]> gcc.gnu.org Git - gcc.git/blame - gcc/ChangeLog
hashtable_policy.h (_Hashtable_alloc): New.
[gcc.git] / gcc / ChangeLog
CommitLineData
eb923e54
OE
12013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
2
3 PR other/12081
4 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5 class insn_gen_fn.
6 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
7 rtx (*) (rtx, ...) with insn_gen_fn.
8 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9 insn_gen_fn::stored_funcptr. Add initializer braces.
10
8ac69a6c
DM
112013-08-05 David Malcolm <dmalcolm@redhat.com>
12
13 Rewrite how instances of passes are cloned to remove assumptions
14 about their sizes (thus allowing pass subclasses to have
15 additional data fields, albeit non-GC-managed ones at this point).
16
17 * passes.c (make_pass_instance): Now that passes have clone
18 methods, rewrite this function to eliminate XNEW and memcpy
19 calls that used hardcoded sizes. Since this function no longer
20 creates pass instances, rename it to...
21 (add_pass_instance): ...this. Document the old way that passes
22 were numbered and flagged, and rework this function to continue
23 using it.
24 (next_pass_1): Add an initial_pass argument for use by
25 add_pass_instance.
26 (position_pass): When adding multiple instances of a pass, use
27 the pass's clone method, rather than relying on the XNEW/memcpy
28 within the former make_pass_instance (now add_pass_instance).
29 (pass_manager::pass_manager): When invoking next_pass_1, also
30 supply the initial instance of the current pass within the
31 pass manager.
32
27a4cd48
DM
332013-08-05 David Malcolm <dmalcolm@redhat.com>
34
35 This is the automated part of the conversion of passes from C
36 structs to C++ classes.
37
38 Patch autogenerated by refactor_passes.py from
39 https://github.com/davidmalcolm/gcc-refactoring-scripts
40 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
41
42 * asan.c (pass_asan): Convert from a global struct to a subclass of
43 gimple_opt_pass along with...
44 (pass_data_asan): ...new pass_data instance and...
45 (make_pass_asan): ...new function.
46 (pass_asan_O0): Convert from a global struct to a subclass of
47 gimple_opt_pass along with...
48 (pass_data_asan_O0): ...new pass_data instance and...
49 (make_pass_asan_O0): ...new function.
50 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
51 subclass of rtl_opt_pass along with...
52 (pass_data_inc_dec): ...new pass_data instance and...
53 (make_pass_inc_dec): ...new function.
54 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
55 a subclass of rtl_opt_pass along with...
56 (pass_data_reorder_blocks): ...new pass_data instance and...
57 (make_pass_reorder_blocks): ...new function.
58 (pass_duplicate_computed_gotos): Convert from a global struct to a
59 subclass of rtl_opt_pass along with...
60 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
61 (make_pass_duplicate_computed_gotos): ...new function.
62 (pass_partition_blocks): Convert from a global struct to a subclass of
63 rtl_opt_pass along with...
64 (pass_data_partition_blocks): ...new pass_data instance and...
65 (make_pass_partition_blocks): ...new function.
66 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
67 struct to a subclass of rtl_opt_pass along with...
68 (pass_data_branch_target_load_optimize1): ...new pass_data instance
69 and...
70 (make_pass_branch_target_load_optimize1): ...new function.
71 (pass_branch_target_load_optimize2): Convert from a global struct to a
72 subclass of rtl_opt_pass along with...
73 (pass_data_branch_target_load_optimize2): ...new pass_data instance
74 and...
75 (make_pass_branch_target_load_optimize2): ...new function.
76 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
77 of rtl_opt_pass along with...
78 (pass_data_jump): ...new pass_data instance and...
79 (make_pass_jump): ...new function.
80 (pass_jump2): Convert from a global struct to a subclass of
81 rtl_opt_pass along with...
82 (pass_data_jump2): ...new pass_data instance and...
83 (make_pass_jump2): ...new function.
84 * cfgexpand.c (pass_expand): Convert from a global struct to a
85 subclass of rtl_opt_pass along with...
86 (pass_data_expand): ...new pass_data instance and...
87 (make_pass_expand): ...new function.
88 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
89 of rtl_opt_pass along with...
90 (pass_data_free_cfg): ...new pass_data instance and...
91 (make_pass_free_cfg): ...new function.
92 (pass_into_cfg_layout_mode): Convert from a global struct to a
93 subclass of rtl_opt_pass along with...
94 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
95 (make_pass_into_cfg_layout_mode): ...new function.
96 (pass_outof_cfg_layout_mode): Convert from a global struct to a
97 subclass of rtl_opt_pass along with...
98 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
99 (make_pass_outof_cfg_layout_mode): ...new function.
100 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
101 struct to a subclass of gimple_opt_pass along with...
102 (pass_data_build_cgraph_edges): ...new pass_data instance and...
103 (make_pass_build_cgraph_edges): ...new function.
104 (pass_rebuild_cgraph_edges): Convert from a global struct to a
105 subclass of gimple_opt_pass along with...
106 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
107 (make_pass_rebuild_cgraph_edges): ...new function.
108 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
109 subclass of gimple_opt_pass along with...
110 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
111 and...
112 (make_pass_remove_cgraph_callee_edges): ...new function.
113 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
114 struct to a subclass of rtl_opt_pass along with...
115 (pass_data_stack_adjustments): ...new pass_data instance and...
116 (make_pass_stack_adjustments): ...new function.
117 * combine.c (pass_combine): Convert from a global struct to a subclass
118 of rtl_opt_pass along with...
119 (pass_data_combine): ...new pass_data instance and...
120 (make_pass_combine): ...new function.
121 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
122 global struct to a subclass of rtl_opt_pass along with...
123 (pass_data_compare_elim_after_reload): ...new pass_data instance
124 and...
125 (make_pass_compare_elim_after_reload): ...new function.
126 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
127 of rtl_opt_pass along with...
128 (pass_data_rtl_cprop): ...new pass_data instance and...
129 (make_pass_rtl_cprop): ...new function.
130 * cse.c (pass_cse): Convert from a global struct to a subclass of
131 rtl_opt_pass along with...
132 (pass_data_cse): ...new pass_data instance and...
133 (make_pass_cse): ...new function.
134 (pass_cse2): Convert from a global struct to a subclass of
135 rtl_opt_pass along with...
136 (pass_data_cse2): ...new pass_data instance and...
137 (make_pass_cse2): ...new function.
138 (pass_cse_after_global_opts): Convert from a global struct to a
139 subclass of rtl_opt_pass along with...
140 (pass_data_cse_after_global_opts): ...new pass_data instance and...
141 (make_pass_cse_after_global_opts): ...new function.
142 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
143 of rtl_opt_pass along with...
144 (pass_data_ud_rtl_dce): ...new pass_data instance and...
145 (make_pass_ud_rtl_dce): ...new function.
146 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
147 rtl_opt_pass along with...
148 (pass_data_fast_rtl_dce): ...new pass_data instance and...
149 (make_pass_fast_rtl_dce): ...new function.
150 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
151 a subclass of rtl_opt_pass along with...
152 (pass_data_df_initialize_opt): ...new pass_data instance and...
153 (make_pass_df_initialize_opt): ...new function.
154 (pass_df_initialize_no_opt): Convert from a global struct to a
155 subclass of rtl_opt_pass along with...
156 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
157 (make_pass_df_initialize_no_opt): ...new function.
158 (pass_df_finish): Convert from a global struct to a subclass of
159 rtl_opt_pass along with...
160 (pass_data_df_finish): ...new pass_data instance and...
161 (make_pass_df_finish): ...new function.
162 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
163 rtl_opt_pass along with...
164 (pass_data_rtl_dse1): ...new pass_data instance and...
165 (make_pass_rtl_dse1): ...new function.
166 (pass_rtl_dse2): Convert from a global struct to a subclass of
167 rtl_opt_pass along with...
168 (pass_data_rtl_dse2): ...new pass_data instance and...
169 (make_pass_rtl_dse2): ...new function.
170 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
171 subclass of rtl_opt_pass along with...
172 (pass_data_dwarf2_frame): ...new pass_data instance and...
173 (make_pass_dwarf2_frame): ...new function.
174 * except.c (pass_set_nothrow_function_flags): Convert from a global
175 struct to a subclass of rtl_opt_pass along with...
176 (pass_data_set_nothrow_function_flags): ...new pass_data instance
177 and...
178 (make_pass_set_nothrow_function_flags): ...new function.
179 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
180 subclass of rtl_opt_pass along with...
181 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
182 and...
183 (make_pass_convert_to_eh_region_ranges): ...new function.
184 * final.c (pass_compute_alignments): Convert from a global struct to a
185 subclass of rtl_opt_pass along with...
186 (pass_data_compute_alignments): ...new pass_data instance and...
187 (make_pass_compute_alignments): ...new function.
188 (pass_final): Convert from a global struct to a subclass of
189 rtl_opt_pass along with...
190 (pass_data_final): ...new pass_data instance and...
191 (make_pass_final): ...new function.
192 (pass_shorten_branches): Convert from a global struct to a subclass of
193 rtl_opt_pass along with...
194 (pass_data_shorten_branches): ...new pass_data instance and...
195 (make_pass_shorten_branches): ...new function.
196 (pass_clean_state): Convert from a global struct to a subclass of
197 rtl_opt_pass along with...
198 (pass_data_clean_state): ...new pass_data instance and...
199 (make_pass_clean_state): ...new function.
200 * function.c (pass_instantiate_virtual_regs): Convert from a global
201 struct to a subclass of rtl_opt_pass along with...
202 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
203 (make_pass_instantiate_virtual_regs): ...new function.
204 (pass_leaf_regs): Convert from a global struct to a subclass of
205 rtl_opt_pass along with...
206 (pass_data_leaf_regs): ...new pass_data instance and...
207 (make_pass_leaf_regs): ...new function.
208 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
209 subclass of rtl_opt_pass along with...
210 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
211 and...
212 (make_pass_thread_prologue_and_epilogue): ...new function.
213 (pass_match_asm_constraints): Convert from a global struct to a
214 subclass of rtl_opt_pass along with...
215 (pass_data_match_asm_constraints): ...new pass_data instance and...
216 (make_pass_match_asm_constraints): ...new function.
217 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
218 subclass of rtl_opt_pass along with...
219 (pass_data_rtl_fwprop): ...new pass_data instance and...
220 (make_pass_rtl_fwprop): ...new function.
221 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
222 rtl_opt_pass along with...
223 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
224 (make_pass_rtl_fwprop_addr): ...new function.
225 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
226 rtl_opt_pass along with...
227 (pass_data_rtl_pre): ...new pass_data instance and...
228 (make_pass_rtl_pre): ...new function.
229 (pass_rtl_hoist): Convert from a global struct to a subclass of
230 rtl_opt_pass along with...
231 (pass_data_rtl_hoist): ...new pass_data instance and...
232 (make_pass_rtl_hoist): ...new function.
233 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
234 subclass of gimple_opt_pass along with...
235 (pass_data_lower_cf): ...new pass_data instance and...
236 (make_pass_lower_cf): ...new function.
237 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
238 from a global struct to a subclass of gimple_opt_pass along with...
239 (pass_data_strength_reduction): ...new pass_data instance and...
240 (make_pass_strength_reduction): ...new function.
241 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
242 of rtl_opt_pass along with...
243 (pass_data_rtl_ifcvt): ...new pass_data instance and...
244 (make_pass_rtl_ifcvt): ...new function.
245 (pass_if_after_combine): Convert from a global struct to a subclass of
246 rtl_opt_pass along with...
247 (pass_data_if_after_combine): ...new pass_data instance and...
248 (make_pass_if_after_combine): ...new function.
249 (pass_if_after_reload): Convert from a global struct to a subclass of
250 rtl_opt_pass along with...
251 (pass_data_if_after_reload): ...new pass_data instance and...
252 (make_pass_if_after_reload): ...new function.
253 * init-regs.c (pass_initialize_regs): Convert from a global struct to
254 a subclass of rtl_opt_pass along with...
255 (pass_data_initialize_regs): ...new pass_data instance and...
256 (make_pass_initialize_regs): ...new function.
257 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
258 of ipa_opt_pass_d along with...
259 (pass_data_ipa_cp): ...new pass_data instance and...
260 (make_pass_ipa_cp): ...new function.
261 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
262 global struct to a subclass of gimple_opt_pass along with...
263 (pass_data_inline_parameters): ...new pass_data instance and...
264 (make_pass_inline_parameters): ...new function.
265 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
266 subclass of gimple_opt_pass along with...
267 (pass_data_early_inline): ...new pass_data instance and...
268 (make_pass_early_inline): ...new function.
269 (pass_ipa_inline): Convert from a global struct to a subclass of
270 ipa_opt_pass_d along with...
271 (pass_data_ipa_inline): ...new pass_data instance and...
272 (make_pass_ipa_inline): ...new function.
273 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
274 struct to a subclass of gimple_opt_pass along with...
275 (pass_data_local_pure_const): ...new pass_data instance and...
276 (make_pass_local_pure_const): ...new function.
277 (pass_ipa_pure_const): Convert from a global struct to a subclass of
278 ipa_opt_pass_d along with...
279 (pass_data_ipa_pure_const): ...new pass_data instance and...
280 (make_pass_ipa_pure_const): ...new function.
281 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
282 to a subclass of ipa_opt_pass_d along with...
283 (pass_data_ipa_reference): ...new pass_data instance and...
284 (make_pass_ipa_reference): ...new function.
285 * ipa-split.c (pass_split_functions): Convert from a global struct to
286 a subclass of gimple_opt_pass along with...
287 (pass_data_split_functions): ...new pass_data instance and...
288 (make_pass_split_functions): ...new function.
289 (pass_feedback_split_functions): Convert from a global struct to a
290 subclass of gimple_opt_pass along with...
291 (pass_data_feedback_split_functions): ...new pass_data instance and...
292 (make_pass_feedback_split_functions): ...new function.
293 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
294 global struct to a subclass of simple_ipa_opt_pass along with...
295 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
296 instance and...
297 (make_pass_ipa_function_and_variable_visibility): ...new function.
298 (pass_ipa_free_inline_summary): Convert from a global struct to a
299 subclass of simple_ipa_opt_pass along with...
300 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
301 (make_pass_ipa_free_inline_summary): ...new function.
302 (pass_ipa_whole_program_visibility): Convert from a global struct to a
303 subclass of ipa_opt_pass_d along with...
304 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
305 and...
306 (make_pass_ipa_whole_program_visibility): ...new function.
307 (pass_ipa_profile): Convert from a global struct to a subclass of
308 ipa_opt_pass_d along with...
309 (pass_data_ipa_profile): ...new pass_data instance and...
310 (make_pass_ipa_profile): ...new function.
311 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
312 ipa_opt_pass_d along with...
313 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
314 (make_pass_ipa_cdtor_merge): ...new function.
315 * ira.c (pass_ira): Convert from a global struct to a subclass of
316 rtl_opt_pass along with...
317 (pass_data_ira): ...new pass_data instance and...
318 (make_pass_ira): ...new function.
319 (pass_reload): Convert from a global struct to a subclass of
320 rtl_opt_pass along with...
321 (pass_data_reload): ...new pass_data instance and...
322 (make_pass_reload): ...new function.
323 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
324 subclass of rtl_opt_pass along with...
325 (pass_data_cleanup_barriers): ...new pass_data instance and...
326 (make_pass_cleanup_barriers): ...new function.
327 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
328 of rtl_opt_pass along with...
329 (pass_data_loop2): ...new pass_data instance and...
330 (make_pass_loop2): ...new function.
331 (pass_rtl_loop_init): Convert from a global struct to a subclass of
332 rtl_opt_pass along with...
333 (pass_data_rtl_loop_init): ...new pass_data instance and...
334 (make_pass_rtl_loop_init): ...new function.
335 (pass_rtl_loop_done): Convert from a global struct to a subclass of
336 rtl_opt_pass along with...
337 (pass_data_rtl_loop_done): ...new pass_data instance and...
338 (make_pass_rtl_loop_done): ...new function.
339 (pass_rtl_move_loop_invariants): Convert from a global struct to a
340 subclass of rtl_opt_pass along with...
341 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
342 (make_pass_rtl_move_loop_invariants): ...new function.
343 (pass_rtl_unswitch): Convert from a global struct to a subclass of
344 rtl_opt_pass along with...
345 (pass_data_rtl_unswitch): ...new pass_data instance and...
346 (make_pass_rtl_unswitch): ...new function.
347 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
348 subclass of rtl_opt_pass along with...
349 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
350 and...
351 (make_pass_rtl_unroll_and_peel_loops): ...new function.
352 (pass_rtl_doloop): Convert from a global struct to a subclass of
353 rtl_opt_pass along with...
354 (pass_data_rtl_doloop): ...new pass_data instance and...
355 (make_pass_rtl_doloop): ...new function.
356 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
357 a subclass of rtl_opt_pass along with...
358 (pass_data_lower_subreg): ...new pass_data instance and...
359 (make_pass_lower_subreg): ...new function.
360 (pass_lower_subreg2): Convert from a global struct to a subclass of
361 rtl_opt_pass along with...
362 (pass_data_lower_subreg2): ...new pass_data instance and...
363 (make_pass_lower_subreg2): ...new function.
364 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
365 struct to a subclass of ipa_opt_pass_d along with...
366 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
367 (make_pass_ipa_lto_gimple_out): ...new function.
368 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
369 of ipa_opt_pass_d along with...
370 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
371 (make_pass_ipa_lto_finish_out): ...new function.
372 * mode-switching.c (pass_mode_switching): Convert from a global struct
373 to a subclass of rtl_opt_pass along with...
374 (pass_data_mode_switching): ...new pass_data instance and...
375 (make_pass_mode_switching): ...new function.
376 * modulo-sched.c (pass_sms): Convert from a global struct to a
377 subclass of rtl_opt_pass along with...
378 (pass_data_sms): ...new pass_data instance and...
379 (make_pass_sms): ...new function.
380 * omp-low.c (pass_expand_omp): Convert from a global struct to a
381 subclass of gimple_opt_pass along with...
382 (pass_data_expand_omp): ...new pass_data instance and...
383 (make_pass_expand_omp): ...new function.
384 (pass_lower_omp): Convert from a global struct to a subclass of
385 gimple_opt_pass along with...
386 (pass_data_lower_omp): ...new pass_data instance and...
387 (make_pass_lower_omp): ...new function.
388 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
389 of gimple_opt_pass along with...
390 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
391 (make_pass_diagnose_omp_blocks): ...new function.
392 * passes.c (pass_early_local_passes): Convert from a global struct to
393 a subclass of simple_ipa_opt_pass along with...
394 (pass_data_early_local_passes): ...new pass_data instance and...
395 (make_pass_early_local_passes): ...new function.
396 (pass_all_early_optimizations): Convert from a global struct to a
397 subclass of gimple_opt_pass along with...
398 (pass_data_all_early_optimizations): ...new pass_data instance and...
399 (make_pass_all_early_optimizations): ...new function.
400 (pass_all_optimizations): Convert from a global struct to a subclass
401 of gimple_opt_pass along with...
402 (pass_data_all_optimizations): ...new pass_data instance and...
403 (make_pass_all_optimizations): ...new function.
404 (pass_all_optimizations_g): Convert from a global struct to a subclass
405 of gimple_opt_pass along with...
406 (pass_data_all_optimizations_g): ...new pass_data instance and...
407 (make_pass_all_optimizations_g): ...new function.
408 (pass_rest_of_compilation): Convert from a global struct to a subclass
409 of rtl_opt_pass along with...
410 (pass_data_rest_of_compilation): ...new pass_data instance and...
411 (make_pass_rest_of_compilation): ...new function.
412 (pass_postreload): Convert from a global struct to a subclass of
413 rtl_opt_pass along with...
414 (pass_data_postreload): ...new pass_data instance and...
415 (make_pass_postreload): ...new function.
416 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
417 subclass of rtl_opt_pass along with...
418 (pass_data_gcse2): ...new pass_data instance and...
419 (make_pass_gcse2): ...new function.
420 * postreload.c (pass_postreload_cse): Convert from a global struct to
421 a subclass of rtl_opt_pass along with...
422 (pass_data_postreload_cse): ...new pass_data instance and...
423 (make_pass_postreload_cse): ...new function.
424 * predict.c (pass_profile): Convert from a global struct to a subclass
425 of gimple_opt_pass along with...
426 (pass_data_profile): ...new pass_data instance and...
427 (make_pass_profile): ...new function.
428 (pass_strip_predict_hints): Convert from a global struct to a subclass
429 of gimple_opt_pass along with...
430 (pass_data_strip_predict_hints): ...new pass_data instance and...
431 (make_pass_strip_predict_hints): ...new function.
432 * recog.c (pass_peephole2): Convert from a global struct to a subclass
433 of rtl_opt_pass along with...
434 (pass_data_peephole2): ...new pass_data instance and...
435 (make_pass_peephole2): ...new function.
436 (pass_split_all_insns): Convert from a global struct to a subclass of
437 rtl_opt_pass along with...
438 (pass_data_split_all_insns): ...new pass_data instance and...
439 (make_pass_split_all_insns): ...new function.
440 (pass_split_after_reload): Convert from a global struct to a subclass
441 of rtl_opt_pass along with...
442 (pass_data_split_after_reload): ...new pass_data instance and...
443 (make_pass_split_after_reload): ...new function.
444 (pass_split_before_regstack): Convert from a global struct to a
445 subclass of rtl_opt_pass along with...
446 (pass_data_split_before_regstack): ...new pass_data instance and...
447 (make_pass_split_before_regstack): ...new function.
448 (pass_split_before_sched2): Convert from a global struct to a subclass
449 of rtl_opt_pass along with...
450 (pass_data_split_before_sched2): ...new pass_data instance and...
451 (make_pass_split_before_sched2): ...new function.
452 (pass_split_for_shorten_branches): Convert from a global struct to a
453 subclass of rtl_opt_pass along with...
454 (pass_data_split_for_shorten_branches): ...new pass_data instance
455 and...
456 (make_pass_split_for_shorten_branches): ...new function.
457 * ree.c (pass_ree): Convert from a global struct to a subclass of
458 rtl_opt_pass along with...
459 (pass_data_ree): ...new pass_data instance and...
460 (make_pass_ree): ...new function.
461 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
462 subclass of rtl_opt_pass along with...
463 (pass_data_stack_regs): ...new pass_data instance and...
464 (make_pass_stack_regs): ...new function.
465 (pass_stack_regs_run): Convert from a global struct to a subclass of
466 rtl_opt_pass along with...
467 (pass_data_stack_regs_run): ...new pass_data instance and...
468 (make_pass_stack_regs_run): ...new function.
469 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
470 subclass of rtl_opt_pass along with...
471 (pass_data_cprop_hardreg): ...new pass_data instance and...
472 (make_pass_cprop_hardreg): ...new function.
473 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
474 subclass of rtl_opt_pass along with...
475 (pass_data_reginfo_init): ...new pass_data instance and...
476 (make_pass_reginfo_init): ...new function.
477 * regmove.c (pass_regmove): Convert from a global struct to a subclass
478 of rtl_opt_pass along with...
479 (pass_data_regmove): ...new pass_data instance and...
480 (make_pass_regmove): ...new function.
481 * regrename.c (pass_regrename): Convert from a global struct to a
482 subclass of rtl_opt_pass along with...
483 (pass_data_regrename): ...new pass_data instance and...
484 (make_pass_regrename): ...new function.
485 * reorg.c (pass_delay_slots): Convert from a global struct to a
486 subclass of rtl_opt_pass along with...
487 (pass_data_delay_slots): ...new pass_data instance and...
488 (make_pass_delay_slots): ...new function.
489 (pass_machine_reorg): Convert from a global struct to a subclass of
490 rtl_opt_pass along with...
491 (pass_data_machine_reorg): ...new pass_data instance and...
492 (make_pass_machine_reorg): ...new function.
493 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
494 of rtl_opt_pass along with...
495 (pass_data_sched): ...new pass_data instance and...
496 (make_pass_sched): ...new function.
497 (pass_sched2): Convert from a global struct to a subclass of
498 rtl_opt_pass along with...
499 (pass_data_sched2): ...new pass_data instance and...
500 (make_pass_sched2): ...new function.
501 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
502 to a subclass of rtl_opt_pass along with...
503 (pass_data_stack_ptr_mod): ...new pass_data instance and...
504 (make_pass_stack_ptr_mod): ...new function.
505 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
506 to a subclass of rtl_opt_pass along with...
507 (pass_data_rtl_store_motion): ...new pass_data instance and...
508 (make_pass_rtl_store_motion): ...new function.
509 * tracer.c (pass_tracer): Convert from a global struct to a subclass
510 of gimple_opt_pass along with...
511 (pass_data_tracer): ...new pass_data instance and...
512 (make_pass_tracer): ...new function.
513 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
514 to a subclass of gimple_opt_pass along with...
515 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
516 (make_pass_diagnose_tm_blocks): ...new function.
517 (pass_lower_tm): Convert from a global struct to a subclass of
518 gimple_opt_pass along with...
519 (pass_data_lower_tm): ...new pass_data instance and...
520 (make_pass_lower_tm): ...new function.
521 (pass_tm_init): Convert from a global struct to a subclass of
522 gimple_opt_pass along with...
523 (pass_data_tm_init): ...new pass_data instance and...
524 (make_pass_tm_init): ...new function.
525 (pass_tm_mark): Convert from a global struct to a subclass of
526 gimple_opt_pass along with...
527 (pass_data_tm_mark): ...new pass_data instance and...
528 (make_pass_tm_mark): ...new function.
529 (pass_tm_edges): Convert from a global struct to a subclass of
530 gimple_opt_pass along with...
531 (pass_data_tm_edges): ...new pass_data instance and...
532 (make_pass_tm_edges): ...new function.
533 (pass_tm_memopt): Convert from a global struct to a subclass of
534 gimple_opt_pass along with...
535 (pass_data_tm_memopt): ...new pass_data instance and...
536 (make_pass_tm_memopt): ...new function.
537 (pass_ipa_tm): Convert from a global struct to a subclass of
538 simple_ipa_opt_pass along with...
539 (pass_data_ipa_tm): ...new pass_data instance and...
540 (make_pass_ipa_tm): ...new function.
541 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
542 subclass of gimple_opt_pass along with...
543 (pass_data_call_cdce): ...new pass_data instance and...
544 (make_pass_call_cdce): ...new function.
545 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
546 subclass of gimple_opt_pass along with...
547 (pass_data_build_cfg): ...new pass_data instance and...
548 (make_pass_build_cfg): ...new function.
549 (pass_split_crit_edges): Convert from a global struct to a subclass of
550 gimple_opt_pass along with...
551 (pass_data_split_crit_edges): ...new pass_data instance and...
552 (make_pass_split_crit_edges): ...new function.
553 (pass_warn_function_return): Convert from a global struct to a
554 subclass of gimple_opt_pass along with...
555 (pass_data_warn_function_return): ...new pass_data instance and...
556 (make_pass_warn_function_return): ...new function.
557 (pass_warn_function_noreturn): Convert from a global struct to a
558 subclass of gimple_opt_pass along with...
559 (pass_data_warn_function_noreturn): ...new pass_data instance and...
560 (make_pass_warn_function_noreturn): ...new function.
561 (pass_warn_unused_result): Convert from a global struct to a subclass
562 of gimple_opt_pass along with...
563 (pass_data_warn_unused_result): ...new pass_data instance and...
564 (make_pass_warn_unused_result): ...new function.
565 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
566 a subclass of gimple_opt_pass along with...
567 (pass_data_merge_phi): ...new pass_data instance and...
568 (make_pass_merge_phi): ...new function.
569 * tree-complex.c (pass_lower_complex): Convert from a global struct to
570 a subclass of gimple_opt_pass along with...
571 (pass_data_lower_complex): ...new pass_data instance and...
572 (make_pass_lower_complex): ...new function.
573 (pass_lower_complex_O0): Convert from a global struct to a subclass of
574 gimple_opt_pass along with...
575 (pass_data_lower_complex_O0): ...new pass_data instance and...
576 (make_pass_lower_complex_O0): ...new function.
577 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
578 subclass of gimple_opt_pass along with...
579 (pass_data_lower_eh): ...new pass_data instance and...
580 (make_pass_lower_eh): ...new function.
581 (pass_refactor_eh): Convert from a global struct to a subclass of
582 gimple_opt_pass along with...
583 (pass_data_refactor_eh): ...new pass_data instance and...
584 (make_pass_refactor_eh): ...new function.
585 (pass_lower_resx): Convert from a global struct to a subclass of
586 gimple_opt_pass along with...
587 (pass_data_lower_resx): ...new pass_data instance and...
588 (make_pass_lower_resx): ...new function.
589 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
590 of gimple_opt_pass along with...
591 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
592 (make_pass_lower_eh_dispatch): ...new function.
593 (pass_cleanup_eh): Convert from a global struct to a subclass of
594 gimple_opt_pass along with...
595 (pass_data_cleanup_eh): ...new pass_data instance and...
596 (make_pass_cleanup_eh): ...new function.
597 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
598 to a subclass of simple_ipa_opt_pass along with...
599 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
600 (make_pass_ipa_lower_emutls): ...new function.
601 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
602 a subclass of gimple_opt_pass along with...
603 (pass_data_if_conversion): ...new pass_data instance and...
604 (make_pass_if_conversion): ...new function.
605 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
606 subclass of gimple_opt_pass along with...
607 (pass_data_build_ssa): ...new pass_data instance and...
608 (make_pass_build_ssa): ...new function.
609 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
610 global struct to a subclass of gimple_opt_pass along with...
611 (pass_data_loop_distribution): ...new pass_data instance and...
612 (make_pass_loop_distribution): ...new function.
613 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
614 subclass of gimple_opt_pass along with...
615 (pass_data_mudflap_1): ...new pass_data instance and...
616 (make_pass_mudflap_1): ...new function.
617 (pass_mudflap_2): Convert from a global struct to a subclass of
618 gimple_opt_pass along with...
619 (pass_data_mudflap_2): ...new pass_data instance and...
620 (make_pass_mudflap_2): ...new function.
621 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
622 subclass of gimple_opt_pass along with...
623 (pass_data_mudflap_1): ...new pass_data instance and...
624 (make_pass_mudflap_1): ...new function.
625 (pass_mudflap_2): Convert from a global struct to a subclass of
626 gimple_opt_pass along with...
627 (pass_data_mudflap_2): ...new pass_data instance and...
628 (make_pass_mudflap_2): ...new function.
629 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
630 gimple_opt_pass along with...
631 (pass_data_nrv): ...new pass_data instance and...
632 (make_pass_nrv): ...new function.
633 (pass_return_slot): Convert from a global struct to a subclass of
634 gimple_opt_pass along with...
635 (pass_data_return_slot): ...new pass_data instance and...
636 (make_pass_return_slot): ...new function.
637 * tree-object-size.c (pass_object_sizes): Convert from a global struct
638 to a subclass of gimple_opt_pass along with...
639 (pass_data_object_sizes): ...new pass_data instance and...
640 (make_pass_object_sizes): ...new function.
641 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
642 global struct to a subclass of gimple_opt_pass along with...
643 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
644 and...
645 (make_pass_cleanup_cfg_post_optimizing): ...new function.
646 (pass_fixup_cfg): Convert from a global struct to a subclass of
647 gimple_opt_pass along with...
648 (pass_data_fixup_cfg): ...new pass_data instance and...
649 (make_pass_fixup_cfg): ...new function.
650 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
651 (make_pass_mudflap_1): ...new function.
652 (pass_mudflap_2): Replace declaration with that of...
653 (make_pass_mudflap_2): ...new function.
654 (pass_asan): Replace declaration with that of...
655 (make_pass_asan): ...new function.
656 (pass_asan_O0): Replace declaration with that of...
657 (make_pass_asan_O0): ...new function.
658 (pass_tsan): Replace declaration with that of...
659 (make_pass_tsan): ...new function.
660 (pass_tsan_O0): Replace declaration with that of...
661 (make_pass_tsan_O0): ...new function.
662 (pass_lower_cf): Replace declaration with that of...
663 (make_pass_lower_cf): ...new function.
664 (pass_refactor_eh): Replace declaration with that of...
665 (make_pass_refactor_eh): ...new function.
666 (pass_lower_eh): Replace declaration with that of...
667 (make_pass_lower_eh): ...new function.
668 (pass_lower_eh_dispatch): Replace declaration with that of...
669 (make_pass_lower_eh_dispatch): ...new function.
670 (pass_lower_resx): Replace declaration with that of...
671 (make_pass_lower_resx): ...new function.
672 (pass_build_cfg): Replace declaration with that of...
673 (make_pass_build_cfg): ...new function.
674 (pass_early_tree_profile): Replace declaration with that of...
675 (make_pass_early_tree_profile): ...new function.
676 (pass_cleanup_eh): Replace declaration with that of...
677 (make_pass_cleanup_eh): ...new function.
678 (pass_sra): Replace declaration with that of...
679 (make_pass_sra): ...new function.
680 (pass_sra_early): Replace declaration with that of...
681 (make_pass_sra_early): ...new function.
682 (pass_early_ipa_sra): Replace declaration with that of...
683 (make_pass_early_ipa_sra): ...new function.
684 (pass_tail_recursion): Replace declaration with that of...
685 (make_pass_tail_recursion): ...new function.
686 (pass_tail_calls): Replace declaration with that of...
687 (make_pass_tail_calls): ...new function.
688 (pass_tree_loop): Replace declaration with that of...
689 (make_pass_tree_loop): ...new function.
690 (pass_tree_loop_init): Replace declaration with that of...
691 (make_pass_tree_loop_init): ...new function.
692 (pass_lim): Replace declaration with that of...
693 (make_pass_lim): ...new function.
694 (pass_tree_unswitch): Replace declaration with that of...
695 (make_pass_tree_unswitch): ...new function.
696 (pass_predcom): Replace declaration with that of...
697 (make_pass_predcom): ...new function.
698 (pass_iv_canon): Replace declaration with that of...
699 (make_pass_iv_canon): ...new function.
700 (pass_scev_cprop): Replace declaration with that of...
701 (make_pass_scev_cprop): ...new function.
702 (pass_empty_loop): Replace declaration with that of...
703 (make_pass_empty_loop): ...new function.
704 (pass_record_bounds): Replace declaration with that of...
705 (make_pass_record_bounds): ...new function.
706 (pass_graphite): Replace declaration with that of...
707 (make_pass_graphite): ...new function.
708 (pass_graphite_transforms): Replace declaration with that of...
709 (make_pass_graphite_transforms): ...new function.
710 (pass_if_conversion): Replace declaration with that of...
711 (make_pass_if_conversion): ...new function.
712 (pass_loop_distribution): Replace declaration with that of...
713 (make_pass_loop_distribution): ...new function.
714 (pass_vectorize): Replace declaration with that of...
715 (make_pass_vectorize): ...new function.
716 (pass_slp_vectorize): Replace declaration with that of...
717 (make_pass_slp_vectorize): ...new function.
718 (pass_complete_unroll): Replace declaration with that of...
719 (make_pass_complete_unroll): ...new function.
720 (pass_complete_unrolli): Replace declaration with that of...
721 (make_pass_complete_unrolli): ...new function.
722 (pass_parallelize_loops): Replace declaration with that of...
723 (make_pass_parallelize_loops): ...new function.
724 (pass_loop_prefetch): Replace declaration with that of...
725 (make_pass_loop_prefetch): ...new function.
726 (pass_iv_optimize): Replace declaration with that of...
727 (make_pass_iv_optimize): ...new function.
728 (pass_tree_loop_done): Replace declaration with that of...
729 (make_pass_tree_loop_done): ...new function.
730 (pass_ch): Replace declaration with that of...
731 (make_pass_ch): ...new function.
732 (pass_ccp): Replace declaration with that of...
733 (make_pass_ccp): ...new function.
734 (pass_phi_only_cprop): Replace declaration with that of...
735 (make_pass_phi_only_cprop): ...new function.
736 (pass_build_ssa): Replace declaration with that of...
737 (make_pass_build_ssa): ...new function.
738 (pass_build_alias): Replace declaration with that of...
739 (make_pass_build_alias): ...new function.
740 (pass_build_ealias): Replace declaration with that of...
741 (make_pass_build_ealias): ...new function.
742 (pass_dominator): Replace declaration with that of...
743 (make_pass_dominator): ...new function.
744 (pass_dce): Replace declaration with that of...
745 (make_pass_dce): ...new function.
746 (pass_dce_loop): Replace declaration with that of...
747 (make_pass_dce_loop): ...new function.
748 (pass_cd_dce): Replace declaration with that of...
749 (make_pass_cd_dce): ...new function.
750 (pass_call_cdce): Replace declaration with that of...
751 (make_pass_call_cdce): ...new function.
752 (pass_merge_phi): Replace declaration with that of...
753 (make_pass_merge_phi): ...new function.
754 (pass_split_crit_edges): Replace declaration with that of...
755 (make_pass_split_crit_edges): ...new function.
756 (pass_pre): Replace declaration with that of...
757 (make_pass_pre): ...new function.
758 (pass_profile): Replace declaration with that of...
759 (make_pass_profile): ...new function.
760 (pass_strip_predict_hints): Replace declaration with that of...
761 (make_pass_strip_predict_hints): ...new function.
762 (pass_lower_complex_O0): Replace declaration with that of...
763 (make_pass_lower_complex_O0): ...new function.
764 (pass_lower_complex): Replace declaration with that of...
765 (make_pass_lower_complex): ...new function.
766 (pass_lower_vector): Replace declaration with that of...
767 (make_pass_lower_vector): ...new function.
768 (pass_lower_vector_ssa): Replace declaration with that of...
769 (make_pass_lower_vector_ssa): ...new function.
770 (pass_lower_omp): Replace declaration with that of...
771 (make_pass_lower_omp): ...new function.
772 (pass_diagnose_omp_blocks): Replace declaration with that of...
773 (make_pass_diagnose_omp_blocks): ...new function.
774 (pass_expand_omp): Replace declaration with that of...
775 (make_pass_expand_omp): ...new function.
776 (pass_expand_omp_ssa): Replace declaration with that of...
777 (make_pass_expand_omp_ssa): ...new function.
778 (pass_object_sizes): Replace declaration with that of...
779 (make_pass_object_sizes): ...new function.
780 (pass_strlen): Replace declaration with that of...
781 (make_pass_strlen): ...new function.
782 (pass_fold_builtins): Replace declaration with that of...
783 (make_pass_fold_builtins): ...new function.
784 (pass_stdarg): Replace declaration with that of...
785 (make_pass_stdarg): ...new function.
786 (pass_early_warn_uninitialized): Replace declaration with that of...
787 (make_pass_early_warn_uninitialized): ...new function.
788 (pass_late_warn_uninitialized): Replace declaration with that of...
789 (make_pass_late_warn_uninitialized): ...new function.
790 (pass_cse_reciprocals): Replace declaration with that of...
791 (make_pass_cse_reciprocals): ...new function.
792 (pass_cse_sincos): Replace declaration with that of...
793 (make_pass_cse_sincos): ...new function.
794 (pass_optimize_bswap): Replace declaration with that of...
795 (make_pass_optimize_bswap): ...new function.
796 (pass_optimize_widening_mul): Replace declaration with that of...
797 (make_pass_optimize_widening_mul): ...new function.
798 (pass_warn_function_return): Replace declaration with that of...
799 (make_pass_warn_function_return): ...new function.
800 (pass_warn_function_noreturn): Replace declaration with that of...
801 (make_pass_warn_function_noreturn): ...new function.
802 (pass_cselim): Replace declaration with that of...
803 (make_pass_cselim): ...new function.
804 (pass_phiopt): Replace declaration with that of...
805 (make_pass_phiopt): ...new function.
806 (pass_forwprop): Replace declaration with that of...
807 (make_pass_forwprop): ...new function.
808 (pass_phiprop): Replace declaration with that of...
809 (make_pass_phiprop): ...new function.
810 (pass_tree_ifcombine): Replace declaration with that of...
811 (make_pass_tree_ifcombine): ...new function.
812 (pass_dse): Replace declaration with that of...
813 (make_pass_dse): ...new function.
814 (pass_nrv): Replace declaration with that of...
815 (make_pass_nrv): ...new function.
816 (pass_rename_ssa_copies): Replace declaration with that of...
817 (make_pass_rename_ssa_copies): ...new function.
818 (pass_sink_code): Replace declaration with that of...
819 (make_pass_sink_code): ...new function.
820 (pass_fre): Replace declaration with that of...
821 (make_pass_fre): ...new function.
822 (pass_check_data_deps): Replace declaration with that of...
823 (make_pass_check_data_deps): ...new function.
824 (pass_copy_prop): Replace declaration with that of...
825 (make_pass_copy_prop): ...new function.
826 (pass_vrp): Replace declaration with that of...
827 (make_pass_vrp): ...new function.
828 (pass_uncprop): Replace declaration with that of...
829 (make_pass_uncprop): ...new function.
830 (pass_return_slot): Replace declaration with that of...
831 (make_pass_return_slot): ...new function.
832 (pass_reassoc): Replace declaration with that of...
833 (make_pass_reassoc): ...new function.
834 (pass_rebuild_cgraph_edges): Replace declaration with that of...
835 (make_pass_rebuild_cgraph_edges): ...new function.
836 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
837 (make_pass_remove_cgraph_callee_edges): ...new function.
838 (pass_build_cgraph_edges): Replace declaration with that of...
839 (make_pass_build_cgraph_edges): ...new function.
840 (pass_local_pure_const): Replace declaration with that of...
841 (make_pass_local_pure_const): ...new function.
842 (pass_tracer): Replace declaration with that of...
843 (make_pass_tracer): ...new function.
844 (pass_warn_unused_result): Replace declaration with that of...
845 (make_pass_warn_unused_result): ...new function.
846 (pass_diagnose_tm_blocks): Replace declaration with that of...
847 (make_pass_diagnose_tm_blocks): ...new function.
848 (pass_lower_tm): Replace declaration with that of...
849 (make_pass_lower_tm): ...new function.
850 (pass_tm_init): Replace declaration with that of...
851 (make_pass_tm_init): ...new function.
852 (pass_tm_mark): Replace declaration with that of...
853 (make_pass_tm_mark): ...new function.
854 (pass_tm_memopt): Replace declaration with that of...
855 (make_pass_tm_memopt): ...new function.
856 (pass_tm_edges): Replace declaration with that of...
857 (make_pass_tm_edges): ...new function.
858 (pass_split_functions): Replace declaration with that of...
859 (make_pass_split_functions): ...new function.
860 (pass_feedback_split_functions): Replace declaration with that of...
861 (make_pass_feedback_split_functions): ...new function.
862 (pass_strength_reduction): Replace declaration with that of...
863 (make_pass_strength_reduction): ...new function.
864 (pass_ipa_lower_emutls): Replace declaration with that of...
865 (make_pass_ipa_lower_emutls): ...new function.
866 (pass_ipa_function_and_variable_visibility): Replace declaration with
867 that of...
868 (make_pass_ipa_function_and_variable_visibility): ...new function.
869 (pass_ipa_tree_profile): Replace declaration with that of...
870 (make_pass_ipa_tree_profile): ...new function.
871 (pass_early_local_passes): Replace declaration with that of...
872 (make_pass_early_local_passes): ...new function.
873 (pass_ipa_whole_program_visibility): Replace declaration with that
874 of...
875 (make_pass_ipa_whole_program_visibility): ...new function.
876 (pass_ipa_lto_gimple_out): Replace declaration with that of...
877 (make_pass_ipa_lto_gimple_out): ...new function.
878 (pass_ipa_increase_alignment): Replace declaration with that of...
879 (make_pass_ipa_increase_alignment): ...new function.
880 (pass_ipa_inline): Replace declaration with that of...
881 (make_pass_ipa_inline): ...new function.
882 (pass_ipa_free_lang_data): Replace declaration with that of...
883 (make_pass_ipa_free_lang_data): ...new function.
884 (pass_ipa_free_inline_summary): Replace declaration with that of...
885 (make_pass_ipa_free_inline_summary): ...new function.
886 (pass_ipa_cp): Replace declaration with that of...
887 (make_pass_ipa_cp): ...new function.
888 (pass_ipa_reference): Replace declaration with that of...
889 (make_pass_ipa_reference): ...new function.
890 (pass_ipa_pure_const): Replace declaration with that of...
891 (make_pass_ipa_pure_const): ...new function.
892 (pass_ipa_pta): Replace declaration with that of...
893 (make_pass_ipa_pta): ...new function.
894 (pass_ipa_lto_finish_out): Replace declaration with that of...
895 (make_pass_ipa_lto_finish_out): ...new function.
896 (pass_ipa_tm): Replace declaration with that of...
897 (make_pass_ipa_tm): ...new function.
898 (pass_ipa_profile): Replace declaration with that of...
899 (make_pass_ipa_profile): ...new function.
900 (pass_ipa_cdtor_merge): Replace declaration with that of...
901 (make_pass_ipa_cdtor_merge): ...new function.
902 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
903 of...
904 (make_pass_cleanup_cfg_post_optimizing): ...new function.
905 (pass_init_datastructures): Replace declaration with that of...
906 (make_pass_init_datastructures): ...new function.
907 (pass_fixup_cfg): Replace declaration with that of...
908 (make_pass_fixup_cfg): ...new function.
909 (pass_expand): Replace declaration with that of...
910 (make_pass_expand): ...new function.
911 (pass_instantiate_virtual_regs): Replace declaration with that of...
912 (make_pass_instantiate_virtual_regs): ...new function.
913 (pass_rtl_fwprop): Replace declaration with that of...
914 (make_pass_rtl_fwprop): ...new function.
915 (pass_rtl_fwprop_addr): Replace declaration with that of...
916 (make_pass_rtl_fwprop_addr): ...new function.
917 (pass_jump): Replace declaration with that of...
918 (make_pass_jump): ...new function.
919 (pass_jump2): Replace declaration with that of...
920 (make_pass_jump2): ...new function.
921 (pass_lower_subreg): Replace declaration with that of...
922 (make_pass_lower_subreg): ...new function.
923 (pass_cse): Replace declaration with that of...
924 (make_pass_cse): ...new function.
925 (pass_fast_rtl_dce): Replace declaration with that of...
926 (make_pass_fast_rtl_dce): ...new function.
927 (pass_ud_rtl_dce): Replace declaration with that of...
928 (make_pass_ud_rtl_dce): ...new function.
929 (pass_rtl_dce): Replace declaration with that of...
930 (make_pass_rtl_dce): ...new function.
931 (pass_rtl_dse1): Replace declaration with that of...
932 (make_pass_rtl_dse1): ...new function.
933 (pass_rtl_dse2): Replace declaration with that of...
934 (make_pass_rtl_dse2): ...new function.
935 (pass_rtl_dse3): Replace declaration with that of...
936 (make_pass_rtl_dse3): ...new function.
937 (pass_rtl_cprop): Replace declaration with that of...
938 (make_pass_rtl_cprop): ...new function.
939 (pass_rtl_pre): Replace declaration with that of...
940 (make_pass_rtl_pre): ...new function.
941 (pass_rtl_hoist): Replace declaration with that of...
942 (make_pass_rtl_hoist): ...new function.
943 (pass_rtl_store_motion): Replace declaration with that of...
944 (make_pass_rtl_store_motion): ...new function.
945 (pass_cse_after_global_opts): Replace declaration with that of...
946 (make_pass_cse_after_global_opts): ...new function.
947 (pass_rtl_ifcvt): Replace declaration with that of...
948 (make_pass_rtl_ifcvt): ...new function.
949 (pass_into_cfg_layout_mode): Replace declaration with that of...
950 (make_pass_into_cfg_layout_mode): ...new function.
951 (pass_outof_cfg_layout_mode): Replace declaration with that of...
952 (make_pass_outof_cfg_layout_mode): ...new function.
953 (pass_loop2): Replace declaration with that of...
954 (make_pass_loop2): ...new function.
955 (pass_rtl_loop_init): Replace declaration with that of...
956 (make_pass_rtl_loop_init): ...new function.
957 (pass_rtl_move_loop_invariants): Replace declaration with that of...
958 (make_pass_rtl_move_loop_invariants): ...new function.
959 (pass_rtl_unswitch): Replace declaration with that of...
960 (make_pass_rtl_unswitch): ...new function.
961 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
962 (make_pass_rtl_unroll_and_peel_loops): ...new function.
963 (pass_rtl_doloop): Replace declaration with that of...
964 (make_pass_rtl_doloop): ...new function.
965 (pass_rtl_loop_done): Replace declaration with that of...
966 (make_pass_rtl_loop_done): ...new function.
967 (pass_web): Replace declaration with that of...
968 (make_pass_web): ...new function.
969 (pass_cse2): Replace declaration with that of...
970 (make_pass_cse2): ...new function.
971 (pass_df_initialize_opt): Replace declaration with that of...
972 (make_pass_df_initialize_opt): ...new function.
973 (pass_df_initialize_no_opt): Replace declaration with that of...
974 (make_pass_df_initialize_no_opt): ...new function.
975 (pass_reginfo_init): Replace declaration with that of...
976 (make_pass_reginfo_init): ...new function.
977 (pass_inc_dec): Replace declaration with that of...
978 (make_pass_inc_dec): ...new function.
979 (pass_stack_ptr_mod): Replace declaration with that of...
980 (make_pass_stack_ptr_mod): ...new function.
981 (pass_initialize_regs): Replace declaration with that of...
982 (make_pass_initialize_regs): ...new function.
983 (pass_combine): Replace declaration with that of...
984 (make_pass_combine): ...new function.
985 (pass_if_after_combine): Replace declaration with that of...
986 (make_pass_if_after_combine): ...new function.
987 (pass_ree): Replace declaration with that of...
988 (make_pass_ree): ...new function.
989 (pass_partition_blocks): Replace declaration with that of...
990 (make_pass_partition_blocks): ...new function.
991 (pass_match_asm_constraints): Replace declaration with that of...
992 (make_pass_match_asm_constraints): ...new function.
993 (pass_regmove): Replace declaration with that of...
994 (make_pass_regmove): ...new function.
995 (pass_split_all_insns): Replace declaration with that of...
996 (make_pass_split_all_insns): ...new function.
997 (pass_fast_rtl_byte_dce): Replace declaration with that of...
998 (make_pass_fast_rtl_byte_dce): ...new function.
999 (pass_lower_subreg2): Replace declaration with that of...
1000 (make_pass_lower_subreg2): ...new function.
1001 (pass_mode_switching): Replace declaration with that of...
1002 (make_pass_mode_switching): ...new function.
1003 (pass_sms): Replace declaration with that of...
1004 (make_pass_sms): ...new function.
1005 (pass_sched): Replace declaration with that of...
1006 (make_pass_sched): ...new function.
1007 (pass_ira): Replace declaration with that of...
1008 (make_pass_ira): ...new function.
1009 (pass_reload): Replace declaration with that of...
1010 (make_pass_reload): ...new function.
1011 (pass_clean_state): Replace declaration with that of...
1012 (make_pass_clean_state): ...new function.
1013 (pass_branch_prob): Replace declaration with that of...
1014 (make_pass_branch_prob): ...new function.
1015 (pass_value_profile_transformations): Replace declaration with that
1016 of...
1017 (make_pass_value_profile_transformations): ...new function.
1018 (pass_postreload_cse): Replace declaration with that of...
1019 (make_pass_postreload_cse): ...new function.
1020 (pass_gcse2): Replace declaration with that of...
1021 (make_pass_gcse2): ...new function.
1022 (pass_split_after_reload): Replace declaration with that of...
1023 (make_pass_split_after_reload): ...new function.
1024 (pass_branch_target_load_optimize1): Replace declaration with that
1025 of...
1026 (make_pass_branch_target_load_optimize1): ...new function.
1027 (pass_thread_prologue_and_epilogue): Replace declaration with that
1028 of...
1029 (make_pass_thread_prologue_and_epilogue): ...new function.
1030 (pass_stack_adjustments): Replace declaration with that of...
1031 (make_pass_stack_adjustments): ...new function.
1032 (pass_peephole2): Replace declaration with that of...
1033 (make_pass_peephole2): ...new function.
1034 (pass_if_after_reload): Replace declaration with that of...
1035 (make_pass_if_after_reload): ...new function.
1036 (pass_regrename): Replace declaration with that of...
1037 (make_pass_regrename): ...new function.
1038 (pass_cprop_hardreg): Replace declaration with that of...
1039 (make_pass_cprop_hardreg): ...new function.
1040 (pass_reorder_blocks): Replace declaration with that of...
1041 (make_pass_reorder_blocks): ...new function.
1042 (pass_branch_target_load_optimize2): Replace declaration with that
1043 of...
1044 (make_pass_branch_target_load_optimize2): ...new function.
1045 (pass_leaf_regs): Replace declaration with that of...
1046 (make_pass_leaf_regs): ...new function.
1047 (pass_split_before_sched2): Replace declaration with that of...
1048 (make_pass_split_before_sched2): ...new function.
1049 (pass_compare_elim_after_reload): Replace declaration with that of...
1050 (make_pass_compare_elim_after_reload): ...new function.
1051 (pass_sched2): Replace declaration with that of...
1052 (make_pass_sched2): ...new function.
1053 (pass_stack_regs): Replace declaration with that of...
1054 (make_pass_stack_regs): ...new function.
1055 (pass_stack_regs_run): Replace declaration with that of...
1056 (make_pass_stack_regs_run): ...new function.
1057 (pass_df_finish): Replace declaration with that of...
1058 (make_pass_df_finish): ...new function.
1059 (pass_compute_alignments): Replace declaration with that of...
1060 (make_pass_compute_alignments): ...new function.
1061 (pass_duplicate_computed_gotos): Replace declaration with that of...
1062 (make_pass_duplicate_computed_gotos): ...new function.
1063 (pass_variable_tracking): Replace declaration with that of...
1064 (make_pass_variable_tracking): ...new function.
1065 (pass_free_cfg): Replace declaration with that of...
1066 (make_pass_free_cfg): ...new function.
1067 (pass_machine_reorg): Replace declaration with that of...
1068 (make_pass_machine_reorg): ...new function.
1069 (pass_cleanup_barriers): Replace declaration with that of...
1070 (make_pass_cleanup_barriers): ...new function.
1071 (pass_delay_slots): Replace declaration with that of...
1072 (make_pass_delay_slots): ...new function.
1073 (pass_split_for_shorten_branches): Replace declaration with that of...
1074 (make_pass_split_for_shorten_branches): ...new function.
1075 (pass_split_before_regstack): Replace declaration with that of...
1076 (make_pass_split_before_regstack): ...new function.
1077 (pass_convert_to_eh_region_ranges): Replace declaration with that
1078 of...
1079 (make_pass_convert_to_eh_region_ranges): ...new function.
1080 (pass_shorten_branches): Replace declaration with that of...
1081 (make_pass_shorten_branches): ...new function.
1082 (pass_set_nothrow_function_flags): Replace declaration with that of...
1083 (make_pass_set_nothrow_function_flags): ...new function.
1084 (pass_dwarf2_frame): Replace declaration with that of...
1085 (make_pass_dwarf2_frame): ...new function.
1086 (pass_final): Replace declaration with that of...
1087 (make_pass_final): ...new function.
1088 (pass_rtl_seqabstr): Replace declaration with that of...
1089 (make_pass_rtl_seqabstr): ...new function.
1090 (pass_release_ssa_names): Replace declaration with that of...
1091 (make_pass_release_ssa_names): ...new function.
1092 (pass_early_inline): Replace declaration with that of...
1093 (make_pass_early_inline): ...new function.
1094 (pass_inline_parameters): Replace declaration with that of...
1095 (make_pass_inline_parameters): ...new function.
1096 (pass_update_address_taken): Replace declaration with that of...
1097 (make_pass_update_address_taken): ...new function.
1098 (pass_convert_switch): Replace declaration with that of...
1099 (make_pass_convert_switch): ...new function.
1100 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1101 to a subclass of simple_ipa_opt_pass along with...
1102 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1103 (make_pass_ipa_tree_profile): ...new function.
1104 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1105 subclass of gimple_opt_pass along with...
1106 (pass_data_sra_early): ...new pass_data instance and...
1107 (make_pass_sra_early): ...new function.
1108 (pass_sra): Convert from a global struct to a subclass of
1109 gimple_opt_pass along with...
1110 (pass_data_sra): ...new pass_data instance and...
1111 (make_pass_sra): ...new function.
1112 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1113 gimple_opt_pass along with...
1114 (pass_data_early_ipa_sra): ...new pass_data instance and...
1115 (make_pass_early_ipa_sra): ...new function.
1116 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1117 subclass of gimple_opt_pass along with...
1118 (pass_data_ccp): ...new pass_data instance and...
1119 (make_pass_ccp): ...new function.
1120 (pass_fold_builtins): Convert from a global struct to a subclass of
1121 gimple_opt_pass along with...
1122 (pass_data_fold_builtins): ...new pass_data instance and...
1123 (make_pass_fold_builtins): ...new function.
1124 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1125 subclass of gimple_opt_pass along with...
1126 (pass_data_copy_prop): ...new pass_data instance and...
1127 (make_pass_copy_prop): ...new function.
1128 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1129 global struct to a subclass of gimple_opt_pass along with...
1130 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1131 (make_pass_rename_ssa_copies): ...new function.
1132 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1133 subclass of gimple_opt_pass along with...
1134 (pass_data_dce): ...new pass_data instance and...
1135 (make_pass_dce): ...new function.
1136 (pass_dce_loop): Convert from a global struct to a subclass of
1137 gimple_opt_pass along with...
1138 (pass_data_dce_loop): ...new pass_data instance and...
1139 (make_pass_dce_loop): ...new function.
1140 (pass_cd_dce): Convert from a global struct to a subclass of
1141 gimple_opt_pass along with...
1142 (pass_data_cd_dce): ...new pass_data instance and...
1143 (make_pass_cd_dce): ...new function.
1144 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1145 subclass of gimple_opt_pass along with...
1146 (pass_data_dominator): ...new pass_data instance and...
1147 (make_pass_dominator): ...new function.
1148 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1149 gimple_opt_pass along with...
1150 (pass_data_phi_only_cprop): ...new pass_data instance and...
1151 (make_pass_phi_only_cprop): ...new function.
1152 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1153 subclass of gimple_opt_pass along with...
1154 (pass_data_dse): ...new pass_data instance and...
1155 (make_pass_dse): ...new function.
1156 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1157 a subclass of gimple_opt_pass along with...
1158 (pass_data_forwprop): ...new pass_data instance and...
1159 (make_pass_forwprop): ...new function.
1160 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1161 struct to a subclass of gimple_opt_pass along with...
1162 (pass_data_tree_ifcombine): ...new pass_data instance and...
1163 (make_pass_tree_ifcombine): ...new function.
1164 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1165 subclass of gimple_opt_pass along with...
1166 (pass_data_ch): ...new pass_data instance and...
1167 (make_pass_ch): ...new function.
1168 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1169 subclass of gimple_opt_pass along with...
1170 (pass_data_tree_loop): ...new pass_data instance and...
1171 (make_pass_tree_loop): ...new function.
1172 (pass_tree_loop_init): Convert from a global struct to a subclass of
1173 gimple_opt_pass along with...
1174 (pass_data_tree_loop_init): ...new pass_data instance and...
1175 (make_pass_tree_loop_init): ...new function.
1176 (pass_lim): Convert from a global struct to a subclass of
1177 gimple_opt_pass along with...
1178 (pass_data_lim): ...new pass_data instance and...
1179 (make_pass_lim): ...new function.
1180 (pass_tree_unswitch): Convert from a global struct to a subclass of
1181 gimple_opt_pass along with...
1182 (pass_data_tree_unswitch): ...new pass_data instance and...
1183 (make_pass_tree_unswitch): ...new function.
1184 (pass_predcom): Convert from a global struct to a subclass of
1185 gimple_opt_pass along with...
1186 (pass_data_predcom): ...new pass_data instance and...
1187 (make_pass_predcom): ...new function.
1188 (pass_vectorize): Convert from a global struct to a subclass of
1189 gimple_opt_pass along with...
1190 (pass_data_vectorize): ...new pass_data instance and...
1191 (make_pass_vectorize): ...new function.
1192 (pass_graphite): Convert from a global struct to a subclass of
1193 gimple_opt_pass along with...
1194 (pass_data_graphite): ...new pass_data instance and...
1195 (make_pass_graphite): ...new function.
1196 (pass_graphite_transforms): Convert from a global struct to a subclass
1197 of gimple_opt_pass along with...
1198 (pass_data_graphite_transforms): ...new pass_data instance and...
1199 (make_pass_graphite_transforms): ...new function.
1200 (pass_check_data_deps): Convert from a global struct to a subclass of
1201 gimple_opt_pass along with...
1202 (pass_data_check_data_deps): ...new pass_data instance and...
1203 (make_pass_check_data_deps): ...new function.
1204 (pass_iv_canon): Convert from a global struct to a subclass of
1205 gimple_opt_pass along with...
1206 (pass_data_iv_canon): ...new pass_data instance and...
1207 (make_pass_iv_canon): ...new function.
1208 (pass_scev_cprop): Convert from a global struct to a subclass of
1209 gimple_opt_pass along with...
1210 (pass_data_scev_cprop): ...new pass_data instance and...
1211 (make_pass_scev_cprop): ...new function.
1212 (pass_record_bounds): Convert from a global struct to a subclass of
1213 gimple_opt_pass along with...
1214 (pass_data_record_bounds): ...new pass_data instance and...
1215 (make_pass_record_bounds): ...new function.
1216 (pass_complete_unroll): Convert from a global struct to a subclass of
1217 gimple_opt_pass along with...
1218 (pass_data_complete_unroll): ...new pass_data instance and...
1219 (make_pass_complete_unroll): ...new function.
1220 (pass_complete_unrolli): Convert from a global struct to a subclass of
1221 gimple_opt_pass along with...
1222 (pass_data_complete_unrolli): ...new pass_data instance and...
1223 (make_pass_complete_unrolli): ...new function.
1224 (pass_parallelize_loops): Convert from a global struct to a subclass
1225 of gimple_opt_pass along with...
1226 (pass_data_parallelize_loops): ...new pass_data instance and...
1227 (make_pass_parallelize_loops): ...new function.
1228 (pass_loop_prefetch): Convert from a global struct to a subclass of
1229 gimple_opt_pass along with...
1230 (pass_data_loop_prefetch): ...new pass_data instance and...
1231 (make_pass_loop_prefetch): ...new function.
1232 (pass_iv_optimize): Convert from a global struct to a subclass of
1233 gimple_opt_pass along with...
1234 (pass_data_iv_optimize): ...new pass_data instance and...
1235 (make_pass_iv_optimize): ...new function.
1236 (pass_tree_loop_done): Convert from a global struct to a subclass of
1237 gimple_opt_pass along with...
1238 (pass_data_tree_loop_done): ...new pass_data instance and...
1239 (make_pass_tree_loop_done): ...new function.
1240 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1241 struct to a subclass of gimple_opt_pass along with...
1242 (pass_data_cse_reciprocals): ...new pass_data instance and...
1243 (make_pass_cse_reciprocals): ...new function.
1244 (pass_cse_sincos): Convert from a global struct to a subclass of
1245 gimple_opt_pass along with...
1246 (pass_data_cse_sincos): ...new pass_data instance and...
1247 (make_pass_cse_sincos): ...new function.
1248 (pass_optimize_bswap): Convert from a global struct to a subclass of
1249 gimple_opt_pass along with...
1250 (pass_data_optimize_bswap): ...new pass_data instance and...
1251 (make_pass_optimize_bswap): ...new function.
1252 (pass_optimize_widening_mul): Convert from a global struct to a
1253 subclass of gimple_opt_pass along with...
1254 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1255 (make_pass_optimize_widening_mul): ...new function.
1256 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1257 subclass of gimple_opt_pass along with...
1258 (pass_data_phiopt): ...new pass_data instance and...
1259 (make_pass_phiopt): ...new function.
1260 (pass_cselim): Convert from a global struct to a subclass of
1261 gimple_opt_pass along with...
1262 (pass_data_cselim): ...new pass_data instance and...
1263 (make_pass_cselim): ...new function.
1264 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1265 subclass of gimple_opt_pass along with...
1266 (pass_data_phiprop): ...new pass_data instance and...
1267 (make_pass_phiprop): ...new function.
1268 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1269 subclass of gimple_opt_pass along with...
1270 (pass_data_pre): ...new pass_data instance and...
1271 (make_pass_pre): ...new function.
1272 (pass_fre): Convert from a global struct to a subclass of
1273 gimple_opt_pass along with...
1274 (pass_data_fre): ...new pass_data instance and...
1275 (make_pass_fre): ...new function.
1276 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1277 subclass of gimple_opt_pass along with...
1278 (pass_data_reassoc): ...new pass_data instance and...
1279 (make_pass_reassoc): ...new function.
1280 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1281 subclass of gimple_opt_pass along with...
1282 (pass_data_sink_code): ...new pass_data instance and...
1283 (make_pass_sink_code): ...new function.
1284 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1285 subclass of gimple_opt_pass along with...
1286 (pass_data_strlen): ...new pass_data instance and...
1287 (make_pass_strlen): ...new function.
1288 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1289 struct to a subclass of gimple_opt_pass along with...
1290 (pass_data_build_alias): ...new pass_data instance and...
1291 (make_pass_build_alias): ...new function.
1292 (pass_build_ealias): Convert from a global struct to a subclass of
1293 gimple_opt_pass along with...
1294 (pass_data_build_ealias): ...new pass_data instance and...
1295 (make_pass_build_ealias): ...new function.
1296 (pass_ipa_pta): Convert from a global struct to a subclass of
1297 simple_ipa_opt_pass along with...
1298 (pass_data_ipa_pta): ...new pass_data instance and...
1299 (make_pass_ipa_pta): ...new function.
1300 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1301 subclass of gimple_opt_pass along with...
1302 (pass_data_uncprop): ...new pass_data instance and...
1303 (make_pass_uncprop): ...new function.
1304 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1305 global struct to a subclass of gimple_opt_pass along with...
1306 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1307 (make_pass_late_warn_uninitialized): ...new function.
1308 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1309 to a subclass of gimple_opt_pass along with...
1310 (pass_data_init_datastructures): ...new pass_data instance and...
1311 (make_pass_init_datastructures): ...new function.
1312 (pass_early_warn_uninitialized): Convert from a global struct to a
1313 subclass of gimple_opt_pass along with...
1314 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1315 (make_pass_early_warn_uninitialized): ...new function.
1316 (pass_update_address_taken): Convert from a global struct to a
1317 subclass of gimple_opt_pass along with...
1318 (pass_data_update_address_taken): ...new pass_data instance and...
1319 (make_pass_update_address_taken): ...new function.
1320 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1321 struct to a subclass of gimple_opt_pass along with...
1322 (pass_data_release_ssa_names): ...new pass_data instance and...
1323 (make_pass_release_ssa_names): ...new function.
1324 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1325 subclass of gimple_opt_pass along with...
1326 (pass_data_stdarg): ...new pass_data instance and...
1327 (make_pass_stdarg): ...new function.
1328 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1329 global struct to a subclass of gimple_opt_pass along with...
1330 (pass_data_convert_switch): ...new pass_data instance and...
1331 (make_pass_convert_switch): ...new function.
1332 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1333 to a subclass of gimple_opt_pass along with...
1334 (pass_data_tail_recursion): ...new pass_data instance and...
1335 (make_pass_tail_recursion): ...new function.
1336 (pass_tail_calls): Convert from a global struct to a subclass of
1337 gimple_opt_pass along with...
1338 (pass_data_tail_calls): ...new pass_data instance and...
1339 (make_pass_tail_calls): ...new function.
1340 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1341 struct to a subclass of gimple_opt_pass along with...
1342 (pass_data_lower_vector): ...new pass_data instance and...
1343 (make_pass_lower_vector): ...new function.
1344 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1345 gimple_opt_pass along with...
1346 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1347 (make_pass_lower_vector_ssa): ...new function.
1348 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
1349 to a subclass of gimple_opt_pass along with...
1350 (pass_data_slp_vectorize): ...new pass_data instance and...
1351 (make_pass_slp_vectorize): ...new function.
1352 (pass_ipa_increase_alignment): Convert from a global struct to a
1353 subclass of simple_ipa_opt_pass along with...
1354 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
1355 (make_pass_ipa_increase_alignment): ...new function.
1356 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
1357 gimple_opt_pass along with...
1358 (pass_data_vrp): ...new pass_data instance and...
1359 (make_pass_vrp): ...new function.
1360 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
1361 subclass of simple_ipa_opt_pass along with...
1362 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
1363 (make_pass_ipa_free_lang_data): ...new function.
1364 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
1365 gimple_opt_pass along with...
1366 (pass_data_tsan): ...new pass_data instance and...
1367 (make_pass_tsan): ...new function.
1368 (pass_tsan_O0): Convert from a global struct to a subclass of
1369 gimple_opt_pass along with...
1370 (pass_data_tsan_O0): ...new pass_data instance and...
1371 (make_pass_tsan_O0): ...new function.
1372 * var-tracking.c (pass_variable_tracking): Convert from a global
1373 struct to a subclass of rtl_opt_pass along with...
1374 (pass_data_variable_tracking): ...new pass_data instance and...
1375 (make_pass_variable_tracking): ...new function.
1376 * web.c (pass_web): Convert from a global struct to a subclass of
1377 rtl_opt_pass along with...
1378 (pass_data_web): ...new pass_data instance and...
1379 (make_pass_web): ...new function.
1380 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
1381 declaration with that of...
1382 (make_pass_mode_switch_use): ...new function.
1383 (pass_resolve_sw_modes): Replace declaration with that of...
1384 (make_pass_resolve_sw_modes): ...new function.
1385 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
1386 from a global struct to a subclass of rtl_opt_pass along with...
1387 (pass_data_mode_switch_use): ...new pass_data instance and...
1388 (make_pass_mode_switch_use): ...new function.
1389 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
1390 from a global struct to a subclass of rtl_opt_pass along with...
1391 (pass_data_resolve_sw_modes): ...new pass_data instance and...
1392 (make_pass_resolve_sw_modes): ...new function.
1393 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
1394 struct to a subclass of rtl_opt_pass along with...
1395 (pass_data_insert_vzeroupper): ...new pass_data instance and...
1396 (make_pass_insert_vzeroupper): ...new function.
1397 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
1398 global struct to a subclass of rtl_opt_pass along with...
1399 (pass_data_work_around_errata): ...new pass_data instance and...
1400 (make_pass_work_around_errata): ...new function.
1401 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
1402 struct to a subclass of rtl_opt_pass along with...
1403 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
1404 (make_pass_mips_machine_reorg2): ...new function.
1405
6a389ed5
DM
14062013-08-05 David Malcolm <dmalcolm@redhat.com>
1407
1408 * passes.c (pass_manager::operator new): New.
1409
f7695dbf
DM
14102013-08-05 David Malcolm <dmalcolm@redhat.com>
1411
1412 Handwritten part of conversion of passes to C++ classes.
1413
1414 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
1415 (toplev.o): Add dep on PASS_MANAGER_H.
1416 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
1417 of early local pases to reflect this moving from a global to a
1418 member of gcc::pass_manager.
1419 (cgraph_add_new_function): Likewise.
1420 * lto-cgraph.c (lto_output_node): Update for conversion of
1421 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
1422 * passes.c (opt_pass::clone): New.
1423 (opt_pass::gate): New.
1424 (opt_pass::execute): New.
1425 (opt_pass::opt_pass): New.
1426 (pass_manager::execute_early_local_passes): New.
1427 (pass_manager::execute_pass_mode_switching): new.
1428 (finish_optimization_passes): Convert to...
1429 (pass_manager::finish_optimization_passes): ...this.
1430 (finish_optimization_passes): Update for conversion of passes to
1431 C++ classes.
1432 (register_dump_files_1): Use has_gate since we cannot portably
1433 check a vtable entry against NULL.
1434 (dump_one_pass): Likewise.
1435 (ipa_write_summaries_2): Likewise.
1436 (ipa_write_optimization_summaries_1): Likewise.
1437 (ipa_read_summaries_1): Likewise.
1438 (ipa_read_optimization_summaries_1): Likewise.
1439 (execute_ipa_stmt_fixups): Likewise.
1440 (pass_manager::pass_manager): Rewrite pass-creation, invoking
1441 pass-creation functions rather than wiring up globals, and
1442 storing the results in fields of pass_manager generated using
1443 pass-instances.def.
1444 (pass_manager::dump_profile_report): Update for conversion of
1445 passes to C++ classes.
1446 (pass_manager::execute_ipa_summary_passes): Likewise.
1447 (execute_one_ipa_transform_pass): Likewise.
1448 (execute_one_pass): Use has_gate and has_execute since we cannot
1449 portably check a vtable entry against NULL.
1450 * pass_manager.h (pass_manager::finish_optimization_passes): New.
1451 (pass_manager): Use pass-instances.def to add fields for the
1452 various pass instances.
1453 * toplev.c (finalize): Update for move of
1454 finish_optimization_passes to a method of gcc::pass_manager.
1455 * toplev.h (finish_optimization_passes): Move to method of class
1456 pass_manager.
1457 * tree-pass.h (struct pass_data): New.
1458 (opt_pass): Convert to C++ class, make it a subclass of
1459 pass_data.
1460 (opt_pass::gate): Convert to virtual function.
1461 (opt_pass::~opt_pass): New.
1462 (opt_pass::clone): New.
1463 (opt_pass::execute): Convert to virtual function.
1464 (opt_pass::opt_pass): New.
1465 (opt_pass::ctxt_): new.
1466 (gimple_opt_pass): Convert to subclass of opt_pass.
1467 (gimple_opt_pass::gimple_opt_pass): New.
1468 (rtl_opt_pass): Convert to subclass of opt_pass.
1469 (rtl_opt_pass::rtl_opt_pass): New.
1470 (ipa_opt_pass_d): Convert to subclass of opt_pass.
1471 (ipa_opt_pass_d::ipa_opt_pass_d): New.
1472 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
1473 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
1474 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
1475 invocation of pass_mode_switching to reflect this moving from a
1476 global to a member of gcc::pass_manager.
1477 (ix86_option_override): Rework how pass_insert_vzeroupper is
1478 added to the pass_manager to reflect autogenerated changes.
1479 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and
1480 PASS_MANAGER_H.
1481
d0b2f831
RE
14822013-08-05 Richard Earnshaw <rearnsha@arm.com>
1483
1484 PR rtl-optimization/57708
1485 * recog.c (peep2_find_free_register): Validate all regs in a
1486 multi-reg mode.
1487
51a5c0c2
JH
14882013-08-05 Jan Hubicka <jh@suse.cz>
1489
1490 PR lto/57602
1491 * cgraph.c (verify_cgraph_node): Accept local flags from other partitions.
1492 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
1493 (function_and_variable_visibility): Likewise.
1494 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
1495
65f0a120
GDR
14962013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1497
1498 * graph.c (init_graph_slim_pretty_print): Remove.
1499 (print_graph_cfg): Do not call it. Use local pretty printer.
1500 (start_graph_dump): Likewise.
1501
b3f80694
GDR
15022013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1503
1504 * gimple-pretty-print.c (buffer): Remove.
1505 (initialized): Likewise.
1506 (maybe_init_pretty_print): Likewise.
1507 (print_gimple_stmt): Do not call it. Use non-static local
1508 pretty_printer variable.
1509 (print_gimple_expr): Likewise.
1510 (print_gimple_seq): Likewise.
1511 (gimple_dump_bb): Likewise.
1512
11a877b3
GDR
15132013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1514
1515 * asan.c (asan_pp): Remove.
1516 (asan_pp_initialized): Likewise.
1517 (asan_pp_initialize): Likewise.
1518 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
1519 (asan_emit_stack_protection): Tidy. Use local pretty printer.
1520 (asan_add_global): Likewise.
1521
b066401f
GDR
15222013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
1523
1524 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
1525 * diagnostic.h (diagnostic_flush_buffer): Adjust.
1526 * pretty-print.c (pp_formatted_text_data): Likewise.
1527 (pp_indent): Rename from pp_base_indent.
1528 (pp_format): Rename from pp_base_format.
1529 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
1530 (pp_format_verbatim): Rename from pp_base_format_verbatim.
1531 (pp_flush): Rename from pp_base_flush.
1532 (pp_set_line_maximum_length): Rename from
1533 pp_base_set_line_maximum_length.
1534 (pp_clear_output_area): Rename from pp_base_clear_output_area.
1535 (pp_set_prefix): Rename from pp_base_set_prefix.
1536 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
1537 (pp_emit_prefix): Rename from pp_base_emit_prefix.
1538 (pp_append_text): Rename from pp_base_append_text.
1539 (pp_formatted_text): Rename from pp_base_formatted_text.
1540 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
1541 (pp_remaining_character_count_for_line): Rename from
1542 pp_base_remaining_character_count_for_line.
1543 (pp_newline): Rename from pp_base_newline.
1544 (pp_character): Rename from pp_base_character.
1545 (pp_string): Rename from pp_base_string.
1546 (pp_maybe_space): Rename from pp_base_maybe_space.
1547 * asan.c (asan_pp_string): Adjust.
1548 (asan_emit_stack_protection): Likewise.
1549 (asan_add_global): Likewise.
1550 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
1551 * tree-mudflap.c (mf_varname_tree): Likewise.
1552 * tree-pretty-print.c (pp_tree_identifier): Rename from
1553 pp_base_tree_identifier.
1554 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
1555 Declare as function.
1556
137a1a27
GDR
15572013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1558
1559 * pretty-print.h (pp_bar_bar): New.
1560 (pp_ampersand_ampersand): Likewise.
1561 (pp_less_equal): Likewise.
1562 (pp_greater_equal): Likewise.
1563 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
1564 printer functions instead of pp_string or operators and punctuators.
1565 (dump_gimple_call): Likewise.
1566 (dump_gimple_omp_for): Likewise.
1567 (dump_gimple_transaction): Likewise.
1568 (dump_gimple_phi): Likewise.
1569 (pp_gimple_stmt_1): Likewise.
1570 * sched-vis.c (print_insn): Likewise.
1571 * tree-mudflap.c (mf_varname_tree): Likewise.
1572 * tree-pretty-print.c (dump_block_node): Likewise.
1573 (dump_generic_node): Likewise.
1574
815effe1
JH
15752013-08-02 Jan Hubicka <jh@suse.cz>
1576
1577 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
1578 boundaries.
1579 * lto-streamer-out.c (tree_is_indexable): Results decls and
1580 parm decls are not indexable.
1581 (DFS_write_tree_body): Do not follow args and results.
1582 (hash_tree): Likewise.
1583 (output_functions): Rearrange so struct function is needed
1584 only when real body is output; be able to also ouptut abstract
1585 functions; output DECL_ARGUMENTS and DECL_RESULT.
1586 (lto_output): When not in WPA, ale store abstract functions.
1587 (write_symbol): Do not care about RESULT_DECL.
1588 (output_symbol_p): Handle correctly sbtract decls.
1589 * lto-streamer-in.c (input_function): Rearrange so struct
1590 function can be NULL at entry; allow streaming of
1591 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
1592 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
1593 sanity check during LTO.
1594 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
1595 RESULT_DECl and DECL_ARGUMENTS.
1596 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
1597 Likewise.
1598
07838b13
GDR
15992013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1600
1601 * pretty-print.h (pp_underscore): New.
1602 (pp_comma): Tidy.
1603 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
1604 printer functions instead of pp_character.
1605 (dump_binary_rhs): Likewise.
1606 (dump_ternary_rhs): Likewise.
1607 (dump_gimple_call_args): Likewise.
1608 (pp_points_to_solution): Likewise.
1609 (dump_gimple_call): Likewise.
1610 (dump_gimple_switch): Likewise.
1611 (dump_gimple_cond): Likewise.
1612 (dump_gimple_bind): Likewise.
1613 (dump_gimple_try): Likewise.
1614 (dump_gimple_omp_for): Likewise.
1615 (dump_gimple_omp_continue): Likewise.
1616 (dump_gimple_omp_single): Likewise.
1617 (dump_gimple_omp_sections): Likewise.
1618 (dump_gimple_omp_block): Likewise.
1619 (dump_gimple_omp_critical): Likewise.
1620 (dump_gimple_transaction): Likewise.
1621 (dump_gimple_asm): Likewise.
1622 (dump_gimple_phi): Likewise.
1623 (dump_gimple_omp_parallel): Likewise.
1624 (dump_gimple_omp_task): Likewise.
1625 (dump_gimple_omp_atomic_load): Likewise.
1626 (dump_gimple_omp_atomic_store): Likewise.
1627 (dump_gimple_mem_ops): Likewise.
1628 (pp_gimple_stmt_1): Likewise.
1629 (pp_cfg_jump): Likewise.
1630 (dump_implicit_edges): Likewise.
1631 (gimple_dump_bb_for_graph): Likewise.
1632 * graph.c (draw_cfg_node): Likewise.
1633 * langhooks.c (lhd_print_error_function): Likewise.
1634 * sched-vis.c (print_exp): Likewise.
1635 (print_value): Likewise.
1636 (print_pattern): Likewise.
1637 (print_insn): Likewise.
1638 (rtl_dump_bb_for_graph): Likewise.
1639 * tree-pretty-print.c (dump_function_declaration): Likewise.
1640 (dump_array_domain): Likewise.
1641 (dump_omp_clause): Likewise.
1642 (dump_location): Likewise.
1643 (dump_generic_node): Likewise.
1644 (print_struct_decl): Likewise.
1645 * diagnostic.c (diagnostic_show_locus): Use pp_space.
1646
bb0d2039
BS
16472013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1648
1649 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
1650 candidate table when replacing a candidate statement.
1651 (replace_rhs_if_not_dup): Likewise.
1652 (replace_one_candidate): Likewise.
1653
bc0ec027 16542013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 1655 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH
1656
1657 * cgraphunit.c (add_new_function): Fix logic when adding from
1658 late IPA pass.
1659 (assemble_thunk): Rename to ...
1660 (expand_thunk); .. this one; export; get it working with general functions;
1661 make produced gimple valid.
1662 * cgraph.h (expand_thunk): Declare.
1663
0e8853ee
JH
16642013-08-02 Jan Hubicka <jh@suse.cz>
1665
1666 * ipa-cp.c (gather_context_independent_values): Use ipa_get_param_move_cost.
1667 (get_replacement_map): Remove PARAM; move parameter folding into tree-inline.c
1668 (create_specialized_node): Update.
1669 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
1670 assert that we have gimple body; update move_cost.
1671 (count_formal_params): Assert that we have gimple body.
1672 (ipa_dump_param): New function.
1673 (ipa_alloc_node_params): Break out from ...
1674 (ipa_initialize_node_params): ... here.
1675 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
1676 (ipa_write_node_info): Stream move costs.
1677 (ipa_read_node_info): Read move costs.
1678 (ipa_update_after_lto_read): Do not recompute node params.
1679 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
1680 (ipa_get_param): Check we are not in WPA.
1681 (ipa_get_param_move_cost): New.
1682 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
1683 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
1684 parm numbers to be present.
1685
ecee672b
VM
16862013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1687
1688 PR rtl-optimization/58048
1689 * lra-constraints.c (process_alt_operands): Don't check asm
1690 operand on register.
1691
13ec0527
EB
16922013-08-02 Eric Botcazou <ebotcazou@adacore.com>
1693
1694 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
1695 the implied StoreLoad barrier for atomic operations if before.
1696
af15184a
JH
16972013-08-02 Jan Hubicka <jh@suse.cz>
1698 Martin Liska <marxin.liska@gmail.com>
1699
1700 * cgraph.c (cgraph_function_body_availability): Do not check cgrpah flags.
1701 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
1702 symtab_node_availability): Declare.
1703 * ipa.c (can_replace_by_local_alias): New.
1704 (function_and_variable_visibility): Use it.
1705 * symtab.c (symtab_for_node_and_aliases, symtab_nonoverwritable_alias_1,
1706 symtab_nonoverwritable_alias): New.
1707
01e54ef8
VM
17082013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1709
1710 PR rtl-optimization/57963
1711 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p):
1712 New.
1713 (lra_constraints): Use them.
1714
e2fd7ca7
SN
17152013-08-02 Sofiane Naci <sofiane.naci@arm.com>
1716
1717 * config/arm/types.md (define_attr "type"): Add "load_acq" and "store_rel".
1718 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
1719 changes.
1720 (cortex_a53_store1): Likewise.
1721
0dcc26c3
JH
17222013-08-01 Jan Hubicka <jh@suse.cz>
1723
1724 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other partitions are
1725 not needed.
1726
0bd72901
UB
17272013-08-01 Uros Bizjak <ubizjak@gmail.com>
1728
1729 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
1730 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
1731 MAYBE_NON_Q_CLASS_P where appropriate.
1732
12123452
JH
17332013-08-01 Jan Hubicka <jh@suse.cz>
1734
1735 * cgraph.h (release_function_body): Declare.
1736 * tree.c (free_lang_data_in_decl): Free, parameters and return values
1737 of unused delcarations.
1738
ce852f9c
KT
17392013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1740
1741 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical RTL form
1742 when subtracting a constant.
1743
c743b246
KT
17442013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1745
1746 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
1747 Generate canonical plus rtx with negated immediate instead of minus
1748 where appropriate.
1749 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
1750
c0c123ef
JH
17512013-08-01 Jan Hubicka <jh@suse.cz>
1752
1753 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
1754 (cgraph_release_function_body): Likewise.
1755 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1756 * cgraph.h (cgrpah_node): Rename abstract_and_needed
1757 to used_as_abstract_origin.
1758 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
1759 symbols used as abstract origins.
1760 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB
1761 * ipa.c (symtab_remove_unreachable_nodes): Recompute
1762 used_as_abstract_origin.
c0c123ef 1763 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB
1764 used_as_abstract_origin; be ready for DECL_RESULT and
1765 DECL_ARGUMENTS to be NULL.
c0c123ef 1766
0bd72901
UB
1767 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
1768 for abstract functions.
1769 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
1770 real symbols.
c0c123ef 1771
1f029433
JH
17722013-08-01 Jan Hubicka <jh@suse.cz>
1773
1774 * profile.c (compute_value_histograms): Fix thinko.
1775
c451f4d6
SN
17762013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1777
1778 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
1779 aarch-common-protos.h to extra_headers.
1780 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1781 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
1782 * config/aarch64/t-aarch64 (aarch-common.o): Define.
1783
34b512f5
SN
17842013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1785
1786 * config/aarch64/aarch64.md (define_attr "type"): Delete.
1787 Include "../arm/types.md". Define "type" attribute for all patterns.
1788 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
1789 attribute changes.
1790
d86e633a
MM
17912013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
1792
1793 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
1794 to support power8 load fusion.
1795 (fusion_gpr_mem_load): Likewise.
1796
1797 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
1798
1799 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
1800 declarations for power8 load fusion.
1801 (emit_fusion_gpr_load): Likewise.
1802
1803 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
1804 tuning for power8, turn on fusion mode by default. Turn on sign
1805 extending fusion mode if normal fusion mode is on, and we are at
1806 -O2 or -O3.
1807 (fusion_gpr_load_p): New function, return true if we can fuse an
1808 addis instruction with a dependent load to a GPR.
1809 (emit_fusion_gpr_load): Emit the instructions for power8 load
1810 fusion to GPRs.
1811
0bd72901 1812 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM
1813 (VSX load fusion peepholes): New peepholes to fuse together an
1814 addi instruction with a VSX load instruction.
1815
1816 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
1817 peepholes to fuse an addis instruction with a load to a GPR base
1818 register. If we are supporting sign extending fusions, convert
1819 sign extending loads to zero extending loads and add an explicit
1820 sign extension.
1821
c3f35647
SN
18222013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1823
1824 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
1825 aarch-common-protos.h to extra_headers.
1826 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1827 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
1828 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
1829 (arm_no_early_alu_shift_dep): Likewise.
1830 (arm_no_early_alu_shift_value_dep): Likewise.
1831 (arm_no_early_mul_dep): Likewise.
1832 (arm_no_early_store_addr_dep): Likewise.
1833 (arm_mac_accumulator_is_mul_result): Likewise.
1834 (arm_mac_accumulator_is_result): Likewise.
c3f35647 1835 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB
1836 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
1837 here to ...
c3f35647 1838 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
1839 (arm_no_early_alu_shift_dep): Likewise.
1840 (arm_no_early_alu_shift_value_dep): Likewise.
1841 (arm_no_early_mul_dep): Likewise.
1842 (arm_no_early_store_addr_dep): Likewise.
1843 (arm_mac_accumulator_is_mul_result): Likewise.
1844 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN
1845 * config/arm/aarch-common-protos.h: ... here. New file.
1846 * config/arm/t-arm (aarch-common.o): Define.
1847
18482013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1849
1850 * config/arm/arm.md: Include new file "types.md".
1851 (define_attr "type"): Move from here to ...
1852 (define_attr "mul32"): Likewise.
1853 (define_attr "mul64"): Likewise.
1854 * config/arm/types.md: ... here. New file.
1855
228c1313
SH
18562013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
1857
1858 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
1859 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
1860
d6e9046d 18612013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG
1862
1863 * gen-pass-instances.awk: Fix offset of substr().
1864
a167b052
DM
18652013-07-31 David Malcolm <dmalcolm@redhat.com>
1866
1867 * Makefile.in (pass-instances.def): New.
1868 (passes.o): Replace dependency on passes.def with one on
1869 pass-instances.def
1870
1871 * gen-pass-instances.awk: New.
1872
1873 * passes.c (pass_manager::pass_manager): Use pass-instances.def
1874 rather than passes.def, updating local definition of NEXT_PASS
1875 macro to add an extra NUM parameter (currently unused).
1876
315f8c0e
DM
18772013-07-30 David Malcolm <dmalcolm@redhat.com>
1878
1879 * Makefile.in (PASS_MANAGER_H): New.
1880 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
1881 (passes.o): Likewise.
1882 (statistics.o): Likewise.
1883 (cgraphunit.o): Likewise.
1884 (context.o): Depend on PASS_MANAGER_H.
1885
1886 * pass_manager.h: New.
1887
1888 * cgraphunit.c (cgraph_add_new_function): Update for moves
1889 of globals to fields of pass_manager.
1890 (analyze_function): Likewise.
1891 (expand_function): Likewise.
1892 (ipa_passes): Likewise.
1893 (compile): Likewise.
1894
1895 * context.c (context::context): New.
1896 * context.h (context::context): New.
1897 (context::get_passes): New.
1898 (context::passes_): New.
1899
1900 * lto-cgraph.c (input_node): Update for moves of globals to
1901 fields of pass_manager.
1902
1903 * passes.c (all_passes): Remove, in favor of a field of the
1904 same name within the new class pass_manager.
1905 (all_small_ipa_passes): Likewise.
1906 (all_lowering_passes): Likewise.
1907 (all_regular_ipa_passes): Likewise.
1908 (all_late_ipa_passes): Likewise.
1909 (all_lto_gen_passes): Likewise.
1910 (passes_by_id): Likewise.
1911 (passes_by_id_size): Likewise.
1912 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
1913 the new class pass_manager.
1914 (set_pass_for_id): Convert to...
1915 (pass_manager::set_pass_for_id): ...method.
1916 (get_pass_for_id): Convert to...
1917 (pass_manager::get_pass_for_id): ...method.
1918 (register_one_dump_file): Move body of implementation into...
1919 (pass_manager::register_one_dump_file): ...here.
1920 (register_dump_files_1): Convert to...
1921 (pass_manager::register_dump_files_1): ...method.
1922 (register_dump_files): Convert to...
1923 (pass_manager::register_dump_files): ...method.
1924 (create_pass_tab): Update for moves of globals to fields of
1925 pass_manager.
1926 (dump_passes): Move body of implementation into...
1927 (pass_manager::dump_passes): ...here.
1928 (register_pass): Move body of implementation into...
1929 (pass_manager::register_pass): ...here.
1930 (init_optimization_passes): Convert into...
1931 (pass_manager::pass_manager): ...constructor for new
1932 pass_manager class, and initialize the pass_lists array.
1933 (check_profile_consistency): Update for moves of globals to
1934 fields of pass_manager.
1935 (dump_profile_report): Move body of implementation into...
1936 (pass_manager::dump_profile_report): ...here.
1937 (ipa_write_summaries_1): Update for moves of pass lists from
1938 being globals to fields of pass_manager.
1939 (ipa_write_optimization_summaries): Likewise.
1940 (ipa_read_summaries): Likewise.
1941 (ipa_read_optimization_summaries): Likewise.
1942 (execute_all_ipa_stmt_fixups): Likewise.
1943
1944 * statistics.c (statistics_fini): Update for moves of globals to
1945 fields of pass_manager.
1946
1947 * toplev.c (general_init): Replace call to
1948 init_optimization_passes with construction of the pass_manager
1949 instance.
1950
1951 * tree-pass.h (all_passes): Remove, in favor of a field of the
1952 same name within the new class pass_manager.
1953 (all_small_ipa_passes): Likewise.
1954 (all_lowering_passes): Likewise.
1955 (all_regular_ipa_passes): Likewise.
1956 (all_lto_gen_passes): Likewise.
1957 (all_late_ipa_passes): Likewise.
1958 (passes_by_id): Likewise.
1959 (passes_by_id_size): Likewise.
1960 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
1961 the new class pass_manager.
1962 (get_pass_for_id): Remove.
1963
02cab1c1
RE
19642013-07-30 Richard Earnshaw <rearnsha@arm.com>
1965
1966 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
1967 configs.
1968
b2d3f886
RE
19692013-07-30 Richard Earnshaw <rearnsha@arm.com>
1970
1971 * arm.md (mulhi3): New expand pattern.
1972
605e86fa
JH
19732013-07-30 Jan Hubicka <jh@suse.cz>
1974 Martin Liska <marxin.liska@gmail.com>
1975
1976 * profile.c (compute_value_histograms): Do not ICE when
1977 there is mismatch only on some counters.
1978
14a87636
ZC
19792013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1980
1981 PR rtl-optimization/57637
1982 * function.c (move_insn_for_shrink_wrap): Also check the
1983 GEN set of the LIVE problem for the liveness analysis
1984 if it exists, otherwise give up.
1985
0100cd3f
BS
19862013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1987
1988 PR tree-optimization/57993
1989 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
1990 replaced statement in the candidate table.
1991 (phi_add_costs): Return infinite cost when the hidden basis does
1992 not dominate all phis on which the candidate is dependent.
1993 (replace_one_candidate): Record replaced statement in the
1994 candidate table.
1995
f0d811f7
JR
19962013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
1997
1998 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
1999 (ashlv2si3): New expander.
2000 (*ashlv2si3_i): New define_insn_and_split.
2001 * predicates.md (float_operation): Allow patterns with three
2002 basic sub-patterns.
2003
bba33211
JR
2004 PR rtl-optimization/58021
2005 * mode-switching.c (create_pre_exit): Always split off preceding
2006 insns if we are not at the basic block head.
2007
ff3f3951
MR
20082013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2009
2010 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2011 (UCLIBC_DYNAMIC_LINKER): New macro.
2012 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2013 `-mnan=2008'.
2014 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2015 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2016 `-mnan=2008'.
2017 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2018 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2019 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2020 for SF and DF modes. Use ieee_quad_format for TF mode.
2021 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2022 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2023 (mips_option_override): Handle `-mnan=legacy'.
2024 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2025 `-mabs=2008' and `-mnan=2008'.
2026 (OPTION_DEFAULT_SPECS): Add "nan" default.
2027 (ASM_SPEC): Handle `-mnan='.
2028 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2029 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2030 comment accordingly.
2031 (neg<mode>2): Likewise.
2032 * config/mips/mips.opt (mabs, mnan): New options.
2033 * doc/install.texi (Configuration): Document `--with-nan=' option.
2034 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2035 `-mnan=' options.
2036 (MIPS Options): Document them.
2037 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2038 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2039 * configure: Regenerate.
2040 * config.in: Regenerate.
2041
576e2f37
UB
20422013-07-29 Uros Bizjak <ubizjak@gmail.com>
2043
2044 * config/i386/i386.md (float post-reload splitters): Do not check
2045 for subregs of SSE registers.
2046
20472013-07-29 Uros Bizjak <ubizjak@gmail.com>
2048 H.J. Lu <hongjiu.lu@intel.com>
2049
2050 PR target/57954
2051 PR target/57988
2052 * config/i386/i386.md (post-reload splitter
2053 to avoid partial SSE reg dependency stalls): New pattern.
2054
a71f0749
DV
20552013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2056
2057 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2058
4c97f1cc
JR
20592013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2060
2061 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2062 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2063
b0c733d8
CC
20642013-07-26 Cary Coutant <ccoutant@google.com>
2065
2066 * dwarf2out.c (die_checksum_ordered): Don't include template
2067 instantiations in signature.
2068 (is_template_parameter): New function.
2069 (is_template_instantiation): New function.
2070 (generate_skeleton_bottom_up): Don't include template instantiations
2071 in type unit DIE.
2072 (generate_skeleton): Likewise.
2073 (break_out_comdat_types): Move recursive call to break out nested
2074 types earlier.
2075 (prune_unused_types_mark_generic_parms_dies): Call
2076 is_template_parameter.
2077
040d8a1c
IB
20782013-07-26 Ian Bolton <ian.bolton@arm.com>
2079
2080 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2081 uses vector registers.
89fdc743 2082 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 2083
85bd4ac6 20842013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 2085 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT
2086
2087 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 2088 where appropriate.
85bd4ac6
KT
2089 * config/arm/arm.md (movcond_addsi): New splitter.
2090
b6051207
SA
20912013-07-25 Sterling Augustine <saugustine@google.com>
2092
2093 * dwarf2out.c (size_of_pubnames): Move code to...
2094 (include_pubname_in_output): ...here. New.
2095 (want_pubnames): Rearrange.
2096 (output_pubnames): Call include_pubname_in_output. Move assertion.
2097
41a7c215
CM
20982013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2099
2100 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2101
21022013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2103
2104 PR target/38836
2105 * doc/extend.texi: Remove obsolete builtins. Fix
2106 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2107
49bde175
JH
21082013-07-25 Jan Hubicka <jh@suse.cz>
2109
2110 * cgraph.c (release_function_body): Break out from ...
2111 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2112 and DECL_ARGUMENTS.
2113 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2114 old_tree in the map.
2115 (create_specialized_node): Update.
2116 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2117 into index.
41a7c215
CM
2118 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2119 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH
2120 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2121 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2122 DECL_RESULT.
2123
4870352d
KT
21242013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2125
2126 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2127 addsi3_carryin_alt2_<optab>): Correct output template.
2128
1d6aee1c
KT
21292013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2130
2131 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2132 Adjust for arm_restrict_it.
2133 Remove trailing whitespace.
2134
c7cafd75
MK
21352013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2136
017d38f5
MK
2137 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2138 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2139
c7cafd75
MK
2140 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2141
00b7527b
VM
21422013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2143
2144 PR rtl-optimization/57960
2145 * lra-constraints.c (process_alt_operands): Use the right mode
2146 when checking strict_low.
2147
4f63dfc6
JH
21482013-07-25 Jan Hubicka <jh@suse.cz>
2149
2150 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2151 * cgraph.c (cgraph_remove_node): Do not release function body
2152 when in cgraph streaming.
2153 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM
2154 in other partitions are not considered reachable; fix handling of
2155 clones.
4f63dfc6 2156
e6dcfa18
RR
21572013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2158
2159 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2160
0b93d3b6
RR
21612013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2162
2163 PR target/19599
2164 PR target/57731
bb80c2eb 2165 PR target/57837
e6dcfa18 2166 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 2167 Ss with US. Adjust output for v5 and v4t.
41a7c215 2168 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 2169
0b93d3b6
RR
2170 * config/arm/constraints.md ("Ss"): Rename to US.
2171
690688b3
TG
21722013-07-25 Terry Guo <terry.guo@arm.com>
2173
2174 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2175 shift_add/shift_sub0/shift_sub1 RTXs.
2176
a4ad093b
BS
21772013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2178 Anton Blanchard <anton@au1.ibm.com>
2179
2180 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2181 (altivec_vpks<VI_char>ss): Likewise.
2182 (altivec_vpks<VI_char>us): Likewise.
2183 (altivec_vpku<VI_char>us): Likewise.
2184 (altivec_vpku<VI_char>um): Likewise.
2185
8c5005ce
DM
21862013-07-24 David Malcolm <dmalcolm@redhat.com>
2187
2188 Introduce context class.
2189
2190 * Makefile.in (CONTEXT_H): New.
2191 (OBJS): Add context.o.
2192 (toplev.o): Add CONTEXT_H to dependencies.
2193 (context.o): New.
2194
41a7c215 2195 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM
2196
2197 * context.c: New.
2198
2199 * context.h: New.
2200
ce4a9422
JR
22012013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2202
2203 PR rtl-optimization/57968
2204 * mode-switching.c (create_pre_exit): Allow instructions that
2205 don't set a return register to need a non-exit mode.
2206
493f4c9e
BS
22072013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2208 Anton Blanchard <anton@au1.ibm.com>
2209
b5860fd3
WS
2210 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2211 operands to vperm for little endian.
2212 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2213 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 2214
73792b92
BS
22152013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2216 Anton Blanchard <anton@au1.ibm.com>
2217
2218 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2219 two operands for little-endian.
2220
14c21302
SE
22212013-07-23 Steve Ellcey <sellcey@mips.com>
2222
2223 * config/mips/mips.c (mips_case_values_threshold): New.
2224 (TARGET_CASE_VALUES_THRESHOLD): Define.
2225
a6056198
BS
22262013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2227 Anton Blanchard <anton@au1.ibm.com>
2228
2229 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2230 selection of field for vector splat in little endian mode.
2231
dd7a40e1
MM
22322013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2233
2234 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2235 expanders to rs6000.md.
2236 (ior<mode>3): Likewise.
2237 (and<mode>3): Likewise.
2238 (one_cmpl<mode>2): Likewise.
2239 (nor<mode>3): Likewise.
2240 (andc<mode>3): Likewise.
2241 (eqv<mode>3): Likewise.
2242 (nand<mode>3): Likewise.
2243 (orc<mode>3): Likewise.
2244
2245 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2246 declaration.
2247
2248 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2249 to split multi-word logical operations.
2250 (rs6000_split_logical_di): Likewise.
2251 (rs6000_split_logical): Likewise.
2252
2253 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2254 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2255 and allow TImode operations in 32-bit.
2256 (vsx_and<mode>3_64bit): Likewise.
2257 (vsx_ior<mode>3_32bit): Likewise.
2258 (vsx_ior<mode>3_64bit): Likewise.
2259 (vsx_xor<mode>3_32bit): Likewise.
2260 (vsx_xor<mode>3_64bit): Likewise.
2261 (vsx_one_cmpl<mode>2_32bit): Likewise.
2262 (vsx_one_cmpl<mode>2_64bit): Likewise.
2263 (vsx_nor<mode>3_32bit): Likewise.
2264 (vsx_nor<mode>3_64bit): Likewise.
2265 (vsx_andc<mode>3_32bit): Likewise.
2266 (vsx_andc<mode>3_64bit): Likewise.
2267 (vsx_eqv<mode>3_32bit): Likewise.
2268 (vsx_eqv<mode>3_64bit): Likewise.
2269 (vsx_nand<mode>3_32bit): Likewise.
2270 (vsx_nand<mode>3_64bit): Likewise.
2271 (vsx_orc<mode>3_32bit): Likewise.
2272 (vsx_orc<mode>3_64bit): Likewise.
2273
2274 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2275 logical types in GPRs.
2276
2277 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2278 logical insns to rs6000.md, and allow TImode operations in
2279 32-bit.
2280 (altivec_ior<mode>3): Likewise.
2281 (altivec_xor<mode>3): Likewise.
2282 (altivec_one_cmpl<mode>2): Likewise.
2283 (altivec_nor<mode>3): Likewise.
2284 (altivec_andc<mode>3): Likewise.
2285
2286 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2287 attributes for moving the 128-bit logical operations into
2288 rs6000.md.
2289 (BOOL_REGS_OUTPUT): Likewise.
2290 (BOOL_REGS_OP1): Likewise.
2291 (BOOL_REGS_OP2): Likewise.
2292 (BOOL_REGS_UNARY): Likewise.
2293 (BOOL_REGS_AND_CR0): Likewise.
2294 (one_cmpl<mode>2): Add support for DI logical operations on
2295 32-bit, splitting the operations to 32-bit.
2296 (anddi3): Likewise.
2297 (iordi3): Likewise.
2298 (xordi3): Likewise.
2299 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2300 changes to combine the 32/64-bit code, allow logical operations on
2301 TI mode in 32-bit, and to use similar match_operator patterns like
2302 scalar mode uses. Combine the Altivec and VSX code for logical
2303 operations, and move it here.
2304 (ior<mode>3, 128-bit types): Likewise.
2305 (xor<mode>3, 128-bit types): Likewise.
2306 (one_cmpl<mode>3, 128-bit types): Likewise.
2307 (nor<mode>3, 128-bit types): Likewise.
2308 (andc<mode>3, 128-bit types): Likewise.
2309 (eqv<mode>3, 128-bit types): Likewise.
2310 (nand<mode>3, 128-bit types): Likewise.
2311 (orc<mode>3, 128-bit types): Likewise.
2312 (and<mode>3_internal): Likewise.
2313 (bool<mode>3_internal): Likewise.
2314 (boolc<mode>3_internal1): Likewise.
2315 (boolc<mode>3_internal2): Likewise.
2316 (boolcc<mode>3_internal1): Likewise.
2317 (boolcc<mode>3_internal2): Likewise.
2318 (eqv<mode>3_internal1): Likewise.
2319 (eqv<mode>3_internal2): Likewise.
2320 (one_cmpl1<mode>3_internal): Likewise.
2321
9b69d4b4
DH
23222013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2323
2324 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2325 Rename flag_stack_usage to flag_stack_usage_info.
2326
450b0ebb
DH
23272013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2328
a5f257fd
ME
2329 * config/microblaze/sync.md: New file.
2330 * config/microblaze/microblaze.md: Include sync.md
2331 * config/microblaze/microblaze.c: Add print_operand 'y'.
2332 * config/microblaze/constraints.md: Add memory_contraint
2333 'Q' which is a single register.
450b0ebb 2334
49fba14f
EB
23352013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2336
2337 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2338
c7b3b99f
PCC
23392013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2340
2341 * reload.c (find_reloads): Exit loop once we find this operand
2342 cannot be reloaded somehow for this alternative.
2343
2344 * reload.c (find_reloads): Exit loop once we find a hard register.
2345
2346 * rtlanal.c (computed_jump_p): Exit loop once we find label
2347 reference is used.
2348
2349 * i386.c (ix86_pad_returns): Exit loop after setting replace.
2350
2351 * cfgloopmanip.c (remove_path): Exit loop after setting
2352 irred_invalidated.
2353
2354 * gensupport.c (subst_dup): Avoid loop if code is not
2355 MATCH_DUP nor MATCH_OP_DUP.
2356
c8fbf1fa
NBJ
23572013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
2358
2359 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
2360
9259db42
YZ
23612013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2362
2363 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
2364 true for SP_REGNUM if mode == ptr_mode.
2365 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
2366 with value R0_REGNUM + 31.
2367
00edcfbe
YZ
23682013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2369
2370 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
2371 pad pointer-typed argument downward.
2372
43be9a95
YZ
23732013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2374
2375 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
2376 and __ILP32__ when the ILP32 model is in use.
2377
28514dda
YZ
23782013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2379
2380 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
2381 (aarch64_load_symref_appropriately): In the case of
2382 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
2383 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
2384 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
2385 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
2386 if 'mode' doesn't equal to 'ptr_mode'.
2387 (aarch64_output_mi_thunk): Add an assertion on the alignment of
2388 'vcall_offset'; change to call aarch64_emit_move differently depending
2389 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
2390 to calculate the upper bound of 'vcall_offset'.
2391 (aarch64_cannot_force_const_mem): Change to also return true if
2392 mode != ptr_mode.
2393 (aarch64_legitimize_reload_address): In the case of large
2394 displacements, add new local variable 'xmode' and an assertion
2395 based on it; change to use 'xmode' to generate the new rtx and
2396 reload.
2397 (aarch64_asm_trampoline_template): Change to generate the template
2398 differently depending on TARGET_ILP32 or not; change to use
2399 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
2400 (aarch64_trampoline_size): Removed.
2401 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
2402 and replace immediate literals with it. Change to use 'ptr_mode'
2403 instead of 'DImode' and call convert_memory_address if the mode
2404 of 'fnaddr' doesn't equal to 'ptr_mode'.
2405 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
2406 to output symbol.
2407 (aarch64_elf_asm_destructor): Likewise.
2408 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
2409 on TARGET_ILP32 instead of aarch64_trampoline_size.
2410 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
2411 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
2412 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
2413 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
2414 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
2415 (storewb_pair<GPI:mode>_<P:mode>): ... this.
2416 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
2417 depending on the value of 'mode'.
2418 (add_losym_<mode>): New.
2419 (ldr_got_small_<mode>): New, based on ldr_got_small.
2420 (ldr_got_small): Remove.
2421 (ldr_got_small_sidi): New.
2422 * config/aarch64/iterators.md (P): New.
2423 (PTR): Change to 'ptr_mode' in the condition.
2424
17a819cb
YZ
24252013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2426
2427 * config.gcc (aarch64*-*-*): Support --with-abi.
2428 (aarch64*-*-elf): Support --with-multilib-list.
2429 (aarch64*-*-linux*): Likewise.
2430 (supported_defaults): Add abi to aarch64*-*-*.
2431 * configure.ac: Mention AArch64 for --with-multilib-list.
2432 * configure: Re-generated.
2433 * config/aarch64/biarchilp32.h: New file.
2434 * config/aarch64/biarchlp64.h: New file.
2435 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
2436 (ABI_SPEC): Ditto.
2437 (MULTILIB_DEFAULTS): Ditto.
2438 (DRIVER_SELF_SPECS): Ditto.
2439 (ASM_SPEC): Update to also substitute -mabi.
2440 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
2441 file whose name depends on -mabi= and -mbig-endian.
2442 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
2443 TARGET_ILP32.
2444 (POINTER_SIZE): New define.
2445 (POINTERS_EXTEND_UNSIGNED): Ditto.
2446 (enum aarch64_abi_type): New enumeration tag.
2447 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
2448 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
2449 (TARGET_ILP32): New define.
2450 * config/aarch64/aarch64.opt (mabi): New.
2451 (aarch64_abi): New.
2452 (ilp32, lp64): New values for -mabi.
2453 * config/aarch64/t-aarch64 (comma): New define.
2454 (MULTILIB_OPTIONS): Ditto.
2455 (MULTILIB_DIRNAMES): Ditto.
2456 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
2457 * doc/invoke.texi: Document -mabi for AArch64.
2458
9f9cbdce
GJL
24592013-07-23 Georg-Johann Lay <avr@gjlay.de>
2460
2461 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
2462
7061977b
BS
24632013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2464 Anton Blanchard <anton@au1.ibm.com>
2465
2466 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
2467 endianness when selecting field to splat.
2468
75a41878
EC
24692013-07-22 Eric Christopher <echristo@gmail.com>
2470
2471 * dwarf2out.c (die_odr_checksum): New function to use
2472 CHECKSUM_ macros and ULEB128 for DIE tag.
2473 (generate_type_signature): Use.
41a7c215 2474
38ae58ca
EB
24752013-07-22 Eric Botcazou <ebotcazou@adacore.com>
2476
2477 * config.gcc (sparc*-*-*): Accept leon3 processor.
2478 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
2479 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
2480 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
2481 * config/sparc/sparc.opt (enum processor_type): Add leon3.
2482 (mfix-ut699): Adjust comment.
2483 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
2484 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
2485 (CPP_CPU_SPEC): Likewise.
2486 (ASM_CPU_SPEC): Likewise.
2487 * config/sparc/sparc.c (leon3_cost): New constant.
2488 (sparc_option_override): Add leon3 support.
2489 (mem_ref): New function.
2490 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
2491 (sparc_do_work_around_errata): Look into the instruction in the delay
2492 slot and adjust accordingly. Add fix for the data cache nullify issues
2493 of the UT699. Change insertion position for the NOP.
2494 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
2495 (leon3_load): New reservation.
2496 (leon_store): Bump latency to 2.
2497 (grfpu): New automaton.
2498 (grfpu_alu): New unit.
2499 (grfpu_ds): Likewise.
2500 (leon_fp_alu): Adjust.
2501 (leon_fp_mult): Delete.
2502 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
2503 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
2504 * config/sparc/sparc.md (cpu): Add leon3.
2505 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
2506 (swapsi): Likewise.
2507 (atomic_test_and_set): Likewise.
2508 (ldstub): Likewise.
2509
04dfc6df
JU
25102013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
2511
2512 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
2513 default for R5900 targets.
2514 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
2515 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
2516 * config/mips/mips.c (mips_option_override): Report an error for
2517 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
2518 for -march=r5900 -mhard-float.
2519
c7b3b99f 25202013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC
2521
2522 * df-problems.c (can_move_insns_across): Exit loop once we
2523 find a non-fixed, non-global register.
2524
2525 * ipa-pure-const.c (propagate_nothrow): Exit loop after
2526 setting can_throw.
2527
2528 * omega.c (omega_eliminate_red): Break after setting red_found.
2529 (omega_problem_has_red_equations): Similarly after setting found.
2530 (omega_query_variable): Similarly after setting coupled.
2531
9b6e6981
MP
25322013-07-22 Marek Polacek <polacek@redhat.com>
2533
2534 * gimplify.c: Don't include gimple.h twice.
2535
5e5f7673
KT
25362013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2537
2538 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
2539 instead of TARGET_THUMB1.
2540 (Pz): New constraint.
2541 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
2542 encodings.
2543 (compare_negsi_si): Likewise.
2544 (compare_addsi2_op0): Likewise.
2545 (compare_addsi2_op1): Likewise.
2546 (addsi3_carryin_<optab>): Likewise.
2547 (addsi3_carryin_alt2_<optab>): Likewise.
2548 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
2549 for arm_restrict_it.
2550 (subsi3_carryin): Likewise.
2551 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
2552 (minmax_arithsi): Disable for arm_restrict_it.
2553 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
2554 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
2555 (satsi_<SAT:code>_shift): Likewise.
2556 (arm_shiftsi3): Add alternative for 16-bit encoding.
2557 (arm32_movhf): Disable for arm_restrict_it.
2558 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
2559 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
2560
859abddd
SN
25612013-07-22 Sofiane Naci <sofiane.naci@arm.com>
2562
2563 * config/arm/arm.md (attribute "insn"): Delete.
2564 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
2565 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
2566 (not_shiftsi): Update for attribute change.
2567 (not_shiftsi_compare0): Likewise.
2568 (not_shiftsi_compare0_scratch): Likewise.
2569 (arm_one_cmplsi2): Likewise.
2570 (thumb1_one_cmplsi2): Likewise.
2571 (notsi_compare0): Likewise.
2572 (notsi_compare0_scratch): Likewise.
2573 (thumb1_movdi_insn): Likewise.
2574 (arm_movsi_insn): Likewise.
2575 (movhi_insn_arch4): Likewise.
2576 (movhi_bytes): Likewise.
2577 (arm_movqi_insn): Likewise.
2578 (thumb1_movqi_insn): Likewise.
2579 (arm32_movhf): Likewise.
2580 (thumb1_movhf): Likewise.
2581 (arm_movsf_soft_insn): Likewise.
2582 (thumb1_movsf_insn): Likewise.
2583 (thumb_movdf_insn): Likewise.
2584 (movsicc_insn): Likewise.
2585 (movsfcc_soft_insn): Likewise.
2586 (and_scc): Likewise.
2587 (cond_move): Likewise.
2588 (if_move_not): Likewise.
2589 (if_not_move): Likewise.
2590 (if_shift_move): Likewise.
2591 (if_move_shift): Likewise.
2592 (if_shift_shift): Likewise.
2593 (if_not_arith): Likewise.
2594 (if_arith_not): Likewise.
2595 (cond_move_not): Likewise.
2596 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
2597 (neon_mov<mode>): Likewise.
2598 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
2599 (thumb2_movsi_vfp): Likewise.
2600 (movsf_vfp): Likewise.
2601 (thumb2_movsf_vfp): Likewise.
371e77e3
UB
2602 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
2603 change.
859abddd
SN
2604 (cortexa7_older_only): Likewise.
2605 (cortexa7_younger): Likewise.
2606 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
2607 (1020alu_shift_op): Likewise.
2608 (1020alu_shift_reg_op): Likewise.
2609 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
2610 (alu_shift_op): Likewise.
2611 (alu_shift_reg_op): Likewise.
2612 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
2613 (11_alu_shift_op): Likewise.
2614 (11_alu_shift_reg_op): Likewise.
2615 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
2616 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
2617 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2618 change.
859abddd
SN
2619 (cortex_a15_alu_shift): Likewise.
2620 (cortex_a15_alu_shift_reg): Likewise.
2621 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
2622 (cortex_a5_alu_shift): Likewise.
371e77e3
UB
2623 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
2624 change.
859abddd 2625 (cortex_a53_alu_shift): Likewise.
371e77e3
UB
2626 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
2627 change.
859abddd
SN
2628 (cortex_a7_alu_reg): Likewise.
2629 (cortex_a7_alu_shift): Likewise.
2630 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2631 (cortex_a8_alu_shift): Likewise.
2632 (cortex_a8_alu_shift_reg): Likewise.
2633 (cortex_a8_mov): Likewise.
2634 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2635 (cortex_a9_dp_shift): Likewise.
2636 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
2637 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
2638 (cortex_r4_mov): Likewise.
2639 (cortex_r4_alu_shift): Likewise.
2640 (cortex_r4_alu_shift_reg): Likewise.
2641 * config/arm/fa526.md (526_alu_op): Update for attribute change.
2642 (526_alu_shift_op): Likewise.
2643 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
2644 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
2645 (626te_alu_shift_op): Likewise.
2646 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
2647 (726te_alu_op): Likewise.
2648 (726te_alu_shift_op): Likewise.
2649 (726te_alu_shift_reg_op): Likewise.
2650 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
2651 (mp626_alu_shift_op): Likewise.
2652 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
2653 (pj4_alu_e1_conds): Likewise.
2654 (pj4_alu): Likewise.
2655 (pj4_alu_conds): Likewise.
2656 (pj4_shift): Likewise.
2657 (pj4_shift_conds): Likewise.
2658 (pj4_alu_shift): Likewise.
2659 (pj4_alu_shift_conds): Likewise.
2660
95b97fac
KT
26612013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2662
2663 * config/arm/predicates.md (shiftable_operator_strict_it):
2664 New predicate.
2665 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
2666 Disable cond_exec version for arm_restrict_it.
2667 (thumb2_smaxsi3): Convert to generate cond_exec.
2668 (thumb2_sminsi3): Likewise.
2669 (thumb32_umaxsi3): Likewise.
2670 (thumb2_uminsi3): Likewise.
2671 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
2672 (thumb2_neg_abssi2): Likewise.
2673 (thumb2_mov_scc): Add alternative for 16-bit encoding.
2674 (thumb2_movsicc_insn): Adjust alternatives.
2675 (thumb2_mov_negscc): Disable for arm_restrict_it.
2676 (thumb2_mov_negscc_strict_it): New pattern.
2677 (thumb2_mov_notscc_strict_it): New pattern.
2678 (thumb2_mov_notscc): Disable for arm_restrict_it.
2679 (thumb2_ior_scc): Likewise.
2680 (thumb2_ior_scc_strict_it): New pattern.
2681 (thumb2_cond_move): Adjust for arm_restrict_it.
2682 (thumb2_cond_arith): Disable for arm_restrict_it.
2683 (thumb2_cond_arith_strict_it): New pattern.
2684 (thumb2_cond_sub): Adjust for arm_restrict_it.
2685 (thumb2_movcond): Likewise.
2686 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
2687 (thumb2_zero_extendhisi2_v6): Likewise.
2688 (thumb2_zero_extendqisi2_v6): Likewise.
2689 (orsi_notsi_si): Likewise.
2690 (orsi_not_shiftsi_si): Likewise.
2691
d5a6ef82
GJL
26922013-07-22 Georg-Johann Lay <avr@gjlay.de>
2693
2694 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
2695 instruction sequence is 1 byte shorter.
2696
d9e71154
UB
26972013-07-22 Uros Bizjak <ubizjak@gmail.com>
2698
2699 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
2700 it is not needed after split.
2701
8c6381ec
IS
27022013-07-20 Iain Sandoe <iain@codesourcery.com>
2703
2704 PR target/51784
2705 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
2706 second label for nonlocal goto receivers. Don't output pic base labels
2707 unless we're producing PIC; mark that action unreachable().
2708 (ix86_save_reg): If the function contains a nonlocal label, save the
2709 PIC base reg.
2710 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
371e77e3 2711 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS
2712 (update_pic_label_number_if_needed): New.
2713 (machopic_output_function_base_name): Adjust for nonlocal receiver
2714 case.
2715 (machopic_should_output_picbase_label): New.
2716 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
2717 (nonlocal_goto_receiver): New insn and split.
2718
d05d0709
JG
27192013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
2720
2721 * config/aarch64/aarch64-builtins.c
2722 (aarch64_fold_builtin): Fold abs in all modes.
2723 * config/aarch64/aarch64-simd-builtins.def
2724 (abs): Enable for all modes.
2725 * config/aarch64/arm_neon.h
2726 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
2727 (vabs_f64): Add missing intrinsic.
2728
ba590e63
IB
27292013-07-19 Ian Bolton <ian.bolton@arm.com>
2730
2731 * config/aarch64/arm_neon.h (vabs_s64): New function
2732
02371798
GJL
27332013-07-19 Georg-Johann Lay <avr@gjlay.de>
2734
2735 PR target/57516
2736 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
2737 * config/avr/avr.md (adjust_len): Add `round'.
2738 * config/avr/avr-protos.h (avr_out_round): New prototype.
2739 (avr_out_plus): Add `out_label' argument.
2740 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
2741 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
2742 Handle the case where `insn' is just a pattern.
2743 (avr_out_bitop): Handle the case where `insn' is just a pattern.
2744 (avr_out_round): New function.
2745 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
2746
b950ac25
DH
27472013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
2748
2749 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2750 Add check for flag_stack_usage to handle -fstack-usage support
2751
ed1fd999
PH
27522013-07-18 Pat Haugen <pthaugen@us.ibm.com>
2753
2754 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
2755 interaction for new Power8 flags and VSX.
2756
df9dda2d
ST
27572013-07-18 Sriraman Tallam <tmsriram@google.com>
2758
2759 PR middle-end/57698
2760 * tree-inline.c (expand_call_inline): Emit errors during
2761 early_inlining only if optimization is not turned on.
2762
0f9cdef4
DM
27632013-07-18 David Malcolm <dmalcolm@redhat.com>
2764
2765 * passes.def: New.
2766
2767 * passes.c (init_optimization_passes): Move the construction of
2768 the pass hierarchy into a new passes.def file.
2769
2770 * Makefile.in (passes.o): Add dependency on passes.def.
2771
2efa4087
DM
27722013-07-18 David Malcolm <dmalcolm@redhat.com>
2773
2774 * passes.c (init_optimization_passes): Introduce macros for
2775 constructing the tree of passes (INSERT_PASSES_AFTER,
2776 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
2777 TERMINATE_PASS_LIST).
2778
bc404e1b
VM
27792013-07-18 Vladimir Makarov <vmakarov@redhat.com>
2780 Wei Mi <wmi@google.com>
2781
2782 PR rtl-optimization/57878
2783 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
2784 top.
2785 (reload_pseudo_compare_func): Check nregs first for reload
2786 pseudos.
2787
d675843f
DM
27882013-07-18 David Malcolm <dmalcolm@redhat.com>
2789
2790 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
2791
c7b3b99f 27922013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC
2793
2794 * read-rtl.c (validate_const_int): Once an invalid character is
2795 seen, quit the loop.
2796
2797 * gengtype.c (write_roots): Similarly once we find the "deletable"
2798 or "if_marked" option.
2799
52fceb44
SN
28002013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2801
2802 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
2803 "xtab" and "sat". Move value "clz" from here to ...
2804 (attriubte "type"): ... here.
2805 (satsi_<SAT:code>): Delete "insn" attribute.
2806 (satsi_<SAT:code>_shift): Likewise.
2807 (arm_zero_extendqisi2addsi): Likewise.
2808 (arm_extendqisi2addsi): Likewise.
2809 (clzsi2): Update for attribute changes.
2810 (rbitsi2): Likewise.
371e77e3
UB
2811 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
2812 attribute.
52fceb44
SN
2813 (arm_usatsihi): Likewise.
2814 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2815
006bd006
SN
28162013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2817
2818 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB
2819 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
2820 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN
2821 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
2822 in alphabetical order.
2823 (attribute "core_cycles"): Update for attribute changes.
2824 (arm_addsi3): Likewise.
2825 (addsi3_compare0): Likewise.
2826 (addsi3_compare0_scratch): Likewise.
2827 (addsi3_compare_op1): Likewise.
2828 (addsi3_compare_op2): Likewise.
2829 (compare_addsi2_op0): Likewise.
2830 (compare_addsi2_op1): Likewise.
2831 (addsi3_carryin_shift_<optab>): Likewise.
2832 (subsi3_carryin_shift): Likewise.
2833 (rsbsi3_carryin_shift): Likewise.
2834 (arm_subsi3_insn): Likewise.
2835 (subsi3_compare0): Likewise.
2836 (subsi3_compare): Likewise.
2837 (arm_andsi3_insn): Likewise.
2838 (thumb1_andsi3_insn): Likewise.
2839 (andsi3_compare0): Likewise.
2840 (andsi3_compare0_scratch): Likewise.
2841 (zeroextractsi_compare0_scratch
2842 (andsi_not_shiftsi_si): Likewise.
2843 (iorsi3_insn): Likewise.
2844 (iorsi3_compare0): Likewise.
2845 (iorsi3_compare0_scratch): Likewise.
2846 (arm_xorsi3): Likewise.
2847 (thumb1_xorsi3_insn): Likewise.
2848 (xorsi3_compare0): Likewise.
2849 (xorsi3_compare0_scratch): Likewise.
2850 (satsi_<SAT:code>_shift): Likewise.
2851 (rrx): Likewise.
2852 (arm_shiftsi3): Likewise.
2853 (shiftsi3_compare0): Likewise.
2854 (not_shiftsi): Likewise.
2855 (not_shiftsi_compare0): Likewise.
2856 (not_shiftsi_compare0_scratch): Likewise.
2857 (arm_one_cmplsi2): Likewise.
2858 (thumb_one_complsi2): Likewise.
2859 (notsi_compare0): Likewise.
2860 (notsi_compare0_scratch): Likewise.
2861 (thumb1_zero_extendhisi2): Likewise.
2862 (arm_zero_extendhisi2): Likewise.
2863 (arm_zero_extendhisi2_v6): Likewise.
2864 (arm_zero_extendhisi2addsi): Likewise.
2865 (thumb1_zero_extendqisi2): Likewise.
2866 (thumb1_zero_extendqisi2_v6): Likewise.
2867 (arm_zero_extendqisi2): Likewise.
2868 (arm_zero_extendqisi2_v6): Likewise.
2869 (arm_zero_extendqisi2addsi): Likewise.
2870 (thumb1_extendhisi2): Likewise.
2871 (arm_extendhisi2): Likewise.
2872 (arm_extendhisi2_v6): Likewise.
2873 (arm_extendqisi): Likewise.
2874 (arm_extendqisi_v6): Likewise.
2875 (arm_extendqisi2addsi): Likewise.
2876 (thumb1_extendqisi2): Likewise.
2877 (thumb1_movdi_insn): Likewise.
2878 (arm_movsi_insn): Likewise.
2879 (movsi_compare0): Likewise.
2880 (movhi_insn_arch4): Likewise.
2881 (movhi_bytes): Likewise.
2882 (arm_movqi_insn): Likewise.
2883 (thumb1_movqi_insn): Likewise.
2884 (arm32_movhf): Likewise.
2885 (thumb1_movhf): Likewise.
2886 (arm_movsf_soft_insn): Likewise.
2887 (thumb1_movsf_insn): Likewise.
2888 (movdf_soft_insn): Likewise.
2889 (thumb_movdf_insn): Likewise.
2890 (arm_cmpsi_insn): Likewise.
2891 (cmpsi_shiftsi): Likewise.
2892 (cmpsi_shiftsi_swp): Likewise.
2893 (arm_cmpsi_negshiftsi_si): Likewise.
2894 (movsicc_insn): Likewise.
2895 (movsfcc_soft_insn): Likewise.
2896 (arith_shiftsi): Likewise.
2897 (arith_shiftsi_compare0
2898 (arith_shiftsi_compare0_scratch
2899 (sub_shiftsi): Likewise.
2900 (sub_shiftsi_compare0
2901 (sub_shiftsi_compare0_scratch
2902 (and_scc): Likewise.
2903 (cond_move): Likewise.
2904 (if_plus_move): Likewise.
2905 (if_move_plus): Likewise.
2906 (if_move_not): Likewise.
2907 (if_not_move): Likewise.
2908 (if_shift_move): Likewise.
2909 (if_move_shift): Likewise.
2910 (if_shift_shift): Likewise.
2911 (if_not_arith): Likewise.
2912 (if_arith_not): Likewise.
2913 (cond_move_not): Likewise.
2914 (thumb1_ashlsi3): Set type attribute.
2915 (thumb1_ashrsi3): Likewise.
2916 (thumb1_lshrsi3): Likewise.
2917 (thumb1_rotrsi3): Likewise.
2918 (shiftsi3_compare0_scratch): Likewise.
2919 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
2920 (neon_mov<mode>): Likewise.
371e77e3
UB
2921 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
2922 attribute changes.
006bd006
SN
2923 (thumb2_movsi_insn): Likewise.
2924 (thumb2_cmpsi_neg_shiftsi): Likewise.
2925 (thumb2_extendqisi_v6): Likewise.
2926 (thumb2_zero_extendhisi2_v6): Likewise.
2927 (thumb2_zero_extendqisi2_v6): Likewise.
2928 (thumb2_shiftsi3_short): Likewise.
2929 (thumb2_addsi3_compare0_scratch): Likewise.
2930 (orsi_not_shiftsi_si): Likewise.
2931 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
2932 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
2933 changes.
2934 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2935 (1020alu_shift_op): Likewise.
2936 (1020alu_shift_reg_op): Likewise.
2937 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2938 (alu_shift_op): Likewise.
2939 (alu_shift_reg_op): Likewise.
2940 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2941 (11_alu_shift_op): Likewise.
2942 (11_alu_shift_reg_op): Likewise.
2943 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2944 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
2945 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2946 changes.
006bd006
SN
2947 (cortex_a15_alu_shift): Likewise.
2948 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB
2949 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
2950 changes.
006bd006
SN
2951 (cortex_a5_alu_shift): Likewise.
2952 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
2953 changes.
2954 (cortex_a53_alu_shift): Likewise.
2955 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
2956 changes.
2957 (cortex_a7_alu_reg): Likewise.
2958 (cortex_a7_alu_shift): Likewise.
371e77e3
UB
2959 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
2960 changes.
006bd006
SN
2961 (cortex_a8_alu_shift): Likewise.
2962 (cortex_a8_alu_shift_reg): Likewise.
2963 (cortex_a8_mov): Likewise.
2964 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2965 (cortex_a9_dp_shift): Likewise.
371e77e3
UB
2966 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
2967 changes.
2968 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
2969 changes.
006bd006
SN
2970 (cortex_r4_mov): Likewise.
2971 (cortex_r4_alu_shift): Likewise.
2972 (cortex_r4_alu_shift_reg): Likewise.
2973 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2974 (526_alu_shift_op): Likewise.
2975 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2976 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2977 (626te_alu_shift_op): Likewise.
2978 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
2979 (726te_alu_op): Likewise.
2980 (726te_alu_shift_op): Likewise.
2981 (726te_alu_shift_reg_op): Likewise.
2982 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2983 (mp626_alu_shift_op): Likewise.
2984 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
2985 (pj4_alu_e1_conds): Likewise.
2986 (pj4_alu): Likewise.
2987 (pj4_alu_conds): Likewise.
2988 (pj4_shift): Likewise.
2989 (pj4_shift_conds): Likewise.
2990 (pj4_alu_shift): Likewise.
2991 (pj4_alu_shift_conds): Likewise.
371e77e3
UB
2992 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
2993 changes.
006bd006
SN
2994 (cortexa7_older_only): Likewise.
2995 (cortexa7_younger): Likewise.
2996
651df1b2
DM
29972013-07-18 David Malcolm <dmalcolm@redhat.com>
2998
2999 * ipa-pure-const.c (generate_summary): Rename to...
3000 (pure_const_generate_summary): ... this.
3001
ea717bd7
IS
30022013-07-17 Iain Sandoe <iain@codesourcery.com>
3003
3004 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3005
a1259a13
YR
30062013-07-17 Yvan Roux <yvan.roux@linaro.org>
3007
3008 PR target/57909
3009 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3010 usage in HI mode.
3011
167f68ed
AK
30122013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3013
3014 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3015 enabled without -march=zEC12.
3016 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3017 flags to be set.
3018
f900a982
MR
30192013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3020
3021 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3022 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3023 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3024 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3025 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3026 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3027 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3028 ISA_HAS_FP4.
3029 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3030 and ISA_HAS_NMADD3_NMSUB3.
3031 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3032 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3033 (nmsub4<mode>, nmsub3<mode>): Likewise.
3034 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3035
d4ed27eb
MR
30362013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3037
3038 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3039 TARGET_MIPS5400 checking.
3040
7a5add18
PB
30412013-07-16 Jakub Jelinek <jakub@redhat.com>
3042 Peter Bergner <bergner@vnet.ibm.com>
3043
3044 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3045 registers in the comment.
3046 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3047 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3048 rather than FIRST_PSEUDO_REGISTERS.
3049
30502013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB
3051
3052 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3053 enable extra ISA flags with TARGET_HTM.
3054
d2ab0929
MR
30552013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3056
3057 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3058 Fix comment typos.
3059
839c74bc
CH
30602013-07-15 Cong Hou <congh@google.com>
3061
3062 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3063 in compare function for sorting.
3064
7a5add18 30652013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB
3066
3067 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3068 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3069 * config/rs6000/rs6000.opt: Add -mhtm option.
3070 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3071 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3072 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3073 __HTM__ if the HTM instructions are available.
3074 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3075 htm_spr_reg_operand): New define_predicates.
3076 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3077 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3078 Include htm.md.
3079 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3080 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3081 HTM builtin functions.
3082 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3083 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3084 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3085 (rs6000_builtin_mask_calculate): Likewise.
3086 (rs6000_option_override_internal): Likewise.
3087 (bdesc_htm): Add new HTM builtin support.
3088 (htm_spr_num): New function.
3089 (htm_spr_regno): Likewise.
3090 (rs6000_htm_spr_icode): Likewise.
3091 (htm_expand_builtin): Likewise.
3092 (htm_init_builtins): Likewise.
3093 (rs6000_expand_builtin): Add support for HTM builtin functions.
3094 (rs6000_init_builtins): Likewise.
371e77e3
UB
3095 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3096 option.
0258b6e4
PB
3097 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3098 (TARGET_HTM, MASK_HTM): Define macros.
3099 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3100 (FIXED_REGISTERS): Likewise.
3101 (CALL_USED_REGISTERS): Likewise.
3102 (CALL_REALLY_USED_REGISTERS): Likewise.
3103 (REG_ALLOC_ORDER): Likewise.
3104 (enum reg_class): Likewise.
3105 (REG_CLASS_NAMES): Likewise.
3106 (REG_CLASS_CONTENTS): Likewise.
3107 (REGISTER_NAMES): Likewise.
3108 (ADDITIONAL_REGISTER_NAMES): Likewise.
3109 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3110 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3111 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3112 * config/rs6000/htm.md: New file.
3113 * config/rs6000/htmintrin.h: New file.
3114 * config/rs6000/htmxlintrin.h: New file.
3115
87dd8ab0
MS
31162013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3117
3118 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3119 Define SYMBOL_TINY_GOT, update comment.
3120 * config/aarch64/aarch64.c
3121 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3122 (aarch64_expand_mov_immediate): Likewise.
3123 (aarch64_print_operand): Likewise.
3124 (aarch64_classify_symbol): Likewise.
3125 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3126 (ldr_got_tiny): Define.
3127
52d676b6
TG
31282013-07-13 Tobias Grosser <tobias@grosser.es>
3129
3130 PR tree-optimization/54094
3131 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3132 scheduling dimension for the parallelism check from the polyhedral
3133 information in the AST.
3134 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3135 in 2D + 1 form.
3136
286e8fc1
JM
31372013-07-13 Jason Merrill <jason@redhat.com>
3138
3139 * print-tree.c (debug_vec_tree): Use debug_raw.
3140 (debug_raw (vec<tree, va_gc> &)): New.
3141 (debug_raw (vec<tree, va_gc> *)): New.
3142 * tree.h: Declare them.
3143
e43257e8
BC
31442013-07-13 Bin Cheng <bin.cheng@arm.com>
3145
3146 * ifcvt.c (ifcvt_after_combine): New static variable.
3147 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3148 for size.
3149 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3150 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3151 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3152
da582d46
MR
31532013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3154
3155 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3156
18b0ea8f
MM
31572013-07-12 Michael Matz <matz@suse.de>
3158
3159 PR middle-end/55771
3160 * convert.c (convert_to_real): Reject non-float inner types.
3161
e4f0f84d
TB
31622013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3163
3164 * config/aarch64/aarch64-protos.h
3165 (aarch64_simd_immediate_valid_for_move): Remove.
3166 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3167 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3168 cases.
3169 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3170
76500022
SE
31712013-07-11 Steve Ellcey <sellcey@mips.com>
3172
3173 * config/mips/mips.c (mips_conditional_register_usage): Do not
3174 use t[0-7] registers in MIPS16 mode when optimizing for size.
3175
78039734
ST
31762013-07-11 Sriraman Tallam <tmsriram@google.com>
3177
3178 * config/i386/i386.c (dispatch_function_versions): Fix array
3179 indexing of function_version_info to match actual_versions.
3180
7b24b675
TJ
31812013-07-11 Teresa Johnson <tejohnson@google.com>
3182
3183 * vec.h (struct va_gc): Move release out-of-line.
3184 (va_gc::release): Call ggc_free on released vec.
3185
85d4cbb8
UW
31862013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3187
3188 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3189 Require GOT register as additional operand in UNSPEC.
3190 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3191 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3192 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3193 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3194 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3195 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3196 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3197
52b3f9e2
GJL
31982013-07-11 Georg-Johann Lay <avr@gjlay.de>
3199
3200 PR target/57631
3201 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3202 name seen by assembler/linker rather if available.
3203
3f3e266e
AS
32042013-07-11 Andreas Schwab <schwab@suse.de>
3205
3206 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3207
21231ca6
VM
32082013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3209
371e77e3 3210 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 3211
d331d374
JM
32122013-07-10 Joseph Myers <joseph@codesourcery.com>
3213
673c2f63
JM
3214 * doc/tm.texi.in: Move hook documentation to ....
3215 * target.def: ... here.
3216
d331d374
JM
3217 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3218 text on @hook line.
3219 * doc/tm.texi: Regenerate.
3220
dfeadaa0
PC
32212013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3222
3223 PR c++/57869
3224 * doc/invoke.texi: Document -Wconditionally-supported.
3225
bb6f2bac
GJL
32262013-07-10 Georg-Johann Lay <avr@gjlay.de>
3227
3228 PR target/57844
3229 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3230 of my_fp.
3231
5219b232
GJL
32322013-07-10 Georg-Johann Lay <avr@gjlay.de>
3233
3234 PR target/57506
3235 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3236 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3237 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3238 Remove duplicate devices.
3239 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3240 * config/avr/t-multilib: Regenerate.
3241 * config/avr/avr-tables.opt: Regenerate.
3242 * doc/avr-mmcu.texi: Regenerate.
3243
283045c0
GJL
32442013-07-10 Georg-Johann Lay <avr@gjlay.de>
3245
3246 PR target/56987
3247 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3248
ceb2d59f 32492013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 3250
ceb2d59f
GS
3251 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3252 the cost of MULT when optimizing for size.
3253
055e0a99
JBG
32542013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3255
3256 * config/cr16/cr16-protos.h: Don't include target.h.
3257
ef43e868
JM
32582013-07-09 Joseph Myers <joseph@codesourcery.com>
3259
3260 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3261 adjust register size for TDmode and TFmode for VSX registers.
3262
4a283fdf
KT
32632013-07-08 Kai Tietz <ktietz@redhat.com>
3264
3265 PR target/56892
3266 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3267 hook_bool_const_tree_true.
3268
2cf4c39e
AK
32692013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3270
3271 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3272 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3273 * config/s390/s390.md: Define FPR*_REGNUM constants.
3274 Fix FPR2_REGNUM constant (18 -> 17).
3275 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3276 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3277 ("trunc<BFP:mode><DFP_ALL:mode>2")
3278 ("trunc<DFP_ALL:mode><BFP:mode>2")
3279 ("*extend<BFP:mode><DFP_ALL:mode>2")
3280 ("*extend<DFP_ALL:mode><BFP:mode>2")
3281 ("extend<BFP:mode><DFP_ALL:mode>2")
3282 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3283 FPR4_REGNUM.
3284
e6ac0270 32852013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 3286
6a2fa4b2 3287 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 3288
b89b22fc
AK
32892013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3290
3291 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3292 and cfun_fpr_bit_p to cfun_fpr_save_p.
3293 (s390_frame_area, s390_register_info, s390_frame_info)
3294 (s390_emit_prologue, s390_emit_epilogue)
3295 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3296 register numbers.
3297 * config/s390/s390.h: Define *_REGNUM macros for floating point
3298 register numbers.
3299
07f398aa
EB
33002013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3301
3302 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3303
75fefa91
PCC
33042013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3305
3306 PR rtl-optimization/57786
3307 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3308 and break out of the loop when it is set to false.
3309
046f1eee
JJ
33102013-07-08 Jakub Jelinek <jakub@redhat.com>
3311
8140c065
JJ
3312 PR target/57819
3313 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3314 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3315 (const_int 63)) 0)).
3316 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3317 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3318 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3319
046f1eee
JJ
3320 PR rtl-optimization/57829
3321 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3322 mask bits outside of mode are just sign-extension from mode to HWI.
3323
930b700b
MZ
33242013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3325
3326 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3327 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3328 adjust_address instead of change_address to keep info about alignment.
3329 (emit_strmov): Remove.
3330 (emit_memmov): New function.
3331 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3332 (expand_movmem_epilogue): Likewise and return updated rtx for
3333 destination.
3334 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3335 destination and source.
3336 (decide_alignment): Refactor, handle vector_loop.
3337 (ix86_expand_movmem): Likewise.
3338 (ix86_expand_setmem): Likewise.
3339 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3340
7b9d1bd8
UB
33412013-07-07 Uros Bizjak <ubizjak@gmail.com>
3342
3343 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3344 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
3345
eabb5f48
UB
33462013-07-06 Uros Bizjak <ubizjak@gmail.com>
3347
3348 * config/i386/sse.md (sse_movlhps): Change alternative 3
3349 of operand 2 to "m".
3350
33512013-07-06 Uros Bizjak <ubizjak@gmail.com>
3352
3353 PR target/57807
3354 * config/i386/sse.md (iptr): New mode attribute.
3355 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
3356 (<sse>_vm<plusminus_insn><mode>3): Ditto.
3357 (<sse>_vmmul<mode>3): Ditto.
3358 (<sse>_vmdiv<mode>3): Ditto.
3359 (sse_vmrcpv4sf2): Ditto.
3360 (<sse>_vmsqrt<mode>2): Ditto.
3361 (sse_vmrsqrtv4sf2): Ditto.
3362 (<sse>_vm<code><mode>3): Ditto.
3363 (avx_vmcmp<mode>3): Ditto.
3364 (<sse>_vmmaskcmp<mode>3): Ditto.
3365 (<sse>_comi): Ditto.
3366 (<sse>_ucomi): Ditto.
3367 (*xop_vmfrcz_<mode>): Ditto.
3368 (*fmai_fmadd_<mode>): Ditto.
3369 (*fmai_fmsub_<mode>): Ditto.
3370 (*fmai_fnmadd_<mode>): Ditto.
3371 (*fmai_fnmsub_<mode>): Ditto.
3372 (*fma4i_vmfmadd_<mode>): Ditto.
3373 (*fma4i_vmfmsub_<mode>): Ditto.
3374 (*fma4i_vmfnmadd_<mode>): Ditto.
3375 (*fma4i_vmfnmsub_<mode>): Ditto.
3376 (*xop_vmfrcz_<mode>): Ditto.
3377 (sse_cvtps2pi): Ditto.
3378 (sse_cvttps2pi): Ditto.
3379 (sse_cvtss2si): Ditto.
3380 (sse_cvtss2si_2): Ditto.
3381 (sse_cvtss2siq_2): Ditto.
3382 (sse_cvttss2si): Ditto.
3383 (sse_cvttss2siq): Ditto.
3384 (sse_cvtsd2si): Ditto.
3385 (sse_cvtsd2si_2): Ditto.
3386 (sse_cvtsd2siq_2): Ditto.
3387 (sse_cvttsd2si): Ditto.
3388 (sse_cvttsd2siq): Ditto.
3389 (sse_cvtsd2ss): Ditto.
3390 (sse_cvtss2sd): Ditto.
3391 (avx2_pbroadcast<mode>): Ditto.
3392 (avx2_pbroadcast<mode>_1): Ditto.
3393 (*avx_vperm_broadcast_v4sf): Ditto.
3394
3395 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
3396 (sse_movlhps): Ditto.
3397 (sse_storehps): Ditto.
3398 (sse_loadhps): Ditto.
3399 (sse_storelps): Ditto.
3400 (sse_loadlps): Ditto.
3401 (*vec_concatv4sf): Ditto.
3402 (*vec_interleave_highv2df): Ditto.
3403 (*vec_interleave_lowv2df): Ditto.
3404 (*vec_extractv2df_1_sse): Ditto.
3405 (*vec_extractv2df_0_sse): Ditto.
3406 (sse2_storelpd): Ditto.
3407 (sse2_loadlpd): Ditto.
3408 (sse2_movsd): Ditto.
3409 (*vec_concatv4si): Ditto.
3410 (vec_concatv2di): Ditto.
3411
3412 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
3413 for Intel asm dialect.
3414 (mmx_punpcklwd): Ditto.
3415 (mmx_punpckldq): Ditto.
3416
3417 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
3418 for intel assembler dialect.
3419
1f6eac90
JJ
34202013-07-06 Jakub Jelinek <jakub@redhat.com>
3421
3422 PR target/29776
3423 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
3424 for BUILT_IN_C{LZ,LRSB}*.
3425 * tree.h (CASE_INT_FN): Add FN##IMAX case.
3426 * tree-vrp.c (extract_range_basic): Handle
3427 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
3428 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
3429 fall thru to code calling set_value*.
3430 * builtins.c (expand_builtin): Remove *IMAX cases.
3431 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
3432 if width is bigger than 2*HWI.
3433
2b778c9d
VM
34342013-07-05 Vladimir Makarov <vmakarov@redhat.com>
3435
3436 PR rtl-optimization/55342
3437 * lra-int.h (lra_subreg_reload_pseudos): New.
3438 * lra.c: Add undoing optional reloads to the block diagram.
3439 (lra_subreg_reload_pseudos): New.
3440 (lra_optional_reload_pseudos): Change comments.
3441 (lra): Init and clear lra_subreg_reload_pseudos. Clear
3442 lra_optional_reload_pseudos after undo transformations.
3443 * lra-assigns.c (pseudo_prefix_title): New.
3444 (lra_setup_reg_renumber): Use it.
3445 (spill_for): Ditto. Check subreg reload pseudos too.
3446 (assign_by_spills): Consider subreg reload pseudos too.
3447 * lra-constraints.c (simplify_operand_subreg): Use
3448 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
3449 (curr_insn_transform): Recognize and do optional reloads.
3450 (undo_optional_reloads): New.
3451 (lra_undo_inheritance): Call undo_optional_reloads.
3452
207156e4
TQ
34532013-07-05 Thomas Quinot <quinot@adacore.com>
3454
3455 * tree-complex.c (expand_complex_operations_1): Fix typo.
3456
8990e73a
TB
34572013-07-04 Tejas Belagod <tejas.belagod@arm.com>
3458
3459 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
3460 (tune_params): New member 'const vec_costs'.
3461 * config/aarch64/aarch64.c (generic_vector_cost): New.
3462 (generic_tunings): New member 'generic_vector_cost'.
3463 (aarch64_builtin_vectorization_cost): New.
3464 (aarch64_add_stmt_cost): New.
3465 (TARGET_VECTORIZE_ADD_STMT_COST): New.
3466 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
3467
f7bc421d
JJ
34682013-07-03 Jakub Jelinek <jakub@redhat.com>
3469
3470 PR target/57777
3471 * config/i386/predicates.md (vsib_address_operand): Disallow
3472 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
3473
e90d1568
HPN
34742013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
3475
3476 PR middle-end/55030
3477 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
3478 expand_builtin_setjmp_receiver.
3479 (expand_label): Adjust, call expand_builtin_setjmp_receiver
3480 with NULL for the label parameter.
3481 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
3482 the frame-pointer. Adjust comments.
3483 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
3484 only if LABEL is non-NULL.
3485
cbd7413b
YZ
34862013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
3487
3488 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
3489 (ARM_ABI_AAPCS64): Ditto.
3490 (arm_abi): Ditto.
3491 (ARM_DEFAULT_ABI): Ditto.
3492
dec11868
JG
34932013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
3494
3495 * config/aarch64/aarch64-builtins.c
3496 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
3497 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
3498 (st1): Likewise.
3499 * config/aarch64/aarch64-simd.md
3500 (aarch64_ld1<VALL:mode>): New.
3501 (aarch64_st1<VALL:mode>): Likewise.
3502 * config/aarch64/arm_neon.h
3503 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
3504
786fb9b6
ST
35052013-07-02 Sriraman Tallam <tmsriram@google.com>
3506
0ba6aff3
UB
3507 * config/i386/i386.c (gate_insert_vzeroupper): Check if
3508 target ISA is AVX.
786fb9b6
ST
3509 (ix86_option_override_internal):Turn on all -mavx target flags by
3510 default as they are dependent on AVX anyway.
3511
4b043553
CC
35122013-07-02 Cary Coutant <ccoutant@google.com>
3513
3514 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
3515 deterministic hash.
3516 (loc_checksum_ordered): Likewise.
3517 (hash_loc_operands): Remove inline keyword.
3518
c4e87a13
JJ
35192013-07-02 Jakub Jelinek <jakub@redhat.com>
3520
3521 PR tree-optimization/57741
3522 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
3523 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
3524 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
3525 Allow REAL_CST step_exprs if flag_associative_math.
3526 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
3527
26366d28
IB
35282013-07-02 Ian Bolton <ian.bolton@arm.com>
3529
0ba6aff3 3530 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 3531
2879bb2b
IB
35322013-07-02 Ian Bolton <ian.bolton@arm.com>
3533
3534 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
3535
afc5e8a6
KT
35362013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3537
3538 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
3539 encoding.
3540 (iorsi3_insn): Likewise.
3541 (arm_xorsi3): Likewise.
3542
31f8442b
SN
35432013-07-01 Sofiane Naci <sofiane.naci@arm.com>
3544
3545 * arm.md (attribute "wtype"): Delete. Move attribute values from here
3546 to ...
3547 (attribute "type"): ... here, and prefix with "wmmx_".
3548 (attribute "core_cycles"): Update for attribute changes.
3549 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
3550 (tbcstv4hi): Likewise.
3551 (tbcstv2si): Likewise.
3552 (iwmmxt_iordi3): Likewise.
3553 (iwmmxt_xordi3): Likewise.
3554 (iwmmxt_anddi3): Likewise.
3555 (iwmmxt_nanddi3): Likewise.
3556 (iwmmxt_arm_movdi): Likewise.
3557 (iwmmxt_movsi_insn): Likewise.
3558 (mov<mode>_internal): Likewise.
3559 (and<mode>3_iwmmxt): Likewise.
3560 (ior<mode>3_iwmmxt): Likewise.
3561 (xor<mode>3_iwmmxt): Likewise.
3562 (add<mode>3_iwmmxt): Likewise.
3563 (ssaddv8qi3): Likewise.
3564 (ssaddv4hi3): Likewise.
3565 (ssaddv2si3): Likewise.
3566 (usaddv8qi3): Likewise.
3567 (usaddv4hi3): Likewise.
3568 (usaddv2si3): Likewise.
3569 (sub<mode>3_iwmmxt): Likewise.
3570 (sssubv8qi3): Likewise.
3571 (sssubv4hi3): Likewise.
3572 (sssubv2si3): Likewise.
3573 (ussubv8qi3): Likewise.
3574 (ussubv4hi3): Likewise.
3575 (ussubv2si3): Likewise.
3576 (mulv4hi3_iwmmxt): Likewise.
3577 (smulv4hi3_highpart): Likewise.
3578 (umulv4hi3_highpart): Likewise.
3579 (iwmmxt_wmacs): Likewise.
3580 (iwmmxt_wmacsz): Likewise.
3581 (iwmmxt_wmacu): Likewise.
3582 (iwmmxt_wmacuz): Likewise.
3583 (iwmmxt_clrdi): Likewise.
3584 (iwmmxt_clrv8qi): Likewise.
3585 (iwmmxt_clr4hi): Likewise.
3586 (iwmmxt_clr2si): Likewise.
3587 (iwmmxt_uavgrndv8qi3): Likewise.
3588 (iwmmxt_uavgrndv4hi3): Likewise.
3589 (iwmmxt_uavgv8qi3): Likewise.
3590 (iwmmxt_uavgv4hi3): Likewise.
3591 (iwmmxt_tinsrb): Likewise.
3592 (iwmmxt_tinsrh): Likewise.
3593 (iwmmxt_tinsrw): Likewise.
3594 (iwmmxt_textrmub): Likewise.
3595 (iwmmxt_textrmsb): Likewise.
3596 (iwmmxt_textrmuh): Likewise.
3597 (iwmmxt_textrmsh): Likewise.
3598 (iwmmxt_textrmw): Likewise.
3599 (iwmxxt_wshufh): Likewise.
3600 (eqv8qi3): Likewise.
3601 (eqv4hi3): Likewise.
3602 (eqv2si3): Likewise.
3603 (gtuv8qi3): Likewise.
3604 (gtuv4hi3): Likewise.
3605 (gtuv2si3): Likewise.
3606 (gtv8qi3): Likewise.
3607 (gtv4hi3): Likewise.
3608 (gtv2si3): Likewise.
3609 (smax<mode>3_iwmmxt): Likewise.
3610 (umax<mode>3_iwmmxt): Likewise.
3611 (smin<mode>3_iwmmxt): Likewise.
3612 (umin<mode>3_iwmmxt): Likewise.
3613 (iwmmxt_wpackhss): Likewise.
3614 (iwmmxt_wpackwss): Likewise.
3615 (iwmmxt_wpackdss): Likewise.
3616 (iwmmxt_wpackhus): Likewise.
3617 (iwmmxt_wpackwus): Likewise.
3618 (iwmmxt_wpackdus): Likewise.
3619 (iwmmxt_wunpckihb): Likewise.
3620 (iwmmxt_wunpckihh): Likewise.
3621 (iwmmxt_wunpckihw): Likewise.
3622 (iwmmxt_wunpckilb): Likewise.
3623 (iwmmxt_wunpckilh): Likewise.
3624 (iwmmxt_wunpckilw): Likewise.
3625 (iwmmxt_wunpckehub): Likewise.
3626 (iwmmxt_wunpckehuh): Likewise.
3627 (iwmmxt_wunpckehuw): Likewise.
3628 (iwmmxt_wunpckehsb): Likewise.
3629 (iwmmxt_wunpckehsh): Likewise.
3630 (iwmmxt_wunpckehsw): Likewise.
3631 (iwmmxt_wunpckelub): Likewise.
3632 (iwmmxt_wunpckeluh): Likewise.
3633 (iwmmxt_wunpckeluw): Likewise.
3634 (iwmmxt_wunpckelsb): Likewise.
3635 (iwmmxt_wunpckelsh): Likewise.
3636 (iwmmxt_wunpckelsw): Likewise.
3637 (ror<mode>3): Likewise.
3638 (ashr<mode>3_iwmmxt): Likewise.
3639 (lshr<mode>3_iwmmxt): Likewise.
3640 (ashl<mode>3_iwmmxt): Likewise.
3641 (ror<mode>3_di): Likewise.
3642 (ashr<mode>3_di): Likewise.
3643 (lshr<mode>3_di): Likewise.
3644 (ashl<mode>3_di): Likewise.
3645 (iwmmxt_wmadds): Likewise.
3646 (iwmmxt_wmaddu): Likewise.
3647 (iwmmxt_tmia): Likewise.
3648 (iwmmxt_tmiaph): Likewise.
3649 (iwmmxt_tmiabb): Likewise.
3650 (iwmmxt_tmiatb): Likewise.
3651 (iwmmxt_tmiabt): Likewise.
3652 (iwmmxt_tmiatt): Likewise.
3653 (iwmmxt_tmovmskb): Likewise.
3654 (iwmmxt_tmovmskh): Likewise.
3655 (iwmmxt_tmovmskw): Likewise.
3656 (iwmmxt_waccb): Likewise.
3657 (iwmmxt_wacch): Likewise.
3658 (iwmmxt_waccw): Likewise.
3659 (iwmmxt_waligni): Likewise.
3660 (iwmmxt_walignr): Likewise.
3661 (iwmmxt_walignr0): Likewise.
3662 (iwmmxt_walignr1): Likewise.
3663 (iwmmxt_walignr2): Likewise.
3664 (iwmmxt_walignr3): Likewise.
3665 (iwmmxt_wsadb): Likewise.
3666 (iwmmxt_wsadh): Likewise.
3667 (iwmmxt_wsadbz): Likewise.
3668 (iwmmxt_wsadhz): Likewise.
3669 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
3670 (iwmmxt_wabsdiffb): Likewise.
3671 (iwmmxt_wabsdiffh): Likewise.
3672 (iwmmxt_wabsdiffw): Likewise.
3673 (iwmmxt_waddsubhx): Likewise
3674 (iwmmxt_wsubaddhx): Likewise.
3675 (addc<mode>3): Likewise.
3676 (iwmmxt_avg4): Likewise.
3677 (iwmmxt_avg4r): Likewise.
3678 (iwmmxt_wmaddsx): Likewise.
3679 (iwmmxt_wmaddux): Likewise.
3680 (iwmmxt_wmaddsn): Likewise.
3681 (iwmmxt_wmaddun): Likewise.
3682 (iwmmxt_wmulwsm): Likewise.
3683 (iwmmxt_wmulwum): Likewise.
3684 (iwmmxt_wmulsmr): Likewise.
3685 (iwmmxt_wmulumr): Likewise.
3686 (iwmmxt_wmulwsmr): Likewise.
3687 (iwmmxt_wmulwumr): Likewise.
3688 (iwmmxt_wmulwl): Likewise.
3689 (iwmmxt_wqmulm): Likewise.
3690 (iwmmxt_wqmulwm): Likewise.
3691 (iwmmxt_wqmulmr): Likewise.
3692 (iwmmxt_wqmulwmr): Likewise.
3693 (iwmmxt_waddbhusm): Likewise.
3694 (iwmmxt_waddbhusl): Likewise.
3695 (iwmmxt_wqmiabb): Likewise.
3696 (iwmmxt_wqmiabt): Likewise.
3697 (iwmmxt_wqmiatb): Likewise.
3698 (iwmmxt_wqmiatt): Likewise.
3699 (iwmmxt_wqmiabbn): Likewise.
3700 (iwmmxt_wqmiabtn): Likewise.
3701 (iwmmxt_wqmiatbn): Likewise.
3702 (iwmmxt_wqmiattn): Likewise.
3703 (iwmmxt_wmiabb): Likewise.
3704 (iwmmxt_wmiabt): Likewise.
3705 (iwmmxt_wmiatb): Likewise.
3706 (iwmmxt_wmiatt): Likewise.
3707 (iwmmxt_wmiabbn): Likewise.
3708 (iwmmxt_wmiabtn): Likewise.
3709 (iwmmxt_wmiatbn): Likewise.
3710 (iwmmxt_wmiattn): Likewise.
3711 (iwmmxt_wmiawbb): Likewise.
3712 (iwmmxt_wmiawbt): Likewise.
3713 (iwmmxt_wmiawtb): Likewise.
3714 (iwmmxt_wmiawtt): Likewise.
3715 (iwmmxt_wmiawbbn): Likewise.
3716 (iwmmxt_wmiawbtn): Likewise.
3717 (iwmmxt_wmiawtbn): Likewise.
3718 (iwmmxt_wmiawttn): Likewise.
3719 (iwmmxt_wmerge): Likewise.
3720 (iwmmxt_tandc<mode>3): Likewise.
3721 (iwmmxt_torc<mode>3): Likewise.
3722 (iwmmxt_torvsc<mode>3): Likewise.
3723 (iwmmxt_textrc<mode>3): Likewise.
3724 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
3725 (wmmxt_pack): Likewise.
3726 (wmmxt_mult_c1): Likewise.
3727 (wmmxt_mult_c2): Likewise.
3728 (wmmxt_alu_c1): Likewise.
3729 (wmmxt_alu_c2): Likewise.
3730 (wmmxt_alu_c3): Likewise.
3731 (wmmxt_transfer_c1): Likewise.
3732 (wmmxt_transfer_c2): Likewise.
3733 (wmmxt_transfer_c3): Likewise.
3734 (marvell_f_iwmmxt_wstr): Likewise.
3735 (marvell_f_iwmmxt_wldr): Likewise.
3736
bdb7bf8a
YZ
37372013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
3738
0ba6aff3 3739 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 3740
e6631d38
VM
37412013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3742
3743 Revert:
3744 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3745 * lra-constraints.c (need_for_split_p): Check call used hard regs
3746 living through calls.
3747
3748 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
3749 call used regs for call insn.
3750
53c1275b
JJ
37512013-06-28 Jakub Jelinek <jakub@redhat.com>
3752
3753 PR target/57736
0ba6aff3
UB
3754 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
3755 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 3756
1141ed3f
BI
37572013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
3758
3759 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 3760
e32e4c4a
VM
37612013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3762
3763 * lra-constraints.c (need_for_split_p): Check call used hard regs
3764 living through calls.
3765
f161bfd3
MM
37662013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
3767
3768 PR target/57744
3769 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
3770 to tie with any other modes. Eliminate Altivec vector mode tests,
3771 since these are a subset of ALTIVEC or VSX vector modes. Simplify
3772 code, to return 0 if testing MODE2 for a condition, if we've
3773 already tested MODE1 for the same condition.
3774
7eda14e1
MS
37752013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3776
3777 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
3778 layout.
3779
c822f852
MS
37802013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3781
3782 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3783 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
3784
da4f13a4
MS
37852013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3786
0ba6aff3
UB
3787 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
3788 Define.
da4f13a4
MS
3789 (aarch64_symbolic_constant_p): Remove.
3790 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
3791 static. Fix line length and white space.
3792 (aarch64_symbolic_constant_p): Remove.
3793 * config/aarch64/predicates.md (aarch64_valid_symref):
3794 Use aarch64_classify_symbol_expression.
3795
956a95a5
KT
37962013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3797
3798 * config/arm/constraints.md (Ts): New constraint.
3799 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
3800 16-bit encodings.
3801 (compare_scc): Use "Ts" constraint for operand 0.
3802 (ior_scc_scc): Likewise.
3803 (and_scc_scc): Likewise.
3804 (and_scc_scc_nodom): Likewise.
3805 (ior_scc_scc_cmp): Likewise for operand 7.
3806 (and_scc_scc_cmp): Likewise.
3807 * config/arm/thumb2.md (thumb2_movsi_insn):
3808 Add alternatives for 16-bit encodings.
3809 (thumb2_movhi_insn): Likewise.
3810 (thumb2_movsicc_insn): Likewise.
3811 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
3812 (thumb2_negscc): Use "Ts" constraint.
3813 Move mvn instruction outside cond_exec block.
3814 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
3815 for 16-bit encodings.
3816
31dfce10
KT
38172013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3818
3819 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
3820 encoding.
3821 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
3822 (mulsi3subsi): Likewise.
3823 (mulsidi3adddi): Likewise.
3824 (mulsidi3_v6): Likewise.
3825 (umulsidi3_v6): Likewise.
3826 (umulsidi3adddi_v6): Likewise.
3827 (smulsi3_highpart_v6): Likewise.
3828 (umulsi3_highpart_v6): Likewise.
3829 (mulhisi3tb): Likewise.
3830 (mulhisi3bt): Likewise.
3831 (mulhisi3tt): Likewise.
3832 (maddhisi4): Likewise.
3833 (maddhisi4tb): Likewise.
3834 (maddhisi4tt): Likewise.
3835 (maddhidi4): Likewise.
3836 (maddhidi4tb): Likewise.
3837 (maddhidi4tt): Likewise.
3838 (zeroextractsi_compare0_scratch): Likewise.
3839 (insv_zero): Likewise.
3840 (insv_t2): Likewise.
3841 (anddi_notzesidi_di): Likewise.
3842 (anddi_notsesidi_di): Likewise.
3843 (andsi_notsi_si): Likewise.
3844 (iordi_zesidi_di): Likewise.
3845 (xordi_zesidi_di): Likewise.
3846 (andsi_iorsi3_notsi): Likewise.
3847 (smax_0): Likewise.
3848 (smax_m1): Likewise.
3849 (smin_0): Likewise.
3850 (not_shiftsi): Likewise.
3851 (unaligned_loadsi): Likewise.
3852 (unaligned_loadhis): Likewise.
3853 (unaligned_loadhiu): Likewise.
3854 (unaligned_storesi): Likewise.
3855 (unaligned_storehi): Likewise.
3856 (extv_reg): Likewise.
3857 (extzv_t2): Likewise.
3858 (divsi3): Likewise.
3859 (udivsi3): Likewise.
3860 (arm_zero_extendhisi2addsi): Likewise.
3861 (arm_zero_extendqisi2addsi): Likewise.
3862 (compareqi_eq0): Likewise.
3863 (arm_extendhisi2_v6): Likewise.
3864 (arm_extendqisi2addsi): Likewise.
3865 (arm_movt): Likewise.
3866 (thumb2_ldrd): Likewise.
3867 (thumb2_ldrd_base): Likewise.
3868 (thumb2_ldrd_base_neg): Likewise.
3869 (thumb2_strd): Likewise.
3870 (thumb2_strd_base): Likewise.
3871 (thumb2_strd_base_neg): Likewise.
3872 (arm_negsi2): Add alternative for 16-bit encoding.
3873 (arm_one_cmplsi2): Likewise.
3874
1572e697
KT
38752013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3876
3877 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
3878 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
3879 (movdfcc): Likewise.
3880 * config/arm/vfp.md (*thumb2_movsf_vfp):
3881 Disable predication for arm_restrict_it.
3882 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
3883 (*thumb2_movdfcc_vfp): Likewise.
3884 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
3885 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
3886 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
3887 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
3888 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
3889 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
3890 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
3891 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
3892 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
3893 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
3894 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
3895 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
3896 Disable predication for arm_restrict_it.
3897
ebf8f0ea
KY
38982013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
3899
3900 * config/i386/bmiintrin.h (_bextr_u32): New.
3901 (_bextr_u64): Ditto.
3902
a290fcda
RS
39032013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
3904
3905 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
3906 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
3907 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
3908 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
3909 * config/mips/n32-elf.h: ...this new file.
3910
ea169996
MG
39112013-06-27 Marc Glisse <marc.glisse@inria.fr>
3912
3913 PR target/57224
3914 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
3915 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
3916
8e932114
CM
39172013-06-27 Catherine Moore <clm@codesourcery.com>
3918
3919 * config/mips/mips-tables.opt: Regenerate.
3920 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 3921 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 3922 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 3923
2444244e
JJ
39242013-06-27 Jakub Jelinek <jakub@redhat.com>
3925
cd06d2a2
JJ
3926 PR target/57623
3927 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
3928 constraints of operand 1 and 2.
3929
2444244e
JJ
3930 PR target/57623
3931 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
3932 to match RTL canonicalization. Swap predicates and
3933 constraints of operand 1 and 2.
3934
2f259720
VM
39352013-06-27 Vladimir Makarov <vmakarov@redhat.com>
3936
0ba6aff3
UB
3937 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
3938 Process OP_INOUT regs for splitting too.
2f259720 3939
fef4d2b3
JJ
39402013-06-27 Jakub Jelinek <jakub@redhat.com>
3941
3942 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
3943 decl before the loop, initialize to NULL.
3944 (vectorizable_load): Initialize ptr_incr to NULL.
3945
79ee9826
MJ
39462013-06-27 Martin Jambor <mjambor@suse.cz>
3947
3948 PR lto/57208
3949 * ipa-ref.h (ipa_maybe_record_reference): Declare.
3950 * ipa-ref.c (ipa_maybe_record_reference): New function.
3951 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
3952 * ipa-cp.c (create_specialized_node): Record potential references from
3953 aggvals.
3954 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
3955
e18b4a81
YZ
39562013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
3957
3958 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
3959 parameter 'mode' of type 'enum machine_mode mode'; change to pass
3960 'mode' to force_reg.
3961 (aarch64_add_offset): Update calls to aarch64_force_temporary.
3962 (aarch64_expand_mov_immediate): Likewise.
3963
9c023bf0
YZ
39642013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
3965
3966 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
3967 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
3968
5a3fe9b6
AK
39692013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3970
3971 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
3972 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
3973 (struct machine_function): Add tbegin_p.
3974 (s390_canonicalize_comparison): Fold CC mode compares to
3975 conditional jump if possible.
3976 (s390_emit_jump): Return the emitted jump.
3977 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
3978 Handle CCRAWmode compares.
3979 (s390_option_override): Default to -mhtm if available.
3980 (s390_reg_clobbered_rtx): Handle floating point regs as well.
3981 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
3982 FPRs instead of df_regs_ever_live_p.
3983 (s390_optimize_nonescaping_tx): New function.
3984 (s390_init_frame_layout): Extend clobbered_regs array to cover
3985 FPRs as well.
3986 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
3987 (s390_expand_tbegin): New function.
3988 (enum s390_builtin): New enum definition.
3989 (code_for_builtin): New array definition.
3990 (s390_init_builtins): New function.
3991 (s390_expand_builtin): New function.
3992 (TARGET_INIT_BUILTINS): Define.
3993 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 3994 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK
3995 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
3996 (s390_alc_comparison): Likewise.
3997 * config/s390/s390-modes.def: Add CCRAWmode.
3998 * config/s390/s390.h (processor_flags): Add PF_TX.
3999 (TARGET_CPU_HTM): Define macro.
4000 (TARGET_HTM): Define macro.
4001 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4002 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4003 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4004 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 4005 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK
4006 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4007 ("*cc_to_int"): Move up.
4008 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4009 constants other than 0.
4010 ("*ccraw_to_int"): New insn and splitter definition.
4011 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4012 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4013 ("tx_assist"): New expander.
4014 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4015 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4016 * config/s390/s390.opt: Add -mhtm option.
4017 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4018 * config/s390/htmxlintrin.h: New file.
4019 * config/s390/htmintrin.h: New file.
4020 * config/s390/s390intrin.h: New file.
4021 * doc/extend.texi: Document htm builtins.
4022 * config.gcc: Add the new header files to extra_headers.
4023
9e216629
TS
40242013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4025
4026 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 4027 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 4028
fd8c65e7
MM
40292013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4030 Pat Haugen <pthaugen@us.ibm.com>
4031 Peter Bergner <bergner@vnet.ibm.com>
4032
4033 * config/rs6000/power8.md: New.
4034 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4035 setting for power8 entry.
4036 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4037 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4038 test for Power4/Power5 only.
4039 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4040 support.
4041 (force_new_group): Adjust comment.
4042 * config/rs6000/rs6000.md: Include power8.md.
4043
b24a2ce5
GY
40442013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4045
4046 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4047 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4048 declaration.
4049 (tune_params): Update comment.
4050 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4051 (arm_max_conditional_execute): New function.
4052 (thumb2_final_prescan_insn): Use max_insn_skipped and
4053 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4054
f6fdeb62
JJ
40552013-06-25 Jakub Jelinek <jakub@redhat.com>
4056
4057 PR tree-optimization/57705
4058 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4059 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 4060 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ
4061 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4062
c13bc3d9
MJ
40632013-06-25 Martin Jambor <mjambor@suse.cz>
4064
4065 PR middle-end/57670
4066 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4067 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4068 calls in the dump.
4069 (ipa_note_param_call): Initialize member_ptr flag.
4070 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4071 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4072 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4073 (ipa_read_indirect_edge_info): Likewise.
4074
8a845901
RB
40752013-06-25 Richard Biener <rguenther@suse.de>
4076
4077 PR middle-end/56977
4078 * passes.c (init_optimization_passes): Move pass_fold_builtins
4079 and pass_dce earlier with -Og.
4080
33be0bec
EB
40812013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4082
4083 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4084 <BIT_FIELD_REF>: Remove trailing TAB.
4085 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4086 remove blank line.
4087
5fe8e757
MJ
40882013-06-24 Martin Jambor <mjambor@suse.cz>
4089
4090 PR tree-optimization/57358
4091 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4092 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4093 (ipa_analyze_params_uses): Generate pessimistic info when true.
4094
44a60244
MJ
40952013-06-24 Martin Jambor <mjambor@suse.cz>
4096
4097 PR tree-optimization/57539
4098 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4099 global.inlined_to of the new node to it. All callers changed.
4100 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4101 inlining_into, pass it to cgraph_clone_node.
4102 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4103 ipa_free_edge_args_substructures.
4104 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4105 rdesc linked list. Do not assert rdesc edges have inlined caller.
4106 Assert we have found an rdesc in the rdesc list.
4107
7c5848b8
RB
41082013-06-24 Richard Biener <rguenther@suse.de>
4109
0ba6aff3 4110 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB
4111 (pointer_set_lookup): Declare.
4112 (class pointer_map): New template class implementing a
4113 generic pointer to T map.
4114 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4115 pointer_map<T>::contains, pointer_map<T>::insert,
4116 pointer_map<T>::traverse): New functions.
4117 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4118 (pointer_set_lookup): New function.
4119 (pointer_set_contains): Use pointer_set_lookup.
4120 (pointer_set_insert): Likewise.
4121 (insert_aux): Remove.
4122 (struct pointer_map_t): Embed a pointer_set_t.
4123 (pointer_map_create): Adjust.
4124 (pointer_map_destroy): Likewise.
4125 (pointer_map_contains): Likewise.
4126 (pointer_map_insert): Likewise.
4127 (pointer_map_traverse): Likewise.
4128 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4129 pointer_map<unsigned> instead of a pointer_map_t.
4130 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4131 (streamer_tree_cache_lookup): Likewise.
4132 (streamer_tree_cache_create): Likewise.
4133 (streamer_tree_cache_delete): Likewise.
4134 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4135 pointer_map<unsigned> instead of a pointer_map_t.
4136 (lto_init_tree_ref_encoder): Adjust.
4137 (lto_destroy_tree_ref_encoder): Likewise.
4138 * lto-section-out.c (lto_output_decl_index): Likewise.
4139 (lto_record_function_out_decl_state): Likewise.
4140 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4141
3dbc97a9
RB
41422013-06-24 Richard Biener <rguenther@suse.de>
4143
4144 PR tree-optimization/57488
4145 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4146
8345965a
AM
41472013-06-24 Alan Modra <amodra@gmail.com>
4148
4149 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4150 (gen_easy_altivec_constant): Likewise.
4151 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4152 easy_vector_constant_msb): Likewise.
4153
b7df379f
JJ
41542013-06-23 Jakub Jelinek <jakub@redhat.com>
4155
4156 PR target/57688
4157 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4158 add missing return true.
4159
0bcf9a09
OE
41602013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4161
4162 PR target/52483
4163 * config/sh/predicates.md (general_extend_operand): Invoke
4164 general_movsrc_operand for memory operands.
4165 (general_movsrc_operand): Allow reg+reg addressing, do not use
4166 general_operand for memory operands.
4167
97db2bf7
ST
41682013-06-23 Sriraman Tallam <tmsriram@google.com>
4169
4170 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4171 when current target options does not apply.
4172 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4173 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB
4174 * config/i386/bmiintrin.h: Pass appropriate target
4175 attributes to header.
97db2bf7
ST
4176 * config/i386/mmintrin.h: Ditto.
4177 * config/i386/nmmintrin.h: Ditto.
4178 * config/i386/avx2intrin.h: Ditto.
4179 * config/i386/fxsrintrin.h: Ditto.
4180 * config/i386/tbmintrin.h: Ditto.
4181 * config/i386/xsaveintrin.h: Ditto.
4182 * config/i386/f16cintrin.h: Ditto.
4183 * config/i386/xtestintrin.h: Ditto.
4184 * config/i386/xsaveoptintrin.h: Ditto.
4185 * config/i386/bmi2intrin.h: Ditto.
4186 * config/i386/lzcntintrin.h: Ditto.
4187 * config/i386/smmintrin.h: Ditto.
4188 * config/i386/wmmintrin.h: Ditto.
4189 * config/i386/x86intrin.h: Remove all header include guards.
4190 * config/i386/prfchwintrin.h: Ditto.
4191 * config/i386/pmmintrin.h: Ditto.
4192 * config/i386/tmmintrin.h: Ditto.
4193 * config/i386/xmmintrin.h: Ditto.
4194 * config/i386/popcntintrin.h: Ditto.
4195 * config/i386/rdseedintrin.h: Ditto.
4196 * config/i386/ammintrin.h: Ditto.
4197 * config/i386/emmintrin.h: Ditto.
4198 * config/i386/immintrin.h: Remove all header include guards.
4199 * config/i386/fma4intrin.h: Ditto.
4200 * config/i386/lwpintrin.h: Ditto.
4201 * config/i386/xopintrin.h: Ditto.
4202 * config/i386/ia32intrin.h: Ditto.
4203 * config/i386/avxintrin.h: Ditto.
4204 * config/i386/rtmintrin.h: Ditto.
4205 * config/i386/fmaintrin.h: Ditto.
4206 * config/i386/mm3dnow.h: Ditto.
4207
495e6879
ST
42082013-06-22 Sriraman Tallam <tmsriram@google.com>
4209
4210 * common/config/i386/i386-common.c: Handle LZCNT.
4211
11e8f687
AK
42122013-06-22 Andi Kleen <ak@linux.intel.com>
4213
4214 * doc/extend.texi: Use __atomic_store_n instead of
4215 __atomic_store in HLE example.
4216
664dd721
OE
42172013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4218
4219 * config/sh/sh.c: Remove <cstdlib> workaround.
4220
927f908b
AK
42212013-06-21 Andi Kleen <ak@linux.intel.com>
4222
0ba6aff3 4223 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 4224
cca410b6
AK
42252013-06-21 Andi Kleen <ak@linux.intel.com>
4226
4227 * doc/extend.texi: Document that __atomic_clear and
4228 __atomic_test_and_set should only be used with bool.
4229
d84db1ed
JH
42302013-06-20 Jan Hubicka <jh@suse.cz>
4231
4232 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4233 types_same_for_odr.
4234 * tree.c (decls_same_for_odr): New function.
4235 (same_for_edr): New function.
4236 (types_same_for_odr): New function.
4237 (get_binfo_at_offset): Use it.
4238 * tree.h (types_same_for_odr): Declare.
4239
664dd721 42402013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE
4241 Jason Merrill <jason@redhat.com>
4242
4243 * system.h: Include <cstdlib> as well as <stdlib.h>.
4244
49b4ff63
UB
42452013-06-20 Uros Bizjak <ubizjak@gmail.com>
4246
4247 PR target/57655
4248 * config/i386/i386.c (construct_container): Report error if
4249 long double is used with disabled x87 float returns.
4250
7a89ad00
JH
42512013-06-20 Jan Hubicka <jh@suse.cz>
4252
4253 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4254
bd280792
JR
42552013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4256
4257 PR rtl-optimization/57425
4258 PR rtl-optimization/57569
4259 * alias.c (write_dependence_p): Remove parameters mem_mode and
4260 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4261 Changed all callers.
4262 (canon_anti_dependence): Get comments and semantics in sync.
4263 Add parameter mem_canonicalized. Changed all callers.
4264 * rtl.h (canon_anti_dependence): Update prototype.
4265
d16e9a99
RB
42662013-06-20 Richard Biener <rguenther@suse.de>
4267
4268 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4269 byte case, inline streamer_read_uchar and defer section
4270 overrun check.
4271
c3a9b91b
RB
42722013-06-20 Richard Biener <rguenther@suse.de>
4273
4274 PR tree-optimization/57584
4275 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4276 SSA names into the expanded expression that take part in
4277 abnormal coalescing.
4278
371e77e3 42792013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS
4280
4281 * gcov.c (print_usage): Handle new option.
4282 (process_args): Ditto.
4283 (get_gcov_intermediate_filename): New function.
4284 (output_intermediate_file): New function.
4285 (output_gcov_file): New function
4286 (generate_results): Handle new option.
4287 (release_function): Relase demangled name.
4288 (read_graph_file): Handle demangled name.
4289 (output_lines): Ditto.
4290 * doc/gcov.texi: Document gcov intermediate format.
4291
11f5b71a
VM
42922013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4293
4294 PR bootstrap/57604
4295 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4296 (lra_emit_add): Use the functions. Add comment about Y as an
4297 address segment.
4298
e1f9340b
DE
42992013-06-19 David Edelsohn <dje.gcc@gmail.com>
4300
4301 PR driver/57652
4302 * collect2.c (collect_atexit): New.
4303 (collect_exit): Delete.
4304 (main): Register collect_atexit with atexit.
4305 (collect_wait): Change collect_exit to exit.
4306 (do_wait): Same.
4307 * collect2.h (collect_exit): Delete.
4308 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4309
e3f9e0ac
WM
43102013-06-19 Wei Mi <wmi@google.com>
4311
4312 PR rtl-optimization/57518
4313 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4314 if regno is used in paradoxical subreg.
4315 (update_equiv_regs): Check pdx_subregs[regno] before
4316 set a reg to be equivalent with a mem.
4317
a09f1a76
MK
43182013-06-19 Matthias Klose <doko@ubuntu.com>
4319
4320 PR driver/57651
4321 * file-find.h (find_a_file): Add a mode parameter.
4322 * file-find.c (find_a_file): Likewise.
4323 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4324 with X_OK for the executables.
4325 * collect2.c (main): Call find_a_file with X_OK.
4326
7d18b0ad
SE
43272013-06-19 Steve Ellcey <sellcey@mips.com>
4328
4329 PR target/56942
4330 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4331 Use NEXT_INSN instead of next_real_insn.
4332
6a6dac52
JH
43332013-06-19 Jan Hubicka <jh@suse.cz>
4334
4335 * cgraph.h (const_value_known_p): Replace by ...
4336 (ctor_for_folding): .. this one.
4337 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4338 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4339 * expr.c (expand_expr_real_1): Likewise.
4340 (string_constant): Likewise.
4341 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4342 * ipa.c (process_references): Likewise.
4343 (symtab_remove_unreachable_nodes): Likewise.
4344 * ipa-inline-analysis.c (param_change_prob): Likewise.
4345 * gimple-fold.c (canonicalize_constructor_val): Likewise.
4346 (get_base_constructor): Likwise.
4347 * varpool.c (varpool_remove_node): Likewise.
4348 (varpool_remove_initializer): LIkewise.
4349 (dump_varpool_node): LIkwise.
4350 (const_value_known_p): Rewrite to ...
4351 (ctor_for_folding): ... this one.
4352
216c12ab
JJ
43532013-06-19 Jakub Jelinek <jakub@redhat.com>
4354
4355 PR driver/57651
4356 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
4357 PERSONALITY in $PATH derived prefixes.
4358
d53e2f99
JL
43592013-06-19 Jeff Law <law@redhat.com>
4360
ecdbf306
JL
4361 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
4362 in comment.
4363
d53e2f99
JL
4364 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
4365 (simplify_bitwise_binary): Use it to simpify certain binary ops on
4366 booleans.
4367
89e7fc04
SN
43682013-06-19 Sofiane Naci <sofiane.naci@arm.com>
4369
4370 * config/arm/vfp.md: Move VFP instruction classification documentation
4371 to ...
4372 * config/arm/arm.md: ... here. Update instruction classification
4373 documentation.
4374
5e7f6aaa
RE
43752013-06-19 Richard Earnshaw <rearnsha@arm.com>
4376
4377 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
4378 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
4379 pattern. Use more efficient sequences on ARMv5 and Thumb2.
4380
5fdc1e5d
SB
43812013-06-19 Steven Bosscher <steven@gcc.gnu.org>
4382
4383 PR target/57609
4384 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4385 with NEXT_INSN. Use tablejump_p to check for jump table data
4386 insns.
4387
e952afba
PC
43882013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
4389
4390 PR c++/56544
4391 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
4392 that now in C++ the value is correct per the C++ standards.
4393
3fd0303a
RB
43942013-06-19 Richard Biener <rguenther@suse.de>
4395
4396 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
4397 for global context.
4398
b04e4b08
AK
43992013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4400
4401 Revert:
4402 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4403
4404 PR target/57609
4405 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4406 with next_active_insn.
4407
bef8491a
ST
44082013-06-18 Sriraman Tallam <tmsriram@google.com>
4409
4410 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
4411 functions are inlined during failures to flag an error.
4412 * tree-inline.c (expand_call_inline): Allow the error to be flagged
4413 in early inline pass.
22fc7d3d 4414
c21b257d
L
44152013-06-18 H.J. Lu <hongjiu.lu@intel.com>
4416
4417 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
4418 in comments.
4419
33255ae3
JB
44202013-06-18 Julian Brown <julian@codesourcery.com>
4421
4422 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
4423 Permit virtual register pre-reload if !strict.
4424 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
4425 change.
4426 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
4427 prototype.
4428 * config/arm/neon.md (movmisalign<mode>): Use
4429 neon_perm_struct_or_reg_operand instead of
4430 neon_struct_or_register_operand.
4431 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
4432 neon_permissive_struct_operand instead of neon_struct_operand.
4433 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
4434 neon_vector_mem_operand.
4435 * config/arm/predicates.md (neon_struct_operand): Adjust call to
4436 neon_vector_mem_operand.
4437 (neon_permissive_struct_operand): New.
4438 (neon_struct_or_register_operand): Rename to...
4439 (neon_perm_struct_or_reg_operand): This. Adjust call to
4440 neon_vector_mem_operand.
4441
d579fcda
RB
44422013-06-18 Richard Biener <rguenther@suse.de>
4443
4444 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
4445 * lto-streamer.h: Include pointer-set.h.
4446 (struct lto_decl_slot): Remove.
4447 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
4448 Remove next_index entry.
4449 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4450 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4451 (lto_init_tree_ref_encoder): Adjust.
4452 (lto_destroy_tree_ref_encoder): Likewise.
4453 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4454 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4455 (lto_output_decl_index): Adjust.
4456 (lto_new_out_decl_state): Likewise.
4457 (lto_record_function_out_decl_state): Likewise.
4458 * lto-streamer-out.c (copy_function): Likewise.
4459
31ee20ba
RB
44602013-06-18 Richard Biener <rguenther@suse.de>
4461
4462 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
4463 * cgraphunit.c: Include cfgloop.h.
4464 (init_lowered_empty_function): Initialize the loop tree.
4465 (assemble_thunk): Insert new BBs into loops.
4466
ac0511f2
RB
44672013-06-18 Richard Biener <rguenther@suse.de>
4468
4469 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
4470 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
4471 the map from cache entry to cache index optional.
4472 (streamer_tree_cache_replace_tree): Adjust accordingly.
4473 (streamer_tree_cache_append): Likewise.
4474 (streamer_tree_cache_delete): Likewise.
4475 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
4476 streamer cache map from cache entry to cache index.
4477 * lto-streamer-out.c (create_output_block): Adjust.
4478
09485a08
SN
44792013-06-18 Sofiane Naci <sofiane.naci@arm.com>
4480
22fc7d3d
UB
4481 * config/arm/arm.md (attribute "insn"): Move multiplication and
4482 division attributes to...
09485a08
SN
4483 (attribute "type"): ... here. Remove mult.
4484 (attribute "mul32"): New attribute.
4485 (attribute "mul64"): Add umaal.
4486 (*arm_mulsi3): Update attributes.
4487 (*arm_mulsi3_v6): Likewise.
4488 (*thumb_mulsi3): Likewise.
4489 (*thumb_mulsi3_v6): Likewise.
4490 (*mulsi3_compare0): Likewise.
4491 (*mulsi3_compare0_v6): Likewise.
4492 (*mulsi_compare0_scratch): Likewise.
4493 (*mulsi_compare0_scratch_v6): Likewise.
4494 (*mulsi3addsi): Likewise.
4495 (*mulsi3addsi_v6): Likewise.
4496 (*mulsi3addsi_compare0): Likewise.
4497 (*mulsi3addsi_compare0_v6): Likewise.
4498 (*mulsi3addsi_compare0_scratch): Likewise.
4499 (*mulsi3addsi_compare0_scratch_v6): Likewise.
4500 (*mulsi3subsi): Likewise.
4501 (*mulsidi3adddi): Likewise.
4502 (*mulsi3addsi_v6): Likewise.
4503 (*mulsidi3adddi_v6): Likewise.
4504 (*mulsidi3_nov6): Likewise.
4505 (*mulsidi3_v6): Likewise.
4506 (*umulsidi3_nov6): Likewise.
4507 (*umulsidi3_v6): Likewise.
4508 (*umulsidi3adddi): Likewise.
4509 (*umulsidi3adddi_v6): Likewise.
4510 (*smulsi3_highpart_nov6): Likewise.
4511 (*smulsi3_highpart_v6): Likewise.
4512 (*umulsi3_highpart_nov6): Likewise.
4513 (*umulsi3_highpart_v6): Likewise.
4514 (mulhisi3): Likewise.
4515 (*mulhisi3tb): Likewise.
4516 (*mulhisi3bt): Likewise.
4517 (*mulhisi3tt): Likewise.
4518 (maddhisi4): Likewise.
4519 (*maddhisi4tb): Likewise.
4520 (*maddhisi4tt): Likewise.
4521 (maddhidi4): Likewise.
4522 (*maddhidi4tb): Likewise.
4523 (*maddhidi4tt): Likewise.
4524 (divsi3): Likewise.
4525 (udivsi3): Likewise.
4526 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
4527 (thumb2_mulsi_short_compare0): Likewise.
4528 (thumb2_mulsi_short_compare0_scratch): Likewise.
4529 * config/arm/arm1020e.md (1020mult1): Update attribute change.
4530 (1020mult2): Likewise.
4531 (1020mult3): Likewise.
4532 (1020mult4): Likewise.
4533 (1020mult5): Likewise.
4534 (1020mult6): Likewise.
22fc7d3d
UB
4535 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
4536 change.
09485a08
SN
4537 (cortex_a15_mult64): Likewise.
4538 (cortex_a15_sdiv): Likewise.
4539 (cortex_a15_udiv): Likewise.
4540 * config/arm/arm1026ejs.md (mult1): Update attribute change.
4541 (mult2): Likewise.
4542 (mult3): Likewise.
4543 (mult4): Likewise.
4544 (mult5): Likewise.
4545 (mult6): Likewise.
4546 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
4547 (pj4_ir_div): Likewise.
4548 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
4549 (11_mult2): Likewise.
4550 (11_mult3): Likewise.
4551 (11_mult4): Likewise.
4552 (11_mult5): Likewise.
4553 (11_mult6): Likewise.
4554 (11_mult7): Likewise.
4555 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
4556 (cortex_a8_mla): Likewise.
4557 (cortex_a8_mull): Likewise.
4558 (cortex_a8_smulwy): Likewise.
4559 (cortex_a8_smlald): Likewise.
4560 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
4561 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
4562 (cortex_r4_mul_3): Likewise.
4563 (cortex_r4_mla_4): Likewise.
4564 (cortex_r4_mla_3): Likewise.
4565 (cortex_r4_smlald): Likewise.
4566 (cortex_r4_mull): Likewise.
4567 (cortex_r4_sdiv): Likewise.
4568 (cortex_r4_udiv): Likewise.
4569 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
4570 (cortex_a7_idiv): Likewise.
4571 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
4572 (9_mult2): Likewise.
4573 (9_mult3): Likewise.
4574 (9_mult4): Likewise.
4575 (9_mult5): Likewise.
4576 (9_mult6): Likewise.
4577 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
4578 (cortex_a53_sdiv): Likewise.
4579 (cortex_a53_udiv): Likewise.
4580 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
4581 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
4582 (mp626_mult2): Likewise.
4583 (mp626_mult3): Likewise.
4584 (mp626_mult4): Likewise.
4585 * config/arm/fa526.md (526_mult1): Update attribute change.
4586 (526_mult2): Likewise.
4587 * config/arm/arm-generic.md (mult): Update attribute change.
4588 (mult_ldsched_strongarm): Likewise.
4589 (mult_ldsched): Likewise.
4590 (multi_cycle): Likewise.
4591 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
4592 * config/arm/fa606te.md (606te_mult1): Update attribute change.
4593 (606te_mult2): Likewise.
4594 (606te_mult3): Likewise.
4595 (606te_mult4): Likewise.
4596 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
4597 (cortex_a9_mac16): Likewise.
4598 (cortex_a9_multiply): Likewise.
4599 (cortex_a9_mac): Likewise.
4600 (cortex_a9_multiply_long): Likewise.
4601 * config/arm/fa626te.md (626te_mult1): Update attribute change.
4602 (626te_mult2): Likewise.
4603 (626te_mult3): Likewise.
4604 (626te_mult4): Likewise.
4605
8951f345
RB
46062013-06-18 Richard Biener <rguenther@suse.de>
4607
4608 PR lto/57334
4609 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
4610
d8101d05
AK
46112013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4612
4613 PR target/57609
4614 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4615 with next_active_insn.
4616
69eff9da
AM
46172013-06-18 Alan Modra <amodra@gmail.com>
4618
4619 * config/rs6000/rs6000.h (enum data_align): New.
4620 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
4621 (DATA_ABI_ALIGNMENT): Define.
4622 (CONSTANT_ALIGNMENT): Correct comment.
4623 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
4624 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
4625
8ed00d76
DM
46262013-06-17 David Malcolm <dmalcolm@redhat.com>
4627
4628 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
4629 ATTRIBUTE_UNUSED marking.
4630
3d33d06b
SN
46312013-06-17 Sofiane Naci <sofiane.naci@arm.com>
4632
4633 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
4634 alternative and update.
4635 (aarch64_dup_lanedi): Delete.
4636 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
4637 * config/aarch64/aarch64-simd-builtins.def: Update.
4638
ee03e71d
RB
46392013-06-17 Richard Biener <rguenther@suse.de>
4640
4641 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
4642 (lto_input_scc): Declare.
4643 (lto_input_tree_1): Likewise.
4644 (struct lto_stats_d): Add num_tree_bodies_output and
4645 num_pickle_refs_output.
4646 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
4647 (lto_read_tree_1): Split out from ...
4648 (lto_read_tree): ... this.
4649 (lto_input_scc): New function.
4650 (lto_input_tree_1): Split out from ...
4651 (lto_input_tree): ... this. Handle LTO_tree_scc.
4652 (lto_data_in_create): Create the streamer cache without hashes.
4653 * lto-streamer-out.c (create_output_block): Create the streamer
4654 cache with hashes when not doing WPA.
4655 (lto_write_tree_1): Split out from ...
4656 (lto_write_tree): ... this.
4657 (get_symbol_initial_value): New function.
4658 (lto_output_tree_1): Split out from ...
4659 (lto_output_tree): ... this. Write trees as series of SCCs
4660 using a DFS walk via DFS_write_tree.
4661 (struct sccs, struct scc_entry): New types.
4662 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
4663 (DFS_write_tree_body): New function.
4664 (DFS_write_tree): Likewise.
4665 (hash_tree): Likewise.
4666 (scc_entry_compare): Likewise.
4667 (hash_scc): Likewise.
4668 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
4669 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
4670 TREE_CHAIN as regular reference.
4671 (streamer_read_integer_cst): Remove.
4672 (streamer_get_pickled_tree): Adjust.
4673 * tree-streamer-out.c (streamer_write_chain): Disable streaming
4674 of DECL_EXTERNALs in BLOCK_VARS for now.
4675 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
4676 reference.
4677 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
4678 Add hash value argument and record that if hashes are recorded
4679 in the cache.
4680 (streamer_tree_cache_insert_1): Adjust.
4681 (streamer_tree_cache_insert): Likewise.
4682 (streamer_tree_cache_insert_at): Rename to ...
4683 (streamer_tree_cache_replace_tree): ... this and adjust.
4684 (streamer_tree_cache_append): Adjust.
4685 (record_common_node): Likewise.
4686 (streamer_tree_cache_create): Add argument whether to
4687 record hash values together with trees.
4688 (streamer_tree_cache_delete): Adjust.
4689 * tree-streamer.h (struct streamer_tree_cache_d): Add
4690 vector of hashes.
4691 (streamer_read_integer_cst): Remove.
4692 (streamer_tree_cache_insert): Adjust.
4693 (streamer_tree_cache_append): Likewise.
4694 (streamer_tree_cache_insert_at): Rename to ...
4695 (streamer_tree_cache_replace_tree): ... this and adjust.
4696 (streamer_tree_cache_create): Add argument whether to record hashes.
4697 (streamer_tree_cache_get): Rename to ...
4698 (streamer_tree_cache_get_tree): ... this.
4699 (streamer_tree_cache_get_hash): New function.
4700 * tree.c (cache_integer_cst): New function.
4701 * tree.h (cache_integer_cst): Declare.
4702 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
4703 * lto-symtab.c (lto_varpool_replace_node): Only release
4704 DECL_INITIAL of non-prevailing decls.
4705 * varpool.c (varpool_remove_initializer): Do not release
4706 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
4707
0de86a92
JU
47082013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4709
4710 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
4711 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
4712 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
4713 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
4714 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
4715 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
4716 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
4717 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
4718 instead of TARGET_64BIT.
4719 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
4720 Require ISA_HAS_<D>DIV.
4721
74240413
RS
47222013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
4723
4724 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
4725 (mips*-*-linux*): Move default with_llsc setting to where other
4726 defaults are set.
4727 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
4728 with_arch block.
4729 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
4730 Likewise. Remove default with_tune setting. Move default float
4731 setting to its own block. Handle with_llsc in the same block as above.
4732
393f9fed
JR
47332013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
4734
4735 PR rtl-optimization/57425
4736 PR rtl-optimization/57569
299a5f6a 4737 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR
4738 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
4739 Changed all callers.
4740 (canon_anti_dependence): New function.
4741 * cse.c (check_dependence): Use canon_anti_dependence.
4742 * cselib.c (cselib_invalidate_mem): Likewise.
4743 * rtl.h (canon_anti_dependence): Declare.
4744
b259d352
JU
47452013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4746
4747 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
4748 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
4749 ".set mips3" for 64-bit targets.
4750
4de09b85
DC
47512013-06-15 Dehao Chen <dehao@google.com>
4752
4753 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
4754 * gimple-low.c (gimple_check_call_matching_types): Likewise.
4755 (gimple_check_call_args): Likewise.
4756 * value-prof.c (check_ic_target): Likewise.
4757 * ipa-inline.c (early_inliner): Likewise.
4758 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
4759 * cgraph.c (cgraph_create_edge_1): Likewise.
4760 (cgraph_make_edge_direct): Likewise.
4761
d1aa4a76
MM
47622013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
4763
4764 PR target/57615
4765 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
4766 rs6000_output_move_128bit to handle emitting quad memory
4767 operations. Set attribute length to 8 bytes.
4768
24244e4d
VP
47692013-06-14 Vidya Praveen <vidyapraveen@arm.com>
4770
4771 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
4772 New pattern.
4773 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
4774 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
4775 (aarch64_<su>mlsl<mode>): Likewise.
4776
f5273574
MS
47772013-06-14 Mike Stump <mikestump@comcast.net>
4778
4779 * Makefile.in (TARGET_H): Add insn-codes.h.
4780
ef7e7ab2
AM
47812013-06-14 Alan Modra <amodra@gmail.com>
4782
4783 PR middle-end/57134
4784 PR middle-end/57586
4785 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
4786 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
4787 bitfield expansion when EXPAND_MEMORY.
4788 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
4789
135faab6
MM
47902013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
4791
4792 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
4793 test for clearing quad memory on 32-bit later.
4794
948a1fd9
MG
47952013-06-13 Marc Glisse <marc.glisse@inria.fr>
4796
4797 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
4798 (fold_negate_expr): Likewise.
4799 (fold_real_zero_addition_p): Handle vectors.
4800 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
4801
0597fb9c
AM
48022013-06-14 Alan Modra <amodra@gmail.com>
4803
4804 * varasm.c (force_const_mem): Revert 2013-06-07 change.
4805
e5b962d0
JH
48062013-06-13 Jan Hubicka <jh@suse.cz>
4807
4808 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
4809 Local comdats are not externally visible.
4810 * symtab.c (dump_symtab_base): Dump externally visible.
4811 (verify_symtab_base): Verify back links in the symtab hash.
4812
a8355e51
BC
48132013-06-13 Bin Cheng <bin.cheng@arm.com>
4814
4815 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
4816 CONVERT_EXPR as equal nodes.
4817
8f7e6e33
BC
48182013-06-13 Bin Cheng <bin.cheng@arm.com>
4819
22fc7d3d 4820 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 4821
5d418483
MG
48222013-06-13 Marc Glisse <marc.glisse@inria.fr>
4823
4824 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
4825 Generalize to complex and vector.
4826 * tree.c (build_all_ones_cst): New function.
4827 * tree.h (build_all_ones_cst): Declare it.
4828
50751417
AM
48292013-06-13 Alan Modra <amodra@gmail.com>
4830
4831 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
4832 * config/rs6000/rs6000.md (signbittf2): New insn.
4833 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
4834 (abstf2_internal, cmptf_internal2): Likewise.
4835 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
4836
4b02c962
MM
48372013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4838 Pat Haugen <pthaugen@us.ibm.com>
4839 Peter Bergner <bergner@vnet.ibm.com>
4840
4841 * config/rs6000/rs6000.c (emit_load_locked): Add support for
4842 power8 byte, half-word, and quad-word atomic instructions.
4843 (emit_store_conditional): Likewise.
4844 (rs6000_expand_atomic_compare_and_swap): Likewise.
4845 (rs6000_expand_atomic_op): Likewise.
4846
4847 * config/rs6000/sync.md (larx): Add new modes for power8.
4848 (stcx): Likewise.
4849 (AINT): New mode iterator to include TImode as well as normal
4850 integer modes on power8.
4851 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
4852 that VSX registers are not considered. Use AINT mode iterator
4853 instead of INT1 to allow inclusion of quad word atomic operations
4854 on power8.
4855 (load_locked<mode>): Likewise.
4856 (store_conditional<mode>): Likewise.
4857 (atomic_compare_and_swap<mode>): Likewise.
4858 (atomic_exchange<mode>): Likewise.
4859 (atomic_nand<mode>): Likewise.
4860 (atomic_fetch_<fetchop_name><mode>): Likewise.
4861 (atomic_nand_fetch<mode>): Likewise.
4862 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
4863 each type.
4864 (ATOMIC): On power8, add QImode, HImode modes.
4865 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
4866 modes that promote to SImode.
4867 (load_lockedti): Convert TImode arguments to PTImode, so that we
4868 get a guaranteed even/odd register pair.
4869 (load_lockedpti): Likewise.
4870 (store_conditionalti): Likewise.
4871 (store_conditionalpti): Likewise.
4872
4873 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
4874 atomic load/store instructions.
4875 (HSI): Likewise.
4876
710b6cc1
RS
48772013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
4878
4879 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
4880 loads.
4881 (insn_count): New attribute, with most cases extracted from...
4882 (length): ...here. Redefine most cases in terms of insn_count.
4883 (single_insn): Delete.
4884 (can_delay): Use insn_count to check for single instructions.
4885 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
4886 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
4887 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
4888 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
4889 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
4890 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
4891 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
4892 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
4893 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
4894 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
4895 rather than "length".
4896 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
4897 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
4898 Use "insn_count" rather than "length".
4899 * config/mips/mips-dsp.md
4900 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
4901 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
4902 length attributes.
4903
1951f101
MG
49042013-06-12 Marc Glisse <marc.glisse@inria.fr>
4905
4906 PR tree-optimization/57361
4907 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
4908
8b033a8a
SN
49092013-06-12 Sofiane Naci <sofiane.naci@arm.com>
4910
22fc7d3d
UB
4911 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
4912 to split.
8b033a8a
SN
4913 (aarch64_simd_combine<mode>): New instruction expansion.
4914 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
4915 function prototype.
4916 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
4917 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
4918
ca0f62a8
JH
49192013-06-12 Jan Hubicka <jh@suse.cz>
4920
4921 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
4922 decl has when in streaming stage.
4923 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
4924 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
4925
cddd69eb
RS
49262013-06-12 Roland Stigge <stigge@antcom.de>
4927
4928 PR target/57578
4929 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
4930
bfdeda2c
JJ
49312013-06-12 Jakub Jelinek <jakub@redhat.com>
4932
4933 PR tree-optimization/57537
4934 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
4935 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
4936
a4fa02d1
RB
49372013-06-12 Richard Biener <rguenther@suse.de>
4938
4939 * data-streamer.h (streamer_write_char_stream): CSE
4940 obs->current_pointer.
4941 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
4942 streamer_write_char_stream manually and optimize the resulting loop.
4943 (streamer_write_hwi_stream): Likewise.
4944
bbf9ad07
JH
49452013-06-12 Jan Hubicka <jh@suse.cz>
4946
4947 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
4948 * cgraph.h (varpool_create_empty_node): Declare.
4949 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
4950 duplicated nodes.
4951 * symtab.c (symtab_unregister_node): Be lax about missin entries
4952 in node hash.
4953 (symtab_get_node): Update comment.
4954 * varpool.c (varpool_create_empty_node): Break out from ...
4955 (varpool_node_for_decl): ... here.
4956 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
4957
a53f90ad
EB
49582013-06-12 Eric Botcazou <ebotcazou@adacore.com>
4959
4960 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
4961 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
4962 part. Use straight-line flow at the end.
4963 <COMPONENT_REF>: Remove superfluous else.
4964 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
4965
946f9306
JJ
49662013-06-12 Jakub Jelinek <jakub@redhat.com>
4967
4968 PR target/56564
4969 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
4970 target hook even for !TREE_PUBLIC decls. If no resolution info
4971 is available, return false for common and external decls.
4972
a7fca6f0
KP
49732013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
4974
4975 * config/rl78/constraints.md (U): New constraint.
4976 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
4977 valloc attribute.
4978
c7d8f446
MM
49792013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
4980
4981 PR target/57589
4982 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
4983 to allow returning address to AT_PLATFORM name.
4984
08346abd
JH
49852013-06-11 Jan Hubicka <jh@suse.cz>
4986
4987 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
4988 * cgraph.h (symtab_node_base): Add weakref flag.
4989 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
4990 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
4991 (output_weakrefs): Use weakref flag.
4992 * fold-const.c (simple_operand_p): Handle WEAK.
4993 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
4994 * ipa.c (varpool_externally_visible_p): Drop weakref.
4995 (function_and_variable_visibility): Update comment; fix weakref
4996 sanity checks; do not clear DECL_WEAK on them.
4997 * lto-cgraph.c (lto_output_node): update.
4998 (lto_output_varpool_node): Update.
4999 (input_overwrite_node): Update.
5000 (input_node): Update.
5001 (input_varpool_node): Update.
5002 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5003 (lto_symtab_merge_symbols): Add sanity check.
5004 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5005 * passes.c (rest_of_decl_compilation): Set static flag, too.
5006 * symtab.c (dump_symtab_base): Dump weakref.
5007 (verify_symtab_base): Sanity check weakrefs.
5008 (symtab_make_decl_local): Remove duplicated code.
5009 (symtab_alias_ultimate_target): Simplify.
5010 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 5011
d9af3652
DD
50122013-06-11 DJ Delorie <dj@redhat.com>
5013
5014 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5015 (rl78_unwind_word_mode): New.
5016
ffa66012
DM
50172013-06-11 David Malcolm <dmalcolm@redhat.com>
5018
5019 * final.c (debug_prefix_maps): Make static.
5020
783f2d35
DM
50212013-06-11 David Malcolm <dmalcolm@redhat.com>
5022
5023 * function.c (initial_trampoline): Remove stray copy.
5024
7af26def
SN
50252013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5026
5027 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5028
97756c0e
MJ
50292013-06-11 Martin Jambor <mjambor@suse.cz>
5030
5031 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5032 within bounds at the beginning of the function.
5033
2c7b8bf6
AM
50342013-06-11 Alan Modra <amodra@gmail.com>
5035
5036 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5037 reporting.
5038 (get_named_section): Don't NULL !DECL_P decl.
5039
77cea46e
IZ
50402013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5041
5042 * doc/invoke.texi (core-avx2): Document.
5043 (slm): Likewise.
5044 (atom): Updated with MOVBE.
5045
888f0920
RB
50462013-06-11 Richard Biener <rguenther@suse.de>
5047
22fc7d3d 5048 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 5049
6b438d58
AB
50502013-06-11 Anton Blanchard <anton@samba.org>
5051
5052 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5053 correct shift value in little-endian mode.
5054
3bc66938
JJ
50552013-06-11 Jakub Jelinek <jakub@redhat.com>
5056
5057 PR target/56564
5058 * varasm.c (get_variable_align): Move #endif to the right place.
5059
a5dc7f88
CC
50602013-06-10 Cary Coutant <ccoutant@google.com>
5061
5062 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5063 for hash so that hash table traversal order is deterministic.
5064
19be72ab
MM
50652013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5066 Pat Haugen <pthaugen@us.ibm.com>
5067 Peter Bergner <bergner@vnet.ibm.com>
5068
5069 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5070 of vectors in GPRS if they are direct moves or quad word load or
5071 store moves.
5072
5073 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5074 declaration.
5075 (direct_move_p): Likewise.
5076 (quad_load_store_p): Likewise.
5077
5078 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5079 classes into bins based on the physical register type.
5080 (reg_class_to_reg_type): Likewise.
5081 (IS_STD_REG_TYPE): Likewise.
5082 (IS_FP_VECT_REG_TYPE): Likewise.
5083 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5084 use direct move instructions.
5085 (reload_gpr_vsx): Likewise.
5086 (reload_vsx_gpr): Likewise.
5087 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5088 information that is a simplification of register classes. Also
5089 precalculate direct move reload helpers.
5090 (direct_move_p): New function to return true if the operation can
5091 be done as a direct move instruciton.
5092 (quad_load_store_p): New function to return true if the operation
5093 is a quad memory operation.
5094 (rs6000_legitimize_address): If quad memory, only allow register
5095 indirect for TImode addresses.
5096 (rs6000_legitimate_address_p): Likewise.
5097 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5098 (rs6000_reload_register_type): Likewise.
5099 (register_to_reg_type): Return register type.
5100 (rs6000_secondary_reload_simple_move): New helper function for
5101 secondary reload and secondary memory needed to identify anything
5102 that is a simple move, and does not need reloading.
5103 (rs6000_secondary_reload_direct_move): New helper function for
5104 secondary reload to identify cases that can be done with several
5105 instructions via the direct move instructions.
5106 (rs6000_secondary_reload_move): New helper function for secondary
5107 reload to identify moves between register types that can be done.
5108 (rs6000_secondary_reload): Add support for quad memory operations
5109 and for direct move.
5110 (rs6000_secondary_memory_needed): Likewise.
5111 (rs6000_debug_secondary_memory_needed): Change argument names.
5112 (rs6000_output_move_128bit): New function to return the move to
5113 use for 128-bit moves, including knowing about the various
5114 limitations of quad memory operations.
5115
5116 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5117 memory operations. call rs6000_output_move_128bit for the actual
5118 instruciton(s) to generate.
5119 (vsx_movti_64bit): Likewise.
5120
5121 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5122 (UNSPEC_P8V_MTVSRWZ): Likewise.
5123 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5124 (UNSPEC_P8V_MTVSRD): Likewise.
5125 (UNSPEC_P8V_XXPERMDI): Likewise.
5126 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5127 (UNSPEC_FUSION_GPR): Likewise.
5128 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 5129 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM
5130 (f32_sv): Likewise.
5131 (f32_dm): Likewise.
5132 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5133 loads and direct move instructions.
5134 (zero_extendsidi2_lfiwzx): Likewise.
5135 (extendsidi2_lfiwax): Likewise.
5136 (extendsidi2_nocell): Likewise.
5137 (floatsi<mode>2_lfiwax): Likewise.
5138 (lfiwax): Likewise.
5139 (floatunssi<mode>2_lfiwzx): Likewise.
5140 (lfiwzx): Likewise.
5141 (fix_trunc<mode>_stfiwx): Likewise.
5142 (fixuns_trunc<mode>_stfiwx): Likewise.
5143 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5144 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5145 (parity<mode>2_cmpb): Set length/type attr.
5146 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5147 for 'mr.' to fast_compare.
5148 (bpermd_<mode>): Change type attr to popcnt.
5149 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5150 (p8_mtvsrwz_1): Likewise.
5151 (p8_mtvsrwz_2): Likewise.
5152 (reload_fpr_from_gpr<mode>): Likewise.
5153 (p8_mtvsrd_1): Likewise.
5154 (p8_mtvsrd_2): Likewise.
5155 (p8_xxpermdi_<mode>): Likewise.
5156 (reload_vsx_from_gpr<mode>): Likewise.
5157 (reload_vsx_from_gprsf): Likewise.
5158 (p8_mfvsrd_3_<mode>): LIkewise.
5159 (reload_gpr_from_vsx<mode>): Likewise.
5160 (reload_gpr_from_vsxsf): Likewise.
5161 (p8_mfvsrd_4_disf): Likewise.
5162 (multi-word GPR splits): Do not split direct moves or quad memory
5163 operations.
5164
16876bdc
DM
51652013-06-10 David Malcolm <dmalcolm@redhat.com>
5166
5167 * tree-into-ssa.c (interesting_blocks): Make static.
5168
df8a1d28
JJ
51692013-06-10 Jakub Jelinek <jakub@redhat.com>
5170
5171 PR target/56564
5172 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5173 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5174 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5175 (get_variable_align): New function.
5176 (get_variable_section, emit_bss, emit_common,
5177 assemble_variable_contents, place_block_symbol): Use
5178 get_variable_align instead of DECL_ALIGN.
5179 (assemble_noswitch_variable): Add align argument, use it
5180 instead of DECL_ALIGN.
5181 (assemble_variable): Adjust caller. Use get_variable_align
5182 instead of DECL_ALIGN.
5183 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5184 caller.
5185 (DATA_ABI_ALIGNMENT): Define.
5186 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5187 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5188 opt is false, only return the psABI mandated alignment increase.
5189 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5190 (DATA_ABI_ALIGNMENT): ... this.
5191 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5192 (DATA_ABI_ALIGNMENT): ... this.
5193 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5194 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5195 (DATA_ABI_ALIGNMENT): ... this.
5196 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5197 * doc/tm.texi: Regenerated.
5198
57b29ca6
UB
51992013-06-10 Uros Bizjak <ubizjak@gmail.com>
5200
5201 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5202 cmp_code to construct REG_EQUAL note.
5203
3e56ed50
JJ
52042013-06-09 Jakub Jelinek <jakub@redhat.com>
5205
5206 PR target/57568
5207 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5208 that operands[2] doesn't overlap with operands[0].
5209
74fe2a1d
DE
52102013-06-09 David Edelsohn <dje.gcc@gmail.com>
5211 Jan Hubicka <jh@suse.cz>
5212
5213 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5214 hack to mark symbols as used.
5215
ed52a84e
VM
52162013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5217
5218 PR rtl-optimization/57559
5219 * lra-constraints.c (process_alt_operands): Don't discourage
5220 memory with known offset for offsetable memory constraint.
5221 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5222
299316ed
EB
52232013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5224
5225 * varasm.c (struct oc_local_state): Reorder fields.
5226 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5227 and adjust accordingly.
5228 (output_constructor): Reorder initialization code and adjust call to
5229 output_constructor_bitfield.
5230
88ac60d3
JH
52312013-06-07 Jan Hubicka <jh@suse.cz>
5232
5233 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5234
4cdce1a8
DM
52352013-06-07 David Malcolm <dmalcolm@redhat.com>
5236
5237 * tree-object-size.c (unknown): Make const.
5238
b5e0425c
AK
52392013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5240
5241 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5242 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5243 for last alternative in the cpu_facility attribute.
5244
79678d04
KT
52452013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5246
5247 PR target/56315
5248 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5249 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5250 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5251 * config/arm/constraints.md (Dg): New constraint.
5252 * config/arm/neon.md (xordi3_neon): Remove.
5253 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5254 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5255
b31ddbdb
KT
52562013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5257
5258 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5259 Clean up alternatives.
5260
8f90b7d4
AM
52612013-06-07 Alan Modra <amodra@gmail.com>
5262
5263 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5264 va_list_gpr_size.
5265
961ce119
AM
52662013-06-07 Alan Modra <amodra@gmail.com>
5267
5268 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5269
33e49835
KT
52702013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5271
5272 * config/arm/constraints.md (Df): New constraint.
5273 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5274 Correct length attribute for last two alternatives.
5275
acd17ae6
AM
52762013-06-07 Alan Modra <amodra@gmail.com>
5277
5278 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5279 override user -mfp-in-toc.
5280 (offsettable_ok_by_alignment): Consider just the current access
5281 rather than the whole object, unless BLKmode. Handle
5282 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5283 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5284 for -mcmodel=medium.
5285 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5286 override user -mfp-in-toc or -msum-in-toc. Default to
5287 -mno-fp-in-toc for -mcmodel=medium.
5288
73310b0e
DD
52892013-06-06 DJ Delorie <dj@redhat.com>
5290
5291 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5292 TARGET_VALID_POINTER_MODE.
5293
0bd62dca
MM
52942013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5295 Pat Haugen <pthaugen@us.ibm.com>
5296 Peter Bergner <bergner@vnet.ibm.com>
5297
5298 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5299 Document new power8 builtins.
5300
5301 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5302 condition code register, to allow 128-bit logical operations to be
5303 done in the VSX or GPR registers.
5304 (nor<mode>3): Use the canonical form for nor.
5305 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5306 vclz*, and vpopcnt* vector instructions.
5307 (nand<mode>3): Likewise.
5308 (orc<mode>3): Likewise.
5309 (clz<mode>2): LIkewise.
5310 (popcount<mode>2): Likewise.
5311
5312 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5313 that only the GPRs are recognized.
5314
5315 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5316 support for new power8 builtins.
5317
5318 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5319 builtin functions.
5320 (xscvdpspn): Likewise.
5321 (vclz): Likewise.
5322 (vclzb): Likewise.
5323 (vclzh): Likewise.
5324 (vclzw): Likewise.
5325 (vclzd): Likewise.
5326 (vpopcnt): Likewise.
5327 (vpopcntb): Likewise.
5328 (vpopcnth): Likewise.
5329 (vpopcntw): Likewise.
5330 (vpopcntd): Likewise.
5331 (vgbbd): Likewise.
5332 (vmrgew): Likewise.
5333 (vmrgow): Likewise.
5334 (eqv): Likewise.
5335 (eqv_v16qi3): Likewise.
5336 (eqv_v8hi3): Likewise.
5337 (eqv_v4si3): Likewise.
5338 (eqv_v2di3): Likewise.
5339 (eqv_v4sf3): Likewise.
5340 (eqv_v2df3): Likewise.
5341 (nand): Likewise.
5342 (nand_v16qi3): Likewise.
5343 (nand_v8hi3): Likewise.
5344 (nand_v4si3): Likewise.
5345 (nand_v2di3): Likewise.
5346 (nand_v4sf3): Likewise.
5347 (nand_v2df3): Likewise.
5348 (orc): Likewise.
5349 (orc_v16qi3): Likewise.
5350 (orc_v8hi3): Likewise.
5351 (orc_v4si3): Likewise.
5352 (orc_v2di3): Likewise.
5353 (orc_v4sf3): Likewise.
5354 (orc_v2df3): Likewise.
5355
5356 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
5357 allow power8 quad mode in 64-bit.
5358 (rs6000_builtin_vectorized_function): Add support to vectorize
5359 ISA 2.07 count leading zeros, population count builtins.
5360 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
5361 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
5362 (builtin_function_type): Add vgbbd builtin function which takes an
5363 unsigned argument.
5364 (altivec_expand_vec_perm_const): Add support for new power8 merge
5365 instructions.
5366
5367 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
5368 that does not include TImdoe for use with 32-bit.
5369 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
5370 instructions.
5371 (UNSPEC_VSX_CVDPSPN): Likewise.
5372 (vsx_xscvdpspn): Likewise.
5373 (vsx_xscvspdpn): Likewise.
5374 (vsx_xscvdpspn_scalar): Likewise.
5375 (vsx_xscvspdpn_directmove): Likewise.
5376 (vsx_and<mode>3): Split logical operations into 32-bit and
5377 64-bit. Add support to do logical operations on TImode as well as
5378 VSX vector types. Allow logical operations to be done in either
5379 VSX registers or in general purpose registers in 64-bit mode. Add
5380 splitters if GPRs were used. For AND, add clobber of CCmode to
5381 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
5382 encoding.
5383 (vsx_and<mode>3_32bit): Likewise.
5384 (vsx_and<mode>3_64bit): Likewise.
5385 (vsx_ior<mode>3): Likewise.
5386 (vsx_ior<mode>3_32bit): Likewise.
5387 (vsx_ior<mode>3_64bit): Likewise.
5388 (vsx_xor<mode>3): Likewise.
5389 (vsx_xor<mode>3_32bit): Likewise.
5390 (vsx_xor<mode>3_64bit): Likewise.
5391 (vsx_one_cmpl<mode>2): Likewise.
5392 (vsx_one_cmpl<mode>2_32bit): Likewise.
5393 (vsx_one_cmpl<mode>2_64bit): Likewise.
5394 (vsx_nor<mode>3): Likewise.
5395 (vsx_nor<mode>3_32bit): Likewise.
5396 (vsx_nor<mode>3_64bit): Likewise.
5397 (vsx_andc<mode>3): Likewise.
5398 (vsx_andc<mode>3_32bit): Likewise.
5399 (vsx_andc<mode>3_64bit): Likewise.
5400 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
5401 and xxlorc instructions.
5402 (vsx_eqv<mode>3_64bit): Likewise.
5403 (vsx_nand<mode>3_32bit): Likewise.
5404 (vsx_nand<mode>3_64bit): Likewise.
5405 (vsx_orc<mode>3_32bit): Likewise.
5406 (vsx_orc<mode>3_64bit): Likewise.
5407
5408 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
5409
5410 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
5411 instruction.
5412 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
5413 (p8_vmrgow): Likewise.
5414 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
5415 GPRs to be split under VSX.
5416 (p8v_clz<mode>2): Add power8 count leading zero support.
5417 (p8v_popcount<mode>2): Add power8 population count support.
5418 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
5419 support.
5420
5421 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
5422 instruction.
5423
5424 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
5425 builtin functions.
5426 (vec_nand): Likewise.
5427 (vec_vclz): Likewise.
5428 (vec_vclzb): Likewise.
5429 (vec_vclzd): Likewise.
5430 (vec_vclzh): Likewise.
5431 (vec_vclzw): Likewise.
5432 (vec_vgbbd): Likewise.
5433 (vec_vmrgew): Likewise.
5434 (vec_vmrgow): Likewise.
5435 (vec_vpopcnt): Likewise.
5436 (vec_vpopcntb): Likewise.
5437 (vec_vpopcntd): Likewise.
5438 (vec_vpopcnth): Likewise.
5439 (vec_vpopcntw): Likewise.
5440
37684c46
VM
54412013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5442
5443 PR rtl-optimization/57468
5444 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
5445 spilled pseudos.
5446
d89ae27c
VM
54472013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5448
5449 PR rtl-optimization/57459
5450 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
5451 type when setting live regs.
5452
3597e113
VM
54532013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5454
5455 * config/s390/s390.opt (mlra): New option.
5456 * config/s390/s390.c (s390_decompose_address): Check displacement
5457 for all registers for LRA.
5458 (s390_secondary_reload): Don't used secondary reloads for LRA.
5459 (s390_lra_p): New function.
5460 (TARGET_LRA_P): Define.
5461 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
5462 of attribute cpu_facility to zarch for the last alternative.
5463 (*cmpmem_short): Ditto.
5464
01037aeb
EB
54652013-06-06 Eric Botcazou <ebotcazou@adacore.com>
5466
5467 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
5468 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
5469 (arm_expand_prologue): Likewise.
5470
3371a64f
TJ
54712013-06-06 Teresa Johnson <tejohnson@google.com>
5472
5473 PR c++/53743
5474 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
5475 as this is now done by redirect_edge_and_branch_force.
5476 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
5477 barriers, and fix interaction with splitting.
5478 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
5479 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
5480 reflect changes made in the routine.
5481 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
5482 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
5483 since this is called in cfglayout mode, and replace partition fixup
5484 with assert as that is now done by force_nonfallthru_and_redirect.
5485 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
5486 already be marked with region crossing note.
5487 (insert_section_boundary_note): Make non-static, gate on flag
5488 has_bb_partition, rewrite to also check for multiple partitions.
5489 (rest_of_handle_reorder_blocks): Remove call to
5490 insert_section_boundary_note, now done later during free_cfg.
5491 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
5492 * bb-reorder.h (insert_section_boundary_note): Declare.
5493 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
5494 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
5495 invoke insert_section_boundary_note.
5496 (try_redirect_by_replacing_jump): Remove unnecessary
5497 check for region crossing note.
5498 (fixup_partition_crossing): New function.
5499 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
5500 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
5501 in non-cfglayout mode.
5502 (force_nonfallthru_and_redirect): Fixup partition boundaries,
5503 remove old code that tried to do this. Emit barrier correctly
5504 when we are in cfglayout mode.
5505 (last_bb_in_partition): New function.
5506 (rtl_split_edge): Correctly fixup partition boundaries.
5507 (commit_one_edge_insertion): Remove old code that tried to
5508 fixup region crossing edge since this is now handled in
5509 split_block, and set up insertion point correctly since
5510 block may now end in a jump.
5511 (verify_hot_cold_block_grouping): Guard against checking when not in
5512 linearized RTL mode.
5513 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
5514 notes.
5515 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
5516 rtl_verify_flow_info, so not called in cfglayout mode.
5517 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
5518 (fixup_reorder_chain): Remove old code that attempted to fixup region
5519 crossing note as this is now handled in force_nonfallthru_and_redirect.
5520 (duplicate_insn_chain): Don't duplicate switch section notes.
5521 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
5522 note.
5523 * basic-block.h (emit_barrier_after_bb): Declare.
5524
66071e10
KT
55252013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5526
5527 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
5528 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
5529 arm_usatsihi): Adjust alternatives for arm_restrict_it.
5530
0e26bf3d
KT
55312013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5532
5533 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
5534 where appropriate.
5535 * config/arm/ldmstm.md: Regenerate.
5536
12b4e7ef
KT
55372013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5538
5539 * config/arm/sync.md (atomic_loaddi_1):
5540 Disable predication for arm_restrict_it.
5541 (arm_load_exclusive<mode>): Likewise.
5542 (arm_load_exclusivesi): Likewise.
5543 (arm_load_exclusivedi): Likewise.
5544 (arm_load_acquire_exclusive<mode>): Likewise.
5545 (arm_load_acquire_exclusivesi): Likewise.
5546 (arm_load_acquire_exclusivedi): Likewise.
5547 (arm_store_exclusive<mode>): Likewise.
5548 (arm_store_exclusive<mode>): Likewise.
5549 (arm_store_release_exclusivedi): Likewise.
5550 (arm_store_release_exclusive<mode>): Likewise.
5551
15d16c8a
RB
55522013-06-06 Richard Biener <rguenther@suse.de>
5553
5554 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
5555 after LTO_null.
5556 (lto_tag_is_tree_code_p): Adjust.
5557 (lto_tag_is_gimple_code_p): Likewise.
5558 (lto_gimple_code_to_tag): Likewise.
5559 (lto_tag_to_gimple_code): Likewise.
5560 (lto_tree_code_to_tag): Likewise.
5561 (lto_tag_to_tree_code): Likewise.
5562 * data-streamer.h (streamer_write_hwi_in_range): Use
5563 uhwi streaming to stream the normalized range.
5564 (streamer_read_hwi_in_range): Likewise.
5565
17a7fc37
KT
55662013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5567
5568 * config/arm/arm.md (enabled_for_depr_it): New attribute.
5569 (predicable_short_it): Likewise.
5570 (predicated): Likewise.
5571 (enabled): Handle above.
5572 (define_cond_exec): Set predicated attribute to yes.
5573
b57ca59b
MS
55742013-06-05 Mike Stump <mikestump@comcast.net>
5575
5576 * gdbinit.in (__FUNCTION__): Add.
5577
c1e183a9
UB
55782013-06-05 Uros Bizjak <ubizjak@gmail.com>
5579
5580 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
5581 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
5582
0378bddb
JH
55832013-06-05 Jan Hubicka <jh@suse.cz>
5584
5585 * varasm.c (mark_decl_referenced): Revert the removal until targets
5586 are fixed.
5587
8f8a2057
DE
55882013-06-05 David Edelsohn <dje.gcc@gmail.com>
5589
5590 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
5591 instead of mark_decl_referenced.
5592
edb983b2
JH
55932013-06-05 Jan Hubicka <jh@suse.cz>
5594
5595 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
5596 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
5597 and symtab_used_from_object_file_p.
5598 (cgraph_make_node_local_1): Clear forced_by_abi.
5599 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
5600 * cgraph.h (symtab_node_base): Add forced_by_abi.
5601 (decide_is_variable_needed): Remove.
5602 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
5603 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
5604 (decide_is_symbol_needed): ... this one; handle symbols in general;
5605 always analyze virtuals; honnor forced_by_abi.
5606 (cgraph_finalize_function): Update.
5607 (varpool_finalize_decl): Update.
5608 (symbol_defined_and_needed): Remove.
5609 (analyze_functions): Update.
5610 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
5611 output_refs, input_overwrite_node): Handle forced_by_abi.
5612 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
5613 (address_taken_from_non_vtable_p): ... this one.
5614 (comdat_can_be_unshared_p_1): New function.
5615 (cgraph_comdat_can_be_unshared_p): Rename to ...
5616 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
5617 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
5618 (function_and_variable_visibility): Clear forced_by_abi as needed.
5619 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
5620 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH
5621 * symtab.c (dump_symtab_base): Dump forced_by_abi.
5622 * varpool.c (decide_is_variable_needed): Remove.
5623
9912dbe5
KT
56242013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5625
5626 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
5627 (arm_option_override): Override arm_restrict_it where appropriate.
5628 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
5629 * config/arm/arm.opt (mrestrict-it): New command-line option.
5630 * doc/invoke.texi: Document -mrestrict-it.
5631
34c136b6
DM
56322013-06-05 David Malcolm <dmalcolm@redhat.com>
5633
5634 * tsan.c (tsan_atomic_table): Make const.
5635
a367df53
RB
56362013-06-05 Richard Biener <rguenther@suse.de>
5637
5638 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
5639 index associated with the tree we are supposed to replace.
5640 * tree-streamer-out.c (pack_ts_base_value_fields): Output
5641 TREE_ASM_WRITTEN as zero for everything but SSA names.
5642
70d51a19
DM
56432013-06-05 David Malcolm <dmalcolm@redhat.com>
5644
5645 * tree-ssa-structalias.c (call_stmt_vars): Make static.
5646
b9bd2075
JH
56472013-06-04 Jan Hubicka <jh@suse.cz>
5648
5649 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
5650 (input_node, input_varpool_node): Handle correctly external same
5651 body aliases.
5652 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
5653 nodes at ltrans stage.
5654
aaae719d
JH
56552013-06-04 Jan Hubicka <jh@suse.cz>
5656
5657 * ipa-inline.c (update_caller_keys): Fix availability test.
5658 (update_callee_keys): Likewise.
5659 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
5660 to follow ELF standard.
5661
107eea2c
JU
56622013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
5663
5664 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
5665 (mips64r5900el-*-elf*): New configurations.
5666 * config/mips/mips-cpus.def (r5900): New processor.
5667 * config/mips/mips-tables.opt: Regenerate.
5668 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
5669 (mips_issue_rate): Handle PROCESSOR_R5900.
5670 (mips_reorg_process_insns): Force reorder mode for the R5900.
5671 * config/mips/mips.h (TARGET_MIPS5900): Define.
5672 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
5673 TARGET_MIPS5900.
5674 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
5675 TARGET_MIPS5900.
5676 * config/mips/mips.md (processor): Add r5900.
5677 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
5678
b7342d25
IB
56792013-06-04 Ian Bolton <ian.bolton@arm.com>
5680
5681 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
5682 into function to generate MOVI instruction.
c1e183a9 5683 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB
5684 (aarch64_preferred_simd_mode): Turn into wrapper.
5685 (aarch64_output_scalar_simd_mov_immediate): New function.
5686 * config/aarch64/aarch64-protos.h: Add prototype for above.
5687
81c2dfb9
IB
56882013-06-04 Ian Bolton <ian.bolton@arm.com>
5689
5690 * config/aarch64/aarch64.c (simd_immediate_info): Remove
5691 element_char member.
5692 (sizetochar): Return signed char.
5693 (aarch64_simd_valid_immediate): Remove elchar and other
5694 unnecessary variables.
5695 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
5696 Calculate element_char as required.
5697 * config/aarch64/aarch64-protos.h: Update and move prototype
5698 for aarch64_output_simd_mov_immediate.
5699 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
5700 Update arguments.
5701
48063b9d
IB
57022013-06-04 Ian Bolton <ian.bolton@arm.com>
5703
5704 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
5705 information completed by aarch64_simd_valid_immediate.
5706 (aarch64_legitimate_constant_p): Update arguments.
5707 (aarch64_simd_valid_immediate): Work with struct rather than many
5708 pointers.
5709 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
5710 (aarch64_simd_make_constant): Update arguments.
5711 (aarch64_output_simd_mov_immediate): Work with struct rather than
5712 many pointers. Output immediate directly rather than as operand.
5713 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5714 Update prototype.
5715 * config/aarch64/constraints.md (Dn): Update arguments.
5716
3ea63f60
IB
57172013-06-04 Ian Bolton <ian.bolton@arm.com>
5718
5719 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
5720 longer static.
5721 (aarch64_simd_immediate_valid_for_move): Remove.
5722 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
5723 (aarch64_simd_make_constant): Update call.
5724 (aarch64_output_simd_mov_immediate): Update call.
5725 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5726 Add prototype.
5727 * config/aarch64/constraints.md (Dn): Update call.
5728
d8edd899
IB
57292013-06-04 Ian Bolton <ian.bolton@arm.com>
5730
5731 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
5732 return type to bool for prototype.
5733 (aarch64_legitimate_constant_p): Check for true instead of not -1.
5734 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
5735 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
5736
44b20bb8
CM
57372013-06-04 Catherine Moore <clm@codesourcery.com>
5738
5739 * config/mips/mips.opt (meva): New.
c1e183a9 5740 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM
5741 (ASM_SPEC): Handle -meva.
5742 * doc/invoke.texi (meva): Document.
5743
52befbd8
AM
57442013-06-04 Alan Modra <amodra@gmail.com>
5745
5746 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
5747 constant output.
5748
aadaf24e
KT
57492013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5750
5751 * rtl.def: Add extra fourth optional field to define_cond_exec.
5752 * gensupport.c (process_one_cond_exec): Process attributes from
5753 define_cond_exec.
5754 * doc/md.texi: Document fourth field in define_cond_exec.
5755
0cd9e9ee
EB
57562013-06-04 Eric Botcazou <ebotcazou@adacore.com>
5757
5758 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
5759 out the processing order as in store_bit_field_1.
5760
cb0f665d
JH
57612013-06-04 Jan Hubicka <jh@suse.cz>
5762
5763 PR middle-end/57500
5764 * cgraphunit.c (cgraph_process_same_body_aliases): Create
5765 non-VAR_DECL node if it does not exist yet.
5766
53984b9b
RS
57672013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
5768
5769 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
5770 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
5771 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
5772 target_cpu_default setting.
5773
5d318fd4
TJ
57742013-06-03 Teresa Johnson <tejohnson@google.com>
5775
5776 * dumpfile.c (opt_info_switch_p): Change -fopt-info
5777 default to -fopt-info=optimized instead of all.
5778 * doc/invoke.texi: Ditto.
5779 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
5780 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
5781 (execute_vect_slp): Emit BB vectorization success under
5782 MSG_OPTIMIZED_LOCATIONS.
5783 * tree-vect-slp.c (vect_slp_transform_bb): Change
5784 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
5785 * tree-vect-loop.c (vect_transform_loop): Ditto.
5786
ec3af349
JM
57872013-06-03 Jason Merrill <jason@redhat.com>
5788
5789 PR c++/57415
5790 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
5791 Use TARGET_EXPR for C++.
5792
1baf9159
JJ
57932013-06-03 Jakub Jelinek <jakub@redhat.com>
5794
5795 PR rtl-optimization/57268
0cd9e9ee 5796 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ
5797 if DEBUG_INSN_P (insn).
5798
5799 Reapply
5800 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
5801
5802 PR rtl-optimization/57268
5803 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5804 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5805
0f1d3965
YR
58062013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
5807
5808 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
5809 (ix86_avoid_lea_for_addr): Likewise.
5810 (exact_dependency_1): Likewise.
5811 (ix86_adjust_cost): Likewise.
5812 (swap_top_of_ready_list): Fix formatting and !reload_completed check
5813 removed.
5814 (do_reorder_for_imul): Fix typo, formatting and
5815 !reload_completed check removed.
5816 (ix86_sched_reorder): Fix typo and formatting.
5817 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
5818 list.
5819
3bc30db4
SN
58202013-06-03 Sofiane Naci <sofiane.naci@arm.com>
5821
5822 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
5823
19cc6fac
EB
58242013-06-03 Eric Botcazou <ebotcazou@adacore.com>
5825
5826 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
5827 <STRING_CST>: Likewise.
5828 <VECTOR_CST>: Likewise.
5829
c68a6e08
JW
58302013-06-01 Janus Weil <janus@gcc.gnu.org>
5831 Mikael Morin <mikael@gcc.gnu.org>
5832
5833 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
5834 * config.in: Regenerated.
5835 * configure: Regenerated.
5836
38e55e5c
JH
58372013-06-01 Jan Hubicka <jh@suse.cz>
5838
5839 PR middle-end/57366
5840 * cgraphunit.c (compile): When weakref is not supported,
5841 set up transparent aliases before final output pass.
5842 * varasm.c (assemble_alias): Do not try to do it here.
5843
eb51d2ff
JH
58442013-06-01 Jan Hubicka <jh@suse.cz>
5845
5846 PR middle-end/57467
5847 * passes.c (for_per_function): Skip unanalyzed functions.
5848
40a7fe1e
JH
58492013-06-01 Jan Hubicka <jh@suse.cz>
5850
5851 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
5852 (lto_symtab_merge_symbols_1): ... this one.
5853 (lto_symtab_merge_cgraph_nodes): Rename to ...
5854 (lto_symtab_merge_symbols): ... this one; simplify.
5855 * cgraph.c (same_body_aliases_done): Rename to ...
5856 (cpp_implicit_aliases_done): ... this one.
5857 (cgraph_create_function_alias): Update.
5858 (cgraph_same_body_alias): Update.
c1e183a9 5859 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 5860 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 5861 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH
5862 (cgraph_node): Remove same_body_alias.
5863 (varpool_node): Remove alias_of and extra_name_alias.
5864 (same_body_aliases_done): Rename to ..
5865 (cpp_implicit_aliases_done): ... this one.
5866 (symtab_alias_ultimate_target): Add default parameter.
5867 (symtab_resolve_alias): New function.
5868 (fixup_same_cpp_alias_visibility): Declare.
5869 (cgraph_function_node): Add default parameter.
5870 (cgraph_node_asm_name): Likewise.
5871 (cgraph_function_or_thunk_node): Add default parameter; do
5872 not ICE when it is NULL.
5873 (varpool_variable_node): Likewise.
5874 * tree-emutls.c (create_emultls_var): Update.
5875 (ipa_lower_emutls): Update.
5876 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
5877 (cgraph_reset_node): Reset alias info.
5878 (cgraph_finalize_function): Update.
5879 (fixup_same_cpp_alias_visibility): Move to symtab.c.
5880 (analyze_function): Simplify.
5881 (cgraph_process_same_body_aliases): Simplify.
5882 (analyze_functions): Fixup same body aliases.
5883 (handle_alias_pairs): Simplify.
5884 (assemble_thunk): Update.
5885 (assemble_thunks_and_aliases): Update.
5886 (output_weakrefs): Rewrite.
5887 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
5888 (lto_output_varpool_node): Likewise.
5889 (compute_ltrans_boundary): Remve assert.
5890 (get_alias_symbol): New functoin.
5891 (input_node): Rewrite alias handling.
5892 (input_varpool_node): Likewise.
5893 * ipa-pure-const.c (propagate_pure_const): Fix formating.
5894 * ipa.c (process_references): Handle weakrefs correctly.
5895 (symtab_remove_unreachable_nodes): Likewise.
5896 * trans-mem.c (get_cg_data): Update.
5897 (ipa_tm_create_version_alias): Update.
5898 (ipa_tm_execute): Update.
5899 * symtab.c (dump_symtab_base): Dump aliases.
5900 (verify_symtab_base): Verify aliases.
5901 (symtab_node_availability): New function.
5902 (symtab_alias_ultimate_target): Simplify.
5903 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
5904 handle all the fixup cases.
5905 (symtab_resolve_alias): New function.
5906 * passes.c (ipa_write_summaries): Handle weakrefs.
5907 * varpool.c (varpool_analyze_node): Simplify.
5908 (assemble_aliases): Update.
5909 (varpool_create_variable_alias): Simplify.
5910 (varpool_extra_name_alias): Simplify.
5911 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
5912 (lto_symtab_merge_symbols): ... this one.
5913
21759881
DT
59142013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
5915
5916 Revert
5917 PR rtl-optimization/57268
5918 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5919 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5920
daf4e940
TB
59212013-06-01 Tobias Burnus <burnus@net-b.de>
5922
5923 Partially reverted:
5924 2013-05-31 Tobias Burnus <burnus@net-b.de>
5925
5926 PR middle-end/57073
5927 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
5928 further up.
5929
c359d8d0
DT
59302013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
5931
5932 PR rtl-optimization/57268
5933 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5934 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5935
4172215d
EB
59362013-05-31 Eric Botcazou <ebotcazou@adacore.com>
5937
5938 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
5939 unordered comparison operators when -fno-trapping-math is in effect
5940 on the e500.
5941 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
5942 and implement unordered comparison operators properly on the e500.
5943
a8c50132
EB
59442013-05-31 Eric Botcazou <ebotcazou@adacore.com>
5945
5946 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
5947 for constant scalar integers.
5948 (simplify_relational_operation_1): Likewise.
5949
44626634
SB
59502013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
5951
5952 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
5953 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
5954 Fix comment.
5955
55805e54
YR
59562013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
5957 Igor Zamyatin <igor.zamyatin@intel.com>
5958
5959 Silvermont (SLM) architecture performance tuning.
5960 * config/i386/i386.h (enum ix86_tune_indices): Add
5961 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
5962 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
5963
5964 * config/i386/i386.c (initial_ix86_tune_features)
5965 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
5966 (ix86_lea_outperforms): Handle Silvermont tuning.
5967 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
5968 call.
5969 (ix86_use_lea_for_mov): Likewise.
5970 (ix86_avoid_lea_for_addr): Likewise.
5971 (ix86_lea_for_add_ok): Likewise.
5972 (exact_dependency_1): New function.
5973 (exact_store_load_dependency): Likewise.
5974 (ix86_adjust_cost): Handle Silvermont tuning.
5975 (do_reoder_for_imul): Likewise.
5976 (swap_top_of_ready_list): New function.
5977 (ix86_sched_reorder): Changed to handle Silvermont tuning.
5978
5979 * config/i386/i386.md (peepholes that split memory operand in fp
5980 converts): New.
5981
e19c9de2
MS
59822013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
5983
5984 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5985 Remove un-necessary braces.
5986
38e6c9a6
MS
59872013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
5988
5989 * config/aarch64/aarch64.c (aarch64_classify_symbol):
5990 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
5991
e3530904
TB
59922013-05-31 Tobias Burnus <burnus@net-b.de>
5993
5994 PR middle-end/57073
c1e183a9 5995 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 5996
b6af05a9
KT
59972013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5998
5999 PR target/56315
6000 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6001 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6002 * config/arm/neon.md (iordi3_neon): Remove.
6003 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6004 * config/arm/predicates.md (imm_for_neon_logic_operand):
6005 Move to earlier in the file.
6006 (neon_logic_op2): Likewise.
6007 (arm_iordi_operand_neon): New predicate.
6008
f800c166
RB
60092013-05-31 Richard Biener <rguenther@suse.de>
6010
6011 PR tree-optimization/57478
6012 PR tree-optimization/57453
6013 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6014 are life as well.
6015
e571fa59
KP
60162013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6017
6018 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6019 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6020
0fa6e0ef
TB
60212013-05-30 Tobias Burnus <burnus@net-b.de>
6022 Thomas Koenig <tkoenig@gcc.gnu.org>
6023
6024 PR middle-end/57073
6025 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6026 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6027
4688ddf5
SB
60282013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6029
0fa6e0ef 6030 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 6031
3b9ceb4b
VM
60322013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6033
6034 * target.def (register_usage_leveling_p): New hook.
6035 * targhooks.c (default_register_usage_leveling_p): New.
6036 * targhooks.h (default_register_usage_leveling_p): New prototype.
6037 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6038 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6039 * doc/tm.texi: Update.
6040 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6041
65f9e789
IB
60422013-05-30 Ian Bolton <ian.bolton@arm.com>
6043
6044 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6045 (*insv_reg<mode>): New define_insn.
6046
ca035367
JR
60472013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6048
6049 PR rtl-optimization/57439
6050 * postreload.c (move2add_valid_value_p): Check that we have
6051 a zero subreg_regno_offset when accessing the register in
6052 the requested mode.
6053
0b871ccf
YR
60542013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6055 Igor Zamyatin <igor.zamyatin@intel.com>
6056
6057 Silvermont (SLM) architecture pipeline model, tuning and
6058 insn selection.
6059 * config.gcc: Add slm config options and target.
6060
6061 * config/i386/slm.md: New.
6062
6063 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6064
6065 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6066 PROCESSOR_SLM.
6067 (ix86_target_macros_internal): Likewise.
6068
6069 * gcc/config/i386/i386.c (slm_cost): New cost.
6070 (m_SLM): New macro flag.
6071 (initial_ix86_tune_features): Set m_SLM.
6072 (x86_accumulate_outgoing_args): Likewise.
6073 (x86_arch_always_fancy_math_387): Likewise.
6074 (processor_target_table): Add slm cost.
6075 (cpu_names): Add slm cpu name.
6076 (x86_option_override_internal): Set SLM ISA.
6077 (ix86_issue_rate): New case PROCESSOR_SLM.
6078 (ia32_multipass_dfa_lookahead): Likewise.
6079 (fold_builtin_cpu): Add slm.
6080
6081 * config/i386/i386.h (TARGET_SLM): New target macro.
6082 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6083 (processor_type): Add PROCESSOR_SLM.
6084
6085 * config/i386/i386.md (cpu): Add new value "slm".
6086 (slm.md): Include slm.md.
6087
24d5b097
XG
60882013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6089 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6090
6091 * config/arm/arm-protos.h: Add and update function protos.
6092 * config/arm/arm.c (use_simple_return_p): New added.
6093 (thumb2_expand_return): Check simple_return flag.
6094 * config/arm/arm.md: Add simple_return and conditional simple_return.
6095 * config/arm/iterators.md: Add iterator for return and simple_return.
6096
c1cccc15
ZC
60972013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6098
6099 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6100 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6101 (arm_emit_vfp_multi_reg_pop): Likewise.
6102 (thumb2_emit_ldrd_pop): Likewise.
6103 (arm_expand_epilogue): Add misc REG_CFA notes.
6104 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6105
3a4f280b
LC
61062013-05-29 Lawrence Crowl <crowl@google.com>
6107
6108 * config/arm/t-arm: Update for below.
6109
6110 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6111 Change type to hash_table. Update dependent calls and types.
6112
6113 * config/i386/t-cygming: Update for below.
6114
6115 * config/i386/t-interix: Update for below.
6116
6117 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6118 Change type to hash_table. Update dependent calls and types.
6119 (i386_find_on_wrapper_list::wrappers): Likewise.
6120
6121 * config/ia64/t-ia64: Update for below.
6122
6123 * config/ia64/ia64.c (bundle_state_table):
6124 Change type to hash_table. Update dependent calls and types.
6125
6126 * config/mips/mips.c (mips_reorg_process_insns::htab):
6127 Change type to hash_table. Update dependent calls and types.
6128
6129 * config/sol2.c (solaris_comdat_htab):
6130 Change type to hash_table. Update dependent calls and types.
6131
6132 * config/t-sol2: Update for above.
6133
1388a0e3
TJ
61342013-05-29 Teresa Johnson <tejohnson@google.com>
6135
6136 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6137 functions are not yet marked as defined.
6138
a5965b52
MM
61392013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6140 Pat Haugen <pthaugen@us.ibm.com>
6141 Peter Bergner <bergner@vnet.ibm.com>
6142
6143 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6144 instructions.
6145 (VEC_A): Likewise.
6146 (VEC_C): Likewise.
6147 (vrotl<mode>3): Likewise.
6148 (vashl<mode>3): Likewise.
6149 (vlshr<mode>3): Likewise.
6150 (vashr<mode>3): Likewise.
6151
6152 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6153 support for power8 V2DI builtins.
6154
6155 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6156 power8 V2DI builtins.
6157 (vupkhsw): Likewise.
6158 (vupklsw): Likewise.
6159 (vaddudm): Likewise.
6160 (vminsd): Likewise.
6161 (vmaxsd): Likewise.
6162 (vminud): Likewise.
6163 (vmaxud): Likewise.
6164 (vpkudum): Likewise.
6165 (vpksdss): Likewise.
6166 (vpkudus): Likewise.
6167 (vpksdus): Likewise.
6168 (vrld): Likewise.
6169 (vsld): Likewise.
6170 (vsrd): Likewise.
6171 (vsrad): Likewise.
6172 (vsubudm): Likewise.
6173 (vcmpequd): Likewise.
6174 (vcmpgtsd): Likewise.
6175 (vcmpgtud): Likewise.
6176 (vcmpequd_p): Likewise.
6177 (vcmpgtsd_p): Likewise.
6178 (vcmpgtud_p): Likewise.
6179 (vupkhsw): Likewise.
6180 (vupklsw): Likewise.
6181 (vaddudm): Likewise.
6182 (vmaxsd): Likewise.
6183 (vmaxud): Likewise.
6184 (vminsd): Likewise.
6185 (vminud): Likewise.
6186 (vpksdss): Likewise.
6187 (vpksdus): Likewise.
6188 (vpkudum): Likewise.
6189 (vpkudus): Likewise.
6190 (vrld): Likewise.
6191 (vsld): Likewise.
6192 (vsrad): Likewise.
6193 (vsrd): Likewise.
6194 (vsubudm): Likewise.
6195
6196 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6197 support for power8 V2DI instructions.
6198
6199 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6200 power8 V2DI instructions. Combine pack and unpack insns to use an
6201 iterator for each mode. Check whether a particular mode supports
6202 Altivec instructions instead of just checking TARGET_ALTIVEC.
6203 (UNSPEC_VPKUWUM): Likewise.
6204 (UNSPEC_VPKSHSS): Likewise.
6205 (UNSPEC_VPKSWSS): Likewise.
6206 (UNSPEC_VPKUHUS): Likewise.
6207 (UNSPEC_VPKSHUS): Likewise.
6208 (UNSPEC_VPKUWUS): Likewise.
6209 (UNSPEC_VPKSWUS): Likewise.
6210 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6211 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6212 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6213 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6214 (UNSPEC_VUPKHSB): Likewise.
6215 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6216 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6217 (UNSPEC_VUPKHSH): Likewise.
6218 (UNSPEC_VUPKLSB): Likewise.
6219 (UNSPEC_VUPKLSH): Likewise.
6220 (VI2): Likewise.
6221 (VI_char): Likewise.
6222 (VI_scalar): Likewise.
6223 (VI_unit): Likewise.
6224 (VP): Likewise.
6225 (VP_small): Likewise.
6226 (VP_small_lc): Likewise.
6227 (VU_char): Likewise.
6228 (add<mode>3): Likewise.
6229 (altivec_vaddcuw): Likewise.
6230 (altivec_vaddu<VI_char>s): Likewise.
6231 (altivec_vadds<VI_char>s): Likewise.
6232 (sub<mode>3): Likewise.
6233 (altivec_vsubcuw): Likewise.
6234 (altivec_vsubu<VI_char>s): Likewise.
6235 (altivec_vsubs<VI_char>s): Likewise.
6236 (altivec_vavgs<VI_char>): Likewise.
6237 (altivec_vcmpbfp): Likewise.
6238 (altivec_eq<mode>): Likewise.
6239 (altivec_gt<mode>): Likewise.
6240 (altivec_gtu<mode>): Likewise.
6241 (umax<mode>3): Likewise.
6242 (smax<mode>3): Likewise.
6243 (umin<mode>3): Likewise.
6244 (smin<mode>3): Likewise.
6245 (altivec_vpkuhum): Likewise.
6246 (altivec_vpkuwum): Likewise.
6247 (altivec_vpkshss): Likewise.
6248 (altivec_vpkswss): Likewise.
6249 (altivec_vpkuhus): Likewise.
6250 (altivec_vpkshus): Likewise.
6251 (altivec_vpkuwus): Likewise.
6252 (altivec_vpkswus): Likewise.
6253 (altivec_vpks<VI_char>ss): Likewise.
6254 (altivec_vpks<VI_char>us): Likewise.
6255 (altivec_vpku<VI_char>us): Likewise.
6256 (altivec_vpku<VI_char>um): Likewise.
6257 (altivec_vrl<VI_char>): Likewise.
6258 (altivec_vsl<VI_char>): Likewise.
6259 (altivec_vsr<VI_char>): Likewise.
6260 (altivec_vsra<VI_char>): Likewise.
6261 (altivec_vsldoi_<mode>): Likewise.
6262 (altivec_vupkhsb): Likewise.
6263 (altivec_vupkhs<VU_char>): Likewise.
6264 (altivec_vupkls<VU_char>): Likewise.
6265 (altivec_vupkhsh): Likewise.
6266 (altivec_vupklsb): Likewise.
6267 (altivec_vupklsh): Likewise.
6268 (altivec_vcmpequ<VI_char>_p): Likewise.
6269 (altivec_vcmpgts<VI_char>_p): Likewise.
6270 (altivec_vcmpgtu<VI_char>_p): Likewise.
6271 (abs<mode>2): Likewise.
6272 (vec_unpacks_hi_v16qi): Likewise.
6273 (vec_unpacks_hi_v8hi): Likewise.
6274 (vec_unpacks_lo_v16qi): Likewise.
6275 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6276 (vec_unpacks_lo_v8hi): Likewise.
6277 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6278 (vec_pack_trunc_v8h): Likewise.
6279 (vec_pack_trunc_v4si): Likewise.
6280 (vec_pack_trunc_<mode>): Likewise.
6281
6282 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6283 V2DI builtins.
6284 (vec_vmaxsd): Likewise.
6285 (vec_vmaxud): Likewise.
6286 (vec_vminsd): Likewise.
6287 (vec_vminud): Likewise.
6288 (vec_vpksdss): Likewise.
6289 (vec_vpksdus): Likewise.
6290 (vec_vpkudum): Likewise.
6291 (vec_vpkudus): Likewise.
6292 (vec_vrld): Likewise.
6293 (vec_vsld): Likewise.
6294 (vec_vsrad): Likewise.
6295 (vec_vsrd): Likewise.
6296 (vec_vsubudm): Likewise.
6297 (vec_vupkhsw): Likewise.
6298 (vec_vupklsw): Likewise.
6299
e70670cf
JH
63002013-05-29 Jan Hubicka <jh@suse.cz>
6301
6302 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6303 flags; reorder rest of fields in more consistent way.
6304 (varpool_node): Remove analyzed, finalized and alias.
6305 (cgraph_ndoe): Likewise.
6306 (symtab_alias_ultimate_target): New function.
6307 (cgraph_function_node): Move offline.
6308 (cgraph_reset_node): Declare.
6309 (cgraph_comdat_can_be_unshared_p): Remove.
6310 (varpool_remove_initializer): Declare.
6311 (varpool_first_defined_variable, varpool_next_defined_variable
6312 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6313 (cgraph_function_with_gimple_body_p): Update.
6314 (varpool_all_refs_explicit_p): Update.
6315 (symtab_alias_target): New function.
6316 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6317 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB
6318 (cgraph_function_or_thunk_node): Simplify using
6319 symtab_alias_ultimate_target.
e70670cf
JH
6320 (varpool_variable_node): Likewise.
6321 * cgraph.c (cgraph_create_function_alias): Update.
6322 (cgraph_add_thunk): Update.
6323 (cgraph_remove_node): Update.
6324 (dump_cgraph_node): Do not dump removed flags.
6325 (cgraph_function_body_availability): Update.
6326 (cgraph_propagate_frequency): Update.
6327 (verify_cgraph_node): Check sanity of local flag.
6328 (cgraph_function_node): Move here from cgraph.h; revamp for
6329 cgraph_function_or_thunk_node.
6330 * lto-symtab.c (lto_varpool_replace_node): Update.
6331 (lto_symtab_resolve_can_prevail_p): Update.
6332 (lto_symtab_merge_cgraph_nodes): Update.
6333 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6334 propagate_constants_accross_call, devirtualization_time_bonus,
6335 ipcp_propagate_stage): Update.
6336 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB
6337 * ipa-inline-transform.c (clone_inlined_nodes,
6338 preserve_function_body_p): Update.
e70670cf
JH
6339 * ipa-reference.c (propagate): Update.
6340 (write_node_summary_p): Update.
6341 * toplev.c (wrapup_global_declaration_2): Update.
6342 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6343 (analyze_function) ... this one.
6344 (cgraph_process_new_functions): Update.
6345 (cgraph_reset_node): Export.
6346 (cgraph_finalize_function): Update.
6347 (cgraph_add_new_function): Update.
6348 (process_function_and_variable_attributes): Update.
6349 (varpool_finalize_decl): Update.
6350 (symbol_finalized): Remove.
6351 (symbol_finalized_and_needed): Rename to ...
6352 (symbol_defined_and_needed): ... update.
6353 (cgraph_analyze_functions): Update.
6354 (handle_alias_pairs): Update.
6355 (mark_functions_to_output): Update.
6356 (assemble_thunk): Update.
6357 (output_in_order): Update.
6358 (output_weakrefs): Update.
6359 (finalize_compilation_unit): Update.
6360 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
6361 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
6362 input_node, input_varpool_node): Update.
6363 * dbxout.c (dbxout_expand_expr): Update.
6364 * cgraphclones.c (cgraph_clone_node): Update.
6365 (cgraph_copy_node_for_versioning): Update.
6366 (cgraph_materialize_clone): Update.
6367 (cgraph_materialize_all_clones): Update.
6368 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
6369 propagate_pure_const, propagate_nothrow): Update.
6370 * lto-streamer-out.c (lto_output, write_symbol): Update.
6371 * ipa-utils.c (ipa_reverse_postorder): Update.
6372 * ipa-inline.c (can_inline_edge_p): Update.
6373 (update_caller_keys, ipa_inline): Update.
6374 * dwarf2out.c (reference_to_unused,
6375 premark_types_used_by_global_vars_helper): Update.
6376 * tree-eh.c (tree_could_trap_p): Update.
6377 * ipa-split.c (consider_split, execute_split_functions): Update.
6378 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB
6379 has_addr_references_p): Update; move ahead in file for better
6380 readability.
e70670cf
JH
6381 (process_references): Simplify.
6382 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
6383 bodies are removed.
6384 (cgraph_comdat_can_be_unshared_p): Make static.
6385 (cgraph_externally_visible_p): Update.
6386 (varpool_externally_visible_p): Update.
6387 (function_and_variable_visibility): Update.
6388 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
6389 ipa_tm_mark_force_output_node): Update.
6390 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
6391 estimate_edge_devirt_benefit, inline_generate_summary,
6392 inline_write_summary): Update.
6393 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
6394 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB
6395 (ipa_print_node_params, ipa_prop_read_section,
6396 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH
6397 * varasm.c (mark_decl_referenced): Update.
6398 (assemble_alias, dump_tm_clone_pairs): Update.
6399 * tree-inline.c (copy_bb): Update.
6400 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
6401 Update.
6402 * symtab.c (dump_symtab_base): Print new flags.
6403 (verify_symtab_base): Verify new flags.
6404 (symtab_alias_ultimate_target): New function.
6405 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
6406 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
6407 Update.
c1e183a9
UB
6408 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
6409 Update.
e70670cf
JH
6410 * i386.c (ix86_get_function_versions_dispatcher,
6411 ix86_generate_version_dispatcher_body): Update.
6412 (fold_builtin_cpu): Use varpool_add_new_variable.
6413 * varpool.c (varpool_remove_initializer): Break out from ...
6414 (varpool_remove_node): ... this one.
6415 (dump_varpool_node, varpool_node_for_asm,
6416 cgraph_variable_initializer_availability, varpool_analyze_node,
6417 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB
6418 varpool_finalize_named_section_flags, varpool_create_variable_alias):
6419 Update.
e70670cf 6420
182802ad
JH
64212013-05-29 Jan Hubicka <jh@suse.cz>
6422
6423 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
6424
8190b609
ER
64252013-05-29 Easwaran Raman <eraman@google.com>
6426
6427 PR tree-optimization/57442
6428 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
6429 when control exits the main loop.
6430
69f5aa9b
SKS
64312013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
6432
6433 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
6434 and RX600.
c1e183a9 6435 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS
6436 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
6437 * rx/t-rx: Add rx100 under multi library matches option for nofpu
6438 option.
6439
4b847da9
BS
64402013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6441
6442 PR tree-optimization/57441
6443 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
6444 Don't limit size of incr_vec to number of candidates.
6445
4fc43c39
SE
64462013-05-29 Steve Ellcey <sellcey@imgtec.com>
6447
6448 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
6449 and mips16 directories.
c1e183a9 6450 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE
6451 (MULTILIB_DIRNAMES): Ditto.
6452 (MULTILIB_EXCEPTIONS): Add new exceptions.
6453 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
6454 (MULTILIB_DIRNAMES): Ditto.
6455 (MULTILIB_EXCEPTIONS): Add new exceptions.
6456
a5350ddc
CSS
64572012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6458 Marcus Shawcroft <marcus.shawcroft@arm.com>
6459
6460 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
6461 SYMBOL_TINY_ABSOLUTE.
6462 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
6463 SYMBOL_TINY_ABSOLUTE.
6464 (aarch64_expand_mov_immediate): Likewise.
6465 (aarch64_classify_symbol): Likewise.
6466 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
6467 Permit SYMBOL_TINY_ABSOLUTE.
6468 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
6469
17f4d4bf
CSS
64702013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6471 Marcus Shawcroft <marcus.shawcroft@arm.com>
6472
6473 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
6474 Refactor if/switch. Replace gcc_assert with if.
6475
c0186656
GG
64762013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
6477
6478 * config/i386/i386.c (initial_ix86_tune_features): Enable
6479 FP Reassociation for AMD bdver1 and bdver2.
6480
d20188f3
MJ
64812013-05-29 Martin Jambor <mjambor@suse.cz>
6482
6483 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
6484 and IMAGPART_EXPR do not occur within other handled_components.
6485
292cba13
RB
64862013-05-29 Richard Biener <rguenther@suse.de>
6487
6488 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
6489 access on whether the use is in the BB we currently try to
6490 vectorize.
6491 (vect_bb_vectorization_profitable_p): Pass the BB we currently
6492 vectorize to vect_bb_slp_scalar_cost.
6493
6eddf228
RB
64942013-05-29 Richard Biener <rguenther@suse.de>
6495
6496 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
6497 computing scalar cost offsetted by stmts that are kept live
6498 by scalar uses.
6499 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
6500 for computation of scalar cost.
6501
7df36117
SE
65022013-05-28 Steve Ellcey <sellcey@mips.com>
6503
6504 * config/mips/mips-cpus.def (mips32r2): Change processor type.
6505
36536d79
BI
65062013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6507
6508 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
6509 array notation built-in reduction functions.
6510 * doc/passes.texi (Passes): Added documentation about changes done
6511 for Cilk Plus.
6512 * doc/invoke.texi (C Dialect Options): Added documentation about
6513 the -fcilkplus flag.
6514 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
6515 (BUILTINS_DEF): Depend on cilkplus.def.
6516 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
6517 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
6518 * cilkplus.def: New file.
6519
b49eefa5
JR
65202013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
6521
6522 PR rtl-optimization/57439
6523 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
6524
9daf6dbb
ER
65252013-05-28 Easwaran Raman <eraman@google.com>
6526
6527 PR tree-optimization/57337
6528 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
6529 (find_insert_point): Correctly identify the insertion point
6530 when two statements with the same UID is compared.
6531
fbd7e877
RB
65322013-05-28 Richard Biener <rguenther@suse.de>
6533
6534 PR tree-optimization/56787
6535 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
6536 from the list of data references.
6537 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
6538 clobbers.
6539 (vect_analyze_loop_operations): Likewise.
6540 (vect_transform_loop): Remove clobbers.
6541
bbba1117
MJ
65422013-05-28 Martin Jambor <mjambor@suse.cz>
6543
6544 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
6545 and REALPART_EXPRs have scalar type.
6546
bd388c2a
RB
65472013-05-28 Richard Biener <rguenther@suse.de>
6548
6549 PR tree-optimization/57411
6550 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
6551 virtual operands.
6552 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
6553 virtual operand propagation.
6554
2f56a311
EB
65552013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6556
6557 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
6558 destination register for bmasksi_vis.
6559 (vector_init_bshuffle): Likewise.
6560 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
6561
5c3eacbb
EB
65622013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6563
6564 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
6565 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
6566 mode if the instruction isn't available in the original mode.
6567 * config/sparc/sparc.opt (mfix-ut699): New option.
6568 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
6569 (divdf3): Turn into expander.
6570 (divdf3_nofix): New insn.
6571 (divdf3_fix): Likewise.
6572 (divsf3): Disable if -mfix-ut699.
6573 (sqrtdf2): Turn into expander.
6574 (sqrtdf2_nofix): New insn.
6575 (sqrtdf2_fix): Likewise.
6576 (sqrtsf2): Disable if -mfix-ut699.
6577
a1756c0a
RB
65782013-05-27 Richard Biener <rguenther@suse.de>
6579
6580 PR middle-end/57412
6581 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
6582 block for the new loop.
6583
5a892248
RB
65842013-05-27 Richard Biener <rguenther@suse.de>
6585
6586 PR tree-optimization/57343
6587 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
6588 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
6589 (number_of_iterations_cond): Do not build the folded tree.
6590
d1de852b
RB
65912013-05-27 Richard Biener <rguenther@suse.de>
6592
6593 Revert
6594 PR middle-end/57381
6595 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6596 OEP_CONSTANT_ADDRESS_OF retained.
6597
6598 PR tree-optimization/57417
6599 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
6600 for unchanged base.
6601 (set_ssa_val_to): Compare addresses using
6602 get_addr_base_and_unit_offset.
6603
7894bc6b
JR
66042013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
6605
6606 PR rtl-optimization/56833
6607 * postreload.c (move2add_record_mode): New function.
6608 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
6609 (move2add_use_add2_insn): Use move2add_record_sym_value.
6610 (move2add_use_add3_insn): Likewise.
6611 (reload_cse_move2add): Use move2add_valid_value_p and
6612 move2add_record_mode. Invalidate call-clobbered and REG_INC
6613 affected regs by setting reg_mode to VOIDmode.
6614 (move2add_note_store): Don't pretend the inside of a SUBREG is
6615 the actual destination. Invalidate single/leading registers by
6616 setting reg_mode to VOIDmode.
6617 Use move2add_record_sym_value, move2add_valid_value_p and
6618 move2add_record_mode.
6619
b03be25f
RB
66202013-05-27 Richard Biener <rguenther@suse.de>
6621
6622 PR tree-optimization/57396
6623 * tree-affine.c (double_int_constant_multiple_p): Properly
6624 return false for val == 0 and div != 0.
6625
44e88db2
RS
66262013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6627
6628 * config/mips/mips.h: Use #elif in preprocessor conditions.
6629
3b859704
RS
66302013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6631
6632 PR target/53916
6633 * config/mips/constraints.md (kl): New constraint.
6634 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
6635 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
6636 constraint for operand 0. Split after CSE for MIPS16. Emit a move
6637 from LO for MIPS16.
6638 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
6639
c979d5f5
RS
66402013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6641
6642 PR target/55777
6643 * config/mips/mips.c (mips_can_inline_p): New function.
6644 (TARGET_CAN_INLINE_P): Define.
6645
8e90de43
SB
66462013-05-25 Steven Bosscher <steven@gcc.gnu.org>
6647
6648 * sched-int.h (ds_t, dw_t): Make unsigned int.
6649 Fix documentation that describes how all the ds_t bits are used.
6650 Reserve the last bit for delayed-branch scheduling.
6651 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
6652 (BITS_PER_DEP_WEAK): Fix definition and documentation.
6653 (gen_dep_weak_1): Remove prototype.
6654 * sched-deps.c (get_dep_weak_1): Make static.
6655 * target.def (speculate_insn, needs_block_p, gen_spec_check,
6656 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
6657 * doc/tm.texi: Regenerate.
6658 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
6659
cb5cb194
SB
66602013-05-24 Steven Bosscher <steven@gcc.gnu.org>
6661
6662 PR debug/56950
6663 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
6664
5bd93ff6
NS
66652013-05-24 Nathan Sidwell <nathan@codesourcery.com>
6666 Sandra Loosemore <sandra@codesourcery.com>
6667
6668 * config.gcc (powerpc-*): Allow native for with-cpu.
6669
2343af65
JL
66702013-05-24 Jeff Law <law@redhat.com>
6671
6672 PR tree-optimization/57124
6673 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
6674 conversion feeding a condition if the range has an overflow
6675 if -fstrict-overflow. Add warnings for when we do make the
6676 transformation.
6677
3ad695b9
DC
66782013-05-24 Dehao Chen <dehao@google.com>
6679
04960246 6680 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC
6681 (locus_discrim_hasher): Likewise.
6682 (locus_discrim_hasher::hash): Likewise.
6683 (locus_discrim_hasher::equal): Likewise.
6684
82338059
MJ
66852013-05-24 Martin Jambor <mjambor@suse.cz>
6686
6687 PR tree-optimization/57294
6688 * cgraph.h (ipa_record_stmt_references): Declare.
6689 * cgraphbuild.c (ipa_record_stmt_references): New function.
6690 (build_cgraph_edges): Use ipa_record_stmt_references.
6691 (rebuild_cgraph_edges): Likewise.
6692 (cgraph_rebuild_references): Likewise.
6693 * ipa-prop.c (ipa_modify_call_arguments): Discard references
6694 associated with the old statement and build references from the
6695 newly built statements.
6696 * ipa-ref.c (ipa_remove_stmt_references): New function.
6697 * ipa-ref.h (ipa_remove_stmt_references): Declare.
6698
1ccd4874
VM
66992013-05-24 Vladimir Makarov <vmakarov@redhat.com>
6700
55805e54 6701 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 6702 mem-mem moves.
55805e54 6703 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 6704 too.
55805e54 6705 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM
6706 secondary memory moves.
6707 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
6708 reg set up in the current insn.
6709
25e25c73
DC
67102013-05-24 Dehao Chen <dehao@google.com>
6711
04960246 6712 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC
6713 hash function.
6714 (locus_descrim_hasher::equal): Likewise.
04960246 6715 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC
6716 (make_edges): Likewise.
6717 (next_discriminator_for_locus): Likewise.
6718 (same_line_p): Likewise.
6719 (assign_discriminators): Likewise.
6720 (make_cond_expr_edges): Likewise.
6721 (make_gimple_switch_edges): Likewise.
6722 (make_goto_expr_edges): Likewise.
6723 (make_gimple_asm_edges): Likewise.
6724
50d38551
IB
67252013-05-24 Ian Bolton <ian.bolton@arm.com>
6726
6727 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
6728 X format specifier to only display bottom 16 bits.
6729 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
6730 immediate to match for operand 2, since it will be masked.
6731
aea0101d
RB
67322013-05-24 Richard Biener <rguenther@suse.de>
6733
6734 PR tree-optimization/57287
6735 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
6736 all SSA names that occur in abnormal PHIs.
6737
634e03d3
AI
67382013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
6739
6740 PR tree-ssa/57385
6741 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
6742 that index is not negative.
6743
b17c024f
EB
67442013-05-24 Eric Botcazou <ebotcazou@adacore.com>
6745
6746 PR rtl-optimization/55177
6747 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
6748 (simplify_byte_swapping_operation): New.
6749 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
6750 (simplify_relational_operation_1): Deal with BSWAP.
6751
46aeac1b
RH
67522013-05-23 Richard Henderson <rth@redhat.com>
6753
6754 PR target/56742
6755 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
6756 (ix86_reorg): Call it.
6757
70cc1536
UB
67582013-05-23 Uros Bizjak <ubizjak@gmail.com>
6759
6760 PR target/57379
6761 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
6762 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
6763 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
6764
b3851501
CB
67652013-05-23 Christian Bruel <christian.bruel@st.com>
6766
6767 PR debug/57351
6768 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
6769
4e194912
CSS
67702013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6771 Marcus Shawcroft <marcus.shawcroft@arm.com>
6772
6773 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
6774 * config/aarch64/constraints.md (Usa): Remove.
6775 * doc/md.texi (AArch64 Usa): Remove.
6776
83f8c414
CSS
67772013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6778 Marcus Shawcroft <marcus.shawcroft@arm.com>
6779
6780 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
6781 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
6782 * config/aarch64/predicates.md (aarch64_const_address): Remove.
6783 (aarch64_mov_operand): Use aarch64_mov_operand_p.
6784
b5574232
VP
67852013-05-23 Vidya Praveen <vidyapraveen@arm.com>
6786
6787 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
6788 instruction (AdvSIMD).
6789 * config/aarch64/aarch64-builtins.c
6790 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
6791 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
6792
76c36cb1 67932013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ
6794
6795 PR middle-end/57347
6796 * tree.h (contains_bitfld_component_ref_p): Declare.
6797 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB
6798 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
6799 caller.
7d2fb524
MJ
6800 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
6801 not access a bit-field. Assert all final offsets are byte-aligned.
6802
ce521ae6
RB
68032013-05-23 Richard Biener <rguenther@suse.de>
6804
6805 PR tree-optimization/57380
6806 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
6807 least one invariant or re-used load.
6808 * passes.c (init_optimization_passes): Move pass_phiprop before
6809 pass_forwprop.
6810
75c7257f
JG
68112013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
6812
6813 * config/aarch64/aarch64-simd.md
6814 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
6815
af6d44b5
RB
68162013-05-23 Richard Biener <rguenther@suse.de>
6817
6818 PR middle-end/57381
6819 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6820 OEP_CONSTANT_ADDRESS_OF retained.
6821
bd3647bf
JJ
68222013-05-23 Jakub Jelinek <jakub@redhat.com>
6823
6824 PR middle-end/57344
70cc1536
UB
6825 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
6826 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 6827
a55757ea
RB
68282013-05-23 Richard Biener <rguenther@suse.de>
6829
6830 PR rtl-optimization/57341
6831 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
6832 instead of true_dependence.
6833
cfb00b41
DM
68342013-05-22 David Malcolm <dmalcolm@redhat.com>
6835
dd1c676f
DM
6836 * bb-reorder.c (branch_threshold): Make const.
6837 (exec_threshold): Ditto.
cfb00b41 6838
f62511da
MM
68392013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6840 Pat Haugen <pthaugen@us.ibm.com>
6841 Peter Bergner <bergner@vnet.ibm.com>
6842
6843 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
6844 documentation for the power8 crypto builtins.
6845
6846 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
6847
6848 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
6849 macros for defining power8 builtin functions.
6850 (BU_P8V_AV_2): Likewise.
6851 (BU_P8V_AV_P): Likewise.
6852 (BU_P8V_VSX_1): Likewise.
6853 (BU_P8V_OVERLOAD_1): Likewise.
6854 (BU_P8V_OVERLOAD_2): Likewise.
6855 (BU_CRYPTO_1): Likewise.
6856 (BU_CRYPTO_2): Likewise.
6857 (BU_CRYPTO_3): Likewise.
6858 (BU_CRYPTO_OVERLOAD_1): Likewise.
6859 (BU_CRYPTO_OVERLOAD_2): Likewise.
6860 (XSCVSPDP): Fix typo, point to the correct instruction.
6861 (VCIPHER): Add power8 crypto builtins.
6862 (VCIPHERLAST): Likewise.
6863 (VNCIPHER): Likewise.
6864 (VNCIPHERLAST): Likewise.
6865 (VPMSUMB): Likewise.
6866 (VPMSUMH): Likewise.
6867 (VPMSUMW): Likewise.
6868 (VPERMXOR_V2DI): Likewise.
6869 (VPERMXOR_V4SI: Likewise.
6870 (VPERMXOR_V8HI: Likewise.
6871 (VPERMXOR_V16QI: Likewise.
6872 (VSHASIGMAW): Likewise.
6873 (VSHASIGMAD): Likewise.
6874 (VPMSUM): Likewise.
6875 (VPERMXOR): Likewise.
6876 (VSHASIGMA): Likewise.
6877
6878 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6879 __CRYPTO__ if the crypto instructions are available.
6880 (altivec_overloaded_builtins): Add support for overloaded power8
6881 builtins.
6882
6883 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
6884 support for power8 crypto builtins.
6885 (builtin_function_type): Likewise.
6886 (altivec_init_builtins): Add support for builtins that take vector
6887 long long (V2DI) arguments.
6888
6889 * config/rs6000/crypto.md: New file, define power8 crypto
6890 instructions.
6891
68922013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6893 Pat Haugen <pthaugen@us.ibm.com>
6894 Peter Bergner <bergner@vnet.ibm.com>
6895
6896 * doc/invoke.texi (Option Summary): Add power8 options.
6897 (RS/6000 and PowerPC Options): Likewise.
6898
6899 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
6900 constraints.md instead of rs6000.h. Reorder w* constraints. Add
6901 wm, wn, wr documentation.
6902
6903 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
6904 registers if direct move instructions are enabled.
6905 (wn): New constraint for no registers.
6906 (wq): New constraint for quad word even GPR registers.
6907 (wr): New constraint if 64-bit instructions are enabled.
6908 (wv): New constraint if power8 vector instructions are enabled.
6909 (wQ): New constraint for quad word memory locations.
6910
6911 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
6912 constraint for 0..15 for crypto instructions.
6913 (gpc_reg_operand): If VSX allow registers in VSX registers as well
6914 as GPR and floating point registers.
6915 (int_reg_operand): New predicate to match only GPR registers.
6916 (base_reg_operand): New predicate to match base registers.
6917 (quad_int_reg_operand): New predicate to match even GPR registers
6918 for quad memory operations.
6919 (vsx_reg_or_cint_operand): New predicate to allow vector logical
6920 operations in both GPR and VSX registers.
6921 (quad_memory_operand): New predicate for quad memory operations.
6922 (reg_or_indexed_operand): New predicate for direct move support.
6923
6924 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
6925 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
6926 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
6927 (POWERPC_MASKS): Add power8 options.
6928 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
6929 various options.
6930
6931 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6932 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
6933
6934 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
6935 (-mpower8-fusion): New power8 options.
6936 (-mpower8-fusion-sign): Likewise.
6937 (-mpower8-vector): Likewise.
6938 (-mcrypto): Likewise.
6939 (-mdirect-move): Likewise.
6940 (-mquad-memory): Likewise.
6941
6942 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
6943 power8.
6944 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
6945 registers.
70cc1536 6946 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM
6947 (rs6000_debug_vector_unit): Add p8_vector.
6948 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
6949 definitions. Also print fusion state.
6950 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
6951 (rs6000_builtin_mask_calculate): Add power8 builtin support.
6952 (rs6000_option_override_internal): Add support for power8.
6953 (rs6000_common_init_builtins): Add debugging for skipped builtins
6954 if -mdebug=builtin.
6955 (rs6000_adjust_cost): Add power8 support.
6956 (rs6000_issue_rate): Likewise.
6957 (insn_must_be_first_in_group): Likewise.
6958 (insn_must_be_last_in_group): Likewise.
6959 (force_new_group): Likewise.
6960 (rs6000_register_move_cost): Likewise.
6961 (rs6000_opt_masks): Likewise.
6962
6963 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
6964 power8 capable assembler, default to power7 options.
6965 (TARGET_DIRECT_MOVE): Likewise.
6966 (TARGET_CRYPTO): Likewise.
6967 (TARGET_P8_VECTOR): Likewise.
6968 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
6969 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
6970 (VECTOR_MEM_P8_VECTOR_P): Likewise.
6971 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
6972 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
6973 (TARGET_XSCVDPSPN): Likewise.
6974 (TARGET_XSCVSPDPN): Likewsie.
6975 (TARGET_SYNC_HI_QI): Likewise.
6976 (TARGET_SYNC_TI): Likewise.
6977 (MASK_CRYPTO): Likewise.
6978 (MASK_DIRECT_MOVE): Likewise.
6979 (MASK_P8_FUSION): Likewise.
6980 (MASK_P8_VECTOR): Likewise.
70cc1536
UB
6981 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
6982 temporary used by some of the direct move instructions to get two FP
6983 temporary registers does not force creation of a stack frame.
f62511da
MM
6984 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
6985 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
6986 that any VSX registers are tieable, even if they are also an
6987 Altivec vector mode.
6988 (r6000_reg_class_enum): Add wm, wr, wv constraints.
6989 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
6990 (RS6000_BTM_CRYPTO): Likewise.
6991 (RS6000_BTM_COMMON): Likewise.
6992
6993 * config/rs6000/rs6000.md (cpu attribute): Add power8.
6994 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
6995 (enum rs6000_vector): Add power8 vector support.
6996
73a1a707
RR
69972013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6998
6999 PR target/19599
7000 PR target/57340
7001 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7002 (any_sibcall_could_use_r3): this and handle indirect calls.
7003 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7004
d28073d4
BS
70052013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7006
7007 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7008
bee0b10c
RB
70092013-05-22 Richard Biener <rguenther@suse.de>
7010
7011 PR middle-end/57349
7012 * profile.c (branch_prob): Do not split blocks that are
7013 abnormally receiving from ECF_RETURNS_TWICE functions.
7014
98409b51
RS
70152013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7016
7017 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7018 address mode. Move pointer mode initialization to the same place.
7019
c0602ab8
MZ
70202013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7021
7022 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7023 while it has any effect.
7024
4f7a634e
ER
70252013-05-21 Easwaran Raman <eraman@google.com>
7026
7027 PR tree-optimization/57322
9daf6dbb
ER
7028 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7029 UID of the statement added to the BB to be 1.
4f7a634e 7030
a4ce1258
JJ
70312013-05-21 Jakub Jelinek <jakub@redhat.com>
7032
7033 PR tree-optimization/57331
70cc1536
UB
7034 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7035 of conversion from pointer type to integral type with integer.
a4ce1258 7036
1b14621a
MJ
70372013-05-21 Martin Jambor <mjambor@suse.cz>
7038
7039 PR lto/57289
7040 * ipa-prop.c (ipa_read_node_info): Process param_used and
7041 controlled_uses in the same order as when writing.
7042
e60661f0
MG
70432013-05-21 Magnus Granberg <baldrick@free.fr>
7044
7045 PR plugins/56754
ee49aa34 7046 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 7047
4a61be9a
RB
70482013-05-21 Richard Biener <rguenther@suse.de>
7049
7050 PR tree-optimization/57318
7051 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7052 estimate stmts with side-effects as likely eliminated.
7053
c52da5f7
RB
70542013-05-21 Richard Biener <rguenther@suse.de>
7055
7056 PR tree-optimization/57330
7057 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7058 preserve the call stmts fntype.
7059
7ec67e2a
RB
70602013-05-21 Richard Biener <rguenther@suse.de>
7061
7062 PR tree-optimization/57303
7063 * tree-ssa-sink.c (statement_sink_location): Improve killing
7064 stmt detection and properly handle self-assignments.
7065
b112d513
CB
70662013-05-21 Christian Bruel <christian.bruel@st.com>
7067
55805e54
YR
7068 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7069 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB
7070 registers. Set register size out of the PARALLEL loop.
7071
14c2ec26
OE
70722013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7073
7074 PR target/56547
7075 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7076 (*fmasf4, *fmasf4_media): New insns.
7077
da734fa1
RS
70782013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7079
7080 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7081 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7082 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7083 (mips_idiv_insns): Update the comments to say that the returned
7084 instruction counts are in units of BASE_INSN_LENGTH.
7085 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7086 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7087 using 2 rather than 4 as the length of indirect MIPS16 and
7088 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7089 length of a NOP. Don't divide MIPS16 lengths by 2.
7090 (mips16_split_long_branches): Assume a branch is long if the
7091 length is greater than 4 rather than 8.
7092 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7093 rather than multiplying them by 2. Multiply instruction counts
7094 by BASE_INSN_LENGTH rather than 4.
7095 (*jump_mips16, tls_get_tp_mips16_<mode>)
7096 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7097
13719e8b
RS
70982013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7099
7100 * config/mips/mips.md (extended_mips16): Remove branch case.
7101 (length): Remove duplicated extended_mips16 test.
7102
c3850d14
RS
71032013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7104
7105 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7106
8da2e059
RS
71072013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7108
7109 * recog.h (Recog_data): Rename to...
7110 (recog_data_d): ...this.
7111 (recog_data): Update accordingly.
7112 * recog.c (recog_data): Likewise.
7113 * reload.c (save_recog_data): Likewise.
7114 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7115 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7116
2df013f3
JB
71172013-05-17 Julian Brown <julian@codesourcery.com>
7118
7119 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7120 found in a REG_EQUAL note, invalidate it.
7121
371e77e3 71222013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER
7123
7124 * tree-ssa-reassoc.c (find_insert_point): New function.
7125 (insert_stmt_after): Likewise.
7126 (get_def_stmt): Likewise.
7127 (ensure_ops_are_available): Likewise.
7128 (not_dominated_by): Likewise.
7129 (rewrite_expr_tree): Do not move statements beyond what is
7130 necessary. Remove call to swap_ops_for_binary_stmt...
7131 (reassociate_bb): ... and move it here.
7132 (build_and_add_sum): Assign UIDs for new statements.
7133 (linearize_expr): Likewise.
7134 (do_reassoc): Renumber gimple statement UIDs.
7135
e01c7cca
JH
71362013-05-17 Jan Hubicka <jh@suse.cz>
7137
7138 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7139 weakrefs.
7140 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB
7141 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7142 weakrefs.
e01c7cca
JH
7143 (output_weakrefs): Update.
7144
c3272a92
PCC
71452013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7146 Martin Jambor <mjambor@suse.cz>
7147
7148 PR middle-end/57276
7149 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7150 value that corresponds to the given aggval is found in values vector.
7151
11c2aa39
UB
71522013-05-17 Uros Bizjak <ubizjak@gmail.com>
7153
7154 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7155 sse, sse2, sse3, ssse3 and sse4a flags to options.
7156
abecc8c6
DM
71572013-05-17 David Malcolm <dmalcolm@redhat.com>
7158
7159 * gengtype-state.c: (s_expr_writer): New class, to handle
7160 prettifying of output layout of s-expressions.
7161 (state_writer): New class, to write out gtype.state.
7162 (state_written_type_count): Move this variable into member data of
7163 state_writer.
7164 (s_expr_writer::s_expr_writer): New code: constructor for new class
7165 (state_writer::state_writer(): ditto
7166 (s_expr_writer::write_new_line): New function
7167 (s_expr_writer::write_any_indent): ditto
7168 (s_expr_writer::begin_s_expr): ditto
7169 (s_expr_writer::end_s_expr): ditto
7170 (write_state_fileloc): convert to method of state_writer...
7171 (state_writer:: write_state_fileloc): ...and use methods of
7172 s_expr_writer to write indentation into the gtype.state output file
7173 to visually represent the hierarchical structure of the list
7174 structures
7175 (write_state_fields): ditto, renaming to...
7176 (state_writer::write_state_fields)
7177 (write_state_a_string): ditto, renaming to...
7178 (state_writer::write_state_a_string)
7179 (write_state_string_option): ditto, renaming to...
7180 (state_writer::write_state_string_option)
7181 (write_state_type_option): ditto, renaming to...
7182 (state_writer::write_state_type_option)
7183 (write_state_nested_option): ditto, renaming to...
7184 (state_writer::write_state_nested_option)
7185 (write_state_option): ditto, renaming to...
7186 (state_writer::write_state_option)
7187 (write_state_options): ditto, renaming to...
7188 (state_writer::write_state_options)
7189 (write_state_lang_bitmap): ditto, renaming to...
7190 (state_writer::write_state_lang_bitmap)
7191 (write_state_version): ditto, renaming to...
7192 (state_writer::write_state_version)
7193 (write_state_scalar_type): ditto, renaming to...
7194 (state_writer::write_state_scalar_type)
7195 (write_state_string_type): ditto, renaming to...
7196 (state_writer::write_state_string_type)
7197 (write_state_undefined_type): ditto, renaming to...
7198 (state_writer::write_state_undefined_type)
7199 (write_state_struct_union_type): ditto, renaming to...
7200 (state_writer::write_state_struct_union_type)
7201 (write_state_struct_type): ditto, renaming to...
7202 (state_writer::write_state_struct_type)
7203 (write_state_user_struct_type): ditto, renaming to...
7204 (state_writer::write_state_user_struct_type)
7205 (write_state_lang_struct_type): ditto, renaming to...
7206 (state_writer::write_state_lang_struct_type)
7207 (write_state_param_struct_type): ditto, renaming to...
7208 (state_writer::write_state_param_struct_type)
7209 (write_state_pointer_type): ditto, renaming to...
7210 (state_writer::write_state_pointer_type)
7211 (write_state_array_type): ditto, renaming to...
7212 (state_writer::write_state_array_type)
7213 (write_state_gc_used): ditto, renaming to...
7214 (state_writer::write_state_gc_used)
7215 (write_state_common_type_content): ditto, renaming to...
7216 (state_writer::write_state_common_type_content)
7217 (write_state_type): ditto, renaming to...
7218 (state_writer::write_state_type)
7219 (write_state_pair_list): ditto, renaming to...
7220 (state_writer::write_state_pair_list)
7221 (write_state_pair): ditto, renaming to...
7222 (state_writer::write_state_pair)
7223 (write_state_typedefs): ditto, renaming to...
7224 (state_writer::write_state_typedefs)
7225 (write_state_structures): ditto, renaming to...
7226 (state_writer::write_state_structures)
7227 (write_state_param_structs): ditto, renaming to...
7228 (state_writer::write_state_param_structs)
7229 (write_state_variables): ditto, renaming to...
7230 (state_writer::write_state_variables)
7231 (write_state_srcdir): ditto, renaming to...
7232 (state_writer::write_state_srcdir)
7233 (write_state_files_list): ditto, renaming to...
7234 (state_writer::write_state_files_list)
7235 (write_state_languages): ditto, renaming to...
7236 (state_writer::write_state_languages)
7237 (write_state): create a state_writer instance and use it when
7238 writing out the state file
7239
d6545f29
MS
72402013-05-17 Mike Stump <mikestump@comcast.net>
7241
816a3f73 7242 PR rtl-optimization/57304
d6545f29
MS
7243 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7244 accessing DF_REF_REAL_LOC.
7245
38c821cf
JJ
72462013-05-17 Jakub Jelinek <jakub@redhat.com>
7247
7248 PR rtl-optimization/57281
7249 PR rtl-optimization/57300
7250 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7251 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7252 what the other splitter did if the registers are dead.
7253
2a293391
RB
72542013-05-17 Richard Biener <rguenther@suse.de>
7255
7256 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7257 MEM_REF offsets.
7258
17042d2b
JJ
72592013-05-17 Jakub Jelinek <jakub@redhat.com>
7260
7261 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7262 linking.
7263
5b115c1f
MP
72642013-05-17 Marek Polacek <polacek@redhat.com>
7265
11c2aa39
UB
7266 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7267 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 7268
52d84413
JJ
72692013-05-17 Jakub Jelinek <jakub@redhat.com>
7270
68119618
JJ
7271 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7272 vect_external_def oprnd1 with loop_vinfo, try to emit
7273 optional cast, negation and and stmts on the loop preheader
7274 edge instead of into the pattern def seq.
7275
52d84413
JJ
7276 PR tree-optimization/57051
7277 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7278 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7279
17b962bd
NC
72802013-05-16 Nick Clifton <nickc@redhat.com>
7281
7282 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7283 (rl78_is_naked_func): New function.
56aefbf7
UB
7284 (rl78_expand_prologue): Skip prologue generation for naked functions.
7285 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC
7286 * doc/extend.texi (naked): Add RL78 to the list of processors
7287 that supports this attribute.
7288
b1a0f84e
JL
72892013-05-16 Jeff Law <law@redhat.com>
7290
7291 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7292
af0e415b
UB
72932013-05-16 Uros Bizjak <ubizjak@gmail.com>
7294
7295 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7296 cache parameters using detect_caches_amd also for CYRIX,
7297 NSC and TM2 signatures.
7298
19db293a
UB
72992013-05-16 Uros Bizjak <ubizjak@gmail.com>
7300 Dzianis Kahanovich <mahatma@eu.by>
7301
7302 PR target/45359
7303 PR target/46396
7304 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7305 VIA/Centaur processors and determine their cache parameters
7306 using detect_caches_amd.
7307
251a41b9
TJ
73082013-05-16 Teresa Johnson <tejohnson@google.com>
7309
7310 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7311 (rtl_verify_edges): New function.
7312 (rtl_verify_bb_insns): Ditto.
7313 (rtl_verify_bb_pointers): Ditto.
7314 (rtl_verify_bb_insn_chain): Ditto.
7315 (rtl_verify_fallthru): Ditto.
7316 (rtl_verify_bb_layout): Ditto.
7317 (rtl_verify_flow_info_1): Outline checks into new functions.
7318 (rtl_verify_flow_info): Ditto.
7319
f14540b6
SE
73202013-05-16 Steve Ellcey <sellcey@imgtec.com>
7321
7322 * cfghooks.c (copy_bbs): Add update_dominance argument.
7323 * cfghooks.h (copy_bbs): Update prototype.
7324 * tree-cfg.c (gimple_duplicate_sese_region):
7325 Add update_dominance argument.
7326 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7327 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7328 gimple_duplicate_sese_region call.
7329 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7330 Update copy_bbs call.
7331 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7332 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7333
7e9a3abb
JJ
73342013-05-16 Jakub Jelinek <jakub@redhat.com>
7335
7336 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7337 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7338 vect_recog_rotate_pattern.
7339 (vect_recog_rotate_pattern): New function.
7340
427b248d
JM
73412013-05-16 Jason Merrill <jason@redhat.com>
7342
7343 * Makefile.in (LLINKER): New variable.
7344 (mostlyclean): Remove link mutex.
7345 * configure.ac: Handle --enable-link-mutex.
7346 * lock-and-run.sh: New script.
7347
b871e3d2
RR
73482013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7349
7350 PR target/19599
7351 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
7352 for NULL decl.
7353
ce858126
RO
73542013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7355
7356 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
7357
798d3d04
GY
73582013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
7359
7360 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
7361 * config/arm/arm.c (next_consecutive_mem): New function.
7362 (gen_movmem_ldrd_strd): Likewise.
7363 * config/arm/arm.md (movmemqi): Update condition and code.
7364 (unaligned_loaddi, unaligned_storedi): New patterns.
7365
0baddc45
RO
73662013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7367
7368 * config.gcc: Obsolete *-*-solaris2.9*.
7369 * doc/install.texi (Specific, *-*-solaris2*): Document it.
7370
aa9d5bcf
RB
73712013-05-16 Richard Biener <rguenther@suse.de>
7372
7373 * passes.c (init_optimization_passes): Move pass_parallelize_loops
7374 earlier, after GRAPHITE transforms and IV canonicalization.
7375
5a0f4dd3
JJ
73762013-05-16 Jakub Jelinek <jakub@redhat.com>
7377
7378 * omp-low.c (extract_omp_for_data): For collapsed loops,
7379 if at least one of the loops is known at compile time to
7380 iterate zero times, set count to 0.
7381 (expand_omp_regimplify_p): New function.
7382 (expand_omp_for_generic): For collapsed loops, if at least
7383 one of the loops isn't known to iterate at least once,
7384 add runtime check with setting count to 0.
7385 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7386 For unsigned types if it isn't known at compile time that
7387 the loop will iterate at least once, add runtime check to bypass
7388 the whole loop if initial condition isn't true.
7389
e3753785
NS
73902013-05-16 Nathan Sidwell <nathan@codesourcery.com>
7391
7392 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
7393
43bb4dd1
MG
73942013-05-16 Marc Glisse <marc.glisse@inria.fr>
7395
7396 PR middle-end/57286
7397 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
7398 transformations to avoid an infinite loop.
7399
3571dde6
MP
74002013-05-16 Marek Polacek <polacek@redhat.com>
7401
7402 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
7403
b8b3f0ca
LE
74042013-05-15 Leif Ekblad <leif@rdos.net>
7405
7406 * config/i386/i386.c (ix86_decompose_address): Use
7407 DEFAULT_TLS_SEG_REG to access TLS segment register.
7408 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
7409 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
7410 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
7411
e299a383
RS
74122013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
7413
7414 PR target/57260
7415 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
7416 sibling calls to functions that would normally be lazily bound,
7417 unless $gp is call-clobbered.
7418
e7208ea3
UB
74192013-05-15 Uros Bizjak <ubizjak@gmail.com>
7420
19db293a 7421 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 7422 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB
7423 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
7424 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB
7425 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
7426 of TARGET_3DNOW.
7427 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
7428
e7413f3d
AS
74292013-05-15 Andreas Schwab <schwab@suse.de>
7430
7431 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
7432 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
7433 third operand.
7434
ad4db775
TJ
74352013-05-15 Teresa Johnson <tejohnson@google.com>
7436
7437 * loop-unroll.c (report_unroll_peel): Check decision before
7438 emitting unroll/peel message.
7439
af205f67
TJ
74402013-05-15 Teresa Johnson <tejohnson@google.com>
7441
7442 * function.h (has_bb_partition): New rtl_data flag.
7443 (bb_reorder_complete): Ditto.
7444 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
7445 instead of flag_reorder_blocks_and_partition.
7446 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
7447 with some enhancements.
7448 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
7449 * bb-reorder.c (connect_traces): Check for has_bb_partition
7450 instead of flag_reorder_blocks_and_partition.
7451 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
7452 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
7453 verify_hot_cold_block_grouping.
7454 (partition_hot_cold_basic_blocks): Set has_bb_partition.
7455
9adcfa3c
RR
74562013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7457
7458 PR target/19599
7459 * config/arm/predicates.md (call_insn_operand): New predicate.
7460 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
7461 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
7462 if insn is not a tail call.
7463 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
7464 registers.
7465 * config/arm/arm.h (enum reg_class): New caller save register class.
7466 (REG_CLASS_NAMES): Likewise.
7467 (REG_CLASS_CONTENTS): Likewise.
7468 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
7469 without decls.
7470
ccb3ad87
RB
74712013-05-15 Richard Biener <rguenther@suse.de>
7472
7473 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
7474 of MSG_OPTIMIZED_LOCATIONS.
7475 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
7476 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
7477 message.
7478 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
7479 of MSG_OPTIMIZED_LOCATIONS.
7480 (execute_vect_slp): Likewise.
7481 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
7482 (vect_create_cond_for_alias_checks): Likewise.
7483 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
7484 (vect_recog_widen_mult_pattern): Likewise.
7485 (vect_recog_widen_sum_pattern): Likewise.
7486 (vect_recog_over_widening_pattern): Likewise.
7487 (vect_recog_widen_shift_pattern): Likewise.
7488 (vect_recog_vector_vector_shift_pattern): Likewise.
7489 (vect_recog_divmod_pattern): Likewise.
7490 (vect_recog_mixed_size_cond_pattern): Likewise.
7491 (vect_recog_bool_pattern): Likewise.
7492 (vect_pattern_recog_1): Likewise.
7493
48b1474e
MJ
74942013-05-15 Martin Jambor <mjambor@suse.cz>
7495
7496 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
7497 non-functions to builtin_unreachable.
7498 * ipa-inline-transform.c (inline_call): Do not assert estimates were
7499 correct when new direct edges were discovered.
7500
9de04252
MJ
75012013-05-15 Martin Jambor <mjambor@suse.cz>
7502
7503 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
7504 header, print symbol order instead of node uid, print more information
7505 about indirect edge targets.
7506 (ipa_make_edge_direct_to_target): Print symbol order instead of node
7507 uids.
7508 (ipa_make_edge_direct_to_target): Likewise.
7509 (remove_described_reference): Likewise.
7510 (propagate_controlled_uses): Likewise.
7511 (ipa_print_node_params): Also print symbol order.
7512 (ipcp_transform_function): Print symbol order instead of node uids.
7513 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
7514 (cgraph_get_create_real_symbol_node): Likewise.
7515 * ipa-cp.c (print_lattice): Likewise.
7516 (print_all_lattices): Likewise.
7517 (determine_versionability): Likewise.
7518 (initialize_node_lattices): Likewise.
7519 (estimate_local_effects): Likewise.
7520 (update_profiling_info): Likewise.
7521 (create_specialized_node): Likewise.
7522 (perhaps_add_new_callers): Likewise.
7523 (decide_about_value): Likewise.
7524 (decide_whether_version_node): Likewise.
7525 (identify_dead_nodes): Likewise.
7526 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
7527 (dump_inline_summary): Likewise.
7528 (estimate_node_size_and_time): Likewise.
7529 (inline_analyze_function): Likewise.
7530 * ipa-inline.c (report_inline_failed_reason): Likewise.
7531 (want_early_inline_function_p): Likewise.
7532 (edge_badness): Likewise.
7533 (update_edge_key): Likewise.
7534 (inline_small_functions): Likewise. Add dumping of order to two other
7535 dumps.
7536 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
7537 instead of node uids.
7538 (propagate_pure_const): Likewise.
7539 (propagate_pure_const): Likewise.
7540 * ipa-utils.c (dump_cgraph_node_set): Likewise.
7541 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
7542 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
7543 of node uids.
7544 * tree-pretty-print.c (dump_function_header): Likewise.
7545 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
7546 Print symbol order instead of node uids.
7547
1dd03b91
AK
75482013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7549
7550 * config/s390/s390.c (s390_register_move_cost): Don't impose the
7551 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
7552
fdf6a7b9
RB
75532013-05-15 Richard Biener <rguenther@suse.de>
7554
7555 PR tree-optimization/57275
7556 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
7557 return value for fail to do runtime alias checks for gather loads.
7558
2d6e4603
JH
75592013-05-15 Jan Hubicka <jh@suse.cz>
7560
7561 PR lto/57038
7562 PR lto/47375
e7208ea3
UB
7563 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
7564 weakrefs are not external.
7565 (lto_symtab_merge_decls): Fix thinko when dealing with
7566 non-lto_symtab decls.
2d6e4603
JH
7567 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
7568 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
7569 * varpool.c (dump_varpool_node): Dump more flags.
7570
83f44b39
GG
75712013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7572
7573 * config/i386/i386.c (processor_alias_table): Add instruction
7574 FSGSBASE for AMD bdver3 architecture.
7575
2e55d062
JJ
75762013-05-14 Jakub Jelinek <jakub@redhat.com>
7577
7578 * tree.c (warn_deprecated_use): Print file:line using locus color.
7579 * diagnostic.c (diagnostic_report_current_module): Print file:line
7580 and file:line:column using locus color.
7581
0bfdb81e
MS
75822013-05-14 Mike Stump <mikestump@comcast.net>
7583
7584 * gdbinit.in: Add __null.
7585
a508ef22
MS
75862013-05-14 Mike Stump <mikestump@comcast.net>
7587
7588 * recog.h: Rename struct recog_data to Recog_data.
7589 * recog.c: Likewise.
7590 * reload.c (can_reload_into): Likewise.
7591 * config/picochip/picochip.c: Likewise.
7592
e7180acb
MS
75932013-05-14 Mike Stump <mikestump@comcast.net>
7594
7595 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
7596
61216c26
SB
75972013-05-14 Steven Bosscher <steven@gcc.gnu.org>
7598
fd6beed4
SB
7599 * resource.h (struct resources): Remove unch_memory member.
7600 (CLEAR_RESOURCE): Don't clear unch_memory.
7601 * resource.c (mark_referenced_resources): Don't set it.
7602 (mark_set_resources): Likewise.
7603 (mark_target_live_regs): Don't clear it.
7604 (init_resource_info): Likewise.
7605 * reorg.c (resource_conflicts_p): Don't compare it.
7606 (redundant_insn): Don't set it.
7607
61216c26
SB
7608 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
7609 Remove prototypes.
7610 * emit-rtl.c (next_label): Remove unused function.
7611 (skip_consecutive_labels, link_cc0_insns): Move to ...
7612 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
7613 only place where these functions are used, and make them static.
7614
418d1b87
MG
76152013-05-14 Marc Glisse <marc.glisse@inria.fr>
7616
7617 * fold-const.c (fold_negate_expr): Handle vectors.
7618 (fold_truth_not_expr): Make it static.
7619 (fold_invert_truthvalue): New static function.
7620 (invert_truthvalue_loc): Handle vectors. Do not call
7621 fold_truth_not_expr directly.
7622 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
7623 <TRUTH_NOT_EXPR>: Do not cast to boolean.
7624 (fold_comparison): Handle vector constants.
7625 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
7626 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
7627 * tree.h (fold_truth_not_expr): Remove declaration.
7628
fc21784d
JG
76292013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
7630
7631 * config/aarch64/aarch64-simd.md
7632 (aarch64_vcond_internal<mode>): Rename to...
7633 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
7634 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
7635 float modes. Clarify all iterator modes.
7636 (vcond<mode><mode>): Use new name for vcond expanders.
7637 (vcond<v_cmp_result><mode>): Likewise.
7638 (vcondu<mode><mode>: Likewise.
7639 * config/aarch64/iterators.md (VDQF_COND): New.
7640
d4c52634
MG
76412013-05-14 Marc Glisse <marc.glisse@inria.fr>
7642
7643 PR bootstrap/57266
7644 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
7645 variable for the shift amount. Check that we shift by non-negative
7646 amounts.
7647
2b261262
CLT
76482013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
7649
7650 PR target/42017
7651 * config/arm/arm.h (EPILOGUE_USES): Only return true
7652 for LR_REGNUM after epilogue_completed.
7653
6e022d7b
JR
76542013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
7655
7656 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 7657 is error_mark_node.
6e022d7b 7658
56cf7859
RO
76592013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7660
7661 PR target/57261
7662 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
7663 and Solaris 11+/x86 with gld.
7664 * configure: Regenerate.
7665
75776c6d
JJ
76662013-05-14 Jakub Jelinek <jakub@redhat.com>
7667
7668 * expmed.c (expand_shift_1): Canonicalize rotates by
7669 constant bitsize / 2 to bitsize - 1.
7f998021 7670 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ
7671 case ROTATERT>: Likewise.
7672
7673 Revert:
7674 2013-05-10 Jakub Jelinek <jakub@redhat.com>
7675
7676 * config/i386/i386.md (rotateinv): New code attr.
7677 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7678 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7679 roll $31, %eax, etc.
7680
df35498a
RB
76812013-05-14 Richard Biener <rguenther@suse.de>
7682
7683 PR middle-end/57235
7684 * tree-eh.c (sink_clobbers): Give up for successors with
7685 multiple predecessors and no virtual uses.
7686
cc6e7ece
EB
76872013-05-14 Eric Botcazou <ebotcazou@adacore.com>
7688
7689 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
7690 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
7691
f778dd4d
JJ
76922013-05-14 Jakub Jelinek <jakub@redhat.com>
7693
7694 PR middle-end/57251
7695 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
7696 the case when both op0 and op1 have VOIDmode.
7697
bad4df9b
KP
76982013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7699
7700 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
7701 in multiply-accumulate mode.
7702
56f3e9ac
GW
77032013-05-13 Guozhi Wei <carrot@google.com>
7704
7705 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
7706
a3d7ab92
KT
77072013-05-13 Kai Tietz <ktietz@redhat.com>
7708
7709 PR target/56975
7710 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 7711 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 7712 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 7713 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT
7714 (ix86_expand_prologue): Likewise.
7715 (ix86_expand_split_stack_prologue): Likewise.
7716 (legitimate_pic_address_disp_p): Likewise.
7717 (legitimize_pic_address): Likewise.
7718 (legitimize_tls_address): Likewise.
7719 (legitimize_pe_coff_symbol): Likewise.
7720 (output_pic_addr_const): Likewise.
7721 (construct_plt_address): Likewise.
7722 (ix86_expand_call): Likewise.
7723 (x86_output_mi_thunk): Likewise.
7724 (x86_function_profiler): Likewise.
7725
c59b7e28
SN
77262013-05-13 Sofiane Naci <sofiane.naci@arm.com>
7727
7728 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
7729 similar switch cases.
7730 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
7731 (aarch64_simd_mov_to_<mode>low): Delete.
7732 (aarch64_simd_mov_to_<mode>high): Delete.
7733 (move_lo_quad_<mode>): Add w<-r alternative.
7734 (aarch64_simd_move_hi_quad_<mode>): Likewise.
7735 (aarch64_simd_mov_from_*): Update type attribute.
7736 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
7737 statement.
7738
5f28524a
JH
77392013-05-13 Jan Hubicka <jh@suse.cz>
7740
7741 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
7742 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB
7743 ix86_expand_epilogue, emit_i387_cw_initialization,
7744 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
7745 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 7746
ae6fa899
JJ
77472013-05-13 Jakub Jelinek <jakub@redhat.com>
7748
7749 PR tree-optimization/45216
7750 PR tree-optimization/57157
7751 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
7752 the (-Y) & (B - 1) variant if OP is |.
7753 * expmed.c (expand_shift_1): For rotations by const0_rtx just
7754 return shifted. Use (-op1) & (prec - 1) as other_amount
7755 instead of prec - op1.
7756
4502fe8d
MJ
77572013-05-13 Martin Jambor <mjambor@suse.cz>
7758
7759 PR middle-end/42371
7760 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
7761 (ipa_constant_data): New type.
7762 (ipa_jump_func): Use ipa_constant_data to hold information about
7763 constant jump functions.
7764 (ipa_get_jf_constant): Adjust to jump function type changes.
7765 (ipa_get_jf_constant_rdesc): New function.
7766 (ipa_param_descriptor): New field controlled_uses.
7767 (ipa_get_controlled_uses): New function.
7768 (ipa_set_controlled_uses): Likewise.
7769 * ipa-ref.h (ipa_find_reference): Declare.
7770 * ipa-prop.c (ipa_cst_ref_desc): New type.
7771 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
7772 changes.
7773 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
7774 New parameter cs. Adjust all callers.
7775 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
7776 (remove_described_reference): New function.
7777 (jfunc_rdesc_usable): Likewise.
7778 (try_make_edge_direct_simple_call): Decrement controlled use count,
7779 attempt to remove reference if it hits zero.
7780 (combine_controlled_uses_counters): New function.
7781 (propagate_controlled_uses): Likewise.
7782 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
7783 (ipa_edge_duplication_hook): Duplicate reference descriptions.
7784 (ipa_print_node_params): Print described use counter.
7785 (ipa_write_jump_function): Adjust to jump function type changes.
7786 (ipa_read_jump_function): New parameter CS, pass it to
7787 ipa_set_jf_constant. Adjust caller.
7788 (ipa_write_node_info): Stream controlled use count
7789 (ipa_read_node_info): Likewise.
7790 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
7791 asserting.
7792 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
7793 count. Remove cloning-added reference if it reaches zero.
7794 * ipa-ref.c (ipa_find_reference): New function.
7795
0864bfc2
GG
77962013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7797
e7208ea3 7798 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG
7799 alignment values for AMD BD and BT architectures.
7800
640bfeb2
MG
78012013-05-13 Marc Glisse <marc.glisse@inria.fr>
7802
7803 * tree-vect-generic.c (uniform_vector_p): Move ...
7804 * tree.c (uniform_vector_p): ... here.
7805 * tree.h (uniform_vector_p): Declare it.
7806 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
7807 into a scalar.
7808
3a60f32b
JJ
78092013-05-13 Jakub Jelinek <jakub@redhat.com>
7810
198fe1bf
JJ
7811 PR tree-optimization/57230
7812 * tree-ssa-strlen.c (handle_char_store): Record length for
7813 array store from STRING_CST.
7814
3a60f32b
JJ
7815 PR tree-optimization/57230
7816 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
7817 check.
7818
566be57c
JR
78192013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
7820
7821 * config/epiphany/epiphany.c (epiphany_init): Check size of
7822 NUM_MODES_FOR_MODE_SWITCHING.
7823 (epiphany_expand_prologue):
7824 Remove CONFIG_REGNUM initial value handling code.
7825 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
7826 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 7827 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR
7828 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
7829 Don't return 1 for FP_MODE_NONE.
7830 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
7831 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
7832 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
7833 * config/epiphany/epiphany.md (save_config): New pattern.
7834
0f2c2331
UB
78352013-05-12 Uros Bizjak <ubizjak@gmail.com>
7836
7837 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
7838
5b3f93c7
UB
78392013-05-10 Uros Bizjak <ubizjak@gmail.com>
7840
7841 * config/i386/i386.md (memory): Handle sseishft1.
7842 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
7843 (*vec_extractv2di_1): Ditto.
7844
1f873f0e
VM
78452013-05-10 Vladimir Makarov <vmakarov@redhat.com>
7846
7847 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
7848 saved registers.
7849
341427fa 78502013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH
7851
7852 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
7853 Add mthumb/march=armv7-a multilib.
7854 Add mthumb/march=armv7-r multilib.
7855 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
7856
9e69bdde
RC
78572013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
7858
7859 * config/v850/t-rtems: Add more multilibs.
7860
9ff09a22
RB
78612013-05-10 Richard Biener <rguenther@suse.de>
7862
7863 PR tree-optimization/57214
7864 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
7865 not propagate from SSA names that occur in abnormal PHI nodes.
7866
a5e0cd1d
MG
78672013-05-10 Marc Glisse <marc.glisse@inria.fr>
7868
7869 * stor-layout.c (element_precision): New function.
7870 * machmode.h (element_precision): Declare it.
7871 * tree.c (build_minus_one_cst): New function.
7872 (element_precision): Likewise.
7873 * tree.h (build_minus_one_cst): Declare new function.
7874 (element_precision): Likewise.
7875 * fold-const.c (operand_equal_p): Use element_precision.
7876 (fold_binary_loc): Handle vector types.
7877 * convert.c (convert_to_integer): Use element_precision.
7878 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
7879 separately.
7880
cb2558bc
RS
78812013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
7882
7883 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
7884 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
7885 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
7886 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7887 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
7888 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
7889 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
7890 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7891 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
7892 (Uuw8): New constraints.
7893 (Usb4): Move into alphabetical order.
7894 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
7895 (sd8_operand, ub8_operand, uw8_operand): New predicates.
7896 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
7897 previously unnamed patterns.
7898 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
7899 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
7900 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
7901 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
7902 of set_attr_alternative/if_then_else. Use extended_mips16 instead
7903 of specific lengths.
7904
cb3b8d33
JJ
79052013-05-10 Jakub Jelinek <jakub@redhat.com>
7906
6f93c008
JJ
7907 * config/i386/i386.md (rotateinv): New code attr.
7908 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7909 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7910 roll $31, %eax, etc.
7911
cb3b8d33
JJ
7912 PR tree-optimization/45216
7913 PR tree-optimization/57157
7914 * tree-ssa-forwprop.c (simplify_rotate): New function.
7915 (ssa_forward_propagate_and_combine): Call it.
7916
afb119be
RB
79172013-05-10 Richard Biener <rguenther@suse.de>
7918
7919 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
7920 disable peeling when we version for aliasing.
7921 (vector_alignment_reachable_p): Honor explicit user alignment.
7922 (vect_supportable_dr_alignment): Likewise.
7923 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
7924 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
7925 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
7926 then peeling to arrange for the cost-model check to come first.
7927
01ae4861
AM
79282013-05-10 Alan Modra <amodra@gmail.com>
7929
7930 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
7931 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
7932 * configure: Regenerate.
7933
ebc9a431
AM
79342013-05-10 Alan Modra <amodra@gmail.com>
7935
7936 PR target/55033
7937 * varasm.c (default_elf_select_section): Move !DECL_P check..
7938 (get_named_section): ..to here before calling get_section_name.
7939 Adjust assertion.
7940 (default_section_type_flags): Add DECL_P check.
7941 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
7942 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
7943
d4bca93c
JR
79442013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
7945
7946 * config/epiphany/epiphany.c (epiphany_expand_prologue):
7947 When using gen_stack_adjust_str with a register offset, add a
7948 REG_FRAME_RELATED_EXPR note.
7949
60ca9a65
UB
79502013-05-09 Uros Bizjak <ubizjak@gmail.com>
7951
7952 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
7953 (*vec_extractv4si_zext_mem): Ditto.
7954 (*vec_extractv2di): Add 0->x and x->x alternatives.
7955 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
7956 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
7957
a3409c02
JM
79582013-05-09 Jason Merrill <jason@redhat.com>
7959
0138d6b2
JM
7960 N3639 C++1y VLA support
7961 * gimplify.c (gimplify_vla_decl): Don't touch an existing
7962 DECL_VALUE_EXPR.
7963
a3409c02
JM
7964 * tree.c (build_constructor_va): New.
7965 * tree.h: Declare it.
7966
66e6b990
MJ
79672013-05-09 Martin Jambor <mjambor@suse.cz>
7968
7969 PR lto/57084
7970 * gimple-fold.c (canonicalize_constructor_val): Call
7971 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
7972
64cfa6c0
JH
79732013-05-09 Jan Hubicka <jh@suse.cz>
7974 Richard Biener <rguenther@suse.de>
7975
7976 PR lto/54095
7977 * symtab.c (symtab_make_decl_local): Do not add private names.
7978
c3167b00
JH
79792013-05-09 Jan Hubicka <jh@suse.cz>
7980
7981 PR lto/54095
7982 * symtab.c (insert_to_assembler_name_hash): Handle clones.
7983 (unlink_from_assembler_name_hash): Likewise.
7984 (symtab_prevail_in_asm_name_hash, symtab_register_node,
7985 symtab_unregister_node, symtab_initialize_asm_name_hash,
7986 change_decl_assembler_name): Update.
7987
12dc6974
SN
79882013-05-09 Sofiane Naci <sofiane.naci@arm.com>
7989
7990 * config/aarch64/aarch64.md: New movtf split.
7991 (*movtf_aarch64): Update.
7992 (aarch64_movdi_tilow): Handle TF modes and rename to
7993 aarch64_movdi_<mode>low.
7994 (aarch64_movdi_tihigh): Handle TF modes and rename to
7995 aarch64_movdi_<mode>high
7996 (aarch64_movtihigh_di): Handle TF modes and rename to
7997 aarch64_mov<mode>high_di
7998 (aarch64_movtilow_di): Handle TF modes and rename to
7999 aarch64_mov<mode>low_di
8000 (aarch64_movtilow_tilow): Remove spurious whitespace.
8001 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8002 splits.
8003 (aarch64_print_operand): Update.
8004
227eb343
AM
80052013-05-09 Alan Modra <amodra@gmail.com>
8006
8007 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8008 powerpc64le.
8009 * configure: Regenerate.
8010
0b013847
UB
80112013-05-08 Uros Bizjak <ubizjak@gmail.com>
8012
8013 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8014 splitter preparation statements.
8015 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8016 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8017 adjust_address_nv.
8018
1dc3d6e9
BS
80192013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8020
8021 * gimple-ssa-strength-reduction.c (count_candidates): Change
8022 return value to int.
8023 (analyze_candidates_and_replace): Change type of length to int.
8024
e61e7d28
UB
80252013-05-08 Uros Bizjak <ubizjak@gmail.com>
8026
8027 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8028 (*vec_extract<mode>): Use VI12_128 mode iterator.
8029 (*vec_extract<mode>_mem): Ditto.
8030 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8031 attribute.
8032
4fbfcf44
DN
80332013-05-08 Diego Novillo <dnovillo@google.com>
8034
8035 PR bootstrap/54659
8036
8037 Revert:
4fbfcf44
DN
8038 2012-08-17 Diego Novillo <dnovillo@google.com>
8039
e61e7d28
UB
8040 PR bootstrap/54281
8041 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8042 * config.in: Regenerate.
8043 * configure: Regenerate.
8044 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 8045
702d8703
JH
80462013-05-08 Jan Hubicka <jh@suse.cz>
8047
8048 PR lto/54095
8049 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8050 * cgraph.h (symtab_node_base): Add unique_name.
8051 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8052 input_overwrite_node, input_varpool_node): Stream unique_name.
8053 * cgraphclones.c (cgraph_create_virtual_clone,
8054 cgraph_function_versioning): Set unique_name.
8055 * ipa.c (function_and_variable_visibility): Set unique_name.
8056
8b28cf47
BS
80572013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8058
8059 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8060 (alloc_cand_and_find_basis): Restrict conditional candidate
8061 processing to CAND_MULTs.
8062
e86074fd
JH
80632013-05-08 Jan Hubicka <jh@suse.cz>
8064
8065 PR lto/54095
8066 lto-symtab.c (lto_symtab_symbol_p): New function.
8067 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8068 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8069 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8070 Skip static symbols.
8071
44398cbe
PC
80722013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8073
8074 PR tree-optimization/57200
8075 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8076 Only call inform if the preceding warning_at returns true.
8077
f6bc1c4a
HS
80782013-05-07 Han Shen <shenhan@google.com>
8079
8080 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8081 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8082 * common.opt (fstack-protector-strong): New option.
8083 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8084 * doc/invoke.texi (Optimization Options): Document
8085 "-fstack-protector-strong".
8086 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8087
4ffecb1f
SB
80882013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8089
8090 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8091
5a107a0f
VM
80922013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8093
8094 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8095 it is not lhs of the insn.
8096 (setup_reg_equiv): Remove insn having equiv memory which it is not
8097 lhs of the insn.
8098 * lra-constraints.c (process_address): Try to improve generation
8099 code for address base + disp.
8100 (lra_constraints): Make correct the code for checking insn setting
8101 up backward equivalence. Remove insn only if it is in the init
8102 insn list.
8103 * lra-eliminations.c (update_reg_eliminate): Change return value.
8104 (lra_eliminate): Use the result.
8105
3f5783ea
UB
81062013-05-07 Uros Bizjak <ubizjak@gmail.com>
8107
8108 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8109 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8110 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8111 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8112 register target operands.
8113 (*vec_extractv8hi_sse2): New pattern.
8114 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8115 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8116 (*vec_extract<mode>_mem): New insn and split pattern.
8117
8a5800b8
CL
81182013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8119
8120 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8121 (TARGET_ASAN_SHADOW_OFFSET): Define.
8122 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8123 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8124
7bf55a70
BS
81252013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8126
8127 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8128 (incr_vec_index): Return -1 if increment not found.
8129 (create_add_on_incoming_edge): Assert if increment not found.
8130 (record_increment): Limit number of increments recorded.
8131 (all_phi_incrs_profitable): Return false if an increment not found.
8132 (replace_profitable_candidates): Don't process increments that were
8133 not recorded.
8134 (analyze_candidates_and_replace): Limit size of incr_vec.
8135
3f8825c0
RB
81362013-05-07 Richard Biener <rguenther@suse.de>
8137
8138 * calls.c (special_function_p): setjmp-like functions are leaf.
8139 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8140 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8141
fd4842cd
SN
81422013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8143
8144 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8145 (aarch64_simd_mov<mode>): New expander.
8146 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8147 (aarch64_simd_mov_to_<mode>high): Likewise.
8148 (aarch64_simd_mov_from_<mode>low): Likewise.
8149 (aarch64_simd_mov_from_<mode>high): Likewise.
8150 (aarch64_dup_lane<mode>): Update.
8151 (aarch64_dup_lanedi): New instruction pattern.
8152 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8153 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8154
a7a7d10e
BS
81552013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8156
8157 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8158 (replace_mult_candidate): Remove unnecessary argument; remove
8159 unnecessary parameter from call to introduce_cast_before_cand.
8160 (replace_unconditional_candidate): Remove unnecessary parameter
8161 from call to replace_mult_candidate.
8162 (replace_conditional_candidate): Likewise.
8163 (insert_initializers): Use make_temp_ssa_name.
8164 (introduce_cast_before_cand): Remove unnecessary argument; use
8165 make_temp_ssa_name.
8166 (replace_one_candidate): Remove unnecessary argument; remove
8167 unnecessary parameter from calls to introduce_cast_before_cand.
8168 (replace_profitable_candidates): Remove unnecessary parameters
8169 from calls to replace_one_candidate.
8170
29105868
BS
81712013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8172
8173 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8174 phi def as possibly hiding a basis for a CAND_ADD whose operands
8175 have been commuted in the analysis.
8176 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8177
4095f9fa
N
81782013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8179
8180 * config/aarch64/aarch64.md
8181 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8182 shift value between 0-4.
8183
39e843e8
RB
81842013-05-07 Richard Biener <rguenther@suse.de>
8185
8186 * double-int.h (rshift): New overload.
8187 * double-int.c (rshift): New function.
8188 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8189 (create_reference_ops_from_ref): Remove.
8190 (vn_reference_insert): Use shared ops for constructing the
8191 reference and copy it.
8192
0a1a83cb
RB
81932013-05-07 Richard Biener <rguenther@suse.de>
8194
8195 PR middle-end/57190
8196 * tree-eh.c (sink_clobbers): Properly propagate
8197 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8198
43303d6f
JJ
81992013-05-07 Jakub Jelinek <jakub@redhat.com>
8200
ba7e83f8
JJ
8201 PR tree-optimization/57149
8202 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8203 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8204 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8205 uninit_undefined_value_p instead of ssa_undefined_value_p.
8206
43303d6f
JJ
8207 PR debug/57184
8208 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8209 for modifier == EXPAND_INITIALIZER.
8210
14523c25
AB
82112013-05-07 Anton Blanchard <anton@samba.org>
8212
8213 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8214 for powerpc64 little endian.
8215 * configure: Regenerate.
8216
cb7c8be9
GS
82172013-05-06 Graham Stott <grahams@btinternet.com>
8218
8219 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8220 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8221 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8222 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8223
82242013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS
8225
8226 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8227 codes which allow non-lvalues.
8228
9a0ee7b0
MG
82292013-05-06 Marc Glisse <marc.glisse@inria.fr>
8230
8231 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8232 components are all 1s.
8233 (integer_minus_onep): New function.
8234 * tree.h (integer_minus_onep): Declare it.
8235 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8236 integer_minus_onep instead of integer_all_onesp.
8237
f2c17ea9
OE
82382013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8239
8240 PR target/52933
8241 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8242 variations of these patterns.
8243
f75e6a51
UB
82442013-05-06 Uros Bizjak <ubizjak@gmail.com>
8245
8246 * config/i386/i386.md (isa): Add x64_sse4 member.
8247 (enabled): Handle x64_sse4.
8248 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8249 instruction for 64bit SSE4_1 targets. Update insn attributes.
8250 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8251 instruction for SSE4_1 targets. Update insn attributes.
8252 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8253 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8254 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8255 const_1 selector.
8256 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8257 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8258 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8259
a986d468
OE
82602013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8261
8262 PR target/57108
8263 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8264
382522cb
MK
82652013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8266
8267 * final.c (do_assembler_dialects): Don't handle curly braces and
8268 vertical bar escaped by % as dialect delimiters.
8269 (output_asm_insn): Print curly braces and vertical bar if escaped
8270 by % and ASSEMBLER_DIALECT defined.
8271 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8272 * doc/tm.texi: Regenerated.
8273
fb0d5c60
SB
82742013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8275
fb0d5c60
SB
8276 * config/mips/mips.c: Include tree-pass.h.
8277 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8278 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8279 (pass_mips_machine_reorg2): New machine specific pass.
8280 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8281 (mips_option_override): Register the new pass.
8282 * rtl.h (cleanup_barriers): Remove prototype.
8283 (dbr_schedule): Likewise.
8284 * jump.c (cleanup_barriers): Make static.
8285 * reorg.c (dbr_schedule): Likewise.
8286
aa06a978
RB
82872013-05-06 Richard Biener <rguenther@suse.de>
8288
8289 PR tree-optimization/57185
8290 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8291 SSA names properly.
8292
0b953bec
UB
82932013-05-06 Uros Bizjak <ubizjak@gmail.com>
8294
8295 PR target/57106
8296 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8297 between "first_arg" and "insn", not "last" and "insn".
8298
28708525
WS
82992013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8300
8301 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8302 (find_candidates_in_block): Re-enable slsr_process_phi.
8303 (create_phi_basis): Fix double counting of candidate adjustment.
8304
0107dca2
RB
83052013-05-06 Richard Biener <rguenther@suse.de>
8306
8307 PR middle-end/57147
8308 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8309 the edge is also fallthru, preserve it and just clear the
8310 abnormal flag.
8311 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8312 also complex, preserve that and just clear the fallthru flag.
8313 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8314 update virtual operands.
8315
470d4d13
AM
83162013-05-06 Alan Modra <amodra@gmail.com>
8317
8318 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8319 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8320 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8321 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8322 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8323 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8324
8f1dbf8d
AM
83252013-05-06 Alan Modra <amodra@gmail.com>
8326
8327 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8328 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8329 (DEFAULT_ASM_ENDIAN): Define.
8330 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8331 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8332 Update -K PIC clause from sysv4.h.
8333 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8334 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8335
54890767
AM
83362013-05-06 Alan Modra <amodra@gmail.com>
8337
8338 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8339 twice for little-endian.
8340 (ashrdi3_no_power, ashrdi3): Support little-endian.
8341
2353515d
OE
83422013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8343
8344 PR target/55303
8345 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
8346 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
8347 related expanders.
8348 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
8349 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
8350 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
8351 New predicates.
8352
33e67557
SB
83532013-05-05 Steven Bosscher <steven@gcc.gnu.org>
8354 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
8355
8356 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
8357 * config/pa/pa.opt: Make mbig-switch a no-op.
8358 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
8359 (CASE_VECTOR_MODE): Always return SImode.
8360 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
8361 for the !TARGET_BIG_SWITCH case.
8362 * config/pa/pa-linux.h: Likewise.
8363 * config/pa/pa-openbsd.h: Likewise.
8364 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
8365 * config/pa/pa.md (short_jump): Remove define_insn.
8366 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
8367 (casesi0): Remove define_insn.
8368 (type): Remove btable_branch.
8369 (pa_combine_type): Likewise.
8370 (in_nullified_branch_delay): Likewise.
8371 (in_call_delay): Likewise.
8372 (define_delay): Likewise.
8373 (define_insn_reservation "Z3"): Likewise.
8374 (define_insn_reservation "Z4"): Likewise.
8375 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
8376 (pa_adjust_insn_length): Remove adjustment for btable branches.
8377 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
8378 and mno-big-switch
8379
3095685e
UB
83802013-05-05 Uros Bizjak <ubizjak@gmail.com>
8381
8382 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8383 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
8384 Add m->r,x alternatives.
8385 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
8386 splitters using SWI48x mode iterator.
8387 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
8388 TARGET_64BIT. Add m->x alternative.
8389 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
8390 Add o->x alternative. Enable for TARGET_SSE.
8391 (sse_storeq): Remove expander.
8392 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
8393 with memory input operand.
8394 (*vec_extractv2di_1 splitter): New.
8395 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
8396 * config/i386/i386.md (ssevecmodelower): New mode attribute.
8397
4b36ae28
SB
83982013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8399
8400 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
8401 (INT_LOWPART): Delete.
8402 (extract_MB): Adjust.
8403 (extract_ME): Adjust.
8404 (print_operand): Adjust.
8405
da226db2
SB
84062013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8407
8408 * config/rs6000/predicates.md (reg_or_add_cint_operand,
8409 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8410 (reg_or_logical_cint_operand, easy_fp_constant,
8411 logical_const_operand): Delete "CONST_DOUBLE" case.
8412 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
8413 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 8414 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB
8415 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
8416 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
8417 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
8418 test.
8419 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
8420 CONST_DOUBLE DImode/VOIDmode case.
8421 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
8422 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8423 CONST_DOUBLE VOIDmode case.
8424 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8425 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
8426 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
8427 Delete CONST_DOUBLE case.
8428 (splitters for mov FMOVE64 const_double): Delete
8429 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8430 "HOST_BITS_PER_WIDE_INT >= 64" test.
8431 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
8432 case.
8433 (mov DI const_double): Delete.
8434
40de22d6
JJ
84352013-05-04 Jakub Jelinek <jakub@redhat.com>
8436
8437 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
8438 on op shows all bits zero in mode of a lowpart subreg, return zero.
8439
5ec6aff2
MM
84402013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8441
8442 PR target/57150
8443 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
8444 to save TFmode registers and DImode to save TImode registers for
8445 caller save operations.
8446 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
8447 mark being partially clobbered since they only use the first
8448 double word.
8449
8450 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
8451 and TDmode only use the upper 64-bits of each VSX register.
8452
2cefad90
BS
84532013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8454
8455 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
8456 (find_candidates_in_block): Disable slsr_process_phi.
8457
d6d7eee1
GW
84582013-05-03 Guozhi Wei <carrot@google.com>
8459
8460 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB
8461 constructor to ...
8462 (build_init_ctor): ... here.
d6d7eee1 8463
9b92d12b
BS
84642013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8465
8466 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
8467 (slsr_cand_d): Redefine def_phi.
8468 (stride_status, phi_adjust_status, count_phis_status): New enums.
8469 (find_phi_def): New.
8470 (find_basis_for_base_expr): New.
8471 (find_basis_for_candidate): Handle hidden bases.
8472 (alloc_cand_and_find_basis): Handle phi candidates.
8473 (slsr_process_phi): New.
8474 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
8475 (create_mul_imm_cand): Likewise.
8476 (create_add_ssa_cand): Exclude phi base candidates.
8477 (create_add_imm_cand): Likewise.
8478 (slsr_process_cast): Likewise.
8479 (slsr_process_copy): Likewise.
8480 (find_candidates_in_block): Handle phi candidates.
8481 (dump_candidate): Likewise.
8482 (unconditional_cands): Delete.
8483 (unconditional_cands_with_known_stride_p): Delete.
8484 (phi_dependent_cand_p): New.
8485 (cand_increment): Handle phi-dependent candidates.
8486 (replace_dependent): Delete.
8487 (replace_mult_candidate): New.
8488 (replace_unconditional_candidate): New.
8489 (incr_vec_index): Move to avoid forward reference.
8490 (create_add_on_incoming_edge): New.
8491 (create_phi_basis): New.
8492 (replace_dependents): Delete.
8493 (replace_conditional_candidate): New.
8494 (phi_add_costs): New.
8495 (replace_uncond_cands_and_profitable_phis): New.
8496 (record_increment): Handle phi adjustments.
8497 (record_phi_increments): New.
8498 (record_increments): Handle phi adjustments.
8499 (phi_incr_cost): New.
8500 (lowest_cost_path): Handle phis.
8501 (total_savings): Likewise.
8502 (analyze_increments): Likewise.
8503 (ncd_with_phi): New.
8504 (ncd_of_cand_and_phis): New.
8505 (nearest_common_dominator_for_cands): Handle phi increments.
8506 (all_phi_incrs_profitable): New.
8507 (replace_profitable_candidates): Handle phi-dependent candidates.
8508 (analyze_candidates_and_replace): Likewise.
8509
68f073d4
TJ
85102013-05-03 Teresa Johnson <tejohnson@google.com>
8511
8512 PR bootstrap/57154
8513 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
8514 do not exceed REG_BR_PROB_BASE.
8515
a4ee7cb9
JL
85162013-05-03 Jeff Law <law@redhat.com>
8517
ade67f70 8518 PR tree-optimization/57144
a4ee7cb9
JL
8519 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
8520 operand of the condition will bit into the new type when eliminating
8521 a cast feeding a condition.
8522
47954c4e
JJ
85232013-05-03 Jakub Jelinek <jakub@redhat.com>
8524
8525 PR rtl-optimization/57130
3095685e
UB
8526 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
8527 of COMPARE as in_code to the recursive call if needed.
47954c4e 8528
3c21604f
UB
85292013-05-03 Uros Bizjak <ubizjak@gmail.com>
8530
8531 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
8532 (enabled): Handle new members.
8533 * config/i386/sse.md (*vec_concatv2si): Merge from
8534 *vec_concatv2si_sse2 and vec_concatv2si_sse.
8535 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
8536
95c03b36
JR
85372013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
8538
8539 PR tree-optimization/57027
8540 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
8541 for fnms opportunity, check we got the prerequisite kind
8542 of tree / gimple before using accessor functions.
8543
07bfc9ec
RB
85442013-05-03 Richard Biener <rguenther@suse.de>
8545
8546 * double-int.h (lshift): New overload without precision
8547 and arith argument.
8548 (operator *=, operator +=, operator -=): Move ...
8549 * double-int.c (operator *=, operator +=, operator -=): ... here
8550 and implement more efficiently.
8551 (mul_double_with_sign): Remove.
8552 (lshift_double): Adjust to take unsinged shift argument, push
8553 dispatching code to callers.
8554 (mul_double_wide_with_sign): Add early out for callers that
8555 are not interested in high parts or overflow.
8556 (lshift): New function.
8557 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
8558 dispatch code here.
8559 (lrotate, rrotate): Use logical shifts.
8560 * expr.c (get_inner_reference): Use lshift.
8561 * fixed-value.c (do_fixed_divide): Likewise.
8562 * tree-dfa.c (get_ref_base_and_extent): Likewise.
8563 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
8564 (indirect_refs_may_alias_p): Likewise.
8565 (stmt_kills_ref_p_1): Likewise.
8566
7769bb64
VP
85672013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8568
8569 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
8570
1d0c8e5c
VP
85712013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8572
8573 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
8574 scalar form of FABD instruction.
8575
f15feaf9
VM
85762013-05-02 Vladimir Makarov <vmakarov@redhat.com>
8577
8578 * lra-constraints.c (process_alt_operands): Add checking alt
8579 number to choose the best alternative.
8580
d90e76d4
RB
85812013-05-02 Richard Biener <rguenther@suse.de>
8582
8583 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
8584 bitmap and its handling.
8585 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
8586
9f8e7a96
RB
85872013-05-02 Richard Biener <rguenther@suse.de>
8588
8589 PR middle-end/57140
8590 * tree-inline.c (copy_loops): Properly handle removed loops.
8591 (copy_cfg_body): Mark destination loops for fixup if source
8592 loops needed fixup.
8593
f3a81b39
GY
85942013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
8595
8596 PR target/56732
8597 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
8598 generating simple_return for naked functions.
8599
7b920a9a
MJ
86002013-05-02 Martin Jambor <mjambor@suse.cz>
8601
8602 PR middle-end/56988
8603 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
8604 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
8605 flags match.
8606 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
8607 ipa_agg_replacement_value structures.
8608 (known_aggs_to_agg_replacement_list): Likewise.
8609 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
8610 (read_agg_replacement_chain): Likewise.
8611 (ipcp_transform_function): Also check that by_ref flags match.
8612
2c41c19d
RB
86132013-05-02 Richard Biener <rguenther@suse.de>
8614
8615 * graphds.h (struct graph): Add obstack member.
8616 * graphds.c (new_graph): Initialize obstack and allocate
8617 vertices from it.
8618 (add_edge): Allocate edge from the obstack.
3c21604f 8619 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 8620
8b47039c
TJ
86212013-05-02 Teresa Johnson <tejohnson@google.com>
8622
8623 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
8624 divides.
8625 * cfg.c (update_bb_profile_for_threading): Ditto.
8626 * tree-inline.c (copy_bb): Ditto.
8627 (copy_edges_for_bb): Ditto.
8628 (initialize_cfun): Ditto.
8629 (copy_cfg_body): Ditto.
8630 (expand_call_inline): Ditto.
8631 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
8632 (estimate_node_size_and_time): Ditto.
8633 (inline_merge_summary): Ditto.
8634 * cgraphclones.c (cgraph_clone_edge): Ditto.
8635 (cgraph_clone_node): Ditto.
8636 * sched-rgn.c (compute_dom_prob_ps): Ditto.
8637 (compute_trg_info): Ditto.
8638
da65928c
IB
86392013-05-02 Ian Bolton <ian.bolton@arm.com>
8640
8641 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
8642 S reg when fp attribute set.
8643 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
8644
add3c965
IB
86452013-05-02 Ian Bolton <ian.bolton@arm.com>
8646
8647 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
8648 New pattern.
8649 (*and_one_cmplsi3_compare0_uxtw): Likewise.
8650 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
8651 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
8652
fdd43ac4
RB
86532013-05-02 Richard Biener <rguenther@suse.de>
8654
8655 * tree-scalar-evolution.c (scev_info_hasher): Remove.
8656 (struct instantiate_cache_entry): New type.
8657 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
8658 (struct instantiate_cache_type): New type.
8659 (set_instantiated_value, get_instantiated_value): Remove.
8660 (get_instantiated_value_entry): New function.
8661 (instantiate_scev_name): Use the new cache and adjust.
8662 (instantiate_scev_poly): Adjust.
8663 (instantiate_scev_binary): Likewise.
8664 (instantiate_array_ref): Likewise.
8665 (instantiate_scev_convert): Likewise.
8666 (instantiate_scev_not): Likewise.
8667 (instantiate_scev_3): Likewise.
8668 (instantiate_scev_2): Likewise.
8669 (instantiate_scev_r): Likewise.
8670 (instantiate_scev): Likewise.
8671 (resolve_mixers): Likewise.
8672
36ff9dfb
VM
86732013-05-01 Vladimir Makarov <vmakarov@redhat.com>
8674
8675 PR target/57091
8676 * lra-constraints.c (best_small_class_operands_num): Remove.
8677 (process_alt_operands): Remove small_class_operands_num. Take
8678 small classes operands into losers and only if the operand is not
8679 matched. Modify debugging output.
8680 (curr_insn_transform): Remove best_small_class_operands_num.
8681 Print insn name.
8682
36054fab
JG
86832013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8684
8685 * config/aarch64/aarch64-builtins.c
8686 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
8687 * config/aarch64/aarch64-simd-builtins.def
8688 (reduc_splus_): Add new modes.
8689 (reduc_uplus_): New.
8690 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
8691 (reduc_uplus_v4sf): Likewise.
8692 (reduc_splus_v4sf): Likewise.
8693 (aarch64_addv<mode>): Likewise.
8694 (reduc_uplus_<mode>): Likewise.
8695 (reduc_splus_<mode>): Likewise.
8696 (aarch64_addvv2di): Likewise.
8697 (reduc_uplus_v2di): Likewise.
8698 (reduc_splus_v2di): Likewise.
8699 (aarch64_addvv2si): Likewise.
8700 (reduc_uplus_v2si): Likewise.
8701 (reduc_splus_v2si): Likewise.
8702 (reduc_<sur>plus_<mode>): New.
8703 (reduc_<sur>plus_v2di): Likewise.
8704 (reduc_<sur>plus_v2si): Likewise.
8705 (reduc_<sur>plus_v4sf): Likewise.
8706 (aarch64_addpv4sf): Likewise.
8707 * config/aarch64/arm_neon.h
8708 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
8709 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
8710 add UNSPEC_SADDV, UNSPEC_UADDV.
8711 (SUADDV): New.
8712 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
8713
6dce23a8
JG
87142013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8715
8716 * config/aarch64/arm_neon.h
8717 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
8718
1598945b
JG
87192013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8720
8721 * config/aarch64/aarch64-builtins
8722 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
8723
998eaf97
JG
87242013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8725
8726 * config/aarch64/aarch64-simd-builtins.def
8727 (reduc_smax_): New.
8728 (reduc_smin_): Likewise.
8729 (reduc_umax_): Likewise.
8730 (reduc_umin_): Likewise.
8731 (reduc_smax_nan_): Likewise.
8732 (reduc_smin_nan_): Likewise.
8733 (fmax): Remove.
8734 (fmin): Likewise.
8735 (smax): Update for V2SF, V4SF and V2DF modes.
8736 (smin): Likewise.
8737 (smax_nan): New.
8738 (smin_nan): Likewise.
8739 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
8740 (<su><maxmin><mode>3): ...This, refactor.
8741 (s<maxmin><mode>3): New.
8742 (<maxmin_uns><mode>3): Likewise.
8743 (reduc_<maxmin_uns>_<mode>): Refactor.
8744 (reduc_<maxmin_uns>_v4sf): Likewise.
8745 (reduc_<maxmin_uns>_v2si): Likewise.
8746 (aarch64_<fmaxmin><mode>: Remove.
8747 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
8748 new builtin names.
8749 (vmin<q>_f<32,64>): Likewise.
8750 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8751 (FMAXMIN): New.
8752 (su): Add mappings for smax, smin, umax, umin.
8753 (maxmin): New.
8754 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8755 (FMAXMIN): Rename as...
8756 (FMAXMIN_UNS): ...This.
8757 (maxminv): Remove.
8758 (fmaxminv): Likewise.
8759 (fmaxmin): Likewise.
8760 (maxmin_uns): New.
8761 (maxmin_uns_op): Likewise.
8762
bd11644e
JG
87632013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8764
8765 * config/aarch64/arm_neon.h
8766 (vac<ge, gt><sd>_f<32, 64>): Rename to...
8767 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
8768 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
8769
75dd5ace
JG
87702013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8771
8772 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
8773 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
8774
7c19979f
JG
87752013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8776
8777 * config/aarch64/aarch64-simd.md
8778 (vcond<mode>_internal): Handle special cases for constant masks.
8779 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
8780 (vcondu<mode><mode>): Likewise.
8781 (vcond<v_cmp_result><mode>): New.
8782
bb60efd9
JG
87832013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8784
8785 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
8786 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
8787 * config/aarch64/aarch64-simd-builtins.def
8788 (cmeq): Update to BUILTIN_VALLDI.
8789 (cmgt): Likewise.
8790 (cmge): Likewise.
8791 (cmle): Likewise.
8792 (cmlt): Likewise.
8793 * config/aarch64/arm_neon.h
8794 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
8795 to builtins or C as appropriate.
8796
889b9412
JG
87972013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8798
8799 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
8800 (cmgeu): ...This.
8801 (cmhi): Rename to...
8802 (cmgtu): ...This.
8803 * config/aarch64/aarch64-simd.md
8804 (simd_mode): Add SF.
8805 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
8806 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
8807 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
8808 (cstore<mode>_neg): ...This.
8809 * config/aarch64/iterators.md
8810 (VALLF): new.
8811 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
8812 (COMPARISONS): New.
8813 (UCOMPARISONS): Likewise.
8814 (optab): Add missing comparisons.
8815 (n_optab): New.
8816 (cmp_1): Likewise.
8817 (cmp_2): Likewise.
8818 (CMP): Likewise.
8819 (cmp): Remove.
8820 (VCMP_S): Likewise.
8821 (VCMP_U): Likewise.
8822 (V_cmp_result): Add DF, SF modes.
8823 (v_cmp_result): Likewise.
8824 (v): Likewise.
8825 (vmtype): Likewise.
8826 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
8827
0a7dbb76
GY
88282013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
8829
8830 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
8831 define_insn to define_insn_and_split.
8832 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
8833 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
8834 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
8835 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
8836 (thumb2_negscc): Likewise.
8837
fb614ca6
GY
88382013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8839
8840 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
8841
9e64a0bf
GY
88422013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8843
8844 * config/arm/thumb2.md: Remove trailing whitespaces.
8845
d6b28156
RS
88462013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8847
8848 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
8849 Use gen_int_mode rather than GEN_INT.
8850
f91674c3
L
88512013-04-30 H.J. Lu <hongjiu.lu@intel.com>
8852
b0dec607 8853 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L
8854 debug_gimple_stmt.
8855
3551257c
RB
88562013-04-30 Richard Biener <rguenther@suse.de>
8857
8858 PR middle-end/57122
3c21604f 8859 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 8860
4c1aff1c
RB
88612013-04-30 Richard Biener <rguenther@suse.de>
8862
8863 PR middle-end/57107
8864 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
8865
54b8379a
AB
88662013-04-30 Andrey Belevantsev <abel@ispras.ru>
8867
8868 PR rtl-optimization/56957
8869 PR rtl-optimization/57105
54b8379a
AB
8870 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
8871 variable. Use just INSN_UID for determining whether an insn
8872 should be only disconnected from the insn stream.
8873 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
8874
abc27962
JJ
88752013-04-30 Jakub Jelinek <jakub@redhat.com>
8876
8877 PR tree-optimization/57104
8878 * tsan.c (instrument_expr): Don't instrument accesses to
8879 DECL_HARD_REGISTER VAR_DECLs.
8880
0fc822d0
RB
88812013-04-30 Richard Biener <rguenther@suse.de>
8882
8883 * function.h (loops_for_fn): New inline function.
8884 (set_loops_for_fn): Likewise.
8885 * cfgloop.h (place_new_loop): Add struct function parameter.
8886 (get_loop): Likewise.
8887 (get_loops): Likewise.
8888 (number_of_loops): Likewise.
8889 (fel_next): Adjust.
8890 (fel_init): Likewise.
8891 * cfg.c (get_loop_copy): Adjust.
8892 * cfgloop.c (flow_loops_dump): Likewise.
8893 (record_loop_exits): Likewise.
8894 (verify_loop_structure): Likewise.
8895 * cfgloopanal.c (mark_irreducible_loops): Likewise.
8896 (estimate_reg_pressure_cost): Likewise.
8897 (mark_loop_exit_edges): Likewise.
8898 * cfgloopmanip.c (place_new_loop): Likewise.
8899 (add_loop): Likewise.
8900 (duplicate_loop): Likewise.
8901 * graph.c (draw_cfg_nodes): Likewise.
8902 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
8903 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
8904 (extract_affine_chrec): Likewise.
8905 (build_scop_iteration_domain): Likewise.
8906 * graphite.c (graphite_initialize): Likewise.
8907 * ira-build.c (create_loop_tree_nodes): Likewise.
8908 (more_one_region_p): Likewise.
8909 (rebuild_regno_allocno_maps): Likewise.
8910 (mark_loops_for_removal): Likewise.
8911 (mark_all_loops_for_removal): Likewise.
8912 (remove_unnecessary_regions): Likewise.
8913 (ira_build): Likewise.
8914 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
8915 * loop-init.c (fix_loop_structure): Likewise.
8916 (gate_rtl_move_loop_invariants): Likewise.
8917 (gate_rtl_unswitch): Likewise.
8918 (gate_rtl_unroll_and_peel_loops): Likewise.
8919 (rtl_doloop): Likewise.
8920 * lto-streamer-in.c (input_cfg): Likewise.
8921 * lto-streamer-out.c (output_cfg): Likewise.
8922 * modulo-sched.c (sms_schedule): Likewise.
8923 * predict.c (tree_estimate_probability): Likewise.
8924 (tree_estimate_probability_driver): Likewise.
8925 (estimate_loops): Likewise.
8926 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
8927 (move_sese_region_to_fn): Likewise.
8928 (debug_loop_num): Likewise.
8929 * tree-chrec.c (chrec_evaluate): Likewise.
8930 (hide_evolution_in_other_loops_than_loop): Likewise.
8931 (chrec_component_in_loop_num): Likewise.
8932 (reset_evolution_in_loop): Likewise.
8933 (evolution_function_is_invariant_rec_p): Likewise.
8934 * tree-if-conv.c (main_tree_if_conversion): Likewise.
8935 * tree-inline.c (copy_loops): Likewise.
8936 (copy_cfg_body): Likewise.
8937 (tree_function_versioning): Likewise.
8938 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
8939 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
8940 Likewise.
8941 (add_to_evolution_1): Likewise.
8942 (scev_const_prop): Likewise.
8943 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
8944 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
8945 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
8946 (tree_ssa_lim_initialize): Likewise.
8947 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
8948 (verify_loop_closed_ssa): Likewise.
8949 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
8950 (tree_ssa_loop_im): Likewise.
8951 (tree_ssa_loop_unswitch): Likewise.
8952 (tree_vectorize): Likewise.
8953 (check_data_deps): Likewise.
8954 (tree_ssa_loop_ivcanon): Likewise.
8955 (tree_ssa_loop_bounds): Likewise.
8956 (tree_complete_unroll): Likewise.
8957 (tree_complete_unroll_inner): Likewise.
8958 (tree_parallelize_loops): Likewise.
8959 (tree_ssa_loop_prefetch): Likewise.
8960 (tree_ssa_loop_ivopts): Likewise.
8961 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
8962 * tree-vectorizer.c (vectorize_loops): Likewise.
8963
37953bd3
MF
89642013-04-29 Mike Frysinger <vapier@gentoo.org>
8965
8966 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
8967 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
8968 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
8969 with EABI_LINK_SPEC.
8970
f9ed28db
UB
89712013-04-29 Uros Bizjak <ubizjak@gmail.com>
8972
8973 PR target/44578
8974 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
8975 alternative.
8976
deca73f5
VM
89772013-04-29 Vladimir Makarov <vmakarov@redhat.com>
8978
8979 PR target/57097
37953bd3 8980 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM
8981 using memory for pseudos. Print cost dump for alternatives.
8982 Modify cost values for conflicts with early clobbers.
8983 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
8984
fdca7d03
UB
89852013-04-29 Uros Bizjak <ubizjak@gmail.com>
8986
8987 PR target/57098
8988 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
8989
0b064172
IB
89902013-04-29 Ian Bolton <ian.bolton@arm.com>
8991
8992 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
8993 from/to S register.
8994 (movdi_aarch64): Support LDR/STR from/to D register.
8995
473cec55
IB
89962013-04-29 Ian Bolton <ian.bolton@arm.com>
8997
8998 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
8999 or higher by default.
9000
a6f30e66
RB
90012013-04-29 Richard Biener <rguenther@suse.de>
9002
9003 PR middle-end/57075
9004 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9005 even if not adding abnormal edges for calls that can make
9006 abnormal gotos.
9007
0c2b2040
RB
90082013-04-29 Richard Biener <rguenther@suse.de>
9009
9010 PR middle-end/57103
9011 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9012 TREE_BLOCK.
9013 (move_stmt_r): Remove redundant checking.
9014
f41f80f9
TJ
90152013-04-29 Teresa Johnson <tejohnson@google.com>
9016
9017 PR bootstrap/57077
9018 * basic-block.h (apply_scale): New function.
9019 (apply_probability): Use apply_scale.
9020 * gimple-streamer-in.c (input_bb): Ditto.
9021 * lto-streamer-in.c (input_cfg): Ditto.
9022 * lto-cgraph.c (merge_profile_summaries): Ditto.
9023 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 9024 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ
9025 (copy_edges_for_bb): Ditto.
9026 (copy_cfg_body): Ditto.
9027
315bbd2e
TV
90282013-04-29 Tom de Vries <tom@codesourcery.com>
9029
9030 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9031 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9032 (tail_merge_optimize): Handle current_loops == NULL.
9033
ebbd90d8
JL
90342013-04-26 Jeff Law <law@redhat.com>
9035
9036 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9037 (simplify_cond_using_ranges): Generalize code to simplify
9038 COND_EXPRs where one argument is a constant and the other
9039 is an SSA_NAME created by an integral type conversion.
9040
8b9b57eb
KT
90412013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9042
9043 * config/arm/arm.md (store_minmaxsi): Use only when
9044 optimize_insn_for_size_p.
9045
9498e5dc
CB
90462013-04-29 Christian Bruel <christian.bruel@st.com>
9047
9048 PR target/57108
9049 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9050
6093bc06
RB
90512013-04-29 Richard Biener <rguenther@suse.de>
9052
9053 PR middle-end/57089
fdca7d03
UB
9054 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9055 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB
9056 (expand_omp_for_generic): Properly add loops.
9057 (expand_omp_for_static_nochunk): Likewise.
9058 (expand_omp_for_static_chunk): Likewise.
9059 (expand_omp_for): For the degenerate case fixup loops.
9060 (expand_omp_sections): Fix default bb placement in loops.
9061 (expand_omp_atomic_pipeline): Properly add loops.
9062
84aacbfd
KT
90632013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9064
9065 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9066
8dee4479
TV
90672013-04-29 Tom de Vries <tom@codesourcery.com>
9068
9069 * tree-ssa-tail-merge.c: Update header comment.
9070
47934dc4
JG
90712013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9072
9073 * config/aarch64/arm_neon.h
9074 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9075 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9076 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9077 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9078 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9079 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9080 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9081 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9082
384be29f
JG
90832013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9084
9085 * config/aarch64/aarch64-simd.md
9086 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9087 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9088 fix_trunc, fixuns_trunc.
9089 (ftrunc<VDQF:mode>2): New.
9090 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9091 (fix_trunc_optab): New.
9092
0386b123
JG
90932013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9094
9095 * config/aarch64/aarch64-builtins.c
9096 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9097 iceilf, lround, iroundf.
9098
00fcb892
UB
90992013-04-29 Uros Bizjak <ubizjak@gmail.com>
9100
9101 PR target/54349
9102 * config/i386/i386.h (enum ix86_tune_indices)
9103 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9104 New, split from X86_TUNE_INTER_UNIT_MOVES.
9105 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9106 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9107 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9108 (TARGET_INTER_UNIT_MOVES): Remove.
9109 * config/i386/i386.c (initial_ix86_tune_features): Update.
9110 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9111 (ix86_expand_convert_uns_didf_sse): Use
9112 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9113 (ix86_expand_vector_init_one_nonzero): Ditto.
9114 (ix86_expand_vector_init_interleave): Ditto.
9115 (inline_secondary_memory_needed): Return true for moves from SSE class
9116 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9117 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9118 * config/i386/constraints.md (Yi, Ym): Depend on
9119 TARGET_INTER_UNIT_MOVES_TO_VEC.
9120 (Yj, Yn): New constraints.
9121 * config/i386/i386.md (*movdi_internal): Change constraints of
9122 operand 1 from Yi to Yj and from Ym to Yn.
9123 (*movsi_internal): Ditto.
9124 (*movdf_internal): Ditto.
9125 (*movsf_internal): Ditto.
9126 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9127 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9128 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9129 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9130 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9131 * config/i386/sse.md (movdi_to_sse): Ditto.
9132 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9133 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9134 TARGET_INTER_UNIT_MOVES.
9135 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9136 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9137 instead of TARGET_INTER_UNIT_MOVES.
9138 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9139 operand 1 from Yi to Yj and from Ym to Yn.
9140
4c871069
JG
91412013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9142
9143 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9144 (float_truncate_hi_): Likewise.
9145 (float_extend_lo_): Likewise.
9146 (float_truncate_lo_): Likewise.
9147 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9148 (aarch64_float_extend_lo_v2df): Likewise.
9149 (vec_unpacks_hi_v4sf): Likewise.
9150 (aarch64_float_truncate_lo_v2sf): Likewise.
9151 (aarch64_float_truncate_hi_v4sf): Likewise.
9152 (vec_pack_trunc_v2df): Likewise.
9153 (vec_pack_trunc_df): Likewise.
9154
1709ff9b
JG
91552013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9156
9157 * config/aarch64/aarch64-builtins.c
9158 (aarch64_fold_builtin): Fold float conversions.
9159 * config/aarch64/aarch64-simd-builtins.def
9160 (floatv2si, floatv4si, floatv2di): New.
9161 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9162 * config/aarch64/aarch64-simd.md
9163 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9164 * config/aarch64/iterators.md (FLOATUORS): New.
9165 (optab): Add float, floatuns.
9166 (su_optab): Likewise.
9167
ce966824
JG
91682013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9169
9170 * config/aarch64/aarch64-builtins.c
9171 (aarch64_builtin_vectorized_function): Use new names for
9172 fcvt builtins.
9173 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9174 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9175 (fcvtzu): Split as...
9176 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9177 (fcvtas): Split as...
9178 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9179 (fcvtau): Split as...
9180 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9181 (fcvtps): Split as...
9182 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9183 (fcvtpu): Split as...
9184 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9185 (fcvtms): Split as...
9186 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9187 (fcvtmu): Split as...
9188 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9189 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9190 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9191 (lfrintnusf, lfrintnudf): Likewise.
9192 * config/aarch64/aarch64-simd.md
9193 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9194 define_insn.
9195 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9196 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9197 (fcvt_pattern): Likewise.
9198
b9de24fe
JG
91992013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9200
9201 * config/aarch64/aarch64-simd.md
9202 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9203 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9204
77a205be
JG
92052013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9206
9207 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9208 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9209 (vrnd<a,m,n,p>_f32): Implement using builtins.
9210 (vrnd<i,x><q>_f<32, 64>): New.
9211
0659ce6f
JG
92122013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9213
9214 * config/aarch64/aarch64-builtins.c
9215 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9216 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9217 (frintz): Rename to...
9218 (btrunc): ...this.
9219 (frintp): Rename to...
9220 (ceil): ...this.
9221 (frintm): Rename to...
9222 (floor): ...this.
9223 (frinti): Rename to...
9224 (nearbyint): ...this.
9225 (frintx): Rename to...
9226 (rint): ...this.
9227 (frinta): Rename to...
9228 (round): ...this.
9229 * config/aarch64/aarch64-simd.md
9230 (aarch64_frint<frint_suffix><mode>): Delete.
9231 (<frint_pattern><mode>2): Convert to insn.
9232 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9233 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9234 (frint_pattern): Likewise.
9235 (frint_suffix): Likewise.
9236
ea78906a
RB
92372013-04-29 Richard Biener <rguenther@suse.de>
9238
9239 PR tree-optimization/57081
9240 * loop-init.c: Include tree-flow.h.
9241 (loop_optimizer_finalize): Free number of iteration estimates.
9242 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9243
baee1763
JJ
92442013-04-29 Jakub Jelinek <jakub@redhat.com>
9245
94dc5332
JJ
9246 PR tree-optimization/57083
9247 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9248 non-singleton shift count range, zero extend low_bound for uns case.
9249
baee1763
JJ
9250 * config/i386/predicates.md (general_vector_operand): New predicate.
9251 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9252 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9253 if they aren't nonimmediate operands. If their original values
9254 satisfy const_vector_equal_evenodd_p, don't shift them.
9255 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9256 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9257 if not nonimmediate_operand.
9258 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9259 instead of register_operand.
9260 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9261
a9073727 92622013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB
9263
9264 * stor-layout.c (finalize_size_functions): Allocate a structure and
9265 reset cfun before dumping the functions.
9266
ba8011e6
JJ
92672013-04-27 Jakub Jelinek <jakub@redhat.com>
9268
d6fde69e
JJ
9269 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9270
ba8011e6
JJ
9271 PR target/56866
9272 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9273 use xop_pmacsdqh if uns_p.
9274 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9275 the immediate rotate count.
9276
2c62cbaa
VM
92772013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9278
9279 * rtl.h (struct rtx_def): Add comment for field jump.
9280 (LRA_SUBREG_P): New macro.
9281 * recog.c (register_operand): Check LRA_SUBREG_P.
9282 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9283 stack frame.
9284 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9285 (lra_final_code_change): Skip subreg change for operators.
9286 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9287 if there are no operand changes.
9288 * lra-constraints.c (curr_insn_set): New.
9289 (match_reload): Set LRA_SUBREG_P.
9290 (emit_spill_move): Ditto.
9291 (check_and_process_move): Use curr_insn_set. Process only single
9292 set insns. Don't initialize sec_mem_p and change_p.
9293 (simplify_operand_subreg): Use LRA_SUBREG_P.
9294 (reg_in_class_p): New function.
9295 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9296 of #ifdef. Add code to remove cycling.
9297 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9298 non-null disp. Reload inner instead of disp when base and index
9299 are null. Try to put lo_sum into register.
9300 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9301 (check_and_process_move): Move code for move cost check to
9302 simple_move_p. Remove equiv_substitution.
9303 (simple_move_p): New function.
9304 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9305 curr_insn_set. Call check_and_process_move only for single set
9306 insns. Use the new function. Move call of check_and_process_move
9307 after operand equiv substitution and address process.
9308
e7d764f3
JJ
93092013-04-26 Jakub Jelinek <jakub@redhat.com>
9310
9311 PR go/57045
9312 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9313 with nonlocal goto receivers or returns twice calls, ignore
9314 unininitialized values from abnormal edges to nl goto receiver
9315 or returns twice call.
9316
41e10689
JJ
93172013-04-26 Jakub Jelinek <jakub@redhat.com>
9318
9319 PR tree-optimization/57051
9320 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9321 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9322 bitsize.
9323
d7ed20db
RB
93242013-04-26 Richard Biener <rguenther@suse.de>
9325
9326 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9327 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9328 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9329 (fixup_loop_arrays_after_move): New function.
9330 (move_sese_region_to_fn): Properly outline the loop tree parts
9331 of the SESE region.
9332
df93505e
UB
93332013-04-26 Uros Bizjak <ubizjak@gmail.com>
9334
9335 * config/i386/i386.md (type, unit): Fix long lines.
9336
dd366ec3
RB
93372013-04-26 Richard Biener <rguenther@suse.de>
9338
9339 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9340 (lto-streamer-out.o): Likewise.
9341 * cfgloop.c (init_loops_structure): Export, add struct function
9342 argument and adjust.
9343 (flow_loops_find): Adjust.
9344 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9345 (init_loops_structure): Declare.
9346 * lto-streamer-in.c: Include cfgloop.h.
9347 (input_cfg): Input the loop tree.
9348 * lto-streamer-out.c: Include cfgloop.h.
9349 (output_cfg): Output the loop tree.
9350 (output_struct_function_base): Do not drop PROP_loops.
9351
a9e0d843
RB
93522013-03-26 Richard Biener <rguenther@suse.de>
9353
9354 * tree-cfg.c (execute_build_cfg): Build the loop tree.
9355 (pass_build_cfg): Provide PROP_loops.
9356 (move_sese_region_to_fn): Remove loops that are outlined into fn
9357 for now.
9358 * tree-inline.c: Include cfgloop.h.
9359 (initialize_cfun): Do not drop PROP_loops.
9360 (copy_loops): New function.
9361 (copy_cfg_body): Copy loop structure.
9362 (tree_function_versioning): Initialize destination loop tree.
9363 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
9364 (pass_parallelize_loops): Do IL verification.
9365 * loop-init.c (loop_optimizer_init): Fixup loops if required.
9366 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
9367 the CFG make sure we fixup loops as well.
9368 * tree-ssa-tail-merge.c: Include cfgloop.h.
9369 (replace_block_by): When merging loop latches mark loops for fixup.
9370 * lto-streamer-out.c (output_struct_function_base): Drop
9371 PROP_loops for now.
9372 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
9373 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
9374 * ipa-split.c: Include cfgloop.h.
9375 (split_function): Add the new return block to the loop tree root.
9376 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
9377 whether we have removed the forwarder block.
9378 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
9379 * cfgloop.h (place_new_loop): Declare.
9380 * cfgloopmanip.c (place_new_loop): Export.
9381 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
9382 (tree-switch-conversion.o): Likewise.
9383 (tree-complex.o): Likewise.
9384 (tree-inline.o): Likewise.
9385 (tree-ssa-tailmerge.o): Likewise.
9386 (ipa-split.o): Likewise.
9387 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
9388 (tree-ssa-copy.o): Likewise.
9389 * tree-switch-conversion.c: Include cfgloop.h
9390 (process_switch): If we emit a bit-test cascade, schedule loops
9391 for fixup.
9392 * tree-complex.c: Include cfgloop.h.
9393 (expand_complex_div_wide): Properly add new basic-blocks to loops.
9394 * asan.c: Include cfgloop.h.
9395 (create_cond_insert_point): Properly add new basic-blocks to
9396 loops, schedule loop fixup.
9397 * cfgloop.c (verify_loop_structure): Check that looks are not
9398 marked for fixup.
9399 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
9400 to loops.
9401 (expand_omp_for_generic): Likewise.
9402 (expand_omp_sections): Likewise.
9403 (expand_omp_atomic_pipeline): Schedule loops for fixup.
9404 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
9405 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
9406 is initialized, not when loops are present.
9407 * tree-parloops.c (parallelize_loops): Remove checking here.
9408 * passes.c (init_optimization_passes): Schedule a copy-propagation
9409 pass before complete unrolling of inner loops.
9410
e78e8a0b
JJ
94112013-04-26 Jakub Jelinek <jakub@redhat.com>
9412
a2e836b2
JJ
9413 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
9414 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
9415 (colorize_init): Add argument to _WIN32 version.
9416 * toplev.c: Include diagnostic-color.h.
9417 (process_options): Default to -fdiagnostics-color=auto if
9418 GCC_COLORS env var is in the environment.
9419 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB
9420 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
9421 env var is in the environment, the default is auto rather than never.
a2e836b2 9422
e78e8a0b
JJ
9423 * diagnostic.h (file_name_as_prefix): Add context argument.
9424 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
9425 the string as locus.
9426 * langhooks.c (lhd_print_error_function): Adjust caller.
9427
013e5ef9
LC
94282013-04-25 Lawrence Crowl <crowl@google.com>
9429
9430 * var-tracking.c (shared_hash_def::htab):
9431 Change type to hash_table. Update dependent calls and types.
9432
4a8fb1a1
LC
94332013-04-25 Lawrence Crowl <crowl@google.com>
9434
9435 * Makefile.in: Update as needed below.
9436
9437 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
9438 Move declaration to after the type's method definitons.
9439
9440 * attribs.c (htab_t scoped_attributes::attribute_hash):
9441 Change type to hash_table. Update dependent calls and types.
9442
9443 * bitmap.c (htab_t bitmap_desc_hash):
9444 Change type to hash_table. Update dependent calls and types.
9445
9446 * cselib.c (htab_t cselib_hash_table):
9447 Change type to hash_table. Update dependent calls and types.
9448
9449 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
9450 (hash_string_slot_node): Move implementation into lto-streamer.h
9451 struct string_slot_hasher.
9452 (eq_string_slot_node): Likewise.
9453
9454 * data-streamer-out.c: Update output_block::string_hash_table
9455 dependent calls and types.
9456
9457 * dwarf2cfi.c (htab_t trace_index):
9458 Change type to hash_table. Update dependent calls and types.
9459
9460 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
9461 Change type to hash_table. Update dependent calls and types.
9462 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
9463 (htab_t optimize_external_refs::map): Likewise.
9464 (htab_t output_comp_unit::extern_map): Likewise.
9465 (htab_t output_comdat_type_unit::extern_map): Likewise.
9466 (htab_t output_macinfo::macinfo_htab): Likewise.
9467 (htab_t optimize_location_lists::htab): Likewise.
9468 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
9469
9470 * except.c (htab_t ehspec_hash_type):
9471 Change type to hash_table. Update dependent calls and types.
9472 (assign_filter_values::ttypes): Likewise.
9473 (assign_filter_values::ehspec): Likewise.
9474 (sjlj_assign_call_site_values::ar_hash): Likewise.
9475 (convert_to_eh_region_ranges::ar_hash): Likewise.
9476
9477 * gcse.c (htab_t pre_ldst_table):
9478 Change type to hash_table. Update dependent calls and types.
9479
9480 * ggc-common.c (htab_t saving_htab):
9481 Change type to hash_table. Update dependent calls and types.
9482 (htab_t loc_hash): Likewise.
9483 (htab_t ptr_hash): Likewise.
9484 (call_count): Rename ggc_call_count.
9485 (call_alloc): Rename ggc_call_alloc.
9486 (loc_descriptor): Rename make_loc_descriptor.
9487 (add_statistics): Rename ggc_add_statistics.
9488
9489 * ggc-common.c (saving_htab):
9490 Change type to hash_table. Update dependent calls and types.
9491
9492 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
9493 (push_gimplify_context): Likewise.
9494 (pop_gimplify_context): Likewise.
9495 (struct gimple_temp_hash_elt): Added.
9496 (struct gimplify_hasher): Likewise.
9497 (struct gimplify_ctx.temp_htab):
9498 Change type to hash_table. Update dependent calls and types.
9499
9500 * gimple-fold.c: Include gimplify-ctx.h.
9501
9502 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
9503 Change type to hash_table. Update dependent calls and types.
9504 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
9505 avoid potential global name collision.
9506
9507 * gimplify.c: Include gimplify-ctx.h.
9508 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
9509 (htab_t gimplify_ctx::temp_htab):
9510 Update dependent calls and types for new type hash_table.
9511 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
9512 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
9513
9514 * gimplify-ctx.h: New.
9515 (struct gimple_temp_hash_elt): Move from gimplify.c.
9516 (class gimplify_hasher): New.
9517 (struct gimplify_ctx): Move from gimple.h.
9518 (htab_t gimplify_ctx::temp_htab):
9519 Change type to hash_table. Update dependent calls and types.
9520
9521 * graphite-clast-to-gimple.c: Include graphite-htab.h.
9522 (htab_t ivs_params::newivs_index):
9523 Change type to hash_table. Update dependent calls and types.
9524 (htab_t ivs_params::params_index): Likewise.
9525 (htab_t print_generated_program::params_index): Likewise.
9526 (htab_t gloog::newivs_index): Likewise.
9527 (htab_t gloog::params_index): Likewise.
9528
9529 * graphite.c: Include graphite-htab.h.
9530 4htab_t graphite_transform_loops::bb_pbb_mapping):
9531 Change type to hash_table. Update dependent calls and types.
9532
9533 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
9534 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
9535 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
9536
9537 * graphite-dependences.c: Include graphite-htab.h.
9538 (loop_is_parallel_p): Change hash table type of parameter.
9539
9540 * graphite-htab.h: New.
9541 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
9542 (extern find_pbb_via_hash): Move from graphite-poly.h.
9543 (extern loop_is_parallel_p): Move from graphite-poly.h.
9544 (extern get_loop_body_pbbs): Move from graphite-poly.h.
9545
9546 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
9547 (extern loop_is_parallel_p): Move to graphite-htab.h.
9548 (extern get_loop_body_pbbs): Move to graphite-htab.h.
9549
9550 * haifa-sched.c (htab_t delay_htab):
9551 Change type to hash_table. Update dependent calls and types.
9552 (htab_t delay_htab_i2): Likewise.
9553
9554 * ira-color.c (htab_t allocno_hard_regs_htab):
9555 Change type to hash_table. Update dependent calls and types.
9556
9557 * ira-costs.c (htab_t cost_classes_htab):
9558 Change type to hash_table. Update dependent calls and types.
9559
9560 * loop-invariant.c (htab_t merge_identical_invariants::eq):
9561 Change type to hash_table. Update dependent calls and types.
9562
9563 * loop-iv.c (htab_t bivs):
9564 Change type to hash_table. Update dependent calls and types.
9565
9566 * loop-unroll.c (htab_t opt_info::insns_to_split):
9567 Change type to hash_table. Update dependent calls and types.
9568 (htab_t opt_info::insns_with_var_to_expand): Likewise.
9569
9570 * lto-streamer.h (struct string_slot): Move from data-streamer.h
9571 (struct string_slot_hasher): New.
9572 (htab_t output_block::string_hash_table):
9573 Change type to hash_table. Update dependent calls and types.
9574
9575 * lto-streamer-in.c (freeing_string_slot_hasher): New.
9576 (htab_t file_name_hash_table):
9577 Change type to hash_table. Update dependent calls and types.
9578
9579 * lto-streamer-out.c: Update output_block::string_hash_table dependent
9580 calls and types.
9581
9582 * lto-streamer.c (htab_t tree_htab):
9583 Change type to hash_table. Update dependent calls and types.
9584
9585 * omp-low.c: Include gimplify-ctx.h.
9586
9587 * passes.c (htab_t name_to_pass_map):
9588 Change type to hash_table. Update dependent calls and types.
9589 (pass_traverse): Rename to passes_pass_traverse.
9590
9591 * plugin.c (htab_t event_tab):
9592 Change type to hash_table. Update dependent calls and types.
9593
9594 * postreload-gcse.c (htab_t expr_table):
9595 Change type to hash_table. Update dependent calls and types.
9596 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
9597
9598 * sese.c (debug_rename_map_1): Make extern.
9599 (htab_t copy_bb_and_scalar_dependences::rename_map):
9600 Change type to hash_table. Update dependent calls and types.
9601
9602 * sese.h (extern debug_rename_map): Move to .c file.
9603
9604 * store-motion.c (htab_t store_motion_mems_table):
9605 Change type to hash_table. Update dependent calls and types.
9606
9607 * trans-mem.c (htab_t tm_new_mem_hash):
9608 Change type to hash_table. Update dependent calls and types.
9609
9610 * tree-browser.c (htab_t TB_up_ht):
9611 Change type to hash_table. Update dependent calls and types.
9612
9613 * tree-cfg.c (htab_t discriminator_per_locus):
9614 Change type to hash_table. Update dependent calls and types.
9615
9616 * tree-complex.c: Include tree-hasher.h
9617 (htab_t complex_variable_components):
9618 Change type to hash_table. Update dependent calls and types.
9619
9620 * tree-eh.c (htab_t finally_tree):
9621 Change type to hash_table. Update dependent calls and types.
9622
9623 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
9624 struct int_tree_hasher.
9625 (extern int_tree_map_eq): Likewise.
9626 (uid_decl_map_hash): Removed.
9627 (extern decl_tree_map_eq): Likewise.
9628
9629 * tree-hasher.h: New.
9630 (struct int_tree_hasher): New.
9631 (typedef int_tree_htab_type): New.
9632
9633 * tree-inline.c: Include gimplify-ctx.h.
9634
9635 * tree-mudflap.c: Include gimplify-ctx.h.
9636
9637 * tree-parloops.c: Include tree-hasher.h.
9638 (htab_t eliminate_local_variables_stmt::decl_address):
9639 Change type to hash_table. Update dependent calls and types.
9640 (htab_t separate_decls_in_region::decl_copies): Likewise.
9641
9642 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
9643 Change type to hash_table. Update dependent calls and types.
9644
9645 * tree-sra.c (candidates):
9646 Change type to hash_table. Update dependent calls and types.
9647
9648 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
9649 in tree-flow.h.
9650 (int_tree_map_hash): Likewise.
9651
9652 * tree-ssa-dom.c (htab_t avail_exprs):
9653 Change type to hash_table. Update dependent calls and types.
9654
9655 * tree-ssa-live.c (var_map_base_init::tree_to_index):
9656 Change type to hash_table. Update dependent calls and types.
9657
9658 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
9659 Change type to hash_table. Update dependent calls and types.
9660
9661 * tree-ssa-phiopt.c (seen_ssa_names):
9662 Change type to hash_table. Update dependent calls and types.
9663
9664 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
9665 Change type to hash_table. Update dependent calls and types.
9666
9667 * tree-ssa-uncprop.c (equiv):
9668 Change type to hash_table. Update dependent calls and types.
9669
c5a44004
JJ
96702013-04-25 Jakub Jelinek <jakub@redhat.com>
9671
9672 PR rtl-optimization/57003
9673 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
9674 call note_stores with kill_clobbered_value callback again after
9675 killing regs_invalidated_by_call.
9676
09962a4a
JG
96772013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9678
9679 * config/aarch64/aarch64-simd.md
9680 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
9681 (aarch64_simd_bsl<mode>): Likewise.
9682 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
9683
ea28bb0b
MP
96842013-04-25 Marek Polacek <polacek@redhat.com>
9685
9686 PR tree-optimization/57066
3c21604f 9687 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 9688
96659611
JG
96892013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
9690
9691 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
9692
9697e620
JG
96932013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9694
9695 * config/aarch64/aarch64-builtins.c
9696 (aarch64_fold_builtin): New.
9697 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
9698 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
9699 * config/aarch64/aarch64-simd-builtins.def (abs): New.
9700 * config/aarch64/arm_neon.h
9701 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
9702
0ac198d3
JG
97032013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9704 Tejas Belagod <tejas.belagod@arm.com>
9705
9706 * config/aarch64/aarch64-builtins.c
9707 (aarch64_gimple_fold_builtin): New.
9708 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
9709 * config/aarch64/aarch64-simd-builtins.def (addv): New.
9710 * config/aarch64/aarch64-simd.md (addpv4sf): New.
9711 (addvv4sf): Update.
9712 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
9713
58cff58c
N
97142013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9715
df93505e 9716 * config/aarch64/aarch64.md
58cff58c
N
9717 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
9718
7e0228bf
N
97192013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9720
9721 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
9722 (*ngcsi_uxtw): New pattern.
9723
5819f96f 97242013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 9725 Julian Brown <julian@codesourcery.com>
5819f96f
KT
9726
9727 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
9728 (TB_DREG): Add T_V4HF.
9729 (v4hf_UP): New macro.
9730 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 9731 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
9732 Handle initialisation of V4HF. Adjust initialisation of reinterpret
9733 built-ins.
df93505e 9734 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
9735 (arm_vector_mode_supported_p): Handle V4HF.
9736 (arm_mangle_map): Handle V4HFmode.
9737 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
9738 * config/arm/arm_neon_builtins.def: Add entries for
9739 vcvtv4hfv4sf, vcvtv4sfv4hf.
9740 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
9741 (neon_vcvtv4hfv4sf): Likewise.
9742 * config/arm/neon-gen.ml: Handle half-precision floating point
9743 features.
9744 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
9745 * config/arm/arm_neon.h: Regenerate.
9746 * config/arm/neon.ml (type elts): Add F16.
9747 (type vectype): Add T_float16x4, T_floatHF.
9748 (type vecmode): Add V4HF.
9749 (type features): Add Requires_FP_bit feature.
9750 (elt_width): Handle F16.
9751 (elt_class): Likewise.
9752 (elt_of_class_width): Likewise.
9753 (mode_of_elt): Refactor.
9754 (type_for_elt): Handle F16, fix error messages.
9755 (vectype_size): Handle T_float16x4.
9756 (vcvt_sh): New function.
9757 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
9758 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
9759 (string_of_mode): Handle V4HF.
9760 * doc/arm-neon-intrinsics.texi: Regenerate.
9761
1ef395e4
JG
97622013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9763
9764 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
9765 format specifier in 'X' case.
9766
41c34e94
AM
97672013-04-25 Alan Modra <amodra@gmail.com>
9768
9769 PR target/57052
9770 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
9771 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
9772 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
9773 Repeat for many other rotate/shift and mask patterns using subregs.
9774 Name lshiftrt insns.
9775 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
9776 on WORDS_BIG_ENDIAN.
9777
b9a7eb5d
AM
97782013-04-25 Alan Modra <amodra@gmail.com>
9779
9780 * config.gcc: Support little-endian powerpc-linux targets.
9781 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
9782 (LINK_OS_LINUX_SPEC): Define.
9783 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
9784 Preserve MASK_LITTLE_ENDIAN.
9785 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
9786 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
9787 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
9788 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
9789 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
9790 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
9791 Correct fp word order for little-endian. Don't shift toc entries
9792 smaller than a word for little-endian.
9793 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
9794 (bswapdi2 splits): Correct low-part subreg for little-endian.
9795 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
9796 low/high where such is correct only for be.
9797 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
9798 little-endian for -mcall-aixdesc.
9799
87f73374
AM
98002013-04-25 Alan Modra <amodra@gmail.com>
9801
9802 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9803 replace_equiv_address_nv.
9804
cabf91cd
AM
98052013-04-25 Alan Modra <amodra@gmail.com>
9806
9807 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
9808
0ae24cc8
VM
98092013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9810
9811 Revert:
9812 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9813 * rtl.h (struct rtx_def): ...
cabf91cd 9814
77bce07c
VM
98152013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9816
9817 PR rtl-optimizations/57046
9818 * lra-constraints (split_reg): Set up lra_risky_transformations_p
9819 for multi-reg splits.
9820
0db63e7f
L
98212013-04-24 H.J. Lu <hongjiu.lu@intel.com>
9822
9823 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
9824
3362b6b6
SA
98252013-04-24 Sterling Augustine <saugustine@google.com>
9826
9827 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
9828 (comp_dir_string, debug_str_dwo_section): New.
9829 (DEBUG_STR_DWO_SECTION): Rename to ...
9830 (DEBUG_DWO_STR_SECTION): ... this.
9831 (DEBUG_NORM_STR_SECTION): Delete.
9832 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
9833 (DEBUG_STR_DWO_SECTION_FLAGS): New.
9834 (find_AT_string): Move most logic to ...
9835 (find_AT_string_in_table): ... here. New.
9836 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
9837 add_skeleton_AT_string. Delete logic.
9838 (output_skeleton_debug_sections): Remove call to
9839 add_top_level_skeleton_die_attrs.
9840 (add_comp_dir_attribute): Move logic to comp_dir_string.
9841 (dwarf2out_init): Initialize debug_str_dwo_section.
9842 (output_indirect_string): Call find_string_form.
9843 (output_indirect_strings): Rewrite.
9844 (prune_unused_types): Empty skeleton_debug_str_hash.
9845 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
9846 (dwarf2out_finish): Call output_indirect_strings.
9847
e93e18e9
PC
98482013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9849
9850 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
9851
f6ce35ac
VM
98522013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9853
cabf91cd 9854 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM
9855 (LRA_SUBREG_P): New macro.
9856 * recog.c (register_operand): Check LRA_SUBREG_P.
9857 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9858 stack frame.
9859 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9860 (lra_final_code_change): Skip subreg change for operators.
9861 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9862 if there are no operand changes.
9863 * lra-constraints.c (curr_insn_set): New.
9864 (match_reload): Set LRA_SUBREG_P.
9865 (emit_spill_move): Ditto.
9866 (check_and_process_move): Use curr_insn_set. Process only single
9867 set insns. Don't initialize sec_mem_p and change_p.
9868 (simplify_operand_subreg): Use LRA_SUBREG_P.
9869 (reg_in_class_p): New function.
9870 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9871 of #ifdef. Add code to remove cycling.
9872 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9873 non-null disp. Reload inner instead of disp when base and index
9874 are null. Try to put lo_sum into register.
9875 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 9876 (check_and_process_move): Move code for move cost check to
f6ce35ac 9877 simple_move_p. Remove equiv_substitution.
cabf91cd 9878 (simple_move_p): New function.
f6ce35ac
VM
9879 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9880 curr_insn_set. Call check_and_process_move only for single set
9881 insns. Use the new function. Move call of check_and_process_move
9882 after operand equiv substitution and address process.
9883
38047d90
JG
98842013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
9885
9886 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
9887 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
9888 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
9889
13f39b2e
PC
98902013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9891
9892 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
9893
97c116dc
MP
98942013-04-24 Marek Polacek <polacek@redhat.com>
9895
9896 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
9897 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
9898 (select_loops_exit_conditions): Likewise.
9899 (number_of_iterations_for_all_loops): Likewise.
9900 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
9901 (scev_analysis): Likewise.
9902
83082391
CM
99032013-04-02 Catherine Moore <clm@codesourcery.com>
9904 Chao-ying Fu <fu@mips.com>
9905
cabf91cd
AM
9906 * config/mips/micromips.md (jraddiusp): New pattern.
9907 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
9908 instruction if possible.
83082391 9909
19e34aa2
AM
99102013-04-24 Alan Modra <amodra@gmail.com>
9911
9912 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
9913
fdb6603c
JB
99142013-04-24 Julian Brown <julian@codesourcery.com>
9915 Chung-Lin Tang <cltang@codesourcery.com>
9916
9917 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB
9918 dependency behavior in enumeration type DIE generation. Add TODO note
9919 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 9920
bf190e8d
LC
99212013-04-23 Lawrence Crowl <crowl@google.com>
9922
9923 * Makefile.in: Update as needed below.
9924
9925 * hash-table.h (class hash_table):
9926 Correct many methods with parameter types compare_type to the correct
9927 value_type. (Correct code was unlikely to notice the change.)
9928 (hash_table::elements_with_deleted) New.
9929 (class hashtable::iterator): New.
9930 (hashtable::begin()): New.
9931 (hashtable::end()): New.
9932 (FOR_EACH_HASH_TABLE_ELEMENT): New.
9933
9934 * statistics.c (statistics_hashes):
9935 Change type to hash_table. Update dependent calls and types.
9936
9937 * tree-into-ssa.c (var_infos):
9938 Change type to hash_table. Update dependent calls and types.
9939
9940 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
9941 Change type to hash_table. Update dependent calls and types.
9942
9943 * tree-ssa-loop-im.c (struct mem_ref.refs):
9944 Change type to hash_table. Update dependent calls and types.
9945
9946 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
9947 Change type to hash_table. Update dependent calls and types.
9948
9949 * tree-ssa-sccvn.c (vn_tables_s::nary):
9950 Change type to hash_table. Update dependent calls and types.
9951 (vn_tables_s::phis): Likewise.
9952 (vn_tables_s::references): Likewise.
9953
9954 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
9955 (vn_reference_eq): Update parameter and return types.
9956
9957 * tree-ssa-structalias.c (pointer_equiv_class_table):
9958 Change type to hash_table. Update dependent calls and types.
9959 (location_equiv_class_table): Likewise.
9960
9961 * tree-vect-data-refs.c: Consequential changes for making
9962 peeling a hash_table.
9963
9964 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
9965 (destroy_loop_vec_info): Dependent hash_table update.
9966
9967 * tree-vectorizer.h (peeling_htab):
9968 Change type to hash_table. Update dependent calls and types.
9969
d70a81dd
SC
99702013-04-23 Shiva Chen <shiva0217@gmail.com>
9971
cabf91cd
AM
9972 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
9973 to check the register content is equal or not.
9974 * lra-constraints.c (match_reload): Use lra_assign_reg_val
9975 to assign register content record.
9976 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 9977 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM
9978 * lra-int.h (struct lra_reg): Add offset member.
9979 (lra_reg_val_equal_p): New static inline function.
9980 (lra_update_reg_val_offset): New static inline function.
9981 (lra_assign_reg_val): New static inline function.
9982 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
9983 to assign register content record.
9984 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 9985
b894a1f3
CM
99862013-04-23 Catherine Moore <clm@codesourcery.com>
9987
9988 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
9989 operands. Record compression.
9990
ecd14de9
XDL
99912013-04-23 Xinliang David Li <davidxl@google.com>
9992
9993 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
9994
92e776e9
RB
99952013-04-23 Richard Biener <rguenther@suse.de>
9996
9997 PR middle-end/57036
9998 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
9999 parameter, only add abnormal goto edges from the copied body
10000 if the call could perform abnormal gotos.
10001 (copy_cfg_body): Adjust.
10002
a15ee567
SN
100032013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10004
10005 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10006
08c52234
AS
100072013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10008
10009 * coretypes.h (gimple_stmt_iterator): Add struct to make
10010 compatible with C.
10011
999c1171
RB
100122013-04-23 Richard Biener <rguenther@suse.de>
10013
10014 PR tree-optimization/57026
10015 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10016 from SSA names occuring in abnormal PHI nodes.
10017
53669259
AK
100182013-04-22 Andi Kleen <ak@linux.intel.com>
10019
10020 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10021
1a0ad150
AK
100222013-04-22 Andi Kleen <ak@linux.intel.com>
10023
10024 * lto/lto.c (print_lto_report_1): Declare early.
10025 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10026
057f8f20
AK
100272013-04-22 Andi Kleen <ak@linux.intel.com>
10028
10029 * common.opt (-flto-report-wpa): Add.
10030 * doc/invoke.texi (-flto-report-wpa): Add.
10031 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10032 (lto_main): dito.
10033
473b1e05
XDL
100342013-04-22 Xinliang David Li <davidxl@google.com>
10035
10036 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10037 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10038 * Makefile.in: New dependency
10039
cabf91cd 10040 David Daney <ddaney.cavm@gmail.com>
b1485a33 10041
cabf91cd
AM
10042 * configure.ac (gcc_cv_as_micromips_support): Use the
10043 --fatal-warnings option.
10044 * configure: Regenerate.
b1485a33 10045
829d0168
MP
100462013-04-22 Marek Polacek <polacek@redhat.com>
10047
10048 PR sanitizer/56990
10049 * tsan.c (instrument_expr): Don't instrument expression
10050 in case its size is zero.
10051
6d9b7208
UB
100522013-04-22 Uros Bizjak <ubizjak@gmail.com>
10053
10054 PR target/57032
10055 Revert:
10056 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10057
10058 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10059
ea679d55
JG
100602013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10061
10062 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10063 (gimple_stmt_iterator): New typedef.
10064 * gimple.h (gimple_stmt_iterator): Rename to...
10065 (gimple_stmt_iterator_d): ... This.
10066 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10067 trees be valid for GIMPLE and GENERIC.
10068 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10069 * gimple-fold.c (gimple_fold_call): Call target hook
10070 gimple_fold_builtin.
10071 * hooks.c (hook_bool_gsiptr_false): New.
10072 * hooks.h (hook_bool_gsiptr_false): New.
10073 * target.def (fold_stmt): New.
10074 * doc/tm.texi: Regenerate.
10075
88a581da
VM
100762013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10077
10078 PR target/57018
10079 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10080 a set sp if no stack realignment.
10081
92be22dc
NC
100822013-04-22 Nick Clifton <nickc@redhat.com>
10083
10084 * config.gcc (tilegx-linux): Extend extra_objs rather than
10085 overwriting it.
10086 (tilepro-linux): Likewise.
10087
0ddec79f
JG
100882013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10089
10090 * config/aarch64/aarch64-builtins.c
10091 (CF): Remove.
10092 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10093 (VAR<1-12>): Add MAP parameter.
10094 (BUILTIN_*): Likewise.
10095 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10096 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10097 (aarch64_ushl_n<mode>): Likewise.
10098 (aarch64_sshr_n<mode>): Likewise.
10099 (aarch64_ushr_n<mode>): Likewise.
10100 (aarch64_<maxmin><mode>): Likewise.
10101 (aarch64_sqrt<mode>): Likewise.
10102 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10103 (vshr<q>_n_*): Likewise.
10104
0050faf8
JG
101052013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10106
10107 * config/aarch64/aarch64-builtins.c
10108 (aarch64_simd_builtin_type_mode): Handle SF types.
10109 (sf_UP): Define.
10110 (BUILTIN_GPF): Define.
10111 (aarch64_init_simd_builtins): Handle SF types.
10112 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10113 (frecps): Likewise.
10114 (frecpx): Likewise.
10115 * config/aarch64/aarch64-simd.md
10116 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10117 (aarch64_frecpe<mode>): New.
10118 (aarch64_frecps<mode>): Likewise.
10119 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10120 (v8type): Add frecp<esx>.
10121 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10122 (aarch64_frecps<mode>): Likewise.
10123 * config/aarch64/iterators.md (FRECP): New.
10124 (frecp_suffix): Likewise.
10125 * config/aarch64/arm_neon.h
10126 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10127
0fad3dbc 101282013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB
10129
10130 PR target/56995
10131 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10132 (REG_CLASS_NAMES): Idem.
10133 (REG_CLASS_CONTENTS): Idem.
10134 (REGCLASS_HAS_FP_REG): Idem.
10135 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10136 (sh_conditional_register_usage): Idem.
10137
3e8a33f9
JL
101382013-04-21 Jeff Law <law@redhat.com>
10139
10140 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10141 (ssa_forward_propagate_and_combine): Use it.
10142
f38e1b0a
VM
101432013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10144
10145 * lra.c: Update the flow chart diagram.
10146
682303da
VM
101472013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10148
10149 PR rtl-optimization/56847
10150 * lra-constraints.c (process_alt_operands): Discourage alternative
10151 with non-matche doffsettable memory constraint fro memory with
10152 known offset.
10153
f6b64c35
RB
101542013-04-19 Richard Biener <rguenther@suse.de>
10155
10156 PR tree-optimization/56982
10157 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10158 function.
10159 * gimplify.c (gimplify_call_expr): Notice special calls.
10160 (gimplify_modify_expr): Likewise.
10161 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10162 abnormal control flow receivers.
10163 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10164 in the same way as cfun->has_nonlocal_labels.
10165 (gimple_purge_dead_abnormal_call_edges): Likewise.
10166 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10167 receivers start a basic-block.
10168
01d8bf07
RB
101692013-04-19 Richard Biener <rguenther@suse.de>
10170
10171 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10172 member ...
10173 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10174 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10175 (SLP_TREE_LOAD_PERMUTATION): Add.
10176 (vect_transform_slp_perm_load): Adjust prototype.
10177 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10178 (vect_free_slp_instance): Likewise.
10179 (vect_create_new_slp_node): Likewise.
10180 (vect_supported_slp_permutation_p): Remove.
10181 (vect_slp_rearrange_stmts): Adjust.
10182 (vect_supported_load_permutation_p): Likewise. Inline
10183 vect_supported_slp_permutation_p here.
10184 (vect_analyze_slp_instance): Compute load permutations per
10185 slp node instead of per instance.
10186 (vect_get_slp_defs): Adjust.
10187 (vect_transform_slp_perm_load): Likewise.
10188 (vect_schedule_slp_instance): Remove redundant code.
10189 (vect_schedule_slp): Remove hack for PR56270, add it ...
10190 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10191 CSE loads for SLP. Adjust.
10192
ede22fc3
GY
101932013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10194
10195 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10196 spelling in two comments.
10197
67bc84fb
GY
101982013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10199
10200 PR target/56797
10201 * config/arm/arm.c (load_multiple_sequence): Require SP
10202 as base register for loads if SP is in the register list.
10203
e248d83f
MJ
102042013-04-19 Martin Jambor <mjambor@suse.cz>
10205
10206 PR tree-optimization/56718
10207 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10208 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10209 and made public. Adjusted all callers.
10210 (ipa_intraprocedural_devirtualization): New function.
10211 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10212 (ipa_intraprocedural_devirtualization): Likewise.
10213 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10214
4891e8f8
RB
102152013-04-19 Richard Biener <rguenther@suse.de>
10216
10217 PR tree-optimization/57000
10218 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10219
dad89f7c
TG
102202013-04-19 Terry Guo <terry.guo@arm.com>
10221
10222 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10223 Replace with ...
10224 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10225 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10226 (cortex_m4_fmacs): Use new reservations.
10227 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10228
72ea0d47
VM
102292013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10230
f1e6512c 10231 PR rtl-optimization/56999
72ea0d47
VM
10232 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10233 related code.
10234 (lra_coalesce): Remove split_origin_bitmap and related code.
10235 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10236 ranges if necessary.
10237
780a5b71
UB
102382013-04-18 Uros Bizjak <ubizjak@gmail.com>
10239
10240 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10241 New array.
10242 (ix86_expand_call): Remove clobbered_registers array and use
10243 x86_64_ms_sysv_extra_clobbered_registers instead.
10244 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10245 Declare here.
10246 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10247 predicate.
10248 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10249 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10250 (*call_value_rex64_ms_sysv): Ditto.
10251
6f5a366a
CC
102522013-04-18 Cary Coutant <ccoutant@google.com>
10253
10254 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10255 parent instead of die_mark.
10256
475b8f37
DN
102572013-04-18 Diego Novillo <dnovillo@google.com>
10258
10259 * gimple.c (create_gimple_tmp): New.
10260 (get_expr_type): New.
10261 (build_assign): New.
10262 (build_type_cast): New.
10263 * gimple.h (enum ssa_mode): Define.
10264 (gimple_seq_set_location): New.
10265 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 10266 to use build_assign and build_type_cast.
475b8f37 10267
08940f33
RB
102682013-04-18 Richard Biener <rguenther@suse.de>
10269
10270 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10271 handle negative step. Remove redundant checks.
10272 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10273 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10274 for negative step and grouped loads fail to vectorize.
10275
0e0f87d4
SB
102762013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10277
10278 * emit-rtl.c (reset_insn_used_flags): New function.
10279 (reset_all_used_flags): Use it.
10280 (verify_insn_sharing): New function.
10281 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10282
4c445590
JJ
102832013-04-18 Jakub Jelinek <jakub@redhat.com>
10284
10285 PR tree-optimization/56984
10286 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10287 and (x >> M) >= N don't register any assertion if N << M is the
10288 minimum value.
10289
6873ecab
SB
102902013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10291
10292 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10293 do not delete_insn insns that have not yet been emitted, only
10294 unlink them with remove_insn.
10295 * df-scan.c (df_insn_delete): Revert r197492.
10296
3ccb989e
SB
102972013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10298
10299 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10300 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10301
103022013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY
10303
10304 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10305 define_insn_and_split.
10306 (and_scc,ior_scc,negscc): Likewise.
10307 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10308
3ccb989e 103092013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY
10310
10311 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10312 can benefit from using a sequence of LDRD instructions in epilogue
10313 instead of a single LDM instruction.
10314
6d10a203
MLI
103152013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10316
10317 PR 45688
10318 * doc/extend.texi: Fix typo.
10319
6983e6b5
RB
103202013-04-17 Richard Biener <rguenther@suse.de>
10321
10322 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10323 (vect_build_slp_tree): ... here.
10324 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10325 match. Remove special-casing of mismatched complex loads.
10326 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10327 re-try the match with swapped commutative operands.
10328 (vect_supported_load_permutation_p): Remove special-casing of
10329 mismatched complex loads.
10330 (vect_analyze_slp_instance): Adjust.
10331
ef23e6a2
RB
103322013-04-17 Richard Biener <rguenther@suse.de>
10333
10334 PR rtl-optimization/56921
10335 * cfgloop.h (struct loop): Add simple_loop_desc member.
10336 (struct niter_desc): Mark with GTY(()).
10337 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10338 * loop-iv.c (get_simple_loop_desc): Likewise.
10339 (free_simple_loop_desc): Likewise.
10340
10341 Revert
10342 2013-04-16 Richard Biener <rguenther@suse.de>
10343
10344 PR rtl-optimization/56921
10345 * loop-init.c (pass_rtl_move_loop_invariants): Add
10346 TODO_do_not_ggc_collect to todo_flags_finish.
10347 (pass_rtl_unswitch): Same.
10348 (pass_rtl_unroll_and_peel_loops): Same.
10349 (pass_rtl_doloop): Same.
10350
fc6f94f5
EB
103512013-04-17 Eric Botcazou <ebotcazou@adacore.com>
10352
10353 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
10354 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
10355 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
10356 references.
10357 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
10358 * tree-streamer.c (record_common_node): Adjust reference in comment.
10359
10a88311
TG
103602013-04-17 Terry Guo <terry.guo@arm.com>
10361
10362 * config/arm/cortex-m4.md: Add a new bypass.
10363
6d9b7208 103642013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N
10365
10366 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
10367 New pattern.
10368 (*subs_<optab><mode>_multp2): New pattern.
10369 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10370 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10371
6d9b7208 103722013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N
10373
10374 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
10375 (*subs_mul_imm_<mode>): New pattern.
10376
18a6701e
DE
103772013-04-16 David Edelsohn <dje.gcc@gmail.com>
10378
10379 PR target/56948
10380 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
10381 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
10382 (vsx_movti_32bit): Change j->wa to O->wa.
10383
07c37b2f
RB
103842013-04-16 Richard Biener <rguenther@suse.de>
10385
10386 PR rtl-optimization/56921
10387 * loop-init.c (pass_rtl_move_loop_invariants): Add
10388 TODO_do_not_ggc_collect to todo_flags_finish.
10389 (pass_rtl_unswitch): Same.
10390 (pass_rtl_unroll_and_peel_loops): Same.
10391 (pass_rtl_doloop): Same.
10392
0e0f87d4 103932013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY
10394
10395 * config/arm/arm.c (emit_multi_reg_push): New declaration
10396 for an existing function.
10397 (arm_emit_strd_push): New function.
10398 (arm_expand_prologue): Used here.
10399 (arm_emit_ldrd_pop): New function.
10400 (arm_expand_epilogue): Used here.
10401 (arm_get_frame_offsets): Update condition.
10402 (arm_emit_multi_reg_pop): Add a special case for load of a single
10403 register with writeback.
10404
5e8e2af4
UB
104052013-04-16 Uros Bizjak <ubizjak@gmail.com>
10406
10407 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
10408 description.
10409
9fd9ccf7
RB
104102013-04-16 Richard Biener <rguenther@suse.de>
10411
10412 PR tree-optimization/56756
10413 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
10414 (first_mem_ref_loc): New.
10415 (execute_sm): Place the load temporarily before a previous
10416 access instead of in the latch edge to ensure its SSA dependencies
10417 are defined at points dominating the load.
10418
96fba521
SB
104192013-04-16 Steven Bosscher <steven@gcc.gnu.org>
10420
4c8af858
SB
10421 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
10422 correct fix by moving header and footer insn to the footer of
10423 the merged basic block. Clear BB_END of the merged-away block.
10424
96fba521
SB
10425 PR middle-end/43631
10426 * emit-rtl.c (make_note_raw): New function.
10427 (link_insn_into_chain): New static inline function.
10428 (add_insn): Use it.
10429 (add_insn_before, add_insn_after): Factor insn chain linking code...
10430 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
10431 using link_insn_into_chain.
10432 (note_outside_basic_block_p): New helper function for emit_note_after
10433 and emit_note_before.
10434 (emit_note_after): Use nobb variant of add_insn_after if the note
10435 should not be contained in a basic block.
10436 (emit_note_before): Use nobb variant of add_insn_before if the note
10437 should not be contained in a basic block.
10438 (emit_note_copy): Use make_note_raw.
10439 (emit_note): Likewise.
10440 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
10441 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
10442 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
10443 the moved barrier the tail of the basic block it follows.
10444 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
10445
7b8265ba
JJ
104462013-04-15 Jakub Jelinek <jakub@redhat.com>
10447
10448 PR tree-optimization/56962
10449 * gimple-ssa-strength-reduction.c (record_increment): Only set
10450 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
10451 either rhs1 or rhs2 is equal to c->base_expr.
10452
5185d248
RB
104532013-04-15 Richard Biener <rguenther@suse.de>
10454
10455 PR tree-optimization/56933
10456 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
10457 member.
10458 (GROUP_READ_WRITE_DEPENDENCE): Remove.
10459 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
10460 * tree-vect-data-refs.c (vect_analyze_group_access): Move
10461 dependence check ...
10462 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
10463 ... here.
10464 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
10465 GROUP_READ_WRITE_DEPENDENCE.
10466
a24243a0
AK
104672013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10468
10469 * emit-rtl.c (reset_all_used_flags): New function.
10470 (verify_rtl_sharing): Call reset_all_used_flags before and after
10471 performing the checks.
10472
1c50eada
KT
104732013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10474
10475 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
10476 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
10477 * config/arm/constraints.md (De): New constraint.
10478 * config/arm/neon.md (anddi3_neon): Delete.
10479 (neon_vand<mode>): Expand to standard anddi3 pattern.
10480 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
10481 Move earlier in the file.
10482 (neon_inv_logic_op2): Likewise.
10483 (arm_anddi_operand_neon): New predicate.
10484
e927b6ad
RO
104852013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10486
10487 * configure.ac (gcc_cv_ld_as_needed): Set
10488 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
10489 Use -z ignore, -z record on *-*-solaris2*.
10490 (HAVE_LD_AS_NEEDED): Update comment.
10491 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
10492 * configure: Regenerate.
10493 * config.in: Regenerate.
10494 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
10495 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
10496 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
10497 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
10498 equivalents. Fix markup.
10499 * doc/tm.texi: Regenerate.
10500
e0ea8797
AH
105012013-04-15 Andrew Hsieh <andrewhsieh.google.com>
10502
10503 * config/i386/i386.opt: New option mstack-protector-guard=.
10504 * config/i386/i386-opts.h: Add enum stack_protector_guard.
10505 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
10506 TARGET_SSP_TLS_GUARD.
10507 * config/i386/i386.c (ix86_option_override_internal): Set
10508 ix86_stack_protector_guard.
10509 * config/i386/i386.md (stack_protect_set): Enable for
10510 TARGET_SSP_TLS_GUARD only.
10511 (stack_protect_set_<mode>): Ditto.
10512 (stack_protect_test): Ditto.
10513 (stack_protect_test_<mode>): Ditto.
10514 * doc/invoke.texi (i386 Option): Document.
10515
811b72f9
EB
105162013-04-15 Eric Botcazou <ebotcazou@adacore.com>
10517
10518 PR target/56890
10519 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
10520 (S_MODES): Set H_MODE bit.
10521 (SF_MODES): Set only S_MODE and SF_MODE bits.
10522 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
10523 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
10524 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
10525 <MODE_FLOAT>: Likewise.
10526
5529fdd6
JY
105272013-04-15 Joey Ye <joey.ye@arm.com>
10528
10529 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
10530
517b1da2
JY
105312013-04-15 Joey Ye <joey.ye@arm.com>
10532
10533 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
10534 for real far jump.
10535 (thumb_far_jump_used_p): Count instruction size and set
10536 far_jump_used.
10537
01007ae0
EB
105382013-04-14 Eric Botcazou <ebotcazou@adacore.com>
10539
10540 * reorg.c (fill_simple_delay_slots): Reindent block of code.
10541 * resource.c (mark_target_live_regs): Reformat conditional block.
10542
c46f6580
SB
105432013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10544
10545 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
10546 notes, they are emitted only just before final.
10547 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
10548
1f397f45
SB
105492013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10550
10551 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
10552 * cfgrtl.c (delete_insn): Call it here instead.
10553 * lra-spills.c (lra_final_code_change): Use delete_insn.
10554 * haifa-sched.c (sched_remove_insn): Likewise.
10555 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
10556 returning to the nop pool.
10557 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
10558 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
10559
58a51369
SB
105602013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10561
10562 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
10563 * doc/tm.texi: Regenerated.
10564
33159866
UB
105652013-04-12 Uros Bizjak <ubizjak@gmail.com>
10566
10567 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
10568 QImode checks.
10569
226e378f
SB
105702013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10571
10572 * df-core.c (df_find_def): Compare register numbers.
10573 (df_find_use): Likewise.
10574
fafb9b18
VM
105752013-04-12 Vladimir Makarov <vmakarov@redhat.com>
10576
10577 PR target/56903
10578 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
10579 lra_in_progress for return.
10580
9a946fd6
GY
105812013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10582
10583 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
10584 define_insn into define_insn_and_split and emit movsicc patterns.
10585
41b83758
GY
105862013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10587
10588 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
10589
d3afd9aa
RB
105902013-04-12 Richard Biener <rguenther@suse.de>
10591
10592 * tree-pass.h (TODO_do_not_ggc_collect): New.
10593 * passes.c (execute_one_ipa_transform_pass): Honor
10594 TODO_do_not_ggc_collect.
10595 (execute_one_pass): Likewise.
10596
10597 Revert
10598 2013-04-10 Richard Biener <rguenther@suse.de>
10599
10600 * passes.c (init_optimization_passes): Remove reload pass.
10601 * ira.c (do_reload): Merge into ...
10602 (ira): ... this.
10603 (rest_of_handle_reload): Remove.
10604 (pass_reload): Likewise.
10605 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10606 of reload for vzeroupper pass placement.
10607
06f9b387
JJ
106082013-04-12 Jakub Jelinek <jakub@redhat.com>
10609
10610 PR tree-optimization/56918
10611 PR tree-optimization/56920
10612 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
10613 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
10614 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
10615 use wide_mul_with_sign method.
10616
953094d2
RB
106172013-04-12 Richard Biener <rguenther@suse.de>
10618
10619 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
10620 not be considered a gimple constant.
10621
bb506982
MG
106222013-04-12 Marc Glisse <marc.glisse@inria.fr>
10623
10624 * fold-const.c (const_binop): Handle vector shifts by a scalar.
10625 (fold_binary_loc): Call const_binop also for mixed vector-scalar
10626 operations.
10627
4b84d650
JJ
106282013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
10629 Jakub Jelinek <jakub@redhat.com>
10630
10631 * opts.c: Include diagnostic-color.h.
10632 (common_handle_option): Handle OPT_fdiagnostics_color_.
10633 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
10634 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
10635 (diagnostic-color.o): New.
10636 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
10637 (diagnostic_color_rule): New enum.
10638 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
10639 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
10640 the location string.
10641 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
10642 either NULL, or color kind.
10643 * diagnostic-color.c: New file.
10644 * diagnostic-color.h: New file.
10645 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
10646 arguments.
10647 * doc/invoke.texi (-fdiagnostics-color): Document.
10648 * pretty-print.h (pp_show_color): Define.
10649 (struct pretty_print_info): Add show_color field.
10650 * diagnostic.c: Include diagnostic-color.h.
10651 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
10652 macros. Colorize error:, warning: etc. strings and also the location
10653 string.
10654 (diagnostic_show_locus): Colorize the caret line.
10655 * pretty-print.c: Include diagnostic-color.h.
10656 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
10657 inside of %< %> quotes or quoted through q format modifier.
10658
067a1e71
AK
106592013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10660
33159866 10661 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 10662
33159866 106632013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N
10664
10665 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
10666 code in CC_NZ mode.
10667 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
10668 pattern.
10669
7b55f98f
MP
106702013-04-11 Marek Polacek <polacek@redhat.com>
10671
10672 PR tree-optimization/48184
33159866 10673 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 10674
966b587e
EB
106752013-04-11 Eric Botcazou <ebotcazou@adacore.com>
10676
10677 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
10678 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
10679 (skip_simple_arithmetic): Tidy up.
10680 * tree.h (skip_simple_constant_arithmetic): Declare.
10681
33159866 106822013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N
10683
10684 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
10685
1d42c1ec
RB
106862013-04-11 Richard Biener <rguenther@suse.de>
10687
10688 * tree-vect-loop.c (get_initial_def_for_induction): Properly
10689 generate vector constants.
10690
4ba5ea11
RB
106912013-04-11 Richard Biener <rguenther@suse.de>
10692
10693 PR tree-optimization/56878
10694 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
10695 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
10696 New function.
10697 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
10698 Prefer to align the DR with the most invariant base address.
10699
f0defe58
SKS
107002013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10701
10702 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
10703 comment.
10704
d07458be
JG
107052013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
10706
10707 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
10708 floating-point vector comparisons against 0.
10709
146b8692
JJ
107102013-04-11 Jakub Jelinek <jakub@redhat.com>
10711
10712 PR tree-optimization/56899
10713 * fold-const.c (extract_muldiv_1): Apply distributive law
10714 only if TYPE_OVERFLOW_WRAPS (ctype).
10715
b8578ff7
BC
107162013-04-11 Bin Cheng <bin.cheng@arm.com>
10717
10718 PR target/56124
10719 * ira-costs.c (scan_one_insn): Check whether the source rtx of
10720 loading has side effect.
10721
0ea8a6f9
SB
107222013-04-10 Steven Bosscher <steven@gcc.gnu.org>
10723
10724 * config/sparc/sparc.c: Include tree-pass.h.
10725 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
10726 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
10727 head of file. Change return type. Split off gate function.
10728 (sparc_gate_work_around_errata): New function.
10729 (pass_work_around_errata): New pass definition.
10730 (insert_pass_work_around_errata) New pass insert definition to
10731 insert pass_work_around_errata just after delayed-branch scheduling.
10732 (sparc_option_override): Insert the pass.
10733 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
10734
42e37616
DM
107352013-04-10 David S. Miller <davem@davemloft.net>
10736
89deeb3b
DM
10737 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
10738 or -mcpu=hypersparc.
10739
42e37616
DM
10740 * target.def (cstore_mode): New hook.
10741 * target.h: Include insn-codes.h
10742 * targhooks.c: Likewise.
10743 (default_cstore_mode): New function.
10744 * targhooks.h: Declare it.
10745 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
10746 * doc/tm.texi: Rebuild.
10747 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
10748 target hook, rather than inspecting the insn_data.
10749 * config/sparc/sparc.c (sparc_cstore_mode): New function.
10750 (TARGET_CSTORE_MODE): Redefine.
10751 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
10752 result patterns.
10753 * config/sparc/predicates.md (cstore_result_operand): New special
10754 predicate.
10755 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
10756 Use it for operand 0.
10757 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
10758 (*snesi_special): Likewise.
10759 (*snesi_zero): Likewise.
10760 (*seqsi_zero): Likewise.
10761 (*sltu_insn): Likewise.
10762 (*sgeu_insn): Likewise.
10763 (*seqdi_special): Make operand 0 and comparison operation be of
10764 DImode.
10765 (*snedi_special): Likewise.
10766 (*snedi_special_vis3): Likewise.
10767 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
10768 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
10769 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
10770 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
10771 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
10772 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
10773 (*sltu_extend_sp64): Likewise.
10774 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
10775 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
10776 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
10777 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
10778 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
10779
95ca411e
YZ
107802013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
10781
10782 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
10783 (aarch64_start_file): Use the new function.
10784
6782438d 107852013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 10786 Jason Merrill <jason@redhat.com>
6782438d
SKS
10787
10788 * common.opt: Add -gdwarf.
10789 * opts.c (common_handle_option): Handle it.
10790 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
10791
bb313b93
RB
107922013-04-10 Richard Biener <rguenther@suse.de>
10793
10794 * passes.c (execute_todo): Do not call ggc_collect conditional here.
10795 (execute_one_ipa_transform_pass): But unconditionally here.
10796 (execute_one_pass): And here.
10797 (init_optimization_passes): Remove reload pass.
10798 * tree-pass.h (TODO_ggc_collect): Remove.
10799 (pass_reload): Likewise.
10800 * ira.c (do_reload): Merge into ...
10801 (ira): ... this.
10802 (rest_of_handle_reload): Remove.
10803 (pass_reload): Likewise.
10804 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10805 of reload for vzeroupper pass placement.
10806 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
10807 and todo_flags_finish of all passes.
10808
793d9a16
RB
108092013-04-10 Richard Biener <rguenther@suse.de>
10810
10811 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
10812 first_const_oprnd field, rename first_def_type to first_op_type.
10813 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
10814 (vect_get_and_check_slp_defs): Always use the type of the
10815 operand. Allow mixed vect_external_def, vect_constant_def types.
10816 (vect_get_constant_vectors): Handle mixed vect_external_def,
10817 vect_constant_def types.
10818
ee8a9b7b
JR
108192013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
10820
10821 PR tree-optimization/55524
10822 * tree-ssa-math-opts.c
10823 (convert_mult_to_fma): Don't use an fms construct
10824 when we don't have an fms operation, but fnma, and it looks
10825 likely that we'll be able to use the latter.
10826
6957a6f6
ZZ
108272013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
10828
10829 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
10830 function.
10831 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
10832 inline fail caused by overwritable functions.
10833
34ab4a5b
CJW
108342013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
10835
10836 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
10837 unnecessary bits in the constant power of two case.
10838
abf9bfbc
RB
108392013-04-10 Richard Biener <rguenther@suse.de>
10840
10841 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
10842 broken code swapping operands.
10843 (vect_build_slp_tree): Do not compute load permutations here.
10844 (vect_analyze_slp_instance): Compute load permutations here,
10845 after building the SLP tree.
10846
f408477e
CB
108472013-04-09 Christian Bruel <christian.bruel@st.com>
10848
10849 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
10850 of next/prev_real_insn.
10851
5ac42672
JH
108522013-04-09 Jan Hubicka <jh@suse.cz>
10853
abf9bfbc
RB
10854 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10855 Drop aliased parameter.
5ac42672
JH
10856 (function_and_variable_visibility): Do not handle alias pairs.
10857 * cgraph.c (varpool_externally_visible_p): Update prototype.
10858 * varpool.c (varpool_add_new_variable): Update.
10859
5017f1d2
KT
108602013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10861
10862 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
10863
48eecbee
SB
108642013-04-09 Steven Bosscher <steven@gcc.gnu.org>
10865
a949cf1c
SB
10866 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
10867
48eecbee
SB
10868 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
10869
75ef8e3d
MP
108702013-04-09 Marek Polacek <polacek@redhat.com>
10871
10872 PR tree-optimization/48762
33159866 10873 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 10874
23847df4
RB
108752013-04-09 Richard Biener <rguenther@suse.de>
10876
10877 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
10878 dealing with cost.
10879 (vect_build_slp_tree): Likewise.
10880 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
10881 calculating the cost of a SLP instance.
10882 (vect_analyze_slp_instance): Use it from here, after building
10883 the SLP tree.
10884
ea3a0fde
JJ
108852013-04-09 Jakub Jelinek <jakub@redhat.com>
10886
10887 PR middle-end/56883
10888 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
10889 expand_omp_for_static_chunk): Use simple_p = true in
10890 force_gimple_operand_gsi calls when assigning to addressable decls.
10891
a32dfe9d
JL
108922013-04-09 Jeff Law <law@redhat.com>
10893
10894 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
10895 when the boolean was created by converting a wider object which
10896 had a boolean range.
10897
d755c7ef
RB
108982013-04-09 Richard Biener <rguenther@suse.de>
10899
10900 * tree-vectorizer.h (slp_void_p): Remove.
10901 (slp_tree): Typedef before _slp_tree declaration.
10902 (struct _slp_tree): Use a vector of slp_tree as children.
10903 (vect_get_place_in_interleaving_chain): Remove.
10904 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
10905 Move ...
10906 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
10907 and make static.
10908 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
10909 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
10910 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
10911 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
10912 Use slp_node instead of slp_void_p and adjust.
10913
3d741091
RB
109142013-04-09 Richard Biener <rguenther@suse.de>
10915
10916 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
10917 work that is not necessary.
10918
39307ba7
JJ
109192013-04-09 Jakub Jelinek <jakub@redhat.com>
10920
10921 PR tree-optimization/56854
10922 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
10923 forward into clobber stmts if it would change MEM_REF lhs into
10924 non-MEM_REF.
10925
343881fd
MK
109262013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
10927
10928 * tree.c (type_hash_lookup, type_hash_add): Make static.
10929 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
10930
3922658a
RB
109312013-04-09 Richard Biener <rguenther@suse.de>
10932
10933 * tree.h (unsave_expr_now): Remove.
10934 * tree-inline.c (mark_local_for_remap_r): Remove.
10935 (unsave_expr_1): Likewise.
10936 (unsave_r): Likewise.
10937 (unsave_expr_now): Likewise.
10938 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
10939 (propagate_tree_value): Likewise.
10940
9fb6b620
SB
109412013-04-08 Steven Bosscher <steven@gcc.gnu.org>
10942
10943 * doc/rtl.texi (sequence): Rewrite documentation to match the
10944 current use of SEQUENCE rtl objects.
10945 * rtl.def (SEQUENCE): Likewise.
10946
10947 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
10948 Update documentation.
10949 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
10950 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
10951
10952 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
10953
8ddb5a29
TJ
109542013-04-08 Teresa Johnson <tejohnson@google.com>
10955
10956 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
10957 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 10958 methods.
8ddb5a29
TJ
10959 (estimate_edge_size_and_time): Add comment to suggest using rounding
10960 methods.
10961 (estimate_node_size_and_time): Ditto.
10962 (remap_edge_change_prob): Use helper rounding divide methods.
10963 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
10964 (gimple_mod_pow2_value_transform): Ditto.
10965 (gimple_mod_subtract_transform): Ditto.
10966 (gimple_ic_transform): Ditto.
10967 (gimple_stringops_transform): Ditto.
10968 * stmt.c (conditional_probability): Ditto.
10969 (emit_case_dispatch_table): Ditto.
10970 * lto-cgraph.c (merge_profile_summaries): Ditto.
10971 * tree-optimize.c (execute_fixup_cfg): Ditto.
10972 * cfgcleanup.c (try_forward_edges): Ditto.
10973 * cfgloopmanip.c (scale_loop_profile): Ditto.
10974 (loopify): Ditto.
10975 (duplicate_loop_to_header_edge): Ditto.
10976 (lv_adjust_loop_entry_edge): Ditto.
10977 * tree-vect-loop.c (vect_transform_loop): Ditto.
10978 * profile.c (compute_branch_probabilities): Ditto.
10979 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
10980 * lto-streamer-in.c (input_cfg): Ditto.
10981 * gimple-streamer-in.c (input_bb): Ditto.
10982 * ipa-cp.c (update_profiling_info): Ditto.
10983 (update_specialized_profile): Ditto.
10984 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
10985 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 10986 rounding methods.
8ddb5a29
TJ
10987 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10988 (compute_trg_info): Ditto.
10989 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
10990 (purge_dead_edges): Ditto.
10991 * loop-unswitch.c (unswitch_loop): Ditto.
10992 * cgraphclones.c (cgraph_clone_edge): Ditto.
10993 (cgraph_clone_node): Ditto.
10994 * tree-inline.c (copy_bb): Ditto.
10995 (copy_edges_for_bb): Ditto.
10996 (initialize_cfun): Ditto.
10997 (copy_cfg_body): Ditto.
10998 (expand_call_inline): Ditto.
10999
661e6bd7
KT
110002013-04-08 Kai Tietz <ktietz@redhat.com>
11001
11002 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11003 TARGET_CYGWIN64 by TARGET_64BIT.
11004
105766f3
JR
110052013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11006
11007 * config/epiphany/epiphany.md (GPR_1): New constant.
11008 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11009 * config/epiphany/epiphany.c (gen_compare_reg):
11010 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11011 is already in place.
11012 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11013 Don't require being called during rtl expansion; If y operlaps r0,
11014 return 0.
11015 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11016 (epiphany_expand_epilogue): Likewise.
11017
8afab237
JR
11018 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11019 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 11020 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 11021
fa7707d6
JR
11022 * config/epiphany/constraints.md (CnL): New constraint.
11023 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11024 * config/epiphany/predicates.md (add_operand): Allow 1024.
11025
5138e18d
JR
11026 * config/epiphany/epiphany.md (logical_op): New code iterator.
11027 (op_mnc): New code attribute.
11028 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11029 (mov_f+1, mov_f+2): New peephole2 patterns.
11030
2ccc703d
JR
11031 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11032 (cstoresi4): Also allow re-use of zero result when doing a NE
11033 comparison to a non-zero operand.
aefb0819 11034 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 11035
093ac0a5
JR
11036 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11037 Use gen_addsi3_i / gen_subsi3_i.
11038
f223bb13
JJ
110392013-04-08 Jakub Jelinek <jakub@redhat.com>
11040
11041 PR c++/34949
11042 PR c++/50243
11043 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11044 contain anything but clobbers, at most one __builtin_stack_restore,
11045 optionally debug stmts and final resx, and if it has at least one
11046 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11047 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11048 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11049 which isn't defaut definition, remove them.
11050 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11051 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11052 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11053 with MEM_REF LHS with SSA_NAME address.
11054
4481581f
JL
110552013-04-08 Jeff Law <law@redhat.com>
11056
11057 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 11058
451dabda
RB
110592013-04-08 Richard Biener <rguenther@suse.de>
11060
11061 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11062 extra newline.
11063 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11064 determined vector type.
11065 (vect_analyze_data_refs): Likewise.
11066 (vect_get_new_vect_var): Adjust.
11067 (vect_create_destination_var): Preserve SSA name versions.
11068 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11069 not dump anything here.
11070
3b088b47
JR
110712013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11072
11073 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11074 Add member lr_slot_known.
11075 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11076 if necessary.
11077 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11078 Remove code that sets lr_slot_offset according to what a previous
11079 version of epiphany_emit_save_restore used to do.
11080 (epiphany_emit_save_restore): When doing an lr save or restore,
11081 set/verify lr_slot_known and lr_slot_offset.
11082
d8484d41
XQ
110832013-04-08 Xinyu Qi <xyqi@marvell.com>
11084
33159866
UB
11085 PR target/54338
11086 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ
11087 in ALL_REGS.
11088
481be1c4
RB
110892013-04-08 Richard Biener <rguenther@suse.de>
11090
11091 * alias.c (find_base_term): Fix thinko in previous change.
11092
401f3a81
JJ
110932013-04-08 Jakub Jelinek <jakub@redhat.com>
11094
11095 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11096 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11097 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11098 if possible to compute val.
11099 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11100 For QImode integers don't require anything about precision. Use
11101 const_with_all_bytes_same to find out if the constant doesn't have
11102 repeated bytes in it.
11103
03ed99a8
AK
111042013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11105
11106 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11107 within mode size.
11108
781b2e62
MP
111092013-04-08 Marek Polacek <polacek@redhat.com>
11110
11111 PR rtl-optimization/48182
11112 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11113 value to 1.
11114
27e430a2
JDA
111152013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11116
11117 PR target/55487
11118 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11119 nuses, make sure we have a label.
11120
4902aa64
BS
111212013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11122
11123 PR target/56843
11124 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11125 (rs6000_emit_swdiv_low_precision): Remove.
11126 (rs6000_emit_swdiv): Rewrite to handle between one and four
11127 iterations of Newton-Raphson generally; modify required number of
11128 iterations for some cases.
11129 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11130
7bca81dc
SB
111312013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11132
11133 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11134 set-but-unused variable.
11135
11136 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11137 basic blocks of released function bodies garbage-collectable.
11138
11139 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11140 (struct rtl_opt_pass): Add TODO_df_finish.
11141
11142 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11143
4542a38a
GY
111442013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11145
11146 * config/arm/constraints.md (q): New constraint.
11147 * config/arm/ldrdstrd.md: New file.
11148 * config/arm/arm.md (ldrdstrd.md) New include.
11149 (arm_movdi): Use "q" instead of "r" constraint
11150 for double-word memory access.
11151 (movdf_soft_insn): Likewise.
11152 * config/arm/vfp.md (movdi_vfp): Likewise.
11153 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 11154 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY
11155 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11156 (mem_ok_for_ldrd_strd): Likewise.
11157 (output_move_double): Update assertion.
11158
2385b218
GY
111592013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11160
11161 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11162
75fe1cb5
GY
111632013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11164
11165 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11166 define_insn_and_split.
11167 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11168
dd660e8e
GY
111692013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11170
11171 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11172 define_insn_and_split.
33159866 11173 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY
11174 (shiftsi3_compare): New pattern.
11175 (rrx): New pattern.
11176 * config/arm/unspecs.md (UNSPEC_RRX): New.
11177
045e472c
GY
111782013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11179
11180 * config/arm/arm.md (negdi_extendsidi): New pattern.
11181 (negdi_zero_extendsidi): Likewise.
11182
3f3bf1a8
GY
111832013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11184
11185 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11186 define_insn_and_split.
11187 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11188 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11189
b0b49556
GY
111902013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11191
11192 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11193 define_insn_and_split.
11194 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11195 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11196
d633dd84
GY
111972013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11198
11199 * config/arm/arm.md (subsi3_carryin): New pattern.
11200 (subsi3_carryin_const): Likewise.
11201 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11202 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11203
f4499066
GY
112042013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11205
11206 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11207
ceef6fd9
GY
112082013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11209
11210 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 11211 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 11212
ddbdd8a7
KT
112132013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11214
11215 * config/arm/arm.c (arm_expand_builtin): Change fcode
11216 type to unsigned int.
11217
8456d78a
RR
112182013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11219
11220 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11221
526c230f
ILT
112222013-04-04 Ian Lance Taylor <iant@google.com>
11223
33159866
UB
11224 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11225 language standard.
526c230f 11226
3cfbe04d
SB
112272013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11228
11229 PR middle-end/56729
11230 * df-scan.c (df_insn_delete): Disable failing assert.
11231
dfa3f8d0
KT
112322013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11233
11234 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11235 New function prototype.
11236 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11237 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11238 (arm_builtin_vectorized_function): New function.
11239
f7837758
KT
112402013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11241
11242 * config/arm/arm_neon_builtins.def: New file.
11243 * config/arm/arm.c (neon_builtin_data): Move contents to
11244 arm_neon_builtins.def.
11245 (enum arm_builtins): Include neon builtin definitions.
11246 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 11247 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 11248
39c1b6db
MP
112492013-04-04 Marek Polacek <polacek@redhat.com>
11250
11251 PR tree-optimization/48186
11252 * predict.c (maybe_hot_frequency_p): Return false if
11253 HOT_BB_FREQUENCY_FRACTION is 0.
11254 (cgraph_maybe_hot_edge_p): Likewise.
11255
314f64eb
RB
112562013-04-04 Richard Biener <rguenther@suse.de>
11257
11258 PR tree-optimization/56826
11259 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11260 more accurately.
11261
51a905b2
RB
112622013-04-04 Richard Biener <rguenther@suse.de>
11263
11264 PR tree-optimization/56213
11265 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 11266 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 11267
f1bf4f3a
RB
112682013-04-04 Richard Biener <rguenther@suse.de>
11269
11270 PR tree-optimization/56837
11271 * tree-loop-distribution.c (classify_partition): For non-zero
11272 values require that the value has the same precision as its
11273 mode to be useful as memset value.
11274
0bca7ded
NC
112752013-04-03 Nick Clifton <nickc@redhat.com>
11276
33159866 11277 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC
11278 (fmssf4): Use fmsf.s on E3V5 architectures.
11279 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11280 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11281
b4019227
JL
112822013-04-03 Jeff Law <law@redhat.com>
11283
11284 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11285 (lra-eliminations.o): Likewise.
11286
f57ddb5b
TJ
112872013-04-03 Teresa Johnson <tejohnson@google.com>
11288
11289 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 11290 compute_working_sets here from profile.c.
f57ddb5b 11291 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB
11292 (gcov_working_set_t): Moved typedef here from basic-block.h
11293 (compute_working_set): Declare.
f57ddb5b
TJ
11294 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11295 (get_working_sets): Renamed from compute_working_set,
33159866 11296 replace most of body with call to new compute_working_sets.
f57ddb5b 11297 (get_exec_counts): Replace call to compute_working_sets
33159866
UB
11298 to get_working_sets.
11299 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 11300 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 11301 to get_working_sets.
f57ddb5b
TJ
11302 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11303 * gcov-dump.c (dump_working_sets): New function.
11304
5e0919f1
KZ
113052013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11306
11307 * hwint.c (sext_hwi, zext_hwi): New functions.
11308 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11309 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11310 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11311 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11312 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11313 (sext_hwi, zext_hwi): New functions.
0bca7ded 11314
be672e08
JL
113152013-04-03 Jeff Law <law@redhat.com>
11316
11317 PR tree-optimization/56799
33159866
UB
11318 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11319 back test for widening conversion erroneously dropped in prior change.
be672e08 11320
9d821fa5
KT
113212013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11322
11323 PR target/56809
11324 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11325 instead of next_real_insn.
11326
71c581e7
MP
113272013-04-03 Marek Polacek <polacek@redhat.com>
11328
11329 PR sanitizer/55702
33159866 11330 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 11331
4a32ef80
KT
113322013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11333
11334 PR target/56809
11335 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11336 next_real_insn.
11337 (thumb1_output_casesi): Likewise.
11338 (thumb2_output_casesi): Likewise.
11339
1b2253d4
RB
113402013-04-03 Richard Biener <rguenther@suse.de>
11341
11342 PR tree-optimization/56817
11343 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11344 Split out ...
11345 (tree_unroll_loops_completely_1): ... new function to manually
11346 walk the loop tree, properly defering outer loops of unrolled
11347 loops to later iterations.
11348
38000232
MG
113492013-04-03 Marc Glisse <marc.glisse@inria.fr>
11350
11351 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
11352 (vectorizable_load): Likewise.
11353 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
11354 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
11355
3fa35298
MG
113562013-04-03 Marc Glisse <marc.glisse@inria.fr>
11357
11358 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
11359 BIT_FIELD_REF.
11360
b3d45ff0
UW
113612013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11362
11363 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
11364
ec9202a8
BC
113652013-04-03 Bin Cheng <bin.cheng@arm.com>
11366
11367 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
11368
6805bd36
MG
113692013-04-03 Marc Glisse <marc.glisse@inria.fr>
11370
11371 PR tree-optimization/56790
33159866
UB
11372 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
11373 folding.
6805bd36 11374
da694a77
MG
113752013-04-03 Marc Glisse <marc.glisse@inria.fr>
11376
11377 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
11378 Handle VEC_MERGE.
11379 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
11380 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
11381 equal arguments.
11382
4e7d7b3d
JJ
113832013-04-03 Jakub Jelinek <jakub@redhat.com>
11384
11385 PR c/19449
11386 * tree.h (force_folding_builtin_constant_p): New decl.
11387 * builtins.c (force_folding_builtin_constant_p): New variable.
11388 (fold_builtin_constant_p): Fold immediately also if
11389 force_folding_builtin_constant_p.
11390
e6c9d234
RB
113912013-04-03 Richard Biener <rguenther@suse.de>
11392
11393 PR tree-optimization/56812
11394 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
11395 DRs of the same interleaving chain are independent.
11396
984d07dd
JM
113972013-04-02 Jason Merrill <jason@redhat.com>
11398
11399 * gdbinit.in (pbb): Use debug fn.
11400
622849c9
LC
114012013-04-02 Lawrence Crowl <crowl@google.com>
11402
11403 * sese.h (struct ivtype_map_elt_s): Remove unused.
11404 (extern debug_ivtype_map): Remove unused.
11405 (extern eq_ivtype_map_elts): Remove unused.
11406 * sese.c (debug_ivtype_map): Removed unused.
11407 (debug_ivtype_map_1): Removed unused.
11408 (debug_ivtype_elt): Remove unused.
11409 (eq_ivtype_map_elts): Remove unused.
11410
11411
82c0e1a0
KT
114122013-04-02 Kai Tietz <ktietz@redhat.com>
11413
11414 PR target/52790
11415 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
11416 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB
11417 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
11418 function.
82c0e1a0
KT
11419 (legitimize_pe_coff_symbol): Likewise.
11420 (is_imported_p): New helper-function.
11421 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
11422 for Windows x64 targets.
11423 (ix86_expand_prologue): Optimize for pe-coff targets.
11424 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
11425 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
11426 medium/large code-model.
11427 (legitimize_pic_address): Likewise.
11428 (legitimize_tls_address): Likewise.
11429 (ix86_expand_call): Likewise.
11430 (x86_output_mi_thunk): Likewise.
11431 (get_dllimport_decl): Add new beimport argument.
11432 (construct_plt_address): Don't assert for x64 pe-coff targets.
11433 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
11434 targets.
11435 (SYMBOL_FLAG_STUBVAR): New macro.
11436 (SYMBOL_REF_STUBVAR_P): Likewise.
11437 * config/i386/winnt.c (stub_list): New structure.
11438 (stub_head): New local variable.
11439 (i386_pe_record_stub): New function.
11440 (i386_pe_file_end): Emit refptr-stubs.
11441
5d751b0c
JJ
114422013-04-02 Jakub Jelinek <jakub@redhat.com>
11443
09bb4c99
JJ
11444 PR rtl-optimization/56745
11445 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
11446 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
11447
a1d8947a
JJ
11448 PR c++/34949
11449 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
11450 and both of them are MEM_REFs, just compare first argument for
11451 equality and attempt to deal even with differing offsets.
11452
5d751b0c
JJ
11453 PR c++/34949
11454 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
11455 of gimple_clobber_p to be MEM_REF.
11456 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
11457 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
11458 after gimplification.
11459 * asan.c (get_mem_ref_of_assignment): Don't instrument
11460 gimple_clobber_p stmts.
11461 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
11462 gimple_clobber_p stmt if they have MEM_REF lhs and
11463 are dead because of another gimple_clobber_p stmt.
11464 * tree-ssa-live.c (clear_unused_block_pointer): Treat
11465 gimple_clobber_p stmts like debug stmts.
11466 (remove_unused_locals): Remove clobbers with MEM_REF lhs
11467 that refer to unused VAR_DECLs or uninitialized values.
11468 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
11469 gimple_clobber_p stmts if they refer to removed parameters.
11470 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
11471 formatting.
11472
e594716a
UB
114732013-04-02 Uros Bizjak <ubizjak@gmail.com>
11474
11475 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
11476 using SWI48 mode attribute.
11477
7a80735b
WM
114782013-04-02 Wei Mi <wmi@google.com>
11479
11480 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
11481 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
11482 *<rotate_insn><mode>3_mask in i386.md.
11483
f423a9e4
AI
114842013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
11485
11486 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
11487
90eb75f2
RB
114882013-04-02 Richard Biener <rguenther@suse.de>
11489
11490 PR tree-optimization/56778
11491 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
11492 Runtime alias tests are not supported for gather loads.
11493 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
11494 stmts referenced from SSA operands before updating SSA form.
11495
d8c69a92
IC
114962013-04-02 Ian Caulfield <ian.caulfield@arm.com>
11497 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11498
11499 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB
11500 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
11501 * config/arm/cortex-a53.md: New file.
11502 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
11503 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
11504 * config/arm/arm.c (arm_issue_rate): Likewise.
11505 * config/arm/arm-tune.md: Regenerate
11506 * config/arm/arm-tables.opt: Regenerate.
11507 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 11508
239eb04c
ZC
115092013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
11510
11511 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
11512 non-static link.
11513
c902d3c8
SN
115142013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11515
d8c69a92
IC
11516 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
11517 scalar load/store operations using B/H registers.
c902d3c8
SN
11518 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
11519
051d0e2f
SN
115202013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11521
11522 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
11523 scalar move.
11524 * config/aarch64/aarch64.c
11525 (aarch64_simd_scalar_immediate_valid_for_move): New.
11526 * config/aarch64/aarch64-protos.h
11527 (aarch64_simd_scalar_immediate_valid_for_move): New.
11528 * config/aarch64/constraints.md (Dh, Dq): New.
11529 * config/aarch64/iterators.md (hq): New.
11530
0ee1e3d9
EB
115312013-04-02 Eric Botcazou <ebotcazou@adacore.com>
11532
11533 * reorg.c (get_branch_condition): Deal with conditional returns.
11534 (fill_simple_delay_slots): Remove dead code dealing with jumps.
11535
136fb3f7
WM
115362013-04-01 Wei Mi <wmi@google.com>
11537
11538 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
11539 Truncate operand 2 using %b asm operand modifier.
11540 (*<shift_insn><mode>3_mask): Ditto.
11541 (*<rotate_insn><mode>3_mask): Ditto.
11542
6388c738
SB
115432013-04-01 Steven Bosscher <steven@gcc.gnu.org>
11544
11545 PR middle-end/56798
11546 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
11547
f7a4d826
KK
115482013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
11549
11550 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
11551 of next_real_insn.
11552 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
11553
4b943a49
LC
115542013-03-30 Lawrence Crowl <crowl@google.com>
11555
11556 * dse.c (clear_alias_sets): Remove never set.
11557 (disqualified_clear_alias_sets): Remove never set.
11558 (clear_alias_mode_pool): Remove never set.
11559 (dse_step0): Remove condition that is never true.
11560 (canon_address): Remove condition that is never true.
11561 (dse_step7): Remove condition that is never true.
11562 (rest_of_handle_dse): Remove condition that is never true.
11563 (rest_of_handle_dse::did_global): Remove never read from above.
11564 (dse_step2_spill): Remove never called from above.
11565 (dse_step5_spill): Remove never called from above.
11566
39718607
SB
115672013-03-30 Steven Bosscher <steven@gcc.gnu.org>
11568
da5c6bde
SB
11569 * doc/md.texi (Standard Names) <casesi>: Update documentation for
11570 JUMP_TABLE_DATA changes.
11571 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
11572 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
11573 (Insns) <jump_table_data>: New entry.
11574 * doc/tm.texi: Regenerate.
11575
39718607
SB
11576 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
11577
11578 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
11579 for table jump at the end of a basic block using tablejump_p.
11580 * targhooks.c (default_invalid_within_doloop): Likewise.
11581 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
11582 target hook implementation that is identical to the default hook.
11583 (rs6000_invalid_within_doloop): Remove.
11584
11585 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
11586 unused variable from tablejump_p call.
11587
11588 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
11589 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
11590 (INSN_DELETED_P): Likewise.
11591 (emit_jump_table_data): New prototype.
11592 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
11593 after 4th as unused.
11594 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
11595 * sched-vis.c (print_insn): Likewise.
11596 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
11597 insn for compatibility with back ends that use next_active_insn to
11598 identify jump table data.
11599 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
11600 (remove_insn): Likewise.
11601 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
11602 to be emitted.
11603 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
11604 (emit_jump_table_data): New function.
11605
11606 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
11607 basic block, a JUMP_TABLE_DATA never is.
11608 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
11609 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
11610 off from code handling real insns.
11611 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
11612 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
11613 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
11614 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
11615 is not a NONDEBUG_INSN_P.
11616 * ira-costs.c (scan_one_insn): Likewise.
11617 * jump.c (mark_all_labels): Likewise.
11618 (mark_jump_label_1): Likewise.
11619 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11620 * lra.c (get_insn_freq): Expect all insns reaching here to be in
11621 a basic block.
11622 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
11623 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
11624 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
11625 JUMP_TABLE_DATA_P insns.
11626 (calculate_elim_costs_all_insns): Likewise.
11627 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
11628 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
11629 (delete_output_reload): Code style fixups.
11630 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
11631 insn flags on this non-insn.
11632 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
11633 as scheduling barriers, for pre-change compatibility.
11634 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
11635 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
11636
11637 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
11638 redundant JUMP_TABLE_DATA_P test.
11639 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
11640 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11641 (frv_for_each_packet): Likewise.
11642 * config/i386/i386.c (min_insn_size): Likewise.
11643 (ix86_avoid_jump_mispredicts): Likewise.
11644 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11645 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
11646 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11647 (mips16_insn_length): Robustify.
11648 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
11649 (mips16_split_long_branches): Likewise.
11650 * config/pa/pa.c (pa_combine_instructions): Likewise.
11651 * config/rs6000/rs6000.c (get_next_active_insn): Treat
11652 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
11653 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
11654 as contributing to pool range lengths.
11655 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
11656 Remove redundant JUMP_TABLE_DATA_P test.
11657 (sh_loop_align): Likewise.
11658 (split_branches): Likewise.
11659 (sh_insn_length_adjustment): Likewise.
11660 * config/spu/spu.c (get_branch_target): Likewise.
11661
0208f7da
JH
116622013-03-29 Jan Hubicka <jh@suse.cz>
11663
11664 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
11665 gcov streaming; stream hot bb threshold to ltrans.
11666 * predict.c (get_hot_bb_threshold): Break out from ....
11667 (maybe_hot_count_p): ... here.
11668 (set_hot_bb_threshold): New function.
11669 * lto-section-in.c (lto_section_name): Add profile.
11670 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
11671 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
11672 and data-streamer.h
11673 (histogram_entry): New structure.
11674 (histogram, histogram_pool): New global vars.
11675 (histogram_hash): New structure.
11676 (histogram_hash::hash): New method.
11677 (histogram_hash::equal): Likewise.
11678 (account_time_size): New function.
11679 (cmp_counts): New function.
11680 (dump_histogram): New function.
11681 (ipa_profile_generate_summary): New function.
11682 (ipa_profile_write_summary): New function.
11683 (ipa_profile_read_summary): New function.
11684 (ipa_profile): Decide on threshold.
e594716a
UB
11685 (pass_ipa_profile): Add ipa_profile_write_summary and
11686 ipa_profile_read_summary.
0208f7da
JH
11687 * Makefile.in (ipa.o): Update dependencies.
11688 * lto-streamer.h (LTO_section_ipa_profile): New section.
11689
5a6ccc94
GDR
116902013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
11691
11692 * tree.h (VAR_P): New.
11693
39385fa6
PC
116942013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
11695
11696 PR lto/56777
11697 * doc/invoke.texi ([-fwhole-program]): Fix typo.
11698
34f0d87a
SB
116992013-03-29 Steven Bosscher <steven@gcc.gnu.org>
11700
11701 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
11702 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11703 (control_flow_insn_p): Likewise.
11704 * cfgrtl.c (duplicate_insn_chain): Likewise.
11705 * final.c (get_attr_length_1): Likewise.
11706 (shorten_branches): Likewise.
11707 (final_scan_insn): Likewise.
11708 * function.c (instantiate_virtual_regs): Likewise.
11709 * gcse.c (insert_insn_end_basic_block): Likewise.
11710 * ira-costs.c (scan_one_insn): Likewise.
11711 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11712 * lra.c (check_rtl): Likewise.
11713 * reload1.c (elimination_costs_in_insn): Likewise.
11714 * reorg.c (follow_jumps): Likewise.
11715
11716 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
11717 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11718 (thumb_far_jump_used_p): Likewise.
11719 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
11720 (workaround_speculation): Likewise.
11721 (add_sched_insns_for_speculation): Likewise.
11722 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
11723 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11724 (frv_for_each_packet): Likewise.
11725 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
11726 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
11727 (final_emit_insn_group_barriers): Likewise.
11728 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11729 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11730 (mips16_insn_length): Likewise.
11731 * config/pa/pa.c (pa_reorg): Likewise.
11732 (pa_combine_instructions): Likewise.
11733 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
11734 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
11735 (sh_reorg): Likewise.
11736 (split_branches): Likewise.
11737 * config/spu/spu.c (get_branch_target): Likewise.
11738
11739 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
11740 JUMP_TABLE_DATA_P.
11741
4ac761b0
KY
117422013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
11743
39385fa6 11744 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY
11745 Fix declaration name.
11746
58a49189
LC
117472013-03-28 Lawrence Crowl <crowl@google.com>
11748
11749 * graphds.h (struct graph.indicies): Remove unused.
11750 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
11751 (SCOP_ORIGINAL_PDDRS): Remove unused.
11752 * sese.h (extern insert_loop_close_phis): Removed unused.
11753 (extern insert_guard_phis): Removed unused.
11754 (extern ivtype_map_elt_info): Removed unused.
11755 (new_ivtype_map_elt): Removed unused.
11756 * sese.c (ivtype_map_elt_info): Removed unused.
11757
7b3b6ae4
LC
117582013-03-28 Lawrence Crowl <crowl@google.com>
11759
11760 * Makefile.in: Add several missing include dependences.
11761 (DUMPFILE_H): New.
11762 (test-dump.o): New. This object is not added to any executable,
11763 but is present for ad-hoc testing.
11764 * bitmap.c
11765 (debug (const bitmap_head_def &)): New.
11766 (debug (const bitmap_head_def *)): New.
11767 * bitmap.h
11768 (extern debug (const bitmap_head_def &)): New.
11769 (extern debug (const bitmap_head_def *)): New.
11770 * cfg.c
11771 (debug (edge_def &)): New.
11772 (debug (edge_def *)): New.
11773 * cfghooks.c
11774 (debug (basic_block_def &)): New.
11775 (debug (basic_block_def *)): New.
11776 * dumpfile.h
11777 (dump_node (const_tree, int, FILE *)): Correct source file.
11778 * dwarf2out.c
11779 (debug (die_struct &)): New.
11780 (debug (die_struct *)): New.
11781 * dwarf2out.h
11782 (extern debug (die_struct &)): New.
11783 (extern debug (die_struct *)): New.
11784 * gimple-pretty-print.c
11785 (debug (gimple_statement_d &)): New.
11786 (debug (gimple_statement_d *)): New.
11787 * gimple-pretty-print.h
11788 (extern debug (gimple_statement_d &)): New.
11789 (extern debug (gimple_statement_d *)): New.
11790 * ira-build.c
11791 (debug (ira_allocno_copy &)): New.
11792 (debug (ira_allocno_copy *)): New.
11793 (debug (ira_allocno &)): New.
11794 (debug (ira_allocno *)): New.
11795 * ira-int.h
11796 (extern debug (ira_allocno_copy &)): New.
11797 (extern debug (ira_allocno_copy *)): New.
11798 (extern debug (ira_allocno &)): New.
11799 (extern debug (ira_allocno *)): New.
11800 * ira-lives.c
11801 (debug (live_range &)): New.
11802 (debug (live_range *)): New.
11803 * lra-int.h
11804 (debug (lra_live_range &)): New.
11805 (debug (lra_live_range *)): New.
11806 * lra-lives.c
11807 (debug (lra_live_range &)): New.
11808 (debug (lra_live_range *)): New.
11809 * omega.c
11810 (debug (omega_pb_d &)): New.
11811 (debug (omega_pb_d *)): New.
11812 * omega.h
11813 (extern debug (omega_pb_d &)): New.
11814 (extern debug (omega_pb_d *)): New.
11815 * print-rtl.c
11816 (debug (const rtx_def &)): New.
11817 (debug (const rtx_def *)): New.
11818 * print-tree.c
11819 (debug_tree (tree): Move within file.
11820 (debug_raw (const tree_node &)): New.
11821 (debug_raw (const tree_node *)): New.
11822 (dump_tree_via_hooks (const tree_node *, int)): New.
11823 (debug (const tree_node &)): New.
11824 (debug (const tree_node *)): New.
11825 (debug_verbose (const tree_node &)): New.
11826 (debug_verbose (const tree_node *)): New.
11827 (debug_head (const tree_node &)): New.
11828 (debug_head (const tree_node *)): New.
11829 (debug_body (const tree_node &)): New.
11830 (debug_body (const tree_node *)): New.
11831 (debug_vec_tree (tree): Move and reimplement in terms of dump.
11832 (debug (vec<tree, va_gc> &)): New.
11833 (debug (vec<tree, va_gc> *)): New.
11834 * rtl.h
11835 (extern debug (const rtx_def &)): New.
11836 (extern debug (const rtx_def *)): New.
11837 * sbitmap.c
11838 (debug_raw (simple_bitmap_def &)): New.
11839 (debug_raw (simple_bitmap_def *)): New.
11840 (debug (simple_bitmap_def &)): New.
11841 (debug (simple_bitmap_def *)): New.
11842 * sbitmap.h
11843 (extern debug (simple_bitmap_def &)): New.
11844 (extern debug (simple_bitmap_def *)): New.
11845 (extern debug_raw (simple_bitmap_def &)): New.
11846 (extern debug_raw (simple_bitmap_def *)): New.
11847 * sel-sched-dump.c
11848 (debug (vinsn_def &)): New.
11849 (debug (vinsn_def *)): New.
11850 (debug_verbose (vinsn_def &)): New.
11851 (debug_verbose (vinsn_def *)): New.
11852 (debug (expr_def &)): New.
11853 (debug (expr_def *)): New.
11854 (debug_verbose (expr_def &)): New.
11855 (debug_verbose (expr_def *)): New.
11856 (debug (vec<rtx> &)): New.
11857 (debug (vec<rtx> *)): New.
11858 * sel-sched-dump.h
11859 (extern debug (vinsn_def &)): New.
11860 (extern debug (vinsn_def *)): New.
11861 (extern debug_verbose (vinsn_def &)): New.
11862 (extern debug_verbose (vinsn_def *)): New.
11863 (extern debug (expr_def &)): New.
11864 (extern debug (expr_def *)): New.
11865 (extern debug_verbose (expr_def &)): New.
11866 (extern debug_verbose (expr_def *)): New.
11867 (extern debug (vec<rtx> &)): New.
11868 (extern debug (vec<rtx> *)): New.
11869 * sel-sched-ir.h
11870 (_list_iter_cond_expr): Make inline instead of static.
11871 * sreal.c
11872 (debug (sreal &)): New.
11873 (debug (sreal *)): New.
11874 * sreal.h
11875 (extern debug (sreal &)): New.
11876 (extern debug (sreal *)): New.
11877 * tree.h
11878 (extern debug_raw (const tree_node &)): New.
11879 (extern debug_raw (const tree_node *)): New.
11880 (extern debug (const tree_node &)): New.
11881 (extern debug (const tree_node *)): New.
11882 (extern debug_verbose (const tree_node &)): New.
11883 (extern debug_verbose (const tree_node *)): New.
11884 (extern debug_head (const tree_node &)): New.
11885 (extern debug_head (const tree_node *)): New.
11886 (extern debug_body (const tree_node &)): New.
11887 (extern debug_body (const tree_node *)): New.
11888 (extern debug (vec<tree, va_gc> &)): New.
11889 (extern debug (vec<tree, va_gc> *)): New.
11890 * tree-cfg.c
11891 (debug (struct loop &)): New.
11892 (debug (struct loop *)): New.
11893 (debug_verbose (struct loop &)): New.
11894 (debug_verbose (struct loop *)): New.
11895 * tree-dump.c: Add header dependence.
11896 * tree-flow.h
11897 (extern debug (struct loop &)): New.
11898 (extern debug (struct loop *)): New.
11899 (extern debug_verbose (struct loop &)): New.
11900 (extern debug_verbose (struct loop *)): New.
11901 * tree-data-ref.c
11902 (debug (data_reference &)): New.
11903 (debug (data_reference *)): New.
11904 (debug (vec<data_reference_p> &)): New.
11905 (debug (vec<data_reference_p> *)): New.
11906 (debug (vec<ddr_p> &)): New.
11907 (debug (vec<ddr_p> *)): New.
11908 * tree-data-ref.h
11909 (extern debug (data_reference &)): New.
11910 (extern debug (data_reference *)): New.
11911 (extern debug (vec<data_reference_p> &)): New.
11912 (extern debug (vec<data_reference_p> *)): New.
11913 (extern debug (vec<ddr_p> &)): New.
11914 (extern debug (vec<ddr_p> *)): New.
11915 * tree-ssa-alias.c
11916 (debug (pt_solution &)): New.
11917 (debug (pt_solution *)): New.
11918 * tree-ssa-alias.h
11919 (extern debug (pt_solution &)): New.
11920 (extern debug (pt_solution *)): New.
11921 * tree-ssa-alias.c
11922 (debug (_var_map &)): New.
11923 (debug (_var_map *)): New.
11924 (debug (tree_live_info_d &)): New.
11925 (debug (tree_live_info_d *)): New.
11926 * tree-ssa-alias.h
11927 (extern debug (_var_map &)): New.
11928 (extern debug (_var_map *)): New.
11929 (extern debug (tree_live_info_d &)): New.
11930 (extern debug (tree_live_info_d *)): New.
11931
be77e1e5
JH
119322013-03-28 Jan Hubicka <jh@suse.cz>
11933
11934 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
11935
777e6976
IB
119362013-03-28 Ian Bolton <ian.bolton@arm.com>
11937
11938 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
11939 record only when desired or required.
11940
9cd347ae
UB
119412013-03-28 Uros Bizjak <ubizjak@gmail.com>
11942
11943 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
11944 *vec_extractv2di_1_rex64. Use x64 isa attribute.
11945
33159866 119462013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N
11947
11948 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
11949 (*andsi3_compare0_uxtw): New pattern.
11950 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
11951 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
11952
89ab31c1
JH
119532013-03-28 Jan Hubicka <jh@suse.cz>
11954
11955 * data-streamer-in.c (streamer_read_gcov_count): New function.
11956 * gimple-streamer-out.c: Include value-prof.h.
11957 (output_gimple_stmt): Output histogram.
11958 (output_bb): Use streamer_write_gcov_count.
11959 * value-prof.c: Include data-streamer.h
11960 (dump_histogram_value): Add HIST_TYPE_MAX.
11961 (stream_out_histogram_value): New function.
11962 (stream_in_histogram_value): New function.
11963 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
11964 (stream_out_histogram_value, stream_in_histogram_value): Declare.
11965 * data-streamer-out.c (streamer_write_gcov_count): New function.
11966 (streamer_write_gcov_count_stream): New function.
11967 * lto-cgraph.c (lto_output_edge): Update counter streaming.
11968 (lto_output_node): Likewise.
11969 (input_node, input_edge): Likewise.
11970 * lto-streamer-out.c (output_cfg): Update streaming.
11971 * lto-streamer-in.c (input_cfg): Likewise.
11972 * data-streamer.h (streamer_write_gcov_count,
11973 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
11974 * gimple-streamer-in.c: Include value-prof.h
11975 (input_gimple_stmt): Input histograms.
11976 (input_bb): Update profile streaming.
11977
e594716a 119782013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 11979
33159866 11980 * genmodes.c (emit_max_int): New function.
8697be17 11981 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC
11982 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
11983 Added doc.
8697be17 11984 * machmode.def: Fixed comment.
89ab31c1 11985
e594716a 119862013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ
11987
11988 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ
11989 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
11990
74946978
MP
119912013-03-28 Marek Polacek <polacek@redhat.com>
11992 Richard Biener <rguenther@suse.de>
11993
11994 PR tree-optimization/56695
11995 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
11996 build signed result of a vector comparison.
11997 * tree-cfg.c (verify_gimple_comparison): Check that a result
11998 of a vector comparison has signed type.
11999
a64b9c26
RB
120002013-03-28 Richard Biener <rguenther@suse.de>
12001
12002 PR tree-optimization/37021
12003 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12004 do not restrict gaps between groups.
12005 * tree-vect-stmts.c (vectorizable_load): Properly account for
12006 a gap between groups.
12007
a9dc2a2f
EB
120082013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12009
12010 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12011 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12012 is not enabled.
12013
53cb97f9
GP
120142013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12015
12016 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12017 * doc/extend.texi (Named Address Spaces): Ditto.
12018 (Variable Attributes): Ditto.
12019
b802ae5c
KT
120202013-03-27 Kai Tietz <ktietz@redhat.com>
12021
eddae10a
KT
12022 * config.build: Add support for cygwin x64 target.
12023 * config.gcc: Likewise.
12024 * config.host: Likewise.
12025 * configure.ac: Likewise
12026 * configure: Regenerated.
b802ae5c 12027
371e77e3 120282013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT
12029
12030 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12031 * config/i386/t-cygwin-w64: New file.
12032 * config/i386/cygwin-w64.h: New file.
12033 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12034 and add support for x64-cygwin target.
12035 (CPP_SPEC): Likewise.
12036 (CXX_WRAP_SPEC_LIST): Undefine before define.
12037 (LIBGCJ_SONAME): Use 15 as version.
12038
f49b33cb
RB
120392013-03-27 Richard Biener <rguenther@suse.de>
12040
12041 PR tree-optimization/56716
12042 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12043 dumping for ref nodes.
12044
b37a6ce5
MJ
120452013-03-27 Martin Jambor <mjambor@suse.cz>
12046
12047 PR tree-optimization/55334
12048 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12049 restricted pointers to arrays.
12050
9469b9b2
GDR
120512013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12052
12053 * Makefile.in (.SUFFIXES): Add .cc.
12054 (.c.o): Apply same recipe for implicit rule .cc.o.
12055
7d24f650
RB
120562013-03-27 Richard Biener <rguenther@suse.de>
12057
12058 PR tree-optimization/37021
12059 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12060 REALPART/IMAGPART_EXPRs around the supported refs.
12061 * tree-ssa-structalias.c (find_func_aliases): Assume that
12062 floating-point values are not used to transfer pointers.
12063
2f251a05
AI
120642013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12065
39385fa6
PC
12066 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12067 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12068 * doc/tm.texi: Regenerate.
12069 * targhooks.h (default_has_ifunc_p): New.
12070 * targhooks.c (default_has_ifunc_p): Ditto.
12071 * config/linux-protos.h: New file.
12072 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
2f251a05
AI
12073 this hook for linux which disables support of indirect functions in
12074 android.
39385fa6
PC
12075 * config/linux-android.c: New file.
12076 * config/t-linux-android.c: Ditto.
12077 * config.gcc: Added new object file linux-android.o.
12078 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 12079 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
39385fa6 12080 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI
12081 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12082 doesn't support indirect functions.
12083 * configure: Regenerate.
12084
78b4e425
BC
120852013-03-27 Bin Cheng <bin.cheng@arm.com>
12086
12087 PR target/56102
12088 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12089 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12090 mult-word mode.
12091
a5ba7b92
AK
120922013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12093
12094 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12095
be624986
TG
120962013-03-27 Terry Guo <terry.guo@arm.com>
12097
12098 * config/arm/arm-cores.def: Added core cortex-r7.
12099 * config/arm/arm-tune.md: Regenerated.
12100 * config/arm/arm-tables.opt: Regenerated.
12101 * doc/invoke.texi: Added entry for core cortex-r7.
12102
0a514f47
WL
121032013-03-27 Walter Lee <walt@tilera.com>
12104
12105 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12106 double-decrement of next_scratch_regno.
12107
9332b0d2
WL
121082013-03-27 Walter Lee <walt@tilera.com>
12109
801d9b2a 12110 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL
12111 input operands.
12112 (insn_v1mulus): Ditto.
12113 (insn_v2muls): Ditto.
12114
f54ea5dd
WL
121152013-03-27 Walter Lee <walt@tilera.com>
12116
e594716a 12117 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL
12118 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12119
bf60f4ca
WL
121202013-03-27 Walter Lee <walt@tilera.com>
12121
e594716a 12122 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL
12123 (*sibcall_value): Ditto.
12124
e3b51eeb
WL
121252013-03-27 Walter Lee <walt@tilera.com>
12126
12127 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12128 (insn_mnz_v8qi): ... this ...
12129 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12130 vector equivalent.
12131 (insn_v<n>mnz): Replaced by ...
12132 (insn_v1mnz): ... this ...
12133 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12134 equivalent.
12135 (insn_mz_<mode>): Replaced by ...
12136 (insn_mz_v8qi): ... this ...
12137 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12138 vector equivalent.
12139 (insn_v<n>mz): Replaced by ...
12140 (insn_v1mz): ... this ...
12141 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12142 equivalent.
12143
4fc7b145
EB
121442013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12145
12146 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12147
6f33817e
RM
121482013-03-26 Roland McGrath <mcgrathr@google.com>
12149
12150 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12151 than fprintf with a non-constant, non-format string.
12152
93a4145b
UB
121532013-03-26 Uros Bizjak <ubizjak@gmail.com>
12154
12155 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12156 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12157 operand 0 predicate.
12158 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12159 attribute. Use general_x64nomem_operand as operand 1 predicate.
12160 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12161 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12162 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12163 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12164 (mov<mode>_insv_1): Remove expander. Merge insn with
12165 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12166 Use general_x64nomem_operand as operand 1 predicate.
12167 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12168 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12169 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12170 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12171 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12172 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12173 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12174 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12175 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12176 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12177 (general_x64nomem_operand): Ditto.
12178
c6a9ed5a 121792013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH
12180
12181 * config/rtems.opt: Add -pthread option.
12182
31b0a960
RB
121832013-03-26 Richard Biener <rguenther@suse.de>
12184
93a4145b
UB
12185 * alias.c (find_base_term): Avoid redundant and not used recursion.
12186 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB
12187 (true_dependence_1): Compute and pass base terms to base_alias_check.
12188 (write_dependence_p): Likewise.
12189 (may_alias_p): Likewise.
12190
79517551
SN
121912013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12192
12193 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12194 PC-relative load in SI modes and above only.
12195
a76213b9
XQ
121962013-03-26 Xinyu Qi <xyqi@marvell.com>
12197
12198 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB
12199 * config/arm/iwmmxt.md (WCGR0): Update.
12200 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 12201
37ff9355
UB
122022013-03-26 Uros Bizjak <ubizjak@gmail.com>
12203
12204 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12205 Use x64 and nox64 isa attributes.
12206
30b0317c
RB
122072013-03-26 Richard Biener <rguenther@suse.de>
12208
12209 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12210 alignment computations and rely on get_object_alignment_1
12211 for the !TYPE_P case.
12212 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12213
cbcdb140
WL
122142013-03-26 Walter Lee <walt@tilera.com>
12215
12216 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12217 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12218
dd884525 122192013-03-25 Jeff Law <law@redhat.com>
be672e08 12220
33159866
UB
12221 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12222 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL
12223 wrong version of prior patch.
12224
0051d3ec
WL
122252013-03-25 Walter Lee <walt@tilera.com>
12226
12227 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12228 TILEGX_INSN_SHUFFLEBYTES1.
12229 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12230 shufflebytes1.
12231 (tilegx_builtins): Ditto.
12232 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12233
065a3605
WL
122342013-03-25 Walter Lee <walt@tilera.com>
12235
12236 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL
12237 (floatunssisf2): New pattern.
12238 (floatsidf2): New pattern.
12239 (floatunssidf2): New pattern.
065a3605 12240
5b2a3c82
WL
122412013-03-25 Walter Lee <walt@tilera.com>
12242
12243 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12244 tests for constraint J, K, N, P.
12245
192ea533
WL
122462013-03-25 Walter Lee <walt@tilera.com>
12247
12248 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12249 Use indirect/pcrel encoding.
12250 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12251 Ditto.
12252
855e0d0b
SE
122532013-03-25 Steve Ellcey <sellcey@mips.com>
12254
12255 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12256 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12257 * config/mips/mips.c (mips_option_override): Set IMADD default.
12258 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12259 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12260 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12261 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12262 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12263
39e45653
JL
122642013-03-25 Jeff Law <law@redhat.com>
12265
12266 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12267 slightly to avoid creating and folding useless trees. Simplify
12268 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12269
7482c470
UB
122702013-03-25 Uros Bizjak <ubizjak@gmail.com>
12271
12272 * config/i386/i386.md (*zero_extendsidi2): Merge with
12273 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12274 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12275 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12276 targets. Clarify comment.
12277
4a53743e
MJ
122782013-03-25 Martin Jambor <mjambor@suse.cz>
12279
12280 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12281 pass-through jump functions differently.
12282 (ipa_read_jump_function): Likewise. Also use setter functions to set
12283 up jump functions.
12284
162712de
MJ
122852013-03-25 Martin Jambor <mjambor@suse.cz>
12286
12287 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12288 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12289 process it.
12290 (ipa_get_indirect_edge_target): New function.
12291 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12292 ipa_get_indirect_edge_target. Update all callers.
12293 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12294 ipa_get_indirect_edge_target_1 instead of calling
12295 ipa_get_indirect_edge_target.
12296 (create_specialized_node): Pass aggvlas to
12297 ipcp_discover_new_direct_edges.
12298
a5a4c20a
KT
122992013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12300
12301 * config/arm/arm.md (f_sels, f_seld): New types.
12302 (*cmov<mode>): New pattern.
12303 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12304 predicate.
12305
f992b9fc
KT
123062013-03-25 Kai Tietz <ktietz@redhat.com>
12307
12308 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12309 POSIX-printf for mingw-hosted builds.
12310
c16fd676
RB
123112013-03-25 Richard Biener <rguenther@suse.de>
12312
12313 PR middle-end/56694
12314 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12315 must-not-throw stmt location.
12316
5ad29f12
KT
123172013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12318
12319 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12320 Emit load-acquire versions when acq is true.
12321 (arm_emit_store_exclusive): Add rel parameter.
12322 Emit store-release versions when rel is true.
12323 (arm_split_compare_and_swap): Use acquire-release instructions
12324 instead.
12325 of barriers when appropriate.
12326 (arm_split_atomic_op): Likewise.
12327 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12328 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12329 (VUNSPEC_SLX): Likewise.
12330 (VUNSPEC_LDA): Likewise.
12331 (VUNSPEC_STL): Likewise.
12332 * config/arm/sync.md (atomic_load<mode>): New pattern.
12333 (atomic_store<mode>): Likewise.
12334 (arm_load_acquire_exclusive<mode>): Likewise.
12335 (arm_load_acquire_exclusivesi): Likewise.
12336 (arm_load_acquire_exclusivedi): Likewise.
12337 (arm_store_release_exclusive<mode>): Likewise.
12338
03a7dddb
CM
123392013-03-25 Catherine Moore <clm@codesourcery.com>
12340
12341 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12342 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12343 * config/mip/predicates.md (lwsp_swsp_operand,
12344 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12345 sb16_operand, db4_operand, db7_operand, ib3_operand,
12346 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
12347 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
12348 andi16_operand): New predicates.
12349 * config/mips/mips.md (compression): New attribute.
12350 (enabled): New attribute.
12351 (length): Consider compression in computing length.
12352 (shift_compression): New code attribute.
12353 (*add<mode>3): New operands. Record compression.
12354 (sub<mode>3): Likewise.
12355 (one_cmpl<mode>2): Likewise.
12356 (*and<mode>3): Likewise.
12357 (*ior<mode>3): Likewise.
12358 (unnamed pattern for xor): Likewise.
12359 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
12360 (*<optab><mode>3): Likewise.
12361 (*mov<mode>_internal: Likewise.
12362 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
12363 (mips_unsigned_immediate_p): New.
12364 (umips_lwsp_swsp_address_p): New.
12365 (m16_based_address_p): New.
12366 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
12367 (mips_unsigned_immediate_p): New prototype.
12368 (lwsp_swsp_address_p): New prototype.
12369 (m16_based_address_p): New prototype.
12370 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
12371 (mips_signed_immediate_p): New function.
12372 (m16_based_address_p): New function.
12373 (lwsp_swsp_address_p): New function.
12374 (mips_print_operand_punctuation): Recognize short delay slot insns
12375 for microMIPS.add<mode>3"
12376
f35c297f 123772013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 12378
f35c297f
KT
12379 PR target/56720
12380 * config/arm/iterators.md (v_cmp_result): New mode attribute.
12381 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
12382
051b9446
RB
123832013-03-25 Richard Biener <rguenther@suse.de>
12384
12385 PR tree-optimization/56689
12386 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
12387 any edge.
12388
374001cb
RB
123892013-03-25 Richard Biener <rguenther@suse.de>
12390
12391 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
12392 of bitmap.
12393 (memory_references): Likewise.
12394 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
12395 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
12396 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
12397 (gather_mem_refs_in_loops): Fold into ...
12398 (analyze_memory_references): ... this. Move initialization
12399 to tree_ssa_lim_initialize.
12400 (fill_always_executed_in): Rename to ...
12401 (fill_always_executed_in_1): ... this.
12402 (fill_always_executed_in): Move contains_call computation to
12403 this new function from ...
12404 (tree_ssa_lim_initialize): ... here.
12405 (tree_ssa_lim): Call fill_always_executed_in.
12406
57895947
EB
124072013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12408
12409 * postreload.c (reload_combine): Fix code detecting returns.
12410
1e1b18c1
EB
124112013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12412
12413 * function.c (emit_use_return_register_into_block): On cc0 targets,
12414 do not emit the sequence between cc0 setter and user.
12415
9216baf1
KT
124162013-03-25 Kai Tietz <ktietz@redhat.com>
12417
1e1b18c1
EB
12418 * config/i386/predicates.md (local_symbolic_operand): Interpret
12419 dll-imported symbols as none-local.
9216baf1 12420
76421b44
RB
124212013-03-25 Richard Biener <rguenther@suse.de>
12422
12423 * tree-ssa-loop-im.c (struct depend): Remove.
12424 (struct lim_aux_data): Make depends a vec of gimples.
12425 (free_lim_aux_data): Adjust.
12426 (add_dependency): Likewise.
12427 (set_level): Likewise.
12428
d154bfa2
RB
124292013-03-25 Richard Biener <rguenther@suse.de>
12430
12431 PR middle-end/56434
12432 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
12433 the pointer returned by calls with ECF_MALLOC set.
12434
2cd9804e
UB
124352013-03-24 Uros Bizjak <ubizjak@gmail.com>
12436
acef5fe0
UB
12437 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
12438
124392013-03-24 Uros Bizjak <ubizjak@gmail.com>
12440
12441 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB
12442 using MMXMODE mode iterator.
12443 (*move<mode>_internal): Merge with *movv2sf_internal and
12444 *movv2sf_internal_rex64 using MMXMODE mode iterator.
12445
225ccc68
SB
124462013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12447
85c0f02d
SB
12448 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
12449 (record_last_mem_set_info): Likewise.
12450
225ccc68
SB
12451 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
12452 of XNEWVEC followed by memset.
12453 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
12454
b64925dc
SB
124552013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12456
12457 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
12458 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
12459 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
12460 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
12461 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
12462 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
12463 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
12464 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
12465 BARRIER_P instead of GET_CODE.
12466
b9aaf52e
EB
124672013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12468
12469 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
12470 inaccuracy in the probing code.
12471
12472 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
12473 (ctrapdi4): Likewise.
12474
66de4d7c
EB
124752013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12476
12477 * calls.c (expand_call): Add missing guard to code handling return
12478 of non-BLKmode structures in MSB.
12479 * function.c (expand_function_end): Likewise.
12480
4f25c8fb
EB
124812013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12482
12483 * combine.c (try_combine): Adjust comment. Do not add the set of
12484 insn #0 if the destination indirectly is set or dies in insn #2.
12485 Tidy up code to distribute a new note.
12486
23b7850d
UB
124872013-03-22 Uros Bizjak <ubizjak@gmail.com>
12488
12489 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
12490 also for alternatives 16 and 17.
12491
bed852cf
UB
124922013-03-22 Uros Bizjak <ubizjak@gmail.com>
12493
12494 * config/i386/sse.md (*mov<mode>_internal): Merge with
12495 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
12496 Emit insn template depending on type attribute. Use
12497 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12498 movd instead of movq mnemonic for interunit moves. Rewrite mode
12499 attribute calculation. Remove unit attribute calculation.
12500 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
12501 Set prefix_data16 attribute for DImode ssemov types.
12502 Use Ym instead of y for SSE-MMX conversion alternatives.
12503 Reorder operand constraints.
12504
1e8a7937
SB
125052013-03-22 Steven Bosscher <steven@gcc.gnu.org>
12506
12507 * df.h (df_insn_delete): Adjust prototype.
12508 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
12509 and let it decide whether mark the basic block dirty.
12510 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
12511 * df-scan.c (df_insn_info_delete): New helper function, split
12512 off from df_insn_delete.
12513 (df_scan_free_bb_info): Use it.
12514 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
12515 Likewise.
12516 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
12517 that the insn is actually an insn and it has a non-NULL basic block.
12518 Do not mark basic block dirty if only deleting a DEBUG_INSN.
12519
e14d094c
RB
125202013-03-22 Richard Biener <rguenther@suse.de>
12521
12522 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
12523 dep_ref members.
12524 (mem_ref_alloc): Do not allocate them.
12525 (refs_independent_p): Do not query or maintain a cache.
12526
e9cf7316
RB
125272013-03-22 Richard Biener <rguenther@suse.de>
12528
12529 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
12530 (gather_mem_refs_in_loops): Do not compute it.
12531 (analyze_memory_references): Do not allocate it.
12532 (tree_ssa_lim_finalize): Do not free it.
12533 (for_all_locs_in_loop): Do not query all_refs_in_loop.
12534
f046e81b
RB
125352013-03-22 Richard Biener <rguenther@suse.de>
12536
12537 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 12538
a05c0ddf
IB
125392013-03-22 Ian Bolton <ian.bolton@arm.com>
12540
12541 * config/aarch64/aarch64.c (aarch64_print_operand): New
12542 format specifier for printing a constant in hex.
12543 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
12544 format specifier for printing second operand.
12545
c00217fc
RB
125462013-03-22 Richard Biener <rguenther@suse.de>
12547
12548 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
12549 bitmaps.
12550 (gather_mem_refs_in_loops): Perform store accumulation here.
12551 (create_vop_ref_mapping_loop): Remove.
12552 (create_vop_ref_mapping): Likewise.
12553 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 12554 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB
12555 (record_indep_loop): Remove.
12556 (record_dep_loop): New function.
12557 (ref_indep_loop_p_1): Adjust to only walk over references
12558 in the loop, not its subloops.
12559 (ref_indep_loop_p): Rename to ...
12560 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
12561 maintaining a more fine-grained cache.
12562 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
12563 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
12564
15d19bf8
RB
125652013-03-22 Richard Biener <rguenther@suse.de>
12566
12567 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
12568 (struct mem_ref): Make accesses_in_loop a vec of a vec of
12569 aggregate mem_ref_loc.
12570 (free_mem_ref_locs): Inline into ...
12571 (memref_free): ... this and adjust.
12572 (mem_ref_alloc): Adjust.
12573 (mem_ref_locs_alloc): Remove.
12574 (record_mem_ref_loc): Adjust.
12575 (get_all_locs_in_loop): Rewrite into ...
12576 (for_all_locs_in_loop): ... this iterator.
12577 (rewrite_mem_ref_loc): New functor.
12578 (rewrite_mem_refs): Use for_all_locs_in_loop.
12579 (sm_set_flag_if_changed): New functor.
12580 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
12581 (ref_always_accessed): New functor.
12582 (ref_always_accessed_p): Use for_all_locs_in_loop.
12583
6f37411d
MG
125842013-03-21 Marc Glisse <marc.glisse@inria.fr>
12585
12586 * tree-pass.h (PROP_gimple_lvec): New.
12587 * passes.c (dump_properties): Handle PROP_gimple_lvec.
12588 (init_optimization_passes): Move pass_lower_vector.
12589 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
12590 PROP_gimple_lvec.
12591 (pass_lower_vector): Provide PROP_gimple_lvec.
12592 (pass_lower_vector_ssa): Likewise.
12593 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
12594
511ef689
MW
125952013-03-21 Mark Wielaard <mjw@redhat.com>
12596
12597 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
12598
cc1df30b
UB
125992013-03-21 Uros Bizjak <ubizjak@gmail.com>
12600
12601 * config/i386/i386.md (*movdi_internal): Disparage slightly
12602 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
12603 conversion alternatives.
12604
c845cfe1
JJ
126052013-03-21 Jakub Jelinek <jakub@redhat.com>
12606
37e99116
JJ
12607 PR middle-end/48087
12608 * diagnostic.def (DK_WERROR): New kind.
12609 * diagnostic.h (werrorcount): Define.
12610 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
12611 promoted to DK_ERROR, increment DK_WERROR counter instead of
12612 DK_ERROR counter.
12613 * toplev.c (toplev_main): Call print_ignored_options even if
12614 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
12615 even if just werrorcount is non-zero.
12616
c845cfe1
JJ
12617 PR debug/55608
12618 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
12619 on failure.
12620 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
12621 (string_cst_pool_decl): New function.
12622 (optimize_one_addr_into_implicit_ptr): New function.
12623 (resolve_addr_in_expr): Optimize DWARF location expression
12624 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
12625 which doesn't live in memory, but has DW_AT_location or
12626 DW_AT_const_value, or refers to a string literal, into
12627 DW_OP_GNU_implicit_pointer.
12628 (optimize_location_into_implicit_ptr): New function.
12629 (resolve_addr): If removing DW_AT_location of a variable because
12630 it was DW_OP_addr of address of the variable, but the variable doesn't
12631 live in memory, try to emit const value attribute for the initializer.
12632
08e0cda6
MG
126332013-03-21 Marc Glisse <marc.glisse@inria.fr>
12634
12635 * tree.h (VECTOR_TYPE_P): New macro.
12636 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
12637 TYPE_MODE): Use it.
12638 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
12639 VEC_COND_EXPR cannot be lvalues.
12640 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
12641
d08633b4
MG
126422013-03-21 Marc Glisse <marc.glisse@inria.fr>
12643
12644 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
12645 Restrict the transformation to equal modes.
12646
e6647190
RB
126472013-03-21 Richard Biener <rguenther@suse.de>
12648
12649 PR tree-optimization/39326
12650 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
12651 (MEM_ANALYZABLE): Adjust.
12652 (record_mem_ref_loc): Move bitmap ops ...
12653 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
12654 unanalyzable refs, do not record locations for it.
12655 (analyze_memory_references): Allocate ref zero as shared
12656 unanalyzable ref.
12657 (refs_independent_p): Do not test for unanalyzed mems here.
12658 (ref_indep_loop_p_1): Special-case disambiguation against
12659 the unanalyzed ref.
cc1df30b 12660 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 12661
65074f54
CL
126622013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
12663
12664 * config/arm/arm-protos.h (tune_params): Add
12665 prefer_neon_for_64bits field.
12666 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
12667 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
12668 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
12669 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
12670 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
12671 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
12672 (arm_option_override): Handle -mneon-for-64bits new option.
12673 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
12674 (prefer_neon_for_64bits): Declare new variable.
12675 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 12676 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL
12677 (arch_enabled): Handle new arch types. Remove support for onlya8
12678 and nota8.
12679 (one_cmpldi2): Use new arch names.
bc5faa5b 12680 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL
12681 * config/arm/arm.opt (mneon-for-64bits): Add option.
12682 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
12683 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
12684 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
12685 of onlya8.
12686 * doc/invoke.texi (-mneon-for-64bits): Document.
12687
5a2d2a79
RB
126882013-03-21 Richard Biener <rguenther@suse.de>
12689
12690 PR tree-optimization/39326
12691 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
12692 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 12693 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 12694
5abe1e05
RB
126952013-03-21 Richard Biener <rguenther@suse.de>
12696
12697 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
12698 (vect_insert_into_interleaving_chain): Likewise.
12699 (vect_drs_dependent_in_basic_block): Inline ...
12700 (vect_slp_analyze_data_ref_dependence): ... here. New function,
12701 split out from ...
12702 (vect_analyze_data_ref_dependence): ... here. Simplify.
12703 (vect_check_interleaving): Simplify.
12704 (vect_analyze_data_ref_dependences): Likewise. Split out ...
12705 (vect_slp_analyze_data_ref_dependences): ... this new function.
12706 (dr_group_sort_cmp): New function.
12707 (vect_analyze_data_ref_accesses): Compute data-reference groups
12708 here instead of in vect_analyze_data_ref_dependence. Use
12709 a more efficient algorithm.
12710 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
12711 vect_slp_analyze_data_ref_dependences. Call
12712 vect_analyze_data_ref_accesses earlier.
12713 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
12714 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
12715 (vect_slp_analyze_data_ref_dependences): New prototype.
12716
cad1735b
RB
127172013-03-21 Richard Biener <rguenther@suse.de>
12718
12719 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
12720 ref is stored in the loop.
12721 (find_refs_for_sm): Walk only over all stores.
12722 (store_motion_loop): Allocate from lim_bitmap_obstack.
12723 (store_motion): Likewise.
12724
141310ef
RB
127252013-03-21 Richard Biener <rguenther@suse.de>
12726
12727 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
12728 Update virtual SSA form.
12729
5022315a
RO
127302013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12731
12732 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
12733 * configure: Regenerate.
12734 * config.in: Regenerate.
12735 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
12736 if !HAVE_LD_EH_FRAME_CIEV3.
12737
50d4421c
RB
127382013-03-21 Richard Biener <rguenther@suse.de>
12739
12740 * tree-cfg.c (verify_expr_no_block): New function.
12741 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
12742 nor DECL_VALUE_EXPR have locations with associated blocks.
12743 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
12744 (clear_unused_block_pointer): Remove code dealing with
12745 blocks in DECL_DEBUG_EXPR locations.
12746
839b422f
RB
127472013-03-21 Richard Biener <rguenther@suse.de>
12748
12749 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
12750 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
12751 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
12752 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
12753 instead of DECL_DEBUG_EXPR_IS_FROM.
12754 * gimplify.c (gimplify_modify_expr): Likewise.
12755 * tree-cfg.c (verify_expr_location_1): Likewise.
12756 * tree-complex.c (create_one_component_var): Likewise.
12757 * tree-sra.c (create_access_replacement): Likewise.
12758 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
12759 (clear_unused_block_pointer): Likewise.
12760 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
12761 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
12762 * var-tracking.c (var_debug_decl): Likewise.
12763 (track_expr_p): Likewise.
12764 * tree-inline.c (add_local_variables): Likewise. Set
12765 DECL_HAS_DEBUG_EXPR_P after copying it.
12766 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
12767 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
12768
fe04878d
UB
127692013-03-21 Uros Bizjak <ubizjak@gmail.com>
12770
12771 PR bootstrap/56656
12772 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
12773 * configure: Regenerate.
12774 * config.in: Regenerate.
12775 * config/i386/i386.md (*movdf_internal): Use
12776 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12777 movd instead of movq mnemonic for interunit moves.
12778 (*movdi_internal): Ditto.
12779
bd059b26 127802013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N
12781
12782 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
12783 (abd<mode>_3): New pattern.
12784 (aba<mode>_3): New pattern.
12785 (fabd<mode>_3): New pattern.
12786
bd059b26 127872013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N
12788
12789 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
12790 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
12791 occurrence of REGISTER_PREFIX as its empty string.
12792
4f1f78b9
JL
127932013-03-20 Jeff Law <law@redhat.com>
12794
12795 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
12796 addititional equivalences for equality comparisons between an SSA_NAME
12797 and a constant where the SSA_NAME was set from a widening conversion.
12798
327a1118
WL
127992013-03-20 Walter Lee <walt@tilera.com>
12800
12801 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
12802
813e0036
UB
128032013-03-20 Uros Bizjak <ubizjak@gmail.com>
12804
12805 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
12806 depending on type attribute.
12807 (*movti_internal): Ditto.
12808 (*movtf_internal): Ditto.
12809 (*movxf_internal): Ditto.
12810 (*movdf_internal): Ditto.
12811 (*movsf_internal): Ditto.
12812
7cf34aae
UB
128132013-03-20 Uros Bizjak <ubizjak@gmail.com>
12814
12815 * config/i386/i386.md (*movti_internal): Set prefix attribute to
12816 maybe_vex for sselog1 and ssemov types.
12817 (*movdi_internal): Reorder operand constraints.
12818 (*movsi_internal): Ditto. Set prefix attribute to
12819 maybe_vex for sselog1 and ssemov types.
12820 (*movtf_internal): Set prefix attribute to maybe_vex
12821 for sselog1 and ssemov types.
12822 (*movdf_internal): Ditto. Set prefix_data16 attribute for
12823 DImode ssemov types. Reorder operand constraints.
12824 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
12825 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
12826 attribute for SImode ssemov types. Reorder operand constraints.
12827
19321415
MJ
128282013-03-20 Martin Jambor <mjambor@suse.cz>
12829
12830 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
12831 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
12832
c6d5ff83
MM
128332013-03-20 Pat Haugen <pthaugen@us.ibm.com>
12834
12835 * config/rs6000/predicates.md (indexed_address, update_address_mem
12836 update_indexed_address_mem): New predicates.
12837 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
12838 attribute for load/store instructions.
12839 * config/rs6000/dfp.md (movsd_store): Likewise.
12840 (movsd_load): Likewise.
12841 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
12842 (unnamed HI->DI extend define_insn): Likewise.
12843 (unnamed SI->DI extend define_insn): Likewise.
12844 (unnamed QI->SI extend define_insn): Likewise.
12845 (unnamed QI->HI extend define_insn): Likewise.
12846 (unnamed HI->SI extend define_insn): Likewise.
12847 (unnamed HI->SI extend define_insn): Likewise.
12848 (extendsfdf2_fpr): Likewise.
12849 (movsi_internal1): Likewise.
12850 (movsi_internal1_single): Likewise.
12851 (movhi_internal): Likewise.
12852 (movqi_internal): Likewise.
12853 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
12854 attribute for load/store instructions.
12855 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
12856 instructions.
12857 (mov<mode>_softfloat): Likewise.
12858 (mov<mode>_hardfloat32): Likewise.
12859 (mov<mode>_hardfloat64): Likewise.
12860 (mov<mode>_softfloat64): Likewise.
12861 (movdi_internal32): Likewise.
12862 (movdi_internal64): Likewise.
12863 (probe_stack_<mode>): Likewise.
12864
128652013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
12866
12867 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
12868 floating point, and decimal floating point to reload iterator.
12869
12870 * config/rs6000/constraints.md (wl constraint): New constraints to
12871 return FLOAT_REGS if certain options are used to reduce the number
12872 of separate patterns that exist in the file.
12873 (wx constraint): Likewise.
12874 (wz constraint): Likewise.
12875
12876 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
12877 -mdebug=reg, print wg, wl, wx, and wz constraints.
12878 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
12879 Initialize the reload functions for 64-bit binary/decimal floating
12880 point types.
12881 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
12882 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
12883 create the buffer on the stack to overcome not having a 32-bit
12884 load and store.
12885 (rs6000_emit_move): Likewise.
12886 (rs6000_secondary_memory_needed_rtx): Likewise.
12887 (rs6000_alloc_sdmode_stack_slot): Likewise.
12888 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
12889 via xxlxor, just like DFmode 0.0.
12890
12891 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
12892 define as 1 if we are running on a power7 or newer.
12893 (enum r6000_reg_class_enum): Add new constraints.
12894
12895 * config/rs6000/dfp.md (movsd): Delete, combine with binary
12896 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
12897 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 12898 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM
12899 (movsd splitter): Likewise.
12900 (movsd_hardfloat): Likewise.
12901 (movsd_softfloat): Likewise.
12902
12903 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
12904 binary and decimal floating point moves.
12905 (fmove_ok): New attributes to combine binary and decimal floating
12906 point moves, and to combine power6x (mfpgpr) moves along normal
12907 floating moves.
12908 (real_value_to_target): Likewise.
12909 (f32_lr): Likewise.
12910 (f32_lm): Likewise.
12911 (f32_li): Likewise.
12912 (f32_sr): Likewise.
12913 (f32_sm): Likewise.
12914 (f32_si): Likewise.
12915 (movsf): Combine binary and decimal floating point moves. Combine
12916 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 12917 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM
12918 (mov<mode> for SFmode/SDmode); Likewise.
12919 (SFmode/SDmode splitters): Likewise.
12920 (movsf_hardfloat): Likewise.
12921 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
12922 (movsf_softfloat): Likewise.
12923 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
12924
12925 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
12926 wx and wz constraints.
12927
12928 * config/rs6000/constraints.md (wg constraint): New constraint to
12929 return FLOAT_REGS if -mmfpgpr (power6x) was used.
12930
12931 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
12932 constraint.
12933
12934 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
12935 -mdebug=reg, print wg, wl, wx, and wz constraints.
12936 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
12937 Initialize the reload functions for 64-bit binary/decimal floating
12938 point types.
12939 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
12940 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
12941 create the buffer on the stack to overcome not having a 32-bit
12942 load and store.
12943 (rs6000_emit_move): Likewise.
12944 (rs6000_secondary_memory_needed_rtx): Likewise.
12945 (rs6000_alloc_sdmode_stack_slot): Likewise.
12946 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
12947 via xxlxor, just like DFmode 0.0.
12948
c6d5ff83
MM
12949 * config/rs6000/dfp.md (movdd): Delete, combine with binary
12950 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
12951 with other moves by using conditional constraits (wg). Use LFIWZX
12952 and STFIWX for loading SDmode on power7.
12953 (movdd splitters): Likewise.
12954 (movdd_hardfloat32): Likewise.
12955 (movdd_softfloat32): Likewise.
12956 (movdd_hardfloat64_mfpgpr): Likewise.
12957 (movdd_hardfloat64): Likewise.
12958 (movdd_softfloat64): Likewise.
12959
12960 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
12961 64-bit binary and decimal floating point moves.
12962 (FMOVE64X): Likewise.
12963 (movdf): Combine 64-bit binary and decimal floating point moves.
12964 Combine power6x (mfpgpr) moves with other moves by using
12965 conditional constraits (wg).
12966 (mov<mode> for DFmode/DDmode): Likewise.
12967 (DFmode/DDmode splitters): Likewise.
12968 (movdf_hardfloat32): Likewise.
12969 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
12970 (movdf_softfloat32): Likewise.
12971 (movdf_hardfloat64_mfpgpr): Likewise.
12972 (movdf_hardfloat64): Likewise.
12973 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
12974 (movdf_softfloat64): Likewise.
12975 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
12976 (reload_<mode>_load): Move to later in the file so they aren't in
12977 the middle of the floating point move insns.
12978 (reload_<mode>_store): Likewise.
12979
12980 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
12981 constraint.
12982
12983 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
12984 constraint if -mdebug=reg.
bd059b26
UB
12985 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
12986 Enable using dd reload support if needed.
c6d5ff83
MM
12987
12988 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
12989 binary and decimal floating point moves in rs6000.md.
12990 (movtd_internal): Likewise.
12991
12992 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
12993 decimal floating point moves.
12994 (movtf): Likewise.
12995 (movtf_internal): Likewise.
12996 (mov<mode>_internal, TDmode/TFmode): Likewise.
12997 (movtf_softfloat): Likewise.
12998 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
12999
13000 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13001 movdi_internal64, using wg constraint for move direct operations.
13002 (movdi_internal64): Likewise.
13003
13004 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13005 MODES_TIEABLE_P for selected modes. Print the numerical value of
13006 the various virtual registers. Use GPR/FPR first/last values,
13007 instead of hard coding the register numbers. Print which modes
13008 have reload functions registered.
bd059b26
UB
13009 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13010 settings before/after setting cpu, target and subtarget settings.
13011 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13012 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM
13013 (rs6000_secondary_reload_fail): Likewise.
13014 (rs6000_secondary_reload_inner): Likewise.
13015
13016 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13017 macros for first/last GPR and FPR registers.
13018 (LAST_GPR_REGNO): Likewise.
13019 (FIRST_FPR_REGNO): Likewise.
13020 (LAST_FPR_REGNO): Likewise.
13021
13022 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13023 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13024 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13025 (vcond<mode><mode>): Likewise.
13026 (vcondu<mode><mode>): Likewise.
13027 (vector_gtu<mode>): Likewise.
13028 (vector_gte<mode>): Likewise.
13029 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 13030 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM
13031 (ior<mode>3): Likewise.
13032 (and<mode>3): Likewise.
13033 (one_cmpl<mode>2): Likewise.
13034 (nor<mode>3): Likewise.
13035 (andc<mode>3): Likewise.
13036
13037 * config/rs6000/constraints.md (wt constraint): New constraint
13038 that returns VSX_REGS if TImode is allowed in VSX registers.
13039
13040 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13041 constant under VSX.
13042
13043 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13044 similar to TImode, but it is restricted to being in the GPRs.
13045
13046 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13047 TImode to occupy a single VSX register.
13048
13049 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13050 -mvsx-timode for power7/power8.
13051 (power7 cpu): Likewise.
13052 (power8 cpu): Likewise.
13053
13054 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13055 sure that TFmode/TDmode take up two registers if they are ever
13056 allowed in the upper VSX registers.
13057 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13058 registers.
13059 (rs6000_init_hard_regno_mode_ok): Likewise.
13060 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13061 constraint. Print if LRA is turned on.
13062 (rs6000_option_override_internal): Give an error if -mvsx-timode
13063 and VSX is not enabled.
13064 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13065 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13066 to reg+offset addressing. Use PTImode when checking offset
13067 addresses for validity.
13068 (reg_offset_addressing_ok_p): Likewise.
13069 (rs6000_legitimate_offset_address_p): Likewise.
13070 (rs6000_legitimize_address): Likewise.
13071 (rs6000_legitimize_reload_address): Likewise.
13072 (rs6000_legitimate_address_p): Likewise.
13073 (rs6000_eliminate_indexed_memrefs): Likewise.
13074 (rs6000_emit_move): Likewise.
13075 (rs6000_secondary_reload): Likewise.
13076 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13077 reloads to fpr registers to continue to use reg+offset addressing,
13078 but 64-bit reloads to altivec registers need reg+reg addressing.
13079 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13080 it. Treat LO_SUM like a PLUS operation.
13081 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 13082 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM
13083 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13084 registers to share a register with a smaller sized type, since VSX
13085 puts scalars in the upper 64-bits.
13086 (print_operand): Add support for PTImode.
13087 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13088 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13089 registers, but don't have arithmetic support.
13090 (rs6000_memory_move_cost): Add test for VSX.
13091 (rs6000_opt_masks): Add -mvsx-timode.
13092
13093 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13094 for TImode.
13095 (VSs): Likewise.
13096 (VSr): Use wt constraint for TImode.
13097 (VSv): Drop TImode support.
13098 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13099 (vsx_movti_64bit): Likewise.
13100 (vsx_movti_32bit): Likewise.
13101 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13102 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13103 one '?' on the appropriate output constraint. Do not allow TImode
13104 logical operations on 32-bit systems.
13105 (vsx_ior<mode>3): Likewise.
13106 (vsx_xor<mode>3): Likewise.
13107 (vsx_one_cmpl<mode>2): Likewise.
13108 (vsx_nor<mode>3): Likewise.
13109 (vsx_andc<mode>3): Likewise.
13110 (vsx_concat_<mode>): Likewise.
13111 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13112
13113 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13114 OPTION_MASK_VSX_TIMODE.
13115 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13116 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13117
13118 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13119 (TI2 iterator): New iterator for TImode, PTImode.
13120 (wd mode attribute): Add values for vector types.
bd059b26
UB
13121 (movti_string): Replace TI move operations with operations for TImode
13122 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM
13123 (mov<mode>_string, TImode/PTImode): Likewise.
13124 (movti_ppc64): Likewise.
13125 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13126 (TI mode splitters): Likewise.
13127
13128 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13129 constraint.
13130
1fc5eced
MG
131312013-03-20 Marc Glisse <marc.glisse@inria.fr>
13132
13133 PR tree-optimization/56355
13134 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13135 Also handle integers with undefined overflow.
13136
22c4c869
CM
131372013-03-20 Catherine Moore <clm@codesourcery.com>
13138 Maciej W. Rozycki <macro@codesourcery.com>
13139 Tom de Vries <tom@codesourcery.com>
13140 Nathan Sidwell <nathan@codesourcery.com>
13141 Iain Sandoe <iain@codesourcery.com>
13142 Nathan Froyd <froydnj@codesourcery.com>
13143 Chao-ying Fu <fu@mips.com>
13144
13145 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 13146 Document new function attributes.
22c4c869
CM
13147 * doc/invoke.texi (minterlink-compressed, mmicromips,
13148 m14k, m14ke, m14kec): Document new options.
13149 (minterlink-mips16): Update documentation.
13150 * doc/md.texi (ZC, ZD): Document new constraints.
13151 * configure.ac (gcc_cv_as_micromips): Check if linker
13152 supports the .set micromips directive.
13153 * configure: Regenerate.
13154 * config.in: Regenerate.
13155 * config/mips/mips-tables.opt: Regenerate.
13156 * config/mips/micromips.md: New file.
13157 * constraints.md (ZC, ZD): New constraints.
13158 * config/mips/predicates.md (movep_src_register): New predicate.
13159 (movep_src_operand): New predicate.
13160 (non_volatile_mem_operand): New predicate.
13161 * config/mips/mips.md (multimem): New type.
13162 (length): Differentiate between 17-bit and 18-bit branch offsets.
13163 (MOVEP1, MOVEP2): New mode iterator.
33159866 13164 (mov_<load>l): Use ZC constraint.
22c4c869
CM
13165 (mov_<load>r): Likewise.
13166 (mov_<store>l): Likewise.
13167 (mov_<store>r): Likewise.
13168 (*branch_equality<mode>_inverted): Add microMIPS support.
13169 (*branch_equality<mode>): Likewise.
13170 (*jump_absolute): Likewise.
13171 (indirect_jump_<mode>): Likewise.
13172 (tablejump_<mode>): Likewise.
13173 (<optab>_internal): Likewise.
13174 (sibcall_internal): Likewise.
13175 (sibcall_value_internal): Likewise.
13176 (prefetch): Use constraint ZD.
13177 * config/mips/mips.opt (minterlink-compressed): New option.
13178 (minterlink-mips16): Now an alias for minterlink-compressed.
13179 (mmicromips): New option.
13180 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13181 (compare_and_swap_12): Likewise.
13182 (sync_add<mode>): Likewise.
13183 (sync_<optab>_12): Likewise.
13184 (sync_old_<optab>_12): Likewise.
13185 (sync_new_<optab>_12): Likewise.
13186 (sync_nand_12): Likewise.
13187 (sync_old_nand_12): Likewise.
13188 (sync_new_nand_12): Likewise.
13189 (sync_sub<mode>): Likewise.
13190 (sync_old_add<mode>): Likewise.
13191 (sync_old_sub<mode>): Likewise.
13192 (sync_new_add<mode>): Likewise.
13193 (sync_new_sub<mode>): Likewise.
13194 (sync_<optab><mode>): Likewise.
13195 (sync_old_<optab><mode>): Likewise.
13196 (sync_new_<optab><mode>): Likewise.
13197 (sync_nand<mode>): Likewise.
13198 (sync_old_nand<mode>): Likewise.
13199 (sync_new_nand<mode>): Likewise.
13200 (sync_lock_test_and_set<mode>): Likewise.
13201 (test_and_set_12): Likewise.
13202 (atomic_compare_and_swap<mode>): Likewise.
13203 (atomic_exchange<mode>_llsc): Likewise.
13204 (atomic_fetch_add<mode>_llsc): Likewise.
13205 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13206 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13207 (umips_save_restore_pattern_p): Likewise.
13208 (umips_load_store_pair_p): Likewise.
13209 (umips_output_load_store_pair): Likewise.
13210 (umips_movep_target_p): Likewise.
13211 (umips_12bit_offset_address_p): Likewise.
13212 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13213 (mips_base_mips16): Rename this...
13214 (mips_base_compression_flags): ...to this. Update all uses.
13215 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13216 (mips_mips16_decl_p): Delete.
13217 (mips_nomips16_decl_p): Delete.
33159866 13218 (mips_get_compress_on_flags): New function.
22c4c869
CM
13219 (mips_get_compress_off_flags): New function.
13220 (mips_get_compress_mode): New function.
13221 (mips_get_compress_on_name): New function.
13222 (mips_get_compress_off_name): New function.
13223 (mips_insert_attributes): Support multiple compression types.
13224 (mips_merge_decl_attributes): Likewise.
13225 (umips_12bit_offset_address_p): New function.
13226 (mips_start_function_definition): Emit .set micromips directive.
13227 (mips_call_may_need_jalx_p): New function.
13228 (mips_function_ok_for_sibcall): Add microMIPS support.
13229 (mips_print_operand_punctuation): Support short delay slots and
13230 compact jumps.
13231 (umips_swm_mask, umips_swm_encoding): New.
13232 (umips_build_save_restore): New function.
13233 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13234 (was_mips16_p): Remove.
13235 (old_compression_mode): New.
13236 (mips_set_compression_mode): New function.
13237 (mips_set_current_function): Add microMIPS support.
13238 (mips_option_override): Likewise.
13239 (umips_save_restore_pattern_p): New function.
13240 (umips_output_save_restore): New function.
13241 (umips_load_store_pair_p_1): New function.
13242 (umips_load_store_pair_p): New function.
13243 (umips_output_load_store_pair_1): New function.
13244 (umips_output_load_store_pair): New function.
13245 (umips_movep_target_p) New function.
13246 (mips_prepare_pch_save): Add microMIPS support.
13247 * config/mips/mips.h (TARGET_COMPRESSION): New.
13248 (TARGET_CPU_CPP_BUILTINS): Update macro
13249 to use new compression flags and to support microMIPS.
13250 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13251 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13252 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13253 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13254 (ASM_SPEC): Support mmicromips and mno-micromips.
13255 (M16STORE_REG_P): New macro.
13256 (MIPS_CALL): Support TARGET_MICROMIPS.
13257 (MICROMIPS_J): New macro.
13258 (mips_base_mips16): Rename this...
13259 (mips_base_compression_flags): ...to this.
13260 (UMIPS_12BIT_OFFSET_P): New macro.
13261 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13262 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB
132632013-03-20 Richard Biener <rguenther@suse.de>
13264
13265 PR tree-optimization/56661
13266 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13267 the result does not have to be distinct.
13268
54714c68
RB
132692013-03-20 Richard Biener <rguenther@suse.de>
13270
13271 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13272 remap_gimple_op_r.
13273
cca1130d
BS
132742013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13275 Steven Bosscher <steven@gcc.gnu.org>
13276
13277 PR rtl-optimization/56605
13278 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13279
2a930377
UB
132802013-03-20 Uros Bizjak <ubizjak@gmail.com>
13281
13282 PR bootstrap/56656
13283 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13284 that require movd instead of movq.
13285
d6d305fe
RB
132862013-03-20 Richard Biener <rguenther@suse.de>
13287
13288 * tree-ssa-structalias.c (struct variable_info): Add pointer
13289 to the first field of an aggregate with sub-vars. Make
13290 this and the pointer to the next subfield its ID.
13291 (vi_next): New function.
13292 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13293 storedanything_id, integer_id): Increment by one.
13294 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13295 get_call_clobber_vi): Adjust.
13296 (solution_set_expand): Simplify and speedup.
13297 (solution_set_add): Inline into ...
13298 (set_union_with_increment): ... this. Adjust accordingly.
13299 (do_sd_constraint): Likewise.
13300 (do_ds_constraint): Likewise.
13301 (do_complex_constraint): Simplify.
13302 (build_pred_graph): Adjust.
13303 (solve_graph): Likewise. Simplify and speedup.
13304 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13305 get_constraint_for_component_ref, get_constraint_for_1,
13306 first_vi_for_offset, first_or_preceding_vi_for_offset,
13307 create_function_info_for, create_variable_info_for_1,
13308 create_variable_info_for, intra_create_variable_infos): Adjust.
13309 (init_base_vars): Push NULL for ID zero.
13310 (compute_points_to_sets): Adjust.
13311
a271b42d
RB
133122013-03-20 Richard Biener <rguenther@suse.de>
13313
13314 * cfgloop.c (verify_loop_structure): Streamline and avoid
13315 ICEing on corrupt loop tree.
13316 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13317 loop tree.
13318
ebd65954
RB
133192013-03-20 Richard Biener <rguenther@suse.de>
13320
13321 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13322 check whether an SSA update is needed.
13323
4547b7ee
RS
133242013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13325
13326 * config/mips/constraints.md (T): Rename to...
13327 (Yf): ...this.
13328 (U): Rename to...
13329 (Yd): ...this.
13330 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13331 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13332
0b8cdc58
IB
133332013-03-19 Ian Bolton <ian.bolton@arm.com>
13334
13335 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13336 (*subsi3_carryin_uxtw): Likewise.
13337
5977a10d
IB
133382013-03-19 Ian Bolton <ian.bolton@arm.com>
13339
13340 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13341 (*rorsi3_insn_uxtw): Likewise.
13342
bd83ff2c
IB
133432013-03-19 Ian Bolton <ian.bolton@arm.com>
13344
13345 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
13346 (*extrsi5_insn_uxtw): Likewise.
13347
ba65123c
RB
133482013-03-19 Richard Biener <rguenther@suse.de>
13349
13350 PR tree-optimization/56273
13351 * passes.c (init_optimization_passes): Move second VRP after DOM.
13352
2eac0476
UB
133532013-03-19 Uros Bizjak <ubizjak@gmail.com>
13354
13355 * config/i386/i386.md (*movti_internal): Merge from
13356 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
13357 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
13358 nox64 isa attributes.
13359
6e55eda7
RB
133602013-03-18 Richard Biener <rguenther@suse.de>
13361
13362 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
13363 (unite): Likewise.
13364 (merge_node_constraints): Likewise.
13365 (build_succ_graph): Likewise.
13366 (valid_graph_edge): Inline into single caller.
13367 (unify_nodes): Likewise. Use bitmap_set_bit return value
13368 and cache varinfo.
13369 (scc_visit): Fix formatting and variable use.
13370 (do_sd_constraint): Use gcc_checking_assert.
13371 (do_ds_constraint): Likewise.
13372 (do_complex_constraint): Likewise.
13373 (condense_visit): Likewise. Cleanup.
13374 (dump_pred_graph): New function.
13375 (perform_var_substitution): Dump the pred-graph before
13376 variable substitution.
13377 (find_equivalent_node): Use gcc_checking_assert.
13378 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
13379
4bdd44c4
RB
133802013-03-18 Richard Biener <rguenther@suse.de>
13381
13382 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13383 Remove cond_expr_stmt_list argument and do not gimplify the
13384 built expression.
13385 (vect_loop_versioning): Adjust.
13386 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
13387 Cleanup to use less temporaries.
13388 (vect_create_data_ref_ptr): Cleanup.
13389
38c56a5b
JJ
133902013-03-18 Jakub Jelinek <jakub@redhat.com>
13391
13392 PR tree-optimization/56635
13393 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
13394 require types_compatible_p types.
13395
20b2e6a0
NC
133962013-03-18 Nick Clifton <nickc@redhat.com>
13397
a6178a25
NC
13398 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
13399 spurious backslash.
13400
20b2e6a0
NC
13401 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
13402 Add missing line to comment describing function.
13403
92608d0e
RB
134042013-03-18 Richard Biener <rguenther@suse.de>
13405
13406 PR tree-optimization/56210
13407 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
13408 Handle string / character search functions.
13409 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
13410
31348d52
RB
134112013-03-18 Richard Biener <rguenther@suse.de>
13412
13413 PR middle-end/56483
13414 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
13415 and implement properly.
13416 * gimple.h (gimple_cond_single_var_p): Remove.
13417
fcac74a1
RB
134182013-03-18 Richard Biener <rguenther@suse.de>
13419
13420 * tree-data-ref.h (find_data_references_in_loop): Declare.
13421 * tree-data-ref.c (get_references_in_stmt): Use a stack
13422 vector pre-allocated in the callers.
13423 (find_data_references_in_stmt): Adjust.
13424 (graphite_find_data_references_in_stmt): Likewise.
13425 (create_rdg_vertices): Likewise.
13426 (find_data_references_in_loop): Export.
13427 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
13428 Compute dependences here...
13429 (vect_analyze_data_refs): ...not here. When we encounter
13430 a non-vectorizable data reference in basic-block vectorization
13431 truncate the data reference vector. Do not bother to
13432 fixup data-dependence information for gather loads.
13433 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
13434 of data references, as reported.
13435
0d5a1b56
RB
134362013-03-18 Richard Biener <rguenther@suse.de>
13437
13438 PR tree-optimization/3713
13439 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
13440 has_constants and expr.
13441 (stmt_has_constants): Properly valueize SSA names when deciding
13442 whether the stmt has constants.
13443
789c34e3
RB
134442013-03-18 Richard Biener <rguenther@suse.de>
13445
13446 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
13447 whole function when there is nothing to do.
13448 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
13449 * tree-vectorizer.c (vectorize_loops): Update virtual and
13450 loop-closed SSA once.
13451 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
13452
076b4605
RB
134532013-03-18 Richard Biener <rguenther@suse.de>
13454
13455 PR middle-end/56113
13456 * domwalk.c (bb_postorder): New global static.
13457 (cmp_bb_postorder): New function.
13458 (walk_dominator_tree): Replace scheme imposing an order for
13459 visiting dominator sons by one sorting them at the time they
13460 are pushed on the stack.
13461
bdb01696
RB
134622013-03-18 Richard Biener <rguenther@suse.de>
13463
13464 PR tree-optimization/39326
13465 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
13466 (struct mem_ref): Replace mem member with ao_ref typed member.
13467 (MEM_ANALYZABLE): Adjust.
13468 (memref_eq): Likewise.
13469 (mem_ref_alloc): Likewise.
13470 (gather_mem_refs_stmt): Likewise.
13471 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
13472 (execute_sm_if_changed_flag_set): Adjust.
13473 (execute_sm): Likewise.
13474 (ref_always_accessed_p): Likewise.
13475 (refs_independent_p): Likewise.
13476 (can_sm_ref_p): Likewise.
13477
12d2dc5e
JJ
134782013-03-18 Jakub Jelinek <jakub@redhat.com>
13479
13480 PR c/56566
13481 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
13482 return 1 even for !unsignedp.
13483
286fb677
UB
134842013-03-17 Uros Bizjak <ubizjak@gmail.com>
13485
13486 * config/i386/i386.md (isa): Add x64 and nox64.
13487 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
13488 (*pushtf): Enable *roF alternative for x64 isa only.
13489 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
13490 mode attribute of integer alternatives to DImode for TARGET_64BIT.
13491 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
13492 (*movtf_internal): Merge from *movtf_internal_rex64 and
13493 *movtf_internal_sse. Use x64 and nox64 isa attributes.
13494 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
13495 nox64 isa attributes.
13496 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
13497 nox64 isa attributes.
13498 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
13499
88b97037
UB
135002013-03-17 Uros Bizjak <ubizjak@gmail.com>
13501
13502 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13503
9f4f1735
JJ
135042013-03-17 Jakub Jelinek <jakub@redhat.com>
13505
13506 PR target/56640
13507 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
13508 class names. Remove trailing comma after "ALL_REGS".
13509
48f4a6fa
JH
135102013-03-16 Jan Hubicka <jh@suse.cz>
13511
13512 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
13513 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
13514 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
13515 of cgraph_get_create_node.
13516 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
13517
98e81123
JM
135182013-03-16 Jason Merrill <jason@redhat.com>
13519
13520 PR debug/49090
13521 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
13522 with DW_AT_default_value.
13523
8a298c04
JJ
135242013-03-16 Jakub Jelinek <jakub@redhat.com>
13525
13526 * BASE-VER: Set to 4.9.0.
13527
4323afa0
AK
135282013-03-14 Andi Kleen <ak@linux.intel.com>
13529
13530 PR target/56619
13531 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
13532 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
13533 Document _x* TSX intrinsics.
13534
b3c0d469
JJ
135352013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
13536 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI
13537
13538 * configure.ac: Add MicroBlaze TLS support detection.
13539 * configure: Regenerate.
88b97037
UB
13540 * config/microblaze/microblaze-protos.h
13541 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
13542 symbol_mentioned_p, label_mentioned_p): Add prototypes.
13543 * config/microblaze/microblaze.c (microblaze_address_type): Add
13544 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI
13545 (microblaze_address_info): Add tls_reloc.
13546 (TARGET_HAVE_TLS): Define.
13547 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
13548 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB
13549 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
13550 load_tls_operand, microblaze_call_tls_get_addr,
13551 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI
13552 (microblaze_classify_unspec): Handle UNSPEC_TLS.
13553 (get_base_reg): Use microblaze_tls_symbol_p.
13554 (microblaze_classify_address): Handle TLS.
88b97037
UB
13555 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
13556 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI
13557 (microblaze_legitimize_address): Handle TLS.
13558 (microblaze_address_insns): Handle ADDRESS_TLS.
13559 (pic_address_needs_scratch): Handle TLS.
13560 (print_operand_address): Handle TLS.
13561 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
13562 (microblaze_expand_move): Handle TLS.
88b97037
UB
13563 (microblaze_legitimate_constant_p): Check
13564 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI
13565 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
13566 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
13567 (PIC_OFFSET_TABLE_REGNUM): Set.
13568 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
13569 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
13570 (addsi3, movsi_internal2, movdf_internal): Update constraints
13571 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB
13572 (move_operand): Redefine as move_src_operand,
13573 check microblaze_tls_referenced_p.
8cc9a5a5 13574
d803a491
IB
135752013-03-14 Ian Bolton <ian.bolton@arm.com>
13576
13577 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
13578 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
13579
a8504f22
IB
135802013-03-14 Ian Bolton <ian.bolton@arm.com>
13581
13582 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
13583 CC mode for AND.
13584
df2dfaea
JJ
135852013-03-14 Jakub Jelinek <jakub@redhat.com>
13586
fbd28bc3
JJ
13587 PR tree-optimization/53265
13588 * common.opt (Waggressive-loop-optimizations): New option.
13589 * tree-ssa-loop-niter.c: Include tree-pass.h.
13590 (do_warn_aggressive_loop_optimizations): New function.
13591 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
13592 if number_of_latch_executions returned constant.
13593 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
13594 early. If number_of_latch_executions returned constant, set
13595 nb_iterations_upper_bound back to it.
13596 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
13597 field.
13598 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
13599 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
13600
df2dfaea
JJ
13601 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
13602 (MULTILIB_OSDIRNAMES): Set.
13603 * genmultilib: If defaultosdirname doesn't start with :: , set
13604 defaultosdirname2 instead, clear it and emit two . multilib_raw
13605 entries instead of just one.
13606
ee0d2b68
KK
136072013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
13608
13609 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13610 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
13611 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13612 (SUBTARGET_OVERRIDE_OPTIONS): New.
13613
decc676e
OE
136142013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
13615
13616 PR target/49880
13617 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
13618 (musermode): Convert to Var(TARGET_USERMODE).
13619 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
13620 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
13621 * config/sh/sh.c (sh_option_override): Use
13622 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
13623 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
13624 condition.
13625 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
13626 TARGET_SH4.
13627 (udivsi3_i4_single, divsi3_i4_single): Use
13628 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
13629
f4b719c7
DK
136302013-03-13 Dave Korn <dave.korn.cygwin@....>
13631
13632 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
13633 default setting.
13634
c40eced0
RB
136352013-03-13 Richard Biener <rguenther@suse.de>
13636
13637 PR tree-optimization/56608
13638 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
13639 calls when vectorizing basic-blocks.
13640
1bfa5973
JJ
136412013-03-13 Jakub Jelinek <jakub@redhat.com>
13642
13643 PR plugins/45078
13644 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
13645 tm_file.
13646
6fcf5434
JJ
136472013-03-12 Jakub Jelinek <jakub@redhat.com>
13648
13649 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
13650
4f38fa8c
JH
136512013-03-11 Jan Hubicka <jh@suse.cz>
13652
13653 PR lto/56557
13654 * lto-streamer-out.c (output_symbol_p): Skip references from
13655 constructors of external variables.
13656
c5c5ba89
JH
136572013-03-11 Jan Hubicka <jh@suse.cz>
13658
13659 PR middle-end/56571
13660 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
13661 from pseudos.
13662 * emit-rtl.c (verify_rtx_sharing): Likewise.
13663 (copy_insn_1): Likewise.
13664 * rtl.c (copy_rtx): Likewise.
13665
c2a939b1
GJL
136662013-03-11 Georg-Johann Lay <avr@gjlay.de>
13667
13668 PR target/56591
13669 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
13670 output_operand_lossage message.
88b97037 13671
3c3279fb
RE
136722013-03-11 Richard Earnshaw <rearnsha@arm.com>
13673
13674 PR target/56470
13675 * arm.c (shift_op): Validate RTL pattern on the fly.
13676 (arm_print_operand, case 'S'): Don't use shift_operator to validate
13677 the RTL.
13678
aef5ef9d
JDA
136792013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13680
13681 PR target/56347
13682 * config/pa/pa.md (call_value): Check for calls to powf and direct to
13683 new call patterns that clobber %fr12.
13684 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
13685 split and postreload patterns.
13686 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
13687 registers %fr12 and %fr12R as call used.
13688
456610d3
SB
136892013-03-09 Steven Bosscher <steven@gcc.gnu.org>
13690
13691 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
13692 (canon_address, record_store, replace_read, check_mem_read_rtx,
13693 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
13694 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
13695 rest_of_handle_dse): Likewise.
13696
4b1baac8
RS
136972013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
13698
13699 PR middle-end/56524
13700 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
13701 Add base_optabs.
13702 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
13703 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
13704 (save_optabs_if_changed): Replace with...
13705 (init_tree_optimization_optabs): ...this.
13706 * optabs.c (save_optabs_if_changed): Rename to...
13707 (init_tree_optimization_optabs): ...this. Take the optimization node
13708 as argument. Do nothing if the base optabs are already correct.
13709 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
13710 to recompute optabs.
13711 * function.h (function): Remove optabs field.
13712 * function.c (invoke_set_current_function_hook): Call
13713 init_tree_optimization_optabs. Use the result to initialize
13714 this_fn_optabs.
13715
b7a78683
AH
137162013-02-27 Aldy Hernandez <aldyh@redhat.com>
13717
13718 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
13719 if GTMA_HAS_NO_INSTRUMENTATION.
13720 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
13721 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
13722 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
13723 * gimple-pretty-print.c (dump_gimple_transaction): Handle
13724 GTMA_HAS_NO_INSTRUMENTATION.
13725
6384c29b
JJ
137262013-03-08 Jakub Jelinek <jakub@redhat.com>
13727
13728 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
13729 libasan_preinit.o.
13730
ecd4f20a
MP
137312013-03-08 Marek Polacek <polacek@redhat.com>
13732 Jakub Jelinek <jakub@redhat.com>
13733
13734 PR tree-optimization/56478
13735 * predict.c (is_comparison_with_loop_invariant_p): Change the
13736 type of loop_step to tree.
13737 (predict_loops): Adjust.
13738 (predict_iv_comparison): Perform the computations on double_ints.
13739
64366d35
RB
137402013-03-08 Richard Biener <rguenther@suse.de>
13741
13742 PR tree-optimization/56570
13743 * tree-cfg.c (verify_expr_location_1): Verify locations for
13744 DECL_DEBUG_EXPR.
13745 * tree-sra.c (create_access_replacement): Strip locations
13746 from DECL_DEBUG_EXPRs.
13747
a9d5a059
RB
137482013-03-08 Richard Biener <rguenther@suse.de>
13749
13750 * tree-inline.c (expand_call_inline): Do not associate
13751 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
13752 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
13753
b7aa4e9a
RB
137542013-03-08 Richard Biener <rguenther@suse.de>
13755
13756 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
13757 or block changes with -Og. Fix for location / block encoding
13758 changes and PHI arguments with locations.
13759
c4c2f9fa
SB
137602013-03-07 Steven Bosscher <steven@gcc.gnu.org>
13761
13762 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
13763 for all counters.
13764 (struct output_info): Likewise.
13765 (register_overhead): Remove bad gcc_assert.
13766 (bitmap_find_bit): If there is only a single bitmap element, do not
13767 count a miss as a search.
13768 (print_statistics): Update for counter type changes.
13769 (dump_bitmap_statistics): Likewise. Print headers such that they
13770 are properly lined up with the printed counters.
13771
5bf6606a
JJ
137722013-03-07 Jakub Jelinek <jakub@redhat.com>
13773
13774 PR tree-optimization/56559
13775 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
13776 check that it has only a single use.
13777
2c653d46
RB
137782013-03-07 Richard Biener <rguenther@suse.de>
13779
13780 * doc/invoke.texi (fwhole-program): Discourage use in combination
13781 with -flto.
13782
a72d8780
JJ
137832013-03-06 Jakub Jelinek <jakub@redhat.com>
13784
01a454df
JJ
13785 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
13786
c1781047
JJ
13787 PR tree-optimization/56539
13788 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
13789 instead of GSI_CONTINUE_LINKING as last argument to
13790 force_gimple_operand_gsi. Adjust function comment.
13791
9772c47a
JJ
13792 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
13793 aarch64-cores.def.
13794
a72d8780
JJ
13795 PR middle-end/56548
13796 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
13797 promoted mode, convert the result back to the original mode.
13798
fa5556de
RB
137992013-03-06 Richard Biener <rguenther@suse.de>
13800
13801 PR middle-end/56294
13802 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
13803 (insert_updated_phi_nodes_compare_uids): New function.
13804 (update_ssa): Sort symbols_to_rename after UID before
13805 traversing it to insert PHI nodes.
13806
010403d1
RB
138072013-03-06 Richard Biener <rguenther@suse.de>
13808
13809 PR middle-end/50494
13810 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
13811 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
13812
13813 Revert
13814 2013-02-13 Richard Biener <rguenther@suse.de>
13815
13816 PR lto/50494
13817 * varasm.c (output_constant_def_1): Get the decl representing
13818 the constant as argument.
13819 (output_constant_def): Wrap output_constant_def_1.
13820 (make_decl_rtl): Use output_constant_def_1 with the decl
13821 representing the constant.
13822 (build_constant_desc): Optionally re-use a decl already
13823 representing the constant.
13824 (tree_output_constant_def): Adjust.
13825
3c27ce4c
JY
138262013-03-06 Joey Ye <joey.ye@arm.com>
13827
13828 PR lto/50293
13829 * gcc.c (convert_white_space): New function.
13830 (main): Handles white space in function name.
13831
8f6d1c86
OE
138322013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
13833
13834 PR target/56529
13835 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
13836 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
13837 to SH_DIV_CALL_TABLE for TARGET_SH2.
13838 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
13839 list.
13840 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
13841 call-table options.
13842
798a209f
SA
138432013-03-05 Sterling Augustine <saugustine@google.com>
13844 Cary Coutant <ccoutant@google.com>
13845
13846 PR debug/55364
13847 * dwarf2out.c (resolve_addr): Don't call
13848 remove_loc_list_addr_table_entries a second time for the same
13849 expression.
13850
6cfa417f 138512013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 13852
6cfa417f
JJ
13853 PR debug/56510
13854 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
13855 (avoid_complex_debug_insns): New function.
13856 (expand_debug_locations): Call it.
13857
13858 PR rtl-optimization/56484
13859 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
13860 lifetimes of hard registers on small register class machines.
13861
138622013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
13863
13864 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 13865 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 13866 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH
13867 fast_interrupt.
13868 (microblaze_fast_interrupt_function_p): New function.
13869 (microblaze_is_interrupt_handler): Rename to
13870 microblaze_is_interrupt_variant and add fast_interrupt check.
13871 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
13872 (save_restore_insns): Likewise.
13873 (compute_frame_size): Likewise.
13874 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
13875 (microblaze_globalize_label): Likewise.
6cfa417f
JJ
13876 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
13877 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH
13878 microblaze_is_interrupt_variant.
13879
731aefac
KT
138802013-03-05 Kai Tietz <ktietz@redhat.com>
13881
13882 * sdbout.c (sdbout_one_type): Switch to current function's section
13883 supporting cold/hot.
13884
a72d8780 138852013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH
13886
13887 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
13888 -mxl-reorder.
13889
334e71e8
JJ
138902013-03-05 Jakub Jelinek <jakub@redhat.com>
13891
0b50e654
JJ
13892 PR middle-end/56461
13893 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
13894 if VALGRIND_GET_VBITS is defined, temporarily make object
13895 memory all defined, and restore previous valgrind addressability
13896 and definability afterwards. Free this_object at the end.
13897
4ccf8f43
JJ
13898 PR middle-end/56461
13899 * lra.c (lra): Call lra_clear_live_ranges if live_p,
13900 right before calling lra_create_live_ranges, also call it
13901 when clearing live_p. Only call lra_clear_live_ranges
13902 at the end if live_p.
13903
334e71e8
JJ
13904 PR middle-end/56461
13905 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
13906
9ca966ca
RB
139072013-03-05 Richard Biener <rguenther@suse.de>
13908
13909 PR tree-optimization/56521
13910 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
13911 value-id.
13912
d273b176
SB
139132013-03-05 Steven Bosscher <steven@gcc.gnu.org>
13914
13915 PR c++/55135
13916 * except.h (remove_unreachable_eh_regions): New prototype.
13917 * except.c (remove_eh_handler_splicer): New function, split out
13918 of remove_eh_handler.
13919 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
13920 warning about running it on many EH regions one at a time.
13921 (remove_unreachable_eh_regions_worker): New function, walk the
13922 EH tree in depth-first order and remove non-marked regions.
13923 (remove_unreachable_eh_regions): New function.
13924 * tree-eh.c (mark_reachable_handlers): New function, split out
13925 from remove_unreachable_handlers.
13926 (remove_unreachable_handlers): Use mark_reachable_handlers and
13927 remove_unreachable_eh_regions.
13928 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
13929 and remove_unreachable_eh_regions.
13930
5e7f1aef
RB
139312013-03-05 Richard Biener <rguenther@suse.de>
13932
13933 PR middle-end/56525
13934 * loop-init.c (fix_loop_structure): Remove loops in two stages,
13935 not freeing them until the end.
13936
f276b762
AK
139372013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13938
13939 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
13940
12a43ab8
RB
139412013-03-05 Richard Biener <rguenther@suse.de>
13942
13943 PR tree-optimization/56270
13944 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
13945 of loads after scheduling an SLP instance.
13946
93675444
JJ
139472013-03-05 Jakub Jelinek <jakub@redhat.com>
13948
db4138e3
JJ
13949 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
13950 tic6x.exp.
13951 (check_gcc_parallelize): Run guality.exp as a separate job from
13952 vect.exp with unsorted.exp and $(dg_target_exps) separately from
13953 struct-layout-1.exp with stackalign.exp.
13954
dd3d1ec0
JJ
13955 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
13956
b8d381a3
JJ
13957 PR middle-end/56461
13958 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
13959 load_index sbitmap even if some bit in it isn't set.
13960
b4f9786b
JJ
13961 PR middle-end/56461
13962 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
13963 (discover_iteration_bound_by_body_walk): Change queues to
13964 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
13965 spelling in comment. Call safe_push on queues[bound_index] directly.
13966 Release queues[queue_index] in every iteration unconditionally.
13967 Release bounds vector.
13968
93675444
JJ
13969 PR middle-end/56461
13970 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
13971 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
13972 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
13973 inner_phis vector.
13974
3e492e9c
RB
139752013-03-05 Richard Biener <rguenther@suse.de>
13976
13977 PR lto/56515
13978 * tree-inline.c (remap_blocks_to_null): New function.
13979 (expand_call_inline): When expanding a call stmt without
13980 an associated block inline remap all callee blocks to NULL.
13981
a5d1569a
JJ
139822013-03-05 Jakub Jelinek <jakub@redhat.com>
13983
86efb5cd
JJ
13984 PR rtl-optimization/56494
13985 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
13986 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
13987 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
13988
85f5dbea
JJ
13989 PR middle-end/56461
13990 * sel-sched-ir.c (free_sched_pools): Release
13991 succs_info_pool.stack[succs_info_pool.max_top] vectors too
13992 if succs_info_pool.max_top isn't -1.
13993
a5d1569a
JJ
13994 PR bootstrap/56509
13995 * opts.c (opts_obstack, opts_concat): Moved to...
13996 * opts-common.c (opts_obstack, opts_concat): ... here.
13997
4432aa6c
JJ
139982013-03-04 Jakub Jelinek <jakub@redhat.com>
13999
14000 PR middle-end/56461
14001 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14002
f3a1fb91
MJ
140032013-03-04 Martin Jambor <mjambor@suse.cz>
14004
14005 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14006 all appropriate places.
14007
22110e6c
EB
140082013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14009
14010 PR tree-optimization/56424
14011 * ipa-split.c (split_function): Do not set the RSO flag if result is
14012 not by reference and its type is a register type.
14013
a72d8780 140142013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 14015
88b97037 14016 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 14017 (microblaze_legitimate_pic_operand): Likewise
88b97037 14018 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH
14019 new function microblaze_legitimate_pic_operand
14020 * config/microblaze/microblaze-protos.h
14021 (microblaze_legitimate_pic_operand): Declare.
14022
a72d8780 140232013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 14024
a72d8780 14025 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 14026 New predicate for supported rtx code types.
a72d8780 14027 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI
14028 call_insn_simple_operand predicate.
14029
541d9ac8
JJ
140302013-03-04 Jakub Jelinek <jakub@redhat.com>
14031
5eb010bc
JJ
14032 PR middle-end/56461
14033 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14034 partitions.ordered_remove.
14035
30862efc
JJ
14036 PR middle-end/56461
14037 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14038 vec_oprnds0.create (1) for modifier == NONE.
14039
8930f723
JJ
14040 PR middle-end/56461
14041 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14042 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14043 vec_oprnds1 right before pushing anything to it for
14044 scalar_shift_arg.
14045
541d9ac8
JJ
14046 PR middle-end/56461
14047 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14048 set nbbs to 0 instead of having separate code path.
14049 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14050 instead of false as last argument if returning NULL.
14051
7aa7f2e3
SL
140522013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14053
14054 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14055 the attribute is now called "target" instead of "option".
14056 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14057 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14058 attribute/pragma name for TARGET_OPTION_VALID_P and
14059 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14060 * doc/tm.texi: Regenerated.
14061
8930f723 140622013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 14063
8930f723 14064 * config/microblaze/microblaze.c:
cb8a1637 14065 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ
14066 * config/microblaze/microblaze.h: Add -mxl-reorder to
14067 DRIVER_SELF_SPECS.
14068 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14069 instructions emitted if TARGET_REORDER.
88b97037
UB
14070 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14071 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 14072 separately.
cb8a1637 14073
4dc1d68c
XDL
140742013-03-01 Xinliang David Li <davidxl@google.com>
14075
14076 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14077 walk length.
14078
689eaba3
JJ
140792013-03-01 Jakub Jelinek <jakub@redhat.com>
14080
9d676bf6
JJ
14081 PR middle-end/56461
14082 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14083 vector even when returning true. Fix up function comment formatting.
14084
caff7edf
JJ
14085 PR middle-end/56461
14086 * ira-build.c (ira_loop_nodes_count): New variable.
14087 (create_loop_tree_nodes): Initialize it.
14088 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14089
b6b9227d
JJ
14090 PR middle-end/56461
14091 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14092 method on dr_chain and result_chain.
14093 * tree-vect-stmts.c (vectorizable_store): Only call
14094 result_chain.create if j == 0.
14095
689eaba3
JJ
14096 PR middle-end/56461
14097 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14098 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14099 before overwriting it.
14100
40bf31ed
TB
141012013-03-01 Tobias Burnus <burnus@net-b.de>
14102
14103 * doc/extended.texi (C Extensions): Change order in @menu
14104 to match @node.
14105 (Other MIPS Built-in Functions): Move last MIPS entry before
14106 "picoChip Built-in Functions".
14107 (SH Built-in Functions): Move after RX Built-in Functions.
14108 * doc/gcc.texi (Introduction): Change order in @menu
14109 to match @node.
14110 * doc/md.texi (Constraints): Ditto.
14111 * gty.texi (Type Information): Ditto.
14112 (User-provided marking routines for template types): Make
14113 subsection.
14114 * doc/invoke.texi (AArch64 Options): Move before
14115 "Adapteva Epiphany Options".
14116
e664c61c
KS
141172013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14118 Jakub Jelinek <jakub@redhat.com>
14119
14120 PR sanitizer/56454
14121 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14122 no_address_safety_analysis attribute.
14123 * doc/extend.texi (no_address_safety_attribute): Rename to
14124 no_sanitize_address attribute, mention no_address_safety_analysis
14125 attribute as deprecated alias.
14126
37b5ec8f
JJ
141272013-02-28 Jakub Jelinek <jakub@redhat.com>
14128
14129 PR middle-end/56461
14130 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14131 type to vec<vec<tree> > *.
14132 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14133 to be vec<tree> instead of vec<tree> *, set vec_defs
14134 to vNULL and call vec_defs.create (number_of_vects), adjust other
14135 uses of vec_defs.
14136 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14137 vectorizable_condition): Adjust vect_get_slp_defs callers.
14138
ba96cdfb
JG
141392013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14140
14141 * config/aarch64/aarch64.c
14142 (aarch64_float_const_representable): Remove unused variable.
14143
6f549691
JG
141442013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14145
14146 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14147
af55e82d
JG
141482013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14149
14150 * config/aarch64/aarch64-builtins.c
14151 (aarch64_init_simd_builtins): Make static.
14152
1df3f464
JG
141532013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14154
14155 * config/aarch64/aarch64.c
14156 (aarch64_simd_make_constant): Make static.
14157
f8f42513
MJ
141582013-02-28 Martin Jambor <mjambor@suse.cz>
14159
14160 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14161 with no initialization to the RHS of debug statements.
14162
b48b3fc4
MJ
141632013-02-28 Martin Jambor <mjambor@suse.cz>
14164
14165 PR tree-optimization/56294
14166 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14167 Adjust dumping.
14168 (get_access_replacement): Do not call create_access_replacement.
14169 Assert a replacement exists.
14170 (get_repl_default_def_ssa_name): Create the replacement declaration
14171 itself.
14172
c3ae224c
RR
141732013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14174
14175 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14176 final_end_function.
14177
45fa0eef
MP
141782013-02-28 Marek Polacek <polacek@redhat.com>
14179
14180 PR rtl-optimization/56466
14181 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14182 if we're changing a loop.
14183 (peel_loops_completely): Likewise.
14184
502c067d
PC
141852013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14186
14187 PR c++/55813
14188 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14189
ff24afc8
GJL
141902013-02-28 Georg-Johann Lay <avr@gjlay.de>
14191
14192 PR target/56445
14193 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14194 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14195 INTX_FTYPE_FX, FX_FTYPE_INTX.
14196 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14197
c7b1fc1b
GJL
141982013-02-28 Georg-Johann Lay <avr@gjlay.de>
14199
14200 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14201 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14202 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14203 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14204 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14205 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14206 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14207 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14208 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14209 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14210 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14211 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14212 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14213 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14214 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14215 (avrxmega6): Increase max flash segments from 5 to 6.
14216 * config/avr/t-multilib: Regenerate.
14217 * config/avr/avr-tables.opt: Regenerate.
14218 * doc/avr-mmcu.texi: Regenerate.
14219
4a0e3cfe
GJL
142202013-02-28 Georg-Johann Lay <avr@gjlay.de>
14221
14222 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14223 (avr_device_to_arch): Rename to avr_device_to_ld.
14224 (avr_device_to_as): New prototype.
14225 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14226 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14227 * config/avr/driver-avr.c (avr_device_to_as): New.
14228 (avr_device_to_arch): Rename to avr_device_to_ld.
14229
97785e52
JJ
142302013-02-27 Jakub Jelinek <jakub@redhat.com>
14231
3f292312
JJ
14232 PR middle-end/56461
14233 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14234 method on dr_chain and result_chain.
14235
a344216b
JJ
14236 PR middle-end/56461
14237 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14238 pointer_set_destroy on not_executed_last_iteration.
14239
f121ad02 14240 PR middle-end/56461
88b97037 14241 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 14242
307f83a3
JJ
14243 PR middle-end/56461
14244 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14245 FOR_EACH_DEFINED_FUNCTION when freeing state.
14246
e19624ee
JJ
14247 PR middle-end/56461
14248 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14249 pool_free.
14250 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14251 overwriting it.
14252
90e709fd
JJ
14253 PR middle-end/56461
14254 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14255 known_aggs[i].items and release known_aggs vector.
14256
97785e52
JJ
14257 PR middle-end/56461
14258 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14259
142602013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 14261
97785e52
JJ
14262 * config/microblaze/microblaze.c (microblaze_emit_compare):
14263 Use xor for EQ/NE comparisions.
14264 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14265 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 14266
6fa5e0ed
JJ
142672013-02-27 Jakub Jelinek <jakub@redhat.com>
14268
14269 PR middle-end/56461
14270 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14271 vector.
14272 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14273 vec_safe_push, always update *slot.
14274 (redirect_edge_var_map_clear): Use vec_free.
14275 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14276 (free_var_map_entry): Use vec_free.
14277 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14278 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14279
436a956a
AB
142802013-02-27 Andrey Belevantsev <abel@ispras.ru>
14281
14282 PR middle-end/45472
436a956a
AB
14283 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14284 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB
14285 Reorder tests for speculativeness in the logical and operator.
14286
f5c2caca
JJ
142872013-02-27 Jakub Jelinek <jakub@redhat.com>
14288
0fcb564b
JJ
14289 * incpath.c (add_standard_paths): Use reconcat instead of concat
14290 where appropriate and avoid leaking memory.
14291
dc357798
JJ
14292 * opts.h: Include obstack.h.
14293 (opts_concat): New prototype.
14294 (opts_obstack): New declaration.
14295 * opts.c (opts_concat): New function.
14296 (opts_obstack): New variable.
14297 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14298 (finish_options): Use opts_concat instead of concat
14299 and XOBNEWVEC instead of XNEWVEC.
14300 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14301 generate_option): Likewise.
14302 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14303 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14304
f5c2caca
JJ
14305 PR target/56455
14306 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14307 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14308
d0163673
JJ
143092013-02-26 Jakub Jelinek <jakub@redhat.com>
14310
14311 PR middle-end/56461
14312 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14313
b5ad2b8e
JR
143142013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14315
14316 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14317 (arm_block_move_unaligned_straight): Likewise.
14318 (arm_adjust_block_mem): Likewise.
14319
9b639e2c
JR
143202013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14321
14322 PR target/48901
14323 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14324 temp, cond and label.
14325 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14326
e13a0ccb
JR
14327 PR target/52500
14328 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14329 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14330
fbe4f171
JR
14331 PR target/52501
14332 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14333 of prologue/epilogue functions.
14334
ae006543
JR
14335 PR target/52550
14336 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14337 Remove unused variable cfa_offset.
14338 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14339
c81369fa
JR
14340 PR target/54639
14341 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14342 type promotion to unsigned.
14343
f8a8fea7
JR
14344 PR target/54640
14345 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
14346 for HOST_WIDE_INT of 32 bit / same size as int.
14347 (arm_block_move_unaligned_straight): Likewise.
14348 (arm_adjust_block_mem): Likewise.
14349
f8be5169
JR
14350 PR target/54662
14351 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
14352 ALL_CFLAGS.
14353
f1ad3354
MP
143542013-02-26 Marek Polacek <polacek@redhat.com>
14355
14356 PR tree-optimization/56426
88b97037 14357 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 14358
a6af0f42
RB
143592013-02-26 Richard Biener <rguenther@suse.de>
14360
14361 PR target/56444
14362 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
14363 unused variable loops.
14364
aca43c6c
JJ
143652013-02-26 Jakub Jelinek <jakub@redhat.com>
14366
cecbe5d9
JJ
14367 PR tree-optimization/56448
14368 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
14369 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
14370 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
14371 later operands of the references, or even first operand for
14372 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
14373
aca43c6c
JJ
14374 PR tree-optimization/56443
14375 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
14376 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
14377 to type_for_mode langhook.
14378
53e2e141
MT
143792013-02-25 Matt Turner <mattst88@gmail.com>
14380
14381 * doc/invoke.texi: Document r4700.
14382
259ee451
RB
143832013-02-25 Richard Biener <rguenther@suse.de>
14384
14385 PR tree-optimization/56175
14386 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
14387 split out from ...
14388 (simplify_bitwise_binary): ... here. Also guard the conversion
14389 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
14390
6d65e8f1
CM
143912013-02-25 Catherine Moore <clm@codesourcery.com>
14392
14393 Revert:
14394 2013-02-24 Catherine Moore <clm@codesourcery.com>
14395 Maciej W. Rozycki <macro@codesourcery.com>
14396 Tom de Vries <tom@codesourcery.com>
a72d8780 14397 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM
14398 Iain Sandoe <iain@codesourcery.com>
14399 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 14400 Chao-ying Fu <fu@mips.com>
88b97037 14401
6d65e8f1 14402 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 14403 Document new function attributes.
6d65e8f1
CM
14404 * doc/invoke.texi (minterlink-compressed, mmicromips,
14405 m14k, m14ke, m14kec): Document new options.
14406 (minterlink-mips16): Update documentation.
14407 * doc/md.texi (ZC, ZD): Document new constraints.
14408 * configure.ac (gcc_cv_as_micromips): Check if linker
14409 supports the .set micromips directive.
14410 * configure: Regenerate.
14411 * config.in: Regenerate.
14412 * config/mips/mips-tables.opt: Regenerate.
14413 * config/mips/micromips.md: New file.
14414 * constraints.md (ZC, AD): New constraints.
14415 * config/mips/predicates.md (movep_src_register): New predicate.
14416 (movep_src_operand): New predicate.
14417 (non_volatile_mem_operand): New predicate.
14418 * config/mips/mips.md (multimem): New type.
14419 (length): Differentiate between 17-bit and 18-bit branch offsets.
14420 (MOVEP1, MOVEP2): New mode iterator.
14421 (mov_<load>l): Use ZC constraint.
14422 (mov_<load>r): Likewise.
14423 (mov_<store>l): Likewise.
14424 (mov_<store>r): Likewise.
14425 (*branch_equality<mode>_inverted): Add microMIPS support.
14426 (*branch_equality<mode>): Likewise.
14427 (*jump_absolute): Likewise.
14428 (indirect_jump_<mode>): Likewise.
14429 (tablejump_<mode>): Likewise.
14430 (<optab>_internal): Likewise.
14431 (sibcall_internal): Likewise.
14432 (sibcall_value_internal): Likewise.
14433 (prefetch): Use constraint ZD.
14434 * config/mips/mips.opt (minterlink-compressed): New option.
14435 (minterlink-mips16): Now an alias for minterlink-compressed.
14436 (mmicromips): New option.
14437 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14438 (compare_and_swap_12): Likewise.
14439 (sync_add<mode>): Likewise.
14440 (sync_<optab>_12): Likewise.
14441 (sync_old_<optab>_12): Likewise.
14442 (sync_new_<optab>_12): Likewise.
14443 (sync_nand_12): Likewise.
14444 (sync_old_nand_12): Likewise.
14445 (sync_new_nand_12): Likewise.
14446 (sync_sub<mode>): Likewise.
14447 (sync_old_add<mode>): Likewise.
14448 (sync_old_sub<mode>): Likewise.
14449 (sync_new_add<mode>): Likewise.
14450 (sync_new_sub<mode>): Likewise.
14451 (sync_<optab><mode>): Likewise.
14452 (sync_old_<optab><mode>): Likewise.
14453 (sync_new_<optab><mode>): Likewise.
14454 (sync_nand<mode>): Likewise.
14455 (sync_old_nand<mode>): Likewise.
14456 (sync_new_nand<mode>): Likewise.
14457 (sync_lock_test_and_set<mode>): Likewise.
14458 (test_and_set_12): Likewise.
14459 (atomic_compare_and_swap<mode>): Likewise.
14460 (atomic_exchange<mode>_llsc): Likewise.
14461 (atomic_fetch_add<mode>_llsc): Likewise.
14462 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14463 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14464 (umips_save_restore_pattern_p): Likewise.
14465 (umips_load_store_pair_p): Likewise.
14466 (umips_output_load_store_pair): Likewise.
14467 (umips_movep_target_p): Likewise.
14468 (umips_12bit_offset_address_p): Likewise.
14469 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14470 (mips_base_mips16): Rename this...
14471 (mips_base_compression_flags): ...to this. Update all uses.
14472 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14473 (mips_mips16_decl_p): Delete.
14474 (mips_nomips16_decl_p): Delete.
14475 (mips_get_compress_on_flags): New function.
14476 (mips_get_compress_off_flags): New function.
14477 (mips_get_compress_mode): New function.
14478 (mips_get_compress_on_name): New function.
14479 (mips_get_compress_off_name): New function.
14480 (mips_insert_attributes): Support multiple compression types.
14481 (mips_merge_decl_attributes): Likewise.
14482 (umips_12bit_offset_address_p): New function.
14483 (mips_start_function_definition): Emit .set micromips directive.
14484 (mips_call_may_need_jalx_p): New function.
14485 (mips_function_ok_for_sibcall): Add microMIPS support.
14486 (mips_print_operand_punctuation): Support short delay slots and
14487 compact jumps.
14488 (umips_swm_mask, umips_swm_encoding): New.
14489 (umips_build_save_restore): New function.
14490 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14491 (was_mips16_p): Remove.
14492 (old_compression_mode): New.
14493 (mips_set_compression_mode): New function.
14494 (mips_set_current_function): Add microMIPS support.
14495 (mips_option_override): Likewise.
14496 (umips_save_restore_pattern_p): New function.
14497 (umips_output_save_restore): New function.
14498 (umips_load_store_pair_p_1): New function.
14499 (umips_load_store_pair_p): New function.
14500 (umips_output_load_store_pair_1): New function.
14501 (umips_output_load_store_pair): New function.
14502 (umips_movep_target_p) New function.
14503 (mips_prepare_pch_save): Add microMIPS support.
14504 * config/mips/mips.h (TARGET_COMPRESSION): New.
14505 (TARGET_CPU_CPP_BUILTINS): Update macro
14506 to use new compression flags and to support microMIPS.
14507 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14508 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14509 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14510 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14511 (ASM_SPEC): Support mmicromips and mno-micromips.
14512 (M16STORE_REG_P): New macro.
14513 (MIPS_CALL): Support TARGET_MICROMIPS.
14514 (MICROMIPS_J): New macro.
14515 (mips_base_mips16): Rename this...
14516 (mips_base_compression_flags): ...to this.
14517 (UMIPS_12BIT_OFFSET_P): New macro.
14518 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14519 (MULTILIB_DIRNAMES): Likewise.
14520
5e5df392
TV
145212013-02-25 Tom de Vries <tom@codesourcery.com>
14522
14523 PR rtl-optimization/56131
14524 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
14525 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
14526 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
14527
3fdb53c1
TB
145282013-02-25 Tobias Burnus <burnus@net-b.de>
14529
14530 * doc/invoke.texi (-fsanitize=): Move from optimization
14531 to debugging options.
14532
ed358aea
AB
145332013-02-25 Andrey Belevantsev <abel@ispras.ru>
14534
14535 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
14536
f45e9053
AB
145372013-02-25 Andrey Belevantsev <abel@ispras.ru>
14538 Alexander Monakov <amonakov@ispras.ru>
14539
14540 PR middle-end/56077
14541 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
14542 flush pending lists also on non-jumps. Adjust comment.
14543
6941b508
CM
145442013-02-24 Catherine Moore <clm@codesourcery.com>
14545 Maciej W. Rozycki <macro@codesourcery.com>
14546 Tom de Vries <tom@codesourcery.com>
a72d8780 14547 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM
14548 Iain Sandoe <iain@codesourcery.com>
14549 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 14550 Chao-ying Fu <fu@mips.com>
6941b508
CM
14551
14552 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 14553 Document new function attributes.
6941b508
CM
14554 * doc/invoke.texi (minterlink-compressed, mmicromips,
14555 m14k, m14ke, m14kec): Document new options.
14556 (minterlink-mips16): Update documentation.
14557 * doc/md.texi (ZC, ZD): Document new constraints.
14558 * configure.ac (gcc_cv_as_micromips): Check if linker
14559 supports the .set micromips directive.
14560 * configure: Regenerate.
14561 * config.in: Regenerate.
14562 * config/mips/mips-tables.opt: Regenerate.
14563 * config/mips/micromips.md: New file.
14564 * constraints.md (ZC, AD): New constraints.
14565 * config/mips/predicates.md (movep_src_register): New predicate.
14566 (movep_src_operand): New predicate.
14567 (non_volatile_mem_operand): New predicate.
14568 * config/mips/mips.md (multimem): New type.
14569 (length): Differentiate between 17-bit and 18-bit branch offsets.
14570 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 14571 (mov_<load>l): Use ZC constraint.
6941b508
CM
14572 (mov_<load>r): Likewise.
14573 (mov_<store>l): Likewise.
14574 (mov_<store>r): Likewise.
14575 (*branch_equality<mode>_inverted): Add microMIPS support.
14576 (*branch_equality<mode>): Likewise.
14577 (*jump_absolute): Likewise.
14578 (indirect_jump_<mode>): Likewise.
14579 (tablejump_<mode>): Likewise.
14580 (<optab>_internal): Likewise.
14581 (sibcall_internal): Likewise.
14582 (sibcall_value_internal): Likewise.
14583 (prefetch): Use constraint ZD.
14584 * config/mips/mips.opt (minterlink-compressed): New option.
14585 (minterlink-mips16): Now an alias for minterlink-compressed.
14586 (mmicromips): New option.
14587 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14588 (compare_and_swap_12): Likewise.
14589 (sync_add<mode>): Likewise.
14590 (sync_<optab>_12): Likewise.
14591 (sync_old_<optab>_12): Likewise.
14592 (sync_new_<optab>_12): Likewise.
14593 (sync_nand_12): Likewise.
14594 (sync_old_nand_12): Likewise.
14595 (sync_new_nand_12): Likewise.
14596 (sync_sub<mode>): Likewise.
14597 (sync_old_add<mode>): Likewise.
14598 (sync_old_sub<mode>): Likewise.
14599 (sync_new_add<mode>): Likewise.
14600 (sync_new_sub<mode>): Likewise.
14601 (sync_<optab><mode>): Likewise.
14602 (sync_old_<optab><mode>): Likewise.
14603 (sync_new_<optab><mode>): Likewise.
14604 (sync_nand<mode>): Likewise.
14605 (sync_old_nand<mode>): Likewise.
14606 (sync_new_nand<mode>): Likewise.
14607 (sync_lock_test_and_set<mode>): Likewise.
14608 (test_and_set_12): Likewise.
14609 (atomic_compare_and_swap<mode>): Likewise.
14610 (atomic_exchange<mode>_llsc): Likewise.
14611 (atomic_fetch_add<mode>_llsc): Likewise.
14612 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14613 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14614 (umips_save_restore_pattern_p): Likewise.
14615 (umips_load_store_pair_p): Likewise.
14616 (umips_output_load_store_pair): Likewise.
14617 (umips_movep_target_p): Likewise.
14618 (umips_12bit_offset_address_p): Likewise.
14619 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14620 (mips_base_mips16): Rename this...
14621 (mips_base_compression_flags): ...to this. Update all uses.
14622 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14623 (mips_mips16_decl_p): Delete.
14624 (mips_nomips16_decl_p): Delete.
3fdb53c1 14625 (mips_get_compress_on_flags): New function.
6941b508
CM
14626 (mips_get_compress_off_flags): New function.
14627 (mips_get_compress_mode): New function.
14628 (mips_get_compress_on_name): New function.
14629 (mips_get_compress_off_name): New function.
14630 (mips_insert_attributes): Support multiple compression types.
14631 (mips_merge_decl_attributes): Likewise.
14632 (umips_12bit_offset_address_p): New function.
14633 (mips_start_function_definition): Emit .set micromips directive.
14634 (mips_call_may_need_jalx_p): New function.
14635 (mips_function_ok_for_sibcall): Add microMIPS support.
14636 (mips_print_operand_punctuation): Support short delay slots and
14637 compact jumps.
14638 (umips_swm_mask, umips_swm_encoding): New.
14639 (umips_build_save_restore): New function.
14640 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14641 (was_mips16_p): Remove.
14642 (old_compression_mode): New.
14643 (mips_set_compression_mode): New function.
14644 (mips_set_current_function): Add microMIPS support.
14645 (mips_option_override): Likewise.
14646 (umips_save_restore_pattern_p): New function.
14647 (umips_output_save_restore): New function.
14648 (umips_load_store_pair_p_1): New function.
14649 (umips_load_store_pair_p): New function.
14650 (umips_output_load_store_pair_1): New function.
14651 (umips_output_load_store_pair): New function.
14652 (umips_movep_target_p) New function.
14653 (mips_prepare_pch_save): Add microMIPS support.
14654 * config/mips/mips.h (TARGET_COMPRESSION): New.
14655 (TARGET_CPU_CPP_BUILTINS): Update macro
14656 to use new compression flags and to support microMIPS.
14657 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14658 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14659 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14660 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14661 (ASM_SPEC): Support mmicromips and mno-micromips.
14662 (M16STORE_REG_P): New macro.
14663 (MIPS_CALL): Support TARGET_MICROMIPS.
14664 (MICROMIPS_J): New macro.
14665 (mips_base_mips16): Rename this...
14666 (mips_base_compression_flags): ...to this.
14667 (UMIPS_12BIT_OFFSET_P): New macro.
14668 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14669 (MULTILIB_DIRNAMES): Likewise.
14670
cdebbc6d
JJ
146712013-02-24 Jakub Jelinek <jakub@redhat.com>
14672
14673 PR target/52555
14674 * target-globals.c (save_target_globals): For init_reg_sets and
14675 target_reinit remporarily set this_fn_optabs to this_target_optabs.
14676
18c63565
JG
146772013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
14678
14679 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
14680 * config/aarch64/t-aarch64
14681 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
14682
2194f7a2
VM
146832013-02-22 Vladimir Makarov <vmakarov@redhat.com>
14684
14685 PR inline-asm/56148
14686 * lra-constraints.c (process_alt_operands): Reload operand
14687 conflicting with earlier clobber only if no more other conflicting
14688 operands.
14689
7d613735
JJ
146902013-02-22 Jakub Jelinek <jakub@redhat.com>
14691
14692 PR sanitizer/56393
14693 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
14694 if not linking a shared library.
14695
ac8d93a7
SL
146962013-02-22 Seth LaForge <sethml@google.com>
14697
14698 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
14699
e0237780
GY
147002013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
14701
3fdb53c1
TB
14702 * config/arm/arm.md (split for extendsidi): Update condition.
14703 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
14704 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
14705 (qhs_zextenddi_cstr): Likewise.
e0237780 14706
d7fde18c
JJ
147072013-02-21 Jakub Jelinek <jakub@redhat.com>
14708
be63b77d
JJ
14709 PR middle-end/56420
14710 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
14711 avoid signed wrapping.
14712 (expand_mult): Handle properly multiplication by
14713 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
14714 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
14715 in the compiler if coeff is HOST_WIDE_INT_MIN.
14716 (expand_divmod): Don't make ext_op1 static, change it's type to
14717 uhwi. Avoid undefined behavior in -INTVAL (op1).
14718
d7fde18c
JJ
14719 PR rtl-optimization/50339
14720 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
14721 field.
14722 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
14723 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
14724 into splitting_ashiftrt field.
14725 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
14726 ASHIFTRT.
14727 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
14728 choices.
14729
6aad4455
AH
147302013-02-20 Aldy Hernandez <aldyh@redhat.com>
14731
14732 PR middle-end/56108
14733 * trans-mem.c (execute_tm_mark): Do not expand transactions that
14734 are sure to go irrevocable.
14735
38fe784d
HPN
147362013-02-21 Hans-Peter Nilsson <hp@axis.com>
14737
14738 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
14739 scalars are valid operands.
14740
0fd44da3
MJ
147412013-02-21 Martin Jambor <mjambor@suse.cz>
14742
14743 PR tree-optimization/56310
14744 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
14745 only matching indices and non-negative final offsets.
14746 (intersect_aggregates_with_edge): Pass src_idx to
14747 agg_replacements_to_vector. Pass src_idx insstead of index to
14748 intersect_with_agg_replacements.
14749
7a92038b
MJ
147502013-02-21 Martin Jambor <mjambor@suse.cz>
14751
14752 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
14753 instead of hard-wired defaults.
14754
c0da9c37
MR
147552013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
14756
14757 * doc/invoke.texi (MIPS Options): Update documentation of the
14758 floating-point multiply-accumulate instruction restrictions.
14759
d247ea0c 147602013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS
14761
14762 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
14763 asan_shadow_offset on x86_64 linux.
14764
22deefcb
RB
147652013-02-21 Richard Biener <rguenther@suse.de>
14766
14767 PR tree-optimization/56415
14768 Revert
14769 2013-02-11 Richard Biener <rguenther@suse.de>
14770
14771 PR tree-optimization/56273
14772 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
14773 first VRP run.
14774
7bcc6e75
JJ
147752013-02-21 Jakub Jelinek <jakub@redhat.com>
14776
a0ad148f
JJ
14777 PR bootstrap/56258
14778 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
14779 instead of @itemx.
14780
7bcc6e75
JJ
14781 PR inline-asm/56405
14782 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
14783 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
14784
a0a7b611
JH
147852013-02-20 Jan Hubicka <jh@suse.cz>
14786
14787 PR tree-optimization/56265
88b97037
UB
14788 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
14789 when target is referenced for first time.
a0a7b611 14790
c0e50f72
RB
147912013-02-20 Richard Biener <rguenther@suse.de>
14792
14793 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
14794 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
14795 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
14796 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
14797 not return anything.
14798 (rename_ssa_copies): Do not remove unused locals.
14799 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 14800 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB
14801 * passes.c (execute_function_todo): Do not schedule unused locals
14802 removal if cleanup_tree_cfg did something.
14803 * tree-ssa-live.c (remove_unused_locals): Dump statistics
14804 about the number of removed locals.
14805
a52ca739
RB
148062013-02-20 Richard Biener <rguenther@suse.de>
14807
14808 PR tree-optimization/56398
88b97037 14809 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 14810
ec9c9d1b
MJ
148112013-02-20 Martin Jambor <mjambor@suse.cz>
14812
14813 PR tree-optimization/55334
14814 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
14815 restricted pointers to arrays.
14816
e91c8ed6 148172013-02-20 Richard Biener <rguenther@suse.de>
88b97037 14818 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB
14819
14820 PR tree-optimization/56396
14821 * tree-ssa-ccp.c (n_const_val): New static variable.
14822 (get_value): Return NULL for SSA names we don't have a lattice
14823 entry for.
14824 (ccp_initialize): Initialize n_const_val.
14825 * tree-ssa-copy.c (n_copy_of): New static variable.
14826 (init_copy_prop): Initialize n_copy_of.
14827 (get_value): Return NULL_TREE for SSA names we don't have a
14828 lattice entry for.
14829
3d916479
MJ
148302013-02-20 Martin Jambor <mjambor@suse.cz>
14831
14832 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
14833
71a86758
RB
148342013-02-20 Richard Biener <rguenther@suse.de>
14835
14836 * genpreds.c (write_lookup_constraint): Do not compare first
14837 letter of the constraint again.
14838
79836a12
RB
148392013-02-20 Richard Biener <rguenther@suse.de>
14840
14841 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
14842 and ceil_log2.
14843 (get_use_iv_cost): Terminate hashtable walk when coming across
14844 an empty entry.
14845
bbe4fb2c
IZ
148462013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
14847
14848 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
14849 reassociation for avx2 targets.
14850
a72d8780 148512012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 14852
c77f83d5
EI
14853 * config/microblaze/microblaze.c: microblaze_has_clz = 0
14854 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 14855 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI
14856 version and TARGET_PATTERN_COMPARE check
14857 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 14858
a72d8780 148592012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 14860
a72d8780 14861 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI
14862 function before branching.
14863
242387fa
AB
148642012-02-19 Andrey Belevantsev <abel@ispras.ru>
14865
14866 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
14867 DUMP_INSN_RTX_UID.
14868 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
14869
b4979ab9
AB
148702012-02-19 Andrey Belevantsev <abel@ispras.ru>
14871
14872 PR middle-end/55889
b4979ab9
AB
14873 * sel-sched.c: Include ira.h.
14874 (implicit_clobber_conflict_p): New function.
14875 (moveup_expr): Use it.
88b97037 14876 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 14877
24d63016
RB
148782013-02-19 Richard Biener <rguenther@suse.de>
14879
14880 PR tree-optimization/56384
14881 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
14882 (vn_hash_type): Split out from ...
14883 (vn_hash_constant_with_type): ... here.
14884 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
14885 (vn_phi_eq): Compare types from vn_phi_s structure.
14886 (vn_phi_lookup): Populate vn_phi_s type.
14887 (vn_phi_insert): Likewise.
14888
a475fd3d
JJ
148892013-02-19 Jakub Jelinek <jakub@redhat.com>
14890
47cc28f5
JJ
14891 PR tree-optimization/56350
14892 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
14893 if haven't found reduction or nested cycle operand, rather than
14894 asserting we must find it.
14895
a475fd3d
JJ
14896 PR tree-optimization/56381
14897 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
14898 to fold_build3.
14899
149002013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH
14901 Jakub Jelinek <jakub@redhat.com>
14902
14903 PR target/52555
14904 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
14905 (swap_optab_enable): Same.
14906 (init_all_optabs): Use argument instead of global.
88b97037 14907 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH
14908 * expr.h (init_all_optabs): Add argument to prototype.
14909 (TREE_OPTIMIZATION_OPTABS): New.
14910 (save_optabs_if_changed): Protoize.
14911 * optabs.h: Declare this_fn_optabs.
14912 * optabs.c (save_optabs_if_changed): New.
14913 Declare this_fn_optabs.
14914 (init_optabs): Add argument to init_all_optabs() call.
14915 * function.c (invoke_set_current_function_hook): Handle per
14916 function optabs.
14917 * function.h (struct function): New field optabs.
14918 * config/mips/mips.c (mips_set_mips16_mode): Handle when
14919 optimization_current_node has changed.
14920 * target-globals.h (save_target_globals_default_opts): Protoize.
14921 * target-globals.c (save_target_globals_default_opts): New.
14922
3f587ca3
JDA
149232013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14924
14925 PR target/56347
14926 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
14927 registers %fr12 and %fr12R as call used.
ceaca33e
JDA
14928
14929 PR target/56214
14930 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB
14931 and HImode, require all displacements to be an integer multiple of
14932 their mode size.
ceaca33e
JDA
14933 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
14934 only allow QImode and HImode when reload is in progress and strict is
14935 true. Likewise for symbolic addresses. Use base14_operand to check
14936 displacements in REG+BASE addresses.
14937
fe0b4796
RB
149382013-02-18 Richard Biener <rguenther@suse.de>
14939
14940 PR tree-optimization/56366
14941 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14942 handle sign-conversion of outer-loop initial induction value.
14943
6aaf596b
RB
149442013-02-18 Richard Biener <rguenther@suse.de>
14945
73db8ff1 14946 PR middle-end/56349
6aaf596b
RB
14947 * cfghooks.c (merge_blocks): If we merge a latch into another
14948 block adjust references to it.
14949 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
14950 (verify_loop_structure): Verify that a recorded latch is in fact
14951 a latch.
14952
5e97dfb6
RB
149532013-02-18 Richard Biener <rguenther@suse.de>
14954
14955 PR tree-optimization/56321
14956 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
14957 order SSA name release and virtual operand unlinking.
14958
825527e8
EI
149592013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14960
14961 * config/microblaze/microblaze.md (save_stack_block): Define.
14962 (restore_stack_block): Likewise.
14963
debd11d9
EI
149642013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14965
14966 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
14967 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
14968 * config/microblaze/microblaze.c (microblaze_option_override):
14969 Bail out early for PIC modes when target does not support PIC.
14970
8ec77be0
EI
149712013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14972
88b97037 14973 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI
14974 Replace with a microblaze version.
14975 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 14976 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI
14977 microblaze.
14978
b41288b3
JJ
149792013-02-16 Jakub Jelinek <jakub@redhat.com>
14980 Dodji Seketeli <dodji@redhat.com>
14981
14982 PR asan/56330
88b97037 14983 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ
14984 (instrument_mem_region_access): Do not forget to always put
14985 instrumentation of the of 'base' and 'base + len' in a "if (len !=
14986 0) statement, even for cases where either 'base' or 'base + len'
14987 are not instrumented -- because they have been previously
14988 instrumented. Simplify the logic by putting all the statements
14989 instrument 'base + len' inside a sequence, and then insert that
14990 sequence right before the current insertion point. Then, to
14991 instrument 'base + len', just get an iterator on that statement.
14992 And do not forget to update the pointer to iterator the function
14993 received as argument.
14994
47918951
VM
149952013-02-15 Vladimir Makarov <vmakarov@redhat.com>
14996
14997 PR rtl-optimization/56348
14998 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
14999
3fb7c699
SB
150002013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15001
15002 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15003 (clean_graph_dump_file): Pass base to start_graph_dump.
15004
576fe41a
RH
150052013-02-14 Richard Henderson <rth@redhat.com>
15006
15007 PR target/55941
15008 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15009
bb3accfa
SB
150102013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15011
15012 * collect2-aix.h: Define F_LOADONLY.
15013
03143140
RB
150142013-02-14 Richard Biener <rguenther@suse.de>
15015
15016 PR lto/50494
15017 * varasm.c (output_constant_def_1): Get the decl representing
15018 the constant as argument.
15019 (output_constant_def): Wrap output_constant_def_1.
15020 (make_decl_rtl): Use output_constant_def_1 with the decl
15021 representing the constant.
15022 (build_constant_desc): Optionally re-use a decl already
15023 representing the constant.
15024 (tree_output_constant_def): Adjust.
15025
8fb06726
DS
150262013-02-14 Dodji Seketeli <dodji@redhat.com>
15027
15028 Fix an asan crash
15029 * asan.c (instrument_builtin_call): Really put the length of the
15030 second source argument into src1_len.
15031
c4bfe8bf
JJ
150322013-02-13 Jakub Jelinek <jakub@redhat.com>
15033
15034 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15035 argument. If it is false, don't create edge from then_bb to
15036 fallthru_bb.
15037 (insert_if_then_before_iter): Pass true to it.
15038 (build_check_stmt): Pass false to it.
15039 (transform_statements): Flush hash table only on extended basic
15040 block boundaries, rather than at the beginning of every bb.
15041 Don't flush hash table on nonfreeing_call_p calls.
15042 * tree-flow.h (nonfreeing_call_p): New prototype.
15043 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15044
7afe2801
DM
150452013-02-13 David S. Miller <davem@davemloft.net>
15046
15047 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15048
70cc3288
VM
150492013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15050
15051 PR target/56184
15052 * ira.c (max_regno_before_ira): Move from ...
15053 (ira): ... here.
15054 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15055 vec_safe_length.
15056
6422242b
JJ
150572013-02-13 Jakub Jelinek <jakub@redhat.com>
15058
15059 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15060
976a81ee
RB
150612013-02-13 Richard Biener <rguenther@suse.de>
15062
15063 PR lto/56295
15064 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15065 globals in MEM_REFs.
15066
c1874a87
RB
150672013-02-13 Richard Biener <rguenther@suse.de>
15068
15069 * loop-init.c (loop_optimizer_init): Clear loop state when
15070 re-initializing preserved loops.
15071 * loop-unswitch.c (unswitch_single_loop): Return whether
15072 we unswitched the loop. Do not verify loop state here.
88b97037 15073 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 15074
b4ab7d34
KS
150752013-02-13 Kostya Serebryany <kcc@google.com>
15076
88b97037
UB
15077 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15078 on x86_64 linux.
b4ab7d34
KS
15079 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15080
bdcbe80c
DS
150812013-02-12 Dodji Seketeli <dodji@redhat.com>
15082
15083 Avoid instrumenting duplicated memory access in the same basic block
15084 * Makefile.in (asan.o): Add new dependency on hash-table.h
15085 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15086 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15087 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15088 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15089 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15090 (get_mem_ref_of_assignment): New functions.
15091 (get_mem_refs_of_builtin_call): Extract from
15092 instrument_builtin_call and tweak a little bit to make it fit with
15093 the new signature.
15094 (instrument_builtin_call): Use the new
15095 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15096 of is_gimple_builtin_call.
15097 (instrument_derefs, instrument_mem_region_access): Insert the
15098 instrumented memory reference into the hash table.
15099 (maybe_instrument_assignment): Renamed instrument_assignment into
15100 this, and change it to advance the iterator when instrumentation
15101 actually happened and return true in that case. This makes it
15102 homogeneous with maybe_instrument_assignment, and thus give a
15103 chance to callers to be more 'regular'.
15104 (transform_statements): Clear the memory reference hash table
15105 whenever we enter a new BB, when we cross a function call, or when
15106 we are done transforming statements. Use
15107 maybe_instrument_assignment instead of instrumentation. No more
15108 need to special case maybe_instrument_assignment and advance the
15109 iterator after calling it; it's now handled just like
15110 maybe_instrument_call. Update comment.
15111
4861a1f7
RB
151122013-02-13 Richard Biener <rguenther@suse.de>
15113
15114 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15115 Fix loop discovery code.
15116
1c86bd80
VM
151172013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15118
15119 PR inline-asm/56148
15120 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 15121 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM
15122 the operand is not reloaded. Prefer to reload conflicting operand
15123 if earlyclobber and matching operands are the same.
15124
7cbda518
RB
151252013-02-12 Richard Biener <rguenther@suse.de>
15126
15127 PR lto/56297
15128 * lto-streamer-out.c (write_symbol): Do not output symbols
15129 for hard register variables.
15130
e68a4ef6
GJL
151312013-02-12 Georg-Johann Lay <avr@gjlay.de>
15132
15133 PR target/54222
15134 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15135 (umulsidi3_insn, mulsidi3_insn): New insns.
15136
a72d8780 151372013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL
15138
15139 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15140 (struct tune_params): Add vec_costs field.
15141 * config/arm/arm.c (arm_builtin_vectorization_cost)
15142 (arm_add_stmt_cost): New functions.
15143 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15144 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15145 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15146 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15147 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15148 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15149 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15150
43320568
RB
151512013-02-12 Richard Biener <rguenther@suse.de>
15152
15153 PR lto/56295
15154 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15155 decls again if possible.
15156
a011aa39
RB
151572013-02-12 Richard Biener <rguenther@suse.de>
15158
15159 PR middle-end/56288
15160 * tree-ssa.c (verify_ssa_name): Fix check, move
15161 SSA_NAME_IN_FREE_LIST check up.
15162
6da26889
JJ
151632013-02-12 Jakub Jelinek <jakub@redhat.com>
15164 Steven Bosscher <steven@gcc.gnu.org>
15165
15166 PR rtl-optimization/56151
15167 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15168 equal to op0 or op1, and last_insn pattern is CODE operation
15169 with MEM dest and one of the operands matches that MEM.
15170
f80e0faf
ST
151712013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15172
15173 * doc/extend.texi: Document Function Multiversioning and "default"
15174 parameter string to target attribute.
15175 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15176 target attribute parameter is "default".
15177 (ix86_compare_version_priority): Remove checks for target attribute.
15178 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15179 Remove check for target attribute equal to NULL. Add assert.
15180 (ix86_generate_version_dispatcher_body): Change error to sorry.
15181
6c59ffd1
IS
151822013-02-11 Iain Sandoe <iain@codesourcery.com>
15183 Jack Howarth <howarth@bromo.med.uc.edu>
15184 Patrick Marlier <patrick.marlier@gmail.com>
15185
15186 PR libitm/55693
15187 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15188 define ENDFILE_SPEC as TM_DESTRUCTOR.
15189 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15190
a72d8780 151912013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP
15192 Jack Howarth <howarth@bromo.med.uc.edu>
15193 Jakub Jelinek <jakub@redhat.com>
15194
15195 PR sanitizer/55617
15196 * config/darwin.c (cdtor_record): Rename ctor_record.
15197 (sort_cdtor_records): Rename sort_ctor_records.
15198 (finalize_dtors): New routine to sort destructors by
15199 priority before use in assemble_integer.
15200 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15201
b63fe007
UB
152022013-02-11 Uros Bizjak <ubizjak@gmail.com>
15203
15204 PR rtl-optimization/56275
15205 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15206 offset is non-negative and less than cmode size before
15207 calling simplify_subreg.
15208
8e89b5b5
RB
152092013-02-11 Richard Biener <rguenther@suse.de>
15210
15211 PR tree-optimization/56264
15212 * cfgloop.h (fix_loop_structure): Adjust prototype.
15213 * loop-init.c (fix_loop_structure): Return the number of
15214 newly discovered loops.
15215 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15216 are discovered, do a full loop-closed SSA rewrite.
15217
b4a4b56d
RB
152182013-02-11 Richard Biener <rguenther@suse.de>
15219
15220 PR tree-optimization/56273
15221 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15222 first VRP run.
15223 (check_array_ref): Fix missing newline in dumps.
15224 (search_for_addr_array): Likewise.
15225
0c885229
DE
152262013-02-09 David Edelsohn <dje.gcc@gmail.com>
15227
15228 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15229
59ac9a55
JJ
152302013-02-09 Jakub Jelinek <jakub@redhat.com>
15231
15232 PR target/56256
15233 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15234
25bb0bb5
VM
152352013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15236
15237 PR rtl-optimization/56246
0c885229 15238 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM
15239 reload pseudo.
15240 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15241 constraints are satisfied.
15242
a698cc03
JL
152432013-02-08 Jeff Law <law@redhat.com>
15244
15245 PR debug/53948
15246 * emit-rtl.c (reg_is_parm_p): New function.
15247 * regs.h (reg_is_parm_p): New prototype.
15248 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15249 callee-clobbered registers.
15250
e1122ddd
MM
152512013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15252
15253 PR target/56043
15254 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15255 If there is no implicit builtin declaration, just return NULL.
15256
19c5f6e6
UB
152572013-02-08 Uros Bizjak <ubizjak@gmail.com>
15258
15259 * config/i386/sse.md (FMAMODEM): New mode iterator.
15260 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15261 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15262
2480f2ca 152632013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 15264
2480f2ca
UB
15265 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15266 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15267 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15268
152692013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15270
15271 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15272 (microblaze*-*-elf): Likewise.
15273 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15274 LINK_SPEC.
15275 * config/microblaze/microblaze-c.c: Add builtin defines for
15276 _LITTLE_ENDIAN and _BIG_ENDIAN.
15277 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15278 add to TARGET_DEFAULT flags.
76ef61fb 15279 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB
15280 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15281 * config/microblaze/microblaze.md: Update extendsidi2 and
15282 movdi_internal instructions to use low-order / high-order reg
15283 print_operands.
15284 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15285 options and inversemask / mask of LITTLE_ENDIAN.
15286 * config/microblaze/t-microblaze: Expand multilib options to
15287 include mlittle-endian (le) and update exceptions patterns.
15288
600a5961
JJ
152892013-02-08 Jakub Jelinek <jakub@redhat.com>
15290
5df81313
JJ
15291 PR rtl-optimization/56195
15292 * lra-constraints.c (get_reload_reg): Don't reuse regs
15293 if they have smaller mode than requested, if they have
15294 wider mode than requested, try to return a SUBREG.
15295
600a5961
JJ
15296 PR tree-optimization/56250
15297 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15298 if type is unsigned and code isn't MULT_EXPR.
15299
ff544649
GJL
153002013-02-08 Georg-Johann Lay <avr@gjlay.de>
15301
15302 PR tree-optimization/56064
15303 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15304 bits according to mode.
15305 * fixed-value.h (fixed_from_double_int)
15306 (const_fixed_from_double_int): Adjust comments.
15307
e45cde98
RB
153082013-02-08 Richard Biener <rguenther@suse.de>
15309
15310 PR lto/56231
15311 * lto-streamer.h (struct data_in): Remove current_file, current_line
15312 and current_col members.
15313 * lto-streamer-out.c (lto_output_location): Stream changed bits
15314 en-block for efficiency.
15315 * lto-streamer-in.c (clear_line_info): Remove.
15316 (lto_input_location): Cache current file, line and column
15317 globally via local statics. Read changed bits en-block.
15318 (input_function): Do not call clear_line_info.
15319 (lto_read_body): Likewise.
15320 (lto_input_toplevel_asms): Likewise.
15321
c1ca73d8
MM
153222013-02-08 Michael Matz <matz@suse.de>
15323
15324 PR tree-optimization/52448
15325 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15326 (nt_call_phase): New static.
15327 (add_or_mark_expr): Only mark accesses with newer phase than any
15328 call seen.
15329 (nonfreeing_call_p): New.
15330 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15331 (nt_fini_block): Keep blocks marked as visited.
15332 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15333
57548aa2
RB
153342013-02-08 Richard Biener <rguenther@suse.de>
15335
15336 * ira.c (ira): Free broken dominator information.
15337
8e10366f
UB
153382013-02-08 Uros Bizjak <ubizjak@gmail.com>
15339
15340 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15341
f64fb0fa
MP
153422013-02-08 Marek Polacek <polacek@redhat.com>
15343
8e10366f 15344 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 15345
0375167b
RB
153462013-02-08 Richard Biener <rguenther@suse.de>
15347
15348 PR middle-end/56181
15349 * cfgloop.h (flow_loops_find): Adjust.
15350 (bb_loop_header_p): Declare.
15351 * cfgloop.c (bb_loop_header_p): New function split out from ...
15352 (flow_loops_find): ... here. Adjust function signature,
15353 support incremental loop structure update.
15354 (verify_loop_structure): Cleanup. Verify a loop is a loop.
15355 * cfgloopmanip.c (fix_loop_structure): Move ...
15356 * loop-init.c (fix_loop_structure): ... here.
15357 (apply_loop_flags): Split out from ...
15358 (loop_optimizer_init): ... here.
15359 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
15360 in incremental mode, only remove dead loops here.
15361
85d768f3
GJL
153622013-02-08 Georg-Johann Lay <avr@gjlay.de>
15363
15364 PR target/54222
15365 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
15366 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
15367 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
15368 (*round<mode>3.libgcc): New insns for fixed-modes.
15369 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
15370 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
15371 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
15372 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
15373 implementations. Define to __builtin_avr_absFX,
15374 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
15375 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
15376 __builtin_avr_countlsFX, respectively.
15377 * config/avr/avr-c.c (target.h): Include it.
15378 (enum avr_builtin_id): New enum.
15379 (avr_resolve_overloaded_builtin): New static function.
15380 (avr_register_target_pragmas): Use it to set
15381 targetm.resolve_overloaded_builtin.
15382 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
15383 tree nodes used by DEF_BUILTIN.
15384 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
15385 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
15386 <AVR_BUILTIN_xxBITS>: Same.
15387
661bc682
RB
153882013-02-08 Richard Biener <rguenther@suse.de>
15389
15390 * cfgloop.c (verify_loop_structure): Properly handle
15391 a loop exiting to another loop header.
15392 * ira-int.h (ira_loops): Remove.
15393 * ira.c (ira_loops): Remove.
15394 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
15395 (do_reload): Use loop_optimizer_finalize.
15396 * ira-build.c (create_loop_tree_nodes): Use get_loops and
15397 number_of_loops to access the loop tree.
15398 (more_one_region_p): Likewise.
15399 (finish_loop_tree_nodes): Likewise.
15400 (rebuild_regno_allocno_maps): Likewise.
15401 (mark_loops_for_removal): Likewise.
15402 (mark_all_loops_for_removal): Likewise.
15403 (remove_unnecessary_regions): Likewise.
15404 (ira_build): Likewise.
15405 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15406
0d5049b2
RB
154072013-02-08 Richard Biener <rguenther@suse.de>
15408
15409 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
15410 * ipa-pure-const.c (analyze_function): Avoid calling
15411 mark_irreducible_loops twice.
8e10366f 15412 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 15413
499f32e8
DM
154142013-02-07 David S. Miller <davem@davemloft.net>
15415
15416 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
15417 on 'reg'.
15418 * var-tracking.c (vt_add_function_parameter): Test the presence of
15419 HAVE_window_save properly and do not remap argument registers when
15420 we have a leaf function.
15421
6edc3e32
UB
154222013-02-07 Uros Bizjak <ubizjak@gmail.com>
15423
15424 PR bootstrap/56227
15425 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
15426 instead of "ll".
15427 * config/i386/i386.c (ix86_print_operand): Ditto.
15428
5306401f
VM
154292013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15430
6edc3e32 15431 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 15432
027ece11
VM
154332013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15434
15435 PR rtl-optimization/56225
15436 * lra-constraints.c (process_alt_operands): Check that reload hard
15437 reg can hold value for strict_low_part.
15438
f980dfdb
JJ
154392013-02-07 Jakub Jelinek <jakub@redhat.com>
15440
15441 PR debug/56154
15442 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
15443 dwarf2out_end_function.
15444 (in_first_function_p, maybe_at_text_label_p,
15445 first_loclabel_num_not_at_text_label): New variables.
15446 (dwarf2out_var_location): In the first function find out
6edc3e32 15447 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ
15448 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
15449 functions.
15450
73dd3123
EB
154512013-02-07 Eric Botcazou <ebotcazou@adacore.com>
15452
15453 PR rtl-optimization/56178
15454 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
15455 SUBREG of a register. Tidy up related block of code.
15456 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
15457 note if the source is a register or a SUBREG of a register.
15458
e3936f47
JJ
154592013-02-07 Jakub Jelinek <jakub@redhat.com>
15460
15461 PR target/56228
15462 * config/rs6000/rs6000.md (ptrm): New mode attr.
15463 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
15464 call_value_indirect_aix<pttrsize>,
15465 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
15466 m in constraints.
15467
d96d674b
MH
154682013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
15469
15470 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
15471 if -bnortl. Convert to strcmp and strncmp.
15472
921f2dee
AM
154732013-02-07 Alan Modra <amodra@gmail.com>
15474
15475 PR target/54009
15476 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
15477 addresses won't wrap when offsetting.
15478 (rs6000_secondary_reload): Provide secondary reloads needed for
15479 wrapping LO_SUM addresses.
15480
d09c7dba
TS
154812013-02-06 Thomas Schwinge <thomas@codesourcery.com>
15482
15483 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
15484 MACH, just __MACH__.
15485
a44bbd48
RB
154862013-02-06 Richard Biener <rguenther@suse.de>
15487
15488 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
15489 instead of calling fix_loop_structure.
15490
6d840d99
JJ
154912013-02-06 Jakub Jelinek <jakub@redhat.com>
15492
15493 PR middle-end/56217
15494 * omp-low.c (use_pointer_for_field): Return false if
15495 lower_send_shared_vars doesn't generate any copy-out code.
15496
0f33baa9
TV
154972013-02-06 Tom de Vries <tom@codesourcery.com>
15498
15499 PR rtl-optimization/56131
15500 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
15501 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
15502 of the label is NULL. Add comment.
15503
11f1e3ab
JJ
155042013-02-05 Jakub Jelinek <jakub@redhat.com>
15505
31502f9f
JJ
15506 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
15507
a74db9bd
JJ
15508 PR sanitizer/55374
15509 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
15510 (STATIC_LIBTSAN_LIBS): Likewise.
15511 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
15512 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
15513 is defined, don't add anything else beyond that.
15514 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
15515 (LINK_COMMAND_SPEC): Use them.
15516
11f1e3ab
JJ
15517 PR tree-optimization/56205
15518 * tree-stdarg.c (check_all_va_list_escapes): Return true if
15519 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
15520 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
15521
14ac6aa2
RB
155222013-02-05 Richard Biener <rguenther@suse.de>
15523
15524 PR tree-optimization/53342
15525 PR tree-optimization/53185
15526 * tree-vectorizer.h (vect_check_strided_load): Remove.
15527 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
15528 not disallow peeling for vectorized strided loads.
15529 (vect_check_strided_load): Make static and simplify.
15530 (vect_analyze_data_refs): Adjust.
15531 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
15532 correctly when vectorizing strided loads.
15533
6f22445a
RB
155342013-02-05 Richard Biener <rguenther@suse.de>
15535
15536 * doc/install.texi: Refer to ISL, not PPL.
15537
39f9719e
JH
155382013-02-05 Jan Hubicka <jh@suse.cz>
15539
ec4224ac
JH
15540 PR tree-optimization/55789
15541 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
15542
155432013-02-05 Jan Hubicka <jh@suse.cz>
15544
15545 PR tree-optimization/55789
39f9719e
JH
15546 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
15547 the dead call anyway.
15548
956d3b33
EB
155492013-02-05 Eric Botcazou <ebotcazou@adacore.com>
15550
15551 PR sanitizer/55374
15552 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
15553
155542013-02-04 Alexander Potapenko <glider@google.com>
15555 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP
15556 Jakub Jelinek <jakub@redhat.com>
15557
15558 PR sanitizer/55617
15559 * config/darwin.c (sort_ctor_records): Stabilized qsort
15560 on constructor priority by using original position.
15561 (finalize_ctors): New routine to sort constructors by
15562 priority before use in assemble_integer.
15563 (machopic_asm_out_constructor): Use finalize_ctors if needed.
15564
7ac3af38
JJ
155652013-02-04 Jakub Jelinek <jakub@redhat.com>
15566
15567 PR libstdc++/54314
15568 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
15569 about visibility on artificial decls.
15570 * config/sol2.c (solaris_assemble_visibility): Likewise.
15571
152689dc
KT
155722013-02-04 Kai Tietz <ktietz@redhat.com>
15573
15574 PR target/56186
15575 * config/i386/i386.c (function_value_ms_64): Add additional valtype
15576 argument and improve checking of return-argument types for 16-byte
15577 modes.
15578 (ix86_function_value_1): Add additional valtype argument on call
15579 of function_value_64.
15580 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
15581 handling infunction_value_64 function.
15582
2b5987b5
MGD
155832013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
15584
7ac3af38 15585 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 15586
14597080
RB
155872013-02-04 Richard Biener <rguenther@suse.de>
15588
15589 PR tree-optimization/56188
15590 * tree-ssa-structalias.c (label_visit): Consider case with
15591 initially non-empty points-to set.
15592 (perform_var_substitution): Dump node mapping and clean up.
15593
ed73881e
RG
155942013-02-04 Richard Guenther <rguenther@suse.de>
15595
15596 PR lto/56168
15597 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
15598 node prevail as last resort.
15599 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 15600 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 15601
139a0707
RB
156022013-02-04 Richard Biener <rguenther@suse.de>
15603
15604 PR tree-optimization/56113
15605 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
15606 Merge into ...
15607 (equiv_class_lookup_or_add): ... this.
15608 (label_visit): Adjust and fix error in previous patch.
15609 (perform_var_substitution): Adjust.
15610
50fe8924
OE
156112013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
15612
15613 * config/sh/divtab.c: Fix formatting and comments throughout the file.
15614 * config/sh/sh4-300.md: Likewise.
15615 * config/sh/sh4a.md: Likewise.
15616 * config/sh/constraints.md: Likewise.
15617 * config/sh/sh.md: Likewise.
15618 * config/sh/netbsd-elf.h: Likewise.
15619 * config/sh/predicates.md: Likewise.
15620 * config/sh/sh-protos.h: Likewise.
15621 * config/sh/ushmedia.h: Likewise.
15622 * config/sh/linux.h: Likewise.
15623 * config/sh/sh.c: Likewise.
15624 * config/sh/superh.h: Likewise.
15625 * config/sh/elf.h: Likewise.
15626 * config/sh/sh4.md: Likewise.
15627 * config/sh/sh.h: Likewise.
15628
1a04ac2b
JDA
156292013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15630
15631 * config/pa/constraints.md: Adjust unused letters. Change "T"
15632 constraint to match_test floating_point_store_memory_operand().
15633 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
15634 (base14_operand): New.
15635 (floating_point_store_memory_operand): New.
15636 (integer_store_memory_operand): Revise to use base14_operand and
15637 reg_plus_base_memory_operand.
15638 (move_dest_operand): Allow symbolic_memory_operands.
15639 (symbolic_memory_operand): Check for LO_SOM.
15640 (symbolic_operand): Change default case to break.
15641 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
15642 CONST_DOUBLE values to be reloaded by putting them into memory when
15643 the destination is a floating point register.
15644 (movdf): Remove code to handle CONST_DOUBLE.
15645 (movsf): Likewise.
15646 (reload_indf_r1): New.
15647 (reload_insf_r1): New.
15648 Consistently use "Q" and "T" constraints with integer and floating
15649 point move instructions, respectively.
15650 (movdi): Remove FAIL.
15651 Change predicate for source operand unamed DImode move from
15652 general_operand to move_src_operand.
15653 (umulsidi3): Change predicate for destination operand to
15654 register_operand.
15655 Likewise for similar unamed patterns.
15656 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
15657 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
15658 (hppa_legitimize_address): Simplify mask calculation.
15659 (pa_emit_move_sequence): Revised handling of secondary reloads from
15660 REG+D addresses for floating point loads and stores. Directly handle
15661 loading CONST0_RTX (mode) to a floating point register.
15662 (pa_secondary_reload): Handle reloading DF and SFmode constant values
15663 to floating point registers. Don't restrict secondary reloads to
15664 floating point registers to integer modes. Revise some comments and
15665 cleanup some code.
15666 (TARGET_LEGITIMATE_ADDRESS_P): Define.
15667 (pa_legitimate_address_p): New.
15668 (pa_legitimize_reload_address): New.
15669 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
15670 (STRICT_REG_OK_FOR_BASE_P): New.
15671 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
15672 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
15673
3bf65591
DE
156742013-02-03 David Edelsohn <dje.gcc@gmail.com>
15675 Andrew Dixie <andrewd@gentrack.com>
15676
15677 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
15678 flag set.
15679
c0a8a3e6
RS
156802013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
15681
15682 * expmed.c (extract_bit_field_1): Pass the full width of the
15683 structure to get_best_reg_extraction_insn.
15684
99113dff
DE
156852013-02-01 David Edelsohn <dje.gcc@gmail.com>
15686
15687 PR target/54601
15688 * configure.ac (use_cxa_atexit): Add AIX.
15689 * configure: Regenerate.
15690
15691 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
15692
ddd84654
JJ
156932013-02-01 Jakub Jelinek <jakub@redhat.com>
15694
15695 PR debug/54793
15696 * final.c (need_profile_function): New variable.
15697 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
15698 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
15699 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
15700 notes, targetm.asm_out.function_prologue doesn't emit anything,
15701 HAVE_prologue and profiler should be emitted before prologue,
15702 set need_profile_function instead of emitting it.
15703 (final_scan_insn): If need_profile_function, emit
15704 profile_function on the first NOTE_INSN_BASIC_BLOCK or
15705 NOTE_INSN_FUNCTION_BEG note.
15706
157072013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH
15708
15709 * config/rs6000/rs6000.md (smulditi3): New.
15710 (umulditi3): New.
15711
ff2a9d88
RH
15712 * config/alpha/alpha.md (umulditi3): New.
15713
14d52b90
DE
157142013-02-01 David Edelsohn <dje.gcc@gmail.com>
15715
15716 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
15717 (ASM_OUTPUT_ALIGNED_LOCAL): New.
15718
8c7ca45c
RB
157192013-02-01 Richard Biener <rguenther@suse.de>
15720
15721 PR tree-optimization/56113
15722 * tree-ssa-structalias.c (label_visit): Reduce work for
15723 single-predecessor nodes.
15724
9f419393
EB
157252013-02-01 Eric Botcazou <ebotcazou@adacore.com>
15726
15727 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
15728 range isn't testing for zero.
15729
c5f4be84
SB
157302013-01-31 Steven Bosscher <steven@gcc.gnu.org>
15731
15732 PR middle-end/56113
15733 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
15734
dbdbd982
NC
157352013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
15736 Nick Clifton <nickc@redhat.com>
15737
15738 * config/v850/constraints.md (Q): Define as a memory constraint.
15739 * config/v850/predicates.md (label_ref_operand): New predicate.
15740 (e3v5_shift_operand): New predicate.
15741 (ior_operator): New predicate.
15742 * config/v850/t-v850: Add e3v5 multilib.
15743 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
15744 (v850_gen_movdi): Prototype.
15745 * config/v850/v850.c: Add support for e3v5 architecture.
15746 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
15747 TARGET_V850E_UP.
15748 (construct_save_jarl): Add e3v5 long JARL support.
15749 (v850_adjust_insn_length): New function. Adjust length of call
15750 insns when using e3v5 instructions.
15751 (v850_gen_movdi): New function: Generate instructions to move a
15752 DImode value.
15753 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
15754 (CPP_SPEC): Define __v850e3v5__ as appropriate.
15755 (TARGET_USE_FPU): Enable for e3v5.
15756 (CONST_OK_FOR_W): New macro.
15757 (ADJUST_INSN_LENGTH): Define.
15758 * config/v850/v850.md (UNSPEC_LOOP): Define.
15759 (attr cpu): Add v850e3v5.
15760 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
15761 (movdi): New pattern.
15762 (movdi_internal): New pattern.
15763 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
15764 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
15765 (cstoresf4): Likewise.
15766 (cstoredf4): Likewise.
15767 (insv): New pattern.
15768 (rotlso3_a): New pattern.
15769 (rotlsi3_b): New pattern
15770 (rotlsi3_v850e3v5): New pattern.
15771 (doloop_begin): New pattern.
15772 (fix_loop_counter): New pattern.
15773 (doloop_end): New pattern.
15774 (branch_normal): Add e3v5 long branch support.
15775 (branch_invert): Likewise.
15776 (branch_z_normal): Likewise.
15777 (branch_z_invert): Likewise.
15778 (branch_nz_normal): Likewise.
15779 (branch_nz_invert): Likewise.
15780 (call_internal_short): Add e3v5 register-indirect JARL support.
15781 (call_internal_long): Likewise.
15782 (call_value_internal_short): Likewise.
15783 (call_value_internal_long): Likewise.
15784 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
15785 (mloop): New option.
15786 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 15787 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 15788
73861a41
PK
157892013-01-31 Paul Koning <ni1d@arrl.net>
15790
15791 PR debug/55059
15792 PR debug/54508
15793 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
15794 children if parent is a class.
15795 (prune_unused_types_prune): Don't add DW_AT_declaration.
15796
e44978dc
RB
157972013-01-31 Richard Biener <rguenther@suse.de>
15798
15799 PR tree-optimization/56157
15800 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
15801 match up operand with SLP child.
15802
90b10dec
JM
158032013-01-31 Jason Merrill <jason@redhat.com>
15804
04d2dadd 15805 PR debug/54410
90b10dec
JM
15806 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
15807 parameters the first time.
15808 (gen_scheduled_generic_parms_dies): Check completeness here.
15809
6e616110
RB
158102013-01-31 Richard Biener <rguenther@suse.de>
15811
15812 PR middle-end/53073
15813 * common.opt (faggressive-loop-optimizations): New flag,
15814 enabled by default.
15815 * doc/invoke.texi (faggressive-loop-optimizations): Document.
15816 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
15817 infer_loop_bounds_from_undefined by it.
15818
636f59cf
RB
158192013-01-31 Richard Biener <rguenther@suse.de>
15820
15821 PR tree-optimization/56150
15822 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
15823 visit virtual operands.
15824 (find_uses_to_rename_bb): Likewise.
15825
158262013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB
15827
15828 PR tree-optimization/56150
15829 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
15830 mixed store non-store stmts.
15831
32887460
JJ
158322013-01-30 Jakub Jelinek <jakub@redhat.com>
15833
e60e09a0
JJ
15834 PR sanitizer/55374
15835 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
15836 LIBASAN_EARLY_SPEC is defined.
15837 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
15838 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
15839 before %o.
15840 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
15841
32887460
JJ
15842 PR c++/55742
15843 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
15844 invalid args instead of ICEing on it.
15845 (ix86_valid_target_attribute_tree): Return error_mark_node if
15846 ix86_valid_target_attribute_inner_p failed.
15847 (ix86_valid_target_attribute_p): Return false only if
15848 ix86_valid_target_attribute_tree returned error_mark_node. Allow
15849 target("default") attribute.
15850 (sorted_attr_string): Change argument from const char * to tree,
15851 merge in all target attribute arguments rather than just one.
15852 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
15853 instead of free. Avoid using strcat.
15854 (ix86_mangle_function_version_assembler_name): Mangle
15855 target("default") as if no target attribute is present. Adjust
15856 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
15857 instead of xmalloc and XDELETEVEC instead of free.
15858 (ix86_function_versions): Don't return true if one of the decls
15859 doesn't have target attribute. If they don't and one of the decls
15860 is DECL_FUNCTION_VERSIONED, report an error. Adjust
15861 sorted_attr_string caller. Use XDELETEVEC instead of free.
15862 (ix86_supports_function_versions): Remove.
15863 (make_name): Fix up formatting.
15864 (make_dispatcher_decl): Remove resolver_name and its initialization.
15865 Avoid leaking memory.
15866 (is_function_default_version): Return true if there is
15867 target("default") attribute rather than no target attribute at all.
15868 (make_resolver_func): Avoid leaking memory.
15869 (ix86_generate_version_dispatcher_body): Likewise.
15870 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
15871 * target.def (supports_function_versions): Remove.
15872 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
15873 * doc/tm.texi: Regenerated.
15874
73cca0cc
VM
158752013-01-30 Vladimir Makarov <vmakarov@redhat.com>
15876
15877 PR rtl-optimization/56144
15878 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
15879 for values with side effects.
15880
11452e7b
RB
158812013-01-30 Richard Biener <rguenther@suse.de>
15882
15883 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
15884 (sparseset_pop): Likewise.
15885 * cfganal.c (compute_idf): Likewise. Increase work-stack size
15886 to be able to use quick_push in the worker loop.
15887
01cb1ef5
MP
158882013-01-30 Marek Polacek <polacek@redhat.com>
15889
15890 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
15891
e892936e
RB
158922013-01-30 Richard Biener <rguenther@suse.de>
15893
15894 PR lto/56147
6edc3e32 15895 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 15896
cc06c01d
GJL
158972013-01-30 Georg-Johann Lay <avr@gjlay.de>
15898
15899 PR tree-optimization/56064
15900 * fixed-value.c (fixed_from_double_int): New function.
15901 * fixed-value.h (fixed_from_double_int): New prototype.
15902 (const_fixed_from_double_int): New static inline function.
15903 * fold-const.c (native_interpret_fixed): New static function.
15904 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
15905 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
15906 (native_encode_fixed): New static function.
15907 (native_encode_expr) <FIXED_CST>: Use it.
15908 (native_interpret_int): Move double_int worker code to...
15909 * double-int.c (double_int::from_buffer): ...this new static method.
15910 * double-int.h (double_int::from_buffer): Prototype it.
15911
d394a308
RB
159122013-01-30 Richard Biener <rguenther@suse.de>
15913
15914 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
15915 New pointer-map and obstack.
15916 (init_alias_vars): Allocate pointer-map and obstack.
15917 (delete_points_to_sets): Free them.
15918 (find_what_var_points_to): Cache result.
15919 (find_what_p_points_to): Adjust for changed interface of
15920 find_what_var_points_to.
15921 (compute_points_to_sets): Likewise.
15922 (ipa_pta_execute): Likewise.
15923
20804d96
RO
159242013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15925
15926 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
15927 * configure: Regenerate.
15928 * config.in: Regenerate.
15929 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
15930 #nobits/#progbits if supported.
15931
a7ad88a2
OE
159322013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
15933
15934 PR target/56121
15935 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
15936 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
15937 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
15938
7af79f92
GY
159392013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15940
91bfca59
OE
15941 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
15942 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 15943
753bcf7b
GY
159442013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15945
91bfca59
OE
15946 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
15947 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 15948
8cbc2ea8
GY
159492013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15950
91bfca59
OE
15951 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
15952 declaration.
15953 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
15954 * config/arm/cortex-a7.md: New bypasses using
15955 arm_mac_accumulator_is_result.
8cbc2ea8 15956
697a3325
GY
159572013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15958
91bfca59 15959 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 15960 (cortex_a7_neon_mla): Likewise.
91bfca59
OE
15961 (cortex_a7_fpfmad): New reservation.
15962 (cortex_a7_fpmacs): Use ffmas and update required units.
15963 (cortex_a7_fpmuld): Update required units and latency.
15964 (cortex_a7_fpmacd): Likewise.
15965 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
15966 (cortex_a7_neon). Likewise.
15967 (bypass) Update participating units.
697a3325 15968
29637783
GY
159692013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15970
91bfca59
OE
15971 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
15972 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
15973 from fmac to ffma.
15974 * config/arm/vfp11.md (vfp_farith): Use ffmas.
15975 (vfp_fmul): Use ffmad.
15976 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
15977 (cortex_r4_fmacd): Use ffmad.
15978 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
15979 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
15980 (cortex_a9_fmacd): Use ffmad.
15981 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
15982 (cortex_a8_vfp_macd): Use ffmad.
15983 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
15984 (cortex_a5_fpmacd): Use ffmad.
15985 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
15986 (cortex_a15_vfp_macd): Use ffmad.
15987 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 15988
2ee8a2d5
JM
159892013-01-29 Jason Merrill <jason@redhat.com>
15990
15991 PR libstdc++/54314
15992 * varasm.c (default_assemble_visibility): Don't warn about
15993 visibility on artificial decls.
15994
78d087bc
RB
159952013-01-29 Richard Biener <rguenther@suse.de>
15996
15997 PR tree-optimization/56113
15998 * tree-ssa-structalias.c (equiv_class_lookup): Also return
15999 the bitmap leader.
16000 (label_visit): Free duplicate bitmaps and record the leader instead.
16001 (perform_var_substitution): Adjust.
16002
83ba4d6f
RB
160032013-01-29 Richard Biener <rguenther@suse.de>
16004
16005 PR tree-optimization/55270
16006 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16007 the CFG, schedule loops for fixup.
16008
66dfe13f
NC
160092013-01-29 Nick Clifton <nickc@redhat.com>
16010
16011 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16012 SP_REG.
16013
5a579c3b
LE
160142013-01-28 Leif Ekblad <leif@rdos.net>
16015
16016 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16017 * config/i386/i386.h (TARGET_RDOS): New macro.
16018 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16019 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16020 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16021 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16022 DEFAULT_LARGE_SECTION_THRESHOLD.
16023 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16024 * config/i386/rdos.h: New file.
16025 * config/i386/rdos64.h: New file.
16026
51e44392
BS
160272013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16028
16029 PR other/54814
16030 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16031 TEST_HARD_REG_BIT.
16032
db1fb332
JJ
160332013-01-28 Jakub Jelinek <jakub@redhat.com>
16034
16035 PR rtl-optimization/56117
16036 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16037 call cselib_lookup_from_insn on the MEM before calling
16038 add_insn_mem_dependence.
16039
16917761
RB
160402013-01-28 Richard Biener <rguenther@suse.de>
16041
16042 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16043 to a stmt that didn't have one.
16044 (copy_phis_for_bb): Likewise for PHI arguments.
16045 (copy_debug_stmt): Likewise for debug stmts.
16046
b9fc0497
RB
160472013-01-28 Richard Biener <rguenther@suse.de>
16048
16049 PR tree-optimization/56034
6edc3e32 16050 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB
16051 (partition_builtin_p): Adjust.
16052 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16053 it is the last partition.
16054 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16055 up the vertex for the definition.
16056 (classify_partition): Classify whether a partition is a
16057 PKIND_REDUCTION, thus has uses outside of the loop.
16058 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16059 Merge all PKIND_REDUCTION partitions into the last partition.
16060 (tree_loop_distribution): Seed partitions from reductions as well.
16061
aa710d25
JJ
160622013-01-28 Jakub Jelinek <jakub@redhat.com>
16063
0bfbca58
JJ
16064 PR tree-optimization/56125
16065 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16066 pow(x,c) into sqrt(x) * powi(x, n/2) or
16067 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16068 optimizing for size.
16069 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16070 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16071 integer.
16072
aa710d25
JJ
16073 PR tree-optimization/56094
16074 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16075 to UNKNOWN_LOCATION while gimplifying expr.
16076
77dc5297
UB
160772013-01-27 Uros Bizjak <ubizjak@gmail.com>
16078
16079 PR target/56114
16080 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16081 operand 0 in movabs insn template for -masm=intel asm alternative.
16082 (*movabs<mode>_2): Ditto for operand 1.
16083
0bfbca58 160842013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH
16085
16086 PR target/54663
16087 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16088 of microblaze-c.o
16089
0bfbca58 160902013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI
16091
16092 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16093 tm_file.
16094
0bfbca58 160952013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N
16096
16097 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16098 Undef to avoid warning.
16099
478f60f9
MH
161002013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16101
16102 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16103 * configure: Regenerate.
16104
d7fa6ee2
JJ
161052013-01-25 Jakub Jelinek <jakub@redhat.com>
16106
16107 PR tree-optimization/56098
16108 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16109 for stmts with volatile ops.
16110 (cond_store_replacement): Don't optimize if assign has volatile ops.
16111 (cond_if_else_store_replacement_1): Don't optimize if either
16112 then_assign or else_assign have volatile ops.
16113 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16114 volatile ops.
16115
f8fe87bd
GJL
161162013-01-25 Georg-Johann Lay <avr@gjlay.de>
16117
16118 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16119
28398d0d
GJL
161202013-01-25 Georg-Johann Lay <avr@gjlay.de>
16121
16122 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16123 missing ':' in asm example.
16124
b7d7d917
TB
161252013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16126
16127 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16128 entries into lane and laneq entries.
77dc5297
UB
16129 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16130 Remove AdvSIMD scalar modes.
b7d7d917
TB
16131 (aarch64_sq<r>dmulh_laneq<mode>): New.
16132 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16133 modes.
16134 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16135 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16136 * config/aarch64/iterators.md (VCOND): New.
16137 (VCONQ): New.
16138
556f9906
GJL
161392013-01-25 Georg-Johann Lay <avr@gjlay.de>
16140
16141 PR target/54222
16142 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16143 Add NULL LIBNAME argument to existing definitions.
16144 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16145 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16146 * config/avr/avr.c (DEF_BUILTIN): Same.
16147 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16148 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16149 implementation is available (DECL_ASSEMBLER_NAME is set).
16150 (avr_fold_absfx): New static function.
16151 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16152 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16153 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16154 AVR_BUILTIN_ABSLLK.
16155 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16156 (abshk, absk, abslk, absllk): Provide as static inline functions.
16157
1f546bbb
MP
161582013-01-25 Marek Polacek <polacek@redhat.com>
16159
16160 PR tree-optimization/56035
16161 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16162
7e184bd7
UB
161632012-01-24 Uros Bizjak <ubizjak@gmail.com>
16164
16165 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16166 (*movtf_internal_rex64): Add (!o,C) alternative
16167 (*movxf_internal_rex64): Ditto.
16168 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16169
f5ba49ac
SM
161702013-01-24 Shenghou Ma <minux.ma@gmail.com>
16171
16172 * doc/invoke.texi: fix typo.
16173 * doc/objc.texi: fix typo.
16174
a6343728
RS
161752013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16176
16177 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16178 for the first two alternatives.
16179
cd030c07
DN
161802013-01-24 Diego Novillo <dnovillo@google.com>
16181
77dc5297 16182 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN
16183 (ggc-zone.o): Remove.
16184 * configure.ac: Remove option --with-gc.
16185 * configure: Re-generate.
16186 * doc/install.texi: Remove documentation for --with-gc.
16187 * gengtype.c (write_enum_defn): Remove. Update all users.
16188 (write_Types_process_field): Remove generation of gt_e_* argument.
16189 (output_type_enum): Remove. Update all users.
16190 (write_enum_defn): Remove. Update all users.
16191 (enum alloc_zone): Remove. Update all users.
77dc5297 16192 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN
16193 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16194 Update all callers.
16195 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 16196 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN
16197 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16198 Update all users.
aaf1e810 16199 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN
16200 (struct alloc_zone): Remove.
16201 (ggc_internal_alloc_zone_stat): Remove.
16202 (ggc_internal_cleared_alloc_zone_stat): Remove.
16203 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB
16204 (ggc_pch_count_object): Remove last argument. Update all users.
16205 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN
16206 (struct alloc_zone): Remove.
16207 * ggc-zone.c: Remove.
77dc5297 16208 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN
16209 (struct alloc_zone): Remove.
16210 (ggc_alloc_typed_stat): Remove.
16211 (ggc_alloc_typed): Remove.
16212 (ggc_splay_alloc): Remove first argument.
16213 (rtl_zone): Remove. Update all users.
16214 (tree_zone): Remove. Update all users.
16215 (tree_id_zone): Remove. Update all users.
16216 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 16217 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 16218 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 16219 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 16220
a861ffa4
GJL
162212013-01-24 Georg-Johann Lay <avr@gjlay.de>
16222
16223 * config/avr/avr.c (avr_out_fract): Make register numbers that
16224 might be outside of source operand signed.
16225
593c0ddd
UB
162262013-01-24 Uros Bizjak <ubizjak@gmail.com>
16227
16228 * config/i386/constraints.md (Yf): New constraint.
16229 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16230 of f constraint to conditionaly disable x87 register preferences.
16231 (*movdf_internal): Ditto.
16232 (*movsf_internal): Ditto.
16233
e86c0101
SB
162342013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16235
16236 PR inline-asm/55934
16237 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16238 that have operands with impossible constraints.
16239 Add a FIXME for a speed-up opportunity.
16240 * lra-constraints.c (process_alt_operands): Verify that a class
16241 selected from constraints on asms is valid for the operand mode.
16242 (curr_insn_transform): Remove incorrect comment.
16243
f6fee35f
DE
162442013-01-23 David Edelsohn <dje.gcc@gmail.com>
16245
16246 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16247 TOC operand is a valid symbol ref in the constant pool.
16248
162492013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 16250
aaf1e810 16251 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 16252
dc62d7d1
GJL
162532013-01-23 Georg-Johann Lay <avr@gjlay.de>
16254
16255 PR target/54222
16256 * config/avr/stdfix.h: New file.
16257 * t-avr (stdfix-gcc.h): New rule to build it.
16258 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16259
2660d12d
KS
162602013-01-23 Kostya Serebryany <kcc@google.com>
16261
77dc5297
UB
16262 * config/darwin.h: remove dependency on
16263 CoreFoundation (asan on Mac OS).
2660d12d 16264
a70418fc
JJ
162652013-01-23 Jakub Jelinek <jakub@redhat.com>
16266
16267 PR target/49069
16268 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16269 instead of cmpdi_operand for first comparison operand.
16270 Don't assert that comparison operands aren't both constants.
16271
47876a2a
JW
162722013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16273
16274 * doc/install.texi (Downloading the Source): Update references to
16275 downloading separate components.
16276
162772013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW
16278
16279 * doc/extend.texi (__int128): Improve grammar.
16280
47876a2a 162812013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB
16282
16283 PR target/56028
16284 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16285 alternative to (o,r).
16286 (*movdi_internal_rex64): Remove (!o,n) alternative.
16287 (DImode immediate->memory splitter): Remove.
16288 (DImode immediate->memory peephole2): Remove.
16289 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16290 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16291 alternative to (!o,*r).
16292 (*movtf_internal_sse): New pattern.
16293 (*movxf_internal_rex64): New pattern.
16294 (*movxf_internal): Disable for TARGET_64BIT.
16295 (*movdf_internal_rex64): Remove (!o,F) alternative.
16296
3a984f10
JJ
162972013-01-22 Jakub Jelinek <jakub@redhat.com>
16298
502498d5
JJ
16299 PR middle-end/56074
16300 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16301 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16302 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16303 stmt locations where LOCATION_LOCUS of the stmt location is
16304 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16305
3a984f10
JJ
16306 PR target/55686
16307 * config/i386/i386.md (UNSPEC_STOS): New.
16308 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16309 *strsetqi_1): Add UNSPEC_STOS.
16310
fa817f7f
PC
163112013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16312
16313 PR c++/56067
16314 * doc/invoke.texi: Remove left over -Wsynth example.
16315
8f498c1b
JJ
163162013-01-21 Jakub Jelinek <jakub@redhat.com>
16317
16318 PR tree-optimization/56051
16319 * fold-const.c (fold_binary_loc): Don't fold
16320 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16321 a narrowing conversion, or widening conversion from signed
16322 to unsigned.
16323
47876a2a 163242013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB
16325
16326 PR rtl-optimization/56023
16327 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16328 dependent on debug instruction.
16329
5b9db1bc
MJ
163302013-01-21 Martin Jambor <mjambor@suse.cz>
16331
16332 PR middle-end/56022
16333 * function.c (allocate_struct_function): Call
16334 invoke_set_current_function_hook earlier.
16335
e8bb7d68
JJ
163362013-01-21 Jakub Jelinek <jakub@redhat.com>
16337
16338 * reload1.c (init_reload): Only initialize reload_obstack
16339 during the first call.
16340
616a4e32
MP
163412013-01-21 Marek Polacek <polacek@redhat.com>
16342
16343 * cfgloop.c (verify_loop_structure): Fix up grammar.
16344
4401981b
YHH
163452013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
16346
16347 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
16348 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
16349
8e87740b
RR
163502013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16351
16352 PR target/56058
16353 * config/arm/marvell-pj4.md: Update copyright year.
16354 Fix up use of alu to alu_reg and simple_alu_imm.
16355
47876a2a 163562013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB
16357
16358 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
16359
89d56d79
VM
163602013-01-20 Vladimir Makarov <vmakarov@redhat.com>
16361
16362 PR target/55433
16363 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 16364 insn for secondary memory move when memory mode should be different.
89d56d79 16365
fe603553
JDA
163662013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16367
16368 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
16369 atomic_storedi_1): New patterns.
16370
01284895
VK
163712013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16372
16373 btver2 pipeline descriptions.
16374 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
16375 descriptions.
16376 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 16377 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK
16378 type attributes.
16379 * config/i386/btver2.md: New file describing btver2 pipelines.
16380
5630e3e1
JL
163812013-01-19 Andrew Pinski <apinski@cavium.com>
16382
16383 PR tree-optimization/52631
16384 * tree-ssa-sccvn (visit_use): Before looking up the original
16385 statement, try looking up the simplified expression.
16386
650ae806
AG
163872013-01-19 Anthony Green <green@moxielogic.com>
16388
16389 * config/moxie/moxie.c (moxie_expand_prologue): Set
16390 current_function_static_stack_size.
16391
e300ec2d
JJ
163922013-01-18 Jakub Jelinek <jakub@redhat.com>
16393
16394 PR tree-optimization/56029
16395 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
16396 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
16397
a3d7af04
SS
163982013-01-18 Sharad Singhai <singhai@google.com>
16399
16400 PR tree-optimization/55995
16401 * dumpfile.c (dump_loc): Print location only if available.
16402 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
16403
66aa7879
VM
164042013-01-18 Vladimir Makarov <vmakarov@redhat.com>
16405
16406 PR target/55433
16407 * lra-constraints.c (curr_insn_transform): Reuse original insn for
16408 secondary memory move.
16409 (inherit_reload_reg): Use rclass instead of cl for
16410 check_secondary_memory_needed_p.
16411
3f0fee7b
JJ
164122013-01-18 Jakub Jelinek <jakub@redhat.com>
16413
16414 PR middle-end/56015
16415 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 16416 the case where writing real complex part of target modifies op1.
3f0fee7b 16417
70c67693
JG
164182013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
16419
16420 * config/aarch64/aarch64-simd.md
16421 (aarch64_vcond_internal<mode>): Handle unordered cases.
16422 * config/aarch64/iterators.md (v_cmp_result): New.
16423
df8de9b3
YHH
164242013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
16425 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16426
16427 * config/arm/marvell-pj4.md: New file.
16428 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
16429 * config/arm/arm.md (generic_sched): Add marvell_pj4.
16430 (generic_vfp): Likewise.
16431 * config/arm/arm-cores.def: Add marvell-pj4.
16432 * config/arm/arm-tune.md: Regenerate.
16433 * config/arm/arm-tables.opt: Regenerate.
16434 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
16435 * doc/invoke.texi: Document marvell-pj4.
16436
be30c356
TB
164372013-01-18 Tejas Belagod <tejas.belagod@arm.com>
16438
16439 * config/aarch64/arm_neon.h: Map scalar types to standard types.
16440
0bfbca58 164412013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO
16442
16443 PR debug/54114
16444 PR debug/54402
16445 PR debug/49888
16446 * var-tracking.c (negative_power_of_two_p): New.
16447 (global_get_addr_cache, local_get_addr_cache): New.
16448 (get_addr_from_global_cache, get_addr_from_local_cache): New.
16449 (vt_canonicalize_addr): Rewrite using the above. Adjust the
16450 heading comment.
16451 (vt_stack_offset_p): Remove.
16452 (vt_canon_true_dep): Always canonicalize loc's address.
16453 (clobber_overlapping_mems): Make sure we have a MEM.
16454 (local_get_addr_clear_given_value): New.
16455 (val_reset): Clear local cached entries.
16456 (compute_bb_dataflow): Create and release the local cache.
16457 Disable duplicate MEMs clobbering.
16458 (emit_notes_in_bb): Clobber MEMs likewise.
16459 (vt_emit_notes): Create and release the local cache.
16460 (vt_initialize, vt_finalize): Create and release the global
16461 cache, respectively.
1f6bc337 16462 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 16463
0bfbca58 164642013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO
16465
16466 PR libmudflap/53359
16467 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
16468 not found in the symtab.
16469
0bfbca58 164702013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 16471
c350ba53 16472 PR debug/56006
3aa03517
AO
16473 PR rtl-optimization/55547
16474 PR rtl-optimization/53827
16475 PR debug/53671
16476 PR debug/49888
16477 * alias.c (offset_overlap_p): New, factored out of...
16478 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
16479 the conservative special case for symbolic constants. Don't
16480 adjust zero sizes on alignment.
16481
c664546f
JL
164822013-01-18 Bernd Schmidt <bernds@codesourcery.com>
16483
16484 PR rtl-optimization/52573
16485 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
16486 REG_UNUSED for the same register.
16487
1bd3f750
MP
164882013-01-17 Richard Biener <rguenther@suse.de>
16489 Marek Polacek <polacek@redhat.com>
16490
16491 PR rtl-optimization/55833
16492 * loop-unswitch.c (unswitch_loops): Move loop verification...
16493 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
16494 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
16495 Set it to true when we're removing a loop from hierarchy tree in
16496 an irreducible region.
16497 (fix_bb_placements): Adjust caller.
16498 (fix_loop_placements): Likewise.
16499
e52a8b71
GJL
165002013-01-17 Georg-Johann Lay <avr@gjlay.de>
16501
16502 * config/avr/builtins.def (DEF_BUILTIN): Factor out
16503 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
16504 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
16505 Remove ID. Adjust comments.
16506 * config/avr/avr-c.c (avr_builtin_name): Remove.
16507 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
16508 * config/avr/avr.c (avr_tolower): New static function.
16509 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
16510 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
16511 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
16512 default expansion.
16513
8386a7ea
JH
165142013-01-17 Jan Hubicka <jh@suse.cz>
16515
610fb637 16516 PR tree-optimization/55273
8386a7ea
JH
16517 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
16518
47876a2a 165192013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB
16520
16521 PR target/55981
16522 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
16523 store through atomic_store<mode>_1.
16524 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
16525
8222c37e
MJ
165262013-01-17 Martin Jambor <mjambor@suse.cz>
16527
16528 PR tree-optimizations/55264
16529 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
16530 for virtual methods.
16531 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
16532 virtual methods before inlining is over.
16533 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
16534 virtual functions.
16535 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
16536 non-virtual.
16537
79f01c76
VM
165382013-01-16 Vladimir Makarov <vmakarov@redhat.com>
16539
16540 PR rtl-optimization/56005
16541 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
16542 pending reads for prefetch.
16543
d0b6bb1b
IB
165442013-01-16 Ian Bolton <ian.bolton@arm.com>
16545
aaf1e810 16546 * config/aarch64/aarch64.md
d0b6bb1b
IB
16547 (*cstoresi_neg_uxtw): New pattern.
16548 (*cmovsi_insn_uxtw): New pattern.
16549 (*<optab>si3_uxtw): New pattern.
16550 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
16551 (*<optab>si3_insn_uxtw): New pattern.
16552 (*bswapsi2_uxtw): New pattern.
16553
cb9cf03b
RB
165542013-01-16 Richard Biener <rguenther@suse.de>
16555
16556 * tree-inline.c (tree_function_versioning): Remove set but
16557 never used variable.
16558
2cfc56b9
RB
165592013-01-16 Richard Biener <rguenther@suse.de>
16560
16561 PR tree-optimization/55964
16562 * tree-flow.h (rename_variables_in_loop): Remove.
16563 (rename_variables_in_bb): Likewise.
16564 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
16565 (copy_loop_before): Adjust and delete update-ssa status.
16566 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 16567 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB
16568 (rename_variables_in_loop): Remove.
16569 (slpeel_update_phis_for_duplicate_loop): Likewise.
16570 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
16571 use available cfg machinery instead of duplicating it.
16572 Update PHI nodes and perform poor-mans SSA update here.
16573 (slpeel_tree_peel_loop_to_edge): Adjust.
16574
c25a0c60
RB
165752013-01-16 Richard Biener <rguenther@suse.de>
16576
16577 PR tree-optimization/54767
16578 PR tree-optimization/53465
16579 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
16580 (vrp_visit_phi_node): For PHI arguments coming via backedges
16581 drop all symbolical range information.
16582 (execute_vrp): Compute backedges.
16583
04b535af
RB
165842013-01-16 Richard Biener <rguenther@suse.de>
16585
16586 * doc/install.texi: Update CLooG and ISL requirements to
16587 0.18.0 and 0.11.1.
16588
8b0a1e0b
CB
165892013-01-16 Christian Bruel <christian.bruel@st.com>
16590
16591 PR target/55301
16592 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
16593 (broken_move): Handle UNSPECV_SP_SWITCH_B.
16594 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
16595
165962013-01-16 DJ Delorie <dj@redhat.com>
16597
16598 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
16599 (UNSPECV_SP_SWITCH_E): New.
16600 (sp_switch_1): Change to an unspec.
16601 (sp_switch_2): Change to an unspec. Don't use post-inc when we
16602 replace $r15.
16603
166042013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB
16605
16606 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
16607 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
16608 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
16609 (expand_mem_thread_fence): Ditto.
16610 (expand_mem_signal_fence): Ditto.
16611 (expand_atomic_load): Ditto.
16612 (expand_atomic_store): Ditto.
16613
0bfbca58 166142013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO
16615
16616 PR rtl-optimization/55547
16617 PR rtl-optimization/53827
16618 PR debug/53671
16619 PR debug/49888
16620 * alias.c (memrefs_conflict_p): Set sizes to negative after
16621 AND adjustments.
16622
305e3ac1
JJ
166232013-01-15 Jakub Jelinek <jakub@redhat.com>
16624
16625 PR target/55940
16626 * function.c (thread_prologue_and_epilogue_insns): Always
16627 add crtl->drap_reg to set_up_by_prologue.set, even if
16628 stack_realign_drap is false.
16629
f78ac4f2
JBG
166302013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16631
16632 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
16633 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
16634 *call): Fix indention.
16635
a78a8cc4
TV
166362013-01-15 Tom de Vries <tom@codesourcery.com>
16637
16638 PR target/55876
16639 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
16640 Update comment.
16641
0e80383f
VM
166422013-01-15 Vladimir Makarov <vmakarov@redhat.com>
16643
305e3ac1 16644 PR rtl-optimization/55153
0e80383f
VM
16645 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
16646
207b5956
MJ
166472013-01-15 Martin Jambor <mjambor@suse.cz>
16648
16649 PR tree-optimization/55920
16650 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
16651 accesses as grp_to_be_debug_replaced.
16652
a7818b54
JJ
166532013-01-15 Jakub Jelinek <jakub@redhat.com>
16654
16655 PR tree-optimization/55920
16656 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
16657 there is non-useless type conversion needed from debug rhs to lhs,
16658 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
16659
b0fe107e
JM
166602013-01-15 Joseph Myers <joseph@codesourcery.com>
16661 Mikael Pettersson <mikpe@it.uu.se>
16662
16663 PR target/43961
16664 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
16665 Thumb.
16666 (ASM_OUTPUT_CASE_LABEL): Remove.
16667 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
16668 * final.c (shorten_branches): Update alignment of labels before
16669 jump tables if CASE_VECTOR_SHORTEN_MODE.
16670
34ab62ee
RB
166712013-01-15 Richard Biener <rguenther@suse.de>
16672
16673 PR bootstrap/55961
16674 * system.h: Do not include gmp.h for building host tools.
16675
783a3a05
RB
166762013-01-15 Richard Biener <rguenther@suse.de>
16677
16678 PR middle-end/55882
16679 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
16680 account for bitpos when computing alignment.
16681
3a579e09
VY
166822013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
16683
16684 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
16685 (ix86_target_macros_internal): Likewise.
16686
16687 * config/i386/i386.c (m_CORE2I7): Removed.
16688 (m_CORE_HASWELL): New macro.
16689 (m_CORE_ALL): Likewise.
16690 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
16691 (initial_ix86_arch_features): Likewise.
16692 (processor_target_table): Initializations for Core avx2.
16693 (cpu_names): New names "core-avx2".
16694 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
16695 PROCESSOR_CORE_HASWELL.
16696 (ix86_issue_rate): New case.
16697 (ia32_multipass_dfa_lookahead): Likewise.
16698 (ix86_sched_init_global): Likewise.
16699
16700 * config/i386/i386.h (TARGET_HASWELL): New macro.
16701 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
16702 (processor_type): New PROCESSOR_HASWELL.
16703
955f5a07
JJ
167042013-01-15 Jakub Jelinek <jakub@redhat.com>
16705
ff784829
JJ
16706 PR tree-optimization/55955
16707 * tree-vect-loop.c (vectorizable_reduction): Give up early on
16708 *SHIFT_EXPR and *ROTATE_EXPR codes.
16709
955f5a07
JJ
16710 PR tree-optimization/48766
16711 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
16712 -ftrapv disable -fwrapv.
16713
aeb8b4e9
GJL
167142013-01-14 Georg-Johann Lay <avr@gjlay.de>
16715
16716 PR target/55974
16717 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
16718 etc. to 1 and not to __flash.
16719 Use LL suffix for __INT24_MAX__ with -mint8.
16720 Use ULL suffix for __UINT24_MAX__ with -mint8.
16721
1c494c6a
GJL
167222013-01-14 Georg-Johann Lay <avr@gjlay.de>
16723
16724 * config/avr/avr-arch.h
16725 (struct base_arch_s): Use typedef avr_arch_t instead.
16726 (struct arch_info_s): Use typedef avr_arch_info_t instead.
16727 (struct mcu_type_s): Use typedef avr_mcu_t instead.
16728 * config/avr/avr.c: Same.
16729 * config/avr/avr-devices.c: Same.
16730 * config/avr/driver-avr.c: Same.
16731 * config/avr/gen-avr-mmcu-texi.c: Same.
16732 * config/avr/avr-mcus.def: Adjust comment.
16733
a50344cb
TB
167342013-01-14 Tejas Belagod <tejas.belagod@arm.com>
16735
88e784e6
UB
16736 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
16737 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 16738
47876a2a 167392013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB
16740 Andi Kleen <ak@linux.intel.com>
16741
16742 PR target/55948
16743 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
16744 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
16745 memmodel flag.
16746
00892272
GJL
167472013-01-14 Georg-Johann Lay <avr@gjlay.de>
16748
16749 * config/avr/avr-stdint.h: Remove trailing blanks.
16750 * config/avr/avr-log.h: Same.
16751 * config/avr/avr-arch.h: Same.
16752 * config/avr/avr-devices.c: Same.
16753 * config/avr/avr-dimode.md: Same.
16754 * config/avr/predicates.md: Same.
16755 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 16756
00892272
GJL
16757 * config/avr/avr-protos.h: Same. And:
16758 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
16759 (init_cumulative_args): Rename to avr_init_cumulative_args.
16760 (expand_prologue): Rename to avr_expand_prologue.
16761 (expand_epilogue): Rename to avr_expand_epilogue.
16762 (adjust_insn_length): Rename to avr_adjust_insn_length.
16763 (notice_update_cc): Rename to avr_notice_update_cc.
16764 (final_prescan_insn): Rename to avr_final_prescan_insn.
16765 * config/avr/avr.c: Same.
16766 * config/avr/avr.h: Same.
16767 * config/avr/avr.md: Remove trailing blanks.
16768 (prologue): Use avr_expand_prologue.
16769 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
16770
7bb01996
RB
167712013-01-14 Richard Biener <rguenther@suse.de>
16772
16773 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
16774 verify_location, collect_subblocks): New functions.
16775 (verify_gimple_in_cfg): Verify that locations only reference
16776 BLOCKs in the functions BLOCK tree.
16777
2724573f
RB
167782013-01-14 Richard Biener <rguenther@suse.de>
16779
16780 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
16781 PHI argument.
16782 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
16783 unshare reference.
16784 (insert_out_of_ssa_copy_on_edge): Likewise.
16785 (rewrite_close_phi_out_of_ssa): Likewise.
16786 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
16787 debug expressions.
16788 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
16789 propagated constants.
16790 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
16791 can not be shared.
16792
9a0bbab6
GJL
167932013-01-14 Georg-Johann Lay <avr@gjlay.de>
16794
16795 * config/avr/avr-modes.def: Add GPL copyright notice.
16796
45805f17
UB
167972013-01-13 Uros Bizjak <ubizjak@gmail.com>
16798
16799 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
16800 MEMMODEL_MASK to determine memory model.
16801 (atomic_store<mode>): Ditto from operands[2].
16802 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
16803
9d60be38
JJ
168042013-01-13 Jakub Jelinek <jakub@redhat.com>
16805
16806 PR fortran/55935
45805f17 16807 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ
16808 (fold_gimple_assign): Don't call unshare_expr here.
16809 (fold_ctor_reference): Call unshare_expr.
16810
e7f49d92
TG
168112013-01-13 Terry Guo <terry.guo@arm.com>
16812
aaf1e810
EB
16813 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
16814 * doc/fragments.texi: Document MULTILIB_REUSE.
16815 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 16816 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 16817 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG
16818 (tmpmultilib4): Ditto.
16819 (multilib_reuse): New multilib argument.
16820
fbd03a27
RS
168212013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
16822
16823 * Makefile.in: Update copyright.
16824
b3681f13
TV
168252013-01-12 Tom de Vries <tom@codesourcery.com>
16826
16827 PR middle-end/55890
16828 * calls.c (expand_call): Check if arg_nr is valid.
16829
3f287e4b
MM
168302013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
16831
16832 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
16833 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM
16834 documentation. Add missing '__' in front of
16835 __builtin_ia32_packssdw256.
3f287e4b 16836
1abcd5eb
AK
168372013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16838
16839 PR target/55719
16840 * config/s390/s390.c (s390_preferred_reload_class): Do not return
16841 NO_REGS for larl operands.
16842 (s390_reload_larl_operand): Use s390_load_address instead of
16843 emit_move_insn.
16844
980d0812
RB
168452013-01-11 Richard Biener <rguenther@suse.de>
16846
16847 * tree-cfg.c (verify_node_sharing_1): Split out from ...
16848 (verify_node_sharing): ... here.
16849 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
16850
7cb9fd07
EB
168512013-01-11 Eric Botcazou <ebotcazou@adacore.com>
16852
16853 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
16854 Substitute TREECHECKING.
16855 * configure: Regenerate.
16856 * Makefile.in (TREECHECKING): New.
16857
47876a2a 168582013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG
16859
16860 PR tree-optimization/44061
16861 * tree-vrp.c (extract_range_basic): Compute zero as
16862 value-range for __builtin_constant_p of function parameters.
16863
d1e082c2
RS
168642013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
16865
45805f17 16866 Update copyright years.
d1e082c2 16867
f85021b0
VM
168682013-01-10 Vladimir Makarov <vmakarov@redhat.com>
16869
9d60be38 16870 PR rtl-optimization/55672
0160442c
VM
16871 * lra-eliminations.c (mark_not_eliminable): Permit addition with
16872 const to be eliminable.
f85021b0 16873
7a8b1ec4
DE
168742013-01-10 David Edelsohn <dje.gcc@gmail.com>
16875
16876 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
16877 * configure: Regenerate.
16878
ad2e5b71
RB
168792013-01-10 Richard Biener <rguenther@suse.de>
16880
ddf9322d 16881 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 16882
ffc5b2bb
RB
168832013-01-10 Richard Biener <rguenther@suse.de>
16884
16885 PR bootstrap/55792
16886 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
16887 locations for virtual PHI arguments.
16888 (rewrite_update_phi_arguments): Likewise.
16889
e1f674e4
JS
168902013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
16891
16892 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
16893 on to assembler.
16894
a57fc743
JJ
168952013-01-10 Jakub Jelinek <jakub@redhat.com>
16896
16897 PR tree-optimization/55921
16898 * tree-complex.c (expand_complex_asm): New function.
16899 (expand_complex_operations_1): Call it for GIMPLE_ASM.
16900
0ff4390d
AK
169012013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16902
16903 PR target/55718
16904 * config/s390/s390.c (s390_symref_operand_p)
16905 (s390_loadrelative_operand_p): Merge the two functions.
16906 (s390_check_qrst_address, print_operand_address): Add parameters
16907 to s390_loadrelative_operand_p invokation.
16908 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
16909 (s390_reload_larl_operand, s390_secondary_reload): Use
16910 s390_loadrelative_operand_p instead of s390_symref_operand_p.
16911 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
16912
6f557e0e
MS
169132013-01-09 Mike Stump <mikestump@comcast.net>
16914
16915 * dse.c (record_store): Remove unnecessary assert.
16916
7770c9e9
JH
169172013-01-09 Jan Hubicka <jh@suse.cz>
16918
16919 PR tree-optimization/55569
16920 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
16921 * cfgloop.h (scale_loop_profile): Likewise.
16922
a19b1432
JH
169232013-01-09 Jan Hubicka <jh@suse.cz>
16924
16925 PR lto/45375
ddf9322d
UB
16926 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
16927 functions.
a19b1432
JH
16928 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
16929
9a002da8
RS
169302013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
16931
16932 PR middle-end/55114
16933 * expr.h (maybe_emit_group_store): Declare.
16934 * expr.c (maybe_emit_group_store): New function.
16935 * builtins.c (expand_builtin_int_roundingfn): Call it.
16936 (expand_builtin_int_roundingfn_2): Likewise.
16937
511dcace
VM
169382013-01-09 Vladimir Makarov <vmakarov@redhat.com>
16939
e1f2b729 16940 PR rtl-optimization/55829
511dcace
VM
16941 * lra-constraints.c (match_reload): Add code for absent output.
16942 (curr_insn_transform): Add code for reloads of matched inputs
16943 without output.
16944
7b0fe4f4
UB
169452013-01-09 Uros Bizjak <ubizjak@gmail.com>
16946
16947 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
16948 attribute of movddup insn to DF.
16949 (*vec_interleave_lowv2df): Ditto.
16950 (vec_dupv2df): Ditto.
16951
870ca331
JH
169522013-01-09 Jan Hubicka <jh@suse.cz>
16953
16954 PR tree-optimiation/55875
16955 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
16956 EVERY_ITERATION parameter.
7b0fe4f4 16957 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH
16958 (idx_infer_loop_bounds): Similarly here.
16959 (n_of_executions_at_most): Simplify
16960 to only test for cases where statement is dominated by the
7b0fe4f4 16961 particular bound; handle correctly the "postdominance" test.
870ca331
JH
16962 (scev_probably_wraps_p): Use max loop iterations info
16963 as a global bound first.
16964
6f575fe4 169652013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD
16966 Nick Clifton <nickc@redhat.com>
16967
16968 * config/v850/v850.md (cbranchsf4): New pattern.
16969 (cstoresf4): New pattern.
16970 (cbranchdf4): New pattern.
16971 (cstoredf4): New pattern.
16972 (movsicc): Disallow floating point comparisons.
16973 (cmpsf_le_insn): Fix order of operators.
16974 (cmpsf_lt_insn): Likewise.
16975 (cmpsf_eq_insn): Likewise.
16976 (cmpdf_le_insn): Likewise.
16977 (cmpdf_lt_insn): Likewise.
16978 (cmpdf_eq_insn): Likewise.
16979 (cmpsf_ge_insn): Use LE comparison.
16980 (cmpdf_ge_insn): Likewise.
16981 (cmpsf_gt_insn): Use LT comparison.
16982 (cmpdf_gt_insn): Likewise.
16983 (cmpsf_ne_insn): Delete pattern.
16984 (cmpdf_ne_insn): Delete pattern.
16985 * config/v850/v850.c (v850_gen_float_compare): Use
16986 gen_cmpdf_eq_insn for NE comparison.
16987 (v850_float_z_comparison_operator)
16988 (v850_float_nz_comparison_operator): Move from here ...
16989 * config/v850/predicates.md: ... to here. Move GT and GE
16990 comparisons into v850_float_z_comparison_operator.
16991 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
16992 Delete prototype.
16993 (v850_float_nz_comparison_operator): Likewise.
16994
f0d54148
JDA
169952013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16996
16997 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
16998 with calls to gen_insvsi/gen_insvdi.
16999
8f01beca
VK
170002013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17001
17002 * config/i386/i386.c (initial_ix86_tune_features): Set up
17003 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17004
2aa579ad
JJ
170052013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17006 Jakub Jelinek <jakub@redhat.com>
17007
17008 PR tree-optimization/48189
17009 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17010 If nitercst is 0, don't predict the exit edge.
17011
6edc3e32 170122013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N
17013
17014 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17015 in asm_fprintf with reg_names.
17016 (aarch64_print_operand_address): Likewise.
17017 (aarch64_return_addr): Likewise.
17018 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17019
f6f94d94
JDA
170202013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17021
17022 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17023 (INT_U6_BITS): Likewise.
17024 * config/pa/predicates.md (uint6_operand): New predicate.
17025 (shift5_operand, shift6_operand): Likewise.
17026 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17027 arith32_operand.
17028 (lshrdi3): Use shift6_operand.
17029 (shrpsi4, shrpdi4): New insn patterns.
17030 (extzv): Delete expander.
17031 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17032 predicates in unamed zero extract patterns. Tighten common constraint.
17033 (extv): Delete expander.
17034 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17035 predicates in unamed sign extract patterns. Tighten common constraint.
17036 (insv): Delete expander.
17037 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17038 predicates in unamed insert patterns. Tighten common constraint.
17039 Change uint32_operand predicate to uint6_operand predicate in unamed
17040 DImode pattern to insert constant values of type 1...1xxxx.
17041
36b72910
JH
170422013-01-04 Jan Hubicka <jh@suse.cz>
17043
17044 PR tree-optimization/55823
7b0fe4f4
UB
17045 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17046 issue.
36b72910 17047
47876a2a 170482013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ
17049 Uros Bizjak <ubizjak@gmail.com>
17050
17051 PR rtl-optimization/55845
17052 * df-problems.c (can_move_insns_across): Stop scanning at
17053 volatile_insn_p source instruction or give up if
17054 across_from .. across_to range contains any volatile_insn_p
17055 instructions.
17056
4369c11e
TB
170572013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17058
7b0fe4f4
UB
17059 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17060 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17061 Declare.
4369c11e 17062 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 17063 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 17064
aeb7e7c1
JJ
170652013-01-08 Jakub Jelinek <jakub@redhat.com>
17066
17067 PR fortran/55341
17068 * asan.c (asan_clear_shadow): New function.
17069 (asan_emit_stack_protection): Use it.
17070
a02ad1aa
TB
170712013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17072
17073 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17074 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17075 with tab instead of space.
17076
f90d87f5
NC
170772013-01-08 Nick Clifton <nickc@redhat.com>
17078
17079 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17080 register bank 0 at the start of an interrupt handler.
83ffd964
NC
17081 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17082 MDBH registers.
f90d87f5 17083
385eb93d
JG
170842013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17085
17086 * config/aarch64/aarch64-simd.md
17087 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17088 (aarch64_simd_bsl): Likewise.
17089 (aarch64_vcond_internal<mode>): Likewise.
17090 (vcond<mode><mode>): Likewise.
17091 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17092 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17093
4dcd1054
JG
170942013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17095
17096 * config/aarch64/aarch64-builtins.c
17097 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17098
4267a4a6
MJ
170992013-01-08 Martin Jambor <mjambor@suse.cz>
17100
17101 PR debug/55579
17102 * tree-sra.c (analyze_access_subtree): Return true also after
17103 potentially creating a debug-only replacement.
17104
5f4e6de3
JJ
171052013-01-08 Jakub Jelinek <jakub@redhat.com>
17106
3138f224
JJ
17107 PR middle-end/55890
17108 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17109
5f4e6de3
JJ
17110 PR tree-optimization/54120
17111 * tree-vrp.c (range_fits_type_p): Don't allow
17112 src_precision < precision from signed vr to unsigned_p
17113 if vr->min or vr->max is negative.
17114 (simplify_float_conversion_using_ranges): Test can_float_p
17115 against CODE_FOR_nothing.
17116
f3ef18ff
JJ
171172013-01-08 Jakub Jelinek <jakub@redhat.com>
17118 Richard Biener <rguenther@suse.de>
17119
17120 PR middle-end/55851
17121 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17122 types instead of just INTEGER_TYPE types.
17123
25c210f9
MK
171242013-01-07 Mark Kettenis <kettenis@openbsd.org>
17125
17126 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17127 TF_SIZE): Define.
f90d87f5 17128
81826a7b
SE
171292013-01-07 Steve Ellcey <sellcey@mips.com>
17130
17131 PR target/42661
17132 * config/mips/mips.opt: Change mad to mmad to match documentation.
17133
35678514
GJL
171342013-01-07 Georg-Johann Lay <avr@gjlay.de>
17135
17136 PR target/55897
17137 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17138 .progmemx.data now.
17139
e5669488
GJL
171402013-01-07 Georg-Johann Lay <avr@gjlay.de>
17141
17142 PR target/55897
17143 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17144 (avr_addrspace_t): Add .section_name field.
17145 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17146 array size.
17147 (avr_addrspace): Same. Initialize .section_name. Remove last
17148 NULL entry. Put __memx into .progmemx.data.
17149 (progmem_section_prefix): Remove.
17150 (avr_asm_init_sections): No need to initialize progmem_section.
17151 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17152 section name prefix.
17153 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17154 retrieve the progmem section.
17155 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17156 boundary to run over avr_addrspace[].
17157 (avr_register_target_pragmas): Ditto.
17158
656e6f37
JJ
171592013-01-06 Jakub Jelinek <jakub@redhat.com>
17160
17161 * varasm.c (output_constant_def_contents): For asan_protect_global
17162 protected strings, adjust DECL_ALIGN if needed, before testing for
17163 anchored symbols.
17164 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17165 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17166 normal decls.
17167 (output_object_block): For asan protected decls, emit asan padding
17168 after their contents.
17169 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17170 (asan_finish_file): Test it here instead.
17171
6edc3e32
UB
171722013-01-07 Nick Clifton <nickc@redhat.com>
17173 Matthias Klose <doko@debian.org>
17174 Doug Kwan <dougkwan@google.com>
17175 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC
17176
17177 PR driver/55470
17178 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17179
17180 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17181
17182 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17183
7b0fe4f4 17184 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC
17185
17186 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17187
2a095093
GJL
171882013-01-07 Georg-Johann Lay <avr@gjlay.de>
17189
f3b503f4 17190 PR target/54461
2a095093
GJL
17191 * doc/install.texi (Cross-Compiler-Specific Options): Document
17192 --with-avrlibc.
17193
383f9b34
TB
171942013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17195
17196 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17197 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17198 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17199 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17200 vqmovun_high_s64): Fix source operand number and update copyright.
17201
3626621a
RB
172022013-01-07 Richard Biener <rguenther@suse.de>
17203
17204 PR middle-end/55890
17205 * gimple.h (gimple_call_builtin_p): New overload.
17206 * gimple.c (validate_call): New function.
17207 (gimple_call_builtin_p): Likewise.
17208 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17209 Use gimple_call_builtin_p.
17210 (find_func_clobbers): Likewise.
17211 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17212 (strlen_optimize_stmt): Likewise.
17213
8b2ea410
JG
172142013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17215
17216 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17217 (vld1q_dup_*): Likewise.
17218 (vld1_*): Likewise.
17219 (vld1q_*): Likewise.
17220 (vld1_lane_*): Likewise.
17221 (vld1q_lane_*): Likewise.
17222
9713d329
RB
172232013-01-07 Richard Biener <rguenther@suse.de>
17224
17225 * lto-streamer.h (LTO_minor_version): Bump to 2.
17226
3520f7cc
JG
172272013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17228
17229 * config/aarch64/aarch64-protos.h
17230 (aarch64_const_double_zero_rtx_p): Rename to...
17231 (aarch64_float_const_zero_rtx_p): ...this.
17232 (aarch64_float_const_representable_p): New.
17233 (aarch64_output_simd_mov_immediate): Likewise.
17234 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17235 move immediate case.
17236 * config/aarch64/aarch64.c
17237 (aarch64_const_double_zero_rtx_p): Rename to...
17238 (aarch64_float_const_zero_rtx_p): ...this.
17239 (aarch64_print_operand): Allow printing of new constants.
17240 (aarch64_valid_floating_const): New.
17241 (aarch64_legitimate_constant_p): Check for valid floating-point
17242 constants.
17243 (aarch64_simd_valid_immediate): Likewise.
17244 (aarch64_vect_float_const_representable_p): New.
17245 (aarch64_float_const_representable_p): Likewise.
17246 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17247 (aarch64_output_simd_mov_immediate): New.
17248 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17249 (*movdf_aarch64): Likewise.
17250 * config/aarch64/constraints.md (Ufc): New.
17251 (Y): call aarch64_float_const_zero_rtx.
17252 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17253
e167c04d
RB
172542013-01-07 Richard Biener <rguenther@suse.de>
17255
17256 PR tree-optimization/55888
17257 PR tree-optimization/55862
17258 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17259 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17260 not if it is contained therein.
17261
0139aaab
GJL
172622013-01-07 Georg-Johann Lay <avr@gjlay.de>
17263
17264 * config/avr/t-avr: Typo.
17265
4a176b23
GJL
172662013-01-07 Georg-Johann Lay <avr@gjlay.de>
17267
17268 PR55243
17269 * config/avr/t-avr: Don't automatically rebuild
17270 $(srcdir)/config/avr/t-multilib
17271 $(srcdir)/config/avr/avr-tables.opt
17272 $(srcdir)/doc/avr-mmcu.texi
17273 (avr-mcus): New phony target to build them on request.
17274 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17275 * avr/avr-mcus.def: Adjust comments.
17276
c7afdc98
UB
172772013-01-07 Uros Bizjak <ubizjak@gmail.com>
17278
17279 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17280
1ab05c31
RS
172812013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17282
17283 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17284
488e3acc
RS
172852013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17286
17287 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17288
a89599d2
DE
172892013-01-05 David Edelsohn <dje.gcc@gmail.com>
17290
17291 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17292 to generate profiling.
17293 * config/rs6000/aix64.h (LIB_SPEC): Same.
17294
70f09188
AP
172952013-01-04 Andrew Pinski <apinski@cavium.com>
17296
17297 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17298 New function.
17299 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17300
918d445f
UB
173012013-01-04 Uros Bizjak <ubizjak@gmail.com>
17302
17303 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17304 unconditionally.
17305 (ix86_expand_move): Ditto.
17306 (ix86_zero_extend_to_Pmode): Ditto.
17307 (ix86_expand_call): Ditto.
17308 (ix86_expand_special_args_builtin): Ditto.
17309 (ix86_expand_builtin): Ditto.
17310
361618ec
RB
173112013-01-04 Richard Biener <rguenther@suse.de>
17312
17313 PR tree-optimization/55862
17314 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17315 translating them through PHI nodes.
17316
87eab554
MJ
173172013-01-04 Martin Jambor <mjambor@suse.cz>
17318
17319 PR tree-optimization/55755
17320 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17321 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17322
33879b9f
MP
173232013-01-04 Marek Polacek <polacek@redhat.com>
17324
17325 PR middle-end/55859
17326 * opts.c (default_options_optimization): Clarify error message.
17327
3068819a
RB
173282013-01-04 Richard Biener <rguenther@suse.de>
17329
17330 PR middle-end/55863
17331 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17332 reassociation.
17333
0e4ae794
JDA
173342013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17335
17336 PR target/53789
17337 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA
17338 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17339 references.
0e4ae794 17340
65c88cef
RH
173412013-01-03 Richard Henderson <rth@redhat.com>
17342
17343 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17344 after eliminating TLS symbols.
17345
8ac16127
MG
173462013-01-03 Marc Glisse <marc.glisse@inria.fr>
17347
17348 PR bootstrap/50167
17349 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
17350 * graphite-poly.c (debug_gmp_value): Likewise.
17351
bb664f09
UB
173522013-01-03 Uros Bizjak <ubizjak@gmail.com>
17353
17354 PR target/55712
17355 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
17356 selected code model, define __code_mode_small__, __code_model_medium__,
17357 __code_model_large__, __code_model_32__ or __code_model_kernel__.
17358 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
17359 xchg temporary register with %k. Declare temporary register as
17360 early clobbered.
17361 [__x86_64__]: For medium and large code models, preserve %rbx register.
17362
bcf1ef00
RB
173632013-01-03 Richard Biener <rguenther@suse.de>
17364
0506634a 17365 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB
17366 (dump_subscript): Adjust.
17367 (finalize_ddr_dependent): Do not dump redundant info.
17368 (analyze_siv_subscript): Adjust.
17369 (subscript_dependence_tester): Likewise.
17370 (compute_affine_dependence): Likewise.
17371
59fd17e3
RB
173722013-01-03 Richard Biener <rguenther@suse.de>
17373
17374 Revert
17375 2013-01-03 Richard Biener <rguenther@suse.de>
17376
17377 PR tree-optimization/55857
17378 * tree-vect-stmts.c (vectorizable_load): Do not setup
17379 re-alignment for invariant loads.
17380
17381 2013-01-02 Richard Biener <rguenther@suse.de>
17382
17383 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 17384 invariant load do not generate a vector load from the scalar location.
59fd17e3 17385
595c2679
RB
173862013-01-03 Richard Biener <rguenther@suse.de>
17387
17388 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
17389 for not vectorizing.
17390 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
17391 not build INDIRECT_REFs, call get_name once only.
17392 (vect_create_data_ref_ptr): Likewise. Dump base object kind
17393 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
17394
90bb7d7a
RB
173952013-01-03 Richard Biener <rguenther@suse.de>
17396
17397 PR tree-optimization/55857
17398 * tree-vect-stmts.c (vectorizable_load): Do not setup
17399 re-alignment for invariant loads.
17400
f09b77ca
RB
174012013-01-03 Richard Biener <rguenther@suse.de>
17402
17403 PR lto/55848
17404 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
17405 prefer a built-in decl.
17406
3461a16e
JJ
174072013-01-03 Jakub Jelinek <jakub@redhat.com>
17408
df8e2b4f
JJ
17409 * gcc.c (process_command): Update copyright notice dates.
17410 * gcov.c (print_version): Likewise.
17411 * gcov-dump.c (print_version): Likewise.
17412
3461a16e
JJ
17413 PR rtl-optimization/55838
17414 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
17415 iv0.step, iv1.step and step.
17416
8b5546d6
JJ
174172013-01-03 Jakub Jelinek <jakub@redhat.com>
17418 Marc Glisse <marc.glisse@inria.fr>
17419
17420 PR tree-optimization/55832
17421 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
17422 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
17423 integer_{one,zero}_node.
17424
8ab1d2e9
JJ
174252013-01-03 Jakub Jelinek <jakub@redhat.com>
17426
17427 PR debug/54402
17428 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
17429 * var-tracking.c (reverse_op): Don't add reverse ops to
17430 VALUEs that have already
0506634a 17431 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 17432
5b9ad1d4
GP
174332013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
17434
17435 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
17436
e25a6711
TJ
174372013-01-02 Teresa Johnson <tejohnson@google.com>
17438
17439 * dumpfile.c (dump_loc): Print filename with location.
17440 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
17441 new location_t parameter to emit complete unroll message with
17442 new dump framework.
17443 (canonicalize_loop_induction_variables): Compute loops location
17444 and pass to try_unroll_loop_completely.
17445 * loop-unroll.c (report_unroll_peel): New function.
17446 (peel_loops_completely): Use new dump format with location
17447 for main dumpfile message, and invoke report_unroll_peel on success.
17448 (decide_unrolling_and_peeling): Ditto.
17449 (decide_peel_once_rolling): Remove old dumpfile message subsumed
17450 by report_unroll_peel.
17451 (decide_peel_completely): Ditto.
17452 (decide_unroll_constant_iterations): Ditto.
17453 (decide_unroll_runtime_iterations): Ditto.
17454 (decide_peel_simple): Ditto.
17455 (decide_unroll_stupid): Ditto.
17456 * cfgloop.c (get_loop_location): New function.
17457 * cfgloop.h (get_loop_location): Declare.
17458
77878621
ST
174592013-01-02 Sriraman Tallam <tmsriram@google.com>
17460
17461 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
17462 NULL.
17463
9e65d03e
JDA
174642013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17465
17466 PR middle-end/55198
17467 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
17468 BLKmode objects when EXPAND_MEMORY is specified.
17469
6a7da30f
ST
174702013-01-02 Sriraman Tallam <tmsriram@google.com>
17471
17472 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
17473 in loop predicate.
17474 (fold_builtin_cpu): Do not share cpu model decls across statements.
17475
e78167f2
JM
174762013-01-02 Jason Merrill <jason@redhat.com>
17477
17478 PR c++/55804
17479 * tree.c (build_array_type_1): Revert earlier change.
17480
8c075fb4
YZ
174812013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
17482
17483 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
17484 "cortex-a57".
17485 * config/aarch64/aarch64-tune.md: Re-generate.
17486
0682ed3e
RB
174872013-01-02 Richard Biener <rguenther@suse.de>
17488
17489 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 17490 invariant load do not generate a vector load from the scalar location.
0682ed3e 17491
b8f6a302
RB
174922013-01-02 Richard Biener <rguenther@suse.de>
17493
17494 PR bootstrap/55784
17495 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
17496 * configure: Regenerate.
17497
04b80dbb
RS
174982013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17499
17500 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
17501 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
17502 (expand_builtin_int_roundingfn_2): Keep the original target around
17503 for the fallback case.
17504
635b0b0c
RS
175052013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17506
17507 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
17508 to be clear for sign changes.
17509
42d57399
JH
175102013-01-01 Jan Hubicka <jh@suse.cz>
17511
17512 * ipa-inline-analysis.c: Fix formatting.
17513
5bb6669d
JJ
175142013-01-01 Jakub Jelinek <jakub@redhat.com>
17515
17516 PR tree-optimization/55831
17517 * tree-vect-loop.c (get_initial_def_for_induction): Use
17518 gsi_after_labels instead of gsi_start_bb.
ad41bd84 17519\f
86a2db33 17520Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM
17521
17522Copying and distribution of this file, with or without modification,
17523are permitted in any medium without royalty provided the copyright
17524notice and this notice are preserved.
This page took 7.164901 seconds and 5 git commands to generate.