameters using detect_caches_amd also for CYRIX,
7322 NSC and TM2 signatures.
7324 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7325 Dzianis Kahanovich <mahatma@eu.by>
7329 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7330 VIA/Centaur processors and determine their cache parameters
7331 using detect_caches_amd.
7333 2013-05-16 Teresa Johnson <tejohnson@google.com>
7335 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7336 (rtl_verify_edges): New function.
7337 (rtl_verify_bb_insns): Ditto.
7338 (rtl_verify_bb_pointers): Ditto.
7339 (rtl_verify_bb_insn_chain): Ditto.
7340 (rtl_verify_fallthru): Ditto.
7341 (rtl_verify_bb_layout): Ditto.
7342 (rtl_verify_flow_info_1): Outline checks into new functions.
7343 (rtl_verify_flow_info): Ditto.
7345 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
7347 * cfghooks.c (copy_bbs): Add update_dominance argument.
7348 * cfghooks.h (copy_bbs): Update prototype.
7349 * tree-cfg.c (gimple_duplicate_sese_region):
7350 Add update_dominance argument.
7351 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7352 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7353 gimple_duplicate_sese_region call.
7354 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7355 Update copy_bbs call.
7356 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7357 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7359 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7361 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7362 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7363 vect_recog_rotate_pattern.
7364 (vect_recog_rotate_pattern): New function.
7366 2013-05-16 Jason Merrill <jason@redhat.com>
7368 * Makefile.in (LLINKER): New variable.
7369 (mostlyclean): Remove link mutex.
7370 * configure.ac: Handle --enable-link-mutex.
7371 * lock-and-run.sh: New script.
7373 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7376 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
7379 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7381 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
7383 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
7385 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
7386 * config/arm/arm.c (next_consecutive_mem): New function.
7387 (gen_movmem_ldrd_strd): Likewise.
7388 * config/arm/arm.md (movmemqi): Update condition and code.
7389 (unaligned_loaddi, unaligned_storedi): New patterns.
7391 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7393 * config.gcc: Obsolete *-*-solaris2.9*.
7394 * doc/install.texi (Specific, *-*-solaris2*): Document it.
7396 2013-05-16 Richard Biener <rguenther@suse.de>
7398 * passes.c (init_optimization_passes): Move pass_parallelize_loops
7399 earlier, after GRAPHITE transforms and IV canonicalization.
7401 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7403 * omp-low.c (extract_omp_for_data): For collapsed loops,
7404 if at least one of the loops is known at compile time to
7405 iterate zero times, set count to 0.
7406 (expand_omp_regimplify_p): New function.
7407 (expand_omp_for_generic): For collapsed loops, if at least
7408 one of the loops isn't known to iterate at least once,
7409 add runtime check with setting count to 0.
7410 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7411 For unsigned types if it isn't known at compile time that
7412 the loop will iterate at least once, add runtime check to bypass
7413 the whole loop if initial condition isn't true.
7415 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
7417 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
7419 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
7422 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
7423 transformations to avoid an infinite loop.
7425 2013-05-16 Marek Polacek <polacek@redhat.com>
7427 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
7429 2013-05-15 Leif Ekblad <leif@rdos.net>
7431 * config/i386/i386.c (ix86_decompose_address): Use
7432 DEFAULT_TLS_SEG_REG to access TLS segment register.
7433 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
7434 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
7435 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
7437 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
7440 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
7441 sibling calls to functions that would normally be lazily bound,
7442 unless $gp is call-clobbered.
7444 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
7446 * config/i386/i386.c (ix86_option_override_internal): Update
7447 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
7448 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
7449 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
7450 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
7452 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
7454 2013-05-15 Andreas Schwab <schwab@suse.de>
7456 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
7457 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
7460 2013-05-15 Teresa Johnson <tejohnson@google.com>
7462 * loop-unroll.c (report_unroll_peel): Check decision before
7463 emitting unroll/peel message.
7465 2013-05-15 Teresa Johnson <tejohnson@google.com>
7467 * function.h (has_bb_partition): New rtl_data flag.
7468 (bb_reorder_complete): Ditto.
7469 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
7470 instead of flag_reorder_blocks_and_partition.
7471 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
7472 with some enhancements.
7473 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
7474 * bb-reorder.c (connect_traces): Check for has_bb_partition
7475 instead of flag_reorder_blocks_and_partition.
7476 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
7477 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
7478 verify_hot_cold_block_grouping.
7479 (partition_hot_cold_basic_blocks): Set has_bb_partition.
7481 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7484 * config/arm/predicates.md (call_insn_operand): New predicate.
7485 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
7486 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
7487 if insn is not a tail call.
7488 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
7490 * config/arm/arm.h (enum reg_class): New caller save register class.
7491 (REG_CLASS_NAMES): Likewise.
7492 (REG_CLASS_CONTENTS): Likewise.
7493 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
7496 2013-05-15 Richard Biener <rguenther@suse.de>
7498 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
7499 of MSG_OPTIMIZED_LOCATIONS.
7500 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
7501 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
7503 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
7504 of MSG_OPTIMIZED_LOCATIONS.
7505 (execute_vect_slp): Likewise.
7506 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
7507 (vect_create_cond_for_alias_checks): Likewise.
7508 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
7509 (vect_recog_widen_mult_pattern): Likewise.
7510 (vect_recog_widen_sum_pattern): Likewise.
7511 (vect_recog_over_widening_pattern): Likewise.
7512 (vect_recog_widen_shift_pattern): Likewise.
7513 (vect_recog_vector_vector_shift_pattern): Likewise.
7514 (vect_recog_divmod_pattern): Likewise.
7515 (vect_recog_mixed_size_cond_pattern): Likewise.
7516 (vect_recog_bool_pattern): Likewise.
7517 (vect_pattern_recog_1): Likewise.
7519 2013-05-15 Martin Jambor <mjambor@suse.cz>
7521 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
7522 non-functions to builtin_unreachable.
7523 * ipa-inline-transform.c (inline_call): Do not assert estimates were
7524 correct when new direct edges were discovered.
7526 2013-05-15 Martin Jambor <mjambor@suse.cz>
7528 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
7529 header, print symbol order instead of node uid, print more information
7530 about indirect edge targets.
7531 (ipa_make_edge_direct_to_target): Print symbol order instead of node
7533 (ipa_make_edge_direct_to_target): Likewise.
7534 (remove_described_reference): Likewise.
7535 (propagate_controlled_uses): Likewise.
7536 (ipa_print_node_params): Also print symbol order.
7537 (ipcp_transform_function): Print symbol order instead of node uids.
7538 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
7539 (cgraph_get_create_real_symbol_node): Likewise.
7540 * ipa-cp.c (print_lattice): Likewise.
7541 (print_all_lattices): Likewise.
7542 (determine_versionability): Likewise.
7543 (initialize_node_lattices): Likewise.
7544 (estimate_local_effects): Likewise.
7545 (update_profiling_info): Likewise.
7546 (create_specialized_node): Likewise.
7547 (perhaps_add_new_callers): Likewise.
7548 (decide_about_value): Likewise.
7549 (decide_whether_version_node): Likewise.
7550 (identify_dead_nodes): Likewise.
7551 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
7552 (dump_inline_summary): Likewise.
7553 (estimate_node_size_and_time): Likewise.
7554 (inline_analyze_function): Likewise.
7555 * ipa-inline.c (report_inline_failed_reason): Likewise.
7556 (want_early_inline_function_p): Likewise.
7557 (edge_badness): Likewise.
7558 (update_edge_key): Likewise.
7559 (inline_small_functions): Likewise. Add dumping of order to two other
7561 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
7562 instead of node uids.
7563 (propagate_pure_const): Likewise.
7564 (propagate_pure_const): Likewise.
7565 * ipa-utils.c (dump_cgraph_node_set): Likewise.
7566 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
7567 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
7569 * tree-pretty-print.c (dump_function_header): Likewise.
7570 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
7571 Print symbol order instead of node uids.
7573 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7575 * config/s390/s390.c (s390_register_move_cost): Don't impose the
7576 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
7578 2013-05-15 Richard Biener <rguenther@suse.de>
7580 PR tree-optimization/57275
7581 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
7582 return value for fail to do runtime alias checks for gather loads.
7584 2013-05-15 Jan Hubicka <jh@suse.cz>
7588 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
7589 weakrefs are not external.
7590 (lto_symtab_merge_decls): Fix thinko when dealing with
7591 non-lto_symtab decls.
7592 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
7593 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
7594 * varpool.c (dump_varpool_node): Dump more flags.
7596 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7598 * config/i386/i386.c (processor_alias_table): Add instruction
7599 FSGSBASE for AMD bdver3 architecture.
7601 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7603 * tree.c (warn_deprecated_use): Print file:line using locus color.
7604 * diagnostic.c (diagnostic_report_current_module): Print file:line
7605 and file:line:column using locus color.
7607 2013-05-14 Mike Stump <mikestump@comcast.net>
7609 * gdbinit.in: Add __null.
7611 2013-05-14 Mike Stump <mikestump@comcast.net>
7613 * recog.h: Rename struct recog_data to Recog_data.
7614 * recog.c: Likewise.
7615 * reload.c (can_reload_into): Likewise.
7616 * config/picochip/picochip.c: Likewise.
7618 2013-05-14 Mike Stump <mikestump@comcast.net>
7620 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
7622 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
7624 * resource.h (struct resources): Remove unch_memory member.
7625 (CLEAR_RESOURCE): Don't clear unch_memory.
7626 * resource.c (mark_referenced_resources): Don't set it.
7627 (mark_set_resources): Likewise.
7628 (mark_target_live_regs): Don't clear it.
7629 (init_resource_info): Likewise.
7630 * reorg.c (resource_conflicts_p): Don't compare it.
7631 (redundant_insn): Don't set it.
7633 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
7635 * emit-rtl.c (next_label): Remove unused function.
7636 (skip_consecutive_labels, link_cc0_insns): Move to ...
7637 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
7638 only place where these functions are used, and make them static.
7640 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
7642 * fold-const.c (fold_negate_expr): Handle vectors.
7643 (fold_truth_not_expr): Make it static.
7644 (fold_invert_truthvalue): New static function.
7645 (invert_truthvalue_loc): Handle vectors. Do not call
7646 fold_truth_not_expr directly.
7647 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
7648 <TRUTH_NOT_EXPR>: Do not cast to boolean.
7649 (fold_comparison): Handle vector constants.
7650 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
7651 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
7652 * tree.h (fold_truth_not_expr): Remove declaration.
7654 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
7656 * config/aarch64/aarch64-simd.md
7657 (aarch64_vcond_internal<mode>): Rename to...
7658 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
7659 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
7660 float modes. Clarify all iterator modes.
7661 (vcond<mode><mode>): Use new name for vcond expanders.
7662 (vcond<v_cmp_result><mode>): Likewise.
7663 (vcondu<mode><mode>: Likewise.
7664 * config/aarch64/iterators.md (VDQF_COND): New.
7666 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
7669 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
7670 variable for the shift amount. Check that we shift by non-negative
7673 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
7676 * config/arm/arm.h (EPILOGUE_USES): Only return true
7677 for LR_REGNUM after epilogue_completed.
7679 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
7681 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
7684 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7687 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
7688 and Solaris 11+/x86 with gld.
7689 * configure: Regenerate.
7691 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7693 * expmed.c (expand_shift_1): Canonicalize rotates by
7694 constant bitsize / 2 to bitsize - 1.
7695 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
7696 case ROTATERT>: Likewise.
7699 2013-05-10 Jakub Jelinek <jakub@redhat.com>
7701 * config/i386/i386.md (rotateinv): New code attr.
7702 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7703 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7704 roll $31, %eax, etc.
7706 2013-05-14 Richard Biener <rguenther@suse.de>
7709 * tree-eh.c (sink_clobbers): Give up for successors with
7710 multiple predecessors and no virtual uses.
7712 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
7714 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
7715 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
7717 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7720 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
7721 the case when both op0 and op1 have VOIDmode.
7723 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7725 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
7726 in multiply-accumulate mode.
7728 2013-05-13 Guozhi Wei <carrot@google.com>
7730 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
7732 2013-05-13 Kai Tietz <ktietz@redhat.com>
7735 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
7736 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
7737 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
7738 * config/i386/i386.c (ix86_option_override_internal): Likewise.
7739 (ix86_expand_prologue): Likewise.
7740 (ix86_expand_split_stack_prologue): Likewise.
7741 (legitimate_pic_address_disp_p): Likewise.
7742 (legitimize_pic_address): Likewise.
7743 (legitimize_tls_address): Likewise.
7744 (legitimize_pe_coff_symbol): Likewise.
7745 (output_pic_addr_const): Likewise.
7746 (construct_plt_address): Likewise.
7747 (ix86_expand_call): Likewise.
7748 (x86_output_mi_thunk): Likewise.
7749 (x86_function_profiler): Likewise.
7751 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
7753 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
7754 similar switch cases.
7755 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
7756 (aarch64_simd_mov_to_<mode>low): Delete.
7757 (aarch64_simd_mov_to_<mode>high): Delete.
7758 (move_lo_quad_<mode>): Add w<-r alternative.
7759 (aarch64_simd_move_hi_quad_<mode>): Likewise.
7760 (aarch64_simd_mov_from_*): Update type attribute.
7761 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
7764 2013-05-13 Jan Hubicka <jh@suse.cz>
7766 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
7767 * config/i386/i386.c (ix86_compute_frame_layout,
7768 ix86_expand_epilogue, emit_i387_cw_initialization,
7769 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
7770 ix86_local_alignment): Fix use of size/speed predicates.
7772 2013-05-13 Jakub Jelinek <jakub@redhat.com>
7774 PR tree-optimization/45216
7775 PR tree-optimization/57157
7776 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
7777 the (-Y) & (B - 1) variant if OP is |.
7778 * expmed.c (expand_shift_1): For rotations by const0_rtx just
7779 return shifted. Use (-op1) & (prec - 1) as other_amount
7780 instead of prec - op1.
7782 2013-05-13 Martin Jambor <mjambor@suse.cz>
7785 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
7786 (ipa_constant_data): New type.
7787 (ipa_jump_func): Use ipa_constant_data to hold information about
7788 constant jump functions.
7789 (ipa_get_jf_constant): Adjust to jump function type changes.
7790 (ipa_get_jf_constant_rdesc): New function.
7791 (ipa_param_descriptor): New field controlled_uses.
7792 (ipa_get_controlled_uses): New function.
7793 (ipa_set_controlled_uses): Likewise.
7794 * ipa-ref.h (ipa_find_reference): Declare.
7795 * ipa-prop.c (ipa_cst_ref_desc): New type.
7796 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
7798 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
7799 New parameter cs. Adjust all callers.
7800 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
7801 (remove_described_reference): New function.
7802 (jfunc_rdesc_usable): Likewise.
7803 (try_make_edge_direct_simple_call): Decrement controlled use count,
7804 attempt to remove reference if it hits zero.
7805 (combine_controlled_uses_counters): New function.
7806 (propagate_controlled_uses): Likewise.
7807 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
7808 (ipa_edge_duplication_hook): Duplicate reference descriptions.
7809 (ipa_print_node_params): Print described use counter.
7810 (ipa_write_jump_function): Adjust to jump function type changes.
7811 (ipa_read_jump_function): New parameter CS, pass it to
7812 ipa_set_jf_constant. Adjust caller.
7813 (ipa_write_node_info): Stream controlled use count
7814 (ipa_read_node_info): Likewise.
7815 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
7817 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
7818 count. Remove cloning-added reference if it reaches zero.
7819 * ipa-ref.c (ipa_find_reference): New function.
7821 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7823 * config/i386/i386.c (processor_target_table): Modified default
7824 alignment values for AMD BD and BT architectures.
7826 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
7828 * tree-vect-generic.c (uniform_vector_p): Move ...
7829 * tree.c (uniform_vector_p): ... here.
7830 * tree.h (uniform_vector_p): Declare it.
7831 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
7834 2013-05-13 Jakub Jelinek <jakub@redhat.com>
7836 PR tree-optimization/57230
7837 * tree-ssa-strlen.c (handle_char_store): Record length for
7838 array store from STRING_CST.
7840 PR tree-optimization/57230
7841 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
7844 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
7846 * config/epiphany/epiphany.c (epiphany_init): Check size of
7847 NUM_MODES_FOR_MODE_SWITCHING.
7848 (epiphany_expand_prologue):
7849 Remove CONFIG_REGNUM initial value handling code.
7850 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
7851 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
7852 (emit_set_fp_mode, epiphany_mode_after): Likewise.
7853 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
7854 Don't return 1 for FP_MODE_NONE.
7855 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
7856 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
7857 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
7858 * config/epiphany/epiphany.md (save_config): New pattern.
7860 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
7862 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
7864 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
7866 * config/i386/i386.md (memory): Handle sseishft1.
7867 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
7868 (*vec_extractv2di_1): Ditto.
7870 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
7872 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
7875 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
7877 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
7878 Add mthumb/march=armv7-a multilib.
7879 Add mthumb/march=armv7-r multilib.
7880 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
7882 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
7884 * config/v850/t-rtems: Add more multilibs.
7886 2013-05-10 Richard Biener <rguenther@suse.de>
7888 PR tree-optimization/57214
7889 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
7890 not propagate from SSA names that occur in abnormal PHI nodes.
7892 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
7894 * stor-layout.c (element_precision): New function.
7895 * machmode.h (element_precision): Declare it.
7896 * tree.c (build_minus_one_cst): New function.
7897 (element_precision): Likewise.
7898 * tree.h (build_minus_one_cst): Declare new function.
7899 (element_precision): Likewise.
7900 * fold-const.c (operand_equal_p): Use element_precision.
7901 (fold_binary_loc): Handle vector types.
7902 * convert.c (convert_to_integer): Use element_precision.
7903 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
7906 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
7908 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
7909 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
7910 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
7911 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7912 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
7913 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
7914 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
7915 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7916 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
7917 (Uuw8): New constraints.
7918 (Usb4): Move into alphabetical order.
7919 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
7920 (sd8_operand, ub8_operand, uw8_operand): New predicates.
7921 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
7922 previously unnamed patterns.
7923 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
7924 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
7925 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
7926 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
7927 of set_attr_alternative/if_then_else. Use extended_mips16 instead
7928 of specific lengths.
7930 2013-05-10 Jakub Jelinek <jakub@redhat.com>
7932 * config/i386/i386.md (rotateinv): New code attr.
7933 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7934 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7935 roll $31, %eax, etc.
7937 PR tree-optimization/45216
7938 PR tree-optimization/57157
7939 * tree-ssa-forwprop.c (simplify_rotate): New function.
7940 (ssa_forward_propagate_and_combine): Call it.
7942 2013-05-10 Richard Biener <rguenther@suse.de>
7944 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
7945 disable peeling when we version for aliasing.
7946 (vector_alignment_reachable_p): Honor explicit user alignment.
7947 (vect_supportable_dr_alignment): Likewise.
7948 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
7949 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
7950 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
7951 then peeling to arrange for the cost-model check to come first.
7953 2013-05-10 Alan Modra <amodra@gmail.com>
7955 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
7956 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
7957 * configure: Regenerate.
7959 2013-05-10 Alan Modra <amodra@gmail.com>
7962 * varasm.c (default_elf_select_section): Move !DECL_P check..
7963 (get_named_section): ..to here before calling get_section_name.
7965 (default_section_type_flags): Add DECL_P check.
7966 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
7967 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
7969 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
7971 * config/epiphany/epiphany.c (epiphany_expand_prologue):
7972 When using gen_stack_adjust_str with a register offset, add a
7973 REG_FRAME_RELATED_EXPR note.
7975 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
7977 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
7978 (*vec_extractv4si_zext_mem): Ditto.
7979 (*vec_extractv2di): Add 0->x and x->x alternatives.
7980 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
7981 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
7983 2013-05-09 Jason Merrill <jason@redhat.com>
7985 N3639 C++1y VLA support
7986 * gimplify.c (gimplify_vla_decl): Don't touch an existing
7989 * tree.c (build_constructor_va): New.
7990 * tree.h: Declare it.
7992 2013-05-09 Martin Jambor <mjambor@suse.cz>
7995 * gimple-fold.c (canonicalize_constructor_val): Call
7996 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
7998 2013-05-09 Jan Hubicka <jh@suse.cz>
7999 Richard Biener <rguenther@suse.de>
8002 * symtab.c (symtab_make_decl_local): Do not add private names.
8004 2013-05-09 Jan Hubicka <jh@suse.cz>
8007 * symtab.c (insert_to_assembler_name_hash): Handle clones.
8008 (unlink_from_assembler_name_hash): Likewise.
8009 (symtab_prevail_in_asm_name_hash, symtab_register_node,
8010 symtab_unregister_node, symtab_initialize_asm_name_hash,
8011 change_decl_assembler_name): Update.
8013 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
8015 * config/aarch64/aarch64.md: New movtf split.
8016 (*movtf_aarch64): Update.
8017 (aarch64_movdi_tilow): Handle TF modes and rename to
8018 aarch64_movdi_<mode>low.
8019 (aarch64_movdi_tihigh): Handle TF modes and rename to
8020 aarch64_movdi_<mode>high
8021 (aarch64_movtihigh_di): Handle TF modes and rename to
8022 aarch64_mov<mode>high_di
8023 (aarch64_movtilow_di): Handle TF modes and rename to
8024 aarch64_mov<mode>low_di
8025 (aarch64_movtilow_tilow): Remove spurious whitespace.
8026 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8028 (aarch64_print_operand): Update.
8030 2013-05-09 Alan Modra <amodra@gmail.com>
8032 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8034 * configure: Regenerate.
8036 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8038 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8039 splitter preparation statements.
8040 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8041 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8044 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8046 * gimple-ssa-strength-reduction.c (count_candidates): Change
8047 return value to int.
8048 (analyze_candidates_and_replace): Change type of length to int.
8050 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8052 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8053 (*vec_extract<mode>): Use VI12_128 mode iterator.
8054 (*vec_extract<mode>_mem): Ditto.
8055 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8058 2013-05-08 Diego Novillo <dnovillo@google.com>
8063 2012-08-17 Diego Novillo <dnovillo@google.com>
8066 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8067 * config.in: Regenerate.
8068 * configure: Regenerate.
8069 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
8071 2013-05-08 Jan Hubicka <jh@suse.cz>
8074 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8075 * cgraph.h (symtab_node_base): Add unique_name.
8076 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8077 input_overwrite_node, input_varpool_node): Stream unique_name.
8078 * cgraphclones.c (cgraph_create_virtual_clone,
8079 cgraph_function_versioning): Set unique_name.
8080 * ipa.c (function_and_variable_visibility): Set unique_name.
8082 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8084 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8085 (alloc_cand_and_find_basis): Restrict conditional candidate
8086 processing to CAND_MULTs.
8088 2013-05-08 Jan Hubicka <jh@suse.cz>
8091 lto-symtab.c (lto_symtab_symbol_p): New function.
8092 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8093 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8094 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8095 Skip static symbols.
8097 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8099 PR tree-optimization/57200
8100 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8101 Only call inform if the preceding warning_at returns true.
8103 2013-05-07 Han Shen <shenhan@google.com>
8105 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8106 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8107 * common.opt (fstack-protector-strong): New option.
8108 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8109 * doc/invoke.texi (Optimization Options): Document
8110 "-fstack-protector-strong".
8111 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8113 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8115 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8117 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8119 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8120 it is not lhs of the insn.
8121 (setup_reg_equiv): Remove insn having equiv memory which it is not
8123 * lra-constraints.c (process_address): Try to improve generation
8124 code for address base + disp.
8125 (lra_constraints): Make correct the code for checking insn setting
8126 up backward equivalence. Remove insn only if it is in the init
8128 * lra-eliminations.c (update_reg_eliminate): Change return value.
8129 (lra_eliminate): Use the result.
8131 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
8133 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8134 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8135 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8136 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8137 register target operands.
8138 (*vec_extractv8hi_sse2): New pattern.
8139 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8140 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8141 (*vec_extract<mode>_mem): New insn and split pattern.
8143 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8145 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8146 (TARGET_ASAN_SHADOW_OFFSET): Define.
8147 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8148 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8150 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8152 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8153 (incr_vec_index): Return -1 if increment not found.
8154 (create_add_on_incoming_edge): Assert if increment not found.
8155 (record_increment): Limit number of increments recorded.
8156 (all_phi_incrs_profitable): Return false if an increment not found.
8157 (replace_profitable_candidates): Don't process increments that were
8159 (analyze_candidates_and_replace): Limit size of incr_vec.
8161 2013-05-07 Richard Biener <rguenther@suse.de>
8163 * calls.c (special_function_p): setjmp-like functions are leaf.
8164 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8165 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8167 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8169 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8170 (aarch64_simd_mov<mode>): New expander.
8171 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8172 (aarch64_simd_mov_to_<mode>high): Likewise.
8173 (aarch64_simd_mov_from_<mode>low): Likewise.
8174 (aarch64_simd_mov_from_<mode>high): Likewise.
8175 (aarch64_dup_lane<mode>): Update.
8176 (aarch64_dup_lanedi): New instruction pattern.
8177 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8178 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8180 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8182 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8183 (replace_mult_candidate): Remove unnecessary argument; remove
8184 unnecessary parameter from call to introduce_cast_before_cand.
8185 (replace_unconditional_candidate): Remove unnecessary parameter
8186 from call to replace_mult_candidate.
8187 (replace_conditional_candidate): Likewise.
8188 (insert_initializers): Use make_temp_ssa_name.
8189 (introduce_cast_before_cand): Remove unnecessary argument; use
8191 (replace_one_candidate): Remove unnecessary argument; remove
8192 unnecessary parameter from calls to introduce_cast_before_cand.
8193 (replace_profitable_candidates): Remove unnecessary parameters
8194 from calls to replace_one_candidate.
8196 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8198 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8199 phi def as possibly hiding a basis for a CAND_ADD whose operands
8200 have been commuted in the analysis.
8201 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8203 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8205 * config/aarch64/aarch64.md
8206 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8207 shift value between 0-4.
8209 2013-05-07 Richard Biener <rguenther@suse.de>
8211 * double-int.h (rshift): New overload.
8212 * double-int.c (rshift): New function.
8213 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8214 (create_reference_ops_from_ref): Remove.
8215 (vn_reference_insert): Use shared ops for constructing the
8216 reference and copy it.
8218 2013-05-07 Richard Biener <rguenther@suse.de>
8221 * tree-eh.c (sink_clobbers): Properly propagate
8222 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8224 2013-05-07 Jakub Jelinek <jakub@redhat.com>
8226 PR tree-optimization/57149
8227 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8228 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8229 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8230 uninit_undefined_value_p instead of ssa_undefined_value_p.
8233 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8234 for modifier == EXPAND_INITIALIZER.
8236 2013-05-07 Anton Blanchard <anton@samba.org>
8238 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8239 for powerpc64 little endian.
8240 * configure: Regenerate.
8242 2013-05-06 Graham Stott <grahams@btinternet.com>
8244 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8245 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8246 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8247 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8249 2013-05-06 Graham Stott <grahams@btinternet.com>
8251 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8252 codes which allow non-lvalues.
8254 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
8256 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8257 components are all 1s.
8258 (integer_minus_onep): New function.
8259 * tree.h (integer_minus_onep): Declare it.
8260 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8261 integer_minus_onep instead of integer_all_onesp.
8263 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8266 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8267 variations of these patterns.
8269 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8271 * config/i386/i386.md (isa): Add x64_sse4 member.
8272 (enabled): Handle x64_sse4.
8273 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8274 instruction for 64bit SSE4_1 targets. Update insn attributes.
8275 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8276 instruction for SSE4_1 targets. Update insn attributes.
8277 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8278 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8279 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8281 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8282 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8283 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8285 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8288 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8290 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8292 * final.c (do_assembler_dialects): Don't handle curly braces and
8293 vertical bar escaped by % as dialect delimiters.
8294 (output_asm_insn): Print curly braces and vertical bar if escaped
8295 by % and ASSEMBLER_DIALECT defined.
8296 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8297 * doc/tm.texi: Regenerated.
8299 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8301 * config/mips/mips.c: Include tree-pass.h.
8302 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8303 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8304 (pass_mips_machine_reorg2): New machine specific pass.
8305 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8306 (mips_option_override): Register the new pass.
8307 * rtl.h (cleanup_barriers): Remove prototype.
8308 (dbr_schedule): Likewise.
8309 * jump.c (cleanup_barriers): Make static.
8310 * reorg.c (dbr_schedule): Likewise.
8312 2013-05-06 Richard Biener <rguenther@suse.de>
8314 PR tree-optimization/57185
8315 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8318 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8321 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8322 between "first_arg" and "insn", not "last" and "insn".
8324 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8326 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8327 (find_candidates_in_block): Re-enable slsr_process_phi.
8328 (create_phi_basis): Fix double counting of candidate adjustment.
8330 2013-05-06 Richard Biener <rguenther@suse.de>
8333 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8334 the edge is also fallthru, preserve it and just clear the
8336 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8337 also complex, preserve that and just clear the fallthru flag.
8338 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8339 update virtual operands.
8341 2013-05-06 Alan Modra <amodra@gmail.com>
8343 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8344 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8345 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8346 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8347 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8348 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8350 2013-05-06 Alan Modra <amodra@gmail.com>
8352 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8353 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8354 (DEFAULT_ASM_ENDIAN): Define.
8355 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8356 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8357 Update -K PIC clause from sysv4.h.
8358 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8359 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8361 2013-05-06 Alan Modra <amodra@gmail.com>
8363 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8364 twice for little-endian.
8365 (ashrdi3_no_power, ashrdi3): Support little-endian.
8367 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8370 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
8371 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
8373 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
8374 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
8375 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
8378 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
8379 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
8381 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
8382 * config/pa/pa.opt: Make mbig-switch a no-op.
8383 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
8384 (CASE_VECTOR_MODE): Always return SImode.
8385 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
8386 for the !TARGET_BIG_SWITCH case.
8387 * config/pa/pa-linux.h: Likewise.
8388 * config/pa/pa-openbsd.h: Likewise.
8389 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
8390 * config/pa/pa.md (short_jump): Remove define_insn.
8391 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
8392 (casesi0): Remove define_insn.
8393 (type): Remove btable_branch.
8394 (pa_combine_type): Likewise.
8395 (in_nullified_branch_delay): Likewise.
8396 (in_call_delay): Likewise.
8397 (define_delay): Likewise.
8398 (define_insn_reservation "Z3"): Likewise.
8399 (define_insn_reservation "Z4"): Likewise.
8400 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
8401 (pa_adjust_insn_length): Remove adjustment for btable branches.
8402 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
8405 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
8407 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8408 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
8409 Add m->r,x alternatives.
8410 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
8411 splitters using SWI48x mode iterator.
8412 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
8413 TARGET_64BIT. Add m->x alternative.
8414 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
8415 Add o->x alternative. Enable for TARGET_SSE.
8416 (sse_storeq): Remove expander.
8417 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
8418 with memory input operand.
8419 (*vec_extractv2di_1 splitter): New.
8420 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
8421 * config/i386/i386.md (ssevecmodelower): New mode attribute.
8423 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8425 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
8426 (INT_LOWPART): Delete.
8427 (extract_MB): Adjust.
8428 (extract_ME): Adjust.
8429 (print_operand): Adjust.
8431 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8433 * config/rs6000/predicates.md (reg_or_add_cint_operand,
8434 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8435 (reg_or_logical_cint_operand, easy_fp_constant,
8436 logical_const_operand): Delete "CONST_DOUBLE" case.
8437 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
8438 "HOST_BITS_PER_WIDE_INT == 64" test.
8439 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
8440 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
8441 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
8442 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
8444 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
8445 CONST_DOUBLE DImode/VOIDmode case.
8446 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
8447 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8448 CONST_DOUBLE VOIDmode case.
8449 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8450 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
8451 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
8452 Delete CONST_DOUBLE case.
8453 (splitters for mov FMOVE64 const_double): Delete
8454 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8455 "HOST_BITS_PER_WIDE_INT >= 64" test.
8456 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
8458 (mov DI const_double): Delete.
8460 2013-05-04 Jakub Jelinek <jakub@redhat.com>
8462 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
8463 on op shows all bits zero in mode of a lowpart subreg, return zero.
8465 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8468 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
8469 to save TFmode registers and DImode to save TImode registers for
8470 caller save operations.
8471 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
8472 mark being partially clobbered since they only use the first
8475 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
8476 and TDmode only use the upper 64-bits of each VSX register.
8478 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8480 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
8481 (find_candidates_in_block): Disable slsr_process_phi.
8483 2013-05-03 Guozhi Wei <carrot@google.com>
8485 * coverage.c (coverage_obj_init): Move the construction of gcov
8487 (build_init_ctor): ... here.
8489 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8491 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
8492 (slsr_cand_d): Redefine def_phi.
8493 (stride_status, phi_adjust_status, count_phis_status): New enums.
8494 (find_phi_def): New.
8495 (find_basis_for_base_expr): New.
8496 (find_basis_for_candidate): Handle hidden bases.
8497 (alloc_cand_and_find_basis): Handle phi candidates.
8498 (slsr_process_phi): New.
8499 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
8500 (create_mul_imm_cand): Likewise.
8501 (create_add_ssa_cand): Exclude phi base candidates.
8502 (create_add_imm_cand): Likewise.
8503 (slsr_process_cast): Likewise.
8504 (slsr_process_copy): Likewise.
8505 (find_candidates_in_block): Handle phi candidates.
8506 (dump_candidate): Likewise.
8507 (unconditional_cands): Delete.
8508 (unconditional_cands_with_known_stride_p): Delete.
8509 (phi_dependent_cand_p): New.
8510 (cand_increment): Handle phi-dependent candidates.
8511 (replace_dependent): Delete.
8512 (replace_mult_candidate): New.
8513 (replace_unconditional_candidate): New.
8514 (incr_vec_index): Move to avoid forward reference.
8515 (create_add_on_incoming_edge): New.
8516 (create_phi_basis): New.
8517 (replace_dependents): Delete.
8518 (replace_conditional_candidate): New.
8519 (phi_add_costs): New.
8520 (replace_uncond_cands_and_profitable_phis): New.
8521 (record_increment): Handle phi adjustments.
8522 (record_phi_increments): New.
8523 (record_increments): Handle phi adjustments.
8524 (phi_incr_cost): New.
8525 (lowest_cost_path): Handle phis.
8526 (total_savings): Likewise.
8527 (analyze_increments): Likewise.
8528 (ncd_with_phi): New.
8529 (ncd_of_cand_and_phis): New.
8530 (nearest_common_dominator_for_cands): Handle phi increments.
8531 (all_phi_incrs_profitable): New.
8532 (replace_profitable_candidates): Handle phi-dependent candidates.
8533 (analyze_candidates_and_replace): Likewise.
8535 2013-05-03 Teresa Johnson <tejohnson@google.com>
8538 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
8539 do not exceed REG_BR_PROB_BASE.
8541 2013-05-03 Jeff Law <law@redhat.com>
8543 PR tree-optimization/57144
8544 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
8545 operand of the condition will bit into the new type when eliminating
8546 a cast feeding a condition.
8548 2013-05-03 Jakub Jelinek <jakub@redhat.com>
8550 PR rtl-optimization/57130
8551 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
8552 of COMPARE as in_code to the recursive call if needed.
8554 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
8556 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
8557 (enabled): Handle new members.
8558 * config/i386/sse.md (*vec_concatv2si): Merge from
8559 *vec_concatv2si_sse2 and vec_concatv2si_sse.
8560 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
8562 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
8564 PR tree-optimization/57027
8565 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
8566 for fnms opportunity, check we got the prerequisite kind
8567 of tree / gimple before using accessor functions.
8569 2013-05-03 Richard Biener <rguenther@suse.de>
8571 * double-int.h (lshift): New overload without precision
8573 (operator *=, operator +=, operator -=): Move ...
8574 * double-int.c (operator *=, operator +=, operator -=): ... here
8575 and implement more efficiently.
8576 (mul_double_with_sign): Remove.
8577 (lshift_double): Adjust to take unsinged shift argument, push
8578 dispatching code to callers.
8579 (mul_double_wide_with_sign): Add early out for callers that
8580 are not interested in high parts or overflow.
8581 (lshift): New function.
8582 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
8584 (lrotate, rrotate): Use logical shifts.
8585 * expr.c (get_inner_reference): Use lshift.
8586 * fixed-value.c (do_fixed_divide): Likewise.
8587 * tree-dfa.c (get_ref_base_and_extent): Likewise.
8588 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
8589 (indirect_refs_may_alias_p): Likewise.
8590 (stmt_kills_ref_p_1): Likewise.
8592 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8594 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
8596 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8598 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
8599 scalar form of FABD instruction.
8601 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
8603 * lra-constraints.c (process_alt_operands): Add checking alt
8604 number to choose the best alternative.
8606 2013-05-02 Richard Biener <rguenther@suse.de>
8608 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
8609 bitmap and its handling.
8610 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
8612 2013-05-02 Richard Biener <rguenther@suse.de>
8615 * tree-inline.c (copy_loops): Properly handle removed loops.
8616 (copy_cfg_body): Mark destination loops for fixup if source
8619 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
8622 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
8623 generating simple_return for naked functions.
8625 2013-05-02 Martin Jambor <mjambor@suse.cz>
8628 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
8629 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
8631 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
8632 ipa_agg_replacement_value structures.
8633 (known_aggs_to_agg_replacement_list): Likewise.
8634 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
8635 (read_agg_replacement_chain): Likewise.
8636 (ipcp_transform_function): Also check that by_ref flags match.
8638 2013-05-02 Richard Biener <rguenther@suse.de>
8640 * graphds.h (struct graph): Add obstack member.
8641 * graphds.c (new_graph): Initialize obstack and allocate
8643 (add_edge): Allocate edge from the obstack.
8644 (free_graph): Free the obstack instead of all edges and vertices.
8646 2013-05-02 Teresa Johnson <tejohnson@google.com>
8648 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
8650 * cfg.c (update_bb_profile_for_threading): Ditto.
8651 * tree-inline.c (copy_bb): Ditto.
8652 (copy_edges_for_bb): Ditto.
8653 (initialize_cfun): Ditto.
8654 (copy_cfg_body): Ditto.
8655 (expand_call_inline): Ditto.
8656 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
8657 (estimate_node_size_and_time): Ditto.
8658 (inline_merge_summary): Ditto.
8659 * cgraphclones.c (cgraph_clone_edge): Ditto.
8660 (cgraph_clone_node): Ditto.
8661 * sched-rgn.c (compute_dom_prob_ps): Ditto.
8662 (compute_trg_info): Ditto.
8664 2013-05-02 Ian Bolton <ian.bolton@arm.com>
8666 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
8667 S reg when fp attribute set.
8668 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
8670 2013-05-02 Ian Bolton <ian.bolton@arm.com>
8672 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
8674 (*and_one_cmplsi3_compare0_uxtw): Likewise.
8675 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
8676 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
8678 2013-05-02 Richard Biener <rguenther@suse.de>
8680 * tree-scalar-evolution.c (scev_info_hasher): Remove.
8681 (struct instantiate_cache_entry): New type.
8682 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
8683 (struct instantiate_cache_type): New type.
8684 (set_instantiated_value, get_instantiated_value): Remove.
8685 (get_instantiated_value_entry): New function.
8686 (instantiate_scev_name): Use the new cache and adjust.
8687 (instantiate_scev_poly): Adjust.
8688 (instantiate_scev_binary): Likewise.
8689 (instantiate_array_ref): Likewise.
8690 (instantiate_scev_convert): Likewise.
8691 (instantiate_scev_not): Likewise.
8692 (instantiate_scev_3): Likewise.
8693 (instantiate_scev_2): Likewise.
8694 (instantiate_scev_r): Likewise.
8695 (instantiate_scev): Likewise.
8696 (resolve_mixers): Likewise.
8698 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
8701 * lra-constraints.c (best_small_class_operands_num): Remove.
8702 (process_alt_operands): Remove small_class_operands_num. Take
8703 small classes operands into losers and only if the operand is not
8704 matched. Modify debugging output.
8705 (curr_insn_transform): Remove best_small_class_operands_num.
8708 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8710 * config/aarch64/aarch64-builtins.c
8711 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
8712 * config/aarch64/aarch64-simd-builtins.def
8713 (reduc_splus_): Add new modes.
8714 (reduc_uplus_): New.
8715 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
8716 (reduc_uplus_v4sf): Likewise.
8717 (reduc_splus_v4sf): Likewise.
8718 (aarch64_addv<mode>): Likewise.
8719 (reduc_uplus_<mode>): Likewise.
8720 (reduc_splus_<mode>): Likewise.
8721 (aarch64_addvv2di): Likewise.
8722 (reduc_uplus_v2di): Likewise.
8723 (reduc_splus_v2di): Likewise.
8724 (aarch64_addvv2si): Likewise.
8725 (reduc_uplus_v2si): Likewise.
8726 (reduc_splus_v2si): Likewise.
8727 (reduc_<sur>plus_<mode>): New.
8728 (reduc_<sur>plus_v2di): Likewise.
8729 (reduc_<sur>plus_v2si): Likewise.
8730 (reduc_<sur>plus_v4sf): Likewise.
8731 (aarch64_addpv4sf): Likewise.
8732 * config/aarch64/arm_neon.h
8733 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
8734 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
8735 add UNSPEC_SADDV, UNSPEC_UADDV.
8737 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
8739 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8741 * config/aarch64/arm_neon.h
8742 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
8744 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8746 * config/aarch64/aarch64-builtins
8747 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
8749 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8751 * config/aarch64/aarch64-simd-builtins.def
8753 (reduc_smin_): Likewise.
8754 (reduc_umax_): Likewise.
8755 (reduc_umin_): Likewise.
8756 (reduc_smax_nan_): Likewise.
8757 (reduc_smin_nan_): Likewise.
8760 (smax): Update for V2SF, V4SF and V2DF modes.
8763 (smin_nan): Likewise.
8764 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
8765 (<su><maxmin><mode>3): ...This, refactor.
8766 (s<maxmin><mode>3): New.
8767 (<maxmin_uns><mode>3): Likewise.
8768 (reduc_<maxmin_uns>_<mode>): Refactor.
8769 (reduc_<maxmin_uns>_v4sf): Likewise.
8770 (reduc_<maxmin_uns>_v2si): Likewise.
8771 (aarch64_<fmaxmin><mode>: Remove.
8772 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
8774 (vmin<q>_f<32,64>): Likewise.
8775 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8777 (su): Add mappings for smax, smin, umax, umin.
8779 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8780 (FMAXMIN): Rename as...
8781 (FMAXMIN_UNS): ...This.
8783 (fmaxminv): Likewise.
8784 (fmaxmin): Likewise.
8786 (maxmin_uns_op): Likewise.
8788 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8790 * config/aarch64/arm_neon.h
8791 (vac<ge, gt><sd>_f<32, 64>): Rename to...
8792 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
8793 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
8795 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8797 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
8798 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
8800 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8802 * config/aarch64/aarch64-simd.md
8803 (vcond<mode>_internal): Handle special cases for constant masks.
8804 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
8805 (vcondu<mode><mode>): Likewise.
8806 (vcond<v_cmp_result><mode>): New.
8808 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8810 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
8811 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
8812 * config/aarch64/aarch64-simd-builtins.def
8813 (cmeq): Update to BUILTIN_VALLDI.
8818 * config/aarch64/arm_neon.h
8819 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
8820 to builtins or C as appropriate.
8822 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8824 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
8826 (cmhi): Rename to...
8828 * config/aarch64/aarch64-simd.md
8829 (simd_mode): Add SF.
8830 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
8831 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
8832 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
8833 (cstore<mode>_neg): ...This.
8834 * config/aarch64/iterators.md
8836 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
8838 (UCOMPARISONS): Likewise.
8839 (optab): Add missing comparisons.
8847 (V_cmp_result): Add DF, SF modes.
8848 (v_cmp_result): Likewise.
8851 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
8853 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
8855 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
8856 define_insn to define_insn_and_split.
8857 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
8858 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
8859 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
8860 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
8861 (thumb2_negscc): Likewise.
8863 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8865 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
8867 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8869 * config/arm/thumb2.md: Remove trailing whitespaces.
8871 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8873 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
8874 Use gen_int_mode rather than GEN_INT.
8876 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
8878 * value-prof.c (stream_in_histogram_value): Remove the strayed
8881 2013-04-30 Richard Biener <rguenther@suse.de>
8884 * cfghooks.c (split_edge): Properly check for the loop latch edge.
8886 2013-04-30 Richard Biener <rguenther@suse.de>
8889 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
8891 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
8893 PR rtl-optimization/56957
8894 PR rtl-optimization/57105
8895 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
8896 variable. Use just INSN_UID for determining whether an insn
8897 should be only disconnected from the insn stream.
8898 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
8900 2013-04-30 Jakub Jelinek <jakub@redhat.com>
8902 PR tree-optimization/57104
8903 * tsan.c (instrument_expr): Don't instrument accesses to
8904 DECL_HARD_REGISTER VAR_DECLs.
8906 2013-04-30 Richard Biener <rguenther@suse.de>
8908 * function.h (loops_for_fn): New inline function.
8909 (set_loops_for_fn): Likewise.
8910 * cfgloop.h (place_new_loop): Add struct function parameter.
8911 (get_loop): Likewise.
8912 (get_loops): Likewise.
8913 (number_of_loops): Likewise.
8915 (fel_init): Likewise.
8916 * cfg.c (get_loop_copy): Adjust.
8917 * cfgloop.c (flow_loops_dump): Likewise.
8918 (record_loop_exits): Likewise.
8919 (verify_loop_structure): Likewise.
8920 * cfgloopanal.c (mark_irreducible_loops): Likewise.
8921 (estimate_reg_pressure_cost): Likewise.
8922 (mark_loop_exit_edges): Likewise.
8923 * cfgloopmanip.c (place_new_loop): Likewise.
8924 (add_loop): Likewise.
8925 (duplicate_loop): Likewise.
8926 * graph.c (draw_cfg_nodes): Likewise.
8927 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
8928 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
8929 (extract_affine_chrec): Likewise.
8930 (build_scop_iteration_domain): Likewise.
8931 * graphite.c (graphite_initialize): Likewise.
8932 * ira-build.c (create_loop_tree_nodes): Likewise.
8933 (more_one_region_p): Likewise.
8934 (rebuild_regno_allocno_maps): Likewise.
8935 (mark_loops_for_removal): Likewise.
8936 (mark_all_loops_for_removal): Likewise.
8937 (remove_unnecessary_regions): Likewise.
8938 (ira_build): Likewise.
8939 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
8940 * loop-init.c (fix_loop_structure): Likewise.
8941 (gate_rtl_move_loop_invariants): Likewise.
8942 (gate_rtl_unswitch): Likewise.
8943 (gate_rtl_unroll_and_peel_loops): Likewise.
8944 (rtl_doloop): Likewise.
8945 * lto-streamer-in.c (input_cfg): Likewise.
8946 * lto-streamer-out.c (output_cfg): Likewise.
8947 * modulo-sched.c (sms_schedule): Likewise.
8948 * predict.c (tree_estimate_probability): Likewise.
8949 (tree_estimate_probability_driver): Likewise.
8950 (estimate_loops): Likewise.
8951 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
8952 (move_sese_region_to_fn): Likewise.
8953 (debug_loop_num): Likewise.
8954 * tree-chrec.c (chrec_evaluate): Likewise.
8955 (hide_evolution_in_other_loops_than_loop): Likewise.
8956 (chrec_component_in_loop_num): Likewise.
8957 (reset_evolution_in_loop): Likewise.
8958 (evolution_function_is_invariant_rec_p): Likewise.
8959 * tree-if-conv.c (main_tree_if_conversion): Likewise.
8960 * tree-inline.c (copy_loops): Likewise.
8961 (copy_cfg_body): Likewise.
8962 (tree_function_versioning): Likewise.
8963 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
8964 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
8966 (add_to_evolution_1): Likewise.
8967 (scev_const_prop): Likewise.
8968 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
8969 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
8970 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
8971 (tree_ssa_lim_initialize): Likewise.
8972 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
8973 (verify_loop_closed_ssa): Likewise.
8974 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
8975 (tree_ssa_loop_im): Likewise.
8976 (tree_ssa_loop_unswitch): Likewise.
8977 (tree_vectorize): Likewise.
8978 (check_data_deps): Likewise.
8979 (tree_ssa_loop_ivcanon): Likewise.
8980 (tree_ssa_loop_bounds): Likewise.
8981 (tree_complete_unroll): Likewise.
8982 (tree_complete_unroll_inner): Likewise.
8983 (tree_parallelize_loops): Likewise.
8984 (tree_ssa_loop_prefetch): Likewise.
8985 (tree_ssa_loop_ivopts): Likewise.
8986 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
8987 * tree-vectorizer.c (vectorize_loops): Likewise.
8989 2013-04-29 Mike Frysinger <vapier@gentoo.org>
8991 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
8992 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
8993 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
8994 with EABI_LINK_SPEC.
8996 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
8999 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
9002 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
9005 * lra-constraints.c (process_alt_operands): Discourage a bit more
9006 using memory for pseudos. Print cost dump for alternatives.
9007 Modify cost values for conflicts with early clobbers.
9008 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
9010 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9013 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
9015 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9017 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
9019 (movdi_aarch64): Support LDR/STR from/to D register.
9021 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9023 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
9024 or higher by default.
9026 2013-04-29 Richard Biener <rguenther@suse.de>
9029 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9030 even if not adding abnormal edges for calls that can make
9033 2013-04-29 Richard Biener <rguenther@suse.de>
9036 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9038 (move_stmt_r): Remove redundant checking.
9040 2013-04-29 Teresa Johnson <tejohnson@google.com>
9043 * basic-block.h (apply_scale): New function.
9044 (apply_probability): Use apply_scale.
9045 * gimple-streamer-in.c (input_bb): Ditto.
9046 * lto-streamer-in.c (input_cfg): Ditto.
9047 * lto-cgraph.c (merge_profile_summaries): Ditto.
9048 * tree-optimize.c (execute_fixup_cfg): Ditto.
9049 * tree-inline.c (copy_bb): Update comment to use apply_scale.
9050 (copy_edges_for_bb): Ditto.
9051 (copy_cfg_body): Ditto.
9053 2013-04-29 Tom de Vries <tom@codesourcery.com>
9055 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9056 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9057 (tail_merge_optimize): Handle current_loops == NULL.
9059 2013-04-26 Jeff Law <law@redhat.com>
9061 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9062 (simplify_cond_using_ranges): Generalize code to simplify
9063 COND_EXPRs where one argument is a constant and the other
9064 is an SSA_NAME created by an integral type conversion.
9066 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9068 * config/arm/arm.md (store_minmaxsi): Use only when
9069 optimize_insn_for_size_p.
9071 2013-04-29 Christian Bruel <christian.bruel@st.com>
9074 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9076 2013-04-29 Richard Biener <rguenther@suse.de>
9079 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9080 loop tree make sure to schedule a fixup for the child as well.
9081 (expand_omp_for_generic): Properly add loops.
9082 (expand_omp_for_static_nochunk): Likewise.
9083 (expand_omp_for_static_chunk): Likewise.
9084 (expand_omp_for): For the degenerate case fixup loops.
9085 (expand_omp_sections): Fix default bb placement in loops.
9086 (expand_omp_atomic_pipeline): Properly add loops.
9088 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9090 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9092 2013-04-29 Tom de Vries <tom@codesourcery.com>
9094 * tree-ssa-tail-merge.c: Update header comment.
9096 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9098 * config/aarch64/arm_neon.h
9099 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9100 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9101 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9102 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9103 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9104 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9105 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9106 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9108 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9110 * config/aarch64/aarch64-simd.md
9111 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9112 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9113 fix_trunc, fixuns_trunc.
9114 (ftrunc<VDQF:mode>2): New.
9115 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9116 (fix_trunc_optab): New.
9118 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9120 * config/aarch64/aarch64-builtins.c
9121 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9122 iceilf, lround, iroundf.
9124 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9127 * config/i386/i386.h (enum ix86_tune_indices)
9128 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9129 New, split from X86_TUNE_INTER_UNIT_MOVES.
9130 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9131 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9132 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9133 (TARGET_INTER_UNIT_MOVES): Remove.
9134 * config/i386/i386.c (initial_ix86_tune_features): Update.
9135 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9136 (ix86_expand_convert_uns_didf_sse): Use
9137 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9138 (ix86_expand_vector_init_one_nonzero): Ditto.
9139 (ix86_expand_vector_init_interleave): Ditto.
9140 (inline_secondary_memory_needed): Return true for moves from SSE class
9141 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9142 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9143 * config/i386/constraints.md (Yi, Ym): Depend on
9144 TARGET_INTER_UNIT_MOVES_TO_VEC.
9145 (Yj, Yn): New constraints.
9146 * config/i386/i386.md (*movdi_internal): Change constraints of
9147 operand 1 from Yi to Yj and from Ym to Yn.
9148 (*movsi_internal): Ditto.
9149 (*movdf_internal): Ditto.
9150 (*movsf_internal): Ditto.
9151 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9152 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9153 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9154 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9155 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9156 * config/i386/sse.md (movdi_to_sse): Ditto.
9157 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9158 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9159 TARGET_INTER_UNIT_MOVES.
9160 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9161 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9162 instead of TARGET_INTER_UNIT_MOVES.
9163 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9164 operand 1 from Yi to Yj and from Ym to Yn.
9166 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9168 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9169 (float_truncate_hi_): Likewise.
9170 (float_extend_lo_): Likewise.
9171 (float_truncate_lo_): Likewise.
9172 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9173 (aarch64_float_extend_lo_v2df): Likewise.
9174 (vec_unpacks_hi_v4sf): Likewise.
9175 (aarch64_float_truncate_lo_v2sf): Likewise.
9176 (aarch64_float_truncate_hi_v4sf): Likewise.
9177 (vec_pack_trunc_v2df): Likewise.
9178 (vec_pack_trunc_df): Likewise.
9180 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9182 * config/aarch64/aarch64-builtins.c
9183 (aarch64_fold_builtin): Fold float conversions.
9184 * config/aarch64/aarch64-simd-builtins.def
9185 (floatv2si, floatv4si, floatv2di): New.
9186 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9187 * config/aarch64/aarch64-simd.md
9188 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9189 * config/aarch64/iterators.md (FLOATUORS): New.
9190 (optab): Add float, floatuns.
9191 (su_optab): Likewise.
9193 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9195 * config/aarch64/aarch64-builtins.c
9196 (aarch64_builtin_vectorized_function): Use new names for
9198 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9199 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9200 (fcvtzu): Split as...
9201 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9202 (fcvtas): Split as...
9203 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9204 (fcvtau): Split as...
9205 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9206 (fcvtps): Split as...
9207 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9208 (fcvtpu): Split as...
9209 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9210 (fcvtms): Split as...
9211 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9212 (fcvtmu): Split as...
9213 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9214 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9215 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9216 (lfrintnusf, lfrintnudf): Likewise.
9217 * config/aarch64/aarch64-simd.md
9218 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9220 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9221 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9222 (fcvt_pattern): Likewise.
9224 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9226 * config/aarch64/aarch64-simd.md
9227 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9228 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9230 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9232 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9233 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9234 (vrnd<a,m,n,p>_f32): Implement using builtins.
9235 (vrnd<i,x><q>_f<32, 64>): New.
9237 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9239 * config/aarch64/aarch64-builtins.c
9240 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9241 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9242 (frintz): Rename to...
9244 (frintp): Rename to...
9246 (frintm): Rename to...
9248 (frinti): Rename to...
9249 (nearbyint): ...this.
9250 (frintx): Rename to...
9252 (frinta): Rename to...
9254 * config/aarch64/aarch64-simd.md
9255 (aarch64_frint<frint_suffix><mode>): Delete.
9256 (<frint_pattern><mode>2): Convert to insn.
9257 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9258 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9259 (frint_pattern): Likewise.
9260 (frint_suffix): Likewise.
9262 2013-04-29 Richard Biener <rguenther@suse.de>
9264 PR tree-optimization/57081
9265 * loop-init.c: Include tree-flow.h.
9266 (loop_optimizer_finalize): Free number of iteration estimates.
9267 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9269 2013-04-29 Jakub Jelinek <jakub@redhat.com>
9271 PR tree-optimization/57083
9272 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9273 non-singleton shift count range, zero extend low_bound for uns case.
9275 * config/i386/predicates.md (general_vector_operand): New predicate.
9276 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9277 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9278 if they aren't nonimmediate operands. If their original values
9279 satisfy const_vector_equal_evenodd_p, don't shift them.
9280 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9281 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9282 if not nonimmediate_operand.
9283 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9284 instead of register_operand.
9285 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9287 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
9289 * stor-layout.c (finalize_size_functions): Allocate a structure and
9290 reset cfun before dumping the functions.
9292 2013-04-27 Jakub Jelinek <jakub@redhat.com>
9294 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9297 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9298 use xop_pmacsdqh if uns_p.
9299 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9300 the immediate rotate count.
9302 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9304 * rtl.h (struct rtx_def): Add comment for field jump.
9305 (LRA_SUBREG_P): New macro.
9306 * recog.c (register_operand): Check LRA_SUBREG_P.
9307 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9309 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9310 (lra_final_code_change): Skip subreg change for operators.
9311 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9312 if there are no operand changes.
9313 * lra-constraints.c (curr_insn_set): New.
9314 (match_reload): Set LRA_SUBREG_P.
9315 (emit_spill_move): Ditto.
9316 (check_and_process_move): Use curr_insn_set. Process only single
9317 set insns. Don't initialize sec_mem_p and change_p.
9318 (simplify_operand_subreg): Use LRA_SUBREG_P.
9319 (reg_in_class_p): New function.
9320 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9321 of #ifdef. Add code to remove cycling.
9322 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9323 non-null disp. Reload inner instead of disp when base and index
9324 are null. Try to put lo_sum into register.
9325 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9326 (check_and_process_move): Move code for move cost check to
9327 simple_move_p. Remove equiv_substitution.
9328 (simple_move_p): New function.
9329 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9330 curr_insn_set. Call check_and_process_move only for single set
9331 insns. Use the new function. Move call of check_and_process_move
9332 after operand equiv substitution and address process.
9334 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9337 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9338 with nonlocal goto receivers or returns twice calls, ignore
9339 unininitialized values from abnormal edges to nl goto receiver
9340 or returns twice call.
9342 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9344 PR tree-optimization/57051
9345 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9346 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9349 2013-04-26 Richard Biener <rguenther@suse.de>
9351 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9352 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9353 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9354 (fixup_loop_arrays_after_move): New function.
9355 (move_sese_region_to_fn): Properly outline the loop tree parts
9358 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
9360 * config/i386/i386.md (type, unit): Fix long lines.
9362 2013-04-26 Richard Biener <rguenther@suse.de>
9364 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9365 (lto-streamer-out.o): Likewise.
9366 * cfgloop.c (init_loops_structure): Export, add struct function
9367 argument and adjust.
9368 (flow_loops_find): Adjust.
9369 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9370 (init_loops_structure): Declare.
9371 * lto-streamer-in.c: Include cfgloop.h.
9372 (input_cfg): Input the loop tree.
9373 * lto-streamer-out.c: Include cfgloop.h.
9374 (output_cfg): Output the loop tree.
9375 (output_struct_function_base): Do not drop PROP_loops.
9377 2013-03-26 Richard Biener <rguenther@suse.de>
9379 * tree-cfg.c (execute_build_cfg): Build the loop tree.
9380 (pass_build_cfg): Provide PROP_loops.
9381 (move_sese_region_to_fn): Remove loops that are outlined into fn
9383 * tree-inline.c: Include cfgloop.h.
9384 (initialize_cfun): Do not drop PROP_loops.
9385 (copy_loops): New function.
9386 (copy_cfg_body): Copy loop structure.
9387 (tree_function_versioning): Initialize destination loop tree.
9388 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
9389 (pass_parallelize_loops): Do IL verification.
9390 * loop-init.c (loop_optimizer_init): Fixup loops if required.
9391 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
9392 the CFG make sure we fixup loops as well.
9393 * tree-ssa-tail-merge.c: Include cfgloop.h.
9394 (replace_block_by): When merging loop latches mark loops for fixup.
9395 * lto-streamer-out.c (output_struct_function_base): Drop
9397 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
9398 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
9399 * ipa-split.c: Include cfgloop.h.
9400 (split_function): Add the new return block to the loop tree root.
9401 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
9402 whether we have removed the forwarder block.
9403 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
9404 * cfgloop.h (place_new_loop): Declare.
9405 * cfgloopmanip.c (place_new_loop): Export.
9406 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
9407 (tree-switch-conversion.o): Likewise.
9408 (tree-complex.o): Likewise.
9409 (tree-inline.o): Likewise.
9410 (tree-ssa-tailmerge.o): Likewise.
9411 (ipa-split.o): Likewise.
9412 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
9413 (tree-ssa-copy.o): Likewise.
9414 * tree-switch-conversion.c: Include cfgloop.h
9415 (process_switch): If we emit a bit-test cascade, schedule loops
9417 * tree-complex.c: Include cfgloop.h.
9418 (expand_complex_div_wide): Properly add new basic-blocks to loops.
9419 * asan.c: Include cfgloop.h.
9420 (create_cond_insert_point): Properly add new basic-blocks to
9421 loops, schedule loop fixup.
9422 * cfgloop.c (verify_loop_structure): Check that looks are not
9424 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
9426 (expand_omp_for_generic): Likewise.
9427 (expand_omp_sections): Likewise.
9428 (expand_omp_atomic_pipeline): Schedule loops for fixup.
9429 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
9430 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
9431 is initialized, not when loops are present.
9432 * tree-parloops.c (parallelize_loops): Remove checking here.
9433 * passes.c (init_optimization_passes): Schedule a copy-propagation
9434 pass before complete unrolling of inner loops.
9436 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9438 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
9439 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
9440 (colorize_init): Add argument to _WIN32 version.
9441 * toplev.c: Include diagnostic-color.h.
9442 (process_options): Default to -fdiagnostics-color=auto if
9443 GCC_COLORS env var is in the environment.
9444 * common.opt (fdiagnostics-color=): Add Var and Init.
9445 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
9446 env var is in the environment, the default is auto rather than never.
9448 * diagnostic.h (file_name_as_prefix): Add context argument.
9449 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
9450 the string as locus.
9451 * langhooks.c (lhd_print_error_function): Adjust caller.
9453 2013-04-25 Lawrence Crowl <crowl@google.com>
9455 * var-tracking.c (shared_hash_def::htab):
9456 Change type to hash_table. Update dependent calls and types.
9458 2013-04-25 Lawrence Crowl <crowl@google.com>
9460 * Makefile.in: Update as needed below.
9462 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
9463 Move declaration to after the type's method definitons.
9465 * attribs.c (htab_t scoped_attributes::attribute_hash):
9466 Change type to hash_table. Update dependent calls and types.
9468 * bitmap.c (htab_t bitmap_desc_hash):
9469 Change type to hash_table. Update dependent calls and types.
9471 * cselib.c (htab_t cselib_hash_table):
9472 Change type to hash_table. Update dependent calls and types.
9474 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
9475 (hash_string_slot_node): Move implementation into lto-streamer.h
9476 struct string_slot_hasher.
9477 (eq_string_slot_node): Likewise.
9479 * data-streamer-out.c: Update output_block::string_hash_table
9480 dependent calls and types.
9482 * dwarf2cfi.c (htab_t trace_index):
9483 Change type to hash_table. Update dependent calls and types.
9485 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
9486 Change type to hash_table. Update dependent calls and types.
9487 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
9488 (htab_t optimize_external_refs::map): Likewise.
9489 (htab_t output_comp_unit::extern_map): Likewise.
9490 (htab_t output_comdat_type_unit::extern_map): Likewise.
9491 (htab_t output_macinfo::macinfo_htab): Likewise.
9492 (htab_t optimize_location_lists::htab): Likewise.
9493 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
9495 * except.c (htab_t ehspec_hash_type):
9496 Change type to hash_table. Update dependent calls and types.
9497 (assign_filter_values::ttypes): Likewise.
9498 (assign_filter_values::ehspec): Likewise.
9499 (sjlj_assign_call_site_values::ar_hash): Likewise.
9500 (convert_to_eh_region_ranges::ar_hash): Likewise.
9502 * gcse.c (htab_t pre_ldst_table):
9503 Change type to hash_table. Update dependent calls and types.
9505 * ggc-common.c (htab_t saving_htab):
9506 Change type to hash_table. Update dependent calls and types.
9507 (htab_t loc_hash): Likewise.
9508 (htab_t ptr_hash): Likewise.
9509 (call_count): Rename ggc_call_count.
9510 (call_alloc): Rename ggc_call_alloc.
9511 (loc_descriptor): Rename make_loc_descriptor.
9512 (add_statistics): Rename ggc_add_statistics.
9514 * ggc-common.c (saving_htab):
9515 Change type to hash_table. Update dependent calls and types.
9517 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
9518 (push_gimplify_context): Likewise.
9519 (pop_gimplify_context): Likewise.
9520 (struct gimple_temp_hash_elt): Added.
9521 (struct gimplify_hasher): Likewise.
9522 (struct gimplify_ctx.temp_htab):
9523 Change type to hash_table. Update dependent calls and types.
9525 * gimple-fold.c: Include gimplify-ctx.h.
9527 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
9528 Change type to hash_table. Update dependent calls and types.
9529 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
9530 avoid potential global name collision.
9532 * gimplify.c: Include gimplify-ctx.h.
9533 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
9534 (htab_t gimplify_ctx::temp_htab):
9535 Update dependent calls and types for new type hash_table.
9536 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
9537 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
9539 * gimplify-ctx.h: New.
9540 (struct gimple_temp_hash_elt): Move from gimplify.c.
9541 (class gimplify_hasher): New.
9542 (struct gimplify_ctx): Move from gimple.h.
9543 (htab_t gimplify_ctx::temp_htab):
9544 Change type to hash_table. Update dependent calls and types.
9546 * graphite-clast-to-gimple.c: Include graphite-htab.h.
9547 (htab_t ivs_params::newivs_index):
9548 Change type to hash_table. Update dependent calls and types.
9549 (htab_t ivs_params::params_index): Likewise.
9550 (htab_t print_generated_program::params_index): Likewise.
9551 (htab_t gloog::newivs_index): Likewise.
9552 (htab_t gloog::params_index): Likewise.
9554 * graphite.c: Include graphite-htab.h.
9555 4htab_t graphite_transform_loops::bb_pbb_mapping):
9556 Change type to hash_table. Update dependent calls and types.
9558 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
9559 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
9560 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
9562 * graphite-dependences.c: Include graphite-htab.h.
9563 (loop_is_parallel_p): Change hash table type of parameter.
9565 * graphite-htab.h: New.
9566 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
9567 (extern find_pbb_via_hash): Move from graphite-poly.h.
9568 (extern loop_is_parallel_p): Move from graphite-poly.h.
9569 (extern get_loop_body_pbbs): Move from graphite-poly.h.
9571 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
9572 (extern loop_is_parallel_p): Move to graphite-htab.h.
9573 (extern get_loop_body_pbbs): Move to graphite-htab.h.
9575 * haifa-sched.c (htab_t delay_htab):
9576 Change type to hash_table. Update dependent calls and types.
9577 (htab_t delay_htab_i2): Likewise.
9579 * ira-color.c (htab_t allocno_hard_regs_htab):
9580 Change type to hash_table. Update dependent calls and types.
9582 * ira-costs.c (htab_t cost_classes_htab):
9583 Change type to hash_table. Update dependent calls and types.
9585 * loop-invariant.c (htab_t merge_identical_invariants::eq):
9586 Change type to hash_table. Update dependent calls and types.
9588 * loop-iv.c (htab_t bivs):
9589 Change type to hash_table. Update dependent calls and types.
9591 * loop-unroll.c (htab_t opt_info::insns_to_split):
9592 Change type to hash_table. Update dependent calls and types.
9593 (htab_t opt_info::insns_with_var_to_expand): Likewise.
9595 * lto-streamer.h (struct string_slot): Move from data-streamer.h
9596 (struct string_slot_hasher): New.
9597 (htab_t output_block::string_hash_table):
9598 Change type to hash_table. Update dependent calls and types.
9600 * lto-streamer-in.c (freeing_string_slot_hasher): New.
9601 (htab_t file_name_hash_table):
9602 Change type to hash_table. Update dependent calls and types.
9604 * lto-streamer-out.c: Update output_block::string_hash_table dependent
9607 * lto-streamer.c (htab_t tree_htab):
9608 Change type to hash_table. Update dependent calls and types.
9610 * omp-low.c: Include gimplify-ctx.h.
9612 * passes.c (htab_t name_to_pass_map):
9613 Change type to hash_table. Update dependent calls and types.
9614 (pass_traverse): Rename to passes_pass_traverse.
9616 * plugin.c (htab_t event_tab):
9617 Change type to hash_table. Update dependent calls and types.
9619 * postreload-gcse.c (htab_t expr_table):
9620 Change type to hash_table. Update dependent calls and types.
9621 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
9623 * sese.c (debug_rename_map_1): Make extern.
9624 (htab_t copy_bb_and_scalar_dependences::rename_map):
9625 Change type to hash_table. Update dependent calls and types.
9627 * sese.h (extern debug_rename_map): Move to .c file.
9629 * store-motion.c (htab_t store_motion_mems_table):
9630 Change type to hash_table. Update dependent calls and types.
9632 * trans-mem.c (htab_t tm_new_mem_hash):
9633 Change type to hash_table. Update dependent calls and types.
9635 * tree-browser.c (htab_t TB_up_ht):
9636 Change type to hash_table. Update dependent calls and types.
9638 * tree-cfg.c (htab_t discriminator_per_locus):
9639 Change type to hash_table. Update dependent calls and types.
9641 * tree-complex.c: Include tree-hasher.h
9642 (htab_t complex_variable_components):
9643 Change type to hash_table. Update dependent calls and types.
9645 * tree-eh.c (htab_t finally_tree):
9646 Change type to hash_table. Update dependent calls and types.
9648 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
9649 struct int_tree_hasher.
9650 (extern int_tree_map_eq): Likewise.
9651 (uid_decl_map_hash): Removed.
9652 (extern decl_tree_map_eq): Likewise.
9654 * tree-hasher.h: New.
9655 (struct int_tree_hasher): New.
9656 (typedef int_tree_htab_type): New.
9658 * tree-inline.c: Include gimplify-ctx.h.
9660 * tree-mudflap.c: Include gimplify-ctx.h.
9662 * tree-parloops.c: Include tree-hasher.h.
9663 (htab_t eliminate_local_variables_stmt::decl_address):
9664 Change type to hash_table. Update dependent calls and types.
9665 (htab_t separate_decls_in_region::decl_copies): Likewise.
9667 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
9668 Change type to hash_table. Update dependent calls and types.
9670 * tree-sra.c (candidates):
9671 Change type to hash_table. Update dependent calls and types.
9673 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
9675 (int_tree_map_hash): Likewise.
9677 * tree-ssa-dom.c (htab_t avail_exprs):
9678 Change type to hash_table. Update dependent calls and types.
9680 * tree-ssa-live.c (var_map_base_init::tree_to_index):
9681 Change type to hash_table. Update dependent calls and types.
9683 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
9684 Change type to hash_table. Update dependent calls and types.
9686 * tree-ssa-phiopt.c (seen_ssa_names):
9687 Change type to hash_table. Update dependent calls and types.
9689 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
9690 Change type to hash_table. Update dependent calls and types.
9692 * tree-ssa-uncprop.c (equiv):
9693 Change type to hash_table. Update dependent calls and types.
9695 2013-04-25 Jakub Jelinek <jakub@redhat.com>
9697 PR rtl-optimization/57003
9698 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
9699 call note_stores with kill_clobbered_value callback again after
9700 killing regs_invalidated_by_call.
9702 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9704 * config/aarch64/aarch64-simd.md
9705 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
9706 (aarch64_simd_bsl<mode>): Likewise.
9707 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
9709 2013-04-25 Marek Polacek <polacek@redhat.com>
9711 PR tree-optimization/57066
9712 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
9714 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
9716 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
9718 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9720 * config/aarch64/aarch64-builtins.c
9721 (aarch64_fold_builtin): New.
9722 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
9723 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
9724 * config/aarch64/aarch64-simd-builtins.def (abs): New.
9725 * config/aarch64/arm_neon.h
9726 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
9728 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9729 Tejas Belagod <tejas.belagod@arm.com>
9731 * config/aarch64/aarch64-builtins.c
9732 (aarch64_gimple_fold_builtin): New.
9733 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
9734 * config/aarch64/aarch64-simd-builtins.def (addv): New.
9735 * config/aarch64/aarch64-simd.md (addpv4sf): New.
9737 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
9739 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9741 * config/aarch64/aarch64.md
9742 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
9744 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9746 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
9747 (*ngcsi_uxtw): New pattern.
9749 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9750 Julian Brown <julian@codesourcery.com>
9752 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
9753 (TB_DREG): Add T_V4HF.
9754 (v4hf_UP): New macro.
9755 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
9756 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
9757 Handle initialisation of V4HF. Adjust initialisation of reinterpret
9759 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
9760 (arm_vector_mode_supported_p): Handle V4HF.
9761 (arm_mangle_map): Handle V4HFmode.
9762 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
9763 * config/arm/arm_neon_builtins.def: Add entries for
9764 vcvtv4hfv4sf, vcvtv4sfv4hf.
9765 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
9766 (neon_vcvtv4hfv4sf): Likewise.
9767 * config/arm/neon-gen.ml: Handle half-precision floating point
9769 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
9770 * config/arm/arm_neon.h: Regenerate.
9771 * config/arm/neon.ml (type elts): Add F16.
9772 (type vectype): Add T_float16x4, T_floatHF.
9773 (type vecmode): Add V4HF.
9774 (type features): Add Requires_FP_bit feature.
9775 (elt_width): Handle F16.
9776 (elt_class): Likewise.
9777 (elt_of_class_width): Likewise.
9778 (mode_of_elt): Refactor.
9779 (type_for_elt): Handle F16, fix error messages.
9780 (vectype_size): Handle T_float16x4.
9781 (vcvt_sh): New function.
9782 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
9783 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
9784 (string_of_mode): Handle V4HF.
9785 * doc/arm-neon-intrinsics.texi: Regenerate.
9787 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9789 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
9790 format specifier in 'X' case.
9792 2013-04-25 Alan Modra <amodra@gmail.com>
9795 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
9796 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
9797 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
9798 Repeat for many other rotate/shift and mask patterns using subregs.
9799 Name lshiftrt insns.
9800 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
9801 on WORDS_BIG_ENDIAN.
9803 2013-04-25 Alan Modra <amodra@gmail.com>
9805 * config.gcc: Support little-endian powerpc-linux targets.
9806 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
9807 (LINK_OS_LINUX_SPEC): Define.
9808 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
9809 Preserve MASK_LITTLE_ENDIAN.
9810 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
9811 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
9812 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
9813 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
9814 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
9815 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
9816 Correct fp word order for little-endian. Don't shift toc entries
9817 smaller than a word for little-endian.
9818 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
9819 (bswapdi2 splits): Correct low-part subreg for little-endian.
9820 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
9821 low/high where such is correct only for be.
9822 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
9823 little-endian for -mcall-aixdesc.
9825 2013-04-25 Alan Modra <amodra@gmail.com>
9827 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9828 replace_equiv_address_nv.
9830 2013-04-25 Alan Modra <amodra@gmail.com>
9832 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
9834 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9837 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9838 * rtl.h (struct rtx_def): ...
9840 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9842 PR rtl-optimizations/57046
9843 * lra-constraints (split_reg): Set up lra_risky_transformations_p
9844 for multi-reg splits.
9846 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
9848 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
9850 2013-04-24 Sterling Augustine <saugustine@google.com>
9852 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
9853 (comp_dir_string, debug_str_dwo_section): New.
9854 (DEBUG_STR_DWO_SECTION): Rename to ...
9855 (DEBUG_DWO_STR_SECTION): ... this.
9856 (DEBUG_NORM_STR_SECTION): Delete.
9857 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
9858 (DEBUG_STR_DWO_SECTION_FLAGS): New.
9859 (find_AT_string): Move most logic to ...
9860 (find_AT_string_in_table): ... here. New.
9861 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
9862 add_skeleton_AT_string. Delete logic.
9863 (output_skeleton_debug_sections): Remove call to
9864 add_top_level_skeleton_die_attrs.
9865 (add_comp_dir_attribute): Move logic to comp_dir_string.
9866 (dwarf2out_init): Initialize debug_str_dwo_section.
9867 (output_indirect_string): Call find_string_form.
9868 (output_indirect_strings): Rewrite.
9869 (prune_unused_types): Empty skeleton_debug_str_hash.
9870 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
9871 (dwarf2out_finish): Call output_indirect_strings.
9873 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9875 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
9877 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9879 * rtl.h (struct rtx_def): Add comment for field jump.
9880 (LRA_SUBREG_P): New macro.
9881 * recog.c (register_operand): Check LRA_SUBREG_P.
9882 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9884 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9885 (lra_final_code_change): Skip subreg change for operators.
9886 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9887 if there are no operand changes.
9888 * lra-constraints.c (curr_insn_set): New.
9889 (match_reload): Set LRA_SUBREG_P.
9890 (emit_spill_move): Ditto.
9891 (check_and_process_move): Use curr_insn_set. Process only single
9892 set insns. Don't initialize sec_mem_p and change_p.
9893 (simplify_operand_subreg): Use LRA_SUBREG_P.
9894 (reg_in_class_p): New function.
9895 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9896 of #ifdef. Add code to remove cycling.
9897 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9898 non-null disp. Reload inner instead of disp when base and index
9899 are null. Try to put lo_sum into register.
9900 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9901 (check_and_process_move): Move code for move cost check to
9902 simple_move_p. Remove equiv_substitution.
9903 (simple_move_p): New function.
9904 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9905 curr_insn_set. Call check_and_process_move only for single set
9906 insns. Use the new function. Move call of check_and_process_move
9907 after operand equiv substitution and address process.
9909 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
9911 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
9912 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
9913 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
9915 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9917 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
9919 2013-04-24 Marek Polacek <polacek@redhat.com>
9921 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
9922 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
9923 (select_loops_exit_conditions): Likewise.
9924 (number_of_iterations_for_all_loops): Likewise.
9925 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
9926 (scev_analysis): Likewise.
9928 2013-04-02 Catherine Moore <clm@codesourcery.com>
9929 Chao-ying Fu <fu@mips.com>
9931 * config/mips/micromips.md (jraddiusp): New pattern.
9932 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
9933 instruction if possible.
9935 2013-04-24 Alan Modra <amodra@gmail.com>
9937 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
9939 2013-04-24 Julian Brown <julian@codesourcery.com>
9940 Chung-Lin Tang <cltang@codesourcery.com>
9942 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
9943 dependency behavior in enumeration type DIE generation. Add TODO note
9944 to comments about future DW_FORM_sdata/udata re-work of related code.
9946 2013-04-23 Lawrence Crowl <crowl@google.com>
9948 * Makefile.in: Update as needed below.
9950 * hash-table.h (class hash_table):
9951 Correct many methods with parameter types compare_type to the correct
9952 value_type. (Correct code was unlikely to notice the change.)
9953 (hash_table::elements_with_deleted) New.
9954 (class hashtable::iterator): New.
9955 (hashtable::begin()): New.
9956 (hashtable::end()): New.
9957 (FOR_EACH_HASH_TABLE_ELEMENT): New.
9959 * statistics.c (statistics_hashes):
9960 Change type to hash_table. Update dependent calls and types.
9962 * tree-into-ssa.c (var_infos):
9963 Change type to hash_table. Update dependent calls and types.
9965 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
9966 Change type to hash_table. Update dependent calls and types.
9968 * tree-ssa-loop-im.c (struct mem_ref.refs):
9969 Change type to hash_table. Update dependent calls and types.
9971 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
9972 Change type to hash_table. Update dependent calls and types.
9974 * tree-ssa-sccvn.c (vn_tables_s::nary):
9975 Change type to hash_table. Update dependent calls and types.
9976 (vn_tables_s::phis): Likewise.
9977 (vn_tables_s::references): Likewise.
9979 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
9980 (vn_reference_eq): Update parameter and return types.
9982 * tree-ssa-structalias.c (pointer_equiv_class_table):
9983 Change type to hash_table. Update dependent calls and types.
9984 (location_equiv_class_table): Likewise.
9986 * tree-vect-data-refs.c: Consequential changes for making
9987 peeling a hash_table.
9989 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
9990 (destroy_loop_vec_info): Dependent hash_table update.
9992 * tree-vectorizer.h (peeling_htab):
9993 Change type to hash_table. Update dependent calls and types.
9995 2013-04-23 Shiva Chen <shiva0217@gmail.com>
9997 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
9998 to check the register content is equal or not.
9999 * lra-constraints.c (match_reload): Use lra_assign_reg_val
10000 to assign register content record.
10001 * lra-eliminations.c (update_reg_eliminate): Use
10002 lra_update_reg_val_offset to update register content offset.
10003 * lra-int.h (struct lra_reg): Add offset member.
10004 (lra_reg_val_equal_p): New static inline function.
10005 (lra_update_reg_val_offset): New static inline function.
10006 (lra_assign_reg_val): New static inline function.
10007 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
10008 to assign register content record.
10009 (initialize_lra_reg_info_element): Initial offset to zero.
10011 2013-04-23 Catherine Moore <clm@codesourcery.com>
10013 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
10014 operands. Record compression.
10016 2013-04-23 Xinliang David Li <davidxl@google.com>
10018 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
10020 2013-04-23 Richard Biener <rguenther@suse.de>
10022 PR middle-end/57036
10023 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
10024 parameter, only add abnormal goto edges from the copied body
10025 if the call could perform abnormal gotos.
10026 (copy_cfg_body): Adjust.
10028 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10030 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10032 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10034 * coretypes.h (gimple_stmt_iterator): Add struct to make
10037 2013-04-23 Richard Biener <rguenther@suse.de>
10039 PR tree-optimization/57026
10040 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10041 from SSA names occuring in abnormal PHI nodes.
10043 2013-04-22 Andi Kleen <ak@linux.intel.com>
10045 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10047 2013-04-22 Andi Kleen <ak@linux.intel.com>
10049 * lto/lto.c (print_lto_report_1): Declare early.
10050 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10052 2013-04-22 Andi Kleen <ak@linux.intel.com>
10054 * common.opt (-flto-report-wpa): Add.
10055 * doc/invoke.texi (-flto-report-wpa): Add.
10056 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10059 2013-04-22 Xinliang David Li <davidxl@google.com>
10061 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10062 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10063 * Makefile.in: New dependency
10065 David Daney <ddaney.cavm@gmail.com>
10067 * configure.ac (gcc_cv_as_micromips_support): Use the
10068 --fatal-warnings option.
10069 * configure: Regenerate.
10071 2013-04-22 Marek Polacek <polacek@redhat.com>
10074 * tsan.c (instrument_expr): Don't instrument expression
10075 in case its size is zero.
10077 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
10081 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10083 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10085 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10087 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10088 (gimple_stmt_iterator): New typedef.
10089 * gimple.h (gimple_stmt_iterator): Rename to...
10090 (gimple_stmt_iterator_d): ... This.
10091 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10092 trees be valid for GIMPLE and GENERIC.
10093 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10094 * gimple-fold.c (gimple_fold_call): Call target hook
10095 gimple_fold_builtin.
10096 * hooks.c (hook_bool_gsiptr_false): New.
10097 * hooks.h (hook_bool_gsiptr_false): New.
10098 * target.def (fold_stmt): New.
10099 * doc/tm.texi: Regenerate.
10101 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10104 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10105 a set sp if no stack realignment.
10107 2013-04-22 Nick Clifton <nickc@redhat.com>
10109 * config.gcc (tilegx-linux): Extend extra_objs rather than
10111 (tilepro-linux): Likewise.
10113 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10115 * config/aarch64/aarch64-builtins.c
10117 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10118 (VAR<1-12>): Add MAP parameter.
10119 (BUILTIN_*): Likewise.
10120 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10121 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10122 (aarch64_ushl_n<mode>): Likewise.
10123 (aarch64_sshr_n<mode>): Likewise.
10124 (aarch64_ushr_n<mode>): Likewise.
10125 (aarch64_<maxmin><mode>): Likewise.
10126 (aarch64_sqrt<mode>): Likewise.
10127 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10128 (vshr<q>_n_*): Likewise.
10130 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10132 * config/aarch64/aarch64-builtins.c
10133 (aarch64_simd_builtin_type_mode): Handle SF types.
10135 (BUILTIN_GPF): Define.
10136 (aarch64_init_simd_builtins): Handle SF types.
10137 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10138 (frecps): Likewise.
10139 (frecpx): Likewise.
10140 * config/aarch64/aarch64-simd.md
10141 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10142 (aarch64_frecpe<mode>): New.
10143 (aarch64_frecps<mode>): Likewise.
10144 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10145 (v8type): Add frecp<esx>.
10146 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10147 (aarch64_frecps<mode>): Likewise.
10148 * config/aarch64/iterators.md (FRECP): New.
10149 (frecp_suffix): Likewise.
10150 * config/aarch64/arm_neon.h
10151 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10153 2013-04-22 Christian Bruel <christian.bruel@st.com>
10156 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10157 (REG_CLASS_NAMES): Idem.
10158 (REG_CLASS_CONTENTS): Idem.
10159 (REGCLASS_HAS_FP_REG): Idem.
10160 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10161 (sh_conditional_register_usage): Idem.
10163 2013-04-21 Jeff Law <law@redhat.com>
10165 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10166 (ssa_forward_propagate_and_combine): Use it.
10168 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10170 * lra.c: Update the flow chart diagram.
10172 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10174 PR rtl-optimization/56847
10175 * lra-constraints.c (process_alt_operands): Discourage alternative
10176 with non-matche doffsettable memory constraint fro memory with
10179 2013-04-19 Richard Biener <rguenther@suse.de>
10181 PR tree-optimization/56982
10182 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10184 * gimplify.c (gimplify_call_expr): Notice special calls.
10185 (gimplify_modify_expr): Likewise.
10186 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10187 abnormal control flow receivers.
10188 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10189 in the same way as cfun->has_nonlocal_labels.
10190 (gimple_purge_dead_abnormal_call_edges): Likewise.
10191 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10192 receivers start a basic-block.
10194 2013-04-19 Richard Biener <rguenther@suse.de>
10196 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10198 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10199 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10200 (SLP_TREE_LOAD_PERMUTATION): Add.
10201 (vect_transform_slp_perm_load): Adjust prototype.
10202 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10203 (vect_free_slp_instance): Likewise.
10204 (vect_create_new_slp_node): Likewise.
10205 (vect_supported_slp_permutation_p): Remove.
10206 (vect_slp_rearrange_stmts): Adjust.
10207 (vect_supported_load_permutation_p): Likewise. Inline
10208 vect_supported_slp_permutation_p here.
10209 (vect_analyze_slp_instance): Compute load permutations per
10210 slp node instead of per instance.
10211 (vect_get_slp_defs): Adjust.
10212 (vect_transform_slp_perm_load): Likewise.
10213 (vect_schedule_slp_instance): Remove redundant code.
10214 (vect_schedule_slp): Remove hack for PR56270, add it ...
10215 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10216 CSE loads for SLP. Adjust.
10218 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10220 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10221 spelling in two comments.
10223 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10226 * config/arm/arm.c (load_multiple_sequence): Require SP
10227 as base register for loads if SP is in the register list.
10229 2013-04-19 Martin Jambor <mjambor@suse.cz>
10231 PR tree-optimization/56718
10232 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10233 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10234 and made public. Adjusted all callers.
10235 (ipa_intraprocedural_devirtualization): New function.
10236 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10237 (ipa_intraprocedural_devirtualization): Likewise.
10238 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10240 2013-04-19 Richard Biener <rguenther@suse.de>
10242 PR tree-optimization/57000
10243 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10245 2013-04-19 Terry Guo <terry.guo@arm.com>
10247 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10249 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10250 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10251 (cortex_m4_fmacs): Use new reservations.
10252 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10254 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10256 PR rtl-optimization/56999
10257 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10259 (lra_coalesce): Remove split_origin_bitmap and related code.
10260 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10261 ranges if necessary.
10263 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
10265 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10267 (ix86_expand_call): Remove clobbered_registers array and use
10268 x86_64_ms_sysv_extra_clobbered_registers instead.
10269 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10271 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10273 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10274 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10275 (*call_value_rex64_ms_sysv): Ditto.
10277 2013-04-18 Cary Coutant <ccoutant@google.com>
10279 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10280 parent instead of die_mark.
10282 2013-04-18 Diego Novillo <dnovillo@google.com>
10284 * gimple.c (create_gimple_tmp): New.
10285 (get_expr_type): New.
10286 (build_assign): New.
10287 (build_type_cast): New.
10288 * gimple.h (enum ssa_mode): Define.
10289 (gimple_seq_set_location): New.
10290 * asan.c (build_check_stmt): Change some gimple_build_* calls
10291 to use build_assign and build_type_cast.
10293 2013-04-18 Richard Biener <rguenther@suse.de>
10295 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10296 handle negative step. Remove redundant checks.
10297 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10298 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10299 for negative step and grouped loads fail to vectorize.
10301 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10303 * emit-rtl.c (reset_insn_used_flags): New function.
10304 (reset_all_used_flags): Use it.
10305 (verify_insn_sharing): New function.
10306 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10308 2013-04-18 Jakub Jelinek <jakub@redhat.com>
10310 PR tree-optimization/56984
10311 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10312 and (x >> M) >= N don't register any assertion if N << M is the
10315 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10317 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10318 do not delete_insn insns that have not yet been emitted, only
10319 unlink them with remove_insn.
10320 * df-scan.c (df_insn_delete): Revert r197492.
10322 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10324 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10325 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10327 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10329 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10330 define_insn_and_split.
10331 (and_scc,ior_scc,negscc): Likewise.
10332 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10334 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10336 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10337 can benefit from using a sequence of LDRD instructions in epilogue
10338 instead of a single LDM instruction.
10340 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10343 * doc/extend.texi: Fix typo.
10345 2013-04-17 Richard Biener <rguenther@suse.de>
10347 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10348 (vect_build_slp_tree): ... here.
10349 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10350 match. Remove special-casing of mismatched complex loads.
10351 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10352 re-try the match with swapped commutative operands.
10353 (vect_supported_load_permutation_p): Remove special-casing of
10354 mismatched complex loads.
10355 (vect_analyze_slp_instance): Adjust.
10357 2013-04-17 Richard Biener <rguenther@suse.de>
10359 PR rtl-optimization/56921
10360 * cfgloop.h (struct loop): Add simple_loop_desc member.
10361 (struct niter_desc): Mark with GTY(()).
10362 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10363 * loop-iv.c (get_simple_loop_desc): Likewise.
10364 (free_simple_loop_desc): Likewise.
10367 2013-04-16 Richard Biener <rguenther@suse.de>
10369 PR rtl-optimization/56921
10370 * loop-init.c (pass_rtl_move_loop_invariants): Add
10371 TODO_do_not_ggc_collect to todo_flags_finish.
10372 (pass_rtl_unswitch): Same.
10373 (pass_rtl_unroll_and_peel_loops): Same.
10374 (pass_rtl_doloop): Same.
10376 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
10378 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
10379 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
10380 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
10382 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
10383 * tree-streamer.c (record_common_node): Adjust reference in comment.
10385 2013-04-17 Terry Guo <terry.guo@arm.com>
10387 * config/arm/cortex-m4.md: Add a new bypass.
10389 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10391 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
10393 (*subs_<optab><mode>_multp2): New pattern.
10394 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10395 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10397 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10399 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
10400 (*subs_mul_imm_<mode>): New pattern.
10402 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
10405 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
10406 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
10407 (vsx_movti_32bit): Change j->wa to O->wa.
10409 2013-04-16 Richard Biener <rguenther@suse.de>
10411 PR rtl-optimization/56921
10412 * loop-init.c (pass_rtl_move_loop_invariants): Add
10413 TODO_do_not_ggc_collect to todo_flags_finish.
10414 (pass_rtl_unswitch): Same.
10415 (pass_rtl_unroll_and_peel_loops): Same.
10416 (pass_rtl_doloop): Same.
10418 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
10420 * config/arm/arm.c (emit_multi_reg_push): New declaration
10421 for an existing function.
10422 (arm_emit_strd_push): New function.
10423 (arm_expand_prologue): Used here.
10424 (arm_emit_ldrd_pop): New function.
10425 (arm_expand_epilogue): Used here.
10426 (arm_get_frame_offsets): Update condition.
10427 (arm_emit_multi_reg_pop): Add a special case for load of a single
10428 register with writeback.
10430 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
10432 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
10435 2013-04-16 Richard Biener <rguenther@suse.de>
10437 PR tree-optimization/56756
10438 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
10439 (first_mem_ref_loc): New.
10440 (execute_sm): Place the load temporarily before a previous
10441 access instead of in the latch edge to ensure its SSA dependencies
10442 are defined at points dominating the load.
10444 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
10446 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
10447 correct fix by moving header and footer insn to the footer of
10448 the merged basic block. Clear BB_END of the merged-away block.
10450 PR middle-end/43631
10451 * emit-rtl.c (make_note_raw): New function.
10452 (link_insn_into_chain): New static inline function.
10453 (add_insn): Use it.
10454 (add_insn_before, add_insn_after): Factor insn chain linking code...
10455 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
10456 using link_insn_into_chain.
10457 (note_outside_basic_block_p): New helper function for emit_note_after
10458 and emit_note_before.
10459 (emit_note_after): Use nobb variant of add_insn_after if the note
10460 should not be contained in a basic block.
10461 (emit_note_before): Use nobb variant of add_insn_before if the note
10462 should not be contained in a basic block.
10463 (emit_note_copy): Use make_note_raw.
10464 (emit_note): Likewise.
10465 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
10466 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
10467 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
10468 the moved barrier the tail of the basic block it follows.
10469 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
10471 2013-04-15 Jakub Jelinek <jakub@redhat.com>
10473 PR tree-optimization/56962
10474 * gimple-ssa-strength-reduction.c (record_increment): Only set
10475 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
10476 either rhs1 or rhs2 is equal to c->base_expr.
10478 2013-04-15 Richard Biener <rguenther@suse.de>
10480 PR tree-optimization/56933
10481 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
10483 (GROUP_READ_WRITE_DEPENDENCE): Remove.
10484 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
10485 * tree-vect-data-refs.c (vect_analyze_group_access): Move
10486 dependence check ...
10487 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
10489 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
10490 GROUP_READ_WRITE_DEPENDENCE.
10492 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10494 * emit-rtl.c (reset_all_used_flags): New function.
10495 (verify_rtl_sharing): Call reset_all_used_flags before and after
10496 performing the checks.
10498 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10500 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
10501 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
10502 * config/arm/constraints.md (De): New constraint.
10503 * config/arm/neon.md (anddi3_neon): Delete.
10504 (neon_vand<mode>): Expand to standard anddi3 pattern.
10505 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
10506 Move earlier in the file.
10507 (neon_inv_logic_op2): Likewise.
10508 (arm_anddi_operand_neon): New predicate.
10510 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10512 * configure.ac (gcc_cv_ld_as_needed): Set
10513 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
10514 Use -z ignore, -z record on *-*-solaris2*.
10515 (HAVE_LD_AS_NEEDED): Update comment.
10516 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
10517 * configure: Regenerate.
10518 * config.in: Regenerate.
10519 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
10520 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
10521 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
10522 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
10523 equivalents. Fix markup.
10524 * doc/tm.texi: Regenerate.
10526 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
10528 * config/i386/i386.opt: New option mstack-protector-guard=.
10529 * config/i386/i386-opts.h: Add enum stack_protector_guard.
10530 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
10531 TARGET_SSP_TLS_GUARD.
10532 * config/i386/i386.c (ix86_option_override_internal): Set
10533 ix86_stack_protector_guard.
10534 * config/i386/i386.md (stack_protect_set): Enable for
10535 TARGET_SSP_TLS_GUARD only.
10536 (stack_protect_set_<mode>): Ditto.
10537 (stack_protect_test): Ditto.
10538 (stack_protect_test_<mode>): Ditto.
10539 * doc/invoke.texi (i386 Option): Document.
10541 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
10544 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
10545 (S_MODES): Set H_MODE bit.
10546 (SF_MODES): Set only S_MODE and SF_MODE bits.
10547 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
10548 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
10549 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
10550 <MODE_FLOAT>: Likewise.
10552 2013-04-15 Joey Ye <joey.ye@arm.com>
10554 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
10556 2013-04-15 Joey Ye <joey.ye@arm.com>
10558 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
10560 (thumb_far_jump_used_p): Count instruction size and set
10563 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
10565 * reorg.c (fill_simple_delay_slots): Reindent block of code.
10566 * resource.c (mark_target_live_regs): Reformat conditional block.
10568 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10570 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
10571 notes, they are emitted only just before final.
10572 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
10574 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10576 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
10577 * cfgrtl.c (delete_insn): Call it here instead.
10578 * lra-spills.c (lra_final_code_change): Use delete_insn.
10579 * haifa-sched.c (sched_remove_insn): Likewise.
10580 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
10581 returning to the nop pool.
10582 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
10583 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
10585 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10587 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
10588 * doc/tm.texi: Regenerated.
10590 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
10592 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
10595 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10597 * df-core.c (df_find_def): Compare register numbers.
10598 (df_find_use): Likewise.
10600 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
10603 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
10604 lra_in_progress for return.
10606 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10608 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
10609 define_insn into define_insn_and_split and emit movsicc patterns.
10611 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10613 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
10615 2013-04-12 Richard Biener <rguenther@suse.de>
10617 * tree-pass.h (TODO_do_not_ggc_collect): New.
10618 * passes.c (execute_one_ipa_transform_pass): Honor
10619 TODO_do_not_ggc_collect.
10620 (execute_one_pass): Likewise.
10623 2013-04-10 Richard Biener <rguenther@suse.de>
10625 * passes.c (init_optimization_passes): Remove reload pass.
10626 * ira.c (do_reload): Merge into ...
10628 (rest_of_handle_reload): Remove.
10629 (pass_reload): Likewise.
10630 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10631 of reload for vzeroupper pass placement.
10633 2013-04-12 Jakub Jelinek <jakub@redhat.com>
10635 PR tree-optimization/56918
10636 PR tree-optimization/56920
10637 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
10638 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
10639 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
10640 use wide_mul_with_sign method.
10642 2013-04-12 Richard Biener <rguenther@suse.de>
10644 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
10645 not be considered a gimple constant.
10647 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
10649 * fold-const.c (const_binop): Handle vector shifts by a scalar.
10650 (fold_binary_loc): Call const_binop also for mixed vector-scalar
10653 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
10654 Jakub Jelinek <jakub@redhat.com>
10656 * opts.c: Include diagnostic-color.h.
10657 (common_handle_option): Handle OPT_fdiagnostics_color_.
10658 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
10659 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
10660 (diagnostic-color.o): New.
10661 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
10662 (diagnostic_color_rule): New enum.
10663 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
10664 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
10665 the location string.
10666 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
10667 either NULL, or color kind.
10668 * diagnostic-color.c: New file.
10669 * diagnostic-color.h: New file.
10670 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
10672 * doc/invoke.texi (-fdiagnostics-color): Document.
10673 * pretty-print.h (pp_show_color): Define.
10674 (struct pretty_print_info): Add show_color field.
10675 * diagnostic.c: Include diagnostic-color.h.
10676 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
10677 macros. Colorize error:, warning: etc. strings and also the location
10679 (diagnostic_show_locus): Colorize the caret line.
10680 * pretty-print.c: Include diagnostic-color.h.
10681 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
10682 inside of %< %> quotes or quoted through q format modifier.
10684 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10686 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
10688 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10690 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
10691 code in CC_NZ mode.
10692 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
10695 2013-04-11 Marek Polacek <polacek@redhat.com>
10697 PR tree-optimization/48184
10698 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
10700 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
10702 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
10703 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
10704 (skip_simple_arithmetic): Tidy up.
10705 * tree.h (skip_simple_constant_arithmetic): Declare.
10707 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10709 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
10711 2013-04-11 Richard Biener <rguenther@suse.de>
10713 * tree-vect-loop.c (get_initial_def_for_induction): Properly
10714 generate vector constants.
10716 2013-04-11 Richard Biener <rguenther@suse.de>
10718 PR tree-optimization/56878
10719 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
10720 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
10722 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
10723 Prefer to align the DR with the most invariant base address.
10725 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10727 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
10730 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
10732 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
10733 floating-point vector comparisons against 0.
10735 2013-04-11 Jakub Jelinek <jakub@redhat.com>
10737 PR tree-optimization/56899
10738 * fold-const.c (extract_muldiv_1): Apply distributive law
10739 only if TYPE_OVERFLOW_WRAPS (ctype).
10741 2013-04-11 Bin Cheng <bin.cheng@arm.com>
10744 * ira-costs.c (scan_one_insn): Check whether the source rtx of
10745 loading has side effect.
10747 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
10749 * config/sparc/sparc.c: Include tree-pass.h.
10750 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
10751 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
10752 head of file. Change return type. Split off gate function.
10753 (sparc_gate_work_around_errata): New function.
10754 (pass_work_around_errata): New pass definition.
10755 (insert_pass_work_around_errata) New pass insert definition to
10756 insert pass_work_around_errata just after delayed-branch scheduling.
10757 (sparc_option_override): Insert the pass.
10758 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
10760 2013-04-10 David S. Miller <davem@davemloft.net>
10762 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
10763 or -mcpu=hypersparc.
10765 * target.def (cstore_mode): New hook.
10766 * target.h: Include insn-codes.h
10767 * targhooks.c: Likewise.
10768 (default_cstore_mode): New function.
10769 * targhooks.h: Declare it.
10770 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
10771 * doc/tm.texi: Rebuild.
10772 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
10773 target hook, rather than inspecting the insn_data.
10774 * config/sparc/sparc.c (sparc_cstore_mode): New function.
10775 (TARGET_CSTORE_MODE): Redefine.
10776 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
10778 * config/sparc/predicates.md (cstore_result_operand): New special
10780 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
10781 Use it for operand 0.
10782 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
10783 (*snesi_special): Likewise.
10784 (*snesi_zero): Likewise.
10785 (*seqsi_zero): Likewise.
10786 (*sltu_insn): Likewise.
10787 (*sgeu_insn): Likewise.
10788 (*seqdi_special): Make operand 0 and comparison operation be of
10790 (*snedi_special): Likewise.
10791 (*snedi_special_vis3): Likewise.
10792 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
10793 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
10794 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
10795 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
10796 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
10797 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
10798 (*sltu_extend_sp64): Likewise.
10799 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
10800 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
10801 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
10802 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
10803 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
10805 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
10807 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
10808 (aarch64_start_file): Use the new function.
10810 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10811 Jason Merrill <jason@redhat.com>
10813 * common.opt: Add -gdwarf.
10814 * opts.c (common_handle_option): Handle it.
10815 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
10817 2013-04-10 Richard Biener <rguenther@suse.de>
10819 * passes.c (execute_todo): Do not call ggc_collect conditional here.
10820 (execute_one_ipa_transform_pass): But unconditionally here.
10821 (execute_one_pass): And here.
10822 (init_optimization_passes): Remove reload pass.
10823 * tree-pass.h (TODO_ggc_collect): Remove.
10824 (pass_reload): Likewise.
10825 * ira.c (do_reload): Merge into ...
10827 (rest_of_handle_reload): Remove.
10828 (pass_reload): Likewise.
10829 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10830 of reload for vzeroupper pass placement.
10831 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
10832 and todo_flags_finish of all passes.
10834 2013-04-10 Richard Biener <rguenther@suse.de>
10836 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
10837 first_const_oprnd field, rename first_def_type to first_op_type.
10838 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
10839 (vect_get_and_check_slp_defs): Always use the type of the
10840 operand. Allow mixed vect_external_def, vect_constant_def types.
10841 (vect_get_constant_vectors): Handle mixed vect_external_def,
10842 vect_constant_def types.
10844 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
10846 PR tree-optimization/55524
10847 * tree-ssa-math-opts.c
10848 (convert_mult_to_fma): Don't use an fms construct
10849 when we don't have an fms operation, but fnma, and it looks
10850 likely that we'll be able to use the latter.
10852 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
10854 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
10856 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
10857 inline fail caused by overwritable functions.
10859 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
10861 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
10862 unnecessary bits in the constant power of two case.
10864 2013-04-10 Richard Biener <rguenther@suse.de>
10866 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
10867 broken code swapping operands.
10868 (vect_build_slp_tree): Do not compute load permutations here.
10869 (vect_analyze_slp_instance): Compute load permutations here,
10870 after building the SLP tree.
10872 2013-04-09 Christian Bruel <christian.bruel@st.com>
10874 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
10875 of next/prev_real_insn.
10877 2013-04-09 Jan Hubicka <jh@suse.cz>
10879 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10880 Drop aliased parameter.
10881 (function_and_variable_visibility): Do not handle alias pairs.
10882 * cgraph.c (varpool_externally_visible_p): Update prototype.
10883 * varpool.c (varpool_add_new_variable): Update.
10885 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10887 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
10889 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
10891 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
10893 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
10895 2013-04-09 Marek Polacek <polacek@redhat.com>
10897 PR tree-optimization/48762
10898 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
10900 2013-04-09 Richard Biener <rguenther@suse.de>
10902 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
10904 (vect_build_slp_tree): Likewise.
10905 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
10906 calculating the cost of a SLP instance.
10907 (vect_analyze_slp_instance): Use it from here, after building
10910 2013-04-09 Jakub Jelinek <jakub@redhat.com>
10912 PR middle-end/56883
10913 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
10914 expand_omp_for_static_chunk): Use simple_p = true in
10915 force_gimple_operand_gsi calls when assigning to addressable decls.
10917 2013-04-09 Jeff Law <law@redhat.com>
10919 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
10920 when the boolean was created by converting a wider object which
10921 had a boolean range.
10923 2013-04-09 Richard Biener <rguenther@suse.de>
10925 * tree-vectorizer.h (slp_void_p): Remove.
10926 (slp_tree): Typedef before _slp_tree declaration.
10927 (struct _slp_tree): Use a vector of slp_tree as children.
10928 (vect_get_place_in_interleaving_chain): Remove.
10929 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
10931 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
10933 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
10934 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
10935 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
10936 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
10937 Use slp_node instead of slp_void_p and adjust.
10939 2013-04-09 Richard Biener <rguenther@suse.de>
10941 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
10942 work that is not necessary.
10944 2013-04-09 Jakub Jelinek <jakub@redhat.com>
10946 PR tree-optimization/56854
10947 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
10948 forward into clobber stmts if it would change MEM_REF lhs into
10951 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
10953 * tree.c (type_hash_lookup, type_hash_add): Make static.
10954 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
10956 2013-04-09 Richard Biener <rguenther@suse.de>
10958 * tree.h (unsave_expr_now): Remove.
10959 * tree-inline.c (mark_local_for_remap_r): Remove.
10960 (unsave_expr_1): Likewise.
10961 (unsave_r): Likewise.
10962 (unsave_expr_now): Likewise.
10963 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
10964 (propagate_tree_value): Likewise.
10966 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
10968 * doc/rtl.texi (sequence): Rewrite documentation to match the
10969 current use of SEQUENCE rtl objects.
10970 * rtl.def (SEQUENCE): Likewise.
10972 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
10973 Update documentation.
10974 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
10975 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
10977 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
10979 2013-04-08 Teresa Johnson <tejohnson@google.com>
10981 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
10982 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
10984 (estimate_edge_size_and_time): Add comment to suggest using rounding
10986 (estimate_node_size_and_time): Ditto.
10987 (remap_edge_change_prob): Use helper rounding divide methods.
10988 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
10989 (gimple_mod_pow2_value_transform): Ditto.
10990 (gimple_mod_subtract_transform): Ditto.
10991 (gimple_ic_transform): Ditto.
10992 (gimple_stringops_transform): Ditto.
10993 * stmt.c (conditional_probability): Ditto.
10994 (emit_case_dispatch_table): Ditto.
10995 * lto-cgraph.c (merge_profile_summaries): Ditto.
10996 * tree-optimize.c (execute_fixup_cfg): Ditto.
10997 * cfgcleanup.c (try_forward_edges): Ditto.
10998 * cfgloopmanip.c (scale_loop_profile): Ditto.
11000 (duplicate_loop_to_header_edge): Ditto.
11001 (lv_adjust_loop_entry_edge): Ditto.
11002 * tree-vect-loop.c (vect_transform_loop): Ditto.
11003 * profile.c (compute_branch_probabilities): Ditto.
11004 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
11005 * lto-streamer-in.c (input_cfg): Ditto.
11006 * gimple-streamer-in.c (input_bb): Ditto.
11007 * ipa-cp.c (update_profiling_info): Ditto.
11008 (update_specialized_profile): Ditto.
11009 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
11010 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
11012 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11013 (compute_trg_info): Ditto.
11014 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
11015 (purge_dead_edges): Ditto.
11016 * loop-unswitch.c (unswitch_loop): Ditto.
11017 * cgraphclones.c (cgraph_clone_edge): Ditto.
11018 (cgraph_clone_node): Ditto.
11019 * tree-inline.c (copy_bb): Ditto.
11020 (copy_edges_for_bb): Ditto.
11021 (initialize_cfun): Ditto.
11022 (copy_cfg_body): Ditto.
11023 (expand_call_inline): Ditto.
11025 2013-04-08 Kai Tietz <ktietz@redhat.com>
11027 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11028 TARGET_CYGWIN64 by TARGET_64BIT.
11030 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11032 * config/epiphany/epiphany.md (GPR_1): New constant.
11033 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11034 * config/epiphany/epiphany.c (gen_compare_reg):
11035 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11036 is already in place.
11037 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11038 Don't require being called during rtl expansion; If y operlaps r0,
11040 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11041 (epiphany_expand_epilogue): Likewise.
11043 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11044 Don't use CC_FPmode for ORDERED / UNORDERED.
11045 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
11047 * config/epiphany/constraints.md (CnL): New constraint.
11048 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11049 * config/epiphany/predicates.md (add_operand): Allow 1024.
11051 * config/epiphany/epiphany.md (logical_op): New code iterator.
11052 (op_mnc): New code attribute.
11053 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11054 (mov_f+1, mov_f+2): New peephole2 patterns.
11056 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11057 (cstoresi4): Also allow re-use of zero result when doing a NE
11058 comparison to a non-zero operand.
11059 Use (clobber (scratch)) for first insn if the gpr output is not needed.
11061 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11062 Use gen_addsi3_i / gen_subsi3_i.
11064 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11068 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11069 contain anything but clobbers, at most one __builtin_stack_restore,
11070 optionally debug stmts and final resx, and if it has at least one
11071 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11072 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11073 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11074 which isn't defaut definition, remove them.
11075 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11076 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11077 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11078 with MEM_REF LHS with SSA_NAME address.
11080 2013-04-08 Jeff Law <law@redhat.com>
11082 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
11084 2013-04-08 Richard Biener <rguenther@suse.de>
11086 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11088 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11089 determined vector type.
11090 (vect_analyze_data_refs): Likewise.
11091 (vect_get_new_vect_var): Adjust.
11092 (vect_create_destination_var): Preserve SSA name versions.
11093 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11094 not dump anything here.
11096 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11098 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11099 Add member lr_slot_known.
11100 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11102 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11103 Remove code that sets lr_slot_offset according to what a previous
11104 version of epiphany_emit_save_restore used to do.
11105 (epiphany_emit_save_restore): When doing an lr save or restore,
11106 set/verify lr_slot_known and lr_slot_offset.
11108 2013-04-08 Xinyu Qi <xyqi@marvell.com>
11111 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
11114 2013-04-08 Richard Biener <rguenther@suse.de>
11116 * alias.c (find_base_term): Fix thinko in previous change.
11118 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11120 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11121 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11122 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11123 if possible to compute val.
11124 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11125 For QImode integers don't require anything about precision. Use
11126 const_with_all_bytes_same to find out if the constant doesn't have
11127 repeated bytes in it.
11129 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11131 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11134 2013-04-08 Marek Polacek <polacek@redhat.com>
11136 PR rtl-optimization/48182
11137 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11140 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11143 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11144 nuses, make sure we have a label.
11146 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11149 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11150 (rs6000_emit_swdiv_low_precision): Remove.
11151 (rs6000_emit_swdiv): Rewrite to handle between one and four
11152 iterations of Newton-Raphson generally; modify required number of
11153 iterations for some cases.
11154 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11156 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11158 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11159 set-but-unused variable.
11161 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11162 basic blocks of released function bodies garbage-collectable.
11164 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11165 (struct rtl_opt_pass): Add TODO_df_finish.
11167 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11169 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11171 * config/arm/constraints.md (q): New constraint.
11172 * config/arm/ldrdstrd.md: New file.
11173 * config/arm/arm.md (ldrdstrd.md) New include.
11174 (arm_movdi): Use "q" instead of "r" constraint
11175 for double-word memory access.
11176 (movdf_soft_insn): Likewise.
11177 * config/arm/vfp.md (movdi_vfp): Likewise.
11178 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
11179 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
11180 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11181 (mem_ok_for_ldrd_strd): Likewise.
11182 (output_move_double): Update assertion.
11184 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11186 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11188 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11190 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11191 define_insn_and_split.
11192 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11194 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11196 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11197 define_insn_and_split.
11198 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
11199 (shiftsi3_compare): New pattern.
11200 (rrx): New pattern.
11201 * config/arm/unspecs.md (UNSPEC_RRX): New.
11203 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11205 * config/arm/arm.md (negdi_extendsidi): New pattern.
11206 (negdi_zero_extendsidi): Likewise.
11208 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11210 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11211 define_insn_and_split.
11212 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11213 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11215 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11217 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11218 define_insn_and_split.
11219 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11220 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11222 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11224 * config/arm/arm.md (subsi3_carryin): New pattern.
11225 (subsi3_carryin_const): Likewise.
11226 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11227 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11229 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11231 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11233 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11235 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
11236 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
11238 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11240 * config/arm/arm.c (arm_expand_builtin): Change fcode
11241 type to unsigned int.
11243 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11245 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11247 2013-04-04 Ian Lance Taylor <iant@google.com>
11249 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11252 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11254 PR middle-end/56729
11255 * df-scan.c (df_insn_delete): Disable failing assert.
11257 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11259 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11260 New function prototype.
11261 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11262 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11263 (arm_builtin_vectorized_function): New function.
11265 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11267 * config/arm/arm_neon_builtins.def: New file.
11268 * config/arm/arm.c (neon_builtin_data): Move contents to
11269 arm_neon_builtins.def.
11270 (enum arm_builtins): Include neon builtin definitions.
11271 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
11272 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
11274 2013-04-04 Marek Polacek <polacek@redhat.com>
11276 PR tree-optimization/48186
11277 * predict.c (maybe_hot_frequency_p): Return false if
11278 HOT_BB_FREQUENCY_FRACTION is 0.
11279 (cgraph_maybe_hot_edge_p): Likewise.
11281 2013-04-04 Richard Biener <rguenther@suse.de>
11283 PR tree-optimization/56826
11284 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11287 2013-04-04 Richard Biener <rguenther@suse.de>
11289 PR tree-optimization/56213
11290 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
11291 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
11293 2013-04-04 Richard Biener <rguenther@suse.de>
11295 PR tree-optimization/56837
11296 * tree-loop-distribution.c (classify_partition): For non-zero
11297 values require that the value has the same precision as its
11298 mode to be useful as memset value.
11300 2013-04-03 Nick Clifton <nickc@redhat.com>
11302 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
11303 (fmssf4): Use fmsf.s on E3V5 architectures.
11304 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11305 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11307 2013-04-03 Jeff Law <law@redhat.com>
11309 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11310 (lra-eliminations.o): Likewise.
11312 2013-04-03 Teresa Johnson <tejohnson@google.com>
11314 * gcov-io.c (compute_working_sets): Moved most of body of old
11315 compute_working_sets here from profile.c.
11316 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
11317 (gcov_working_set_t): Moved typedef here from basic-block.h
11318 (compute_working_set): Declare.
11319 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11320 (get_working_sets): Renamed from compute_working_set,
11321 replace most of body with call to new compute_working_sets.
11322 (get_exec_counts): Replace call to compute_working_sets
11323 to get_working_sets.
11324 * profile.h (get_working_sets): Renamed from compute_working_set.
11325 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
11326 to get_working_sets.
11327 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11328 * gcov-dump.c (dump_working_sets): New function.
11330 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11332 * hwint.c (sext_hwi, zext_hwi): New functions.
11333 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11334 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11335 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11336 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11337 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11338 (sext_hwi, zext_hwi): New functions.
11340 2013-04-03 Jeff Law <law@redhat.com>
11342 PR tree-optimization/56799
11343 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11344 back test for widening conversion erroneously dropped in prior change.
11346 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11349 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11350 instead of next_real_insn.
11352 2013-04-03 Marek Polacek <polacek@redhat.com>
11355 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
11357 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11360 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11362 (thumb1_output_casesi): Likewise.
11363 (thumb2_output_casesi): Likewise.
11365 2013-04-03 Richard Biener <rguenther@suse.de>
11367 PR tree-optimization/56817
11368 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11370 (tree_unroll_loops_completely_1): ... new function to manually
11371 walk the loop tree, properly defering outer loops of unrolled
11372 loops to later iterations.
11374 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11376 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
11377 (vectorizable_load): Likewise.
11378 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
11379 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
11381 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11383 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
11386 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11388 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
11390 2013-04-03 Bin Cheng <bin.cheng@arm.com>
11392 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
11394 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11396 PR tree-optimization/56790
11397 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
11400 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11402 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
11404 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
11405 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
11408 2013-04-03 Jakub Jelinek <jakub@redhat.com>
11411 * tree.h (force_folding_builtin_constant_p): New decl.
11412 * builtins.c (force_folding_builtin_constant_p): New variable.
11413 (fold_builtin_constant_p): Fold immediately also if
11414 force_folding_builtin_constant_p.
11416 2013-04-03 Richard Biener <rguenther@suse.de>
11418 PR tree-optimization/56812
11419 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
11420 DRs of the same interleaving chain are independent.
11422 2013-04-02 Jason Merrill <jason@redhat.com>
11424 * gdbinit.in (pbb): Use debug fn.
11426 2013-04-02 Lawrence Crowl <crowl@google.com>
11428 * sese.h (struct ivtype_map_elt_s): Remove unused.
11429 (extern debug_ivtype_map): Remove unused.
11430 (extern eq_ivtype_map_elts): Remove unused.
11431 * sese.c (debug_ivtype_map): Removed unused.
11432 (debug_ivtype_map_1): Removed unused.
11433 (debug_ivtype_elt): Remove unused.
11434 (eq_ivtype_map_elts): Remove unused.
11437 2013-04-02 Kai Tietz <ktietz@redhat.com>
11440 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
11441 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
11442 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
11444 (legitimize_pe_coff_symbol): Likewise.
11445 (is_imported_p): New helper-function.
11446 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
11447 for Windows x64 targets.
11448 (ix86_expand_prologue): Optimize for pe-coff targets.
11449 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
11450 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
11451 medium/large code-model.
11452 (legitimize_pic_address): Likewise.
11453 (legitimize_tls_address): Likewise.
11454 (ix86_expand_call): Likewise.
11455 (x86_output_mi_thunk): Likewise.
11456 (get_dllimport_decl): Add new beimport argument.
11457 (construct_plt_address): Don't assert for x64 pe-coff targets.
11458 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
11460 (SYMBOL_FLAG_STUBVAR): New macro.
11461 (SYMBOL_REF_STUBVAR_P): Likewise.
11462 * config/i386/winnt.c (stub_list): New structure.
11463 (stub_head): New local variable.
11464 (i386_pe_record_stub): New function.
11465 (i386_pe_file_end): Emit refptr-stubs.
11467 2013-04-02 Jakub Jelinek <jakub@redhat.com>
11469 PR rtl-optimization/56745
11470 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
11471 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
11474 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
11475 and both of them are MEM_REFs, just compare first argument for
11476 equality and attempt to deal even with differing offsets.
11479 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
11480 of gimple_clobber_p to be MEM_REF.
11481 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
11482 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
11483 after gimplification.
11484 * asan.c (get_mem_ref_of_assignment): Don't instrument
11485 gimple_clobber_p stmts.
11486 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
11487 gimple_clobber_p stmt if they have MEM_REF lhs and
11488 are dead because of another gimple_clobber_p stmt.
11489 * tree-ssa-live.c (clear_unused_block_pointer): Treat
11490 gimple_clobber_p stmts like debug stmts.
11491 (remove_unused_locals): Remove clobbers with MEM_REF lhs
11492 that refer to unused VAR_DECLs or uninitialized values.
11493 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
11494 gimple_clobber_p stmts if they refer to removed parameters.
11495 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
11498 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
11500 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
11501 using SWI48 mode attribute.
11503 2013-04-02 Wei Mi <wmi@google.com>
11505 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
11506 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
11507 *<rotate_insn><mode>3_mask in i386.md.
11509 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
11511 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
11513 2013-04-02 Richard Biener <rguenther@suse.de>
11515 PR tree-optimization/56778
11516 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
11517 Runtime alias tests are not supported for gather loads.
11518 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
11519 stmts referenced from SSA operands before updating SSA form.
11521 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
11522 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11524 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
11525 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
11526 * config/arm/cortex-a53.md: New file.
11527 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
11528 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
11529 * config/arm/arm.c (arm_issue_rate): Likewise.
11530 * config/arm/arm-tune.md: Regenerate
11531 * config/arm/arm-tables.opt: Regenerate.
11532 * config/arm/arm-cores.def: Add cortex-a53.
11534 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
11536 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
11539 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11541 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
11542 scalar load/store operations using B/H registers.
11543 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
11545 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11547 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
11549 * config/aarch64/aarch64.c
11550 (aarch64_simd_scalar_immediate_valid_for_move): New.
11551 * config/aarch64/aarch64-protos.h
11552 (aarch64_simd_scalar_immediate_valid_for_move): New.
11553 * config/aarch64/constraints.md (Dh, Dq): New.
11554 * config/aarch64/iterators.md (hq): New.
11556 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
11558 * reorg.c (get_branch_condition): Deal with conditional returns.
11559 (fill_simple_delay_slots): Remove dead code dealing with jumps.
11561 2013-04-01 Wei Mi <wmi@google.com>
11563 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
11564 Truncate operand 2 using %b asm operand modifier.
11565 (*<shift_insn><mode>3_mask): Ditto.
11566 (*<rotate_insn><mode>3_mask): Ditto.
11568 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
11570 PR middle-end/56798
11571 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
11573 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
11575 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
11577 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
11579 2013-03-30 Lawrence Crowl <crowl@google.com>
11581 * dse.c (clear_alias_sets): Remove never set.
11582 (disqualified_clear_alias_sets): Remove never set.
11583 (clear_alias_mode_pool): Remove never set.
11584 (dse_step0): Remove condition that is never true.
11585 (canon_address): Remove condition that is never true.
11586 (dse_step7): Remove condition that is never true.
11587 (rest_of_handle_dse): Remove condition that is never true.
11588 (rest_of_handle_dse::did_global): Remove never read from above.
11589 (dse_step2_spill): Remove never called from above.
11590 (dse_step5_spill): Remove never called from above.
11592 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
11594 * doc/md.texi (Standard Names) <casesi>: Update documentation for
11595 JUMP_TABLE_DATA changes.
11596 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
11597 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
11598 (Insns) <jump_table_data>: New entry.
11599 * doc/tm.texi: Regenerate.
11601 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
11603 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
11604 for table jump at the end of a basic block using tablejump_p.
11605 * targhooks.c (default_invalid_within_doloop): Likewise.
11606 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
11607 target hook implementation that is identical to the default hook.
11608 (rs6000_invalid_within_doloop): Remove.
11610 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
11611 unused variable from tablejump_p call.
11613 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
11614 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
11615 (INSN_DELETED_P): Likewise.
11616 (emit_jump_table_data): New prototype.
11617 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
11618 after 4th as unused.
11619 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
11620 * sched-vis.c (print_insn): Likewise.
11621 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
11622 insn for compatibility with back ends that use next_active_insn to
11623 identify jump table data.
11624 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
11625 (remove_insn): Likewise.
11626 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
11628 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
11629 (emit_jump_table_data): New function.
11631 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
11632 basic block, a JUMP_TABLE_DATA never is.
11633 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
11634 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
11635 off from code handling real insns.
11636 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
11637 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
11638 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
11639 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
11640 is not a NONDEBUG_INSN_P.
11641 * ira-costs.c (scan_one_insn): Likewise.
11642 * jump.c (mark_all_labels): Likewise.
11643 (mark_jump_label_1): Likewise.
11644 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11645 * lra.c (get_insn_freq): Expect all insns reaching here to be in
11647 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
11648 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
11649 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
11650 JUMP_TABLE_DATA_P insns.
11651 (calculate_elim_costs_all_insns): Likewise.
11652 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
11653 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
11654 (delete_output_reload): Code style fixups.
11655 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
11656 insn flags on this non-insn.
11657 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
11658 as scheduling barriers, for pre-change compatibility.
11659 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
11660 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
11662 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
11663 redundant JUMP_TABLE_DATA_P test.
11664 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
11665 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11666 (frv_for_each_packet): Likewise.
11667 * config/i386/i386.c (min_insn_size): Likewise.
11668 (ix86_avoid_jump_mispredicts): Likewise.
11669 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11670 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
11671 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11672 (mips16_insn_length): Robustify.
11673 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
11674 (mips16_split_long_branches): Likewise.
11675 * config/pa/pa.c (pa_combine_instructions): Likewise.
11676 * config/rs6000/rs6000.c (get_next_active_insn): Treat
11677 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
11678 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
11679 as contributing to pool range lengths.
11680 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
11681 Remove redundant JUMP_TABLE_DATA_P test.
11682 (sh_loop_align): Likewise.
11683 (split_branches): Likewise.
11684 (sh_insn_length_adjustment): Likewise.
11685 * config/spu/spu.c (get_branch_target): Likewise.
11687 2013-03-29 Jan Hubicka <jh@suse.cz>
11689 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
11690 gcov streaming; stream hot bb threshold to ltrans.
11691 * predict.c (get_hot_bb_threshold): Break out from ....
11692 (maybe_hot_count_p): ... here.
11693 (set_hot_bb_threshold): New function.
11694 * lto-section-in.c (lto_section_name): Add profile.
11695 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
11696 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
11697 and data-streamer.h
11698 (histogram_entry): New structure.
11699 (histogram, histogram_pool): New global vars.
11700 (histogram_hash): New structure.
11701 (histogram_hash::hash): New method.
11702 (histogram_hash::equal): Likewise.
11703 (account_time_size): New function.
11704 (cmp_counts): New function.
11705 (dump_histogram): New function.
11706 (ipa_profile_generate_summary): New function.
11707 (ipa_profile_write_summary): New function.
11708 (ipa_profile_read_summary): New function.
11709 (ipa_profile): Decide on threshold.
11710 (pass_ipa_profile): Add ipa_profile_write_summary and
11711 ipa_profile_read_summary.
11712 * Makefile.in (ipa.o): Update dependencies.
11713 * lto-streamer.h (LTO_section_ipa_profile): New section.
11715 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
11717 * tree.h (VAR_P): New.
11719 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
11722 * doc/invoke.texi ([-fwhole-program]): Fix typo.
11724 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
11726 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
11727 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11728 (control_flow_insn_p): Likewise.
11729 * cfgrtl.c (duplicate_insn_chain): Likewise.
11730 * final.c (get_attr_length_1): Likewise.
11731 (shorten_branches): Likewise.
11732 (final_scan_insn): Likewise.
11733 * function.c (instantiate_virtual_regs): Likewise.
11734 * gcse.c (insert_insn_end_basic_block): Likewise.
11735 * ira-costs.c (scan_one_insn): Likewise.
11736 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11737 * lra.c (check_rtl): Likewise.
11738 * reload1.c (elimination_costs_in_insn): Likewise.
11739 * reorg.c (follow_jumps): Likewise.
11741 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
11742 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11743 (thumb_far_jump_used_p): Likewise.
11744 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
11745 (workaround_speculation): Likewise.
11746 (add_sched_insns_for_speculation): Likewise.
11747 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
11748 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11749 (frv_for_each_packet): Likewise.
11750 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
11751 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
11752 (final_emit_insn_group_barriers): Likewise.
11753 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11754 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11755 (mips16_insn_length): Likewise.
11756 * config/pa/pa.c (pa_reorg): Likewise.
11757 (pa_combine_instructions): Likewise.
11758 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
11759 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
11760 (sh_reorg): Likewise.
11761 (split_branches): Likewise.
11762 * config/spu/spu.c (get_branch_target): Likewise.
11764 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
11767 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
11769 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
11770 Fix declaration name.
11772 2013-03-28 Lawrence Crowl <crowl@google.com>
11774 * graphds.h (struct graph.indicies): Remove unused.
11775 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
11776 (SCOP_ORIGINAL_PDDRS): Remove unused.
11777 * sese.h (extern insert_loop_close_phis): Removed unused.
11778 (extern insert_guard_phis): Removed unused.
11779 (extern ivtype_map_elt_info): Removed unused.
11780 (new_ivtype_map_elt): Removed unused.
11781 * sese.c (ivtype_map_elt_info): Removed unused.
11783 2013-03-28 Lawrence Crowl <crowl@google.com>
11785 * Makefile.in: Add several missing include dependences.
11787 (test-dump.o): New. This object is not added to any executable,
11788 but is present for ad-hoc testing.
11790 (debug (const bitmap_head_def &)): New.
11791 (debug (const bitmap_head_def *)): New.
11793 (extern debug (const bitmap_head_def &)): New.
11794 (extern debug (const bitmap_head_def *)): New.
11796 (debug (edge_def &)): New.
11797 (debug (edge_def *)): New.
11799 (debug (basic_block_def &)): New.
11800 (debug (basic_block_def *)): New.
11802 (dump_node (const_tree, int, FILE *)): Correct source file.
11804 (debug (die_struct &)): New.
11805 (debug (die_struct *)): New.
11807 (extern debug (die_struct &)): New.
11808 (extern debug (die_struct *)): New.
11809 * gimple-pretty-print.c
11810 (debug (gimple_statement_d &)): New.
11811 (debug (gimple_statement_d *)): New.
11812 * gimple-pretty-print.h
11813 (extern debug (gimple_statement_d &)): New.
11814 (extern debug (gimple_statement_d *)): New.
11816 (debug (ira_allocno_copy &)): New.
11817 (debug (ira_allocno_copy *)): New.
11818 (debug (ira_allocno &)): New.
11819 (debug (ira_allocno *)): New.
11821 (extern debug (ira_allocno_copy &)): New.
11822 (extern debug (ira_allocno_copy *)): New.
11823 (extern debug (ira_allocno &)): New.
11824 (extern debug (ira_allocno *)): New.
11826 (debug (live_range &)): New.
11827 (debug (live_range *)): New.
11829 (debug (lra_live_range &)): New.
11830 (debug (lra_live_range *)): New.
11832 (debug (lra_live_range &)): New.
11833 (debug (lra_live_range *)): New.
11835 (debug (omega_pb_d &)): New.
11836 (debug (omega_pb_d *)): New.
11838 (extern debug (omega_pb_d &)): New.
11839 (extern debug (omega_pb_d *)): New.
11841 (debug (const rtx_def &)): New.
11842 (debug (const rtx_def *)): New.
11844 (debug_tree (tree): Move within file.
11845 (debug_raw (const tree_node &)): New.
11846 (debug_raw (const tree_node *)): New.
11847 (dump_tree_via_hooks (const tree_node *, int)): New.
11848 (debug (const tree_node &)): New.
11849 (debug (const tree_node *)): New.
11850 (debug_verbose (const tree_node &)): New.
11851 (debug_verbose (const tree_node *)): New.
11852 (debug_head (const tree_node &)): New.
11853 (debug_head (const tree_node *)): New.
11854 (debug_body (const tree_node &)): New.
11855 (debug_body (const tree_node *)): New.
11856 (debug_vec_tree (tree): Move and reimplement in terms of dump.
11857 (debug (vec<tree, va_gc> &)): New.
11858 (debug (vec<tree, va_gc> *)): New.
11860 (extern debug (const rtx_def &)): New.
11861 (extern debug (const rtx_def *)): New.
11863 (debug_raw (simple_bitmap_def &)): New.
11864 (debug_raw (simple_bitmap_def *)): New.
11865 (debug (simple_bitmap_def &)): New.
11866 (debug (simple_bitmap_def *)): New.
11868 (extern debug (simple_bitmap_def &)): New.
11869 (extern debug (simple_bitmap_def *)): New.
11870 (extern debug_raw (simple_bitmap_def &)): New.
11871 (extern debug_raw (simple_bitmap_def *)): New.
11873 (debug (vinsn_def &)): New.
11874 (debug (vinsn_def *)): New.
11875 (debug_verbose (vinsn_def &)): New.
11876 (debug_verbose (vinsn_def *)): New.
11877 (debug (expr_def &)): New.
11878 (debug (expr_def *)): New.
11879 (debug_verbose (expr_def &)): New.
11880 (debug_verbose (expr_def *)): New.
11881 (debug (vec<rtx> &)): New.
11882 (debug (vec<rtx> *)): New.
11884 (extern debug (vinsn_def &)): New.
11885 (extern debug (vinsn_def *)): New.
11886 (extern debug_verbose (vinsn_def &)): New.
11887 (extern debug_verbose (vinsn_def *)): New.
11888 (extern debug (expr_def &)): New.
11889 (extern debug (expr_def *)): New.
11890 (extern debug_verbose (expr_def &)): New.
11891 (extern debug_verbose (expr_def *)): New.
11892 (extern debug (vec<rtx> &)): New.
11893 (extern debug (vec<rtx> *)): New.
11895 (_list_iter_cond_expr): Make inline instead of static.
11897 (debug (sreal &)): New.
11898 (debug (sreal *)): New.
11900 (extern debug (sreal &)): New.
11901 (extern debug (sreal *)): New.
11903 (extern debug_raw (const tree_node &)): New.
11904 (extern debug_raw (const tree_node *)): New.
11905 (extern debug (const tree_node &)): New.
11906 (extern debug (const tree_node *)): New.
11907 (extern debug_verbose (const tree_node &)): New.
11908 (extern debug_verbose (const tree_node *)): New.
11909 (extern debug_head (const tree_node &)): New.
11910 (extern debug_head (const tree_node *)): New.
11911 (extern debug_body (const tree_node &)): New.
11912 (extern debug_body (const tree_node *)): New.
11913 (extern debug (vec<tree, va_gc> &)): New.
11914 (extern debug (vec<tree, va_gc> *)): New.
11916 (debug (struct loop &)): New.
11917 (debug (struct loop *)): New.
11918 (debug_verbose (struct loop &)): New.
11919 (debug_verbose (struct loop *)): New.
11920 * tree-dump.c: Add header dependence.
11922 (extern debug (struct loop &)): New.
11923 (extern debug (struct loop *)): New.
11924 (extern debug_verbose (struct loop &)): New.
11925 (extern debug_verbose (struct loop *)): New.
11927 (debug (data_reference &)): New.
11928 (debug (data_reference *)): New.
11929 (debug (vec<data_reference_p> &)): New.
11930 (debug (vec<data_reference_p> *)): New.
11931 (debug (vec<ddr_p> &)): New.
11932 (debug (vec<ddr_p> *)): New.
11934 (extern debug (data_reference &)): New.
11935 (extern debug (data_reference *)): New.
11936 (extern debug (vec<data_reference_p> &)): New.
11937 (extern debug (vec<data_reference_p> *)): New.
11938 (extern debug (vec<ddr_p> &)): New.
11939 (extern debug (vec<ddr_p> *)): New.
11941 (debug (pt_solution &)): New.
11942 (debug (pt_solution *)): New.
11944 (extern debug (pt_solution &)): New.
11945 (extern debug (pt_solution *)): New.
11947 (debug (_var_map &)): New.
11948 (debug (_var_map *)): New.
11949 (debug (tree_live_info_d &)): New.
11950 (debug (tree_live_info_d *)): New.
11952 (extern debug (_var_map &)): New.
11953 (extern debug (_var_map *)): New.
11954 (extern debug (tree_live_info_d &)): New.
11955 (extern debug (tree_live_info_d *)): New.
11957 2013-03-28 Jan Hubicka <jh@suse.cz>
11959 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
11961 2013-03-28 Ian Bolton <ian.bolton@arm.com>
11963 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
11964 record only when desired or required.
11966 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
11968 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
11969 *vec_extractv2di_1_rex64. Use x64 isa attribute.
11971 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11973 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
11974 (*andsi3_compare0_uxtw): New pattern.
11975 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
11976 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
11978 2013-03-28 Jan Hubicka <jh@suse.cz>
11980 * data-streamer-in.c (streamer_read_gcov_count): New function.
11981 * gimple-streamer-out.c: Include value-prof.h.
11982 (output_gimple_stmt): Output histogram.
11983 (output_bb): Use streamer_write_gcov_count.
11984 * value-prof.c: Include data-streamer.h
11985 (dump_histogram_value): Add HIST_TYPE_MAX.
11986 (stream_out_histogram_value): New function.
11987 (stream_in_histogram_value): New function.
11988 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
11989 (stream_out_histogram_value, stream_in_histogram_value): Declare.
11990 * data-streamer-out.c (streamer_write_gcov_count): New function.
11991 (streamer_write_gcov_count_stream): New function.
11992 * lto-cgraph.c (lto_output_edge): Update counter streaming.
11993 (lto_output_node): Likewise.
11994 (input_node, input_edge): Likewise.
11995 * lto-streamer-out.c (output_cfg): Update streaming.
11996 * lto-streamer-in.c (input_cfg): Likewise.
11997 * data-streamer.h (streamer_write_gcov_count,
11998 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
11999 * gimple-streamer-in.c: Include value-prof.h
12000 (input_gimple_stmt): Input histograms.
12001 (input_bb): Update profile streaming.
12003 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12005 * genmodes.c (emit_max_int): New function.
12006 (emit_insn_modes_h): Added call to emit_max_function.
12007 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
12009 * machmode.def: Fixed comment.
12011 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12013 * combine.c (try_combine): Removed useless assert.
12014 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
12016 2013-03-28 Marek Polacek <polacek@redhat.com>
12017 Richard Biener <rguenther@suse.de>
12019 PR tree-optimization/56695
12020 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
12021 build signed result of a vector comparison.
12022 * tree-cfg.c (verify_gimple_comparison): Check that a result
12023 of a vector comparison has signed type.
12025 2013-03-28 Richard Biener <rguenther@suse.de>
12027 PR tree-optimization/37021
12028 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12029 do not restrict gaps between groups.
12030 * tree-vect-stmts.c (vectorizable_load): Properly account for
12031 a gap between groups.
12033 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12035 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12036 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12039 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12041 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12042 * doc/extend.texi (Named Address Spaces): Ditto.
12043 (Variable Attributes): Ditto.
12045 2013-03-27 Kai Tietz <ktietz@redhat.com>
12047 * config.build: Add support for cygwin x64 target.
12048 * config.gcc: Likewise.
12049 * config.host: Likewise.
12050 * configure.ac: Likewise
12051 * configure: Regenerated.
12053 2013-03-27 Kai Tietz <ktietz@redhat.com>
12055 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12056 * config/i386/t-cygwin-w64: New file.
12057 * config/i386/cygwin-w64.h: New file.
12058 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12059 and add support for x64-cygwin target.
12060 (CPP_SPEC): Likewise.
12061 (CXX_WRAP_SPEC_LIST): Undefine before define.
12062 (LIBGCJ_SONAME): Use 15 as version.
12064 2013-03-27 Richard Biener <rguenther@suse.de>
12066 PR tree-optimization/56716
12067 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12068 dumping for ref nodes.
12070 2013-03-27 Martin Jambor <mjambor@suse.cz>
12072 PR tree-optimization/55334
12073 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12074 restricted pointers to arrays.
12076 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12078 * Makefile.in (.SUFFIXES): Add .cc.
12079 (.c.o): Apply same recipe for implicit rule .cc.o.
12081 2013-03-27 Richard Biener <rguenther@suse.de>
12083 PR tree-optimization/37021
12084 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12085 REALPART/IMAGPART_EXPRs around the supported refs.
12086 * tree-ssa-structalias.c (find_func_aliases): Assume that
12087 floating-point values are not used to transfer pointers.
12089 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12091 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12092 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12093 * doc/tm.texi: Regenerate.
12094 * targhooks.h (default_has_ifunc_p): New.
12095 * targhooks.c (default_has_ifunc_p): Ditto.
12096 * config/linux-protos.h: New file.
12097 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
12098 this hook for linux which disables support of indirect functions in
12100 * config/linux-android.c: New file.
12101 * config/t-linux-android.c: Ditto.
12102 * config.gcc: Added new object file linux-android.o.
12103 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
12104 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
12105 * varasm.c (do_assemble_alias): Likewise.
12106 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12107 doesn't support indirect functions.
12108 * configure: Regenerate.
12110 2013-03-27 Bin Cheng <bin.cheng@arm.com>
12113 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12114 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12117 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12119 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12121 2013-03-27 Terry Guo <terry.guo@arm.com>
12123 * config/arm/arm-cores.def: Added core cortex-r7.
12124 * config/arm/arm-tune.md: Regenerated.
12125 * config/arm/arm-tables.opt: Regenerated.
12126 * doc/invoke.texi: Added entry for core cortex-r7.
12128 2013-03-27 Walter Lee <walt@tilera.com>
12130 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12131 double-decrement of next_scratch_regno.
12133 2013-03-27 Walter Lee <walt@tilera.com>
12135 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
12137 (insn_v1mulus): Ditto.
12138 (insn_v2muls): Ditto.
12140 2013-03-27 Walter Lee <walt@tilera.com>
12142 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
12143 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12145 2013-03-27 Walter Lee <walt@tilera.com>
12147 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
12148 (*sibcall_value): Ditto.
12150 2013-03-27 Walter Lee <walt@tilera.com>
12152 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12153 (insn_mnz_v8qi): ... this ...
12154 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12156 (insn_v<n>mnz): Replaced by ...
12157 (insn_v1mnz): ... this ...
12158 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12160 (insn_mz_<mode>): Replaced by ...
12161 (insn_mz_v8qi): ... this ...
12162 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12164 (insn_v<n>mz): Replaced by ...
12165 (insn_v1mz): ... this ...
12166 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12169 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12171 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12173 2013-03-26 Roland McGrath <mcgrathr@google.com>
12175 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12176 than fprintf with a non-constant, non-format string.
12178 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12180 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12181 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12182 operand 0 predicate.
12183 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12184 attribute. Use general_x64nomem_operand as operand 1 predicate.
12185 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12186 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12187 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12188 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12189 (mov<mode>_insv_1): Remove expander. Merge insn with
12190 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12191 Use general_x64nomem_operand as operand 1 predicate.
12192 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12193 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12194 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12195 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12196 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12197 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12198 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12199 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12200 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12201 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12202 (general_x64nomem_operand): Ditto.
12204 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
12206 * config/rtems.opt: Add -pthread option.
12208 2013-03-26 Richard Biener <rguenther@suse.de>
12210 * alias.c (find_base_term): Avoid redundant and not used recursion.
12211 (base_alias_check): Get the initial base term from the caller.
12212 (true_dependence_1): Compute and pass base terms to base_alias_check.
12213 (write_dependence_p): Likewise.
12214 (may_alias_p): Likewise.
12216 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12218 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12219 PC-relative load in SI modes and above only.
12221 2013-03-26 Xinyu Qi <xyqi@marvell.com>
12223 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
12224 * config/arm/iwmmxt.md (WCGR0): Update.
12225 (WCGR1, WCGR2, WCGR3): Likewise.
12227 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12229 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12230 Use x64 and nox64 isa attributes.
12232 2013-03-26 Richard Biener <rguenther@suse.de>
12234 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12235 alignment computations and rely on get_object_alignment_1
12236 for the !TYPE_P case.
12237 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12239 2013-03-26 Walter Lee <walt@tilera.com>
12241 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12242 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12244 2013-03-25 Jeff Law <law@redhat.com>
12246 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12247 check for INTEGRAL_TYPE_P that was missing due to checking in
12248 wrong version of prior patch.
12250 2013-03-25 Walter Lee <walt@tilera.com>
12252 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12253 TILEGX_INSN_SHUFFLEBYTES1.
12254 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12256 (tilegx_builtins): Ditto.
12257 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12259 2013-03-25 Walter Lee <walt@tilera.com>
12261 * config/tilegx/tilegx.md (floatsisf2): New pattern.
12262 (floatunssisf2): New pattern.
12263 (floatsidf2): New pattern.
12264 (floatunssidf2): New pattern.
12266 2013-03-25 Walter Lee <walt@tilera.com>
12268 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12269 tests for constraint J, K, N, P.
12271 2013-03-25 Walter Lee <walt@tilera.com>
12273 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12274 Use indirect/pcrel encoding.
12275 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12278 2013-03-25 Steve Ellcey <sellcey@mips.com>
12280 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12281 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12282 * config/mips/mips.c (mips_option_override): Set IMADD default.
12283 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12284 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12285 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12286 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12287 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12289 2013-03-25 Jeff Law <law@redhat.com>
12291 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12292 slightly to avoid creating and folding useless trees. Simplify
12293 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12295 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
12297 * config/i386/i386.md (*zero_extendsidi2): Merge with
12298 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12299 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12300 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12301 targets. Clarify comment.
12303 2013-03-25 Martin Jambor <mjambor@suse.cz>
12305 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12306 pass-through jump functions differently.
12307 (ipa_read_jump_function): Likewise. Also use setter functions to set
12310 2013-03-25 Martin Jambor <mjambor@suse.cz>
12312 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12313 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12315 (ipa_get_indirect_edge_target): New function.
12316 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12317 ipa_get_indirect_edge_target. Update all callers.
12318 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12319 ipa_get_indirect_edge_target_1 instead of calling
12320 ipa_get_indirect_edge_target.
12321 (create_specialized_node): Pass aggvlas to
12322 ipcp_discover_new_direct_edges.
12324 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12326 * config/arm/arm.md (f_sels, f_seld): New types.
12327 (*cmov<mode>): New pattern.
12328 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12331 2013-03-25 Kai Tietz <ktietz@redhat.com>
12333 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12334 POSIX-printf for mingw-hosted builds.
12336 2013-03-25 Richard Biener <rguenther@suse.de>
12338 PR middle-end/56694
12339 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12340 must-not-throw stmt location.
12342 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12344 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12345 Emit load-acquire versions when acq is true.
12346 (arm_emit_store_exclusive): Add rel parameter.
12347 Emit store-release versions when rel is true.
12348 (arm_split_compare_and_swap): Use acquire-release instructions
12350 of barriers when appropriate.
12351 (arm_split_atomic_op): Likewise.
12352 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12353 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12354 (VUNSPEC_SLX): Likewise.
12355 (VUNSPEC_LDA): Likewise.
12356 (VUNSPEC_STL): Likewise.
12357 * config/arm/sync.md (atomic_load<mode>): New pattern.
12358 (atomic_store<mode>): Likewise.
12359 (arm_load_acquire_exclusive<mode>): Likewise.
12360 (arm_load_acquire_exclusivesi): Likewise.
12361 (arm_load_acquire_exclusivedi): Likewise.
12362 (arm_store_release_exclusive<mode>): Likewise.
12364 2013-03-25 Catherine Moore <clm@codesourcery.com>
12366 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12367 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12368 * config/mip/predicates.md (lwsp_swsp_operand,
12369 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12370 sb16_operand, db4_operand, db7_operand, ib3_operand,
12371 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
12372 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
12373 andi16_operand): New predicates.
12374 * config/mips/mips.md (compression): New attribute.
12375 (enabled): New attribute.
12376 (length): Consider compression in computing length.
12377 (shift_compression): New code attribute.
12378 (*add<mode>3): New operands. Record compression.
12379 (sub<mode>3): Likewise.
12380 (one_cmpl<mode>2): Likewise.
12381 (*and<mode>3): Likewise.
12382 (*ior<mode>3): Likewise.
12383 (unnamed pattern for xor): Likewise.
12384 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
12385 (*<optab><mode>3): Likewise.
12386 (*mov<mode>_internal: Likewise.
12387 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
12388 (mips_unsigned_immediate_p): New.
12389 (umips_lwsp_swsp_address_p): New.
12390 (m16_based_address_p): New.
12391 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
12392 (mips_unsigned_immediate_p): New prototype.
12393 (lwsp_swsp_address_p): New prototype.
12394 (m16_based_address_p): New prototype.
12395 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
12396 (mips_signed_immediate_p): New function.
12397 (m16_based_address_p): New function.
12398 (lwsp_swsp_address_p): New function.
12399 (mips_print_operand_punctuation): Recognize short delay slot insns
12400 for microMIPS.add<mode>3"
12402 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12405 * config/arm/iterators.md (v_cmp_result): New mode attribute.
12406 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
12408 2013-03-25 Richard Biener <rguenther@suse.de>
12410 PR tree-optimization/56689
12411 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
12414 2013-03-25 Richard Biener <rguenther@suse.de>
12416 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
12418 (memory_references): Likewise.
12419 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
12420 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
12421 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
12422 (gather_mem_refs_in_loops): Fold into ...
12423 (analyze_memory_references): ... this. Move initialization
12424 to tree_ssa_lim_initialize.
12425 (fill_always_executed_in): Rename to ...
12426 (fill_always_executed_in_1): ... this.
12427 (fill_always_executed_in): Move contains_call computation to
12428 this new function from ...
12429 (tree_ssa_lim_initialize): ... here.
12430 (tree_ssa_lim): Call fill_always_executed_in.
12432 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12434 * postreload.c (reload_combine): Fix code detecting returns.
12436 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12438 * function.c (emit_use_return_register_into_block): On cc0 targets,
12439 do not emit the sequence between cc0 setter and user.
12441 2013-03-25 Kai Tietz <ktietz@redhat.com>
12443 * config/i386/predicates.md (local_symbolic_operand): Interpret
12444 dll-imported symbols as none-local.
12446 2013-03-25 Richard Biener <rguenther@suse.de>
12448 * tree-ssa-loop-im.c (struct depend): Remove.
12449 (struct lim_aux_data): Make depends a vec of gimples.
12450 (free_lim_aux_data): Adjust.
12451 (add_dependency): Likewise.
12452 (set_level): Likewise.
12454 2013-03-25 Richard Biener <rguenther@suse.de>
12456 PR middle-end/56434
12457 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
12458 the pointer returned by calls with ECF_MALLOC set.
12460 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12462 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
12464 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12466 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
12467 using MMXMODE mode iterator.
12468 (*move<mode>_internal): Merge with *movv2sf_internal and
12469 *movv2sf_internal_rex64 using MMXMODE mode iterator.
12471 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12473 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
12474 (record_last_mem_set_info): Likewise.
12476 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
12477 of XNEWVEC followed by memset.
12478 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
12480 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12482 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
12483 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
12484 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
12485 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
12486 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
12487 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
12488 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
12489 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
12490 BARRIER_P instead of GET_CODE.
12492 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12494 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
12495 inaccuracy in the probing code.
12497 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
12498 (ctrapdi4): Likewise.
12500 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12502 * calls.c (expand_call): Add missing guard to code handling return
12503 of non-BLKmode structures in MSB.
12504 * function.c (expand_function_end): Likewise.
12506 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12508 * combine.c (try_combine): Adjust comment. Do not add the set of
12509 insn #0 if the destination indirectly is set or dies in insn #2.
12510 Tidy up code to distribute a new note.
12512 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12514 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
12515 also for alternatives 16 and 17.
12517 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12519 * config/i386/sse.md (*mov<mode>_internal): Merge with
12520 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
12521 Emit insn template depending on type attribute. Use
12522 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12523 movd instead of movq mnemonic for interunit moves. Rewrite mode
12524 attribute calculation. Remove unit attribute calculation.
12525 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
12526 Set prefix_data16 attribute for DImode ssemov types.
12527 Use Ym instead of y for SSE-MMX conversion alternatives.
12528 Reorder operand constraints.
12530 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
12532 * df.h (df_insn_delete): Adjust prototype.
12533 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
12534 and let it decide whether mark the basic block dirty.
12535 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
12536 * df-scan.c (df_insn_info_delete): New helper function, split
12537 off from df_insn_delete.
12538 (df_scan_free_bb_info): Use it.
12539 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
12541 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
12542 that the insn is actually an insn and it has a non-NULL basic block.
12543 Do not mark basic block dirty if only deleting a DEBUG_INSN.
12545 2013-03-22 Richard Biener <rguenther@suse.de>
12547 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
12549 (mem_ref_alloc): Do not allocate them.
12550 (refs_independent_p): Do not query or maintain a cache.
12552 2013-03-22 Richard Biener <rguenther@suse.de>
12554 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
12555 (gather_mem_refs_in_loops): Do not compute it.
12556 (analyze_memory_references): Do not allocate it.
12557 (tree_ssa_lim_finalize): Do not free it.
12558 (for_all_locs_in_loop): Do not query all_refs_in_loop.
12560 2013-03-22 Richard Biener <rguenther@suse.de>
12562 * is-a.h (as_a): Use gcc_checking_assert.
12564 2013-03-22 Ian Bolton <ian.bolton@arm.com>
12566 * config/aarch64/aarch64.c (aarch64_print_operand): New
12567 format specifier for printing a constant in hex.
12568 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
12569 format specifier for printing second operand.
12571 2013-03-22 Richard Biener <rguenther@suse.de>
12573 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
12575 (gather_mem_refs_in_loops): Perform store accumulation here.
12576 (create_vop_ref_mapping_loop): Remove.
12577 (create_vop_ref_mapping): Likewise.
12578 (analyze_memory_references): Initialize refs_stored_in_loop.
12579 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
12580 (record_indep_loop): Remove.
12581 (record_dep_loop): New function.
12582 (ref_indep_loop_p_1): Adjust to only walk over references
12583 in the loop, not its subloops.
12584 (ref_indep_loop_p): Rename to ...
12585 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
12586 maintaining a more fine-grained cache.
12587 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
12588 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
12590 2013-03-22 Richard Biener <rguenther@suse.de>
12592 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
12593 (struct mem_ref): Make accesses_in_loop a vec of a vec of
12594 aggregate mem_ref_loc.
12595 (free_mem_ref_locs): Inline into ...
12596 (memref_free): ... this and adjust.
12597 (mem_ref_alloc): Adjust.
12598 (mem_ref_locs_alloc): Remove.
12599 (record_mem_ref_loc): Adjust.
12600 (get_all_locs_in_loop): Rewrite into ...
12601 (for_all_locs_in_loop): ... this iterator.
12602 (rewrite_mem_ref_loc): New functor.
12603 (rewrite_mem_refs): Use for_all_locs_in_loop.
12604 (sm_set_flag_if_changed): New functor.
12605 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
12606 (ref_always_accessed): New functor.
12607 (ref_always_accessed_p): Use for_all_locs_in_loop.
12609 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12611 * tree-pass.h (PROP_gimple_lvec): New.
12612 * passes.c (dump_properties): Handle PROP_gimple_lvec.
12613 (init_optimization_passes): Move pass_lower_vector.
12614 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
12616 (pass_lower_vector): Provide PROP_gimple_lvec.
12617 (pass_lower_vector_ssa): Likewise.
12618 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
12620 2013-03-21 Mark Wielaard <mjw@redhat.com>
12622 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
12624 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
12626 * config/i386/i386.md (*movdi_internal): Disparage slightly
12627 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
12628 conversion alternatives.
12630 2013-03-21 Jakub Jelinek <jakub@redhat.com>
12632 PR middle-end/48087
12633 * diagnostic.def (DK_WERROR): New kind.
12634 * diagnostic.h (werrorcount): Define.
12635 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
12636 promoted to DK_ERROR, increment DK_WERROR counter instead of
12638 * toplev.c (toplev_main): Call print_ignored_options even if
12639 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
12640 even if just werrorcount is non-zero.
12643 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
12645 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
12646 (string_cst_pool_decl): New function.
12647 (optimize_one_addr_into_implicit_ptr): New function.
12648 (resolve_addr_in_expr): Optimize DWARF location expression
12649 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
12650 which doesn't live in memory, but has DW_AT_location or
12651 DW_AT_const_value, or refers to a string literal, into
12652 DW_OP_GNU_implicit_pointer.
12653 (optimize_location_into_implicit_ptr): New function.
12654 (resolve_addr): If removing DW_AT_location of a variable because
12655 it was DW_OP_addr of address of the variable, but the variable doesn't
12656 live in memory, try to emit const value attribute for the initializer.
12658 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12660 * tree.h (VECTOR_TYPE_P): New macro.
12661 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
12662 TYPE_MODE): Use it.
12663 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
12664 VEC_COND_EXPR cannot be lvalues.
12665 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
12667 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12669 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
12670 Restrict the transformation to equal modes.
12672 2013-03-21 Richard Biener <rguenther@suse.de>
12674 PR tree-optimization/39326
12675 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
12676 (MEM_ANALYZABLE): Adjust.
12677 (record_mem_ref_loc): Move bitmap ops ...
12678 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
12679 unanalyzable refs, do not record locations for it.
12680 (analyze_memory_references): Allocate ref zero as shared
12682 (refs_independent_p): Do not test for unanalyzed mems here.
12683 (ref_indep_loop_p_1): Special-case disambiguation against
12684 the unanalyzed ref.
12685 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
12687 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
12689 * config/arm/arm-protos.h (tune_params): Add
12690 prefer_neon_for_64bits field.
12691 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
12692 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
12693 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
12694 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
12695 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
12696 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
12697 (arm_option_override): Handle -mneon-for-64bits new option.
12698 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
12699 (prefer_neon_for_64bits): Declare new variable.
12700 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
12701 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
12702 (arch_enabled): Handle new arch types. Remove support for onlya8
12704 (one_cmpldi2): Use new arch names.
12705 (zero_extend<mode>di2, extend<mode>di2): Ditto.
12706 * config/arm/arm.opt (mneon-for-64bits): Add option.
12707 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
12708 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
12709 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
12711 * doc/invoke.texi (-mneon-for-64bits): Document.
12713 2013-03-21 Richard Biener <rguenther@suse.de>
12715 PR tree-optimization/39326
12716 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
12717 (sort_bbs_in_loop_postorder_cmp): New function.
12718 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
12720 2013-03-21 Richard Biener <rguenther@suse.de>
12722 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
12723 (vect_insert_into_interleaving_chain): Likewise.
12724 (vect_drs_dependent_in_basic_block): Inline ...
12725 (vect_slp_analyze_data_ref_dependence): ... here. New function,
12727 (vect_analyze_data_ref_dependence): ... here. Simplify.
12728 (vect_check_interleaving): Simplify.
12729 (vect_analyze_data_ref_dependences): Likewise. Split out ...
12730 (vect_slp_analyze_data_ref_dependences): ... this new function.
12731 (dr_group_sort_cmp): New function.
12732 (vect_analyze_data_ref_accesses): Compute data-reference groups
12733 here instead of in vect_analyze_data_ref_dependence. Use
12734 a more efficient algorithm.
12735 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
12736 vect_slp_analyze_data_ref_dependences. Call
12737 vect_analyze_data_ref_accesses earlier.
12738 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
12739 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
12740 (vect_slp_analyze_data_ref_dependences): New prototype.
12742 2013-03-21 Richard Biener <rguenther@suse.de>
12744 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
12745 ref is stored in the loop.
12746 (find_refs_for_sm): Walk only over all stores.
12747 (store_motion_loop): Allocate from lim_bitmap_obstack.
12748 (store_motion): Likewise.
12750 2013-03-21 Richard Biener <rguenther@suse.de>
12752 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
12753 Update virtual SSA form.
12755 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12757 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
12758 * configure: Regenerate.
12759 * config.in: Regenerate.
12760 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
12761 if !HAVE_LD_EH_FRAME_CIEV3.
12763 2013-03-21 Richard Biener <rguenther@suse.de>
12765 * tree-cfg.c (verify_expr_no_block): New function.
12766 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
12767 nor DECL_VALUE_EXPR have locations with associated blocks.
12768 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
12769 (clear_unused_block_pointer): Remove code dealing with
12770 blocks in DECL_DEBUG_EXPR locations.
12772 2013-03-21 Richard Biener <rguenther@suse.de>
12774 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
12775 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
12776 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
12777 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
12778 instead of DECL_DEBUG_EXPR_IS_FROM.
12779 * gimplify.c (gimplify_modify_expr): Likewise.
12780 * tree-cfg.c (verify_expr_location_1): Likewise.
12781 * tree-complex.c (create_one_component_var): Likewise.
12782 * tree-sra.c (create_access_replacement): Likewise.
12783 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
12784 (clear_unused_block_pointer): Likewise.
12785 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
12786 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
12787 * var-tracking.c (var_debug_decl): Likewise.
12788 (track_expr_p): Likewise.
12789 * tree-inline.c (add_local_variables): Likewise. Set
12790 DECL_HAS_DEBUG_EXPR_P after copying it.
12791 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
12792 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
12794 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
12797 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
12798 * configure: Regenerate.
12799 * config.in: Regenerate.
12800 * config/i386/i386.md (*movdf_internal): Use
12801 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12802 movd instead of movq mnemonic for interunit moves.
12803 (*movdi_internal): Ditto.
12805 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12807 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
12808 (abd<mode>_3): New pattern.
12809 (aba<mode>_3): New pattern.
12810 (fabd<mode>_3): New pattern.
12812 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12814 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
12815 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
12816 occurrence of REGISTER_PREFIX as its empty string.
12818 2013-03-20 Jeff Law <law@redhat.com>
12820 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
12821 addititional equivalences for equality comparisons between an SSA_NAME
12822 and a constant where the SSA_NAME was set from a widening conversion.
12824 2013-03-20 Walter Lee <walt@tilera.com>
12826 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
12828 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
12830 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
12831 depending on type attribute.
12832 (*movti_internal): Ditto.
12833 (*movtf_internal): Ditto.
12834 (*movxf_internal): Ditto.
12835 (*movdf_internal): Ditto.
12836 (*movsf_internal): Ditto.
12838 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
12840 * config/i386/i386.md (*movti_internal): Set prefix attribute to
12841 maybe_vex for sselog1 and ssemov types.
12842 (*movdi_internal): Reorder operand constraints.
12843 (*movsi_internal): Ditto. Set prefix attribute to
12844 maybe_vex for sselog1 and ssemov types.
12845 (*movtf_internal): Set prefix attribute to maybe_vex
12846 for sselog1 and ssemov types.
12847 (*movdf_internal): Ditto. Set prefix_data16 attribute for
12848 DImode ssemov types. Reorder operand constraints.
12849 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
12850 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
12851 attribute for SImode ssemov types. Reorder operand constraints.
12853 2013-03-20 Martin Jambor <mjambor@suse.cz>
12855 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
12856 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
12858 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
12860 * config/rs6000/predicates.md (indexed_address, update_address_mem
12861 update_indexed_address_mem): New predicates.
12862 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
12863 attribute for load/store instructions.
12864 * config/rs6000/dfp.md (movsd_store): Likewise.
12865 (movsd_load): Likewise.
12866 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
12867 (unnamed HI->DI extend define_insn): Likewise.
12868 (unnamed SI->DI extend define_insn): Likewise.
12869 (unnamed QI->SI extend define_insn): Likewise.
12870 (unnamed QI->HI extend define_insn): Likewise.
12871 (unnamed HI->SI extend define_insn): Likewise.
12872 (unnamed HI->SI extend define_insn): Likewise.
12873 (extendsfdf2_fpr): Likewise.
12874 (movsi_internal1): Likewise.
12875 (movsi_internal1_single): Likewise.
12876 (movhi_internal): Likewise.
12877 (movqi_internal): Likewise.
12878 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
12879 attribute for load/store instructions.
12880 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
12882 (mov<mode>_softfloat): Likewise.
12883 (mov<mode>_hardfloat32): Likewise.
12884 (mov<mode>_hardfloat64): Likewise.
12885 (mov<mode>_softfloat64): Likewise.
12886 (movdi_internal32): Likewise.
12887 (movdi_internal64): Likewise.
12888 (probe_stack_<mode>): Likewise.
12890 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
12892 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
12893 floating point, and decimal floating point to reload iterator.
12895 * config/rs6000/constraints.md (wl constraint): New constraints to
12896 return FLOAT_REGS if certain options are used to reduce the number
12897 of separate patterns that exist in the file.
12898 (wx constraint): Likewise.
12899 (wz constraint): Likewise.
12901 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
12902 -mdebug=reg, print wg, wl, wx, and wz constraints.
12903 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
12904 Initialize the reload functions for 64-bit binary/decimal floating
12906 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
12907 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
12908 create the buffer on the stack to overcome not having a 32-bit
12910 (rs6000_emit_move): Likewise.
12911 (rs6000_secondary_memory_needed_rtx): Likewise.
12912 (rs6000_alloc_sdmode_stack_slot): Likewise.
12913 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
12914 via xxlxor, just like DFmode 0.0.
12916 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
12917 define as 1 if we are running on a power7 or newer.
12918 (enum r6000_reg_class_enum): Add new constraints.
12920 * config/rs6000/dfp.md (movsd): Delete, combine with binary
12921 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
12922 with other moves by using conditional constraits (wg). Use LFIWZX
12923 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
12924 (movsd splitter): Likewise.
12925 (movsd_hardfloat): Likewise.
12926 (movsd_softfloat): Likewise.
12928 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
12929 binary and decimal floating point moves.
12930 (fmove_ok): New attributes to combine binary and decimal floating
12931 point moves, and to combine power6x (mfpgpr) moves along normal
12933 (real_value_to_target): Likewise.
12934 (f32_lr): Likewise.
12935 (f32_lm): Likewise.
12936 (f32_li): Likewise.
12937 (f32_sr): Likewise.
12938 (f32_sm): Likewise.
12939 (f32_si): Likewise.
12940 (movsf): Combine binary and decimal floating point moves. Combine
12941 power6x (mfpgpr) moves with other moves by using conditional
12942 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
12943 (mov<mode> for SFmode/SDmode); Likewise.
12944 (SFmode/SDmode splitters): Likewise.
12945 (movsf_hardfloat): Likewise.
12946 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
12947 (movsf_softfloat): Likewise.
12948 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
12950 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
12951 wx and wz constraints.
12953 * config/rs6000/constraints.md (wg constraint): New constraint to
12954 return FLOAT_REGS if -mmfpgpr (power6x) was used.
12956 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
12959 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
12960 -mdebug=reg, print wg, wl, wx, and wz constraints.
12961 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
12962 Initialize the reload functions for 64-bit binary/decimal floating
12964 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
12965 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
12966 create the buffer on the stack to overcome not having a 32-bit
12968 (rs6000_emit_move): Likewise.
12969 (rs6000_secondary_memory_needed_rtx): Likewise.
12970 (rs6000_alloc_sdmode_stack_slot): Likewise.
12971 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
12972 via xxlxor, just like DFmode 0.0.
12974 * config/rs6000/dfp.md (movdd): Delete, combine with binary
12975 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
12976 with other moves by using conditional constraits (wg). Use LFIWZX
12977 and STFIWX for loading SDmode on power7.
12978 (movdd splitters): Likewise.
12979 (movdd_hardfloat32): Likewise.
12980 (movdd_softfloat32): Likewise.
12981 (movdd_hardfloat64_mfpgpr): Likewise.
12982 (movdd_hardfloat64): Likewise.
12983 (movdd_softfloat64): Likewise.
12985 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
12986 64-bit binary and decimal floating point moves.
12987 (FMOVE64X): Likewise.
12988 (movdf): Combine 64-bit binary and decimal floating point moves.
12989 Combine power6x (mfpgpr) moves with other moves by using
12990 conditional constraits (wg).
12991 (mov<mode> for DFmode/DDmode): Likewise.
12992 (DFmode/DDmode splitters): Likewise.
12993 (movdf_hardfloat32): Likewise.
12994 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
12995 (movdf_softfloat32): Likewise.
12996 (movdf_hardfloat64_mfpgpr): Likewise.
12997 (movdf_hardfloat64): Likewise.
12998 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
12999 (movdf_softfloat64): Likewise.
13000 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
13001 (reload_<mode>_load): Move to later in the file so they aren't in
13002 the middle of the floating point move insns.
13003 (reload_<mode>_store): Likewise.
13005 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
13008 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
13009 constraint if -mdebug=reg.
13010 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
13011 Enable using dd reload support if needed.
13013 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
13014 binary and decimal floating point moves in rs6000.md.
13015 (movtd_internal): Likewise.
13017 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
13018 decimal floating point moves.
13020 (movtf_internal): Likewise.
13021 (mov<mode>_internal, TDmode/TFmode): Likewise.
13022 (movtf_softfloat): Likewise.
13023 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
13025 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13026 movdi_internal64, using wg constraint for move direct operations.
13027 (movdi_internal64): Likewise.
13029 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13030 MODES_TIEABLE_P for selected modes. Print the numerical value of
13031 the various virtual registers. Use GPR/FPR first/last values,
13032 instead of hard coding the register numbers. Print which modes
13033 have reload functions registered.
13034 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13035 settings before/after setting cpu, target and subtarget settings.
13036 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13037 and for secondary reload failures in rs6000_secondary_reload_inner.
13038 (rs6000_secondary_reload_fail): Likewise.
13039 (rs6000_secondary_reload_inner): Likewise.
13041 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13042 macros for first/last GPR and FPR registers.
13043 (LAST_GPR_REGNO): Likewise.
13044 (FIRST_FPR_REGNO): Likewise.
13045 (LAST_FPR_REGNO): Likewise.
13047 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13048 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13049 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13050 (vcond<mode><mode>): Likewise.
13051 (vcondu<mode><mode>): Likewise.
13052 (vector_gtu<mode>): Likewise.
13053 (vector_gte<mode>): Likewise.
13054 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
13055 to prevent the compiler from converting DImode operations to TImode.
13056 (ior<mode>3): Likewise.
13057 (and<mode>3): Likewise.
13058 (one_cmpl<mode>2): Likewise.
13059 (nor<mode>3): Likewise.
13060 (andc<mode>3): Likewise.
13062 * config/rs6000/constraints.md (wt constraint): New constraint
13063 that returns VSX_REGS if TImode is allowed in VSX registers.
13065 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13066 constant under VSX.
13068 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13069 similar to TImode, but it is restricted to being in the GPRs.
13071 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13072 TImode to occupy a single VSX register.
13074 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13075 -mvsx-timode for power7/power8.
13076 (power7 cpu): Likewise.
13077 (power8 cpu): Likewise.
13079 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13080 sure that TFmode/TDmode take up two registers if they are ever
13081 allowed in the upper VSX registers.
13082 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13084 (rs6000_init_hard_regno_mode_ok): Likewise.
13085 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13086 constraint. Print if LRA is turned on.
13087 (rs6000_option_override_internal): Give an error if -mvsx-timode
13088 and VSX is not enabled.
13089 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13090 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13091 to reg+offset addressing. Use PTImode when checking offset
13092 addresses for validity.
13093 (reg_offset_addressing_ok_p): Likewise.
13094 (rs6000_legitimate_offset_address_p): Likewise.
13095 (rs6000_legitimize_address): Likewise.
13096 (rs6000_legitimize_reload_address): Likewise.
13097 (rs6000_legitimate_address_p): Likewise.
13098 (rs6000_eliminate_indexed_memrefs): Likewise.
13099 (rs6000_emit_move): Likewise.
13100 (rs6000_secondary_reload): Likewise.
13101 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13102 reloads to fpr registers to continue to use reg+offset addressing,
13103 but 64-bit reloads to altivec registers need reg+reg addressing.
13104 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13105 it. Treat LO_SUM like a PLUS operation.
13106 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
13107 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
13108 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13109 registers to share a register with a smaller sized type, since VSX
13110 puts scalars in the upper 64-bits.
13111 (print_operand): Add support for PTImode.
13112 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13113 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13114 registers, but don't have arithmetic support.
13115 (rs6000_memory_move_cost): Add test for VSX.
13116 (rs6000_opt_masks): Add -mvsx-timode.
13118 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13121 (VSr): Use wt constraint for TImode.
13122 (VSv): Drop TImode support.
13123 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13124 (vsx_movti_64bit): Likewise.
13125 (vsx_movti_32bit): Likewise.
13126 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13127 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13128 one '?' on the appropriate output constraint. Do not allow TImode
13129 logical operations on 32-bit systems.
13130 (vsx_ior<mode>3): Likewise.
13131 (vsx_xor<mode>3): Likewise.
13132 (vsx_one_cmpl<mode>2): Likewise.
13133 (vsx_nor<mode>3): Likewise.
13134 (vsx_andc<mode>3): Likewise.
13135 (vsx_concat_<mode>): Likewise.
13136 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13138 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13139 OPTION_MASK_VSX_TIMODE.
13140 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13141 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13143 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13144 (TI2 iterator): New iterator for TImode, PTImode.
13145 (wd mode attribute): Add values for vector types.
13146 (movti_string): Replace TI move operations with operations for TImode
13147 and PTImode. Add support for TImode being allowed in VSX registers.
13148 (mov<mode>_string, TImode/PTImode): Likewise.
13149 (movti_ppc64): Likewise.
13150 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13151 (TI mode splitters): Likewise.
13153 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13156 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
13158 PR tree-optimization/56355
13159 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13160 Also handle integers with undefined overflow.
13162 2013-03-20 Catherine Moore <clm@codesourcery.com>
13163 Maciej W. Rozycki <macro@codesourcery.com>
13164 Tom de Vries <tom@codesourcery.com>
13165 Nathan Sidwell <nathan@codesourcery.com>
13166 Iain Sandoe <iain@codesourcery.com>
13167 Nathan Froyd <froydnj@codesourcery.com>
13168 Chao-ying Fu <fu@mips.com>
13170 * doc/extend.texi: (micromips, nomicromips, nocompression):
13171 Document new function attributes.
13172 * doc/invoke.texi (minterlink-compressed, mmicromips,
13173 m14k, m14ke, m14kec): Document new options.
13174 (minterlink-mips16): Update documentation.
13175 * doc/md.texi (ZC, ZD): Document new constraints.
13176 * configure.ac (gcc_cv_as_micromips): Check if linker
13177 supports the .set micromips directive.
13178 * configure: Regenerate.
13179 * config.in: Regenerate.
13180 * config/mips/mips-tables.opt: Regenerate.
13181 * config/mips/micromips.md: New file.
13182 * constraints.md (ZC, ZD): New constraints.
13183 * config/mips/predicates.md (movep_src_register): New predicate.
13184 (movep_src_operand): New predicate.
13185 (non_volatile_mem_operand): New predicate.
13186 * config/mips/mips.md (multimem): New type.
13187 (length): Differentiate between 17-bit and 18-bit branch offsets.
13188 (MOVEP1, MOVEP2): New mode iterator.
13189 (mov_<load>l): Use ZC constraint.
13190 (mov_<load>r): Likewise.
13191 (mov_<store>l): Likewise.
13192 (mov_<store>r): Likewise.
13193 (*branch_equality<mode>_inverted): Add microMIPS support.
13194 (*branch_equality<mode>): Likewise.
13195 (*jump_absolute): Likewise.
13196 (indirect_jump_<mode>): Likewise.
13197 (tablejump_<mode>): Likewise.
13198 (<optab>_internal): Likewise.
13199 (sibcall_internal): Likewise.
13200 (sibcall_value_internal): Likewise.
13201 (prefetch): Use constraint ZD.
13202 * config/mips/mips.opt (minterlink-compressed): New option.
13203 (minterlink-mips16): Now an alias for minterlink-compressed.
13204 (mmicromips): New option.
13205 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13206 (compare_and_swap_12): Likewise.
13207 (sync_add<mode>): Likewise.
13208 (sync_<optab>_12): Likewise.
13209 (sync_old_<optab>_12): Likewise.
13210 (sync_new_<optab>_12): Likewise.
13211 (sync_nand_12): Likewise.
13212 (sync_old_nand_12): Likewise.
13213 (sync_new_nand_12): Likewise.
13214 (sync_sub<mode>): Likewise.
13215 (sync_old_add<mode>): Likewise.
13216 (sync_old_sub<mode>): Likewise.
13217 (sync_new_add<mode>): Likewise.
13218 (sync_new_sub<mode>): Likewise.
13219 (sync_<optab><mode>): Likewise.
13220 (sync_old_<optab><mode>): Likewise.
13221 (sync_new_<optab><mode>): Likewise.
13222 (sync_nand<mode>): Likewise.
13223 (sync_old_nand<mode>): Likewise.
13224 (sync_new_nand<mode>): Likewise.
13225 (sync_lock_test_and_set<mode>): Likewise.
13226 (test_and_set_12): Likewise.
13227 (atomic_compare_and_swap<mode>): Likewise.
13228 (atomic_exchange<mode>_llsc): Likewise.
13229 (atomic_fetch_add<mode>_llsc): Likewise.
13230 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13231 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13232 (umips_save_restore_pattern_p): Likewise.
13233 (umips_load_store_pair_p): Likewise.
13234 (umips_output_load_store_pair): Likewise.
13235 (umips_movep_target_p): Likewise.
13236 (umips_12bit_offset_address_p): Likewise.
13237 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13238 (mips_base_mips16): Rename this...
13239 (mips_base_compression_flags): ...to this. Update all uses.
13240 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13241 (mips_mips16_decl_p): Delete.
13242 (mips_nomips16_decl_p): Delete.
13243 (mips_get_compress_on_flags): New function.
13244 (mips_get_compress_off_flags): New function.
13245 (mips_get_compress_mode): New function.
13246 (mips_get_compress_on_name): New function.
13247 (mips_get_compress_off_name): New function.
13248 (mips_insert_attributes): Support multiple compression types.
13249 (mips_merge_decl_attributes): Likewise.
13250 (umips_12bit_offset_address_p): New function.
13251 (mips_start_function_definition): Emit .set micromips directive.
13252 (mips_call_may_need_jalx_p): New function.
13253 (mips_function_ok_for_sibcall): Add microMIPS support.
13254 (mips_print_operand_punctuation): Support short delay slots and
13256 (umips_swm_mask, umips_swm_encoding): New.
13257 (umips_build_save_restore): New function.
13258 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13259 (was_mips16_p): Remove.
13260 (old_compression_mode): New.
13261 (mips_set_compression_mode): New function.
13262 (mips_set_current_function): Add microMIPS support.
13263 (mips_option_override): Likewise.
13264 (umips_save_restore_pattern_p): New function.
13265 (umips_output_save_restore): New function.
13266 (umips_load_store_pair_p_1): New function.
13267 (umips_load_store_pair_p): New function.
13268 (umips_output_load_store_pair_1): New function.
13269 (umips_output_load_store_pair): New function.
13270 (umips_movep_target_p) New function.
13271 (mips_prepare_pch_save): Add microMIPS support.
13272 * config/mips/mips.h (TARGET_COMPRESSION): New.
13273 (TARGET_CPU_CPP_BUILTINS): Update macro
13274 to use new compression flags and to support microMIPS.
13275 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13276 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13277 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13278 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13279 (ASM_SPEC): Support mmicromips and mno-micromips.
13280 (M16STORE_REG_P): New macro.
13281 (MIPS_CALL): Support TARGET_MICROMIPS.
13282 (MICROMIPS_J): New macro.
13283 (mips_base_mips16): Rename this...
13284 (mips_base_compression_flags): ...to this.
13285 (UMIPS_12BIT_OFFSET_P): New macro.
13286 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13287 (MULTILIB_DIRNAMES): Likewise.
13288 2013-03-20 Richard Biener <rguenther@suse.de>
13290 PR tree-optimization/56661
13291 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13292 the result does not have to be distinct.
13294 2013-03-20 Richard Biener <rguenther@suse.de>
13296 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13299 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13300 Steven Bosscher <steven@gcc.gnu.org>
13302 PR rtl-optimization/56605
13303 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13305 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13308 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13309 that require movd instead of movq.
13311 2013-03-20 Richard Biener <rguenther@suse.de>
13313 * tree-ssa-structalias.c (struct variable_info): Add pointer
13314 to the first field of an aggregate with sub-vars. Make
13315 this and the pointer to the next subfield its ID.
13316 (vi_next): New function.
13317 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13318 storedanything_id, integer_id): Increment by one.
13319 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13320 get_call_clobber_vi): Adjust.
13321 (solution_set_expand): Simplify and speedup.
13322 (solution_set_add): Inline into ...
13323 (set_union_with_increment): ... this. Adjust accordingly.
13324 (do_sd_constraint): Likewise.
13325 (do_ds_constraint): Likewise.
13326 (do_complex_constraint): Simplify.
13327 (build_pred_graph): Adjust.
13328 (solve_graph): Likewise. Simplify and speedup.
13329 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13330 get_constraint_for_component_ref, get_constraint_for_1,
13331 first_vi_for_offset, first_or_preceding_vi_for_offset,
13332 create_function_info_for, create_variable_info_for_1,
13333 create_variable_info_for, intra_create_variable_infos): Adjust.
13334 (init_base_vars): Push NULL for ID zero.
13335 (compute_points_to_sets): Adjust.
13337 2013-03-20 Richard Biener <rguenther@suse.de>
13339 * cfgloop.c (verify_loop_structure): Streamline and avoid
13340 ICEing on corrupt loop tree.
13341 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13344 2013-03-20 Richard Biener <rguenther@suse.de>
13346 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13347 check whether an SSA update is needed.
13349 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13351 * config/mips/constraints.md (T): Rename to...
13355 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13356 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13358 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13360 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13361 (*subsi3_carryin_uxtw): Likewise.
13363 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13365 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13366 (*rorsi3_insn_uxtw): Likewise.
13368 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13370 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
13371 (*extrsi5_insn_uxtw): Likewise.
13373 2013-03-19 Richard Biener <rguenther@suse.de>
13375 PR tree-optimization/56273
13376 * passes.c (init_optimization_passes): Move second VRP after DOM.
13378 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
13380 * config/i386/i386.md (*movti_internal): Merge from
13381 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
13382 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
13383 nox64 isa attributes.
13385 2013-03-18 Richard Biener <rguenther@suse.de>
13387 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
13389 (merge_node_constraints): Likewise.
13390 (build_succ_graph): Likewise.
13391 (valid_graph_edge): Inline into single caller.
13392 (unify_nodes): Likewise. Use bitmap_set_bit return value
13394 (scc_visit): Fix formatting and variable use.
13395 (do_sd_constraint): Use gcc_checking_assert.
13396 (do_ds_constraint): Likewise.
13397 (do_complex_constraint): Likewise.
13398 (condense_visit): Likewise. Cleanup.
13399 (dump_pred_graph): New function.
13400 (perform_var_substitution): Dump the pred-graph before
13401 variable substitution.
13402 (find_equivalent_node): Use gcc_checking_assert.
13403 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
13405 2013-03-18 Richard Biener <rguenther@suse.de>
13407 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13408 Remove cond_expr_stmt_list argument and do not gimplify the
13410 (vect_loop_versioning): Adjust.
13411 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
13412 Cleanup to use less temporaries.
13413 (vect_create_data_ref_ptr): Cleanup.
13415 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13417 PR tree-optimization/56635
13418 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
13419 require types_compatible_p types.
13421 2013-03-18 Nick Clifton <nickc@redhat.com>
13423 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
13424 spurious backslash.
13426 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
13427 Add missing line to comment describing function.
13429 2013-03-18 Richard Biener <rguenther@suse.de>
13431 PR tree-optimization/56210
13432 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
13433 Handle string / character search functions.
13434 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
13436 2013-03-18 Richard Biener <rguenther@suse.de>
13438 PR middle-end/56483
13439 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
13440 and implement properly.
13441 * gimple.h (gimple_cond_single_var_p): Remove.
13443 2013-03-18 Richard Biener <rguenther@suse.de>
13445 * tree-data-ref.h (find_data_references_in_loop): Declare.
13446 * tree-data-ref.c (get_references_in_stmt): Use a stack
13447 vector pre-allocated in the callers.
13448 (find_data_references_in_stmt): Adjust.
13449 (graphite_find_data_references_in_stmt): Likewise.
13450 (create_rdg_vertices): Likewise.
13451 (find_data_references_in_loop): Export.
13452 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
13453 Compute dependences here...
13454 (vect_analyze_data_refs): ...not here. When we encounter
13455 a non-vectorizable data reference in basic-block vectorization
13456 truncate the data reference vector. Do not bother to
13457 fixup data-dependence information for gather loads.
13458 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
13459 of data references, as reported.
13461 2013-03-18 Richard Biener <rguenther@suse.de>
13463 PR tree-optimization/3713
13464 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
13465 has_constants and expr.
13466 (stmt_has_constants): Properly valueize SSA names when deciding
13467 whether the stmt has constants.
13469 2013-03-18 Richard Biener <rguenther@suse.de>
13471 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
13472 whole function when there is nothing to do.
13473 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
13474 * tree-vectorizer.c (vectorize_loops): Update virtual and
13475 loop-closed SSA once.
13476 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
13478 2013-03-18 Richard Biener <rguenther@suse.de>
13480 PR middle-end/56113
13481 * domwalk.c (bb_postorder): New global static.
13482 (cmp_bb_postorder): New function.
13483 (walk_dominator_tree): Replace scheme imposing an order for
13484 visiting dominator sons by one sorting them at the time they
13485 are pushed on the stack.
13487 2013-03-18 Richard Biener <rguenther@suse.de>
13489 PR tree-optimization/39326
13490 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
13491 (struct mem_ref): Replace mem member with ao_ref typed member.
13492 (MEM_ANALYZABLE): Adjust.
13493 (memref_eq): Likewise.
13494 (mem_ref_alloc): Likewise.
13495 (gather_mem_refs_stmt): Likewise.
13496 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
13497 (execute_sm_if_changed_flag_set): Adjust.
13498 (execute_sm): Likewise.
13499 (ref_always_accessed_p): Likewise.
13500 (refs_independent_p): Likewise.
13501 (can_sm_ref_p): Likewise.
13503 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13506 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
13507 return 1 even for !unsignedp.
13509 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13511 * config/i386/i386.md (isa): Add x64 and nox64.
13512 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
13513 (*pushtf): Enable *roF alternative for x64 isa only.
13514 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
13515 mode attribute of integer alternatives to DImode for TARGET_64BIT.
13516 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
13517 (*movtf_internal): Merge from *movtf_internal_rex64 and
13518 *movtf_internal_sse. Use x64 and nox64 isa attributes.
13519 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
13520 nox64 isa attributes.
13521 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
13522 nox64 isa attributes.
13523 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
13525 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13527 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13529 2013-03-17 Jakub Jelinek <jakub@redhat.com>
13532 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
13533 class names. Remove trailing comma after "ALL_REGS".
13535 2013-03-16 Jan Hubicka <jh@suse.cz>
13537 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
13538 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
13539 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
13540 of cgraph_get_create_node.
13541 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
13543 2013-03-16 Jason Merrill <jason@redhat.com>
13546 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
13547 with DW_AT_default_value.
13549 2013-03-16 Jakub Jelinek <jakub@redhat.com>
13551 * BASE-VER: Set to 4.9.0.
13553 2013-03-14 Andi Kleen <ak@linux.intel.com>
13556 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
13557 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
13558 Document _x* TSX intrinsics.
13560 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
13561 David Holsgrove <david.holsgrove@xilinx.com>
13563 * configure.ac: Add MicroBlaze TLS support detection.
13564 * configure: Regenerate.
13565 * config/microblaze/microblaze-protos.h
13566 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
13567 symbol_mentioned_p, label_mentioned_p): Add prototypes.
13568 * config/microblaze/microblaze.c (microblaze_address_type): Add
13569 ADDRESS_TLS and tls_reloc address types.
13570 (microblaze_address_info): Add tls_reloc.
13571 (TARGET_HAVE_TLS): Define.
13572 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
13573 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
13574 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
13575 load_tls_operand, microblaze_call_tls_get_addr,
13576 microblaze_legitimize_tls_address): New functions.
13577 (microblaze_classify_unspec): Handle UNSPEC_TLS.
13578 (get_base_reg): Use microblaze_tls_symbol_p.
13579 (microblaze_classify_address): Handle TLS.
13580 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
13581 label_mentioned_p and microblaze_tls_referenced_p.
13582 (microblaze_legitimize_address): Handle TLS.
13583 (microblaze_address_insns): Handle ADDRESS_TLS.
13584 (pic_address_needs_scratch): Handle TLS.
13585 (print_operand_address): Handle TLS.
13586 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
13587 (microblaze_expand_move): Handle TLS.
13588 (microblaze_legitimate_constant_p): Check
13589 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
13590 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
13591 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
13592 (PIC_OFFSET_TABLE_REGNUM): Set.
13593 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
13594 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
13595 (addsi3, movsi_internal2, movdf_internal): Update constraints
13596 * config/microblaze/predicates.md (arith_plus_operand): Define
13597 (move_operand): Redefine as move_src_operand,
13598 check microblaze_tls_referenced_p.
13600 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13602 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
13603 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
13605 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13607 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
13610 2013-03-14 Jakub Jelinek <jakub@redhat.com>
13612 PR tree-optimization/53265
13613 * common.opt (Waggressive-loop-optimizations): New option.
13614 * tree-ssa-loop-niter.c: Include tree-pass.h.
13615 (do_warn_aggressive_loop_optimizations): New function.
13616 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
13617 if number_of_latch_executions returned constant.
13618 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
13619 early. If number_of_latch_executions returned constant, set
13620 nb_iterations_upper_bound back to it.
13621 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
13623 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
13624 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
13626 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
13627 (MULTILIB_OSDIRNAMES): Set.
13628 * genmultilib: If defaultosdirname doesn't start with :: , set
13629 defaultosdirname2 instead, clear it and emit two . multilib_raw
13630 entries instead of just one.
13632 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
13634 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13635 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
13636 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13637 (SUBTARGET_OVERRIDE_OPTIONS): New.
13639 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
13642 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
13643 (musermode): Convert to Var(TARGET_USERMODE).
13644 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
13645 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
13646 * config/sh/sh.c (sh_option_override): Use
13647 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
13648 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
13650 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
13652 (udivsi3_i4_single, divsi3_i4_single): Use
13653 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
13655 2013-03-13 Dave Korn <dave.korn.cygwin@....>
13657 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
13660 2013-03-13 Richard Biener <rguenther@suse.de>
13662 PR tree-optimization/56608
13663 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
13664 calls when vectorizing basic-blocks.
13666 2013-03-13 Jakub Jelinek <jakub@redhat.com>
13669 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
13672 2013-03-12 Jakub Jelinek <jakub@redhat.com>
13674 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
13676 2013-03-11 Jan Hubicka <jh@suse.cz>
13679 * lto-streamer-out.c (output_symbol_p): Skip references from
13680 constructors of external variables.
13682 2013-03-11 Jan Hubicka <jh@suse.cz>
13684 PR middle-end/56571
13685 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
13687 * emit-rtl.c (verify_rtx_sharing): Likewise.
13688 (copy_insn_1): Likewise.
13689 * rtl.c (copy_rtx): Likewise.
13691 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
13694 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
13695 output_operand_lossage message.
13697 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
13700 * arm.c (shift_op): Validate RTL pattern on the fly.
13701 (arm_print_operand, case 'S'): Don't use shift_operator to validate
13704 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13707 * config/pa/pa.md (call_value): Check for calls to powf and direct to
13708 new call patterns that clobber %fr12.
13709 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
13710 split and postreload patterns.
13711 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
13712 registers %fr12 and %fr12R as call used.
13714 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
13716 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
13717 (canon_address, record_store, replace_read, check_mem_read_rtx,
13718 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
13719 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
13720 rest_of_handle_dse): Likewise.
13722 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
13724 PR middle-end/56524
13725 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
13727 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
13728 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
13729 (save_optabs_if_changed): Replace with...
13730 (init_tree_optimization_optabs): ...this.
13731 * optabs.c (save_optabs_if_changed): Rename to...
13732 (init_tree_optimization_optabs): ...this. Take the optimization node
13733 as argument. Do nothing if the base optabs are already correct.
13734 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
13735 to recompute optabs.
13736 * function.h (function): Remove optabs field.
13737 * function.c (invoke_set_current_function_hook): Call
13738 init_tree_optimization_optabs. Use the result to initialize
13741 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
13743 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
13744 if GTMA_HAS_NO_INSTRUMENTATION.
13745 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
13746 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
13747 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
13748 * gimple-pretty-print.c (dump_gimple_transaction): Handle
13749 GTMA_HAS_NO_INSTRUMENTATION.
13751 2013-03-08 Jakub Jelinek <jakub@redhat.com>
13753 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
13756 2013-03-08 Marek Polacek <polacek@redhat.com>
13757 Jakub Jelinek <jakub@redhat.com>
13759 PR tree-optimization/56478
13760 * predict.c (is_comparison_with_loop_invariant_p): Change the
13761 type of loop_step to tree.
13762 (predict_loops): Adjust.
13763 (predict_iv_comparison): Perform the computations on double_ints.
13765 2013-03-08 Richard Biener <rguenther@suse.de>
13767 PR tree-optimization/56570
13768 * tree-cfg.c (verify_expr_location_1): Verify locations for
13770 * tree-sra.c (create_access_replacement): Strip locations
13771 from DECL_DEBUG_EXPRs.
13773 2013-03-08 Richard Biener <rguenther@suse.de>
13775 * tree-inline.c (expand_call_inline): Do not associate
13776 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
13777 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
13779 2013-03-08 Richard Biener <rguenther@suse.de>
13781 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
13782 or block changes with -Og. Fix for location / block encoding
13783 changes and PHI arguments with locations.
13785 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
13787 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
13789 (struct output_info): Likewise.
13790 (register_overhead): Remove bad gcc_assert.
13791 (bitmap_find_bit): If there is only a single bitmap element, do not
13792 count a miss as a search.
13793 (print_statistics): Update for counter type changes.
13794 (dump_bitmap_statistics): Likewise. Print headers such that they
13795 are properly lined up with the printed counters.
13797 2013-03-07 Jakub Jelinek <jakub@redhat.com>
13799 PR tree-optimization/56559
13800 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
13801 check that it has only a single use.
13803 2013-03-07 Richard Biener <rguenther@suse.de>
13805 * doc/invoke.texi (fwhole-program): Discourage use in combination
13808 2013-03-06 Jakub Jelinek <jakub@redhat.com>
13810 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
13812 PR tree-optimization/56539
13813 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
13814 instead of GSI_CONTINUE_LINKING as last argument to
13815 force_gimple_operand_gsi. Adjust function comment.
13817 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
13820 PR middle-end/56548
13821 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
13822 promoted mode, convert the result back to the original mode.
13824 2013-03-06 Richard Biener <rguenther@suse.de>
13826 PR middle-end/56294
13827 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
13828 (insert_updated_phi_nodes_compare_uids): New function.
13829 (update_ssa): Sort symbols_to_rename after UID before
13830 traversing it to insert PHI nodes.
13832 2013-03-06 Richard Biener <rguenther@suse.de>
13834 PR middle-end/50494
13835 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
13836 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
13839 2013-02-13 Richard Biener <rguenther@suse.de>
13842 * varasm.c (output_constant_def_1): Get the decl representing
13843 the constant as argument.
13844 (output_constant_def): Wrap output_constant_def_1.
13845 (make_decl_rtl): Use output_constant_def_1 with the decl
13846 representing the constant.
13847 (build_constant_desc): Optionally re-use a decl already
13848 representing the constant.
13849 (tree_output_constant_def): Adjust.
13851 2013-03-06 Joey Ye <joey.ye@arm.com>
13854 * gcc.c (convert_white_space): New function.
13855 (main): Handles white space in function name.
13857 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
13860 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
13861 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
13862 to SH_DIV_CALL_TABLE for TARGET_SH2.
13863 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
13865 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
13866 call-table options.
13868 2013-03-05 Sterling Augustine <saugustine@google.com>
13869 Cary Coutant <ccoutant@google.com>
13872 * dwarf2out.c (resolve_addr): Don't call
13873 remove_loc_list_addr_table_entries a second time for the same
13876 2013-03-05 Jakub Jelinek <jakub@redhat.com>
13879 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
13880 (avoid_complex_debug_insns): New function.
13881 (expand_debug_locations): Call it.
13883 PR rtl-optimization/56484
13884 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
13885 lifetimes of hard registers on small register class machines.
13887 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
13889 * config/microblaze/microblaze-protos.h: Rename
13890 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
13891 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
13893 (microblaze_fast_interrupt_function_p): New function.
13894 (microblaze_is_interrupt_handler): Rename to
13895 microblaze_is_interrupt_variant and add fast_interrupt check.
13896 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
13897 (save_restore_insns): Likewise.
13898 (compute_frame_size): Likewise.
13899 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
13900 (microblaze_globalize_label): Likewise.
13901 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
13902 * config/microblaze/microblaze.md: Use wrapper
13903 microblaze_is_interrupt_variant.
13905 2013-03-05 Kai Tietz <ktietz@redhat.com>
13907 * sdbout.c (sdbout_one_type): Switch to current function's section
13908 supporting cold/hot.
13910 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
13912 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
13915 2013-03-05 Jakub Jelinek <jakub@redhat.com>
13917 PR middle-end/56461
13918 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
13919 if VALGRIND_GET_VBITS is defined, temporarily make object
13920 memory all defined, and restore previous valgrind addressability
13921 and definability afterwards. Free this_object at the end.
13923 PR middle-end/56461
13924 * lra.c (lra): Call lra_clear_live_ranges if live_p,
13925 right before calling lra_create_live_ranges, also call it
13926 when clearing live_p. Only call lra_clear_live_ranges
13927 at the end if live_p.
13929 PR middle-end/56461
13930 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
13932 2013-03-05 Richard Biener <rguenther@suse.de>
13934 PR tree-optimization/56521
13935 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
13938 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
13941 * except.h (remove_unreachable_eh_regions): New prototype.
13942 * except.c (remove_eh_handler_splicer): New function, split out
13943 of remove_eh_handler.
13944 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
13945 warning about running it on many EH regions one at a time.
13946 (remove_unreachable_eh_regions_worker): New function, walk the
13947 EH tree in depth-first order and remove non-marked regions.
13948 (remove_unreachable_eh_regions): New function.
13949 * tree-eh.c (mark_reachable_handlers): New function, split out
13950 from remove_unreachable_handlers.
13951 (remove_unreachable_handlers): Use mark_reachable_handlers and
13952 remove_unreachable_eh_regions.
13953 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
13954 and remove_unreachable_eh_regions.
13956 2013-03-05 Richard Biener <rguenther@suse.de>
13958 PR middle-end/56525
13959 * loop-init.c (fix_loop_structure): Remove loops in two stages,
13960 not freeing them until the end.
13962 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13964 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
13966 2013-03-05 Richard Biener <rguenther@suse.de>
13968 PR tree-optimization/56270
13969 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
13970 of loads after scheduling an SLP instance.
13972 2013-03-05 Jakub Jelinek <jakub@redhat.com>
13974 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
13976 (check_gcc_parallelize): Run guality.exp as a separate job from
13977 vect.exp with unsorted.exp and $(dg_target_exps) separately from
13978 struct-layout-1.exp with stackalign.exp.
13980 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
13982 PR middle-end/56461
13983 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
13984 load_index sbitmap even if some bit in it isn't set.
13986 PR middle-end/56461
13987 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
13988 (discover_iteration_bound_by_body_walk): Change queues to
13989 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
13990 spelling in comment. Call safe_push on queues[bound_index] directly.
13991 Release queues[queue_index] in every iteration unconditionally.
13992 Release bounds vector.
13994 PR middle-end/56461
13995 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
13996 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
13997 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
14000 2013-03-05 Richard Biener <rguenther@suse.de>
14003 * tree-inline.c (remap_blocks_to_null): New function.
14004 (expand_call_inline): When expanding a call stmt without
14005 an associated block inline remap all callee blocks to NULL.
14007 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14009 PR rtl-optimization/56494
14010 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
14011 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
14012 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
14014 PR middle-end/56461
14015 * sel-sched-ir.c (free_sched_pools): Release
14016 succs_info_pool.stack[succs_info_pool.max_top] vectors too
14017 if succs_info_pool.max_top isn't -1.
14020 * opts.c (opts_obstack, opts_concat): Moved to...
14021 * opts-common.c (opts_obstack, opts_concat): ... here.
14023 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14025 PR middle-end/56461
14026 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14028 2013-03-04 Martin Jambor <mjambor@suse.cz>
14030 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14031 all appropriate places.
14033 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14035 PR tree-optimization/56424
14036 * ipa-split.c (split_function): Do not set the RSO flag if result is
14037 not by reference and its type is a register type.
14039 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
14041 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
14042 (microblaze_legitimate_pic_operand): Likewise
14043 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
14044 new function microblaze_legitimate_pic_operand
14045 * config/microblaze/microblaze-protos.h
14046 (microblaze_legitimate_pic_operand): Declare.
14048 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14050 * config/microblaze/predicates.md (call_insn_simple_operand):
14051 New predicate for supported rtx code types.
14052 * config/microblaze/microblaze.md (call_internal1): Use
14053 call_insn_simple_operand predicate.
14055 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14057 PR middle-end/56461
14058 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14059 partitions.ordered_remove.
14061 PR middle-end/56461
14062 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14063 vec_oprnds0.create (1) for modifier == NONE.
14065 PR middle-end/56461
14066 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14067 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14068 vec_oprnds1 right before pushing anything to it for
14071 PR middle-end/56461
14072 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14073 set nbbs to 0 instead of having separate code path.
14074 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14075 instead of false as last argument if returning NULL.
14077 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14079 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14080 the attribute is now called "target" instead of "option".
14081 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14082 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14083 attribute/pragma name for TARGET_OPTION_VALID_P and
14084 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14085 * doc/tm.texi: Regenerated.
14087 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
14089 * config/microblaze/microblaze.c:
14090 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
14091 * config/microblaze/microblaze.h: Add -mxl-reorder to
14093 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14094 instructions emitted if TARGET_REORDER.
14095 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14096 or 0 for -m/-mno case, but initialises as 2 to detect default use case
14099 2013-03-01 Xinliang David Li <davidxl@google.com>
14101 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14104 2013-03-01 Jakub Jelinek <jakub@redhat.com>
14106 PR middle-end/56461
14107 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14108 vector even when returning true. Fix up function comment formatting.
14110 PR middle-end/56461
14111 * ira-build.c (ira_loop_nodes_count): New variable.
14112 (create_loop_tree_nodes): Initialize it.
14113 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14115 PR middle-end/56461
14116 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14117 method on dr_chain and result_chain.
14118 * tree-vect-stmts.c (vectorizable_store): Only call
14119 result_chain.create if j == 0.
14121 PR middle-end/56461
14122 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14123 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14124 before overwriting it.
14126 2013-03-01 Tobias Burnus <burnus@net-b.de>
14128 * doc/extended.texi (C Extensions): Change order in @menu
14130 (Other MIPS Built-in Functions): Move last MIPS entry before
14131 "picoChip Built-in Functions".
14132 (SH Built-in Functions): Move after RX Built-in Functions.
14133 * doc/gcc.texi (Introduction): Change order in @menu
14135 * doc/md.texi (Constraints): Ditto.
14136 * gty.texi (Type Information): Ditto.
14137 (User-provided marking routines for template types): Make
14139 * doc/invoke.texi (AArch64 Options): Move before
14140 "Adapteva Epiphany Options".
14142 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14143 Jakub Jelinek <jakub@redhat.com>
14146 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14147 no_address_safety_analysis attribute.
14148 * doc/extend.texi (no_address_safety_attribute): Rename to
14149 no_sanitize_address attribute, mention no_address_safety_analysis
14150 attribute as deprecated alias.
14152 2013-02-28 Jakub Jelinek <jakub@redhat.com>
14154 PR middle-end/56461
14155 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14156 type to vec<vec<tree> > *.
14157 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14158 to be vec<tree> instead of vec<tree> *, set vec_defs
14159 to vNULL and call vec_defs.create (number_of_vects), adjust other
14161 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14162 vectorizable_condition): Adjust vect_get_slp_defs callers.
14164 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14166 * config/aarch64/aarch64.c
14167 (aarch64_float_const_representable): Remove unused variable.
14169 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14171 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14173 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14175 * config/aarch64/aarch64-builtins.c
14176 (aarch64_init_simd_builtins): Make static.
14178 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14180 * config/aarch64/aarch64.c
14181 (aarch64_simd_make_constant): Make static.
14183 2013-02-28 Martin Jambor <mjambor@suse.cz>
14185 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14186 with no initialization to the RHS of debug statements.
14188 2013-02-28 Martin Jambor <mjambor@suse.cz>
14190 PR tree-optimization/56294
14191 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14193 (get_access_replacement): Do not call create_access_replacement.
14194 Assert a replacement exists.
14195 (get_repl_default_def_ssa_name): Create the replacement declaration
14198 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14200 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14201 final_end_function.
14203 2013-02-28 Marek Polacek <polacek@redhat.com>
14205 PR rtl-optimization/56466
14206 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14207 if we're changing a loop.
14208 (peel_loops_completely): Likewise.
14210 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14213 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14215 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14218 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14219 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14220 INTX_FTYPE_FX, FX_FTYPE_INTX.
14221 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14223 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14225 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14226 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14227 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14228 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14229 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14230 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14231 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14232 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14233 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14234 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14235 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14236 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14237 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14238 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14239 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14240 (avrxmega6): Increase max flash segments from 5 to 6.
14241 * config/avr/t-multilib: Regenerate.
14242 * config/avr/avr-tables.opt: Regenerate.
14243 * doc/avr-mmcu.texi: Regenerate.
14245 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14247 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14248 (avr_device_to_arch): Rename to avr_device_to_ld.
14249 (avr_device_to_as): New prototype.
14250 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14251 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14252 * config/avr/driver-avr.c (avr_device_to_as): New.
14253 (avr_device_to_arch): Rename to avr_device_to_ld.
14255 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14257 PR middle-end/56461
14258 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14259 method on dr_chain and result_chain.
14261 PR middle-end/56461
14262 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14263 pointer_set_destroy on not_executed_last_iteration.
14265 PR middle-end/56461
14266 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
14268 PR middle-end/56461
14269 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14270 FOR_EACH_DEFINED_FUNCTION when freeing state.
14272 PR middle-end/56461
14273 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14275 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14278 PR middle-end/56461
14279 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14280 known_aggs[i].items and release known_aggs vector.
14282 PR middle-end/56461
14283 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14285 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14287 * config/microblaze/microblaze.c (microblaze_emit_compare):
14288 Use xor for EQ/NE comparisions.
14289 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14290 (cbranchsf4): Adjust operator to comparison_operator.
14292 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14294 PR middle-end/56461
14295 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14297 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14298 vec_safe_push, always update *slot.
14299 (redirect_edge_var_map_clear): Use vec_free.
14300 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14301 (free_var_map_entry): Use vec_free.
14302 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14303 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14305 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
14307 PR middle-end/45472
14308 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14309 when the may_trap_p bit of the exprs being merged differs.
14310 Reorder tests for speculativeness in the logical and operator.
14312 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14314 * incpath.c (add_standard_paths): Use reconcat instead of concat
14315 where appropriate and avoid leaking memory.
14317 * opts.h: Include obstack.h.
14318 (opts_concat): New prototype.
14319 (opts_obstack): New declaration.
14320 * opts.c (opts_concat): New function.
14321 (opts_obstack): New variable.
14322 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14323 (finish_options): Use opts_concat instead of concat
14324 and XOBNEWVEC instead of XNEWVEC.
14325 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14326 generate_option): Likewise.
14327 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14328 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14331 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14332 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14334 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14336 PR middle-end/56461
14337 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14339 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14341 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14342 (arm_block_move_unaligned_straight): Likewise.
14343 (arm_adjust_block_mem): Likewise.
14345 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14348 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14349 temp, cond and label.
14350 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14353 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14354 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14357 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14358 of prologue/epilogue functions.
14361 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14362 Remove unused variable cfa_offset.
14363 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14366 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14367 type promotion to unsigned.
14370 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
14371 for HOST_WIDE_INT of 32 bit / same size as int.
14372 (arm_block_move_unaligned_straight): Likewise.
14373 (arm_adjust_block_mem): Likewise.
14376 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
14379 2013-02-26 Marek Polacek <polacek@redhat.com>
14381 PR tree-optimization/56426
14382 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
14384 2013-02-26 Richard Biener <rguenther@suse.de>
14387 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
14388 unused variable loops.
14390 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14392 PR tree-optimization/56448
14393 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
14394 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
14395 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
14396 later operands of the references, or even first operand for
14397 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
14399 PR tree-optimization/56443
14400 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
14401 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
14402 to type_for_mode langhook.
14404 2013-02-25 Matt Turner <mattst88@gmail.com>
14406 * doc/invoke.texi: Document r4700.
14408 2013-02-25 Richard Biener <rguenther@suse.de>
14410 PR tree-optimization/56175
14411 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
14413 (simplify_bitwise_binary): ... here. Also guard the conversion
14414 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
14416 2013-02-25 Catherine Moore <clm@codesourcery.com>
14419 2013-02-24 Catherine Moore <clm@codesourcery.com>
14420 Maciej W. Rozycki <macro@codesourcery.com>
14421 Tom de Vries <tom@codesourcery.com>
14422 Nathan Sidwell <nathan@codesourcery.com>
14423 Iain Sandoe <iain@codesourcery.com>
14424 Nathan Froyd <froydnj@codesourcery.com>
14425 Chao-ying Fu <fu@mips.com>
14427 * doc/extend.texi: (micromips, nomicromips, nocompression):
14428 Document new function attributes.
14429 * doc/invoke.texi (minterlink-compressed, mmicromips,
14430 m14k, m14ke, m14kec): Document new options.
14431 (minterlink-mips16): Update documentation.
14432 * doc/md.texi (ZC, ZD): Document new constraints.
14433 * configure.ac (gcc_cv_as_micromips): Check if linker
14434 supports the .set micromips directive.
14435 * configure: Regenerate.
14436 * config.in: Regenerate.
14437 * config/mips/mips-tables.opt: Regenerate.
14438 * config/mips/micromips.md: New file.
14439 * constraints.md (ZC, AD): New constraints.
14440 * config/mips/predicates.md (movep_src_register): New predicate.
14441 (movep_src_operand): New predicate.
14442 (non_volatile_mem_operand): New predicate.
14443 * config/mips/mips.md (multimem): New type.
14444 (length): Differentiate between 17-bit and 18-bit branch offsets.
14445 (MOVEP1, MOVEP2): New mode iterator.
14446 (mov_<load>l): Use ZC constraint.
14447 (mov_<load>r): Likewise.
14448 (mov_<store>l): Likewise.
14449 (mov_<store>r): Likewise.
14450 (*branch_equality<mode>_inverted): Add microMIPS support.
14451 (*branch_equality<mode>): Likewise.
14452 (*jump_absolute): Likewise.
14453 (indirect_jump_<mode>): Likewise.
14454 (tablejump_<mode>): Likewise.
14455 (<optab>_internal): Likewise.
14456 (sibcall_internal): Likewise.
14457 (sibcall_value_internal): Likewise.
14458 (prefetch): Use constraint ZD.
14459 * config/mips/mips.opt (minterlink-compressed): New option.
14460 (minterlink-mips16): Now an alias for minterlink-compressed.
14461 (mmicromips): New option.
14462 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14463 (compare_and_swap_12): Likewise.
14464 (sync_add<mode>): Likewise.
14465 (sync_<optab>_12): Likewise.
14466 (sync_old_<optab>_12): Likewise.
14467 (sync_new_<optab>_12): Likewise.
14468 (sync_nand_12): Likewise.
14469 (sync_old_nand_12): Likewise.
14470 (sync_new_nand_12): Likewise.
14471 (sync_sub<mode>): Likewise.
14472 (sync_old_add<mode>): Likewise.
14473 (sync_old_sub<mode>): Likewise.
14474 (sync_new_add<mode>): Likewise.
14475 (sync_new_sub<mode>): Likewise.
14476 (sync_<optab><mode>): Likewise.
14477 (sync_old_<optab><mode>): Likewise.
14478 (sync_new_<optab><mode>): Likewise.
14479 (sync_nand<mode>): Likewise.
14480 (sync_old_nand<mode>): Likewise.
14481 (sync_new_nand<mode>): Likewise.
14482 (sync_lock_test_and_set<mode>): Likewise.
14483 (test_and_set_12): Likewise.
14484 (atomic_compare_and_swap<mode>): Likewise.
14485 (atomic_exchange<mode>_llsc): Likewise.
14486 (atomic_fetch_add<mode>_llsc): Likewise.
14487 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14488 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14489 (umips_save_restore_pattern_p): Likewise.
14490 (umips_load_store_pair_p): Likewise.
14491 (umips_output_load_store_pair): Likewise.
14492 (umips_movep_target_p): Likewise.
14493 (umips_12bit_offset_address_p): Likewise.
14494 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14495 (mips_base_mips16): Rename this...
14496 (mips_base_compression_flags): ...to this. Update all uses.
14497 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14498 (mips_mips16_decl_p): Delete.
14499 (mips_nomips16_decl_p): Delete.
14500 (mips_get_compress_on_flags): New function.
14501 (mips_get_compress_off_flags): New function.
14502 (mips_get_compress_mode): New function.
14503 (mips_get_compress_on_name): New function.
14504 (mips_get_compress_off_name): New function.
14505 (mips_insert_attributes): Support multiple compression types.
14506 (mips_merge_decl_attributes): Likewise.
14507 (umips_12bit_offset_address_p): New function.
14508 (mips_start_function_definition): Emit .set micromips directive.
14509 (mips_call_may_need_jalx_p): New function.
14510 (mips_function_ok_for_sibcall): Add microMIPS support.
14511 (mips_print_operand_punctuation): Support short delay slots and
14513 (umips_swm_mask, umips_swm_encoding): New.
14514 (umips_build_save_restore): New function.
14515 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14516 (was_mips16_p): Remove.
14517 (old_compression_mode): New.
14518 (mips_set_compression_mode): New function.
14519 (mips_set_current_function): Add microMIPS support.
14520 (mips_option_override): Likewise.
14521 (umips_save_restore_pattern_p): New function.
14522 (umips_output_save_restore): New function.
14523 (umips_load_store_pair_p_1): New function.
14524 (umips_load_store_pair_p): New function.
14525 (umips_output_load_store_pair_1): New function.
14526 (umips_output_load_store_pair): New function.
14527 (umips_movep_target_p) New function.
14528 (mips_prepare_pch_save): Add microMIPS support.
14529 * config/mips/mips.h (TARGET_COMPRESSION): New.
14530 (TARGET_CPU_CPP_BUILTINS): Update macro
14531 to use new compression flags and to support microMIPS.
14532 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14533 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14534 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14535 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14536 (ASM_SPEC): Support mmicromips and mno-micromips.
14537 (M16STORE_REG_P): New macro.
14538 (MIPS_CALL): Support TARGET_MICROMIPS.
14539 (MICROMIPS_J): New macro.
14540 (mips_base_mips16): Rename this...
14541 (mips_base_compression_flags): ...to this.
14542 (UMIPS_12BIT_OFFSET_P): New macro.
14543 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14544 (MULTILIB_DIRNAMES): Likewise.
14546 2013-02-25 Tom de Vries <tom@codesourcery.com>
14548 PR rtl-optimization/56131
14549 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
14550 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
14551 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
14553 2013-02-25 Tobias Burnus <burnus@net-b.de>
14555 * doc/invoke.texi (-fsanitize=): Move from optimization
14556 to debugging options.
14558 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14560 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
14562 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14563 Alexander Monakov <amonakov@ispras.ru>
14565 PR middle-end/56077
14566 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
14567 flush pending lists also on non-jumps. Adjust comment.
14569 2013-02-24 Catherine Moore <clm@codesourcery.com>
14570 Maciej W. Rozycki <macro@codesourcery.com>
14571 Tom de Vries <tom@codesourcery.com>
14572 Nathan Sidwell <nathan@codesourcery.com>
14573 Iain Sandoe <iain@codesourcery.com>
14574 Nathan Froyd <froydnj@codesourcery.com>
14575 Chao-ying Fu <fu@mips.com>
14577 * doc/extend.texi: (micromips, nomicromips, nocompression):
14578 Document new function attributes.
14579 * doc/invoke.texi (minterlink-compressed, mmicromips,
14580 m14k, m14ke, m14kec): Document new options.
14581 (minterlink-mips16): Update documentation.
14582 * doc/md.texi (ZC, ZD): Document new constraints.
14583 * configure.ac (gcc_cv_as_micromips): Check if linker
14584 supports the .set micromips directive.
14585 * configure: Regenerate.
14586 * config.in: Regenerate.
14587 * config/mips/mips-tables.opt: Regenerate.
14588 * config/mips/micromips.md: New file.
14589 * constraints.md (ZC, AD): New constraints.
14590 * config/mips/predicates.md (movep_src_register): New predicate.
14591 (movep_src_operand): New predicate.
14592 (non_volatile_mem_operand): New predicate.
14593 * config/mips/mips.md (multimem): New type.
14594 (length): Differentiate between 17-bit and 18-bit branch offsets.
14595 (MOVEP1, MOVEP2): New mode iterator.
14596 (mov_<load>l): Use ZC constraint.
14597 (mov_<load>r): Likewise.
14598 (mov_<store>l): Likewise.
14599 (mov_<store>r): Likewise.
14600 (*branch_equality<mode>_inverted): Add microMIPS support.
14601 (*branch_equality<mode>): Likewise.
14602 (*jump_absolute): Likewise.
14603 (indirect_jump_<mode>): Likewise.
14604 (tablejump_<mode>): Likewise.
14605 (<optab>_internal): Likewise.
14606 (sibcall_internal): Likewise.
14607 (sibcall_value_internal): Likewise.
14608 (prefetch): Use constraint ZD.
14609 * config/mips/mips.opt (minterlink-compressed): New option.
14610 (minterlink-mips16): Now an alias for minterlink-compressed.
14611 (mmicromips): New option.
14612 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14613 (compare_and_swap_12): Likewise.
14614 (sync_add<mode>): Likewise.
14615 (sync_<optab>_12): Likewise.
14616 (sync_old_<optab>_12): Likewise.
14617 (sync_new_<optab>_12): Likewise.
14618 (sync_nand_12): Likewise.
14619 (sync_old_nand_12): Likewise.
14620 (sync_new_nand_12): Likewise.
14621 (sync_sub<mode>): Likewise.
14622 (sync_old_add<mode>): Likewise.
14623 (sync_old_sub<mode>): Likewise.
14624 (sync_new_add<mode>): Likewise.
14625 (sync_new_sub<mode>): Likewise.
14626 (sync_<optab><mode>): Likewise.
14627 (sync_old_<optab><mode>): Likewise.
14628 (sync_new_<optab><mode>): Likewise.
14629 (sync_nand<mode>): Likewise.
14630 (sync_old_nand<mode>): Likewise.
14631 (sync_new_nand<mode>): Likewise.
14632 (sync_lock_test_and_set<mode>): Likewise.
14633 (test_and_set_12): Likewise.
14634 (atomic_compare_and_swap<mode>): Likewise.
14635 (atomic_exchange<mode>_llsc): Likewise.
14636 (atomic_fetch_add<mode>_llsc): Likewise.
14637 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14638 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14639 (umips_save_restore_pattern_p): Likewise.
14640 (umips_load_store_pair_p): Likewise.
14641 (umips_output_load_store_pair): Likewise.
14642 (umips_movep_target_p): Likewise.
14643 (umips_12bit_offset_address_p): Likewise.
14644 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14645 (mips_base_mips16): Rename this...
14646 (mips_base_compression_flags): ...to this. Update all uses.
14647 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14648 (mips_mips16_decl_p): Delete.
14649 (mips_nomips16_decl_p): Delete.
14650 (mips_get_compress_on_flags): New function.
14651 (mips_get_compress_off_flags): New function.
14652 (mips_get_compress_mode): New function.
14653 (mips_get_compress_on_name): New function.
14654 (mips_get_compress_off_name): New function.
14655 (mips_insert_attributes): Support multiple compression types.
14656 (mips_merge_decl_attributes): Likewise.
14657 (umips_12bit_offset_address_p): New function.
14658 (mips_start_function_definition): Emit .set micromips directive.
14659 (mips_call_may_need_jalx_p): New function.
14660 (mips_function_ok_for_sibcall): Add microMIPS support.
14661 (mips_print_operand_punctuation): Support short delay slots and
14663 (umips_swm_mask, umips_swm_encoding): New.
14664 (umips_build_save_restore): New function.
14665 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14666 (was_mips16_p): Remove.
14667 (old_compression_mode): New.
14668 (mips_set_compression_mode): New function.
14669 (mips_set_current_function): Add microMIPS support.
14670 (mips_option_override): Likewise.
14671 (umips_save_restore_pattern_p): New function.
14672 (umips_output_save_restore): New function.
14673 (umips_load_store_pair_p_1): New function.
14674 (umips_load_store_pair_p): New function.
14675 (umips_output_load_store_pair_1): New function.
14676 (umips_output_load_store_pair): New function.
14677 (umips_movep_target_p) New function.
14678 (mips_prepare_pch_save): Add microMIPS support.
14679 * config/mips/mips.h (TARGET_COMPRESSION): New.
14680 (TARGET_CPU_CPP_BUILTINS): Update macro
14681 to use new compression flags and to support microMIPS.
14682 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14683 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14684 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14685 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14686 (ASM_SPEC): Support mmicromips and mno-micromips.
14687 (M16STORE_REG_P): New macro.
14688 (MIPS_CALL): Support TARGET_MICROMIPS.
14689 (MICROMIPS_J): New macro.
14690 (mips_base_mips16): Rename this...
14691 (mips_base_compression_flags): ...to this.
14692 (UMIPS_12BIT_OFFSET_P): New macro.
14693 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14694 (MULTILIB_DIRNAMES): Likewise.
14696 2013-02-24 Jakub Jelinek <jakub@redhat.com>
14699 * target-globals.c (save_target_globals): For init_reg_sets and
14700 target_reinit remporarily set this_fn_optabs to this_target_optabs.
14702 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
14704 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
14705 * config/aarch64/t-aarch64
14706 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
14708 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
14710 PR inline-asm/56148
14711 * lra-constraints.c (process_alt_operands): Reload operand
14712 conflicting with earlier clobber only if no more other conflicting
14715 2013-02-22 Jakub Jelinek <jakub@redhat.com>
14718 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
14719 if not linking a shared library.
14721 2013-02-22 Seth LaForge <sethml@google.com>
14723 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
14725 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
14727 * config/arm/arm.md (split for extendsidi): Update condition.
14728 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
14729 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
14730 (qhs_zextenddi_cstr): Likewise.
14732 2013-02-21 Jakub Jelinek <jakub@redhat.com>
14734 PR middle-end/56420
14735 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
14736 avoid signed wrapping.
14737 (expand_mult): Handle properly multiplication by
14738 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
14739 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
14740 in the compiler if coeff is HOST_WIDE_INT_MIN.
14741 (expand_divmod): Don't make ext_op1 static, change it's type to
14742 uhwi. Avoid undefined behavior in -INTVAL (op1).
14744 PR rtl-optimization/50339
14745 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
14747 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
14748 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
14749 into splitting_ashiftrt field.
14750 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
14752 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
14755 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
14757 PR middle-end/56108
14758 * trans-mem.c (execute_tm_mark): Do not expand transactions that
14759 are sure to go irrevocable.
14761 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
14763 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
14764 scalars are valid operands.
14766 2013-02-21 Martin Jambor <mjambor@suse.cz>
14768 PR tree-optimization/56310
14769 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
14770 only matching indices and non-negative final offsets.
14771 (intersect_aggregates_with_edge): Pass src_idx to
14772 agg_replacements_to_vector. Pass src_idx insstead of index to
14773 intersect_with_agg_replacements.
14775 2013-02-21 Martin Jambor <mjambor@suse.cz>
14777 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
14778 instead of hard-wired defaults.
14780 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
14782 * doc/invoke.texi (MIPS Options): Update documentation of the
14783 floating-point multiply-accumulate instruction restrictions.
14785 2013-02-21 Kostya Serebryany <kcc@google.com>
14787 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
14788 asan_shadow_offset on x86_64 linux.
14790 2013-02-21 Richard Biener <rguenther@suse.de>
14792 PR tree-optimization/56415
14794 2013-02-11 Richard Biener <rguenther@suse.de>
14796 PR tree-optimization/56273
14797 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
14800 2013-02-21 Jakub Jelinek <jakub@redhat.com>
14803 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
14806 PR inline-asm/56405
14807 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
14808 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
14810 2013-02-20 Jan Hubicka <jh@suse.cz>
14812 PR tree-optimization/56265
14813 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
14814 when target is referenced for first time.
14816 2013-02-20 Richard Biener <rguenther@suse.de>
14818 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
14819 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
14820 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
14821 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
14822 not return anything.
14823 (rename_ssa_copies): Do not remove unused locals.
14824 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
14825 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
14826 * passes.c (execute_function_todo): Do not schedule unused locals
14827 removal if cleanup_tree_cfg did something.
14828 * tree-ssa-live.c (remove_unused_locals): Dump statistics
14829 about the number of removed locals.
14831 2013-02-20 Richard Biener <rguenther@suse.de>
14833 PR tree-optimization/56398
14834 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
14836 2013-02-20 Martin Jambor <mjambor@suse.cz>
14838 PR tree-optimization/55334
14839 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
14840 restricted pointers to arrays.
14842 2013-02-20 Richard Biener <rguenther@suse.de>
14843 Jakub Jelinek <jakub@redhat.com>
14845 PR tree-optimization/56396
14846 * tree-ssa-ccp.c (n_const_val): New static variable.
14847 (get_value): Return NULL for SSA names we don't have a lattice
14849 (ccp_initialize): Initialize n_const_val.
14850 * tree-ssa-copy.c (n_copy_of): New static variable.
14851 (init_copy_prop): Initialize n_copy_of.
14852 (get_value): Return NULL_TREE for SSA names we don't have a
14855 2013-02-20 Martin Jambor <mjambor@suse.cz>
14857 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
14859 2013-02-20 Richard Biener <rguenther@suse.de>
14861 * genpreds.c (write_lookup_constraint): Do not compare first
14862 letter of the constraint again.
14864 2013-02-20 Richard Biener <rguenther@suse.de>
14866 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
14868 (get_use_iv_cost): Terminate hashtable walk when coming across
14871 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
14873 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
14874 reassociation for avx2 targets.
14876 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14878 * config/microblaze/microblaze.c: microblaze_has_clz = 0
14879 Add version check for v8.10.a to enable microblaze_has_clz
14880 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
14881 version and TARGET_PATTERN_COMPARE check
14882 * config/microblaze/microblaze.md: New clzsi2 instruction
14884 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14886 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
14887 function before branching.
14889 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
14891 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
14893 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
14895 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
14897 PR middle-end/55889
14898 * sel-sched.c: Include ira.h.
14899 (implicit_clobber_conflict_p): New function.
14900 (moveup_expr): Use it.
14901 * Makefile.in (sel-sched.o): Depend on ira.h.
14903 2013-02-19 Richard Biener <rguenther@suse.de>
14905 PR tree-optimization/56384
14906 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
14907 (vn_hash_type): Split out from ...
14908 (vn_hash_constant_with_type): ... here.
14909 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
14910 (vn_phi_eq): Compare types from vn_phi_s structure.
14911 (vn_phi_lookup): Populate vn_phi_s type.
14912 (vn_phi_insert): Likewise.
14914 2013-02-19 Jakub Jelinek <jakub@redhat.com>
14916 PR tree-optimization/56350
14917 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
14918 if haven't found reduction or nested cycle operand, rather than
14919 asserting we must find it.
14921 PR tree-optimization/56381
14922 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
14925 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
14926 Jakub Jelinek <jakub@redhat.com>
14929 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
14930 (swap_optab_enable): Same.
14931 (init_all_optabs): Use argument instead of global.
14932 * tree.h (struct tree_optimization_option): New field target_optabs.
14933 * expr.h (init_all_optabs): Add argument to prototype.
14934 (TREE_OPTIMIZATION_OPTABS): New.
14935 (save_optabs_if_changed): Protoize.
14936 * optabs.h: Declare this_fn_optabs.
14937 * optabs.c (save_optabs_if_changed): New.
14938 Declare this_fn_optabs.
14939 (init_optabs): Add argument to init_all_optabs() call.
14940 * function.c (invoke_set_current_function_hook): Handle per
14942 * function.h (struct function): New field optabs.
14943 * config/mips/mips.c (mips_set_mips16_mode): Handle when
14944 optimization_current_node has changed.
14945 * target-globals.h (save_target_globals_default_opts): Protoize.
14946 * target-globals.c (save_target_globals_default_opts): New.
14948 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14951 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
14952 registers %fr12 and %fr12R as call used.
14955 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
14956 and HImode, require all displacements to be an integer multiple of
14958 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
14959 only allow QImode and HImode when reload is in progress and strict is
14960 true. Likewise for symbolic addresses. Use base14_operand to check
14961 displacements in REG+BASE addresses.
14963 2013-02-18 Richard Biener <rguenther@suse.de>
14965 PR tree-optimization/56366
14966 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14967 handle sign-conversion of outer-loop initial induction value.
14969 2013-02-18 Richard Biener <rguenther@suse.de>
14971 PR middle-end/56349
14972 * cfghooks.c (merge_blocks): If we merge a latch into another
14973 block adjust references to it.
14974 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
14975 (verify_loop_structure): Verify that a recorded latch is in fact
14978 2013-02-18 Richard Biener <rguenther@suse.de>
14980 PR tree-optimization/56321
14981 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
14982 order SSA name release and virtual operand unlinking.
14984 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14986 * config/microblaze/microblaze.md (save_stack_block): Define.
14987 (restore_stack_block): Likewise.
14989 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14991 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
14992 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
14993 * config/microblaze/microblaze.c (microblaze_option_override):
14994 Bail out early for PIC modes when target does not support PIC.
14996 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14998 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
14999 Replace with a microblaze version.
15000 (microblaze_trampoline_init): Adapt for microblaze.
15001 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
15004 2013-02-16 Jakub Jelinek <jakub@redhat.com>
15005 Dodji Seketeli <dodji@redhat.com>
15008 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
15009 (instrument_mem_region_access): Do not forget to always put
15010 instrumentation of the of 'base' and 'base + len' in a "if (len !=
15011 0) statement, even for cases where either 'base' or 'base + len'
15012 are not instrumented -- because they have been previously
15013 instrumented. Simplify the logic by putting all the statements
15014 instrument 'base + len' inside a sequence, and then insert that
15015 sequence right before the current insertion point. Then, to
15016 instrument 'base + len', just get an iterator on that statement.
15017 And do not forget to update the pointer to iterator the function
15018 received as argument.
15020 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
15022 PR rtl-optimization/56348
15023 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
15025 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15027 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15028 (clean_graph_dump_file): Pass base to start_graph_dump.
15030 2013-02-14 Richard Henderson <rth@redhat.com>
15033 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15035 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15037 * collect2-aix.h: Define F_LOADONLY.
15039 2013-02-14 Richard Biener <rguenther@suse.de>
15042 * varasm.c (output_constant_def_1): Get the decl representing
15043 the constant as argument.
15044 (output_constant_def): Wrap output_constant_def_1.
15045 (make_decl_rtl): Use output_constant_def_1 with the decl
15046 representing the constant.
15047 (build_constant_desc): Optionally re-use a decl already
15048 representing the constant.
15049 (tree_output_constant_def): Adjust.
15051 2013-02-14 Dodji Seketeli <dodji@redhat.com>
15054 * asan.c (instrument_builtin_call): Really put the length of the
15055 second source argument into src1_len.
15057 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15059 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15060 argument. If it is false, don't create edge from then_bb to
15062 (insert_if_then_before_iter): Pass true to it.
15063 (build_check_stmt): Pass false to it.
15064 (transform_statements): Flush hash table only on extended basic
15065 block boundaries, rather than at the beginning of every bb.
15066 Don't flush hash table on nonfreeing_call_p calls.
15067 * tree-flow.h (nonfreeing_call_p): New prototype.
15068 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15070 2013-02-13 David S. Miller <davem@davemloft.net>
15072 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15074 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15077 * ira.c (max_regno_before_ira): Move from ...
15079 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15082 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15084 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15086 2013-02-13 Richard Biener <rguenther@suse.de>
15089 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15090 globals in MEM_REFs.
15092 2013-02-13 Richard Biener <rguenther@suse.de>
15094 * loop-init.c (loop_optimizer_init): Clear loop state when
15095 re-initializing preserved loops.
15096 * loop-unswitch.c (unswitch_single_loop): Return whether
15097 we unswitched the loop. Do not verify loop state here.
15098 (unswitch_loops): When we unswitched a loop discover new loops.
15100 2013-02-13 Kostya Serebryany <kcc@google.com>
15102 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15104 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15106 2013-02-12 Dodji Seketeli <dodji@redhat.com>
15108 Avoid instrumenting duplicated memory access in the same basic block
15109 * Makefile.in (asan.o): Add new dependency on hash-table.h
15110 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15111 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15112 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15113 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15114 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15115 (get_mem_ref_of_assignment): New functions.
15116 (get_mem_refs_of_builtin_call): Extract from
15117 instrument_builtin_call and tweak a little bit to make it fit with
15119 (instrument_builtin_call): Use the new
15120 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15121 of is_gimple_builtin_call.
15122 (instrument_derefs, instrument_mem_region_access): Insert the
15123 instrumented memory reference into the hash table.
15124 (maybe_instrument_assignment): Renamed instrument_assignment into
15125 this, and change it to advance the iterator when instrumentation
15126 actually happened and return true in that case. This makes it
15127 homogeneous with maybe_instrument_assignment, and thus give a
15128 chance to callers to be more 'regular'.
15129 (transform_statements): Clear the memory reference hash table
15130 whenever we enter a new BB, when we cross a function call, or when
15131 we are done transforming statements. Use
15132 maybe_instrument_assignment instead of instrumentation. No more
15133 need to special case maybe_instrument_assignment and advance the
15134 iterator after calling it; it's now handled just like
15135 maybe_instrument_call. Update comment.
15137 2013-02-13 Richard Biener <rguenther@suse.de>
15139 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15140 Fix loop discovery code.
15142 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15144 PR inline-asm/56148
15145 * lra-constraints.c (process_alt_operands): Match early clobber
15146 operand with itself. Check conflicts with earlyclobber only if
15147 the operand is not reloaded. Prefer to reload conflicting operand
15148 if earlyclobber and matching operands are the same.
15150 2013-02-12 Richard Biener <rguenther@suse.de>
15153 * lto-streamer-out.c (write_symbol): Do not output symbols
15154 for hard register variables.
15156 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
15159 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15160 (umulsidi3_insn, mulsidi3_insn): New insns.
15162 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
15164 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15165 (struct tune_params): Add vec_costs field.
15166 * config/arm/arm.c (arm_builtin_vectorization_cost)
15167 (arm_add_stmt_cost): New functions.
15168 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15169 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15170 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15171 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15172 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15173 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15174 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15176 2013-02-12 Richard Biener <rguenther@suse.de>
15179 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15180 decls again if possible.
15182 2013-02-12 Richard Biener <rguenther@suse.de>
15184 PR middle-end/56288
15185 * tree-ssa.c (verify_ssa_name): Fix check, move
15186 SSA_NAME_IN_FREE_LIST check up.
15188 2013-02-12 Jakub Jelinek <jakub@redhat.com>
15189 Steven Bosscher <steven@gcc.gnu.org>
15191 PR rtl-optimization/56151
15192 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15193 equal to op0 or op1, and last_insn pattern is CODE operation
15194 with MEM dest and one of the operands matches that MEM.
15196 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15198 * doc/extend.texi: Document Function Multiversioning and "default"
15199 parameter string to target attribute.
15200 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15201 target attribute parameter is "default".
15202 (ix86_compare_version_priority): Remove checks for target attribute.
15203 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15204 Remove check for target attribute equal to NULL. Add assert.
15205 (ix86_generate_version_dispatcher_body): Change error to sorry.
15207 2013-02-11 Iain Sandoe <iain@codesourcery.com>
15208 Jack Howarth <howarth@bromo.med.uc.edu>
15209 Patrick Marlier <patrick.marlier@gmail.com>
15212 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15213 define ENDFILE_SPEC as TM_DESTRUCTOR.
15214 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15216 2013-02-11 Alexander Potapenko <glider@google.com>
15217 Jack Howarth <howarth@bromo.med.uc.edu>
15218 Jakub Jelinek <jakub@redhat.com>
15221 * config/darwin.c (cdtor_record): Rename ctor_record.
15222 (sort_cdtor_records): Rename sort_ctor_records.
15223 (finalize_dtors): New routine to sort destructors by
15224 priority before use in assemble_integer.
15225 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15227 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
15229 PR rtl-optimization/56275
15230 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15231 offset is non-negative and less than cmode size before
15232 calling simplify_subreg.
15234 2013-02-11 Richard Biener <rguenther@suse.de>
15236 PR tree-optimization/56264
15237 * cfgloop.h (fix_loop_structure): Adjust prototype.
15238 * loop-init.c (fix_loop_structure): Return the number of
15239 newly discovered loops.
15240 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15241 are discovered, do a full loop-closed SSA rewrite.
15243 2013-02-11 Richard Biener <rguenther@suse.de>
15245 PR tree-optimization/56273
15246 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15248 (check_array_ref): Fix missing newline in dumps.
15249 (search_for_addr_array): Likewise.
15251 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
15253 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15255 2013-02-09 Jakub Jelinek <jakub@redhat.com>
15258 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15260 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15262 PR rtl-optimization/56246
15263 * lra-constraints.c (simplify_operand_subreg): Try to reuse
15265 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15266 constraints are satisfied.
15268 2013-02-08 Jeff Law <law@redhat.com>
15271 * emit-rtl.c (reg_is_parm_p): New function.
15272 * regs.h (reg_is_parm_p): New prototype.
15273 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15274 callee-clobbered registers.
15276 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15279 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15280 If there is no implicit builtin declaration, just return NULL.
15282 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15284 * config/i386/sse.md (FMAMODEM): New mode iterator.
15285 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15286 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15288 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15290 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15291 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15292 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15294 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15296 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15297 (microblaze*-*-elf): Likewise.
15298 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15300 * config/microblaze/microblaze-c.c: Add builtin defines for
15301 _LITTLE_ENDIAN and _BIG_ENDIAN.
15302 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15303 add to TARGET_DEFAULT flags.
15304 Expand ASM_SPEC and LINK_SPEC.
15305 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15306 * config/microblaze/microblaze.md: Update extendsidi2 and
15307 movdi_internal instructions to use low-order / high-order reg
15309 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15310 options and inversemask / mask of LITTLE_ENDIAN.
15311 * config/microblaze/t-microblaze: Expand multilib options to
15312 include mlittle-endian (le) and update exceptions patterns.
15314 2013-02-08 Jakub Jelinek <jakub@redhat.com>
15316 PR rtl-optimization/56195
15317 * lra-constraints.c (get_reload_reg): Don't reuse regs
15318 if they have smaller mode than requested, if they have
15319 wider mode than requested, try to return a SUBREG.
15321 PR tree-optimization/56250
15322 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15323 if type is unsigned and code isn't MULT_EXPR.
15325 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15327 PR tree-optimization/56064
15328 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15329 bits according to mode.
15330 * fixed-value.h (fixed_from_double_int)
15331 (const_fixed_from_double_int): Adjust comments.
15333 2013-02-08 Richard Biener <rguenther@suse.de>
15336 * lto-streamer.h (struct data_in): Remove current_file, current_line
15337 and current_col members.
15338 * lto-streamer-out.c (lto_output_location): Stream changed bits
15339 en-block for efficiency.
15340 * lto-streamer-in.c (clear_line_info): Remove.
15341 (lto_input_location): Cache current file, line and column
15342 globally via local statics. Read changed bits en-block.
15343 (input_function): Do not call clear_line_info.
15344 (lto_read_body): Likewise.
15345 (lto_input_toplevel_asms): Likewise.
15347 2013-02-08 Michael Matz <matz@suse.de>
15349 PR tree-optimization/52448
15350 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15351 (nt_call_phase): New static.
15352 (add_or_mark_expr): Only mark accesses with newer phase than any
15354 (nonfreeing_call_p): New.
15355 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15356 (nt_fini_block): Keep blocks marked as visited.
15357 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15359 2013-02-08 Richard Biener <rguenther@suse.de>
15361 * ira.c (ira): Free broken dominator information.
15363 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15365 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15367 2013-02-08 Marek Polacek <polacek@redhat.com>
15369 * cfgloop.c (verify_loop_structure): Add more checking of headers.
15371 2013-02-08 Richard Biener <rguenther@suse.de>
15373 PR middle-end/56181
15374 * cfgloop.h (flow_loops_find): Adjust.
15375 (bb_loop_header_p): Declare.
15376 * cfgloop.c (bb_loop_header_p): New function split out from ...
15377 (flow_loops_find): ... here. Adjust function signature,
15378 support incremental loop structure update.
15379 (verify_loop_structure): Cleanup. Verify a loop is a loop.
15380 * cfgloopmanip.c (fix_loop_structure): Move ...
15381 * loop-init.c (fix_loop_structure): ... here.
15382 (apply_loop_flags): Split out from ...
15383 (loop_optimizer_init): ... here.
15384 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
15385 in incremental mode, only remove dead loops here.
15387 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15390 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
15391 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
15392 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
15393 (*round<mode>3.libgcc): New insns for fixed-modes.
15394 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
15395 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
15396 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
15397 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
15398 implementations. Define to __builtin_avr_absFX,
15399 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
15400 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
15401 __builtin_avr_countlsFX, respectively.
15402 * config/avr/avr-c.c (target.h): Include it.
15403 (enum avr_builtin_id): New enum.
15404 (avr_resolve_overloaded_builtin): New static function.
15405 (avr_register_target_pragmas): Use it to set
15406 targetm.resolve_overloaded_builtin.
15407 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
15408 tree nodes used by DEF_BUILTIN.
15409 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
15410 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
15411 <AVR_BUILTIN_xxBITS>: Same.
15413 2013-02-08 Richard Biener <rguenther@suse.de>
15415 * cfgloop.c (verify_loop_structure): Properly handle
15416 a loop exiting to another loop header.
15417 * ira-int.h (ira_loops): Remove.
15418 * ira.c (ira_loops): Remove.
15419 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
15420 (do_reload): Use loop_optimizer_finalize.
15421 * ira-build.c (create_loop_tree_nodes): Use get_loops and
15422 number_of_loops to access the loop tree.
15423 (more_one_region_p): Likewise.
15424 (finish_loop_tree_nodes): Likewise.
15425 (rebuild_regno_allocno_maps): Likewise.
15426 (mark_loops_for_removal): Likewise.
15427 (mark_all_loops_for_removal): Likewise.
15428 (remove_unnecessary_regions): Likewise.
15429 (ira_build): Likewise.
15430 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15432 2013-02-08 Richard Biener <rguenther@suse.de>
15434 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
15435 * ipa-pure-const.c (analyze_function): Avoid calling
15436 mark_irreducible_loops twice.
15437 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
15439 2013-02-07 David S. Miller <davem@davemloft.net>
15441 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
15443 * var-tracking.c (vt_add_function_parameter): Test the presence of
15444 HAVE_window_save properly and do not remap argument registers when
15445 we have a leaf function.
15447 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
15450 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
15452 * config/i386/i386.c (ix86_print_operand): Ditto.
15454 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15456 * lra-constraints.c (process_alt_operands): Fix recently added comment.
15458 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15460 PR rtl-optimization/56225
15461 * lra-constraints.c (process_alt_operands): Check that reload hard
15462 reg can hold value for strict_low_part.
15464 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15467 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
15468 dwarf2out_end_function.
15469 (in_first_function_p, maybe_at_text_label_p,
15470 first_loclabel_num_not_at_text_label): New variables.
15471 (dwarf2out_var_location): In the first function find out
15472 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
15473 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
15476 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
15478 PR rtl-optimization/56178
15479 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
15480 SUBREG of a register. Tidy up related block of code.
15481 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
15482 note if the source is a register or a SUBREG of a register.
15484 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15487 * config/rs6000/rs6000.md (ptrm): New mode attr.
15488 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
15489 call_value_indirect_aix<pttrsize>,
15490 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
15493 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
15495 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
15496 if -bnortl. Convert to strcmp and strncmp.
15498 2013-02-07 Alan Modra <amodra@gmail.com>
15501 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
15502 addresses won't wrap when offsetting.
15503 (rs6000_secondary_reload): Provide secondary reloads needed for
15504 wrapping LO_SUM addresses.
15506 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
15508 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
15509 MACH, just __MACH__.
15511 2013-02-06 Richard Biener <rguenther@suse.de>
15513 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
15514 instead of calling fix_loop_structure.
15516 2013-02-06 Jakub Jelinek <jakub@redhat.com>
15518 PR middle-end/56217
15519 * omp-low.c (use_pointer_for_field): Return false if
15520 lower_send_shared_vars doesn't generate any copy-out code.
15522 2013-02-06 Tom de Vries <tom@codesourcery.com>
15524 PR rtl-optimization/56131
15525 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
15526 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
15527 of the label is NULL. Add comment.
15529 2013-02-05 Jakub Jelinek <jakub@redhat.com>
15531 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
15534 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
15535 (STATIC_LIBTSAN_LIBS): Likewise.
15536 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
15537 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
15538 is defined, don't add anything else beyond that.
15539 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
15540 (LINK_COMMAND_SPEC): Use them.
15542 PR tree-optimization/56205
15543 * tree-stdarg.c (check_all_va_list_escapes): Return true if
15544 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
15545 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
15547 2013-02-05 Richard Biener <rguenther@suse.de>
15549 PR tree-optimization/53342
15550 PR tree-optimization/53185
15551 * tree-vectorizer.h (vect_check_strided_load): Remove.
15552 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
15553 not disallow peeling for vectorized strided loads.
15554 (vect_check_strided_load): Make static and simplify.
15555 (vect_analyze_data_refs): Adjust.
15556 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
15557 correctly when vectorizing strided loads.
15559 2013-02-05 Richard Biener <rguenther@suse.de>
15561 * doc/install.texi: Refer to ISL, not PPL.
15563 2013-02-05 Jan Hubicka <jh@suse.cz>
15565 PR tree-optimization/55789
15566 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
15568 2013-02-05 Jan Hubicka <jh@suse.cz>
15570 PR tree-optimization/55789
15571 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
15572 the dead call anyway.
15574 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
15577 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
15579 2013-02-04 Alexander Potapenko <glider@google.com>
15580 Jack Howarth <howarth@bromo.med.uc.edu>
15581 Jakub Jelinek <jakub@redhat.com>
15584 * config/darwin.c (sort_ctor_records): Stabilized qsort
15585 on constructor priority by using original position.
15586 (finalize_ctors): New routine to sort constructors by
15587 priority before use in assemble_integer.
15588 (machopic_asm_out_constructor): Use finalize_ctors if needed.
15590 2013-02-04 Jakub Jelinek <jakub@redhat.com>
15593 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
15594 about visibility on artificial decls.
15595 * config/sol2.c (solaris_assemble_visibility): Likewise.
15597 2013-02-04 Kai Tietz <ktietz@redhat.com>
15600 * config/i386/i386.c (function_value_ms_64): Add additional valtype
15601 argument and improve checking of return-argument types for 16-byte
15603 (ix86_function_value_1): Add additional valtype argument on call
15604 of function_value_64.
15605 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
15606 handling infunction_value_64 function.
15608 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
15610 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
15612 2013-02-04 Richard Biener <rguenther@suse.de>
15614 PR tree-optimization/56188
15615 * tree-ssa-structalias.c (label_visit): Consider case with
15616 initially non-empty points-to set.
15617 (perform_var_substitution): Dump node mapping and clean up.
15619 2013-02-04 Richard Guenther <rguenther@suse.de>
15622 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
15623 node prevail as last resort.
15624 (lto_symtab_merge_decls): Remove guard on LTRANS here.
15625 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
15627 2013-02-04 Richard Biener <rguenther@suse.de>
15629 PR tree-optimization/56113
15630 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
15632 (equiv_class_lookup_or_add): ... this.
15633 (label_visit): Adjust and fix error in previous patch.
15634 (perform_var_substitution): Adjust.
15636 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
15638 * config/sh/divtab.c: Fix formatting and comments throughout the file.
15639 * config/sh/sh4-300.md: Likewise.
15640 * config/sh/sh4a.md: Likewise.
15641 * config/sh/constraints.md: Likewise.
15642 * config/sh/sh.md: Likewise.
15643 * config/sh/netbsd-elf.h: Likewise.
15644 * config/sh/predicates.md: Likewise.
15645 * config/sh/sh-protos.h: Likewise.
15646 * config/sh/ushmedia.h: Likewise.
15647 * config/sh/linux.h: Likewise.
15648 * config/sh/sh.c: Likewise.
15649 * config/sh/superh.h: Likewise.
15650 * config/sh/elf.h: Likewise.
15651 * config/sh/sh4.md: Likewise.
15652 * config/sh/sh.h: Likewise.
15654 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15656 * config/pa/constraints.md: Adjust unused letters. Change "T"
15657 constraint to match_test floating_point_store_memory_operand().
15658 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
15659 (base14_operand): New.
15660 (floating_point_store_memory_operand): New.
15661 (integer_store_memory_operand): Revise to use base14_operand and
15662 reg_plus_base_memory_operand.
15663 (move_dest_operand): Allow symbolic_memory_operands.
15664 (symbolic_memory_operand): Check for LO_SOM.
15665 (symbolic_operand): Change default case to break.
15666 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
15667 CONST_DOUBLE values to be reloaded by putting them into memory when
15668 the destination is a floating point register.
15669 (movdf): Remove code to handle CONST_DOUBLE.
15671 (reload_indf_r1): New.
15672 (reload_insf_r1): New.
15673 Consistently use "Q" and "T" constraints with integer and floating
15674 point move instructions, respectively.
15675 (movdi): Remove FAIL.
15676 Change predicate for source operand unamed DImode move from
15677 general_operand to move_src_operand.
15678 (umulsidi3): Change predicate for destination operand to
15680 Likewise for similar unamed patterns.
15681 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
15682 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
15683 (hppa_legitimize_address): Simplify mask calculation.
15684 (pa_emit_move_sequence): Revised handling of secondary reloads from
15685 REG+D addresses for floating point loads and stores. Directly handle
15686 loading CONST0_RTX (mode) to a floating point register.
15687 (pa_secondary_reload): Handle reloading DF and SFmode constant values
15688 to floating point registers. Don't restrict secondary reloads to
15689 floating point registers to integer modes. Revise some comments and
15691 (TARGET_LEGITIMATE_ADDRESS_P): Define.
15692 (pa_legitimate_address_p): New.
15693 (pa_legitimize_reload_address): New.
15694 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
15695 (STRICT_REG_OK_FOR_BASE_P): New.
15696 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
15697 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
15699 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
15700 Andrew Dixie <andrewd@gentrack.com>
15702 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
15705 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
15707 * expmed.c (extract_bit_field_1): Pass the full width of the
15708 structure to get_best_reg_extraction_insn.
15710 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
15713 * configure.ac (use_cxa_atexit): Add AIX.
15714 * configure: Regenerate.
15716 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
15718 2013-02-01 Jakub Jelinek <jakub@redhat.com>
15721 * final.c (need_profile_function): New variable.
15722 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
15723 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
15724 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
15725 notes, targetm.asm_out.function_prologue doesn't emit anything,
15726 HAVE_prologue and profiler should be emitted before prologue,
15727 set need_profile_function instead of emitting it.
15728 (final_scan_insn): If need_profile_function, emit
15729 profile_function on the first NOTE_INSN_BASIC_BLOCK or
15730 NOTE_INSN_FUNCTION_BEG note.
15732 2013-02-01 Richard Henderson <rth@redhat.com>
15734 * config/rs6000/rs6000.md (smulditi3): New.
15737 * config/alpha/alpha.md (umulditi3): New.
15739 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
15741 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
15742 (ASM_OUTPUT_ALIGNED_LOCAL): New.
15744 2013-02-01 Richard Biener <rguenther@suse.de>
15746 PR tree-optimization/56113
15747 * tree-ssa-structalias.c (label_visit): Reduce work for
15748 single-predecessor nodes.
15750 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
15752 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
15753 range isn't testing for zero.
15755 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
15757 PR middle-end/56113
15758 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
15760 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
15761 Nick Clifton <nickc@redhat.com>
15763 * config/v850/constraints.md (Q): Define as a memory constraint.
15764 * config/v850/predicates.md (label_ref_operand): New predicate.
15765 (e3v5_shift_operand): New predicate.
15766 (ior_operator): New predicate.
15767 * config/v850/t-v850: Add e3v5 multilib.
15768 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
15769 (v850_gen_movdi): Prototype.
15770 * config/v850/v850.c: Add support for e3v5 architecture.
15771 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
15773 (construct_save_jarl): Add e3v5 long JARL support.
15774 (v850_adjust_insn_length): New function. Adjust length of call
15775 insns when using e3v5 instructions.
15776 (v850_gen_movdi): New function: Generate instructions to move a
15778 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
15779 (CPP_SPEC): Define __v850e3v5__ as appropriate.
15780 (TARGET_USE_FPU): Enable for e3v5.
15781 (CONST_OK_FOR_W): New macro.
15782 (ADJUST_INSN_LENGTH): Define.
15783 * config/v850/v850.md (UNSPEC_LOOP): Define.
15784 (attr cpu): Add v850e3v5.
15785 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
15786 (movdi): New pattern.
15787 (movdi_internal): New pattern.
15788 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
15789 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
15790 (cstoresf4): Likewise.
15791 (cstoredf4): Likewise.
15792 (insv): New pattern.
15793 (rotlso3_a): New pattern.
15794 (rotlsi3_b): New pattern
15795 (rotlsi3_v850e3v5): New pattern.
15796 (doloop_begin): New pattern.
15797 (fix_loop_counter): New pattern.
15798 (doloop_end): New pattern.
15799 (branch_normal): Add e3v5 long branch support.
15800 (branch_invert): Likewise.
15801 (branch_z_normal): Likewise.
15802 (branch_z_invert): Likewise.
15803 (branch_nz_normal): Likewise.
15804 (branch_nz_invert): Likewise.
15805 (call_internal_short): Add e3v5 register-indirect JARL support.
15806 (call_internal_long): Likewise.
15807 (call_value_internal_short): Likewise.
15808 (call_value_internal_long): Likewise.
15809 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
15810 (mloop): New option.
15811 * config.gcc: Add support for configuring v840e3v5 target.
15812 * doc/invoke.texi: Document new v850 specific command line options.
15814 2013-01-31 Paul Koning <ni1d@arrl.net>
15818 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
15819 children if parent is a class.
15820 (prune_unused_types_prune): Don't add DW_AT_declaration.
15822 2013-01-31 Richard Biener <rguenther@suse.de>
15824 PR tree-optimization/56157
15825 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
15826 match up operand with SLP child.
15828 2013-01-31 Jason Merrill <jason@redhat.com>
15831 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
15832 parameters the first time.
15833 (gen_scheduled_generic_parms_dies): Check completeness here.
15835 2013-01-31 Richard Biener <rguenther@suse.de>
15837 PR middle-end/53073
15838 * common.opt (faggressive-loop-optimizations): New flag,
15839 enabled by default.
15840 * doc/invoke.texi (faggressive-loop-optimizations): Document.
15841 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
15842 infer_loop_bounds_from_undefined by it.
15844 2013-01-31 Richard Biener <rguenther@suse.de>
15846 PR tree-optimization/56150
15847 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
15848 visit virtual operands.
15849 (find_uses_to_rename_bb): Likewise.
15851 2013-01-31 Richard Biener <rguenther@suse.de>
15853 PR tree-optimization/56150
15854 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
15855 mixed store non-store stmts.
15857 2013-01-30 Jakub Jelinek <jakub@redhat.com>
15860 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
15861 LIBASAN_EARLY_SPEC is defined.
15862 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
15863 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
15865 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
15868 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
15869 invalid args instead of ICEing on it.
15870 (ix86_valid_target_attribute_tree): Return error_mark_node if
15871 ix86_valid_target_attribute_inner_p failed.
15872 (ix86_valid_target_attribute_p): Return false only if
15873 ix86_valid_target_attribute_tree returned error_mark_node. Allow
15874 target("default") attribute.
15875 (sorted_attr_string): Change argument from const char * to tree,
15876 merge in all target attribute arguments rather than just one.
15877 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
15878 instead of free. Avoid using strcat.
15879 (ix86_mangle_function_version_assembler_name): Mangle
15880 target("default") as if no target attribute is present. Adjust
15881 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
15882 instead of xmalloc and XDELETEVEC instead of free.
15883 (ix86_function_versions): Don't return true if one of the decls
15884 doesn't have target attribute. If they don't and one of the decls
15885 is DECL_FUNCTION_VERSIONED, report an error. Adjust
15886 sorted_attr_string caller. Use XDELETEVEC instead of free.
15887 (ix86_supports_function_versions): Remove.
15888 (make_name): Fix up formatting.
15889 (make_dispatcher_decl): Remove resolver_name and its initialization.
15890 Avoid leaking memory.
15891 (is_function_default_version): Return true if there is
15892 target("default") attribute rather than no target attribute at all.
15893 (make_resolver_func): Avoid leaking memory.
15894 (ix86_generate_version_dispatcher_body): Likewise.
15895 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
15896 * target.def (supports_function_versions): Remove.
15897 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
15898 * doc/tm.texi: Regenerated.
15900 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
15902 PR rtl-optimization/56144
15903 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
15904 for values with side effects.
15906 2013-01-30 Richard Biener <rguenther@suse.de>
15908 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
15909 (sparseset_pop): Likewise.
15910 * cfganal.c (compute_idf): Likewise. Increase work-stack size
15911 to be able to use quick_push in the worker loop.
15913 2013-01-30 Marek Polacek <polacek@redhat.com>
15915 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
15917 2013-01-30 Richard Biener <rguenther@suse.de>
15920 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
15922 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
15924 PR tree-optimization/56064
15925 * fixed-value.c (fixed_from_double_int): New function.
15926 * fixed-value.h (fixed_from_double_int): New prototype.
15927 (const_fixed_from_double_int): New static inline function.
15928 * fold-const.c (native_interpret_fixed): New static function.
15929 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
15930 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
15931 (native_encode_fixed): New static function.
15932 (native_encode_expr) <FIXED_CST>: Use it.
15933 (native_interpret_int): Move double_int worker code to...
15934 * double-int.c (double_int::from_buffer): ...this new static method.
15935 * double-int.h (double_int::from_buffer): Prototype it.
15937 2013-01-30 Richard Biener <rguenther@suse.de>
15939 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
15940 New pointer-map and obstack.
15941 (init_alias_vars): Allocate pointer-map and obstack.
15942 (delete_points_to_sets): Free them.
15943 (find_what_var_points_to): Cache result.
15944 (find_what_p_points_to): Adjust for changed interface of
15945 find_what_var_points_to.
15946 (compute_points_to_sets): Likewise.
15947 (ipa_pta_execute): Likewise.
15949 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15951 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
15952 * configure: Regenerate.
15953 * config.in: Regenerate.
15954 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
15955 #nobits/#progbits if supported.
15957 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
15960 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
15961 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
15962 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
15964 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15966 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
15967 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
15969 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15971 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
15972 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
15974 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15976 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
15978 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
15979 * config/arm/cortex-a7.md: New bypasses using
15980 arm_mac_accumulator_is_result.
15982 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15984 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
15985 (cortex_a7_neon_mla): Likewise.
15986 (cortex_a7_fpfmad): New reservation.
15987 (cortex_a7_fpmacs): Use ffmas and update required units.
15988 (cortex_a7_fpmuld): Update required units and latency.
15989 (cortex_a7_fpmacd): Likewise.
15990 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
15991 (cortex_a7_neon). Likewise.
15992 (bypass) Update participating units.
15994 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15996 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
15997 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
15999 * config/arm/vfp11.md (vfp_farith): Use ffmas.
16000 (vfp_fmul): Use ffmad.
16001 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
16002 (cortex_r4_fmacd): Use ffmad.
16003 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
16004 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
16005 (cortex_a9_fmacd): Use ffmad.
16006 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
16007 (cortex_a8_vfp_macd): Use ffmad.
16008 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
16009 (cortex_a5_fpmacd): Use ffmad.
16010 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
16011 (cortex_a15_vfp_macd): Use ffmad.
16012 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
16014 2013-01-29 Jason Merrill <jason@redhat.com>
16017 * varasm.c (default_assemble_visibility): Don't warn about
16018 visibility on artificial decls.
16020 2013-01-29 Richard Biener <rguenther@suse.de>
16022 PR tree-optimization/56113
16023 * tree-ssa-structalias.c (equiv_class_lookup): Also return
16025 (label_visit): Free duplicate bitmaps and record the leader instead.
16026 (perform_var_substitution): Adjust.
16028 2013-01-29 Richard Biener <rguenther@suse.de>
16030 PR tree-optimization/55270
16031 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16032 the CFG, schedule loops for fixup.
16034 2013-01-29 Nick Clifton <nickc@redhat.com>
16036 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16039 2013-01-28 Leif Ekblad <leif@rdos.net>
16041 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16042 * config/i386/i386.h (TARGET_RDOS): New macro.
16043 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16044 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16045 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16046 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16047 DEFAULT_LARGE_SECTION_THRESHOLD.
16048 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16049 * config/i386/rdos.h: New file.
16050 * config/i386/rdos64.h: New file.
16052 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16055 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16058 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16060 PR rtl-optimization/56117
16061 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16062 call cselib_lookup_from_insn on the MEM before calling
16063 add_insn_mem_dependence.
16065 2013-01-28 Richard Biener <rguenther@suse.de>
16067 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16068 to a stmt that didn't have one.
16069 (copy_phis_for_bb): Likewise for PHI arguments.
16070 (copy_debug_stmt): Likewise for debug stmts.
16072 2013-01-28 Richard Biener <rguenther@suse.de>
16074 PR tree-optimization/56034
16075 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
16076 (partition_builtin_p): Adjust.
16077 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16078 it is the last partition.
16079 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16080 up the vertex for the definition.
16081 (classify_partition): Classify whether a partition is a
16082 PKIND_REDUCTION, thus has uses outside of the loop.
16083 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16084 Merge all PKIND_REDUCTION partitions into the last partition.
16085 (tree_loop_distribution): Seed partitions from reductions as well.
16087 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16089 PR tree-optimization/56125
16090 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16091 pow(x,c) into sqrt(x) * powi(x, n/2) or
16092 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16093 optimizing for size.
16094 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16095 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16098 PR tree-optimization/56094
16099 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16100 to UNKNOWN_LOCATION while gimplifying expr.
16102 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
16105 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16106 operand 0 in movabs insn template for -masm=intel asm alternative.
16107 (*movabs<mode>_2): Ditto for operand 1.
16109 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
16112 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16115 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16117 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16120 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16122 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16123 Undef to avoid warning.
16125 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16127 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16128 * configure: Regenerate.
16130 2013-01-25 Jakub Jelinek <jakub@redhat.com>
16132 PR tree-optimization/56098
16133 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16134 for stmts with volatile ops.
16135 (cond_store_replacement): Don't optimize if assign has volatile ops.
16136 (cond_if_else_store_replacement_1): Don't optimize if either
16137 then_assign or else_assign have volatile ops.
16138 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16141 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16143 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16145 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16147 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16148 missing ':' in asm example.
16150 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16152 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16153 entries into lane and laneq entries.
16154 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16155 Remove AdvSIMD scalar modes.
16156 (aarch64_sq<r>dmulh_laneq<mode>): New.
16157 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16159 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16160 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16161 * config/aarch64/iterators.md (VCOND): New.
16164 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16167 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16168 Add NULL LIBNAME argument to existing definitions.
16169 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16170 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16171 * config/avr/avr.c (DEF_BUILTIN): Same.
16172 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16173 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16174 implementation is available (DECL_ASSEMBLER_NAME is set).
16175 (avr_fold_absfx): New static function.
16176 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16177 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16178 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16179 AVR_BUILTIN_ABSLLK.
16180 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16181 (abshk, absk, abslk, absllk): Provide as static inline functions.
16183 2013-01-25 Marek Polacek <polacek@redhat.com>
16185 PR tree-optimization/56035
16186 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16188 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
16190 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16191 (*movtf_internal_rex64): Add (!o,C) alternative
16192 (*movxf_internal_rex64): Ditto.
16193 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16195 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
16197 * doc/invoke.texi: fix typo.
16198 * doc/objc.texi: fix typo.
16200 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16202 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16203 for the first two alternatives.
16205 2013-01-24 Diego Novillo <dnovillo@google.com>
16207 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
16208 (ggc-zone.o): Remove.
16209 * configure.ac: Remove option --with-gc.
16210 * configure: Re-generate.
16211 * doc/install.texi: Remove documentation for --with-gc.
16212 * gengtype.c (write_enum_defn): Remove. Update all users.
16213 (write_Types_process_field): Remove generation of gt_e_* argument.
16214 (output_type_enum): Remove. Update all users.
16215 (write_enum_defn): Remove. Update all users.
16216 (enum alloc_zone): Remove. Update all users.
16217 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
16218 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16219 Update all callers.
16220 (struct ptr_data): Remove field TYPE. Update all users.
16221 (gt_pch_note_object): Remove argument TYPE. Update all users.
16222 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16224 * ggc-none.c (ggc_alloc_typed_stat): Remove.
16225 (struct alloc_zone): Remove.
16226 (ggc_internal_alloc_zone_stat): Remove.
16227 (ggc_internal_cleared_alloc_zone_stat): Remove.
16228 * ggc-page.c (ggc_alloc_typed_stat): Remove.
16229 (ggc_pch_count_object): Remove last argument. Update all users.
16230 (ggc_pch_alloc_object): Remove last argument. Update all users.
16231 (struct alloc_zone): Remove.
16232 * ggc-zone.c: Remove.
16233 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
16234 (struct alloc_zone): Remove.
16235 (ggc_alloc_typed_stat): Remove.
16236 (ggc_alloc_typed): Remove.
16237 (ggc_splay_alloc): Remove first argument.
16238 (rtl_zone): Remove. Update all users.
16239 (tree_zone): Remove. Update all users.
16240 (tree_id_zone): Remove. Update all users.
16241 (ggc_internal_zone_alloc_stat): Remove. Update all users.
16242 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
16243 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
16244 * tree-ssanames.c: Remove references to zone allocator in comments.
16246 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
16248 * config/avr/avr.c (avr_out_fract): Make register numbers that
16249 might be outside of source operand signed.
16251 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
16253 * config/i386/constraints.md (Yf): New constraint.
16254 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16255 of f constraint to conditionaly disable x87 register preferences.
16256 (*movdf_internal): Ditto.
16257 (*movsf_internal): Ditto.
16259 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16261 PR inline-asm/55934
16262 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16263 that have operands with impossible constraints.
16264 Add a FIXME for a speed-up opportunity.
16265 * lra-constraints.c (process_alt_operands): Verify that a class
16266 selected from constraints on asms is valid for the operand mode.
16267 (curr_insn_transform): Remove incorrect comment.
16269 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
16271 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16272 TOC operand is a valid symbol ref in the constant pool.
16274 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16276 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
16278 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
16281 * config/avr/stdfix.h: New file.
16282 * t-avr (stdfix-gcc.h): New rule to build it.
16283 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16285 2013-01-23 Kostya Serebryany <kcc@google.com>
16287 * config/darwin.h: remove dependency on
16288 CoreFoundation (asan on Mac OS).
16290 2013-01-23 Jakub Jelinek <jakub@redhat.com>
16293 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16294 instead of cmpdi_operand for first comparison operand.
16295 Don't assert that comparison operands aren't both constants.
16297 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16299 * doc/install.texi (Downloading the Source): Update references to
16300 downloading separate components.
16302 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16304 * doc/extend.texi (__int128): Improve grammar.
16306 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
16309 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16310 alternative to (o,r).
16311 (*movdi_internal_rex64): Remove (!o,n) alternative.
16312 (DImode immediate->memory splitter): Remove.
16313 (DImode immediate->memory peephole2): Remove.
16314 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16315 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16316 alternative to (!o,*r).
16317 (*movtf_internal_sse): New pattern.
16318 (*movxf_internal_rex64): New pattern.
16319 (*movxf_internal): Disable for TARGET_64BIT.
16320 (*movdf_internal_rex64): Remove (!o,F) alternative.
16322 2013-01-22 Jakub Jelinek <jakub@redhat.com>
16324 PR middle-end/56074
16325 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16326 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16327 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16328 stmt locations where LOCATION_LOCUS of the stmt location is
16329 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16332 * config/i386/i386.md (UNSPEC_STOS): New.
16333 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16334 *strsetqi_1): Add UNSPEC_STOS.
16336 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16339 * doc/invoke.texi: Remove left over -Wsynth example.
16341 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16343 PR tree-optimization/56051
16344 * fold-const.c (fold_binary_loc): Don't fold
16345 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16346 a narrowing conversion, or widening conversion from signed
16349 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16351 PR rtl-optimization/56023
16352 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16353 dependent on debug instruction.
16355 2013-01-21 Martin Jambor <mjambor@suse.cz>
16357 PR middle-end/56022
16358 * function.c (allocate_struct_function): Call
16359 invoke_set_current_function_hook earlier.
16361 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16363 * reload1.c (init_reload): Only initialize reload_obstack
16364 during the first call.
16366 2013-01-21 Marek Polacek <polacek@redhat.com>
16368 * cfgloop.c (verify_loop_structure): Fix up grammar.
16370 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
16372 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
16373 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
16375 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16378 * config/arm/marvell-pj4.md: Update copyright year.
16379 Fix up use of alu to alu_reg and simple_alu_imm.
16381 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16383 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
16385 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
16388 * lra-constraints.c (curr_insn_transform): Don't reuse original
16389 insn for secondary memory move when memory mode should be different.
16391 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16393 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
16394 atomic_storedi_1): New patterns.
16396 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16398 btver2 pipeline descriptions.
16399 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
16401 * config/i386/i386.md (btver2_decode): New type attributes.
16402 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
16404 * config/i386/btver2.md: New file describing btver2 pipelines.
16406 2013-01-19 Andrew Pinski <apinski@cavium.com>
16408 PR tree-optimization/52631
16409 * tree-ssa-sccvn (visit_use): Before looking up the original
16410 statement, try looking up the simplified expression.
16412 2013-01-19 Anthony Green <green@moxielogic.com>
16414 * config/moxie/moxie.c (moxie_expand_prologue): Set
16415 current_function_static_stack_size.
16417 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16419 PR tree-optimization/56029
16420 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
16421 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
16423 2013-01-18 Sharad Singhai <singhai@google.com>
16425 PR tree-optimization/55995
16426 * dumpfile.c (dump_loc): Print location only if available.
16427 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
16429 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
16432 * lra-constraints.c (curr_insn_transform): Reuse original insn for
16433 secondary memory move.
16434 (inherit_reload_reg): Use rclass instead of cl for
16435 check_secondary_memory_needed_p.
16437 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16439 PR middle-end/56015
16440 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
16441 the case where writing real complex part of target modifies op1.
16443 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
16445 * config/aarch64/aarch64-simd.md
16446 (aarch64_vcond_internal<mode>): Handle unordered cases.
16447 * config/aarch64/iterators.md (v_cmp_result): New.
16449 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
16450 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16452 * config/arm/marvell-pj4.md: New file.
16453 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
16454 * config/arm/arm.md (generic_sched): Add marvell_pj4.
16455 (generic_vfp): Likewise.
16456 * config/arm/arm-cores.def: Add marvell-pj4.
16457 * config/arm/arm-tune.md: Regenerate.
16458 * config/arm/arm-tables.opt: Regenerate.
16459 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
16460 * doc/invoke.texi: Document marvell-pj4.
16462 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
16464 * config/aarch64/arm_neon.h: Map scalar types to standard types.
16466 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16471 * var-tracking.c (negative_power_of_two_p): New.
16472 (global_get_addr_cache, local_get_addr_cache): New.
16473 (get_addr_from_global_cache, get_addr_from_local_cache): New.
16474 (vt_canonicalize_addr): Rewrite using the above. Adjust the
16476 (vt_stack_offset_p): Remove.
16477 (vt_canon_true_dep): Always canonicalize loc's address.
16478 (clobber_overlapping_mems): Make sure we have a MEM.
16479 (local_get_addr_clear_given_value): New.
16480 (val_reset): Clear local cached entries.
16481 (compute_bb_dataflow): Create and release the local cache.
16482 Disable duplicate MEMs clobbering.
16483 (emit_notes_in_bb): Clobber MEMs likewise.
16484 (vt_emit_notes): Create and release the local cache.
16485 (vt_initialize, vt_finalize): Create and release the global
16486 cache, respectively.
16487 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
16489 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16491 PR libmudflap/53359
16492 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
16493 not found in the symtab.
16495 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16498 PR rtl-optimization/55547
16499 PR rtl-optimization/53827
16502 * alias.c (offset_overlap_p): New, factored out of...
16503 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
16504 the conservative special case for symbolic constants. Don't
16505 adjust zero sizes on alignment.
16507 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
16509 PR rtl-optimization/52573
16510 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
16511 REG_UNUSED for the same register.
16513 2013-01-17 Richard Biener <rguenther@suse.de>
16514 Marek Polacek <polacek@redhat.com>
16516 PR rtl-optimization/55833
16517 * loop-unswitch.c (unswitch_loops): Move loop verification...
16518 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
16519 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
16520 Set it to true when we're removing a loop from hierarchy tree in
16521 an irreducible region.
16522 (fix_bb_placements): Adjust caller.
16523 (fix_loop_placements): Likewise.
16525 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
16527 * config/avr/builtins.def (DEF_BUILTIN): Factor out
16528 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
16529 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
16530 Remove ID. Adjust comments.
16531 * config/avr/avr-c.c (avr_builtin_name): Remove.
16532 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
16533 * config/avr/avr.c (avr_tolower): New static function.
16534 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
16535 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
16536 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
16539 2013-01-17 Jan Hubicka <jh@suse.cz>
16541 PR tree-optimization/55273
16542 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
16544 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
16547 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
16548 store through atomic_store<mode>_1.
16549 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
16551 2013-01-17 Martin Jambor <mjambor@suse.cz>
16553 PR tree-optimizations/55264
16554 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
16555 for virtual methods.
16556 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
16557 virtual methods before inlining is over.
16558 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
16560 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
16563 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
16565 PR rtl-optimization/56005
16566 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
16567 pending reads for prefetch.
16569 2013-01-16 Ian Bolton <ian.bolton@arm.com>
16571 * config/aarch64/aarch64.md
16572 (*cstoresi_neg_uxtw): New pattern.
16573 (*cmovsi_insn_uxtw): New pattern.
16574 (*<optab>si3_uxtw): New pattern.
16575 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
16576 (*<optab>si3_insn_uxtw): New pattern.
16577 (*bswapsi2_uxtw): New pattern.
16579 2013-01-16 Richard Biener <rguenther@suse.de>
16581 * tree-inline.c (tree_function_versioning): Remove set but
16582 never used variable.
16584 2013-01-16 Richard Biener <rguenther@suse.de>
16586 PR tree-optimization/55964
16587 * tree-flow.h (rename_variables_in_loop): Remove.
16588 (rename_variables_in_bb): Likewise.
16589 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
16590 (copy_loop_before): Adjust and delete update-ssa status.
16591 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
16592 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
16593 (rename_variables_in_loop): Remove.
16594 (slpeel_update_phis_for_duplicate_loop): Likewise.
16595 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
16596 use available cfg machinery instead of duplicating it.
16597 Update PHI nodes and perform poor-mans SSA update here.
16598 (slpeel_tree_peel_loop_to_edge): Adjust.
16600 2013-01-16 Richard Biener <rguenther@suse.de>
16602 PR tree-optimization/54767
16603 PR tree-optimization/53465
16604 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
16605 (vrp_visit_phi_node): For PHI arguments coming via backedges
16606 drop all symbolical range information.
16607 (execute_vrp): Compute backedges.
16609 2013-01-16 Richard Biener <rguenther@suse.de>
16611 * doc/install.texi: Update CLooG and ISL requirements to
16614 2013-01-16 Christian Bruel <christian.bruel@st.com>
16617 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
16618 (broken_move): Handle UNSPECV_SP_SWITCH_B.
16619 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
16621 2013-01-16 DJ Delorie <dj@redhat.com>
16623 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
16624 (UNSPECV_SP_SWITCH_E): New.
16625 (sp_switch_1): Change to an unspec.
16626 (sp_switch_2): Change to an unspec. Don't use post-inc when we
16629 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
16631 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
16632 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
16633 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
16634 (expand_mem_thread_fence): Ditto.
16635 (expand_mem_signal_fence): Ditto.
16636 (expand_atomic_load): Ditto.
16637 (expand_atomic_store): Ditto.
16639 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
16641 PR rtl-optimization/55547
16642 PR rtl-optimization/53827
16645 * alias.c (memrefs_conflict_p): Set sizes to negative after
16648 2013-01-15 Jakub Jelinek <jakub@redhat.com>
16651 * function.c (thread_prologue_and_epilogue_insns): Always
16652 add crtl->drap_reg to set_up_by_prologue.set, even if
16653 stack_realign_drap is false.
16655 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16657 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
16658 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
16659 *call): Fix indention.
16661 2013-01-15 Tom de Vries <tom@codesourcery.com>
16664 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
16667 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
16669 PR rtl-optimization/55153
16670 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
16672 2013-01-15 Martin Jambor <mjambor@suse.cz>
16674 PR tree-optimization/55920
16675 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
16676 accesses as grp_to_be_debug_replaced.
16678 2013-01-15 Jakub Jelinek <jakub@redhat.com>
16680 PR tree-optimization/55920
16681 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
16682 there is non-useless type conversion needed from debug rhs to lhs,
16683 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
16685 2013-01-15 Joseph Myers <joseph@codesourcery.com>
16686 Mikael Pettersson <mikpe@it.uu.se>
16689 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
16691 (ASM_OUTPUT_CASE_LABEL): Remove.
16692 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
16693 * final.c (shorten_branches): Update alignment of labels before
16694 jump tables if CASE_VECTOR_SHORTEN_MODE.
16696 2013-01-15 Richard Biener <rguenther@suse.de>
16699 * system.h: Do not include gmp.h for building host tools.
16701 2013-01-15 Richard Biener <rguenther@suse.de>
16703 PR middle-end/55882
16704 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
16705 account for bitpos when computing alignment.
16707 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
16709 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
16710 (ix86_target_macros_internal): Likewise.
16712 * config/i386/i386.c (m_CORE2I7): Removed.
16713 (m_CORE_HASWELL): New macro.
16714 (m_CORE_ALL): Likewise.
16715 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
16716 (initial_ix86_arch_features): Likewise.
16717 (processor_target_table): Initializations for Core avx2.
16718 (cpu_names): New names "core-avx2".
16719 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
16720 PROCESSOR_CORE_HASWELL.
16721 (ix86_issue_rate): New case.
16722 (ia32_multipass_dfa_lookahead): Likewise.
16723 (ix86_sched_init_global): Likewise.
16725 * config/i386/i386.h (TARGET_HASWELL): New macro.
16726 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
16727 (processor_type): New PROCESSOR_HASWELL.
16729 2013-01-15 Jakub Jelinek <jakub@redhat.com>
16731 PR tree-optimization/55955
16732 * tree-vect-loop.c (vectorizable_reduction): Give up early on
16733 *SHIFT_EXPR and *ROTATE_EXPR codes.
16735 PR tree-optimization/48766
16736 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
16737 -ftrapv disable -fwrapv.
16739 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16742 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
16743 etc. to 1 and not to __flash.
16744 Use LL suffix for __INT24_MAX__ with -mint8.
16745 Use ULL suffix for __UINT24_MAX__ with -mint8.
16747 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16749 * config/avr/avr-arch.h
16750 (struct base_arch_s): Use typedef avr_arch_t instead.
16751 (struct arch_info_s): Use typedef avr_arch_info_t instead.
16752 (struct mcu_type_s): Use typedef avr_mcu_t instead.
16753 * config/avr/avr.c: Same.
16754 * config/avr/avr-devices.c: Same.
16755 * config/avr/driver-avr.c: Same.
16756 * config/avr/gen-avr-mmcu-texi.c: Same.
16757 * config/avr/avr-mcus.def: Adjust comment.
16759 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
16761 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
16762 * config/aarch64/iterators.md (VALLDI): New.
16764 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
16765 Andi Kleen <ak@linux.intel.com>
16768 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
16769 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
16772 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16774 * config/avr/avr-stdint.h: Remove trailing blanks.
16775 * config/avr/avr-log.h: Same.
16776 * config/avr/avr-arch.h: Same.
16777 * config/avr/avr-devices.c: Same.
16778 * config/avr/avr-dimode.md: Same.
16779 * config/avr/predicates.md: Same.
16780 * config/avr/avr-c.c: Same. And fix typo.
16782 * config/avr/avr-protos.h: Same. And:
16783 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
16784 (init_cumulative_args): Rename to avr_init_cumulative_args.
16785 (expand_prologue): Rename to avr_expand_prologue.
16786 (expand_epilogue): Rename to avr_expand_epilogue.
16787 (adjust_insn_length): Rename to avr_adjust_insn_length.
16788 (notice_update_cc): Rename to avr_notice_update_cc.
16789 (final_prescan_insn): Rename to avr_final_prescan_insn.
16790 * config/avr/avr.c: Same.
16791 * config/avr/avr.h: Same.
16792 * config/avr/avr.md: Remove trailing blanks.
16793 (prologue): Use avr_expand_prologue.
16794 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
16796 2013-01-14 Richard Biener <rguenther@suse.de>
16798 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
16799 verify_location, collect_subblocks): New functions.
16800 (verify_gimple_in_cfg): Verify that locations only reference
16801 BLOCKs in the functions BLOCK tree.
16803 2013-01-14 Richard Biener <rguenther@suse.de>
16805 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
16807 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
16809 (insert_out_of_ssa_copy_on_edge): Likewise.
16810 (rewrite_close_phi_out_of_ssa): Likewise.
16811 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
16813 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
16814 propagated constants.
16815 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
16818 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
16820 * config/avr/avr-modes.def: Add GPL copyright notice.
16822 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
16824 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
16825 MEMMODEL_MASK to determine memory model.
16826 (atomic_store<mode>): Ditto from operands[2].
16827 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
16829 2013-01-13 Jakub Jelinek <jakub@redhat.com>
16832 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
16833 (fold_gimple_assign): Don't call unshare_expr here.
16834 (fold_ctor_reference): Call unshare_expr.
16836 2013-01-13 Terry Guo <terry.guo@arm.com>
16838 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
16839 * doc/fragments.texi: Document MULTILIB_REUSE.
16840 * gcc.c (multilib_reuse): New internal spec.
16841 (set_multilib_dir): Also search multilib from multilib_reuse.
16842 * genmultilib (tmpmultilib3): Refactor code.
16843 (tmpmultilib4): Ditto.
16844 (multilib_reuse): New multilib argument.
16846 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
16848 * Makefile.in: Update copyright.
16850 2013-01-12 Tom de Vries <tom@codesourcery.com>
16852 PR middle-end/55890
16853 * calls.c (expand_call): Check if arg_nr is valid.
16855 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
16857 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
16858 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
16859 documentation. Add missing '__' in front of
16860 __builtin_ia32_packssdw256.
16862 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16865 * config/s390/s390.c (s390_preferred_reload_class): Do not return
16866 NO_REGS for larl operands.
16867 (s390_reload_larl_operand): Use s390_load_address instead of
16870 2013-01-11 Richard Biener <rguenther@suse.de>
16872 * tree-cfg.c (verify_node_sharing_1): Split out from ...
16873 (verify_node_sharing): ... here.
16874 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
16876 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
16878 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
16879 Substitute TREECHECKING.
16880 * configure: Regenerate.
16881 * Makefile.in (TREECHECKING): New.
16883 2013-01-11 Richard Guenther <rguenther@suse.de>
16885 PR tree-optimization/44061
16886 * tree-vrp.c (extract_range_basic): Compute zero as
16887 value-range for __builtin_constant_p of function parameters.
16889 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
16891 Update copyright years.
16893 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
16895 PR rtl-optimization/55672
16896 * lra-eliminations.c (mark_not_eliminable): Permit addition with
16897 const to be eliminable.
16899 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
16901 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
16902 * configure: Regenerate.
16904 2013-01-10 Richard Biener <rguenther@suse.de>
16906 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
16908 2013-01-10 Richard Biener <rguenther@suse.de>
16911 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
16912 locations for virtual PHI arguments.
16913 (rewrite_update_phi_arguments): Likewise.
16915 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
16917 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
16920 2013-01-10 Jakub Jelinek <jakub@redhat.com>
16922 PR tree-optimization/55921
16923 * tree-complex.c (expand_complex_asm): New function.
16924 (expand_complex_operations_1): Call it for GIMPLE_ASM.
16926 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16929 * config/s390/s390.c (s390_symref_operand_p)
16930 (s390_loadrelative_operand_p): Merge the two functions.
16931 (s390_check_qrst_address, print_operand_address): Add parameters
16932 to s390_loadrelative_operand_p invokation.
16933 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
16934 (s390_reload_larl_operand, s390_secondary_reload): Use
16935 s390_loadrelative_operand_p instead of s390_symref_operand_p.
16936 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
16938 2013-01-09 Mike Stump <mikestump@comcast.net>
16940 * dse.c (record_store): Remove unnecessary assert.
16942 2013-01-09 Jan Hubicka <jh@suse.cz>
16944 PR tree-optimization/55569
16945 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
16946 * cfgloop.h (scale_loop_profile): Likewise.
16948 2013-01-09 Jan Hubicka <jh@suse.cz>
16951 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
16953 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
16955 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
16957 PR middle-end/55114
16958 * expr.h (maybe_emit_group_store): Declare.
16959 * expr.c (maybe_emit_group_store): New function.
16960 * builtins.c (expand_builtin_int_roundingfn): Call it.
16961 (expand_builtin_int_roundingfn_2): Likewise.
16963 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
16965 PR rtl-optimization/55829
16966 * lra-constraints.c (match_reload): Add code for absent output.
16967 (curr_insn_transform): Add code for reloads of matched inputs
16970 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
16972 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
16973 attribute of movddup insn to DF.
16974 (*vec_interleave_lowv2df): Ditto.
16975 (vec_dupv2df): Ditto.
16977 2013-01-09 Jan Hubicka <jh@suse.cz>
16979 PR tree-optimiation/55875
16980 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
16981 EVERY_ITERATION parameter.
16982 (number_of_iterations_exit): Check if exit is executed every iteration.
16983 (idx_infer_loop_bounds): Similarly here.
16984 (n_of_executions_at_most): Simplify
16985 to only test for cases where statement is dominated by the
16986 particular bound; handle correctly the "postdominance" test.
16987 (scev_probably_wraps_p): Use max loop iterations info
16988 as a global bound first.
16990 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
16991 Nick Clifton <nickc@redhat.com>
16993 * config/v850/v850.md (cbranchsf4): New pattern.
16994 (cstoresf4): New pattern.
16995 (cbranchdf4): New pattern.
16996 (cstoredf4): New pattern.
16997 (movsicc): Disallow floating point comparisons.
16998 (cmpsf_le_insn): Fix order of operators.
16999 (cmpsf_lt_insn): Likewise.
17000 (cmpsf_eq_insn): Likewise.
17001 (cmpdf_le_insn): Likewise.
17002 (cmpdf_lt_insn): Likewise.
17003 (cmpdf_eq_insn): Likewise.
17004 (cmpsf_ge_insn): Use LE comparison.
17005 (cmpdf_ge_insn): Likewise.
17006 (cmpsf_gt_insn): Use LT comparison.
17007 (cmpdf_gt_insn): Likewise.
17008 (cmpsf_ne_insn): Delete pattern.
17009 (cmpdf_ne_insn): Delete pattern.
17010 * config/v850/v850.c (v850_gen_float_compare): Use
17011 gen_cmpdf_eq_insn for NE comparison.
17012 (v850_float_z_comparison_operator)
17013 (v850_float_nz_comparison_operator): Move from here ...
17014 * config/v850/predicates.md: ... to here. Move GT and GE
17015 comparisons into v850_float_z_comparison_operator.
17016 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
17018 (v850_float_nz_comparison_operator): Likewise.
17020 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17022 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
17023 with calls to gen_insvsi/gen_insvdi.
17025 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17027 * config/i386/i386.c (initial_ix86_tune_features): Set up
17028 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17030 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17031 Jakub Jelinek <jakub@redhat.com>
17033 PR tree-optimization/48189
17034 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17035 If nitercst is 0, don't predict the exit edge.
17037 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17039 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17040 in asm_fprintf with reg_names.
17041 (aarch64_print_operand_address): Likewise.
17042 (aarch64_return_addr): Likewise.
17043 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17045 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17047 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17048 (INT_U6_BITS): Likewise.
17049 * config/pa/predicates.md (uint6_operand): New predicate.
17050 (shift5_operand, shift6_operand): Likewise.
17051 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17053 (lshrdi3): Use shift6_operand.
17054 (shrpsi4, shrpdi4): New insn patterns.
17055 (extzv): Delete expander.
17056 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17057 predicates in unamed zero extract patterns. Tighten common constraint.
17058 (extv): Delete expander.
17059 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17060 predicates in unamed sign extract patterns. Tighten common constraint.
17061 (insv): Delete expander.
17062 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17063 predicates in unamed insert patterns. Tighten common constraint.
17064 Change uint32_operand predicate to uint6_operand predicate in unamed
17065 DImode pattern to insert constant values of type 1...1xxxx.
17067 2013-01-04 Jan Hubicka <jh@suse.cz>
17069 PR tree-optimization/55823
17070 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17073 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17074 Uros Bizjak <ubizjak@gmail.com>
17076 PR rtl-optimization/55845
17077 * df-problems.c (can_move_insns_across): Stop scanning at
17078 volatile_insn_p source instruction or give up if
17079 across_from .. across_to range contains any volatile_insn_p
17082 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17084 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17085 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17087 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
17088 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
17090 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17093 * asan.c (asan_clear_shadow): New function.
17094 (asan_emit_stack_protection): Use it.
17096 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17098 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17099 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17100 with tab instead of space.
17102 2013-01-08 Nick Clifton <nickc@redhat.com>
17104 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17105 register bank 0 at the start of an interrupt handler.
17106 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17109 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17111 * config/aarch64/aarch64-simd.md
17112 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17113 (aarch64_simd_bsl): Likewise.
17114 (aarch64_vcond_internal<mode>): Likewise.
17115 (vcond<mode><mode>): Likewise.
17116 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17117 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17119 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17121 * config/aarch64/aarch64-builtins.c
17122 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17124 2013-01-08 Martin Jambor <mjambor@suse.cz>
17127 * tree-sra.c (analyze_access_subtree): Return true also after
17128 potentially creating a debug-only replacement.
17130 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17132 PR middle-end/55890
17133 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17135 PR tree-optimization/54120
17136 * tree-vrp.c (range_fits_type_p): Don't allow
17137 src_precision < precision from signed vr to unsigned_p
17138 if vr->min or vr->max is negative.
17139 (simplify_float_conversion_using_ranges): Test can_float_p
17140 against CODE_FOR_nothing.
17142 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17143 Richard Biener <rguenther@suse.de>
17145 PR middle-end/55851
17146 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17147 types instead of just INTEGER_TYPE types.
17149 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
17151 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17154 2013-01-07 Steve Ellcey <sellcey@mips.com>
17157 * config/mips/mips.opt: Change mad to mmad to match documentation.
17159 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17162 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17163 .progmemx.data now.
17165 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17168 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17169 (avr_addrspace_t): Add .section_name field.
17170 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17172 (avr_addrspace): Same. Initialize .section_name. Remove last
17173 NULL entry. Put __memx into .progmemx.data.
17174 (progmem_section_prefix): Remove.
17175 (avr_asm_init_sections): No need to initialize progmem_section.
17176 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17177 section name prefix.
17178 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17179 retrieve the progmem section.
17180 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17181 boundary to run over avr_addrspace[].
17182 (avr_register_target_pragmas): Ditto.
17184 2013-01-06 Jakub Jelinek <jakub@redhat.com>
17186 * varasm.c (output_constant_def_contents): For asan_protect_global
17187 protected strings, adjust DECL_ALIGN if needed, before testing for
17189 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17190 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17192 (output_object_block): For asan protected decls, emit asan padding
17193 after their contents.
17194 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17195 (asan_finish_file): Test it here instead.
17197 2013-01-07 Nick Clifton <nickc@redhat.com>
17198 Matthias Klose <doko@debian.org>
17199 Doug Kwan <dougkwan@google.com>
17200 H.J. Lu <hongjiu.lu@intel.com>
17203 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17205 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17207 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17209 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
17211 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17213 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17216 * doc/install.texi (Cross-Compiler-Specific Options): Document
17219 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17221 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17222 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17223 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17224 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17225 vqmovun_high_s64): Fix source operand number and update copyright.
17227 2013-01-07 Richard Biener <rguenther@suse.de>
17229 PR middle-end/55890
17230 * gimple.h (gimple_call_builtin_p): New overload.
17231 * gimple.c (validate_call): New function.
17232 (gimple_call_builtin_p): Likewise.
17233 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17234 Use gimple_call_builtin_p.
17235 (find_func_clobbers): Likewise.
17236 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17237 (strlen_optimize_stmt): Likewise.
17239 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17241 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17242 (vld1q_dup_*): Likewise.
17243 (vld1_*): Likewise.
17244 (vld1q_*): Likewise.
17245 (vld1_lane_*): Likewise.
17246 (vld1q_lane_*): Likewise.
17248 2013-01-07 Richard Biener <rguenther@suse.de>
17250 * lto-streamer.h (LTO_minor_version): Bump to 2.
17252 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17254 * config/aarch64/aarch64-protos.h
17255 (aarch64_const_double_zero_rtx_p): Rename to...
17256 (aarch64_float_const_zero_rtx_p): ...this.
17257 (aarch64_float_const_representable_p): New.
17258 (aarch64_output_simd_mov_immediate): Likewise.
17259 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17260 move immediate case.
17261 * config/aarch64/aarch64.c
17262 (aarch64_const_double_zero_rtx_p): Rename to...
17263 (aarch64_float_const_zero_rtx_p): ...this.
17264 (aarch64_print_operand): Allow printing of new constants.
17265 (aarch64_valid_floating_const): New.
17266 (aarch64_legitimate_constant_p): Check for valid floating-point
17268 (aarch64_simd_valid_immediate): Likewise.
17269 (aarch64_vect_float_const_representable_p): New.
17270 (aarch64_float_const_representable_p): Likewise.
17271 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17272 (aarch64_output_simd_mov_immediate): New.
17273 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17274 (*movdf_aarch64): Likewise.
17275 * config/aarch64/constraints.md (Ufc): New.
17276 (Y): call aarch64_float_const_zero_rtx.
17277 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17279 2013-01-07 Richard Biener <rguenther@suse.de>
17281 PR tree-optimization/55888
17282 PR tree-optimization/55862
17283 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17284 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17285 not if it is contained therein.
17287 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17289 * config/avr/t-avr: Typo.
17291 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17294 * config/avr/t-avr: Don't automatically rebuild
17295 $(srcdir)/config/avr/t-multilib
17296 $(srcdir)/config/avr/avr-tables.opt
17297 $(srcdir)/doc/avr-mmcu.texi
17298 (avr-mcus): New phony target to build them on request.
17299 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17300 * avr/avr-mcus.def: Adjust comments.
17302 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
17304 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17306 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17308 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17310 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17312 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17314 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
17316 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17317 to generate profiling.
17318 * config/rs6000/aix64.h (LIB_SPEC): Same.
17320 2013-01-04 Andrew Pinski <apinski@cavium.com>
17322 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17324 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17326 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
17328 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17330 (ix86_expand_move): Ditto.
17331 (ix86_zero_extend_to_Pmode): Ditto.
17332 (ix86_expand_call): Ditto.
17333 (ix86_expand_special_args_builtin): Ditto.
17334 (ix86_expand_builtin): Ditto.
17336 2013-01-04 Richard Biener <rguenther@suse.de>
17338 PR tree-optimization/55862
17339 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17340 translating them through PHI nodes.
17342 2013-01-04 Martin Jambor <mjambor@suse.cz>
17344 PR tree-optimization/55755
17345 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17346 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17348 2013-01-04 Marek Polacek <polacek@redhat.com>
17350 PR middle-end/55859
17351 * opts.c (default_options_optimization): Clarify error message.
17353 2013-01-04 Richard Biener <rguenther@suse.de>
17355 PR middle-end/55863
17356 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17359 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17362 * config/pa/pa.md (movsi): Revert previous change.
17363 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17366 2013-01-03 Richard Henderson <rth@redhat.com>
17368 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17369 after eliminating TLS symbols.
17371 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
17374 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
17375 * graphite-poly.c (debug_gmp_value): Likewise.
17377 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
17380 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
17381 selected code model, define __code_mode_small__, __code_model_medium__,
17382 __code_model_large__, __code_model_32__ or __code_model_kernel__.
17383 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
17384 xchg temporary register with %k. Declare temporary register as
17386 [__x86_64__]: For medium and large code models, preserve %rbx register.
17388 2013-01-03 Richard Biener <rguenther@suse.de>
17390 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
17391 (dump_subscript): Adjust.
17392 (finalize_ddr_dependent): Do not dump redundant info.
17393 (analyze_siv_subscript): Adjust.
17394 (subscript_dependence_tester): Likewise.
17395 (compute_affine_dependence): Likewise.
17397 2013-01-03 Richard Biener <rguenther@suse.de>
17400 2013-01-03 Richard Biener <rguenther@suse.de>
17402 PR tree-optimization/55857
17403 * tree-vect-stmts.c (vectorizable_load): Do not setup
17404 re-alignment for invariant loads.
17406 2013-01-02 Richard Biener <rguenther@suse.de>
17408 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17409 invariant load do not generate a vector load from the scalar location.
17411 2013-01-03 Richard Biener <rguenther@suse.de>
17413 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
17414 for not vectorizing.
17415 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
17416 not build INDIRECT_REFs, call get_name once only.
17417 (vect_create_data_ref_ptr): Likewise. Dump base object kind
17418 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
17420 2013-01-03 Richard Biener <rguenther@suse.de>
17422 PR tree-optimization/55857
17423 * tree-vect-stmts.c (vectorizable_load): Do not setup
17424 re-alignment for invariant loads.
17426 2013-01-03 Richard Biener <rguenther@suse.de>
17429 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
17430 prefer a built-in decl.
17432 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17434 * gcc.c (process_command): Update copyright notice dates.
17435 * gcov.c (print_version): Likewise.
17436 * gcov-dump.c (print_version): Likewise.
17438 PR rtl-optimization/55838
17439 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
17440 iv0.step, iv1.step and step.
17442 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17443 Marc Glisse <marc.glisse@inria.fr>
17445 PR tree-optimization/55832
17446 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
17447 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
17448 integer_{one,zero}_node.
17450 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17453 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
17454 * var-tracking.c (reverse_op): Don't add reverse ops to
17455 VALUEs that have already
17456 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
17458 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
17460 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
17462 2013-01-02 Teresa Johnson <tejohnson@google.com>
17464 * dumpfile.c (dump_loc): Print filename with location.
17465 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
17466 new location_t parameter to emit complete unroll message with
17467 new dump framework.
17468 (canonicalize_loop_induction_variables): Compute loops location
17469 and pass to try_unroll_loop_completely.
17470 * loop-unroll.c (report_unroll_peel): New function.
17471 (peel_loops_completely): Use new dump format with location
17472 for main dumpfile message, and invoke report_unroll_peel on success.
17473 (decide_unrolling_and_peeling): Ditto.
17474 (decide_peel_once_rolling): Remove old dumpfile message subsumed
17475 by report_unroll_peel.
17476 (decide_peel_completely): Ditto.
17477 (decide_unroll_constant_iterations): Ditto.
17478 (decide_unroll_runtime_iterations): Ditto.
17479 (decide_peel_simple): Ditto.
17480 (decide_unroll_stupid): Ditto.
17481 * cfgloop.c (get_loop_location): New function.
17482 * cfgloop.h (get_loop_location): Declare.
17484 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17486 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
17489 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17491 PR middle-end/55198
17492 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
17493 BLKmode objects when EXPAND_MEMORY is specified.
17495 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17497 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
17499 (fold_builtin_cpu): Do not share cpu model decls across statements.
17501 2013-01-02 Jason Merrill <jason@redhat.com>
17504 * tree.c (build_array_type_1): Revert earlier change.
17506 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
17508 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
17510 * config/aarch64/aarch64-tune.md: Re-generate.
17512 2013-01-02 Richard Biener <rguenther@suse.de>
17514 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17515 invariant load do not generate a vector load from the scalar location.
17517 2013-01-02 Richard Biener <rguenther@suse.de>
17520 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
17521 * configure: Regenerate.
17523 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17525 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
17526 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
17527 (expand_builtin_int_roundingfn_2): Keep the original target around
17528 for the fallback case.
17530 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17532 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
17533 to be clear for sign changes.
17535 2013-01-01 Jan Hubicka <jh@suse.cz>
17537 * ipa-inline-analysis.c: Fix formatting.
17539 2013-01-01 Jakub Jelinek <jakub@redhat.com>
17541 PR tree-optimization/55831
17542 * tree-vect-loop.c (get_initial_def_for_induction): Use
17543 gsi_after_labels instead of gsi_start_bb.
17545 Copyright (C) 2013 Free Software Foundation, Inc.
17547 Copying and distribution of this file, with or without modification,
17548 are permitted in any medium without royalty provided the copyright
17549 notice and this notice are preserved.