]> gcc.gnu.org Git - gcc.git/blame - gcc/ChangeLog
Automated conversion of passes to C++ classes
[gcc.git] / gcc / ChangeLog
CommitLineData
27a4cd48
DM
12013-08-05 David Malcolm <dmalcolm@redhat.com>
2
3 This is the automated part of the conversion of passes from C
4 structs to C++ classes.
5
6 Patch autogenerated by refactor_passes.py from
7 https://github.com/davidmalcolm/gcc-refactoring-scripts
8 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9
10 * asan.c (pass_asan): Convert from a global struct to a subclass of
11 gimple_opt_pass along with...
12 (pass_data_asan): ...new pass_data instance and...
13 (make_pass_asan): ...new function.
14 (pass_asan_O0): Convert from a global struct to a subclass of
15 gimple_opt_pass along with...
16 (pass_data_asan_O0): ...new pass_data instance and...
17 (make_pass_asan_O0): ...new function.
18 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
19 subclass of rtl_opt_pass along with...
20 (pass_data_inc_dec): ...new pass_data instance and...
21 (make_pass_inc_dec): ...new function.
22 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
23 a subclass of rtl_opt_pass along with...
24 (pass_data_reorder_blocks): ...new pass_data instance and...
25 (make_pass_reorder_blocks): ...new function.
26 (pass_duplicate_computed_gotos): Convert from a global struct to a
27 subclass of rtl_opt_pass along with...
28 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
29 (make_pass_duplicate_computed_gotos): ...new function.
30 (pass_partition_blocks): Convert from a global struct to a subclass of
31 rtl_opt_pass along with...
32 (pass_data_partition_blocks): ...new pass_data instance and...
33 (make_pass_partition_blocks): ...new function.
34 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
35 struct to a subclass of rtl_opt_pass along with...
36 (pass_data_branch_target_load_optimize1): ...new pass_data instance
37 and...
38 (make_pass_branch_target_load_optimize1): ...new function.
39 (pass_branch_target_load_optimize2): Convert from a global struct to a
40 subclass of rtl_opt_pass along with...
41 (pass_data_branch_target_load_optimize2): ...new pass_data instance
42 and...
43 (make_pass_branch_target_load_optimize2): ...new function.
44 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
45 of rtl_opt_pass along with...
46 (pass_data_jump): ...new pass_data instance and...
47 (make_pass_jump): ...new function.
48 (pass_jump2): Convert from a global struct to a subclass of
49 rtl_opt_pass along with...
50 (pass_data_jump2): ...new pass_data instance and...
51 (make_pass_jump2): ...new function.
52 * cfgexpand.c (pass_expand): Convert from a global struct to a
53 subclass of rtl_opt_pass along with...
54 (pass_data_expand): ...new pass_data instance and...
55 (make_pass_expand): ...new function.
56 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
57 of rtl_opt_pass along with...
58 (pass_data_free_cfg): ...new pass_data instance and...
59 (make_pass_free_cfg): ...new function.
60 (pass_into_cfg_layout_mode): Convert from a global struct to a
61 subclass of rtl_opt_pass along with...
62 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
63 (make_pass_into_cfg_layout_mode): ...new function.
64 (pass_outof_cfg_layout_mode): Convert from a global struct to a
65 subclass of rtl_opt_pass along with...
66 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
67 (make_pass_outof_cfg_layout_mode): ...new function.
68 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
69 struct to a subclass of gimple_opt_pass along with...
70 (pass_data_build_cgraph_edges): ...new pass_data instance and...
71 (make_pass_build_cgraph_edges): ...new function.
72 (pass_rebuild_cgraph_edges): Convert from a global struct to a
73 subclass of gimple_opt_pass along with...
74 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
75 (make_pass_rebuild_cgraph_edges): ...new function.
76 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
77 subclass of gimple_opt_pass along with...
78 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
79 and...
80 (make_pass_remove_cgraph_callee_edges): ...new function.
81 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
82 struct to a subclass of rtl_opt_pass along with...
83 (pass_data_stack_adjustments): ...new pass_data instance and...
84 (make_pass_stack_adjustments): ...new function.
85 * combine.c (pass_combine): Convert from a global struct to a subclass
86 of rtl_opt_pass along with...
87 (pass_data_combine): ...new pass_data instance and...
88 (make_pass_combine): ...new function.
89 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
90 global struct to a subclass of rtl_opt_pass along with...
91 (pass_data_compare_elim_after_reload): ...new pass_data instance
92 and...
93 (make_pass_compare_elim_after_reload): ...new function.
94 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
95 of rtl_opt_pass along with...
96 (pass_data_rtl_cprop): ...new pass_data instance and...
97 (make_pass_rtl_cprop): ...new function.
98 * cse.c (pass_cse): Convert from a global struct to a subclass of
99 rtl_opt_pass along with...
100 (pass_data_cse): ...new pass_data instance and...
101 (make_pass_cse): ...new function.
102 (pass_cse2): Convert from a global struct to a subclass of
103 rtl_opt_pass along with...
104 (pass_data_cse2): ...new pass_data instance and...
105 (make_pass_cse2): ...new function.
106 (pass_cse_after_global_opts): Convert from a global struct to a
107 subclass of rtl_opt_pass along with...
108 (pass_data_cse_after_global_opts): ...new pass_data instance and...
109 (make_pass_cse_after_global_opts): ...new function.
110 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
111 of rtl_opt_pass along with...
112 (pass_data_ud_rtl_dce): ...new pass_data instance and...
113 (make_pass_ud_rtl_dce): ...new function.
114 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
115 rtl_opt_pass along with...
116 (pass_data_fast_rtl_dce): ...new pass_data instance and...
117 (make_pass_fast_rtl_dce): ...new function.
118 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
119 a subclass of rtl_opt_pass along with...
120 (pass_data_df_initialize_opt): ...new pass_data instance and...
121 (make_pass_df_initialize_opt): ...new function.
122 (pass_df_initialize_no_opt): Convert from a global struct to a
123 subclass of rtl_opt_pass along with...
124 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
125 (make_pass_df_initialize_no_opt): ...new function.
126 (pass_df_finish): Convert from a global struct to a subclass of
127 rtl_opt_pass along with...
128 (pass_data_df_finish): ...new pass_data instance and...
129 (make_pass_df_finish): ...new function.
130 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
131 rtl_opt_pass along with...
132 (pass_data_rtl_dse1): ...new pass_data instance and...
133 (make_pass_rtl_dse1): ...new function.
134 (pass_rtl_dse2): Convert from a global struct to a subclass of
135 rtl_opt_pass along with...
136 (pass_data_rtl_dse2): ...new pass_data instance and...
137 (make_pass_rtl_dse2): ...new function.
138 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
139 subclass of rtl_opt_pass along with...
140 (pass_data_dwarf2_frame): ...new pass_data instance and...
141 (make_pass_dwarf2_frame): ...new function.
142 * except.c (pass_set_nothrow_function_flags): Convert from a global
143 struct to a subclass of rtl_opt_pass along with...
144 (pass_data_set_nothrow_function_flags): ...new pass_data instance
145 and...
146 (make_pass_set_nothrow_function_flags): ...new function.
147 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
148 subclass of rtl_opt_pass along with...
149 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
150 and...
151 (make_pass_convert_to_eh_region_ranges): ...new function.
152 * final.c (pass_compute_alignments): Convert from a global struct to a
153 subclass of rtl_opt_pass along with...
154 (pass_data_compute_alignments): ...new pass_data instance and...
155 (make_pass_compute_alignments): ...new function.
156 (pass_final): Convert from a global struct to a subclass of
157 rtl_opt_pass along with...
158 (pass_data_final): ...new pass_data instance and...
159 (make_pass_final): ...new function.
160 (pass_shorten_branches): Convert from a global struct to a subclass of
161 rtl_opt_pass along with...
162 (pass_data_shorten_branches): ...new pass_data instance and...
163 (make_pass_shorten_branches): ...new function.
164 (pass_clean_state): Convert from a global struct to a subclass of
165 rtl_opt_pass along with...
166 (pass_data_clean_state): ...new pass_data instance and...
167 (make_pass_clean_state): ...new function.
168 * function.c (pass_instantiate_virtual_regs): Convert from a global
169 struct to a subclass of rtl_opt_pass along with...
170 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
171 (make_pass_instantiate_virtual_regs): ...new function.
172 (pass_leaf_regs): Convert from a global struct to a subclass of
173 rtl_opt_pass along with...
174 (pass_data_leaf_regs): ...new pass_data instance and...
175 (make_pass_leaf_regs): ...new function.
176 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
177 subclass of rtl_opt_pass along with...
178 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
179 and...
180 (make_pass_thread_prologue_and_epilogue): ...new function.
181 (pass_match_asm_constraints): Convert from a global struct to a
182 subclass of rtl_opt_pass along with...
183 (pass_data_match_asm_constraints): ...new pass_data instance and...
184 (make_pass_match_asm_constraints): ...new function.
185 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
186 subclass of rtl_opt_pass along with...
187 (pass_data_rtl_fwprop): ...new pass_data instance and...
188 (make_pass_rtl_fwprop): ...new function.
189 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
190 rtl_opt_pass along with...
191 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
192 (make_pass_rtl_fwprop_addr): ...new function.
193 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
194 rtl_opt_pass along with...
195 (pass_data_rtl_pre): ...new pass_data instance and...
196 (make_pass_rtl_pre): ...new function.
197 (pass_rtl_hoist): Convert from a global struct to a subclass of
198 rtl_opt_pass along with...
199 (pass_data_rtl_hoist): ...new pass_data instance and...
200 (make_pass_rtl_hoist): ...new function.
201 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
202 subclass of gimple_opt_pass along with...
203 (pass_data_lower_cf): ...new pass_data instance and...
204 (make_pass_lower_cf): ...new function.
205 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
206 from a global struct to a subclass of gimple_opt_pass along with...
207 (pass_data_strength_reduction): ...new pass_data instance and...
208 (make_pass_strength_reduction): ...new function.
209 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
210 of rtl_opt_pass along with...
211 (pass_data_rtl_ifcvt): ...new pass_data instance and...
212 (make_pass_rtl_ifcvt): ...new function.
213 (pass_if_after_combine): Convert from a global struct to a subclass of
214 rtl_opt_pass along with...
215 (pass_data_if_after_combine): ...new pass_data instance and...
216 (make_pass_if_after_combine): ...new function.
217 (pass_if_after_reload): Convert from a global struct to a subclass of
218 rtl_opt_pass along with...
219 (pass_data_if_after_reload): ...new pass_data instance and...
220 (make_pass_if_after_reload): ...new function.
221 * init-regs.c (pass_initialize_regs): Convert from a global struct to
222 a subclass of rtl_opt_pass along with...
223 (pass_data_initialize_regs): ...new pass_data instance and...
224 (make_pass_initialize_regs): ...new function.
225 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
226 of ipa_opt_pass_d along with...
227 (pass_data_ipa_cp): ...new pass_data instance and...
228 (make_pass_ipa_cp): ...new function.
229 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
230 global struct to a subclass of gimple_opt_pass along with...
231 (pass_data_inline_parameters): ...new pass_data instance and...
232 (make_pass_inline_parameters): ...new function.
233 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
234 subclass of gimple_opt_pass along with...
235 (pass_data_early_inline): ...new pass_data instance and...
236 (make_pass_early_inline): ...new function.
237 (pass_ipa_inline): Convert from a global struct to a subclass of
238 ipa_opt_pass_d along with...
239 (pass_data_ipa_inline): ...new pass_data instance and...
240 (make_pass_ipa_inline): ...new function.
241 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
242 struct to a subclass of gimple_opt_pass along with...
243 (pass_data_local_pure_const): ...new pass_data instance and...
244 (make_pass_local_pure_const): ...new function.
245 (pass_ipa_pure_const): Convert from a global struct to a subclass of
246 ipa_opt_pass_d along with...
247 (pass_data_ipa_pure_const): ...new pass_data instance and...
248 (make_pass_ipa_pure_const): ...new function.
249 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
250 to a subclass of ipa_opt_pass_d along with...
251 (pass_data_ipa_reference): ...new pass_data instance and...
252 (make_pass_ipa_reference): ...new function.
253 * ipa-split.c (pass_split_functions): Convert from a global struct to
254 a subclass of gimple_opt_pass along with...
255 (pass_data_split_functions): ...new pass_data instance and...
256 (make_pass_split_functions): ...new function.
257 (pass_feedback_split_functions): Convert from a global struct to a
258 subclass of gimple_opt_pass along with...
259 (pass_data_feedback_split_functions): ...new pass_data instance and...
260 (make_pass_feedback_split_functions): ...new function.
261 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
262 global struct to a subclass of simple_ipa_opt_pass along with...
263 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
264 instance and...
265 (make_pass_ipa_function_and_variable_visibility): ...new function.
266 (pass_ipa_free_inline_summary): Convert from a global struct to a
267 subclass of simple_ipa_opt_pass along with...
268 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
269 (make_pass_ipa_free_inline_summary): ...new function.
270 (pass_ipa_whole_program_visibility): Convert from a global struct to a
271 subclass of ipa_opt_pass_d along with...
272 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
273 and...
274 (make_pass_ipa_whole_program_visibility): ...new function.
275 (pass_ipa_profile): Convert from a global struct to a subclass of
276 ipa_opt_pass_d along with...
277 (pass_data_ipa_profile): ...new pass_data instance and...
278 (make_pass_ipa_profile): ...new function.
279 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
280 ipa_opt_pass_d along with...
281 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
282 (make_pass_ipa_cdtor_merge): ...new function.
283 * ira.c (pass_ira): Convert from a global struct to a subclass of
284 rtl_opt_pass along with...
285 (pass_data_ira): ...new pass_data instance and...
286 (make_pass_ira): ...new function.
287 (pass_reload): Convert from a global struct to a subclass of
288 rtl_opt_pass along with...
289 (pass_data_reload): ...new pass_data instance and...
290 (make_pass_reload): ...new function.
291 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
292 subclass of rtl_opt_pass along with...
293 (pass_data_cleanup_barriers): ...new pass_data instance and...
294 (make_pass_cleanup_barriers): ...new function.
295 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
296 of rtl_opt_pass along with...
297 (pass_data_loop2): ...new pass_data instance and...
298 (make_pass_loop2): ...new function.
299 (pass_rtl_loop_init): Convert from a global struct to a subclass of
300 rtl_opt_pass along with...
301 (pass_data_rtl_loop_init): ...new pass_data instance and...
302 (make_pass_rtl_loop_init): ...new function.
303 (pass_rtl_loop_done): Convert from a global struct to a subclass of
304 rtl_opt_pass along with...
305 (pass_data_rtl_loop_done): ...new pass_data instance and...
306 (make_pass_rtl_loop_done): ...new function.
307 (pass_rtl_move_loop_invariants): Convert from a global struct to a
308 subclass of rtl_opt_pass along with...
309 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
310 (make_pass_rtl_move_loop_invariants): ...new function.
311 (pass_rtl_unswitch): Convert from a global struct to a subclass of
312 rtl_opt_pass along with...
313 (pass_data_rtl_unswitch): ...new pass_data instance and...
314 (make_pass_rtl_unswitch): ...new function.
315 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
316 subclass of rtl_opt_pass along with...
317 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
318 and...
319 (make_pass_rtl_unroll_and_peel_loops): ...new function.
320 (pass_rtl_doloop): Convert from a global struct to a subclass of
321 rtl_opt_pass along with...
322 (pass_data_rtl_doloop): ...new pass_data instance and...
323 (make_pass_rtl_doloop): ...new function.
324 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
325 a subclass of rtl_opt_pass along with...
326 (pass_data_lower_subreg): ...new pass_data instance and...
327 (make_pass_lower_subreg): ...new function.
328 (pass_lower_subreg2): Convert from a global struct to a subclass of
329 rtl_opt_pass along with...
330 (pass_data_lower_subreg2): ...new pass_data instance and...
331 (make_pass_lower_subreg2): ...new function.
332 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
333 struct to a subclass of ipa_opt_pass_d along with...
334 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
335 (make_pass_ipa_lto_gimple_out): ...new function.
336 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
337 of ipa_opt_pass_d along with...
338 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
339 (make_pass_ipa_lto_finish_out): ...new function.
340 * mode-switching.c (pass_mode_switching): Convert from a global struct
341 to a subclass of rtl_opt_pass along with...
342 (pass_data_mode_switching): ...new pass_data instance and...
343 (make_pass_mode_switching): ...new function.
344 * modulo-sched.c (pass_sms): Convert from a global struct to a
345 subclass of rtl_opt_pass along with...
346 (pass_data_sms): ...new pass_data instance and...
347 (make_pass_sms): ...new function.
348 * omp-low.c (pass_expand_omp): Convert from a global struct to a
349 subclass of gimple_opt_pass along with...
350 (pass_data_expand_omp): ...new pass_data instance and...
351 (make_pass_expand_omp): ...new function.
352 (pass_lower_omp): Convert from a global struct to a subclass of
353 gimple_opt_pass along with...
354 (pass_data_lower_omp): ...new pass_data instance and...
355 (make_pass_lower_omp): ...new function.
356 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
357 of gimple_opt_pass along with...
358 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
359 (make_pass_diagnose_omp_blocks): ...new function.
360 * passes.c (pass_early_local_passes): Convert from a global struct to
361 a subclass of simple_ipa_opt_pass along with...
362 (pass_data_early_local_passes): ...new pass_data instance and...
363 (make_pass_early_local_passes): ...new function.
364 (pass_all_early_optimizations): Convert from a global struct to a
365 subclass of gimple_opt_pass along with...
366 (pass_data_all_early_optimizations): ...new pass_data instance and...
367 (make_pass_all_early_optimizations): ...new function.
368 (pass_all_optimizations): Convert from a global struct to a subclass
369 of gimple_opt_pass along with...
370 (pass_data_all_optimizations): ...new pass_data instance and...
371 (make_pass_all_optimizations): ...new function.
372 (pass_all_optimizations_g): Convert from a global struct to a subclass
373 of gimple_opt_pass along with...
374 (pass_data_all_optimizations_g): ...new pass_data instance and...
375 (make_pass_all_optimizations_g): ...new function.
376 (pass_rest_of_compilation): Convert from a global struct to a subclass
377 of rtl_opt_pass along with...
378 (pass_data_rest_of_compilation): ...new pass_data instance and...
379 (make_pass_rest_of_compilation): ...new function.
380 (pass_postreload): Convert from a global struct to a subclass of
381 rtl_opt_pass along with...
382 (pass_data_postreload): ...new pass_data instance and...
383 (make_pass_postreload): ...new function.
384 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
385 subclass of rtl_opt_pass along with...
386 (pass_data_gcse2): ...new pass_data instance and...
387 (make_pass_gcse2): ...new function.
388 * postreload.c (pass_postreload_cse): Convert from a global struct to
389 a subclass of rtl_opt_pass along with...
390 (pass_data_postreload_cse): ...new pass_data instance and...
391 (make_pass_postreload_cse): ...new function.
392 * predict.c (pass_profile): Convert from a global struct to a subclass
393 of gimple_opt_pass along with...
394 (pass_data_profile): ...new pass_data instance and...
395 (make_pass_profile): ...new function.
396 (pass_strip_predict_hints): Convert from a global struct to a subclass
397 of gimple_opt_pass along with...
398 (pass_data_strip_predict_hints): ...new pass_data instance and...
399 (make_pass_strip_predict_hints): ...new function.
400 * recog.c (pass_peephole2): Convert from a global struct to a subclass
401 of rtl_opt_pass along with...
402 (pass_data_peephole2): ...new pass_data instance and...
403 (make_pass_peephole2): ...new function.
404 (pass_split_all_insns): Convert from a global struct to a subclass of
405 rtl_opt_pass along with...
406 (pass_data_split_all_insns): ...new pass_data instance and...
407 (make_pass_split_all_insns): ...new function.
408 (pass_split_after_reload): Convert from a global struct to a subclass
409 of rtl_opt_pass along with...
410 (pass_data_split_after_reload): ...new pass_data instance and...
411 (make_pass_split_after_reload): ...new function.
412 (pass_split_before_regstack): Convert from a global struct to a
413 subclass of rtl_opt_pass along with...
414 (pass_data_split_before_regstack): ...new pass_data instance and...
415 (make_pass_split_before_regstack): ...new function.
416 (pass_split_before_sched2): Convert from a global struct to a subclass
417 of rtl_opt_pass along with...
418 (pass_data_split_before_sched2): ...new pass_data instance and...
419 (make_pass_split_before_sched2): ...new function.
420 (pass_split_for_shorten_branches): Convert from a global struct to a
421 subclass of rtl_opt_pass along with...
422 (pass_data_split_for_shorten_branches): ...new pass_data instance
423 and...
424 (make_pass_split_for_shorten_branches): ...new function.
425 * ree.c (pass_ree): Convert from a global struct to a subclass of
426 rtl_opt_pass along with...
427 (pass_data_ree): ...new pass_data instance and...
428 (make_pass_ree): ...new function.
429 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
430 subclass of rtl_opt_pass along with...
431 (pass_data_stack_regs): ...new pass_data instance and...
432 (make_pass_stack_regs): ...new function.
433 (pass_stack_regs_run): Convert from a global struct to a subclass of
434 rtl_opt_pass along with...
435 (pass_data_stack_regs_run): ...new pass_data instance and...
436 (make_pass_stack_regs_run): ...new function.
437 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
438 subclass of rtl_opt_pass along with...
439 (pass_data_cprop_hardreg): ...new pass_data instance and...
440 (make_pass_cprop_hardreg): ...new function.
441 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
442 subclass of rtl_opt_pass along with...
443 (pass_data_reginfo_init): ...new pass_data instance and...
444 (make_pass_reginfo_init): ...new function.
445 * regmove.c (pass_regmove): Convert from a global struct to a subclass
446 of rtl_opt_pass along with...
447 (pass_data_regmove): ...new pass_data instance and...
448 (make_pass_regmove): ...new function.
449 * regrename.c (pass_regrename): Convert from a global struct to a
450 subclass of rtl_opt_pass along with...
451 (pass_data_regrename): ...new pass_data instance and...
452 (make_pass_regrename): ...new function.
453 * reorg.c (pass_delay_slots): Convert from a global struct to a
454 subclass of rtl_opt_pass along with...
455 (pass_data_delay_slots): ...new pass_data instance and...
456 (make_pass_delay_slots): ...new function.
457 (pass_machine_reorg): Convert from a global struct to a subclass of
458 rtl_opt_pass along with...
459 (pass_data_machine_reorg): ...new pass_data instance and...
460 (make_pass_machine_reorg): ...new function.
461 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
462 of rtl_opt_pass along with...
463 (pass_data_sched): ...new pass_data instance and...
464 (make_pass_sched): ...new function.
465 (pass_sched2): Convert from a global struct to a subclass of
466 rtl_opt_pass along with...
467 (pass_data_sched2): ...new pass_data instance and...
468 (make_pass_sched2): ...new function.
469 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
470 to a subclass of rtl_opt_pass along with...
471 (pass_data_stack_ptr_mod): ...new pass_data instance and...
472 (make_pass_stack_ptr_mod): ...new function.
473 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
474 to a subclass of rtl_opt_pass along with...
475 (pass_data_rtl_store_motion): ...new pass_data instance and...
476 (make_pass_rtl_store_motion): ...new function.
477 * tracer.c (pass_tracer): Convert from a global struct to a subclass
478 of gimple_opt_pass along with...
479 (pass_data_tracer): ...new pass_data instance and...
480 (make_pass_tracer): ...new function.
481 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
482 to a subclass of gimple_opt_pass along with...
483 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
484 (make_pass_diagnose_tm_blocks): ...new function.
485 (pass_lower_tm): Convert from a global struct to a subclass of
486 gimple_opt_pass along with...
487 (pass_data_lower_tm): ...new pass_data instance and...
488 (make_pass_lower_tm): ...new function.
489 (pass_tm_init): Convert from a global struct to a subclass of
490 gimple_opt_pass along with...
491 (pass_data_tm_init): ...new pass_data instance and...
492 (make_pass_tm_init): ...new function.
493 (pass_tm_mark): Convert from a global struct to a subclass of
494 gimple_opt_pass along with...
495 (pass_data_tm_mark): ...new pass_data instance and...
496 (make_pass_tm_mark): ...new function.
497 (pass_tm_edges): Convert from a global struct to a subclass of
498 gimple_opt_pass along with...
499 (pass_data_tm_edges): ...new pass_data instance and...
500 (make_pass_tm_edges): ...new function.
501 (pass_tm_memopt): Convert from a global struct to a subclass of
502 gimple_opt_pass along with...
503 (pass_data_tm_memopt): ...new pass_data instance and...
504 (make_pass_tm_memopt): ...new function.
505 (pass_ipa_tm): Convert from a global struct to a subclass of
506 simple_ipa_opt_pass along with...
507 (pass_data_ipa_tm): ...new pass_data instance and...
508 (make_pass_ipa_tm): ...new function.
509 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
510 subclass of gimple_opt_pass along with...
511 (pass_data_call_cdce): ...new pass_data instance and...
512 (make_pass_call_cdce): ...new function.
513 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
514 subclass of gimple_opt_pass along with...
515 (pass_data_build_cfg): ...new pass_data instance and...
516 (make_pass_build_cfg): ...new function.
517 (pass_split_crit_edges): Convert from a global struct to a subclass of
518 gimple_opt_pass along with...
519 (pass_data_split_crit_edges): ...new pass_data instance and...
520 (make_pass_split_crit_edges): ...new function.
521 (pass_warn_function_return): Convert from a global struct to a
522 subclass of gimple_opt_pass along with...
523 (pass_data_warn_function_return): ...new pass_data instance and...
524 (make_pass_warn_function_return): ...new function.
525 (pass_warn_function_noreturn): Convert from a global struct to a
526 subclass of gimple_opt_pass along with...
527 (pass_data_warn_function_noreturn): ...new pass_data instance and...
528 (make_pass_warn_function_noreturn): ...new function.
529 (pass_warn_unused_result): Convert from a global struct to a subclass
530 of gimple_opt_pass along with...
531 (pass_data_warn_unused_result): ...new pass_data instance and...
532 (make_pass_warn_unused_result): ...new function.
533 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
534 a subclass of gimple_opt_pass along with...
535 (pass_data_merge_phi): ...new pass_data instance and...
536 (make_pass_merge_phi): ...new function.
537 * tree-complex.c (pass_lower_complex): Convert from a global struct to
538 a subclass of gimple_opt_pass along with...
539 (pass_data_lower_complex): ...new pass_data instance and...
540 (make_pass_lower_complex): ...new function.
541 (pass_lower_complex_O0): Convert from a global struct to a subclass of
542 gimple_opt_pass along with...
543 (pass_data_lower_complex_O0): ...new pass_data instance and...
544 (make_pass_lower_complex_O0): ...new function.
545 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
546 subclass of gimple_opt_pass along with...
547 (pass_data_lower_eh): ...new pass_data instance and...
548 (make_pass_lower_eh): ...new function.
549 (pass_refactor_eh): Convert from a global struct to a subclass of
550 gimple_opt_pass along with...
551 (pass_data_refactor_eh): ...new pass_data instance and...
552 (make_pass_refactor_eh): ...new function.
553 (pass_lower_resx): Convert from a global struct to a subclass of
554 gimple_opt_pass along with...
555 (pass_data_lower_resx): ...new pass_data instance and...
556 (make_pass_lower_resx): ...new function.
557 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
558 of gimple_opt_pass along with...
559 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
560 (make_pass_lower_eh_dispatch): ...new function.
561 (pass_cleanup_eh): Convert from a global struct to a subclass of
562 gimple_opt_pass along with...
563 (pass_data_cleanup_eh): ...new pass_data instance and...
564 (make_pass_cleanup_eh): ...new function.
565 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
566 to a subclass of simple_ipa_opt_pass along with...
567 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
568 (make_pass_ipa_lower_emutls): ...new function.
569 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
570 a subclass of gimple_opt_pass along with...
571 (pass_data_if_conversion): ...new pass_data instance and...
572 (make_pass_if_conversion): ...new function.
573 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
574 subclass of gimple_opt_pass along with...
575 (pass_data_build_ssa): ...new pass_data instance and...
576 (make_pass_build_ssa): ...new function.
577 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
578 global struct to a subclass of gimple_opt_pass along with...
579 (pass_data_loop_distribution): ...new pass_data instance and...
580 (make_pass_loop_distribution): ...new function.
581 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
582 subclass of gimple_opt_pass along with...
583 (pass_data_mudflap_1): ...new pass_data instance and...
584 (make_pass_mudflap_1): ...new function.
585 (pass_mudflap_2): Convert from a global struct to a subclass of
586 gimple_opt_pass along with...
587 (pass_data_mudflap_2): ...new pass_data instance and...
588 (make_pass_mudflap_2): ...new function.
589 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
590 subclass of gimple_opt_pass along with...
591 (pass_data_mudflap_1): ...new pass_data instance and...
592 (make_pass_mudflap_1): ...new function.
593 (pass_mudflap_2): Convert from a global struct to a subclass of
594 gimple_opt_pass along with...
595 (pass_data_mudflap_2): ...new pass_data instance and...
596 (make_pass_mudflap_2): ...new function.
597 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
598 gimple_opt_pass along with...
599 (pass_data_nrv): ...new pass_data instance and...
600 (make_pass_nrv): ...new function.
601 (pass_return_slot): Convert from a global struct to a subclass of
602 gimple_opt_pass along with...
603 (pass_data_return_slot): ...new pass_data instance and...
604 (make_pass_return_slot): ...new function.
605 * tree-object-size.c (pass_object_sizes): Convert from a global struct
606 to a subclass of gimple_opt_pass along with...
607 (pass_data_object_sizes): ...new pass_data instance and...
608 (make_pass_object_sizes): ...new function.
609 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
610 global struct to a subclass of gimple_opt_pass along with...
611 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
612 and...
613 (make_pass_cleanup_cfg_post_optimizing): ...new function.
614 (pass_fixup_cfg): Convert from a global struct to a subclass of
615 gimple_opt_pass along with...
616 (pass_data_fixup_cfg): ...new pass_data instance and...
617 (make_pass_fixup_cfg): ...new function.
618 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
619 (make_pass_mudflap_1): ...new function.
620 (pass_mudflap_2): Replace declaration with that of...
621 (make_pass_mudflap_2): ...new function.
622 (pass_asan): Replace declaration with that of...
623 (make_pass_asan): ...new function.
624 (pass_asan_O0): Replace declaration with that of...
625 (make_pass_asan_O0): ...new function.
626 (pass_tsan): Replace declaration with that of...
627 (make_pass_tsan): ...new function.
628 (pass_tsan_O0): Replace declaration with that of...
629 (make_pass_tsan_O0): ...new function.
630 (pass_lower_cf): Replace declaration with that of...
631 (make_pass_lower_cf): ...new function.
632 (pass_refactor_eh): Replace declaration with that of...
633 (make_pass_refactor_eh): ...new function.
634 (pass_lower_eh): Replace declaration with that of...
635 (make_pass_lower_eh): ...new function.
636 (pass_lower_eh_dispatch): Replace declaration with that of...
637 (make_pass_lower_eh_dispatch): ...new function.
638 (pass_lower_resx): Replace declaration with that of...
639 (make_pass_lower_resx): ...new function.
640 (pass_build_cfg): Replace declaration with that of...
641 (make_pass_build_cfg): ...new function.
642 (pass_early_tree_profile): Replace declaration with that of...
643 (make_pass_early_tree_profile): ...new function.
644 (pass_cleanup_eh): Replace declaration with that of...
645 (make_pass_cleanup_eh): ...new function.
646 (pass_sra): Replace declaration with that of...
647 (make_pass_sra): ...new function.
648 (pass_sra_early): Replace declaration with that of...
649 (make_pass_sra_early): ...new function.
650 (pass_early_ipa_sra): Replace declaration with that of...
651 (make_pass_early_ipa_sra): ...new function.
652 (pass_tail_recursion): Replace declaration with that of...
653 (make_pass_tail_recursion): ...new function.
654 (pass_tail_calls): Replace declaration with that of...
655 (make_pass_tail_calls): ...new function.
656 (pass_tree_loop): Replace declaration with that of...
657 (make_pass_tree_loop): ...new function.
658 (pass_tree_loop_init): Replace declaration with that of...
659 (make_pass_tree_loop_init): ...new function.
660 (pass_lim): Replace declaration with that of...
661 (make_pass_lim): ...new function.
662 (pass_tree_unswitch): Replace declaration with that of...
663 (make_pass_tree_unswitch): ...new function.
664 (pass_predcom): Replace declaration with that of...
665 (make_pass_predcom): ...new function.
666 (pass_iv_canon): Replace declaration with that of...
667 (make_pass_iv_canon): ...new function.
668 (pass_scev_cprop): Replace declaration with that of...
669 (make_pass_scev_cprop): ...new function.
670 (pass_empty_loop): Replace declaration with that of...
671 (make_pass_empty_loop): ...new function.
672 (pass_record_bounds): Replace declaration with that of...
673 (make_pass_record_bounds): ...new function.
674 (pass_graphite): Replace declaration with that of...
675 (make_pass_graphite): ...new function.
676 (pass_graphite_transforms): Replace declaration with that of...
677 (make_pass_graphite_transforms): ...new function.
678 (pass_if_conversion): Replace declaration with that of...
679 (make_pass_if_conversion): ...new function.
680 (pass_loop_distribution): Replace declaration with that of...
681 (make_pass_loop_distribution): ...new function.
682 (pass_vectorize): Replace declaration with that of...
683 (make_pass_vectorize): ...new function.
684 (pass_slp_vectorize): Replace declaration with that of...
685 (make_pass_slp_vectorize): ...new function.
686 (pass_complete_unroll): Replace declaration with that of...
687 (make_pass_complete_unroll): ...new function.
688 (pass_complete_unrolli): Replace declaration with that of...
689 (make_pass_complete_unrolli): ...new function.
690 (pass_parallelize_loops): Replace declaration with that of...
691 (make_pass_parallelize_loops): ...new function.
692 (pass_loop_prefetch): Replace declaration with that of...
693 (make_pass_loop_prefetch): ...new function.
694 (pass_iv_optimize): Replace declaration with that of...
695 (make_pass_iv_optimize): ...new function.
696 (pass_tree_loop_done): Replace declaration with that of...
697 (make_pass_tree_loop_done): ...new function.
698 (pass_ch): Replace declaration with that of...
699 (make_pass_ch): ...new function.
700 (pass_ccp): Replace declaration with that of...
701 (make_pass_ccp): ...new function.
702 (pass_phi_only_cprop): Replace declaration with that of...
703 (make_pass_phi_only_cprop): ...new function.
704 (pass_build_ssa): Replace declaration with that of...
705 (make_pass_build_ssa): ...new function.
706 (pass_build_alias): Replace declaration with that of...
707 (make_pass_build_alias): ...new function.
708 (pass_build_ealias): Replace declaration with that of...
709 (make_pass_build_ealias): ...new function.
710 (pass_dominator): Replace declaration with that of...
711 (make_pass_dominator): ...new function.
712 (pass_dce): Replace declaration with that of...
713 (make_pass_dce): ...new function.
714 (pass_dce_loop): Replace declaration with that of...
715 (make_pass_dce_loop): ...new function.
716 (pass_cd_dce): Replace declaration with that of...
717 (make_pass_cd_dce): ...new function.
718 (pass_call_cdce): Replace declaration with that of...
719 (make_pass_call_cdce): ...new function.
720 (pass_merge_phi): Replace declaration with that of...
721 (make_pass_merge_phi): ...new function.
722 (pass_split_crit_edges): Replace declaration with that of...
723 (make_pass_split_crit_edges): ...new function.
724 (pass_pre): Replace declaration with that of...
725 (make_pass_pre): ...new function.
726 (pass_profile): Replace declaration with that of...
727 (make_pass_profile): ...new function.
728 (pass_strip_predict_hints): Replace declaration with that of...
729 (make_pass_strip_predict_hints): ...new function.
730 (pass_lower_complex_O0): Replace declaration with that of...
731 (make_pass_lower_complex_O0): ...new function.
732 (pass_lower_complex): Replace declaration with that of...
733 (make_pass_lower_complex): ...new function.
734 (pass_lower_vector): Replace declaration with that of...
735 (make_pass_lower_vector): ...new function.
736 (pass_lower_vector_ssa): Replace declaration with that of...
737 (make_pass_lower_vector_ssa): ...new function.
738 (pass_lower_omp): Replace declaration with that of...
739 (make_pass_lower_omp): ...new function.
740 (pass_diagnose_omp_blocks): Replace declaration with that of...
741 (make_pass_diagnose_omp_blocks): ...new function.
742 (pass_expand_omp): Replace declaration with that of...
743 (make_pass_expand_omp): ...new function.
744 (pass_expand_omp_ssa): Replace declaration with that of...
745 (make_pass_expand_omp_ssa): ...new function.
746 (pass_object_sizes): Replace declaration with that of...
747 (make_pass_object_sizes): ...new function.
748 (pass_strlen): Replace declaration with that of...
749 (make_pass_strlen): ...new function.
750 (pass_fold_builtins): Replace declaration with that of...
751 (make_pass_fold_builtins): ...new function.
752 (pass_stdarg): Replace declaration with that of...
753 (make_pass_stdarg): ...new function.
754 (pass_early_warn_uninitialized): Replace declaration with that of...
755 (make_pass_early_warn_uninitialized): ...new function.
756 (pass_late_warn_uninitialized): Replace declaration with that of...
757 (make_pass_late_warn_uninitialized): ...new function.
758 (pass_cse_reciprocals): Replace declaration with that of...
759 (make_pass_cse_reciprocals): ...new function.
760 (pass_cse_sincos): Replace declaration with that of...
761 (make_pass_cse_sincos): ...new function.
762 (pass_optimize_bswap): Replace declaration with that of...
763 (make_pass_optimize_bswap): ...new function.
764 (pass_optimize_widening_mul): Replace declaration with that of...
765 (make_pass_optimize_widening_mul): ...new function.
766 (pass_warn_function_return): Replace declaration with that of...
767 (make_pass_warn_function_return): ...new function.
768 (pass_warn_function_noreturn): Replace declaration with that of...
769 (make_pass_warn_function_noreturn): ...new function.
770 (pass_cselim): Replace declaration with that of...
771 (make_pass_cselim): ...new function.
772 (pass_phiopt): Replace declaration with that of...
773 (make_pass_phiopt): ...new function.
774 (pass_forwprop): Replace declaration with that of...
775 (make_pass_forwprop): ...new function.
776 (pass_phiprop): Replace declaration with that of...
777 (make_pass_phiprop): ...new function.
778 (pass_tree_ifcombine): Replace declaration with that of...
779 (make_pass_tree_ifcombine): ...new function.
780 (pass_dse): Replace declaration with that of...
781 (make_pass_dse): ...new function.
782 (pass_nrv): Replace declaration with that of...
783 (make_pass_nrv): ...new function.
784 (pass_rename_ssa_copies): Replace declaration with that of...
785 (make_pass_rename_ssa_copies): ...new function.
786 (pass_sink_code): Replace declaration with that of...
787 (make_pass_sink_code): ...new function.
788 (pass_fre): Replace declaration with that of...
789 (make_pass_fre): ...new function.
790 (pass_check_data_deps): Replace declaration with that of...
791 (make_pass_check_data_deps): ...new function.
792 (pass_copy_prop): Replace declaration with that of...
793 (make_pass_copy_prop): ...new function.
794 (pass_vrp): Replace declaration with that of...
795 (make_pass_vrp): ...new function.
796 (pass_uncprop): Replace declaration with that of...
797 (make_pass_uncprop): ...new function.
798 (pass_return_slot): Replace declaration with that of...
799 (make_pass_return_slot): ...new function.
800 (pass_reassoc): Replace declaration with that of...
801 (make_pass_reassoc): ...new function.
802 (pass_rebuild_cgraph_edges): Replace declaration with that of...
803 (make_pass_rebuild_cgraph_edges): ...new function.
804 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
805 (make_pass_remove_cgraph_callee_edges): ...new function.
806 (pass_build_cgraph_edges): Replace declaration with that of...
807 (make_pass_build_cgraph_edges): ...new function.
808 (pass_local_pure_const): Replace declaration with that of...
809 (make_pass_local_pure_const): ...new function.
810 (pass_tracer): Replace declaration with that of...
811 (make_pass_tracer): ...new function.
812 (pass_warn_unused_result): Replace declaration with that of...
813 (make_pass_warn_unused_result): ...new function.
814 (pass_diagnose_tm_blocks): Replace declaration with that of...
815 (make_pass_diagnose_tm_blocks): ...new function.
816 (pass_lower_tm): Replace declaration with that of...
817 (make_pass_lower_tm): ...new function.
818 (pass_tm_init): Replace declaration with that of...
819 (make_pass_tm_init): ...new function.
820 (pass_tm_mark): Replace declaration with that of...
821 (make_pass_tm_mark): ...new function.
822 (pass_tm_memopt): Replace declaration with that of...
823 (make_pass_tm_memopt): ...new function.
824 (pass_tm_edges): Replace declaration with that of...
825 (make_pass_tm_edges): ...new function.
826 (pass_split_functions): Replace declaration with that of...
827 (make_pass_split_functions): ...new function.
828 (pass_feedback_split_functions): Replace declaration with that of...
829 (make_pass_feedback_split_functions): ...new function.
830 (pass_strength_reduction): Replace declaration with that of...
831 (make_pass_strength_reduction): ...new function.
832 (pass_ipa_lower_emutls): Replace declaration with that of...
833 (make_pass_ipa_lower_emutls): ...new function.
834 (pass_ipa_function_and_variable_visibility): Replace declaration with
835 that of...
836 (make_pass_ipa_function_and_variable_visibility): ...new function.
837 (pass_ipa_tree_profile): Replace declaration with that of...
838 (make_pass_ipa_tree_profile): ...new function.
839 (pass_early_local_passes): Replace declaration with that of...
840 (make_pass_early_local_passes): ...new function.
841 (pass_ipa_whole_program_visibility): Replace declaration with that
842 of...
843 (make_pass_ipa_whole_program_visibility): ...new function.
844 (pass_ipa_lto_gimple_out): Replace declaration with that of...
845 (make_pass_ipa_lto_gimple_out): ...new function.
846 (pass_ipa_increase_alignment): Replace declaration with that of...
847 (make_pass_ipa_increase_alignment): ...new function.
848 (pass_ipa_inline): Replace declaration with that of...
849 (make_pass_ipa_inline): ...new function.
850 (pass_ipa_free_lang_data): Replace declaration with that of...
851 (make_pass_ipa_free_lang_data): ...new function.
852 (pass_ipa_free_inline_summary): Replace declaration with that of...
853 (make_pass_ipa_free_inline_summary): ...new function.
854 (pass_ipa_cp): Replace declaration with that of...
855 (make_pass_ipa_cp): ...new function.
856 (pass_ipa_reference): Replace declaration with that of...
857 (make_pass_ipa_reference): ...new function.
858 (pass_ipa_pure_const): Replace declaration with that of...
859 (make_pass_ipa_pure_const): ...new function.
860 (pass_ipa_pta): Replace declaration with that of...
861 (make_pass_ipa_pta): ...new function.
862 (pass_ipa_lto_finish_out): Replace declaration with that of...
863 (make_pass_ipa_lto_finish_out): ...new function.
864 (pass_ipa_tm): Replace declaration with that of...
865 (make_pass_ipa_tm): ...new function.
866 (pass_ipa_profile): Replace declaration with that of...
867 (make_pass_ipa_profile): ...new function.
868 (pass_ipa_cdtor_merge): Replace declaration with that of...
869 (make_pass_ipa_cdtor_merge): ...new function.
870 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
871 of...
872 (make_pass_cleanup_cfg_post_optimizing): ...new function.
873 (pass_init_datastructures): Replace declaration with that of...
874 (make_pass_init_datastructures): ...new function.
875 (pass_fixup_cfg): Replace declaration with that of...
876 (make_pass_fixup_cfg): ...new function.
877 (pass_expand): Replace declaration with that of...
878 (make_pass_expand): ...new function.
879 (pass_instantiate_virtual_regs): Replace declaration with that of...
880 (make_pass_instantiate_virtual_regs): ...new function.
881 (pass_rtl_fwprop): Replace declaration with that of...
882 (make_pass_rtl_fwprop): ...new function.
883 (pass_rtl_fwprop_addr): Replace declaration with that of...
884 (make_pass_rtl_fwprop_addr): ...new function.
885 (pass_jump): Replace declaration with that of...
886 (make_pass_jump): ...new function.
887 (pass_jump2): Replace declaration with that of...
888 (make_pass_jump2): ...new function.
889 (pass_lower_subreg): Replace declaration with that of...
890 (make_pass_lower_subreg): ...new function.
891 (pass_cse): Replace declaration with that of...
892 (make_pass_cse): ...new function.
893 (pass_fast_rtl_dce): Replace declaration with that of...
894 (make_pass_fast_rtl_dce): ...new function.
895 (pass_ud_rtl_dce): Replace declaration with that of...
896 (make_pass_ud_rtl_dce): ...new function.
897 (pass_rtl_dce): Replace declaration with that of...
898 (make_pass_rtl_dce): ...new function.
899 (pass_rtl_dse1): Replace declaration with that of...
900 (make_pass_rtl_dse1): ...new function.
901 (pass_rtl_dse2): Replace declaration with that of...
902 (make_pass_rtl_dse2): ...new function.
903 (pass_rtl_dse3): Replace declaration with that of...
904 (make_pass_rtl_dse3): ...new function.
905 (pass_rtl_cprop): Replace declaration with that of...
906 (make_pass_rtl_cprop): ...new function.
907 (pass_rtl_pre): Replace declaration with that of...
908 (make_pass_rtl_pre): ...new function.
909 (pass_rtl_hoist): Replace declaration with that of...
910 (make_pass_rtl_hoist): ...new function.
911 (pass_rtl_store_motion): Replace declaration with that of...
912 (make_pass_rtl_store_motion): ...new function.
913 (pass_cse_after_global_opts): Replace declaration with that of...
914 (make_pass_cse_after_global_opts): ...new function.
915 (pass_rtl_ifcvt): Replace declaration with that of...
916 (make_pass_rtl_ifcvt): ...new function.
917 (pass_into_cfg_layout_mode): Replace declaration with that of...
918 (make_pass_into_cfg_layout_mode): ...new function.
919 (pass_outof_cfg_layout_mode): Replace declaration with that of...
920 (make_pass_outof_cfg_layout_mode): ...new function.
921 (pass_loop2): Replace declaration with that of...
922 (make_pass_loop2): ...new function.
923 (pass_rtl_loop_init): Replace declaration with that of...
924 (make_pass_rtl_loop_init): ...new function.
925 (pass_rtl_move_loop_invariants): Replace declaration with that of...
926 (make_pass_rtl_move_loop_invariants): ...new function.
927 (pass_rtl_unswitch): Replace declaration with that of...
928 (make_pass_rtl_unswitch): ...new function.
929 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
930 (make_pass_rtl_unroll_and_peel_loops): ...new function.
931 (pass_rtl_doloop): Replace declaration with that of...
932 (make_pass_rtl_doloop): ...new function.
933 (pass_rtl_loop_done): Replace declaration with that of...
934 (make_pass_rtl_loop_done): ...new function.
935 (pass_web): Replace declaration with that of...
936 (make_pass_web): ...new function.
937 (pass_cse2): Replace declaration with that of...
938 (make_pass_cse2): ...new function.
939 (pass_df_initialize_opt): Replace declaration with that of...
940 (make_pass_df_initialize_opt): ...new function.
941 (pass_df_initialize_no_opt): Replace declaration with that of...
942 (make_pass_df_initialize_no_opt): ...new function.
943 (pass_reginfo_init): Replace declaration with that of...
944 (make_pass_reginfo_init): ...new function.
945 (pass_inc_dec): Replace declaration with that of...
946 (make_pass_inc_dec): ...new function.
947 (pass_stack_ptr_mod): Replace declaration with that of...
948 (make_pass_stack_ptr_mod): ...new function.
949 (pass_initialize_regs): Replace declaration with that of...
950 (make_pass_initialize_regs): ...new function.
951 (pass_combine): Replace declaration with that of...
952 (make_pass_combine): ...new function.
953 (pass_if_after_combine): Replace declaration with that of...
954 (make_pass_if_after_combine): ...new function.
955 (pass_ree): Replace declaration with that of...
956 (make_pass_ree): ...new function.
957 (pass_partition_blocks): Replace declaration with that of...
958 (make_pass_partition_blocks): ...new function.
959 (pass_match_asm_constraints): Replace declaration with that of...
960 (make_pass_match_asm_constraints): ...new function.
961 (pass_regmove): Replace declaration with that of...
962 (make_pass_regmove): ...new function.
963 (pass_split_all_insns): Replace declaration with that of...
964 (make_pass_split_all_insns): ...new function.
965 (pass_fast_rtl_byte_dce): Replace declaration with that of...
966 (make_pass_fast_rtl_byte_dce): ...new function.
967 (pass_lower_subreg2): Replace declaration with that of...
968 (make_pass_lower_subreg2): ...new function.
969 (pass_mode_switching): Replace declaration with that of...
970 (make_pass_mode_switching): ...new function.
971 (pass_sms): Replace declaration with that of...
972 (make_pass_sms): ...new function.
973 (pass_sched): Replace declaration with that of...
974 (make_pass_sched): ...new function.
975 (pass_ira): Replace declaration with that of...
976 (make_pass_ira): ...new function.
977 (pass_reload): Replace declaration with that of...
978 (make_pass_reload): ...new function.
979 (pass_clean_state): Replace declaration with that of...
980 (make_pass_clean_state): ...new function.
981 (pass_branch_prob): Replace declaration with that of...
982 (make_pass_branch_prob): ...new function.
983 (pass_value_profile_transformations): Replace declaration with that
984 of...
985 (make_pass_value_profile_transformations): ...new function.
986 (pass_postreload_cse): Replace declaration with that of...
987 (make_pass_postreload_cse): ...new function.
988 (pass_gcse2): Replace declaration with that of...
989 (make_pass_gcse2): ...new function.
990 (pass_split_after_reload): Replace declaration with that of...
991 (make_pass_split_after_reload): ...new function.
992 (pass_branch_target_load_optimize1): Replace declaration with that
993 of...
994 (make_pass_branch_target_load_optimize1): ...new function.
995 (pass_thread_prologue_and_epilogue): Replace declaration with that
996 of...
997 (make_pass_thread_prologue_and_epilogue): ...new function.
998 (pass_stack_adjustments): Replace declaration with that of...
999 (make_pass_stack_adjustments): ...new function.
1000 (pass_peephole2): Replace declaration with that of...
1001 (make_pass_peephole2): ...new function.
1002 (pass_if_after_reload): Replace declaration with that of...
1003 (make_pass_if_after_reload): ...new function.
1004 (pass_regrename): Replace declaration with that of...
1005 (make_pass_regrename): ...new function.
1006 (pass_cprop_hardreg): Replace declaration with that of...
1007 (make_pass_cprop_hardreg): ...new function.
1008 (pass_reorder_blocks): Replace declaration with that of...
1009 (make_pass_reorder_blocks): ...new function.
1010 (pass_branch_target_load_optimize2): Replace declaration with that
1011 of...
1012 (make_pass_branch_target_load_optimize2): ...new function.
1013 (pass_leaf_regs): Replace declaration with that of...
1014 (make_pass_leaf_regs): ...new function.
1015 (pass_split_before_sched2): Replace declaration with that of...
1016 (make_pass_split_before_sched2): ...new function.
1017 (pass_compare_elim_after_reload): Replace declaration with that of...
1018 (make_pass_compare_elim_after_reload): ...new function.
1019 (pass_sched2): Replace declaration with that of...
1020 (make_pass_sched2): ...new function.
1021 (pass_stack_regs): Replace declaration with that of...
1022 (make_pass_stack_regs): ...new function.
1023 (pass_stack_regs_run): Replace declaration with that of...
1024 (make_pass_stack_regs_run): ...new function.
1025 (pass_df_finish): Replace declaration with that of...
1026 (make_pass_df_finish): ...new function.
1027 (pass_compute_alignments): Replace declaration with that of...
1028 (make_pass_compute_alignments): ...new function.
1029 (pass_duplicate_computed_gotos): Replace declaration with that of...
1030 (make_pass_duplicate_computed_gotos): ...new function.
1031 (pass_variable_tracking): Replace declaration with that of...
1032 (make_pass_variable_tracking): ...new function.
1033 (pass_free_cfg): Replace declaration with that of...
1034 (make_pass_free_cfg): ...new function.
1035 (pass_machine_reorg): Replace declaration with that of...
1036 (make_pass_machine_reorg): ...new function.
1037 (pass_cleanup_barriers): Replace declaration with that of...
1038 (make_pass_cleanup_barriers): ...new function.
1039 (pass_delay_slots): Replace declaration with that of...
1040 (make_pass_delay_slots): ...new function.
1041 (pass_split_for_shorten_branches): Replace declaration with that of...
1042 (make_pass_split_for_shorten_branches): ...new function.
1043 (pass_split_before_regstack): Replace declaration with that of...
1044 (make_pass_split_before_regstack): ...new function.
1045 (pass_convert_to_eh_region_ranges): Replace declaration with that
1046 of...
1047 (make_pass_convert_to_eh_region_ranges): ...new function.
1048 (pass_shorten_branches): Replace declaration with that of...
1049 (make_pass_shorten_branches): ...new function.
1050 (pass_set_nothrow_function_flags): Replace declaration with that of...
1051 (make_pass_set_nothrow_function_flags): ...new function.
1052 (pass_dwarf2_frame): Replace declaration with that of...
1053 (make_pass_dwarf2_frame): ...new function.
1054 (pass_final): Replace declaration with that of...
1055 (make_pass_final): ...new function.
1056 (pass_rtl_seqabstr): Replace declaration with that of...
1057 (make_pass_rtl_seqabstr): ...new function.
1058 (pass_release_ssa_names): Replace declaration with that of...
1059 (make_pass_release_ssa_names): ...new function.
1060 (pass_early_inline): Replace declaration with that of...
1061 (make_pass_early_inline): ...new function.
1062 (pass_inline_parameters): Replace declaration with that of...
1063 (make_pass_inline_parameters): ...new function.
1064 (pass_update_address_taken): Replace declaration with that of...
1065 (make_pass_update_address_taken): ...new function.
1066 (pass_convert_switch): Replace declaration with that of...
1067 (make_pass_convert_switch): ...new function.
1068 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1069 to a subclass of simple_ipa_opt_pass along with...
1070 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1071 (make_pass_ipa_tree_profile): ...new function.
1072 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1073 subclass of gimple_opt_pass along with...
1074 (pass_data_sra_early): ...new pass_data instance and...
1075 (make_pass_sra_early): ...new function.
1076 (pass_sra): Convert from a global struct to a subclass of
1077 gimple_opt_pass along with...
1078 (pass_data_sra): ...new pass_data instance and...
1079 (make_pass_sra): ...new function.
1080 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1081 gimple_opt_pass along with...
1082 (pass_data_early_ipa_sra): ...new pass_data instance and...
1083 (make_pass_early_ipa_sra): ...new function.
1084 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1085 subclass of gimple_opt_pass along with...
1086 (pass_data_ccp): ...new pass_data instance and...
1087 (make_pass_ccp): ...new function.
1088 (pass_fold_builtins): Convert from a global struct to a subclass of
1089 gimple_opt_pass along with...
1090 (pass_data_fold_builtins): ...new pass_data instance and...
1091 (make_pass_fold_builtins): ...new function.
1092 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1093 subclass of gimple_opt_pass along with...
1094 (pass_data_copy_prop): ...new pass_data instance and...
1095 (make_pass_copy_prop): ...new function.
1096 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1097 global struct to a subclass of gimple_opt_pass along with...
1098 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1099 (make_pass_rename_ssa_copies): ...new function.
1100 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1101 subclass of gimple_opt_pass along with...
1102 (pass_data_dce): ...new pass_data instance and...
1103 (make_pass_dce): ...new function.
1104 (pass_dce_loop): Convert from a global struct to a subclass of
1105 gimple_opt_pass along with...
1106 (pass_data_dce_loop): ...new pass_data instance and...
1107 (make_pass_dce_loop): ...new function.
1108 (pass_cd_dce): Convert from a global struct to a subclass of
1109 gimple_opt_pass along with...
1110 (pass_data_cd_dce): ...new pass_data instance and...
1111 (make_pass_cd_dce): ...new function.
1112 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1113 subclass of gimple_opt_pass along with...
1114 (pass_data_dominator): ...new pass_data instance and...
1115 (make_pass_dominator): ...new function.
1116 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1117 gimple_opt_pass along with...
1118 (pass_data_phi_only_cprop): ...new pass_data instance and...
1119 (make_pass_phi_only_cprop): ...new function.
1120 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1121 subclass of gimple_opt_pass along with...
1122 (pass_data_dse): ...new pass_data instance and...
1123 (make_pass_dse): ...new function.
1124 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1125 a subclass of gimple_opt_pass along with...
1126 (pass_data_forwprop): ...new pass_data instance and...
1127 (make_pass_forwprop): ...new function.
1128 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1129 struct to a subclass of gimple_opt_pass along with...
1130 (pass_data_tree_ifcombine): ...new pass_data instance and...
1131 (make_pass_tree_ifcombine): ...new function.
1132 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1133 subclass of gimple_opt_pass along with...
1134 (pass_data_ch): ...new pass_data instance and...
1135 (make_pass_ch): ...new function.
1136 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1137 subclass of gimple_opt_pass along with...
1138 (pass_data_tree_loop): ...new pass_data instance and...
1139 (make_pass_tree_loop): ...new function.
1140 (pass_tree_loop_init): Convert from a global struct to a subclass of
1141 gimple_opt_pass along with...
1142 (pass_data_tree_loop_init): ...new pass_data instance and...
1143 (make_pass_tree_loop_init): ...new function.
1144 (pass_lim): Convert from a global struct to a subclass of
1145 gimple_opt_pass along with...
1146 (pass_data_lim): ...new pass_data instance and...
1147 (make_pass_lim): ...new function.
1148 (pass_tree_unswitch): Convert from a global struct to a subclass of
1149 gimple_opt_pass along with...
1150 (pass_data_tree_unswitch): ...new pass_data instance and...
1151 (make_pass_tree_unswitch): ...new function.
1152 (pass_predcom): Convert from a global struct to a subclass of
1153 gimple_opt_pass along with...
1154 (pass_data_predcom): ...new pass_data instance and...
1155 (make_pass_predcom): ...new function.
1156 (pass_vectorize): Convert from a global struct to a subclass of
1157 gimple_opt_pass along with...
1158 (pass_data_vectorize): ...new pass_data instance and...
1159 (make_pass_vectorize): ...new function.
1160 (pass_graphite): Convert from a global struct to a subclass of
1161 gimple_opt_pass along with...
1162 (pass_data_graphite): ...new pass_data instance and...
1163 (make_pass_graphite): ...new function.
1164 (pass_graphite_transforms): Convert from a global struct to a subclass
1165 of gimple_opt_pass along with...
1166 (pass_data_graphite_transforms): ...new pass_data instance and...
1167 (make_pass_graphite_transforms): ...new function.
1168 (pass_check_data_deps): Convert from a global struct to a subclass of
1169 gimple_opt_pass along with...
1170 (pass_data_check_data_deps): ...new pass_data instance and...
1171 (make_pass_check_data_deps): ...new function.
1172 (pass_iv_canon): Convert from a global struct to a subclass of
1173 gimple_opt_pass along with...
1174 (pass_data_iv_canon): ...new pass_data instance and...
1175 (make_pass_iv_canon): ...new function.
1176 (pass_scev_cprop): Convert from a global struct to a subclass of
1177 gimple_opt_pass along with...
1178 (pass_data_scev_cprop): ...new pass_data instance and...
1179 (make_pass_scev_cprop): ...new function.
1180 (pass_record_bounds): Convert from a global struct to a subclass of
1181 gimple_opt_pass along with...
1182 (pass_data_record_bounds): ...new pass_data instance and...
1183 (make_pass_record_bounds): ...new function.
1184 (pass_complete_unroll): Convert from a global struct to a subclass of
1185 gimple_opt_pass along with...
1186 (pass_data_complete_unroll): ...new pass_data instance and...
1187 (make_pass_complete_unroll): ...new function.
1188 (pass_complete_unrolli): Convert from a global struct to a subclass of
1189 gimple_opt_pass along with...
1190 (pass_data_complete_unrolli): ...new pass_data instance and...
1191 (make_pass_complete_unrolli): ...new function.
1192 (pass_parallelize_loops): Convert from a global struct to a subclass
1193 of gimple_opt_pass along with...
1194 (pass_data_parallelize_loops): ...new pass_data instance and...
1195 (make_pass_parallelize_loops): ...new function.
1196 (pass_loop_prefetch): Convert from a global struct to a subclass of
1197 gimple_opt_pass along with...
1198 (pass_data_loop_prefetch): ...new pass_data instance and...
1199 (make_pass_loop_prefetch): ...new function.
1200 (pass_iv_optimize): Convert from a global struct to a subclass of
1201 gimple_opt_pass along with...
1202 (pass_data_iv_optimize): ...new pass_data instance and...
1203 (make_pass_iv_optimize): ...new function.
1204 (pass_tree_loop_done): Convert from a global struct to a subclass of
1205 gimple_opt_pass along with...
1206 (pass_data_tree_loop_done): ...new pass_data instance and...
1207 (make_pass_tree_loop_done): ...new function.
1208 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1209 struct to a subclass of gimple_opt_pass along with...
1210 (pass_data_cse_reciprocals): ...new pass_data instance and...
1211 (make_pass_cse_reciprocals): ...new function.
1212 (pass_cse_sincos): Convert from a global struct to a subclass of
1213 gimple_opt_pass along with...
1214 (pass_data_cse_sincos): ...new pass_data instance and...
1215 (make_pass_cse_sincos): ...new function.
1216 (pass_optimize_bswap): Convert from a global struct to a subclass of
1217 gimple_opt_pass along with...
1218 (pass_data_optimize_bswap): ...new pass_data instance and...
1219 (make_pass_optimize_bswap): ...new function.
1220 (pass_optimize_widening_mul): Convert from a global struct to a
1221 subclass of gimple_opt_pass along with...
1222 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1223 (make_pass_optimize_widening_mul): ...new function.
1224 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1225 subclass of gimple_opt_pass along with...
1226 (pass_data_phiopt): ...new pass_data instance and...
1227 (make_pass_phiopt): ...new function.
1228 (pass_cselim): Convert from a global struct to a subclass of
1229 gimple_opt_pass along with...
1230 (pass_data_cselim): ...new pass_data instance and...
1231 (make_pass_cselim): ...new function.
1232 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1233 subclass of gimple_opt_pass along with...
1234 (pass_data_phiprop): ...new pass_data instance and...
1235 (make_pass_phiprop): ...new function.
1236 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1237 subclass of gimple_opt_pass along with...
1238 (pass_data_pre): ...new pass_data instance and...
1239 (make_pass_pre): ...new function.
1240 (pass_fre): Convert from a global struct to a subclass of
1241 gimple_opt_pass along with...
1242 (pass_data_fre): ...new pass_data instance and...
1243 (make_pass_fre): ...new function.
1244 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1245 subclass of gimple_opt_pass along with...
1246 (pass_data_reassoc): ...new pass_data instance and...
1247 (make_pass_reassoc): ...new function.
1248 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1249 subclass of gimple_opt_pass along with...
1250 (pass_data_sink_code): ...new pass_data instance and...
1251 (make_pass_sink_code): ...new function.
1252 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1253 subclass of gimple_opt_pass along with...
1254 (pass_data_strlen): ...new pass_data instance and...
1255 (make_pass_strlen): ...new function.
1256 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1257 struct to a subclass of gimple_opt_pass along with...
1258 (pass_data_build_alias): ...new pass_data instance and...
1259 (make_pass_build_alias): ...new function.
1260 (pass_build_ealias): Convert from a global struct to a subclass of
1261 gimple_opt_pass along with...
1262 (pass_data_build_ealias): ...new pass_data instance and...
1263 (make_pass_build_ealias): ...new function.
1264 (pass_ipa_pta): Convert from a global struct to a subclass of
1265 simple_ipa_opt_pass along with...
1266 (pass_data_ipa_pta): ...new pass_data instance and...
1267 (make_pass_ipa_pta): ...new function.
1268 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1269 subclass of gimple_opt_pass along with...
1270 (pass_data_uncprop): ...new pass_data instance and...
1271 (make_pass_uncprop): ...new function.
1272 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1273 global struct to a subclass of gimple_opt_pass along with...
1274 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1275 (make_pass_late_warn_uninitialized): ...new function.
1276 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1277 to a subclass of gimple_opt_pass along with...
1278 (pass_data_init_datastructures): ...new pass_data instance and...
1279 (make_pass_init_datastructures): ...new function.
1280 (pass_early_warn_uninitialized): Convert from a global struct to a
1281 subclass of gimple_opt_pass along with...
1282 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1283 (make_pass_early_warn_uninitialized): ...new function.
1284 (pass_update_address_taken): Convert from a global struct to a
1285 subclass of gimple_opt_pass along with...
1286 (pass_data_update_address_taken): ...new pass_data instance and...
1287 (make_pass_update_address_taken): ...new function.
1288 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1289 struct to a subclass of gimple_opt_pass along with...
1290 (pass_data_release_ssa_names): ...new pass_data instance and...
1291 (make_pass_release_ssa_names): ...new function.
1292 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1293 subclass of gimple_opt_pass along with...
1294 (pass_data_stdarg): ...new pass_data instance and...
1295 (make_pass_stdarg): ...new function.
1296 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1297 global struct to a subclass of gimple_opt_pass along with...
1298 (pass_data_convert_switch): ...new pass_data instance and...
1299 (make_pass_convert_switch): ...new function.
1300 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1301 to a subclass of gimple_opt_pass along with...
1302 (pass_data_tail_recursion): ...new pass_data instance and...
1303 (make_pass_tail_recursion): ...new function.
1304 (pass_tail_calls): Convert from a global struct to a subclass of
1305 gimple_opt_pass along with...
1306 (pass_data_tail_calls): ...new pass_data instance and...
1307 (make_pass_tail_calls): ...new function.
1308 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1309 struct to a subclass of gimple_opt_pass along with...
1310 (pass_data_lower_vector): ...new pass_data instance and...
1311 (make_pass_lower_vector): ...new function.
1312 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1313 gimple_opt_pass along with...
1314 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1315 (make_pass_lower_vector_ssa): ...new function.
1316 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
1317 to a subclass of gimple_opt_pass along with...
1318 (pass_data_slp_vectorize): ...new pass_data instance and...
1319 (make_pass_slp_vectorize): ...new function.
1320 (pass_ipa_increase_alignment): Convert from a global struct to a
1321 subclass of simple_ipa_opt_pass along with...
1322 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
1323 (make_pass_ipa_increase_alignment): ...new function.
1324 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
1325 gimple_opt_pass along with...
1326 (pass_data_vrp): ...new pass_data instance and...
1327 (make_pass_vrp): ...new function.
1328 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
1329 subclass of simple_ipa_opt_pass along with...
1330 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
1331 (make_pass_ipa_free_lang_data): ...new function.
1332 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
1333 gimple_opt_pass along with...
1334 (pass_data_tsan): ...new pass_data instance and...
1335 (make_pass_tsan): ...new function.
1336 (pass_tsan_O0): Convert from a global struct to a subclass of
1337 gimple_opt_pass along with...
1338 (pass_data_tsan_O0): ...new pass_data instance and...
1339 (make_pass_tsan_O0): ...new function.
1340 * var-tracking.c (pass_variable_tracking): Convert from a global
1341 struct to a subclass of rtl_opt_pass along with...
1342 (pass_data_variable_tracking): ...new pass_data instance and...
1343 (make_pass_variable_tracking): ...new function.
1344 * web.c (pass_web): Convert from a global struct to a subclass of
1345 rtl_opt_pass along with...
1346 (pass_data_web): ...new pass_data instance and...
1347 (make_pass_web): ...new function.
1348 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
1349 declaration with that of...
1350 (make_pass_mode_switch_use): ...new function.
1351 (pass_resolve_sw_modes): Replace declaration with that of...
1352 (make_pass_resolve_sw_modes): ...new function.
1353 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
1354 from a global struct to a subclass of rtl_opt_pass along with...
1355 (pass_data_mode_switch_use): ...new pass_data instance and...
1356 (make_pass_mode_switch_use): ...new function.
1357 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
1358 from a global struct to a subclass of rtl_opt_pass along with...
1359 (pass_data_resolve_sw_modes): ...new pass_data instance and...
1360 (make_pass_resolve_sw_modes): ...new function.
1361 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
1362 struct to a subclass of rtl_opt_pass along with...
1363 (pass_data_insert_vzeroupper): ...new pass_data instance and...
1364 (make_pass_insert_vzeroupper): ...new function.
1365 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
1366 global struct to a subclass of rtl_opt_pass along with...
1367 (pass_data_work_around_errata): ...new pass_data instance and...
1368 (make_pass_work_around_errata): ...new function.
1369 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
1370 struct to a subclass of rtl_opt_pass along with...
1371 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
1372 (make_pass_mips_machine_reorg2): ...new function.
1373
6a389ed5
DM
13742013-08-05 David Malcolm <dmalcolm@redhat.com>
1375
1376 * passes.c (pass_manager::operator new): New.
1377
f7695dbf
DM
13782013-08-05 David Malcolm <dmalcolm@redhat.com>
1379
1380 Handwritten part of conversion of passes to C++ classes.
1381
1382 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
1383 (toplev.o): Add dep on PASS_MANAGER_H.
1384 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
1385 of early local pases to reflect this moving from a global to a
1386 member of gcc::pass_manager.
1387 (cgraph_add_new_function): Likewise.
1388 * lto-cgraph.c (lto_output_node): Update for conversion of
1389 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
1390 * passes.c (opt_pass::clone): New.
1391 (opt_pass::gate): New.
1392 (opt_pass::execute): New.
1393 (opt_pass::opt_pass): New.
1394 (pass_manager::execute_early_local_passes): New.
1395 (pass_manager::execute_pass_mode_switching): new.
1396 (finish_optimization_passes): Convert to...
1397 (pass_manager::finish_optimization_passes): ...this.
1398 (finish_optimization_passes): Update for conversion of passes to
1399 C++ classes.
1400 (register_dump_files_1): Use has_gate since we cannot portably
1401 check a vtable entry against NULL.
1402 (dump_one_pass): Likewise.
1403 (ipa_write_summaries_2): Likewise.
1404 (ipa_write_optimization_summaries_1): Likewise.
1405 (ipa_read_summaries_1): Likewise.
1406 (ipa_read_optimization_summaries_1): Likewise.
1407 (execute_ipa_stmt_fixups): Likewise.
1408 (pass_manager::pass_manager): Rewrite pass-creation, invoking
1409 pass-creation functions rather than wiring up globals, and
1410 storing the results in fields of pass_manager generated using
1411 pass-instances.def.
1412 (pass_manager::dump_profile_report): Update for conversion of
1413 passes to C++ classes.
1414 (pass_manager::execute_ipa_summary_passes): Likewise.
1415 (execute_one_ipa_transform_pass): Likewise.
1416 (execute_one_pass): Use has_gate and has_execute since we cannot
1417 portably check a vtable entry against NULL.
1418 * pass_manager.h (pass_manager::finish_optimization_passes): New.
1419 (pass_manager): Use pass-instances.def to add fields for the
1420 various pass instances.
1421 * toplev.c (finalize): Update for move of
1422 finish_optimization_passes to a method of gcc::pass_manager.
1423 * toplev.h (finish_optimization_passes): Move to method of class
1424 pass_manager.
1425 * tree-pass.h (struct pass_data): New.
1426 (opt_pass): Convert to C++ class, make it a subclass of
1427 pass_data.
1428 (opt_pass::gate): Convert to virtual function.
1429 (opt_pass::~opt_pass): New.
1430 (opt_pass::clone): New.
1431 (opt_pass::execute): Convert to virtual function.
1432 (opt_pass::opt_pass): New.
1433 (opt_pass::ctxt_): new.
1434 (gimple_opt_pass): Convert to subclass of opt_pass.
1435 (gimple_opt_pass::gimple_opt_pass): New.
1436 (rtl_opt_pass): Convert to subclass of opt_pass.
1437 (rtl_opt_pass::rtl_opt_pass): New.
1438 (ipa_opt_pass_d): Convert to subclass of opt_pass.
1439 (ipa_opt_pass_d::ipa_opt_pass_d): New.
1440 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
1441 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
1442 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
1443 invocation of pass_mode_switching to reflect this moving from a
1444 global to a member of gcc::pass_manager.
1445 (ix86_option_override): Rework how pass_insert_vzeroupper is
1446 added to the pass_manager to reflect autogenerated changes.
1447 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and
1448 PASS_MANAGER_H.
1449
d0b2f831
RE
14502013-08-05 Richard Earnshaw <rearnsha@arm.com>
1451
1452 PR rtl-optimization/57708
1453 * recog.c (peep2_find_free_register): Validate all regs in a
1454 multi-reg mode.
1455
51a5c0c2
JH
14562013-08-05 Jan Hubicka <jh@suse.cz>
1457
1458 PR lto/57602
1459 * cgraph.c (verify_cgraph_node): Accept local flags from other partitions.
1460 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
1461 (function_and_variable_visibility): Likewise.
1462 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
1463
65f0a120
GDR
14642013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1465
1466 * graph.c (init_graph_slim_pretty_print): Remove.
1467 (print_graph_cfg): Do not call it. Use local pretty printer.
1468 (start_graph_dump): Likewise.
1469
b3f80694
GDR
14702013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1471
1472 * gimple-pretty-print.c (buffer): Remove.
1473 (initialized): Likewise.
1474 (maybe_init_pretty_print): Likewise.
1475 (print_gimple_stmt): Do not call it. Use non-static local
1476 pretty_printer variable.
1477 (print_gimple_expr): Likewise.
1478 (print_gimple_seq): Likewise.
1479 (gimple_dump_bb): Likewise.
1480
11a877b3
GDR
14812013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1482
1483 * asan.c (asan_pp): Remove.
1484 (asan_pp_initialized): Likewise.
1485 (asan_pp_initialize): Likewise.
1486 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
1487 (asan_emit_stack_protection): Tidy. Use local pretty printer.
1488 (asan_add_global): Likewise.
1489
b066401f
GDR
14902013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
1491
1492 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
1493 * diagnostic.h (diagnostic_flush_buffer): Adjust.
1494 * pretty-print.c (pp_formatted_text_data): Likewise.
1495 (pp_indent): Rename from pp_base_indent.
1496 (pp_format): Rename from pp_base_format.
1497 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
1498 (pp_format_verbatim): Rename from pp_base_format_verbatim.
1499 (pp_flush): Rename from pp_base_flush.
1500 (pp_set_line_maximum_length): Rename from
1501 pp_base_set_line_maximum_length.
1502 (pp_clear_output_area): Rename from pp_base_clear_output_area.
1503 (pp_set_prefix): Rename from pp_base_set_prefix.
1504 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
1505 (pp_emit_prefix): Rename from pp_base_emit_prefix.
1506 (pp_append_text): Rename from pp_base_append_text.
1507 (pp_formatted_text): Rename from pp_base_formatted_text.
1508 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
1509 (pp_remaining_character_count_for_line): Rename from
1510 pp_base_remaining_character_count_for_line.
1511 (pp_newline): Rename from pp_base_newline.
1512 (pp_character): Rename from pp_base_character.
1513 (pp_string): Rename from pp_base_string.
1514 (pp_maybe_space): Rename from pp_base_maybe_space.
1515 * asan.c (asan_pp_string): Adjust.
1516 (asan_emit_stack_protection): Likewise.
1517 (asan_add_global): Likewise.
1518 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
1519 * tree-mudflap.c (mf_varname_tree): Likewise.
1520 * tree-pretty-print.c (pp_tree_identifier): Rename from
1521 pp_base_tree_identifier.
1522 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
1523 Declare as function.
1524
137a1a27
GDR
15252013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1526
1527 * pretty-print.h (pp_bar_bar): New.
1528 (pp_ampersand_ampersand): Likewise.
1529 (pp_less_equal): Likewise.
1530 (pp_greater_equal): Likewise.
1531 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
1532 printer functions instead of pp_string or operators and punctuators.
1533 (dump_gimple_call): Likewise.
1534 (dump_gimple_omp_for): Likewise.
1535 (dump_gimple_transaction): Likewise.
1536 (dump_gimple_phi): Likewise.
1537 (pp_gimple_stmt_1): Likewise.
1538 * sched-vis.c (print_insn): Likewise.
1539 * tree-mudflap.c (mf_varname_tree): Likewise.
1540 * tree-pretty-print.c (dump_block_node): Likewise.
1541 (dump_generic_node): Likewise.
1542
815effe1
JH
15432013-08-02 Jan Hubicka <jh@suse.cz>
1544
1545 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
1546 boundaries.
1547 * lto-streamer-out.c (tree_is_indexable): Results decls and
1548 parm decls are not indexable.
1549 (DFS_write_tree_body): Do not follow args and results.
1550 (hash_tree): Likewise.
1551 (output_functions): Rearrange so struct function is needed
1552 only when real body is output; be able to also ouptut abstract
1553 functions; output DECL_ARGUMENTS and DECL_RESULT.
1554 (lto_output): When not in WPA, ale store abstract functions.
1555 (write_symbol): Do not care about RESULT_DECL.
1556 (output_symbol_p): Handle correctly sbtract decls.
1557 * lto-streamer-in.c (input_function): Rearrange so struct
1558 function can be NULL at entry; allow streaming of
1559 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
1560 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
1561 sanity check during LTO.
1562 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
1563 RESULT_DECl and DECL_ARGUMENTS.
1564 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
1565 Likewise.
1566
07838b13
GDR
15672013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1568
1569 * pretty-print.h (pp_underscore): New.
1570 (pp_comma): Tidy.
1571 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
1572 printer functions instead of pp_character.
1573 (dump_binary_rhs): Likewise.
1574 (dump_ternary_rhs): Likewise.
1575 (dump_gimple_call_args): Likewise.
1576 (pp_points_to_solution): Likewise.
1577 (dump_gimple_call): Likewise.
1578 (dump_gimple_switch): Likewise.
1579 (dump_gimple_cond): Likewise.
1580 (dump_gimple_bind): Likewise.
1581 (dump_gimple_try): Likewise.
1582 (dump_gimple_omp_for): Likewise.
1583 (dump_gimple_omp_continue): Likewise.
1584 (dump_gimple_omp_single): Likewise.
1585 (dump_gimple_omp_sections): Likewise.
1586 (dump_gimple_omp_block): Likewise.
1587 (dump_gimple_omp_critical): Likewise.
1588 (dump_gimple_transaction): Likewise.
1589 (dump_gimple_asm): Likewise.
1590 (dump_gimple_phi): Likewise.
1591 (dump_gimple_omp_parallel): Likewise.
1592 (dump_gimple_omp_task): Likewise.
1593 (dump_gimple_omp_atomic_load): Likewise.
1594 (dump_gimple_omp_atomic_store): Likewise.
1595 (dump_gimple_mem_ops): Likewise.
1596 (pp_gimple_stmt_1): Likewise.
1597 (pp_cfg_jump): Likewise.
1598 (dump_implicit_edges): Likewise.
1599 (gimple_dump_bb_for_graph): Likewise.
1600 * graph.c (draw_cfg_node): Likewise.
1601 * langhooks.c (lhd_print_error_function): Likewise.
1602 * sched-vis.c (print_exp): Likewise.
1603 (print_value): Likewise.
1604 (print_pattern): Likewise.
1605 (print_insn): Likewise.
1606 (rtl_dump_bb_for_graph): Likewise.
1607 * tree-pretty-print.c (dump_function_declaration): Likewise.
1608 (dump_array_domain): Likewise.
1609 (dump_omp_clause): Likewise.
1610 (dump_location): Likewise.
1611 (dump_generic_node): Likewise.
1612 (print_struct_decl): Likewise.
1613 * diagnostic.c (diagnostic_show_locus): Use pp_space.
1614
bb0d2039
BS
16152013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1616
1617 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
1618 candidate table when replacing a candidate statement.
1619 (replace_rhs_if_not_dup): Likewise.
1620 (replace_one_candidate): Likewise.
1621
bc0ec027 16222013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 1623 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH
1624
1625 * cgraphunit.c (add_new_function): Fix logic when adding from
1626 late IPA pass.
1627 (assemble_thunk): Rename to ...
1628 (expand_thunk); .. this one; export; get it working with general functions;
1629 make produced gimple valid.
1630 * cgraph.h (expand_thunk): Declare.
1631
0e8853ee
JH
16322013-08-02 Jan Hubicka <jh@suse.cz>
1633
1634 * ipa-cp.c (gather_context_independent_values): Use ipa_get_param_move_cost.
1635 (get_replacement_map): Remove PARAM; move parameter folding into tree-inline.c
1636 (create_specialized_node): Update.
1637 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
1638 assert that we have gimple body; update move_cost.
1639 (count_formal_params): Assert that we have gimple body.
1640 (ipa_dump_param): New function.
1641 (ipa_alloc_node_params): Break out from ...
1642 (ipa_initialize_node_params): ... here.
1643 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
1644 (ipa_write_node_info): Stream move costs.
1645 (ipa_read_node_info): Read move costs.
1646 (ipa_update_after_lto_read): Do not recompute node params.
1647 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
1648 (ipa_get_param): Check we are not in WPA.
1649 (ipa_get_param_move_cost): New.
1650 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
1651 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
1652 parm numbers to be present.
1653
ecee672b
VM
16542013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1655
1656 PR rtl-optimization/58048
1657 * lra-constraints.c (process_alt_operands): Don't check asm
1658 operand on register.
1659
13ec0527
EB
16602013-08-02 Eric Botcazou <ebotcazou@adacore.com>
1661
1662 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
1663 the implied StoreLoad barrier for atomic operations if before.
1664
af15184a
JH
16652013-08-02 Jan Hubicka <jh@suse.cz>
1666 Martin Liska <marxin.liska@gmail.com>
1667
1668 * cgraph.c (cgraph_function_body_availability): Do not check cgrpah flags.
1669 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
1670 symtab_node_availability): Declare.
1671 * ipa.c (can_replace_by_local_alias): New.
1672 (function_and_variable_visibility): Use it.
1673 * symtab.c (symtab_for_node_and_aliases, symtab_nonoverwritable_alias_1,
1674 symtab_nonoverwritable_alias): New.
1675
01e54ef8
VM
16762013-08-02 Vladimir Makarov <vmakarov@redhat.com>
1677
1678 PR rtl-optimization/57963
1679 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p):
1680 New.
1681 (lra_constraints): Use them.
1682
e2fd7ca7
SN
16832013-08-02 Sofiane Naci <sofiane.naci@arm.com>
1684
1685 * config/arm/types.md (define_attr "type"): Add "load_acq" and "store_rel".
1686 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
1687 changes.
1688 (cortex_a53_store1): Likewise.
1689
0dcc26c3
JH
16902013-08-01 Jan Hubicka <jh@suse.cz>
1691
1692 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other partitions are
1693 not needed.
1694
0bd72901
UB
16952013-08-01 Uros Bizjak <ubizjak@gmail.com>
1696
1697 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
1698 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
1699 MAYBE_NON_Q_CLASS_P where appropriate.
1700
12123452
JH
17012013-08-01 Jan Hubicka <jh@suse.cz>
1702
1703 * cgraph.h (release_function_body): Declare.
1704 * tree.c (free_lang_data_in_decl): Free, parameters and return values
1705 of unused delcarations.
1706
ce852f9c
KT
17072013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1708
1709 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical RTL form
1710 when subtracting a constant.
1711
c743b246
KT
17122013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1713
1714 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
1715 Generate canonical plus rtx with negated immediate instead of minus
1716 where appropriate.
1717 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
1718
c0c123ef
JH
17192013-08-01 Jan Hubicka <jh@suse.cz>
1720
1721 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
1722 (cgraph_release_function_body): Likewise.
1723 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1724 * cgraph.h (cgrpah_node): Rename abstract_and_needed
1725 to used_as_abstract_origin.
1726 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
1727 symbols used as abstract origins.
1728 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB
1729 * ipa.c (symtab_remove_unreachable_nodes): Recompute
1730 used_as_abstract_origin.
c0c123ef 1731 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB
1732 used_as_abstract_origin; be ready for DECL_RESULT and
1733 DECL_ARGUMENTS to be NULL.
c0c123ef 1734
0bd72901
UB
1735 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
1736 for abstract functions.
1737 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
1738 real symbols.
c0c123ef 1739
1f029433
JH
17402013-08-01 Jan Hubicka <jh@suse.cz>
1741
1742 * profile.c (compute_value_histograms): Fix thinko.
1743
c451f4d6
SN
17442013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1745
1746 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
1747 aarch-common-protos.h to extra_headers.
1748 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1749 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
1750 * config/aarch64/t-aarch64 (aarch-common.o): Define.
1751
34b512f5
SN
17522013-08-01 Sofiane Naci <sofiane.naci@arm.com>
1753
1754 * config/aarch64/aarch64.md (define_attr "type"): Delete.
1755 Include "../arm/types.md". Define "type" attribute for all patterns.
1756 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
1757 attribute changes.
1758
d86e633a
MM
17592013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
1760
1761 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
1762 to support power8 load fusion.
1763 (fusion_gpr_mem_load): Likewise.
1764
1765 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
1766
1767 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
1768 declarations for power8 load fusion.
1769 (emit_fusion_gpr_load): Likewise.
1770
1771 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
1772 tuning for power8, turn on fusion mode by default. Turn on sign
1773 extending fusion mode if normal fusion mode is on, and we are at
1774 -O2 or -O3.
1775 (fusion_gpr_load_p): New function, return true if we can fuse an
1776 addis instruction with a dependent load to a GPR.
1777 (emit_fusion_gpr_load): Emit the instructions for power8 load
1778 fusion to GPRs.
1779
0bd72901 1780 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM
1781 (VSX load fusion peepholes): New peepholes to fuse together an
1782 addi instruction with a VSX load instruction.
1783
1784 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
1785 peepholes to fuse an addis instruction with a load to a GPR base
1786 register. If we are supporting sign extending fusions, convert
1787 sign extending loads to zero extending loads and add an explicit
1788 sign extension.
1789
c3f35647
SN
17902013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1791
1792 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
1793 aarch-common-protos.h to extra_headers.
1794 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
1795 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
1796 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
1797 (arm_no_early_alu_shift_dep): Likewise.
1798 (arm_no_early_alu_shift_value_dep): Likewise.
1799 (arm_no_early_mul_dep): Likewise.
1800 (arm_no_early_store_addr_dep): Likewise.
1801 (arm_mac_accumulator_is_mul_result): Likewise.
1802 (arm_mac_accumulator_is_result): Likewise.
c3f35647 1803 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB
1804 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
1805 here to ...
c3f35647 1806 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
1807 (arm_no_early_alu_shift_dep): Likewise.
1808 (arm_no_early_alu_shift_value_dep): Likewise.
1809 (arm_no_early_mul_dep): Likewise.
1810 (arm_no_early_store_addr_dep): Likewise.
1811 (arm_mac_accumulator_is_mul_result): Likewise.
1812 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN
1813 * config/arm/aarch-common-protos.h: ... here. New file.
1814 * config/arm/t-arm (aarch-common.o): Define.
1815
18162013-07-31 Sofiane Naci <sofiane.naci@arm.com>
1817
1818 * config/arm/arm.md: Include new file "types.md".
1819 (define_attr "type"): Move from here to ...
1820 (define_attr "mul32"): Likewise.
1821 (define_attr "mul64"): Likewise.
1822 * config/arm/types.md: ... here. New file.
1823
228c1313
SH
18242013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
1825
1826 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
1827 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
1828
d6e9046d 18292013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG
1830
1831 * gen-pass-instances.awk: Fix offset of substr().
1832
a167b052
DM
18332013-07-31 David Malcolm <dmalcolm@redhat.com>
1834
1835 * Makefile.in (pass-instances.def): New.
1836 (passes.o): Replace dependency on passes.def with one on
1837 pass-instances.def
1838
1839 * gen-pass-instances.awk: New.
1840
1841 * passes.c (pass_manager::pass_manager): Use pass-instances.def
1842 rather than passes.def, updating local definition of NEXT_PASS
1843 macro to add an extra NUM parameter (currently unused).
1844
315f8c0e
DM
18452013-07-30 David Malcolm <dmalcolm@redhat.com>
1846
1847 * Makefile.in (PASS_MANAGER_H): New.
1848 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
1849 (passes.o): Likewise.
1850 (statistics.o): Likewise.
1851 (cgraphunit.o): Likewise.
1852 (context.o): Depend on PASS_MANAGER_H.
1853
1854 * pass_manager.h: New.
1855
1856 * cgraphunit.c (cgraph_add_new_function): Update for moves
1857 of globals to fields of pass_manager.
1858 (analyze_function): Likewise.
1859 (expand_function): Likewise.
1860 (ipa_passes): Likewise.
1861 (compile): Likewise.
1862
1863 * context.c (context::context): New.
1864 * context.h (context::context): New.
1865 (context::get_passes): New.
1866 (context::passes_): New.
1867
1868 * lto-cgraph.c (input_node): Update for moves of globals to
1869 fields of pass_manager.
1870
1871 * passes.c (all_passes): Remove, in favor of a field of the
1872 same name within the new class pass_manager.
1873 (all_small_ipa_passes): Likewise.
1874 (all_lowering_passes): Likewise.
1875 (all_regular_ipa_passes): Likewise.
1876 (all_late_ipa_passes): Likewise.
1877 (all_lto_gen_passes): Likewise.
1878 (passes_by_id): Likewise.
1879 (passes_by_id_size): Likewise.
1880 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
1881 the new class pass_manager.
1882 (set_pass_for_id): Convert to...
1883 (pass_manager::set_pass_for_id): ...method.
1884 (get_pass_for_id): Convert to...
1885 (pass_manager::get_pass_for_id): ...method.
1886 (register_one_dump_file): Move body of implementation into...
1887 (pass_manager::register_one_dump_file): ...here.
1888 (register_dump_files_1): Convert to...
1889 (pass_manager::register_dump_files_1): ...method.
1890 (register_dump_files): Convert to...
1891 (pass_manager::register_dump_files): ...method.
1892 (create_pass_tab): Update for moves of globals to fields of
1893 pass_manager.
1894 (dump_passes): Move body of implementation into...
1895 (pass_manager::dump_passes): ...here.
1896 (register_pass): Move body of implementation into...
1897 (pass_manager::register_pass): ...here.
1898 (init_optimization_passes): Convert into...
1899 (pass_manager::pass_manager): ...constructor for new
1900 pass_manager class, and initialize the pass_lists array.
1901 (check_profile_consistency): Update for moves of globals to
1902 fields of pass_manager.
1903 (dump_profile_report): Move body of implementation into...
1904 (pass_manager::dump_profile_report): ...here.
1905 (ipa_write_summaries_1): Update for moves of pass lists from
1906 being globals to fields of pass_manager.
1907 (ipa_write_optimization_summaries): Likewise.
1908 (ipa_read_summaries): Likewise.
1909 (ipa_read_optimization_summaries): Likewise.
1910 (execute_all_ipa_stmt_fixups): Likewise.
1911
1912 * statistics.c (statistics_fini): Update for moves of globals to
1913 fields of pass_manager.
1914
1915 * toplev.c (general_init): Replace call to
1916 init_optimization_passes with construction of the pass_manager
1917 instance.
1918
1919 * tree-pass.h (all_passes): Remove, in favor of a field of the
1920 same name within the new class pass_manager.
1921 (all_small_ipa_passes): Likewise.
1922 (all_lowering_passes): Likewise.
1923 (all_regular_ipa_passes): Likewise.
1924 (all_lto_gen_passes): Likewise.
1925 (all_late_ipa_passes): Likewise.
1926 (passes_by_id): Likewise.
1927 (passes_by_id_size): Likewise.
1928 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
1929 the new class pass_manager.
1930 (get_pass_for_id): Remove.
1931
02cab1c1
RE
19322013-07-30 Richard Earnshaw <rearnsha@arm.com>
1933
1934 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
1935 configs.
1936
b2d3f886
RE
19372013-07-30 Richard Earnshaw <rearnsha@arm.com>
1938
1939 * arm.md (mulhi3): New expand pattern.
1940
605e86fa
JH
19412013-07-30 Jan Hubicka <jh@suse.cz>
1942 Martin Liska <marxin.liska@gmail.com>
1943
1944 * profile.c (compute_value_histograms): Do not ICE when
1945 there is mismatch only on some counters.
1946
14a87636
ZC
19472013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1948
1949 PR rtl-optimization/57637
1950 * function.c (move_insn_for_shrink_wrap): Also check the
1951 GEN set of the LIVE problem for the liveness analysis
1952 if it exists, otherwise give up.
1953
0100cd3f
BS
19542013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1955
1956 PR tree-optimization/57993
1957 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
1958 replaced statement in the candidate table.
1959 (phi_add_costs): Return infinite cost when the hidden basis does
1960 not dominate all phis on which the candidate is dependent.
1961 (replace_one_candidate): Record replaced statement in the
1962 candidate table.
1963
f0d811f7
JR
19642013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
1965
1966 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
1967 (ashlv2si3): New expander.
1968 (*ashlv2si3_i): New define_insn_and_split.
1969 * predicates.md (float_operation): Allow patterns with three
1970 basic sub-patterns.
1971
bba33211
JR
1972 PR rtl-optimization/58021
1973 * mode-switching.c (create_pre_exit): Always split off preceding
1974 insns if we are not at the basic block head.
1975
ff3f3951
MR
19762013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
1977
1978 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
1979 (UCLIBC_DYNAMIC_LINKER): New macro.
1980 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
1981 `-mnan=2008'.
1982 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
1983 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
1984 `-mnan=2008'.
1985 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
1986 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
1987 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
1988 for SF and DF modes. Use ieee_quad_format for TF mode.
1989 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
1990 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
1991 (mips_option_override): Handle `-mnan=legacy'.
1992 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
1993 `-mabs=2008' and `-mnan=2008'.
1994 (OPTION_DEFAULT_SPECS): Add "nan" default.
1995 (ASM_SPEC): Handle `-mnan='.
1996 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
1997 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
1998 comment accordingly.
1999 (neg<mode>2): Likewise.
2000 * config/mips/mips.opt (mabs, mnan): New options.
2001 * doc/install.texi (Configuration): Document `--with-nan=' option.
2002 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2003 `-mnan=' options.
2004 (MIPS Options): Document them.
2005 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2006 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2007 * configure: Regenerate.
2008 * config.in: Regenerate.
2009
576e2f37
UB
20102013-07-29 Uros Bizjak <ubizjak@gmail.com>
2011
2012 * config/i386/i386.md (float post-reload splitters): Do not check
2013 for subregs of SSE registers.
2014
20152013-07-29 Uros Bizjak <ubizjak@gmail.com>
2016 H.J. Lu <hongjiu.lu@intel.com>
2017
2018 PR target/57954
2019 PR target/57988
2020 * config/i386/i386.md (post-reload splitter
2021 to avoid partial SSE reg dependency stalls): New pattern.
2022
a71f0749
DV
20232013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2024
2025 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2026
4c97f1cc
JR
20272013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2028
2029 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2030 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2031
b0c733d8
CC
20322013-07-26 Cary Coutant <ccoutant@google.com>
2033
2034 * dwarf2out.c (die_checksum_ordered): Don't include template
2035 instantiations in signature.
2036 (is_template_parameter): New function.
2037 (is_template_instantiation): New function.
2038 (generate_skeleton_bottom_up): Don't include template instantiations
2039 in type unit DIE.
2040 (generate_skeleton): Likewise.
2041 (break_out_comdat_types): Move recursive call to break out nested
2042 types earlier.
2043 (prune_unused_types_mark_generic_parms_dies): Call
2044 is_template_parameter.
2045
040d8a1c
IB
20462013-07-26 Ian Bolton <ian.bolton@arm.com>
2047
2048 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2049 uses vector registers.
89fdc743 2050 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 2051
85bd4ac6 20522013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 2053 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT
2054
2055 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 2056 where appropriate.
85bd4ac6
KT
2057 * config/arm/arm.md (movcond_addsi): New splitter.
2058
b6051207
SA
20592013-07-25 Sterling Augustine <saugustine@google.com>
2060
2061 * dwarf2out.c (size_of_pubnames): Move code to...
2062 (include_pubname_in_output): ...here. New.
2063 (want_pubnames): Rearrange.
2064 (output_pubnames): Call include_pubname_in_output. Move assertion.
2065
41a7c215
CM
20662013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2067
2068 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2069
20702013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2071
2072 PR target/38836
2073 * doc/extend.texi: Remove obsolete builtins. Fix
2074 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2075
49bde175
JH
20762013-07-25 Jan Hubicka <jh@suse.cz>
2077
2078 * cgraph.c (release_function_body): Break out from ...
2079 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2080 and DECL_ARGUMENTS.
2081 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2082 old_tree in the map.
2083 (create_specialized_node): Update.
2084 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2085 into index.
41a7c215
CM
2086 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2087 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH
2088 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2089 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2090 DECL_RESULT.
2091
4870352d
KT
20922013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2093
2094 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2095 addsi3_carryin_alt2_<optab>): Correct output template.
2096
1d6aee1c
KT
20972013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2098
2099 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2100 Adjust for arm_restrict_it.
2101 Remove trailing whitespace.
2102
c7cafd75
MK
21032013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2104
017d38f5
MK
2105 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2106 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2107
c7cafd75
MK
2108 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2109
00b7527b
VM
21102013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2111
2112 PR rtl-optimization/57960
2113 * lra-constraints.c (process_alt_operands): Use the right mode
2114 when checking strict_low.
2115
4f63dfc6
JH
21162013-07-25 Jan Hubicka <jh@suse.cz>
2117
2118 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2119 * cgraph.c (cgraph_remove_node): Do not release function body
2120 when in cgraph streaming.
2121 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM
2122 in other partitions are not considered reachable; fix handling of
2123 clones.
4f63dfc6 2124
e6dcfa18
RR
21252013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2126
2127 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2128
0b93d3b6
RR
21292013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2130
2131 PR target/19599
2132 PR target/57731
bb80c2eb 2133 PR target/57837
e6dcfa18 2134 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 2135 Ss with US. Adjust output for v5 and v4t.
41a7c215 2136 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 2137
0b93d3b6
RR
2138 * config/arm/constraints.md ("Ss"): Rename to US.
2139
690688b3
TG
21402013-07-25 Terry Guo <terry.guo@arm.com>
2141
2142 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2143 shift_add/shift_sub0/shift_sub1 RTXs.
2144
a4ad093b
BS
21452013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2146 Anton Blanchard <anton@au1.ibm.com>
2147
2148 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2149 (altivec_vpks<VI_char>ss): Likewise.
2150 (altivec_vpks<VI_char>us): Likewise.
2151 (altivec_vpku<VI_char>us): Likewise.
2152 (altivec_vpku<VI_char>um): Likewise.
2153
8c5005ce
DM
21542013-07-24 David Malcolm <dmalcolm@redhat.com>
2155
2156 Introduce context class.
2157
2158 * Makefile.in (CONTEXT_H): New.
2159 (OBJS): Add context.o.
2160 (toplev.o): Add CONTEXT_H to dependencies.
2161 (context.o): New.
2162
41a7c215 2163 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM
2164
2165 * context.c: New.
2166
2167 * context.h: New.
2168
ce4a9422
JR
21692013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2170
2171 PR rtl-optimization/57968
2172 * mode-switching.c (create_pre_exit): Allow instructions that
2173 don't set a return register to need a non-exit mode.
2174
493f4c9e
BS
21752013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2176 Anton Blanchard <anton@au1.ibm.com>
2177
b5860fd3
WS
2178 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2179 operands to vperm for little endian.
2180 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2181 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 2182
73792b92
BS
21832013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2184 Anton Blanchard <anton@au1.ibm.com>
2185
2186 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2187 two operands for little-endian.
2188
14c21302
SE
21892013-07-23 Steve Ellcey <sellcey@mips.com>
2190
2191 * config/mips/mips.c (mips_case_values_threshold): New.
2192 (TARGET_CASE_VALUES_THRESHOLD): Define.
2193
a6056198
BS
21942013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2195 Anton Blanchard <anton@au1.ibm.com>
2196
2197 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2198 selection of field for vector splat in little endian mode.
2199
dd7a40e1
MM
22002013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2201
2202 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2203 expanders to rs6000.md.
2204 (ior<mode>3): Likewise.
2205 (and<mode>3): Likewise.
2206 (one_cmpl<mode>2): Likewise.
2207 (nor<mode>3): Likewise.
2208 (andc<mode>3): Likewise.
2209 (eqv<mode>3): Likewise.
2210 (nand<mode>3): Likewise.
2211 (orc<mode>3): Likewise.
2212
2213 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2214 declaration.
2215
2216 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2217 to split multi-word logical operations.
2218 (rs6000_split_logical_di): Likewise.
2219 (rs6000_split_logical): Likewise.
2220
2221 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2222 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2223 and allow TImode operations in 32-bit.
2224 (vsx_and<mode>3_64bit): Likewise.
2225 (vsx_ior<mode>3_32bit): Likewise.
2226 (vsx_ior<mode>3_64bit): Likewise.
2227 (vsx_xor<mode>3_32bit): Likewise.
2228 (vsx_xor<mode>3_64bit): Likewise.
2229 (vsx_one_cmpl<mode>2_32bit): Likewise.
2230 (vsx_one_cmpl<mode>2_64bit): Likewise.
2231 (vsx_nor<mode>3_32bit): Likewise.
2232 (vsx_nor<mode>3_64bit): Likewise.
2233 (vsx_andc<mode>3_32bit): Likewise.
2234 (vsx_andc<mode>3_64bit): Likewise.
2235 (vsx_eqv<mode>3_32bit): Likewise.
2236 (vsx_eqv<mode>3_64bit): Likewise.
2237 (vsx_nand<mode>3_32bit): Likewise.
2238 (vsx_nand<mode>3_64bit): Likewise.
2239 (vsx_orc<mode>3_32bit): Likewise.
2240 (vsx_orc<mode>3_64bit): Likewise.
2241
2242 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2243 logical types in GPRs.
2244
2245 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2246 logical insns to rs6000.md, and allow TImode operations in
2247 32-bit.
2248 (altivec_ior<mode>3): Likewise.
2249 (altivec_xor<mode>3): Likewise.
2250 (altivec_one_cmpl<mode>2): Likewise.
2251 (altivec_nor<mode>3): Likewise.
2252 (altivec_andc<mode>3): Likewise.
2253
2254 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2255 attributes for moving the 128-bit logical operations into
2256 rs6000.md.
2257 (BOOL_REGS_OUTPUT): Likewise.
2258 (BOOL_REGS_OP1): Likewise.
2259 (BOOL_REGS_OP2): Likewise.
2260 (BOOL_REGS_UNARY): Likewise.
2261 (BOOL_REGS_AND_CR0): Likewise.
2262 (one_cmpl<mode>2): Add support for DI logical operations on
2263 32-bit, splitting the operations to 32-bit.
2264 (anddi3): Likewise.
2265 (iordi3): Likewise.
2266 (xordi3): Likewise.
2267 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2268 changes to combine the 32/64-bit code, allow logical operations on
2269 TI mode in 32-bit, and to use similar match_operator patterns like
2270 scalar mode uses. Combine the Altivec and VSX code for logical
2271 operations, and move it here.
2272 (ior<mode>3, 128-bit types): Likewise.
2273 (xor<mode>3, 128-bit types): Likewise.
2274 (one_cmpl<mode>3, 128-bit types): Likewise.
2275 (nor<mode>3, 128-bit types): Likewise.
2276 (andc<mode>3, 128-bit types): Likewise.
2277 (eqv<mode>3, 128-bit types): Likewise.
2278 (nand<mode>3, 128-bit types): Likewise.
2279 (orc<mode>3, 128-bit types): Likewise.
2280 (and<mode>3_internal): Likewise.
2281 (bool<mode>3_internal): Likewise.
2282 (boolc<mode>3_internal1): Likewise.
2283 (boolc<mode>3_internal2): Likewise.
2284 (boolcc<mode>3_internal1): Likewise.
2285 (boolcc<mode>3_internal2): Likewise.
2286 (eqv<mode>3_internal1): Likewise.
2287 (eqv<mode>3_internal2): Likewise.
2288 (one_cmpl1<mode>3_internal): Likewise.
2289
9b69d4b4
DH
22902013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2291
2292 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2293 Rename flag_stack_usage to flag_stack_usage_info.
2294
450b0ebb
DH
22952013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2296
a5f257fd
ME
2297 * config/microblaze/sync.md: New file.
2298 * config/microblaze/microblaze.md: Include sync.md
2299 * config/microblaze/microblaze.c: Add print_operand 'y'.
2300 * config/microblaze/constraints.md: Add memory_contraint
2301 'Q' which is a single register.
450b0ebb 2302
49fba14f
EB
23032013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2304
2305 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2306
c7b3b99f
PCC
23072013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2308
2309 * reload.c (find_reloads): Exit loop once we find this operand
2310 cannot be reloaded somehow for this alternative.
2311
2312 * reload.c (find_reloads): Exit loop once we find a hard register.
2313
2314 * rtlanal.c (computed_jump_p): Exit loop once we find label
2315 reference is used.
2316
2317 * i386.c (ix86_pad_returns): Exit loop after setting replace.
2318
2319 * cfgloopmanip.c (remove_path): Exit loop after setting
2320 irred_invalidated.
2321
2322 * gensupport.c (subst_dup): Avoid loop if code is not
2323 MATCH_DUP nor MATCH_OP_DUP.
2324
c8fbf1fa
NBJ
23252013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
2326
2327 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
2328
9259db42
YZ
23292013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2330
2331 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
2332 true for SP_REGNUM if mode == ptr_mode.
2333 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
2334 with value R0_REGNUM + 31.
2335
00edcfbe
YZ
23362013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2337
2338 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
2339 pad pointer-typed argument downward.
2340
43be9a95
YZ
23412013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2342
2343 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
2344 and __ILP32__ when the ILP32 model is in use.
2345
28514dda
YZ
23462013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2347
2348 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
2349 (aarch64_load_symref_appropriately): In the case of
2350 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
2351 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
2352 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
2353 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
2354 if 'mode' doesn't equal to 'ptr_mode'.
2355 (aarch64_output_mi_thunk): Add an assertion on the alignment of
2356 'vcall_offset'; change to call aarch64_emit_move differently depending
2357 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
2358 to calculate the upper bound of 'vcall_offset'.
2359 (aarch64_cannot_force_const_mem): Change to also return true if
2360 mode != ptr_mode.
2361 (aarch64_legitimize_reload_address): In the case of large
2362 displacements, add new local variable 'xmode' and an assertion
2363 based on it; change to use 'xmode' to generate the new rtx and
2364 reload.
2365 (aarch64_asm_trampoline_template): Change to generate the template
2366 differently depending on TARGET_ILP32 or not; change to use
2367 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
2368 (aarch64_trampoline_size): Removed.
2369 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
2370 and replace immediate literals with it. Change to use 'ptr_mode'
2371 instead of 'DImode' and call convert_memory_address if the mode
2372 of 'fnaddr' doesn't equal to 'ptr_mode'.
2373 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
2374 to output symbol.
2375 (aarch64_elf_asm_destructor): Likewise.
2376 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
2377 on TARGET_ILP32 instead of aarch64_trampoline_size.
2378 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
2379 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
2380 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
2381 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
2382 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
2383 (storewb_pair<GPI:mode>_<P:mode>): ... this.
2384 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
2385 depending on the value of 'mode'.
2386 (add_losym_<mode>): New.
2387 (ldr_got_small_<mode>): New, based on ldr_got_small.
2388 (ldr_got_small): Remove.
2389 (ldr_got_small_sidi): New.
2390 * config/aarch64/iterators.md (P): New.
2391 (PTR): Change to 'ptr_mode' in the condition.
2392
17a819cb
YZ
23932013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2394
2395 * config.gcc (aarch64*-*-*): Support --with-abi.
2396 (aarch64*-*-elf): Support --with-multilib-list.
2397 (aarch64*-*-linux*): Likewise.
2398 (supported_defaults): Add abi to aarch64*-*-*.
2399 * configure.ac: Mention AArch64 for --with-multilib-list.
2400 * configure: Re-generated.
2401 * config/aarch64/biarchilp32.h: New file.
2402 * config/aarch64/biarchlp64.h: New file.
2403 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
2404 (ABI_SPEC): Ditto.
2405 (MULTILIB_DEFAULTS): Ditto.
2406 (DRIVER_SELF_SPECS): Ditto.
2407 (ASM_SPEC): Update to also substitute -mabi.
2408 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
2409 file whose name depends on -mabi= and -mbig-endian.
2410 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
2411 TARGET_ILP32.
2412 (POINTER_SIZE): New define.
2413 (POINTERS_EXTEND_UNSIGNED): Ditto.
2414 (enum aarch64_abi_type): New enumeration tag.
2415 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
2416 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
2417 (TARGET_ILP32): New define.
2418 * config/aarch64/aarch64.opt (mabi): New.
2419 (aarch64_abi): New.
2420 (ilp32, lp64): New values for -mabi.
2421 * config/aarch64/t-aarch64 (comma): New define.
2422 (MULTILIB_OPTIONS): Ditto.
2423 (MULTILIB_DIRNAMES): Ditto.
2424 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
2425 * doc/invoke.texi: Document -mabi for AArch64.
2426
9f9cbdce
GJL
24272013-07-23 Georg-Johann Lay <avr@gjlay.de>
2428
2429 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
2430
7061977b
BS
24312013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2432 Anton Blanchard <anton@au1.ibm.com>
2433
2434 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
2435 endianness when selecting field to splat.
2436
75a41878
EC
24372013-07-22 Eric Christopher <echristo@gmail.com>
2438
2439 * dwarf2out.c (die_odr_checksum): New function to use
2440 CHECKSUM_ macros and ULEB128 for DIE tag.
2441 (generate_type_signature): Use.
41a7c215 2442
38ae58ca
EB
24432013-07-22 Eric Botcazou <ebotcazou@adacore.com>
2444
2445 * config.gcc (sparc*-*-*): Accept leon3 processor.
2446 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
2447 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
2448 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
2449 * config/sparc/sparc.opt (enum processor_type): Add leon3.
2450 (mfix-ut699): Adjust comment.
2451 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
2452 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
2453 (CPP_CPU_SPEC): Likewise.
2454 (ASM_CPU_SPEC): Likewise.
2455 * config/sparc/sparc.c (leon3_cost): New constant.
2456 (sparc_option_override): Add leon3 support.
2457 (mem_ref): New function.
2458 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
2459 (sparc_do_work_around_errata): Look into the instruction in the delay
2460 slot and adjust accordingly. Add fix for the data cache nullify issues
2461 of the UT699. Change insertion position for the NOP.
2462 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
2463 (leon3_load): New reservation.
2464 (leon_store): Bump latency to 2.
2465 (grfpu): New automaton.
2466 (grfpu_alu): New unit.
2467 (grfpu_ds): Likewise.
2468 (leon_fp_alu): Adjust.
2469 (leon_fp_mult): Delete.
2470 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
2471 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
2472 * config/sparc/sparc.md (cpu): Add leon3.
2473 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
2474 (swapsi): Likewise.
2475 (atomic_test_and_set): Likewise.
2476 (ldstub): Likewise.
2477
04dfc6df
JU
24782013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
2479
2480 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
2481 default for R5900 targets.
2482 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
2483 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
2484 * config/mips/mips.c (mips_option_override): Report an error for
2485 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
2486 for -march=r5900 -mhard-float.
2487
c7b3b99f 24882013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC
2489
2490 * df-problems.c (can_move_insns_across): Exit loop once we
2491 find a non-fixed, non-global register.
2492
2493 * ipa-pure-const.c (propagate_nothrow): Exit loop after
2494 setting can_throw.
2495
2496 * omega.c (omega_eliminate_red): Break after setting red_found.
2497 (omega_problem_has_red_equations): Similarly after setting found.
2498 (omega_query_variable): Similarly after setting coupled.
2499
9b6e6981
MP
25002013-07-22 Marek Polacek <polacek@redhat.com>
2501
2502 * gimplify.c: Don't include gimple.h twice.
2503
5e5f7673
KT
25042013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2505
2506 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
2507 instead of TARGET_THUMB1.
2508 (Pz): New constraint.
2509 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
2510 encodings.
2511 (compare_negsi_si): Likewise.
2512 (compare_addsi2_op0): Likewise.
2513 (compare_addsi2_op1): Likewise.
2514 (addsi3_carryin_<optab>): Likewise.
2515 (addsi3_carryin_alt2_<optab>): Likewise.
2516 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
2517 for arm_restrict_it.
2518 (subsi3_carryin): Likewise.
2519 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
2520 (minmax_arithsi): Disable for arm_restrict_it.
2521 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
2522 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
2523 (satsi_<SAT:code>_shift): Likewise.
2524 (arm_shiftsi3): Add alternative for 16-bit encoding.
2525 (arm32_movhf): Disable for arm_restrict_it.
2526 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
2527 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
2528
859abddd
SN
25292013-07-22 Sofiane Naci <sofiane.naci@arm.com>
2530
2531 * config/arm/arm.md (attribute "insn"): Delete.
2532 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
2533 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
2534 (not_shiftsi): Update for attribute change.
2535 (not_shiftsi_compare0): Likewise.
2536 (not_shiftsi_compare0_scratch): Likewise.
2537 (arm_one_cmplsi2): Likewise.
2538 (thumb1_one_cmplsi2): Likewise.
2539 (notsi_compare0): Likewise.
2540 (notsi_compare0_scratch): Likewise.
2541 (thumb1_movdi_insn): Likewise.
2542 (arm_movsi_insn): Likewise.
2543 (movhi_insn_arch4): Likewise.
2544 (movhi_bytes): Likewise.
2545 (arm_movqi_insn): Likewise.
2546 (thumb1_movqi_insn): Likewise.
2547 (arm32_movhf): Likewise.
2548 (thumb1_movhf): Likewise.
2549 (arm_movsf_soft_insn): Likewise.
2550 (thumb1_movsf_insn): Likewise.
2551 (thumb_movdf_insn): Likewise.
2552 (movsicc_insn): Likewise.
2553 (movsfcc_soft_insn): Likewise.
2554 (and_scc): Likewise.
2555 (cond_move): Likewise.
2556 (if_move_not): Likewise.
2557 (if_not_move): Likewise.
2558 (if_shift_move): Likewise.
2559 (if_move_shift): Likewise.
2560 (if_shift_shift): Likewise.
2561 (if_not_arith): Likewise.
2562 (if_arith_not): Likewise.
2563 (cond_move_not): Likewise.
2564 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
2565 (neon_mov<mode>): Likewise.
2566 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
2567 (thumb2_movsi_vfp): Likewise.
2568 (movsf_vfp): Likewise.
2569 (thumb2_movsf_vfp): Likewise.
371e77e3
UB
2570 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
2571 change.
859abddd
SN
2572 (cortexa7_older_only): Likewise.
2573 (cortexa7_younger): Likewise.
2574 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
2575 (1020alu_shift_op): Likewise.
2576 (1020alu_shift_reg_op): Likewise.
2577 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
2578 (alu_shift_op): Likewise.
2579 (alu_shift_reg_op): Likewise.
2580 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
2581 (11_alu_shift_op): Likewise.
2582 (11_alu_shift_reg_op): Likewise.
2583 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
2584 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
2585 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2586 change.
859abddd
SN
2587 (cortex_a15_alu_shift): Likewise.
2588 (cortex_a15_alu_shift_reg): Likewise.
2589 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
2590 (cortex_a5_alu_shift): Likewise.
371e77e3
UB
2591 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
2592 change.
859abddd 2593 (cortex_a53_alu_shift): Likewise.
371e77e3
UB
2594 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
2595 change.
859abddd
SN
2596 (cortex_a7_alu_reg): Likewise.
2597 (cortex_a7_alu_shift): Likewise.
2598 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2599 (cortex_a8_alu_shift): Likewise.
2600 (cortex_a8_alu_shift_reg): Likewise.
2601 (cortex_a8_mov): Likewise.
2602 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2603 (cortex_a9_dp_shift): Likewise.
2604 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
2605 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
2606 (cortex_r4_mov): Likewise.
2607 (cortex_r4_alu_shift): Likewise.
2608 (cortex_r4_alu_shift_reg): Likewise.
2609 * config/arm/fa526.md (526_alu_op): Update for attribute change.
2610 (526_alu_shift_op): Likewise.
2611 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
2612 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
2613 (626te_alu_shift_op): Likewise.
2614 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
2615 (726te_alu_op): Likewise.
2616 (726te_alu_shift_op): Likewise.
2617 (726te_alu_shift_reg_op): Likewise.
2618 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
2619 (mp626_alu_shift_op): Likewise.
2620 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
2621 (pj4_alu_e1_conds): Likewise.
2622 (pj4_alu): Likewise.
2623 (pj4_alu_conds): Likewise.
2624 (pj4_shift): Likewise.
2625 (pj4_shift_conds): Likewise.
2626 (pj4_alu_shift): Likewise.
2627 (pj4_alu_shift_conds): Likewise.
2628
95b97fac
KT
26292013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2630
2631 * config/arm/predicates.md (shiftable_operator_strict_it):
2632 New predicate.
2633 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
2634 Disable cond_exec version for arm_restrict_it.
2635 (thumb2_smaxsi3): Convert to generate cond_exec.
2636 (thumb2_sminsi3): Likewise.
2637 (thumb32_umaxsi3): Likewise.
2638 (thumb2_uminsi3): Likewise.
2639 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
2640 (thumb2_neg_abssi2): Likewise.
2641 (thumb2_mov_scc): Add alternative for 16-bit encoding.
2642 (thumb2_movsicc_insn): Adjust alternatives.
2643 (thumb2_mov_negscc): Disable for arm_restrict_it.
2644 (thumb2_mov_negscc_strict_it): New pattern.
2645 (thumb2_mov_notscc_strict_it): New pattern.
2646 (thumb2_mov_notscc): Disable for arm_restrict_it.
2647 (thumb2_ior_scc): Likewise.
2648 (thumb2_ior_scc_strict_it): New pattern.
2649 (thumb2_cond_move): Adjust for arm_restrict_it.
2650 (thumb2_cond_arith): Disable for arm_restrict_it.
2651 (thumb2_cond_arith_strict_it): New pattern.
2652 (thumb2_cond_sub): Adjust for arm_restrict_it.
2653 (thumb2_movcond): Likewise.
2654 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
2655 (thumb2_zero_extendhisi2_v6): Likewise.
2656 (thumb2_zero_extendqisi2_v6): Likewise.
2657 (orsi_notsi_si): Likewise.
2658 (orsi_not_shiftsi_si): Likewise.
2659
d5a6ef82
GJL
26602013-07-22 Georg-Johann Lay <avr@gjlay.de>
2661
2662 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
2663 instruction sequence is 1 byte shorter.
2664
d9e71154
UB
26652013-07-22 Uros Bizjak <ubizjak@gmail.com>
2666
2667 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
2668 it is not needed after split.
2669
8c6381ec
IS
26702013-07-20 Iain Sandoe <iain@codesourcery.com>
2671
2672 PR target/51784
2673 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
2674 second label for nonlocal goto receivers. Don't output pic base labels
2675 unless we're producing PIC; mark that action unreachable().
2676 (ix86_save_reg): If the function contains a nonlocal label, save the
2677 PIC base reg.
2678 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
371e77e3 2679 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS
2680 (update_pic_label_number_if_needed): New.
2681 (machopic_output_function_base_name): Adjust for nonlocal receiver
2682 case.
2683 (machopic_should_output_picbase_label): New.
2684 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
2685 (nonlocal_goto_receiver): New insn and split.
2686
d05d0709
JG
26872013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
2688
2689 * config/aarch64/aarch64-builtins.c
2690 (aarch64_fold_builtin): Fold abs in all modes.
2691 * config/aarch64/aarch64-simd-builtins.def
2692 (abs): Enable for all modes.
2693 * config/aarch64/arm_neon.h
2694 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
2695 (vabs_f64): Add missing intrinsic.
2696
ba590e63
IB
26972013-07-19 Ian Bolton <ian.bolton@arm.com>
2698
2699 * config/aarch64/arm_neon.h (vabs_s64): New function
2700
02371798
GJL
27012013-07-19 Georg-Johann Lay <avr@gjlay.de>
2702
2703 PR target/57516
2704 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
2705 * config/avr/avr.md (adjust_len): Add `round'.
2706 * config/avr/avr-protos.h (avr_out_round): New prototype.
2707 (avr_out_plus): Add `out_label' argument.
2708 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
2709 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
2710 Handle the case where `insn' is just a pattern.
2711 (avr_out_bitop): Handle the case where `insn' is just a pattern.
2712 (avr_out_round): New function.
2713 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
2714
b950ac25
DH
27152013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
2716
2717 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2718 Add check for flag_stack_usage to handle -fstack-usage support
2719
ed1fd999
PH
27202013-07-18 Pat Haugen <pthaugen@us.ibm.com>
2721
2722 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
2723 interaction for new Power8 flags and VSX.
2724
df9dda2d
ST
27252013-07-18 Sriraman Tallam <tmsriram@google.com>
2726
2727 PR middle-end/57698
2728 * tree-inline.c (expand_call_inline): Emit errors during
2729 early_inlining only if optimization is not turned on.
2730
0f9cdef4
DM
27312013-07-18 David Malcolm <dmalcolm@redhat.com>
2732
2733 * passes.def: New.
2734
2735 * passes.c (init_optimization_passes): Move the construction of
2736 the pass hierarchy into a new passes.def file.
2737
2738 * Makefile.in (passes.o): Add dependency on passes.def.
2739
2efa4087
DM
27402013-07-18 David Malcolm <dmalcolm@redhat.com>
2741
2742 * passes.c (init_optimization_passes): Introduce macros for
2743 constructing the tree of passes (INSERT_PASSES_AFTER,
2744 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
2745 TERMINATE_PASS_LIST).
2746
bc404e1b
VM
27472013-07-18 Vladimir Makarov <vmakarov@redhat.com>
2748 Wei Mi <wmi@google.com>
2749
2750 PR rtl-optimization/57878
2751 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
2752 top.
2753 (reload_pseudo_compare_func): Check nregs first for reload
2754 pseudos.
2755
d675843f
DM
27562013-07-18 David Malcolm <dmalcolm@redhat.com>
2757
2758 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
2759
c7b3b99f 27602013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC
2761
2762 * read-rtl.c (validate_const_int): Once an invalid character is
2763 seen, quit the loop.
2764
2765 * gengtype.c (write_roots): Similarly once we find the "deletable"
2766 or "if_marked" option.
2767
52fceb44
SN
27682013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2769
2770 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
2771 "xtab" and "sat". Move value "clz" from here to ...
2772 (attriubte "type"): ... here.
2773 (satsi_<SAT:code>): Delete "insn" attribute.
2774 (satsi_<SAT:code>_shift): Likewise.
2775 (arm_zero_extendqisi2addsi): Likewise.
2776 (arm_extendqisi2addsi): Likewise.
2777 (clzsi2): Update for attribute changes.
2778 (rbitsi2): Likewise.
371e77e3
UB
2779 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
2780 attribute.
52fceb44
SN
2781 (arm_usatsihi): Likewise.
2782 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2783
006bd006
SN
27842013-07-18 Sofiane Naci <sofiane.naci@arm.com>
2785
2786 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB
2787 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
2788 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN
2789 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
2790 in alphabetical order.
2791 (attribute "core_cycles"): Update for attribute changes.
2792 (arm_addsi3): Likewise.
2793 (addsi3_compare0): Likewise.
2794 (addsi3_compare0_scratch): Likewise.
2795 (addsi3_compare_op1): Likewise.
2796 (addsi3_compare_op2): Likewise.
2797 (compare_addsi2_op0): Likewise.
2798 (compare_addsi2_op1): Likewise.
2799 (addsi3_carryin_shift_<optab>): Likewise.
2800 (subsi3_carryin_shift): Likewise.
2801 (rsbsi3_carryin_shift): Likewise.
2802 (arm_subsi3_insn): Likewise.
2803 (subsi3_compare0): Likewise.
2804 (subsi3_compare): Likewise.
2805 (arm_andsi3_insn): Likewise.
2806 (thumb1_andsi3_insn): Likewise.
2807 (andsi3_compare0): Likewise.
2808 (andsi3_compare0_scratch): Likewise.
2809 (zeroextractsi_compare0_scratch
2810 (andsi_not_shiftsi_si): Likewise.
2811 (iorsi3_insn): Likewise.
2812 (iorsi3_compare0): Likewise.
2813 (iorsi3_compare0_scratch): Likewise.
2814 (arm_xorsi3): Likewise.
2815 (thumb1_xorsi3_insn): Likewise.
2816 (xorsi3_compare0): Likewise.
2817 (xorsi3_compare0_scratch): Likewise.
2818 (satsi_<SAT:code>_shift): Likewise.
2819 (rrx): Likewise.
2820 (arm_shiftsi3): Likewise.
2821 (shiftsi3_compare0): Likewise.
2822 (not_shiftsi): Likewise.
2823 (not_shiftsi_compare0): Likewise.
2824 (not_shiftsi_compare0_scratch): Likewise.
2825 (arm_one_cmplsi2): Likewise.
2826 (thumb_one_complsi2): Likewise.
2827 (notsi_compare0): Likewise.
2828 (notsi_compare0_scratch): Likewise.
2829 (thumb1_zero_extendhisi2): Likewise.
2830 (arm_zero_extendhisi2): Likewise.
2831 (arm_zero_extendhisi2_v6): Likewise.
2832 (arm_zero_extendhisi2addsi): Likewise.
2833 (thumb1_zero_extendqisi2): Likewise.
2834 (thumb1_zero_extendqisi2_v6): Likewise.
2835 (arm_zero_extendqisi2): Likewise.
2836 (arm_zero_extendqisi2_v6): Likewise.
2837 (arm_zero_extendqisi2addsi): Likewise.
2838 (thumb1_extendhisi2): Likewise.
2839 (arm_extendhisi2): Likewise.
2840 (arm_extendhisi2_v6): Likewise.
2841 (arm_extendqisi): Likewise.
2842 (arm_extendqisi_v6): Likewise.
2843 (arm_extendqisi2addsi): Likewise.
2844 (thumb1_extendqisi2): Likewise.
2845 (thumb1_movdi_insn): Likewise.
2846 (arm_movsi_insn): Likewise.
2847 (movsi_compare0): Likewise.
2848 (movhi_insn_arch4): Likewise.
2849 (movhi_bytes): Likewise.
2850 (arm_movqi_insn): Likewise.
2851 (thumb1_movqi_insn): Likewise.
2852 (arm32_movhf): Likewise.
2853 (thumb1_movhf): Likewise.
2854 (arm_movsf_soft_insn): Likewise.
2855 (thumb1_movsf_insn): Likewise.
2856 (movdf_soft_insn): Likewise.
2857 (thumb_movdf_insn): Likewise.
2858 (arm_cmpsi_insn): Likewise.
2859 (cmpsi_shiftsi): Likewise.
2860 (cmpsi_shiftsi_swp): Likewise.
2861 (arm_cmpsi_negshiftsi_si): Likewise.
2862 (movsicc_insn): Likewise.
2863 (movsfcc_soft_insn): Likewise.
2864 (arith_shiftsi): Likewise.
2865 (arith_shiftsi_compare0
2866 (arith_shiftsi_compare0_scratch
2867 (sub_shiftsi): Likewise.
2868 (sub_shiftsi_compare0
2869 (sub_shiftsi_compare0_scratch
2870 (and_scc): Likewise.
2871 (cond_move): Likewise.
2872 (if_plus_move): Likewise.
2873 (if_move_plus): Likewise.
2874 (if_move_not): Likewise.
2875 (if_not_move): Likewise.
2876 (if_shift_move): Likewise.
2877 (if_move_shift): Likewise.
2878 (if_shift_shift): Likewise.
2879 (if_not_arith): Likewise.
2880 (if_arith_not): Likewise.
2881 (cond_move_not): Likewise.
2882 (thumb1_ashlsi3): Set type attribute.
2883 (thumb1_ashrsi3): Likewise.
2884 (thumb1_lshrsi3): Likewise.
2885 (thumb1_rotrsi3): Likewise.
2886 (shiftsi3_compare0_scratch): Likewise.
2887 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
2888 (neon_mov<mode>): Likewise.
371e77e3
UB
2889 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
2890 attribute changes.
006bd006
SN
2891 (thumb2_movsi_insn): Likewise.
2892 (thumb2_cmpsi_neg_shiftsi): Likewise.
2893 (thumb2_extendqisi_v6): Likewise.
2894 (thumb2_zero_extendhisi2_v6): Likewise.
2895 (thumb2_zero_extendqisi2_v6): Likewise.
2896 (thumb2_shiftsi3_short): Likewise.
2897 (thumb2_addsi3_compare0_scratch): Likewise.
2898 (orsi_not_shiftsi_si): Likewise.
2899 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
2900 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
2901 changes.
2902 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2903 (1020alu_shift_op): Likewise.
2904 (1020alu_shift_reg_op): Likewise.
2905 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2906 (alu_shift_op): Likewise.
2907 (alu_shift_reg_op): Likewise.
2908 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2909 (11_alu_shift_op): Likewise.
2910 (11_alu_shift_reg_op): Likewise.
2911 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2912 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
2913 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2914 changes.
006bd006
SN
2915 (cortex_a15_alu_shift): Likewise.
2916 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB
2917 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
2918 changes.
006bd006
SN
2919 (cortex_a5_alu_shift): Likewise.
2920 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
2921 changes.
2922 (cortex_a53_alu_shift): Likewise.
2923 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
2924 changes.
2925 (cortex_a7_alu_reg): Likewise.
2926 (cortex_a7_alu_shift): Likewise.
371e77e3
UB
2927 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
2928 changes.
006bd006
SN
2929 (cortex_a8_alu_shift): Likewise.
2930 (cortex_a8_alu_shift_reg): Likewise.
2931 (cortex_a8_mov): Likewise.
2932 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2933 (cortex_a9_dp_shift): Likewise.
371e77e3
UB
2934 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
2935 changes.
2936 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
2937 changes.
006bd006
SN
2938 (cortex_r4_mov): Likewise.
2939 (cortex_r4_alu_shift): Likewise.
2940 (cortex_r4_alu_shift_reg): Likewise.
2941 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2942 (526_alu_shift_op): Likewise.
2943 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2944 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2945 (626te_alu_shift_op): Likewise.
2946 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
2947 (726te_alu_op): Likewise.
2948 (726te_alu_shift_op): Likewise.
2949 (726te_alu_shift_reg_op): Likewise.
2950 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2951 (mp626_alu_shift_op): Likewise.
2952 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
2953 (pj4_alu_e1_conds): Likewise.
2954 (pj4_alu): Likewise.
2955 (pj4_alu_conds): Likewise.
2956 (pj4_shift): Likewise.
2957 (pj4_shift_conds): Likewise.
2958 (pj4_alu_shift): Likewise.
2959 (pj4_alu_shift_conds): Likewise.
371e77e3
UB
2960 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
2961 changes.
006bd006
SN
2962 (cortexa7_older_only): Likewise.
2963 (cortexa7_younger): Likewise.
2964
651df1b2
DM
29652013-07-18 David Malcolm <dmalcolm@redhat.com>
2966
2967 * ipa-pure-const.c (generate_summary): Rename to...
2968 (pure_const_generate_summary): ... this.
2969
ea717bd7
IS
29702013-07-17 Iain Sandoe <iain@codesourcery.com>
2971
2972 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
2973
a1259a13
YR
29742013-07-17 Yvan Roux <yvan.roux@linaro.org>
2975
2976 PR target/57909
2977 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
2978 usage in HI mode.
2979
167f68ed
AK
29802013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2981
2982 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
2983 enabled without -march=zEC12.
2984 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
2985 flags to be set.
2986
f900a982
MR
29872013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
2988
2989 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
2990 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
2991 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
2992 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
2993 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
2994 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
2995 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
2996 ISA_HAS_FP4.
2997 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
2998 and ISA_HAS_NMADD3_NMSUB3.
2999 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3000 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3001 (nmsub4<mode>, nmsub3<mode>): Likewise.
3002 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3003
d4ed27eb
MR
30042013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3005
3006 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3007 TARGET_MIPS5400 checking.
3008
7a5add18
PB
30092013-07-16 Jakub Jelinek <jakub@redhat.com>
3010 Peter Bergner <bergner@vnet.ibm.com>
3011
3012 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3013 registers in the comment.
3014 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3015 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3016 rather than FIRST_PSEUDO_REGISTERS.
3017
30182013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB
3019
3020 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3021 enable extra ISA flags with TARGET_HTM.
3022
d2ab0929
MR
30232013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3024
3025 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3026 Fix comment typos.
3027
839c74bc
CH
30282013-07-15 Cong Hou <congh@google.com>
3029
3030 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3031 in compare function for sorting.
3032
7a5add18 30332013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB
3034
3035 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3036 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3037 * config/rs6000/rs6000.opt: Add -mhtm option.
3038 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3039 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3040 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3041 __HTM__ if the HTM instructions are available.
3042 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3043 htm_spr_reg_operand): New define_predicates.
3044 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3045 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3046 Include htm.md.
3047 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3048 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3049 HTM builtin functions.
3050 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3051 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3052 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3053 (rs6000_builtin_mask_calculate): Likewise.
3054 (rs6000_option_override_internal): Likewise.
3055 (bdesc_htm): Add new HTM builtin support.
3056 (htm_spr_num): New function.
3057 (htm_spr_regno): Likewise.
3058 (rs6000_htm_spr_icode): Likewise.
3059 (htm_expand_builtin): Likewise.
3060 (htm_init_builtins): Likewise.
3061 (rs6000_expand_builtin): Add support for HTM builtin functions.
3062 (rs6000_init_builtins): Likewise.
371e77e3
UB
3063 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3064 option.
0258b6e4
PB
3065 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3066 (TARGET_HTM, MASK_HTM): Define macros.
3067 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3068 (FIXED_REGISTERS): Likewise.
3069 (CALL_USED_REGISTERS): Likewise.
3070 (CALL_REALLY_USED_REGISTERS): Likewise.
3071 (REG_ALLOC_ORDER): Likewise.
3072 (enum reg_class): Likewise.
3073 (REG_CLASS_NAMES): Likewise.
3074 (REG_CLASS_CONTENTS): Likewise.
3075 (REGISTER_NAMES): Likewise.
3076 (ADDITIONAL_REGISTER_NAMES): Likewise.
3077 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3078 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3079 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3080 * config/rs6000/htm.md: New file.
3081 * config/rs6000/htmintrin.h: New file.
3082 * config/rs6000/htmxlintrin.h: New file.
3083
87dd8ab0
MS
30842013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3085
3086 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3087 Define SYMBOL_TINY_GOT, update comment.
3088 * config/aarch64/aarch64.c
3089 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3090 (aarch64_expand_mov_immediate): Likewise.
3091 (aarch64_print_operand): Likewise.
3092 (aarch64_classify_symbol): Likewise.
3093 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3094 (ldr_got_tiny): Define.
3095
52d676b6
TG
30962013-07-13 Tobias Grosser <tobias@grosser.es>
3097
3098 PR tree-optimization/54094
3099 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3100 scheduling dimension for the parallelism check from the polyhedral
3101 information in the AST.
3102 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3103 in 2D + 1 form.
3104
286e8fc1
JM
31052013-07-13 Jason Merrill <jason@redhat.com>
3106
3107 * print-tree.c (debug_vec_tree): Use debug_raw.
3108 (debug_raw (vec<tree, va_gc> &)): New.
3109 (debug_raw (vec<tree, va_gc> *)): New.
3110 * tree.h: Declare them.
3111
e43257e8
BC
31122013-07-13 Bin Cheng <bin.cheng@arm.com>
3113
3114 * ifcvt.c (ifcvt_after_combine): New static variable.
3115 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3116 for size.
3117 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3118 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3119 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3120
da582d46
MR
31212013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3122
3123 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3124
18b0ea8f
MM
31252013-07-12 Michael Matz <matz@suse.de>
3126
3127 PR middle-end/55771
3128 * convert.c (convert_to_real): Reject non-float inner types.
3129
e4f0f84d
TB
31302013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3131
3132 * config/aarch64/aarch64-protos.h
3133 (aarch64_simd_immediate_valid_for_move): Remove.
3134 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3135 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3136 cases.
3137 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3138
76500022
SE
31392013-07-11 Steve Ellcey <sellcey@mips.com>
3140
3141 * config/mips/mips.c (mips_conditional_register_usage): Do not
3142 use t[0-7] registers in MIPS16 mode when optimizing for size.
3143
78039734
ST
31442013-07-11 Sriraman Tallam <tmsriram@google.com>
3145
3146 * config/i386/i386.c (dispatch_function_versions): Fix array
3147 indexing of function_version_info to match actual_versions.
3148
7b24b675
TJ
31492013-07-11 Teresa Johnson <tejohnson@google.com>
3150
3151 * vec.h (struct va_gc): Move release out-of-line.
3152 (va_gc::release): Call ggc_free on released vec.
3153
85d4cbb8
UW
31542013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3155
3156 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3157 Require GOT register as additional operand in UNSPEC.
3158 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3159 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3160 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3161 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3162 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3163 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3164 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3165
52b3f9e2
GJL
31662013-07-11 Georg-Johann Lay <avr@gjlay.de>
3167
3168 PR target/57631
3169 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3170 name seen by assembler/linker rather if available.
3171
3f3e266e
AS
31722013-07-11 Andreas Schwab <schwab@suse.de>
3173
3174 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3175
21231ca6
VM
31762013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3177
371e77e3 3178 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 3179
d331d374
JM
31802013-07-10 Joseph Myers <joseph@codesourcery.com>
3181
673c2f63
JM
3182 * doc/tm.texi.in: Move hook documentation to ....
3183 * target.def: ... here.
3184
d331d374
JM
3185 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3186 text on @hook line.
3187 * doc/tm.texi: Regenerate.
3188
dfeadaa0
PC
31892013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3190
3191 PR c++/57869
3192 * doc/invoke.texi: Document -Wconditionally-supported.
3193
bb6f2bac
GJL
31942013-07-10 Georg-Johann Lay <avr@gjlay.de>
3195
3196 PR target/57844
3197 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3198 of my_fp.
3199
5219b232
GJL
32002013-07-10 Georg-Johann Lay <avr@gjlay.de>
3201
3202 PR target/57506
3203 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3204 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3205 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3206 Remove duplicate devices.
3207 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3208 * config/avr/t-multilib: Regenerate.
3209 * config/avr/avr-tables.opt: Regenerate.
3210 * doc/avr-mmcu.texi: Regenerate.
3211
283045c0
GJL
32122013-07-10 Georg-Johann Lay <avr@gjlay.de>
3213
3214 PR target/56987
3215 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3216
ceb2d59f 32172013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 3218
ceb2d59f
GS
3219 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3220 the cost of MULT when optimizing for size.
3221
055e0a99
JBG
32222013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3223
3224 * config/cr16/cr16-protos.h: Don't include target.h.
3225
ef43e868
JM
32262013-07-09 Joseph Myers <joseph@codesourcery.com>
3227
3228 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3229 adjust register size for TDmode and TFmode for VSX registers.
3230
4a283fdf
KT
32312013-07-08 Kai Tietz <ktietz@redhat.com>
3232
3233 PR target/56892
3234 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3235 hook_bool_const_tree_true.
3236
2cf4c39e
AK
32372013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3238
3239 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3240 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3241 * config/s390/s390.md: Define FPR*_REGNUM constants.
3242 Fix FPR2_REGNUM constant (18 -> 17).
3243 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3244 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3245 ("trunc<BFP:mode><DFP_ALL:mode>2")
3246 ("trunc<DFP_ALL:mode><BFP:mode>2")
3247 ("*extend<BFP:mode><DFP_ALL:mode>2")
3248 ("*extend<DFP_ALL:mode><BFP:mode>2")
3249 ("extend<BFP:mode><DFP_ALL:mode>2")
3250 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3251 FPR4_REGNUM.
3252
e6ac0270 32532013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 3254
6a2fa4b2 3255 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 3256
b89b22fc
AK
32572013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3258
3259 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3260 and cfun_fpr_bit_p to cfun_fpr_save_p.
3261 (s390_frame_area, s390_register_info, s390_frame_info)
3262 (s390_emit_prologue, s390_emit_epilogue)
3263 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3264 register numbers.
3265 * config/s390/s390.h: Define *_REGNUM macros for floating point
3266 register numbers.
3267
07f398aa
EB
32682013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3269
3270 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3271
75fefa91
PCC
32722013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3273
3274 PR rtl-optimization/57786
3275 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3276 and break out of the loop when it is set to false.
3277
046f1eee
JJ
32782013-07-08 Jakub Jelinek <jakub@redhat.com>
3279
8140c065
JJ
3280 PR target/57819
3281 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3282 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3283 (const_int 63)) 0)).
3284 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3285 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3286 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3287
046f1eee
JJ
3288 PR rtl-optimization/57829
3289 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3290 mask bits outside of mode are just sign-extension from mode to HWI.
3291
930b700b
MZ
32922013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3293
3294 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3295 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3296 adjust_address instead of change_address to keep info about alignment.
3297 (emit_strmov): Remove.
3298 (emit_memmov): New function.
3299 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3300 (expand_movmem_epilogue): Likewise and return updated rtx for
3301 destination.
3302 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3303 destination and source.
3304 (decide_alignment): Refactor, handle vector_loop.
3305 (ix86_expand_movmem): Likewise.
3306 (ix86_expand_setmem): Likewise.
3307 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3308
7b9d1bd8
UB
33092013-07-07 Uros Bizjak <ubizjak@gmail.com>
3310
3311 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3312 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
3313
eabb5f48
UB
33142013-07-06 Uros Bizjak <ubizjak@gmail.com>
3315
3316 * config/i386/sse.md (sse_movlhps): Change alternative 3
3317 of operand 2 to "m".
3318
33192013-07-06 Uros Bizjak <ubizjak@gmail.com>
3320
3321 PR target/57807
3322 * config/i386/sse.md (iptr): New mode attribute.
3323 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
3324 (<sse>_vm<plusminus_insn><mode>3): Ditto.
3325 (<sse>_vmmul<mode>3): Ditto.
3326 (<sse>_vmdiv<mode>3): Ditto.
3327 (sse_vmrcpv4sf2): Ditto.
3328 (<sse>_vmsqrt<mode>2): Ditto.
3329 (sse_vmrsqrtv4sf2): Ditto.
3330 (<sse>_vm<code><mode>3): Ditto.
3331 (avx_vmcmp<mode>3): Ditto.
3332 (<sse>_vmmaskcmp<mode>3): Ditto.
3333 (<sse>_comi): Ditto.
3334 (<sse>_ucomi): Ditto.
3335 (*xop_vmfrcz_<mode>): Ditto.
3336 (*fmai_fmadd_<mode>): Ditto.
3337 (*fmai_fmsub_<mode>): Ditto.
3338 (*fmai_fnmadd_<mode>): Ditto.
3339 (*fmai_fnmsub_<mode>): Ditto.
3340 (*fma4i_vmfmadd_<mode>): Ditto.
3341 (*fma4i_vmfmsub_<mode>): Ditto.
3342 (*fma4i_vmfnmadd_<mode>): Ditto.
3343 (*fma4i_vmfnmsub_<mode>): Ditto.
3344 (*xop_vmfrcz_<mode>): Ditto.
3345 (sse_cvtps2pi): Ditto.
3346 (sse_cvttps2pi): Ditto.
3347 (sse_cvtss2si): Ditto.
3348 (sse_cvtss2si_2): Ditto.
3349 (sse_cvtss2siq_2): Ditto.
3350 (sse_cvttss2si): Ditto.
3351 (sse_cvttss2siq): Ditto.
3352 (sse_cvtsd2si): Ditto.
3353 (sse_cvtsd2si_2): Ditto.
3354 (sse_cvtsd2siq_2): Ditto.
3355 (sse_cvttsd2si): Ditto.
3356 (sse_cvttsd2siq): Ditto.
3357 (sse_cvtsd2ss): Ditto.
3358 (sse_cvtss2sd): Ditto.
3359 (avx2_pbroadcast<mode>): Ditto.
3360 (avx2_pbroadcast<mode>_1): Ditto.
3361 (*avx_vperm_broadcast_v4sf): Ditto.
3362
3363 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
3364 (sse_movlhps): Ditto.
3365 (sse_storehps): Ditto.
3366 (sse_loadhps): Ditto.
3367 (sse_storelps): Ditto.
3368 (sse_loadlps): Ditto.
3369 (*vec_concatv4sf): Ditto.
3370 (*vec_interleave_highv2df): Ditto.
3371 (*vec_interleave_lowv2df): Ditto.
3372 (*vec_extractv2df_1_sse): Ditto.
3373 (*vec_extractv2df_0_sse): Ditto.
3374 (sse2_storelpd): Ditto.
3375 (sse2_loadlpd): Ditto.
3376 (sse2_movsd): Ditto.
3377 (*vec_concatv4si): Ditto.
3378 (vec_concatv2di): Ditto.
3379
3380 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
3381 for Intel asm dialect.
3382 (mmx_punpcklwd): Ditto.
3383 (mmx_punpckldq): Ditto.
3384
3385 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
3386 for intel assembler dialect.
3387
1f6eac90
JJ
33882013-07-06 Jakub Jelinek <jakub@redhat.com>
3389
3390 PR target/29776
3391 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
3392 for BUILT_IN_C{LZ,LRSB}*.
3393 * tree.h (CASE_INT_FN): Add FN##IMAX case.
3394 * tree-vrp.c (extract_range_basic): Handle
3395 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
3396 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
3397 fall thru to code calling set_value*.
3398 * builtins.c (expand_builtin): Remove *IMAX cases.
3399 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
3400 if width is bigger than 2*HWI.
3401
2b778c9d
VM
34022013-07-05 Vladimir Makarov <vmakarov@redhat.com>
3403
3404 PR rtl-optimization/55342
3405 * lra-int.h (lra_subreg_reload_pseudos): New.
3406 * lra.c: Add undoing optional reloads to the block diagram.
3407 (lra_subreg_reload_pseudos): New.
3408 (lra_optional_reload_pseudos): Change comments.
3409 (lra): Init and clear lra_subreg_reload_pseudos. Clear
3410 lra_optional_reload_pseudos after undo transformations.
3411 * lra-assigns.c (pseudo_prefix_title): New.
3412 (lra_setup_reg_renumber): Use it.
3413 (spill_for): Ditto. Check subreg reload pseudos too.
3414 (assign_by_spills): Consider subreg reload pseudos too.
3415 * lra-constraints.c (simplify_operand_subreg): Use
3416 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
3417 (curr_insn_transform): Recognize and do optional reloads.
3418 (undo_optional_reloads): New.
3419 (lra_undo_inheritance): Call undo_optional_reloads.
3420
207156e4
TQ
34212013-07-05 Thomas Quinot <quinot@adacore.com>
3422
3423 * tree-complex.c (expand_complex_operations_1): Fix typo.
3424
8990e73a
TB
34252013-07-04 Tejas Belagod <tejas.belagod@arm.com>
3426
3427 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
3428 (tune_params): New member 'const vec_costs'.
3429 * config/aarch64/aarch64.c (generic_vector_cost): New.
3430 (generic_tunings): New member 'generic_vector_cost'.
3431 (aarch64_builtin_vectorization_cost): New.
3432 (aarch64_add_stmt_cost): New.
3433 (TARGET_VECTORIZE_ADD_STMT_COST): New.
3434 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
3435
f7bc421d
JJ
34362013-07-03 Jakub Jelinek <jakub@redhat.com>
3437
3438 PR target/57777
3439 * config/i386/predicates.md (vsib_address_operand): Disallow
3440 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
3441
e90d1568
HPN
34422013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
3443
3444 PR middle-end/55030
3445 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
3446 expand_builtin_setjmp_receiver.
3447 (expand_label): Adjust, call expand_builtin_setjmp_receiver
3448 with NULL for the label parameter.
3449 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
3450 the frame-pointer. Adjust comments.
3451 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
3452 only if LABEL is non-NULL.
3453
cbd7413b
YZ
34542013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
3455
3456 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
3457 (ARM_ABI_AAPCS64): Ditto.
3458 (arm_abi): Ditto.
3459 (ARM_DEFAULT_ABI): Ditto.
3460
dec11868
JG
34612013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
3462
3463 * config/aarch64/aarch64-builtins.c
3464 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
3465 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
3466 (st1): Likewise.
3467 * config/aarch64/aarch64-simd.md
3468 (aarch64_ld1<VALL:mode>): New.
3469 (aarch64_st1<VALL:mode>): Likewise.
3470 * config/aarch64/arm_neon.h
3471 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
3472
786fb9b6
ST
34732013-07-02 Sriraman Tallam <tmsriram@google.com>
3474
0ba6aff3
UB
3475 * config/i386/i386.c (gate_insert_vzeroupper): Check if
3476 target ISA is AVX.
786fb9b6
ST
3477 (ix86_option_override_internal):Turn on all -mavx target flags by
3478 default as they are dependent on AVX anyway.
3479
4b043553
CC
34802013-07-02 Cary Coutant <ccoutant@google.com>
3481
3482 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
3483 deterministic hash.
3484 (loc_checksum_ordered): Likewise.
3485 (hash_loc_operands): Remove inline keyword.
3486
c4e87a13
JJ
34872013-07-02 Jakub Jelinek <jakub@redhat.com>
3488
3489 PR tree-optimization/57741
3490 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
3491 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
3492 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
3493 Allow REAL_CST step_exprs if flag_associative_math.
3494 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
3495
26366d28
IB
34962013-07-02 Ian Bolton <ian.bolton@arm.com>
3497
0ba6aff3 3498 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 3499
2879bb2b
IB
35002013-07-02 Ian Bolton <ian.bolton@arm.com>
3501
3502 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
3503
afc5e8a6
KT
35042013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3505
3506 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
3507 encoding.
3508 (iorsi3_insn): Likewise.
3509 (arm_xorsi3): Likewise.
3510
31f8442b
SN
35112013-07-01 Sofiane Naci <sofiane.naci@arm.com>
3512
3513 * arm.md (attribute "wtype"): Delete. Move attribute values from here
3514 to ...
3515 (attribute "type"): ... here, and prefix with "wmmx_".
3516 (attribute "core_cycles"): Update for attribute changes.
3517 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
3518 (tbcstv4hi): Likewise.
3519 (tbcstv2si): Likewise.
3520 (iwmmxt_iordi3): Likewise.
3521 (iwmmxt_xordi3): Likewise.
3522 (iwmmxt_anddi3): Likewise.
3523 (iwmmxt_nanddi3): Likewise.
3524 (iwmmxt_arm_movdi): Likewise.
3525 (iwmmxt_movsi_insn): Likewise.
3526 (mov<mode>_internal): Likewise.
3527 (and<mode>3_iwmmxt): Likewise.
3528 (ior<mode>3_iwmmxt): Likewise.
3529 (xor<mode>3_iwmmxt): Likewise.
3530 (add<mode>3_iwmmxt): Likewise.
3531 (ssaddv8qi3): Likewise.
3532 (ssaddv4hi3): Likewise.
3533 (ssaddv2si3): Likewise.
3534 (usaddv8qi3): Likewise.
3535 (usaddv4hi3): Likewise.
3536 (usaddv2si3): Likewise.
3537 (sub<mode>3_iwmmxt): Likewise.
3538 (sssubv8qi3): Likewise.
3539 (sssubv4hi3): Likewise.
3540 (sssubv2si3): Likewise.
3541 (ussubv8qi3): Likewise.
3542 (ussubv4hi3): Likewise.
3543 (ussubv2si3): Likewise.
3544 (mulv4hi3_iwmmxt): Likewise.
3545 (smulv4hi3_highpart): Likewise.
3546 (umulv4hi3_highpart): Likewise.
3547 (iwmmxt_wmacs): Likewise.
3548 (iwmmxt_wmacsz): Likewise.
3549 (iwmmxt_wmacu): Likewise.
3550 (iwmmxt_wmacuz): Likewise.
3551 (iwmmxt_clrdi): Likewise.
3552 (iwmmxt_clrv8qi): Likewise.
3553 (iwmmxt_clr4hi): Likewise.
3554 (iwmmxt_clr2si): Likewise.
3555 (iwmmxt_uavgrndv8qi3): Likewise.
3556 (iwmmxt_uavgrndv4hi3): Likewise.
3557 (iwmmxt_uavgv8qi3): Likewise.
3558 (iwmmxt_uavgv4hi3): Likewise.
3559 (iwmmxt_tinsrb): Likewise.
3560 (iwmmxt_tinsrh): Likewise.
3561 (iwmmxt_tinsrw): Likewise.
3562 (iwmmxt_textrmub): Likewise.
3563 (iwmmxt_textrmsb): Likewise.
3564 (iwmmxt_textrmuh): Likewise.
3565 (iwmmxt_textrmsh): Likewise.
3566 (iwmmxt_textrmw): Likewise.
3567 (iwmxxt_wshufh): Likewise.
3568 (eqv8qi3): Likewise.
3569 (eqv4hi3): Likewise.
3570 (eqv2si3): Likewise.
3571 (gtuv8qi3): Likewise.
3572 (gtuv4hi3): Likewise.
3573 (gtuv2si3): Likewise.
3574 (gtv8qi3): Likewise.
3575 (gtv4hi3): Likewise.
3576 (gtv2si3): Likewise.
3577 (smax<mode>3_iwmmxt): Likewise.
3578 (umax<mode>3_iwmmxt): Likewise.
3579 (smin<mode>3_iwmmxt): Likewise.
3580 (umin<mode>3_iwmmxt): Likewise.
3581 (iwmmxt_wpackhss): Likewise.
3582 (iwmmxt_wpackwss): Likewise.
3583 (iwmmxt_wpackdss): Likewise.
3584 (iwmmxt_wpackhus): Likewise.
3585 (iwmmxt_wpackwus): Likewise.
3586 (iwmmxt_wpackdus): Likewise.
3587 (iwmmxt_wunpckihb): Likewise.
3588 (iwmmxt_wunpckihh): Likewise.
3589 (iwmmxt_wunpckihw): Likewise.
3590 (iwmmxt_wunpckilb): Likewise.
3591 (iwmmxt_wunpckilh): Likewise.
3592 (iwmmxt_wunpckilw): Likewise.
3593 (iwmmxt_wunpckehub): Likewise.
3594 (iwmmxt_wunpckehuh): Likewise.
3595 (iwmmxt_wunpckehuw): Likewise.
3596 (iwmmxt_wunpckehsb): Likewise.
3597 (iwmmxt_wunpckehsh): Likewise.
3598 (iwmmxt_wunpckehsw): Likewise.
3599 (iwmmxt_wunpckelub): Likewise.
3600 (iwmmxt_wunpckeluh): Likewise.
3601 (iwmmxt_wunpckeluw): Likewise.
3602 (iwmmxt_wunpckelsb): Likewise.
3603 (iwmmxt_wunpckelsh): Likewise.
3604 (iwmmxt_wunpckelsw): Likewise.
3605 (ror<mode>3): Likewise.
3606 (ashr<mode>3_iwmmxt): Likewise.
3607 (lshr<mode>3_iwmmxt): Likewise.
3608 (ashl<mode>3_iwmmxt): Likewise.
3609 (ror<mode>3_di): Likewise.
3610 (ashr<mode>3_di): Likewise.
3611 (lshr<mode>3_di): Likewise.
3612 (ashl<mode>3_di): Likewise.
3613 (iwmmxt_wmadds): Likewise.
3614 (iwmmxt_wmaddu): Likewise.
3615 (iwmmxt_tmia): Likewise.
3616 (iwmmxt_tmiaph): Likewise.
3617 (iwmmxt_tmiabb): Likewise.
3618 (iwmmxt_tmiatb): Likewise.
3619 (iwmmxt_tmiabt): Likewise.
3620 (iwmmxt_tmiatt): Likewise.
3621 (iwmmxt_tmovmskb): Likewise.
3622 (iwmmxt_tmovmskh): Likewise.
3623 (iwmmxt_tmovmskw): Likewise.
3624 (iwmmxt_waccb): Likewise.
3625 (iwmmxt_wacch): Likewise.
3626 (iwmmxt_waccw): Likewise.
3627 (iwmmxt_waligni): Likewise.
3628 (iwmmxt_walignr): Likewise.
3629 (iwmmxt_walignr0): Likewise.
3630 (iwmmxt_walignr1): Likewise.
3631 (iwmmxt_walignr2): Likewise.
3632 (iwmmxt_walignr3): Likewise.
3633 (iwmmxt_wsadb): Likewise.
3634 (iwmmxt_wsadh): Likewise.
3635 (iwmmxt_wsadbz): Likewise.
3636 (iwmmxt_wsadhz): Likewise.
3637 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
3638 (iwmmxt_wabsdiffb): Likewise.
3639 (iwmmxt_wabsdiffh): Likewise.
3640 (iwmmxt_wabsdiffw): Likewise.
3641 (iwmmxt_waddsubhx): Likewise
3642 (iwmmxt_wsubaddhx): Likewise.
3643 (addc<mode>3): Likewise.
3644 (iwmmxt_avg4): Likewise.
3645 (iwmmxt_avg4r): Likewise.
3646 (iwmmxt_wmaddsx): Likewise.
3647 (iwmmxt_wmaddux): Likewise.
3648 (iwmmxt_wmaddsn): Likewise.
3649 (iwmmxt_wmaddun): Likewise.
3650 (iwmmxt_wmulwsm): Likewise.
3651 (iwmmxt_wmulwum): Likewise.
3652 (iwmmxt_wmulsmr): Likewise.
3653 (iwmmxt_wmulumr): Likewise.
3654 (iwmmxt_wmulwsmr): Likewise.
3655 (iwmmxt_wmulwumr): Likewise.
3656 (iwmmxt_wmulwl): Likewise.
3657 (iwmmxt_wqmulm): Likewise.
3658 (iwmmxt_wqmulwm): Likewise.
3659 (iwmmxt_wqmulmr): Likewise.
3660 (iwmmxt_wqmulwmr): Likewise.
3661 (iwmmxt_waddbhusm): Likewise.
3662 (iwmmxt_waddbhusl): Likewise.
3663 (iwmmxt_wqmiabb): Likewise.
3664 (iwmmxt_wqmiabt): Likewise.
3665 (iwmmxt_wqmiatb): Likewise.
3666 (iwmmxt_wqmiatt): Likewise.
3667 (iwmmxt_wqmiabbn): Likewise.
3668 (iwmmxt_wqmiabtn): Likewise.
3669 (iwmmxt_wqmiatbn): Likewise.
3670 (iwmmxt_wqmiattn): Likewise.
3671 (iwmmxt_wmiabb): Likewise.
3672 (iwmmxt_wmiabt): Likewise.
3673 (iwmmxt_wmiatb): Likewise.
3674 (iwmmxt_wmiatt): Likewise.
3675 (iwmmxt_wmiabbn): Likewise.
3676 (iwmmxt_wmiabtn): Likewise.
3677 (iwmmxt_wmiatbn): Likewise.
3678 (iwmmxt_wmiattn): Likewise.
3679 (iwmmxt_wmiawbb): Likewise.
3680 (iwmmxt_wmiawbt): Likewise.
3681 (iwmmxt_wmiawtb): Likewise.
3682 (iwmmxt_wmiawtt): Likewise.
3683 (iwmmxt_wmiawbbn): Likewise.
3684 (iwmmxt_wmiawbtn): Likewise.
3685 (iwmmxt_wmiawtbn): Likewise.
3686 (iwmmxt_wmiawttn): Likewise.
3687 (iwmmxt_wmerge): Likewise.
3688 (iwmmxt_tandc<mode>3): Likewise.
3689 (iwmmxt_torc<mode>3): Likewise.
3690 (iwmmxt_torvsc<mode>3): Likewise.
3691 (iwmmxt_textrc<mode>3): Likewise.
3692 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
3693 (wmmxt_pack): Likewise.
3694 (wmmxt_mult_c1): Likewise.
3695 (wmmxt_mult_c2): Likewise.
3696 (wmmxt_alu_c1): Likewise.
3697 (wmmxt_alu_c2): Likewise.
3698 (wmmxt_alu_c3): Likewise.
3699 (wmmxt_transfer_c1): Likewise.
3700 (wmmxt_transfer_c2): Likewise.
3701 (wmmxt_transfer_c3): Likewise.
3702 (marvell_f_iwmmxt_wstr): Likewise.
3703 (marvell_f_iwmmxt_wldr): Likewise.
3704
bdb7bf8a
YZ
37052013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
3706
0ba6aff3 3707 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 3708
e6631d38
VM
37092013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3710
3711 Revert:
3712 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3713 * lra-constraints.c (need_for_split_p): Check call used hard regs
3714 living through calls.
3715
3716 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
3717 call used regs for call insn.
3718
53c1275b
JJ
37192013-06-28 Jakub Jelinek <jakub@redhat.com>
3720
3721 PR target/57736
0ba6aff3
UB
3722 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
3723 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 3724
1141ed3f
BI
37252013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
3726
3727 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 3728
e32e4c4a
VM
37292013-06-28 Vladimir Makarov <vmakarov@redhat.com>
3730
3731 * lra-constraints.c (need_for_split_p): Check call used hard regs
3732 living through calls.
3733
f161bfd3
MM
37342013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
3735
3736 PR target/57744
3737 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
3738 to tie with any other modes. Eliminate Altivec vector mode tests,
3739 since these are a subset of ALTIVEC or VSX vector modes. Simplify
3740 code, to return 0 if testing MODE2 for a condition, if we've
3741 already tested MODE1 for the same condition.
3742
7eda14e1
MS
37432013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3744
3745 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
3746 layout.
3747
c822f852
MS
37482013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3749
3750 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3751 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
3752
da4f13a4
MS
37532013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
3754
0ba6aff3
UB
3755 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
3756 Define.
da4f13a4
MS
3757 (aarch64_symbolic_constant_p): Remove.
3758 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
3759 static. Fix line length and white space.
3760 (aarch64_symbolic_constant_p): Remove.
3761 * config/aarch64/predicates.md (aarch64_valid_symref):
3762 Use aarch64_classify_symbol_expression.
3763
956a95a5
KT
37642013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3765
3766 * config/arm/constraints.md (Ts): New constraint.
3767 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
3768 16-bit encodings.
3769 (compare_scc): Use "Ts" constraint for operand 0.
3770 (ior_scc_scc): Likewise.
3771 (and_scc_scc): Likewise.
3772 (and_scc_scc_nodom): Likewise.
3773 (ior_scc_scc_cmp): Likewise for operand 7.
3774 (and_scc_scc_cmp): Likewise.
3775 * config/arm/thumb2.md (thumb2_movsi_insn):
3776 Add alternatives for 16-bit encodings.
3777 (thumb2_movhi_insn): Likewise.
3778 (thumb2_movsicc_insn): Likewise.
3779 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
3780 (thumb2_negscc): Use "Ts" constraint.
3781 Move mvn instruction outside cond_exec block.
3782 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
3783 for 16-bit encodings.
3784
31dfce10
KT
37852013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3786
3787 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
3788 encoding.
3789 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
3790 (mulsi3subsi): Likewise.
3791 (mulsidi3adddi): Likewise.
3792 (mulsidi3_v6): Likewise.
3793 (umulsidi3_v6): Likewise.
3794 (umulsidi3adddi_v6): Likewise.
3795 (smulsi3_highpart_v6): Likewise.
3796 (umulsi3_highpart_v6): Likewise.
3797 (mulhisi3tb): Likewise.
3798 (mulhisi3bt): Likewise.
3799 (mulhisi3tt): Likewise.
3800 (maddhisi4): Likewise.
3801 (maddhisi4tb): Likewise.
3802 (maddhisi4tt): Likewise.
3803 (maddhidi4): Likewise.
3804 (maddhidi4tb): Likewise.
3805 (maddhidi4tt): Likewise.
3806 (zeroextractsi_compare0_scratch): Likewise.
3807 (insv_zero): Likewise.
3808 (insv_t2): Likewise.
3809 (anddi_notzesidi_di): Likewise.
3810 (anddi_notsesidi_di): Likewise.
3811 (andsi_notsi_si): Likewise.
3812 (iordi_zesidi_di): Likewise.
3813 (xordi_zesidi_di): Likewise.
3814 (andsi_iorsi3_notsi): Likewise.
3815 (smax_0): Likewise.
3816 (smax_m1): Likewise.
3817 (smin_0): Likewise.
3818 (not_shiftsi): Likewise.
3819 (unaligned_loadsi): Likewise.
3820 (unaligned_loadhis): Likewise.
3821 (unaligned_loadhiu): Likewise.
3822 (unaligned_storesi): Likewise.
3823 (unaligned_storehi): Likewise.
3824 (extv_reg): Likewise.
3825 (extzv_t2): Likewise.
3826 (divsi3): Likewise.
3827 (udivsi3): Likewise.
3828 (arm_zero_extendhisi2addsi): Likewise.
3829 (arm_zero_extendqisi2addsi): Likewise.
3830 (compareqi_eq0): Likewise.
3831 (arm_extendhisi2_v6): Likewise.
3832 (arm_extendqisi2addsi): Likewise.
3833 (arm_movt): Likewise.
3834 (thumb2_ldrd): Likewise.
3835 (thumb2_ldrd_base): Likewise.
3836 (thumb2_ldrd_base_neg): Likewise.
3837 (thumb2_strd): Likewise.
3838 (thumb2_strd_base): Likewise.
3839 (thumb2_strd_base_neg): Likewise.
3840 (arm_negsi2): Add alternative for 16-bit encoding.
3841 (arm_one_cmplsi2): Likewise.
3842
1572e697
KT
38432013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3844
3845 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
3846 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
3847 (movdfcc): Likewise.
3848 * config/arm/vfp.md (*thumb2_movsf_vfp):
3849 Disable predication for arm_restrict_it.
3850 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
3851 (*thumb2_movdfcc_vfp): Likewise.
3852 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
3853 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
3854 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
3855 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
3856 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
3857 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
3858 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
3859 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
3860 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
3861 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
3862 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
3863 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
3864 Disable predication for arm_restrict_it.
3865
ebf8f0ea
KY
38662013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
3867
3868 * config/i386/bmiintrin.h (_bextr_u32): New.
3869 (_bextr_u64): Ditto.
3870
a290fcda
RS
38712013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
3872
3873 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
3874 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
3875 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
3876 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
3877 * config/mips/n32-elf.h: ...this new file.
3878
ea169996
MG
38792013-06-27 Marc Glisse <marc.glisse@inria.fr>
3880
3881 PR target/57224
3882 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
3883 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
3884
8e932114
CM
38852013-06-27 Catherine Moore <clm@codesourcery.com>
3886
3887 * config/mips/mips-tables.opt: Regenerate.
3888 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 3889 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 3890 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 3891
2444244e
JJ
38922013-06-27 Jakub Jelinek <jakub@redhat.com>
3893
cd06d2a2
JJ
3894 PR target/57623
3895 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
3896 constraints of operand 1 and 2.
3897
2444244e
JJ
3898 PR target/57623
3899 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
3900 to match RTL canonicalization. Swap predicates and
3901 constraints of operand 1 and 2.
3902
2f259720
VM
39032013-06-27 Vladimir Makarov <vmakarov@redhat.com>
3904
0ba6aff3
UB
3905 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
3906 Process OP_INOUT regs for splitting too.
2f259720 3907
fef4d2b3
JJ
39082013-06-27 Jakub Jelinek <jakub@redhat.com>
3909
3910 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
3911 decl before the loop, initialize to NULL.
3912 (vectorizable_load): Initialize ptr_incr to NULL.
3913
79ee9826
MJ
39142013-06-27 Martin Jambor <mjambor@suse.cz>
3915
3916 PR lto/57208
3917 * ipa-ref.h (ipa_maybe_record_reference): Declare.
3918 * ipa-ref.c (ipa_maybe_record_reference): New function.
3919 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
3920 * ipa-cp.c (create_specialized_node): Record potential references from
3921 aggvals.
3922 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
3923
e18b4a81
YZ
39242013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
3925
3926 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
3927 parameter 'mode' of type 'enum machine_mode mode'; change to pass
3928 'mode' to force_reg.
3929 (aarch64_add_offset): Update calls to aarch64_force_temporary.
3930 (aarch64_expand_mov_immediate): Likewise.
3931
9c023bf0
YZ
39322013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
3933
3934 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
3935 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
3936
5a3fe9b6
AK
39372013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3938
3939 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
3940 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
3941 (struct machine_function): Add tbegin_p.
3942 (s390_canonicalize_comparison): Fold CC mode compares to
3943 conditional jump if possible.
3944 (s390_emit_jump): Return the emitted jump.
3945 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
3946 Handle CCRAWmode compares.
3947 (s390_option_override): Default to -mhtm if available.
3948 (s390_reg_clobbered_rtx): Handle floating point regs as well.
3949 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
3950 FPRs instead of df_regs_ever_live_p.
3951 (s390_optimize_nonescaping_tx): New function.
3952 (s390_init_frame_layout): Extend clobbered_regs array to cover
3953 FPRs as well.
3954 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
3955 (s390_expand_tbegin): New function.
3956 (enum s390_builtin): New enum definition.
3957 (code_for_builtin): New array definition.
3958 (s390_init_builtins): New function.
3959 (s390_expand_builtin): New function.
3960 (TARGET_INIT_BUILTINS): Define.
3961 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 3962 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK
3963 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
3964 (s390_alc_comparison): Likewise.
3965 * config/s390/s390-modes.def: Add CCRAWmode.
3966 * config/s390/s390.h (processor_flags): Add PF_TX.
3967 (TARGET_CPU_HTM): Define macro.
3968 (TARGET_HTM): Define macro.
3969 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
3970 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
3971 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
3972 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 3973 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK
3974 (TBEGIN_MASK, TBEGINC_MASK): New constants.
3975 ("*cc_to_int"): Move up.
3976 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
3977 constants other than 0.
3978 ("*ccraw_to_int"): New insn and splitter definition.
3979 ("tbegin", "tbegin_nofloat", "tbegin_retry")
3980 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
3981 ("tx_assist"): New expander.
3982 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
3983 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
3984 * config/s390/s390.opt: Add -mhtm option.
3985 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
3986 * config/s390/htmxlintrin.h: New file.
3987 * config/s390/htmintrin.h: New file.
3988 * config/s390/s390intrin.h: New file.
3989 * doc/extend.texi: Document htm builtins.
3990 * config.gcc: Add the new header files to extra_headers.
3991
9e216629
TS
39922013-06-26 Thomas Schwinge <thomas@codesourcery.com>
3993
3994 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 3995 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 3996
fd8c65e7
MM
39972013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
3998 Pat Haugen <pthaugen@us.ibm.com>
3999 Peter Bergner <bergner@vnet.ibm.com>
4000
4001 * config/rs6000/power8.md: New.
4002 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4003 setting for power8 entry.
4004 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4005 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4006 test for Power4/Power5 only.
4007 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4008 support.
4009 (force_new_group): Adjust comment.
4010 * config/rs6000/rs6000.md: Include power8.md.
4011
b24a2ce5
GY
40122013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4013
4014 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4015 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4016 declaration.
4017 (tune_params): Update comment.
4018 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4019 (arm_max_conditional_execute): New function.
4020 (thumb2_final_prescan_insn): Use max_insn_skipped and
4021 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4022
f6fdeb62
JJ
40232013-06-25 Jakub Jelinek <jakub@redhat.com>
4024
4025 PR tree-optimization/57705
4026 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4027 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 4028 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ
4029 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4030
c13bc3d9
MJ
40312013-06-25 Martin Jambor <mjambor@suse.cz>
4032
4033 PR middle-end/57670
4034 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4035 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4036 calls in the dump.
4037 (ipa_note_param_call): Initialize member_ptr flag.
4038 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4039 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4040 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4041 (ipa_read_indirect_edge_info): Likewise.
4042
8a845901
RB
40432013-06-25 Richard Biener <rguenther@suse.de>
4044
4045 PR middle-end/56977
4046 * passes.c (init_optimization_passes): Move pass_fold_builtins
4047 and pass_dce earlier with -Og.
4048
33be0bec
EB
40492013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4050
4051 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4052 <BIT_FIELD_REF>: Remove trailing TAB.
4053 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4054 remove blank line.
4055
5fe8e757
MJ
40562013-06-24 Martin Jambor <mjambor@suse.cz>
4057
4058 PR tree-optimization/57358
4059 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4060 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4061 (ipa_analyze_params_uses): Generate pessimistic info when true.
4062
44a60244
MJ
40632013-06-24 Martin Jambor <mjambor@suse.cz>
4064
4065 PR tree-optimization/57539
4066 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4067 global.inlined_to of the new node to it. All callers changed.
4068 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4069 inlining_into, pass it to cgraph_clone_node.
4070 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4071 ipa_free_edge_args_substructures.
4072 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4073 rdesc linked list. Do not assert rdesc edges have inlined caller.
4074 Assert we have found an rdesc in the rdesc list.
4075
7c5848b8
RB
40762013-06-24 Richard Biener <rguenther@suse.de>
4077
0ba6aff3 4078 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB
4079 (pointer_set_lookup): Declare.
4080 (class pointer_map): New template class implementing a
4081 generic pointer to T map.
4082 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4083 pointer_map<T>::contains, pointer_map<T>::insert,
4084 pointer_map<T>::traverse): New functions.
4085 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4086 (pointer_set_lookup): New function.
4087 (pointer_set_contains): Use pointer_set_lookup.
4088 (pointer_set_insert): Likewise.
4089 (insert_aux): Remove.
4090 (struct pointer_map_t): Embed a pointer_set_t.
4091 (pointer_map_create): Adjust.
4092 (pointer_map_destroy): Likewise.
4093 (pointer_map_contains): Likewise.
4094 (pointer_map_insert): Likewise.
4095 (pointer_map_traverse): Likewise.
4096 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4097 pointer_map<unsigned> instead of a pointer_map_t.
4098 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4099 (streamer_tree_cache_lookup): Likewise.
4100 (streamer_tree_cache_create): Likewise.
4101 (streamer_tree_cache_delete): Likewise.
4102 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4103 pointer_map<unsigned> instead of a pointer_map_t.
4104 (lto_init_tree_ref_encoder): Adjust.
4105 (lto_destroy_tree_ref_encoder): Likewise.
4106 * lto-section-out.c (lto_output_decl_index): Likewise.
4107 (lto_record_function_out_decl_state): Likewise.
4108 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4109
3dbc97a9
RB
41102013-06-24 Richard Biener <rguenther@suse.de>
4111
4112 PR tree-optimization/57488
4113 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4114
8345965a
AM
41152013-06-24 Alan Modra <amodra@gmail.com>
4116
4117 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4118 (gen_easy_altivec_constant): Likewise.
4119 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4120 easy_vector_constant_msb): Likewise.
4121
b7df379f
JJ
41222013-06-23 Jakub Jelinek <jakub@redhat.com>
4123
4124 PR target/57688
4125 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4126 add missing return true.
4127
0bcf9a09
OE
41282013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4129
4130 PR target/52483
4131 * config/sh/predicates.md (general_extend_operand): Invoke
4132 general_movsrc_operand for memory operands.
4133 (general_movsrc_operand): Allow reg+reg addressing, do not use
4134 general_operand for memory operands.
4135
97db2bf7
ST
41362013-06-23 Sriraman Tallam <tmsriram@google.com>
4137
4138 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4139 when current target options does not apply.
4140 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4141 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB
4142 * config/i386/bmiintrin.h: Pass appropriate target
4143 attributes to header.
97db2bf7
ST
4144 * config/i386/mmintrin.h: Ditto.
4145 * config/i386/nmmintrin.h: Ditto.
4146 * config/i386/avx2intrin.h: Ditto.
4147 * config/i386/fxsrintrin.h: Ditto.
4148 * config/i386/tbmintrin.h: Ditto.
4149 * config/i386/xsaveintrin.h: Ditto.
4150 * config/i386/f16cintrin.h: Ditto.
4151 * config/i386/xtestintrin.h: Ditto.
4152 * config/i386/xsaveoptintrin.h: Ditto.
4153 * config/i386/bmi2intrin.h: Ditto.
4154 * config/i386/lzcntintrin.h: Ditto.
4155 * config/i386/smmintrin.h: Ditto.
4156 * config/i386/wmmintrin.h: Ditto.
4157 * config/i386/x86intrin.h: Remove all header include guards.
4158 * config/i386/prfchwintrin.h: Ditto.
4159 * config/i386/pmmintrin.h: Ditto.
4160 * config/i386/tmmintrin.h: Ditto.
4161 * config/i386/xmmintrin.h: Ditto.
4162 * config/i386/popcntintrin.h: Ditto.
4163 * config/i386/rdseedintrin.h: Ditto.
4164 * config/i386/ammintrin.h: Ditto.
4165 * config/i386/emmintrin.h: Ditto.
4166 * config/i386/immintrin.h: Remove all header include guards.
4167 * config/i386/fma4intrin.h: Ditto.
4168 * config/i386/lwpintrin.h: Ditto.
4169 * config/i386/xopintrin.h: Ditto.
4170 * config/i386/ia32intrin.h: Ditto.
4171 * config/i386/avxintrin.h: Ditto.
4172 * config/i386/rtmintrin.h: Ditto.
4173 * config/i386/fmaintrin.h: Ditto.
4174 * config/i386/mm3dnow.h: Ditto.
4175
495e6879
ST
41762013-06-22 Sriraman Tallam <tmsriram@google.com>
4177
4178 * common/config/i386/i386-common.c: Handle LZCNT.
4179
11e8f687
AK
41802013-06-22 Andi Kleen <ak@linux.intel.com>
4181
4182 * doc/extend.texi: Use __atomic_store_n instead of
4183 __atomic_store in HLE example.
4184
664dd721
OE
41852013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4186
4187 * config/sh/sh.c: Remove <cstdlib> workaround.
4188
927f908b
AK
41892013-06-21 Andi Kleen <ak@linux.intel.com>
4190
0ba6aff3 4191 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 4192
cca410b6
AK
41932013-06-21 Andi Kleen <ak@linux.intel.com>
4194
4195 * doc/extend.texi: Document that __atomic_clear and
4196 __atomic_test_and_set should only be used with bool.
4197
d84db1ed
JH
41982013-06-20 Jan Hubicka <jh@suse.cz>
4199
4200 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4201 types_same_for_odr.
4202 * tree.c (decls_same_for_odr): New function.
4203 (same_for_edr): New function.
4204 (types_same_for_odr): New function.
4205 (get_binfo_at_offset): Use it.
4206 * tree.h (types_same_for_odr): Declare.
4207
664dd721 42082013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE
4209 Jason Merrill <jason@redhat.com>
4210
4211 * system.h: Include <cstdlib> as well as <stdlib.h>.
4212
49b4ff63
UB
42132013-06-20 Uros Bizjak <ubizjak@gmail.com>
4214
4215 PR target/57655
4216 * config/i386/i386.c (construct_container): Report error if
4217 long double is used with disabled x87 float returns.
4218
7a89ad00
JH
42192013-06-20 Jan Hubicka <jh@suse.cz>
4220
4221 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4222
bd280792
JR
42232013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4224
4225 PR rtl-optimization/57425
4226 PR rtl-optimization/57569
4227 * alias.c (write_dependence_p): Remove parameters mem_mode and
4228 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4229 Changed all callers.
4230 (canon_anti_dependence): Get comments and semantics in sync.
4231 Add parameter mem_canonicalized. Changed all callers.
4232 * rtl.h (canon_anti_dependence): Update prototype.
4233
d16e9a99
RB
42342013-06-20 Richard Biener <rguenther@suse.de>
4235
4236 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4237 byte case, inline streamer_read_uchar and defer section
4238 overrun check.
4239
c3a9b91b
RB
42402013-06-20 Richard Biener <rguenther@suse.de>
4241
4242 PR tree-optimization/57584
4243 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4244 SSA names into the expanded expression that take part in
4245 abnormal coalescing.
4246
371e77e3 42472013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS
4248
4249 * gcov.c (print_usage): Handle new option.
4250 (process_args): Ditto.
4251 (get_gcov_intermediate_filename): New function.
4252 (output_intermediate_file): New function.
4253 (output_gcov_file): New function
4254 (generate_results): Handle new option.
4255 (release_function): Relase demangled name.
4256 (read_graph_file): Handle demangled name.
4257 (output_lines): Ditto.
4258 * doc/gcov.texi: Document gcov intermediate format.
4259
11f5b71a
VM
42602013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4261
4262 PR bootstrap/57604
4263 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4264 (lra_emit_add): Use the functions. Add comment about Y as an
4265 address segment.
4266
e1f9340b
DE
42672013-06-19 David Edelsohn <dje.gcc@gmail.com>
4268
4269 PR driver/57652
4270 * collect2.c (collect_atexit): New.
4271 (collect_exit): Delete.
4272 (main): Register collect_atexit with atexit.
4273 (collect_wait): Change collect_exit to exit.
4274 (do_wait): Same.
4275 * collect2.h (collect_exit): Delete.
4276 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4277
e3f9e0ac
WM
42782013-06-19 Wei Mi <wmi@google.com>
4279
4280 PR rtl-optimization/57518
4281 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4282 if regno is used in paradoxical subreg.
4283 (update_equiv_regs): Check pdx_subregs[regno] before
4284 set a reg to be equivalent with a mem.
4285
a09f1a76
MK
42862013-06-19 Matthias Klose <doko@ubuntu.com>
4287
4288 PR driver/57651
4289 * file-find.h (find_a_file): Add a mode parameter.
4290 * file-find.c (find_a_file): Likewise.
4291 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4292 with X_OK for the executables.
4293 * collect2.c (main): Call find_a_file with X_OK.
4294
7d18b0ad
SE
42952013-06-19 Steve Ellcey <sellcey@mips.com>
4296
4297 PR target/56942
4298 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4299 Use NEXT_INSN instead of next_real_insn.
4300
6a6dac52
JH
43012013-06-19 Jan Hubicka <jh@suse.cz>
4302
4303 * cgraph.h (const_value_known_p): Replace by ...
4304 (ctor_for_folding): .. this one.
4305 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4306 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4307 * expr.c (expand_expr_real_1): Likewise.
4308 (string_constant): Likewise.
4309 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4310 * ipa.c (process_references): Likewise.
4311 (symtab_remove_unreachable_nodes): Likewise.
4312 * ipa-inline-analysis.c (param_change_prob): Likewise.
4313 * gimple-fold.c (canonicalize_constructor_val): Likewise.
4314 (get_base_constructor): Likwise.
4315 * varpool.c (varpool_remove_node): Likewise.
4316 (varpool_remove_initializer): LIkewise.
4317 (dump_varpool_node): LIkwise.
4318 (const_value_known_p): Rewrite to ...
4319 (ctor_for_folding): ... this one.
4320
216c12ab
JJ
43212013-06-19 Jakub Jelinek <jakub@redhat.com>
4322
4323 PR driver/57651
4324 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
4325 PERSONALITY in $PATH derived prefixes.
4326
d53e2f99
JL
43272013-06-19 Jeff Law <law@redhat.com>
4328
ecdbf306
JL
4329 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
4330 in comment.
4331
d53e2f99
JL
4332 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
4333 (simplify_bitwise_binary): Use it to simpify certain binary ops on
4334 booleans.
4335
89e7fc04
SN
43362013-06-19 Sofiane Naci <sofiane.naci@arm.com>
4337
4338 * config/arm/vfp.md: Move VFP instruction classification documentation
4339 to ...
4340 * config/arm/arm.md: ... here. Update instruction classification
4341 documentation.
4342
5e7f6aaa
RE
43432013-06-19 Richard Earnshaw <rearnsha@arm.com>
4344
4345 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
4346 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
4347 pattern. Use more efficient sequences on ARMv5 and Thumb2.
4348
5fdc1e5d
SB
43492013-06-19 Steven Bosscher <steven@gcc.gnu.org>
4350
4351 PR target/57609
4352 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4353 with NEXT_INSN. Use tablejump_p to check for jump table data
4354 insns.
4355
e952afba
PC
43562013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
4357
4358 PR c++/56544
4359 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
4360 that now in C++ the value is correct per the C++ standards.
4361
3fd0303a
RB
43622013-06-19 Richard Biener <rguenther@suse.de>
4363
4364 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
4365 for global context.
4366
b04e4b08
AK
43672013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4368
4369 Revert:
4370 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4371
4372 PR target/57609
4373 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4374 with next_active_insn.
4375
bef8491a
ST
43762013-06-18 Sriraman Tallam <tmsriram@google.com>
4377
4378 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
4379 functions are inlined during failures to flag an error.
4380 * tree-inline.c (expand_call_inline): Allow the error to be flagged
4381 in early inline pass.
22fc7d3d 4382
c21b257d
L
43832013-06-18 H.J. Lu <hongjiu.lu@intel.com>
4384
4385 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
4386 in comments.
4387
33255ae3
JB
43882013-06-18 Julian Brown <julian@codesourcery.com>
4389
4390 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
4391 Permit virtual register pre-reload if !strict.
4392 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
4393 change.
4394 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
4395 prototype.
4396 * config/arm/neon.md (movmisalign<mode>): Use
4397 neon_perm_struct_or_reg_operand instead of
4398 neon_struct_or_register_operand.
4399 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
4400 neon_permissive_struct_operand instead of neon_struct_operand.
4401 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
4402 neon_vector_mem_operand.
4403 * config/arm/predicates.md (neon_struct_operand): Adjust call to
4404 neon_vector_mem_operand.
4405 (neon_permissive_struct_operand): New.
4406 (neon_struct_or_register_operand): Rename to...
4407 (neon_perm_struct_or_reg_operand): This. Adjust call to
4408 neon_vector_mem_operand.
4409
d579fcda
RB
44102013-06-18 Richard Biener <rguenther@suse.de>
4411
4412 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
4413 * lto-streamer.h: Include pointer-set.h.
4414 (struct lto_decl_slot): Remove.
4415 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
4416 Remove next_index entry.
4417 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4418 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4419 (lto_init_tree_ref_encoder): Adjust.
4420 (lto_destroy_tree_ref_encoder): Likewise.
4421 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4422 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4423 (lto_output_decl_index): Adjust.
4424 (lto_new_out_decl_state): Likewise.
4425 (lto_record_function_out_decl_state): Likewise.
4426 * lto-streamer-out.c (copy_function): Likewise.
4427
31ee20ba
RB
44282013-06-18 Richard Biener <rguenther@suse.de>
4429
4430 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
4431 * cgraphunit.c: Include cfgloop.h.
4432 (init_lowered_empty_function): Initialize the loop tree.
4433 (assemble_thunk): Insert new BBs into loops.
4434
ac0511f2
RB
44352013-06-18 Richard Biener <rguenther@suse.de>
4436
4437 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
4438 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
4439 the map from cache entry to cache index optional.
4440 (streamer_tree_cache_replace_tree): Adjust accordingly.
4441 (streamer_tree_cache_append): Likewise.
4442 (streamer_tree_cache_delete): Likewise.
4443 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
4444 streamer cache map from cache entry to cache index.
4445 * lto-streamer-out.c (create_output_block): Adjust.
4446
09485a08
SN
44472013-06-18 Sofiane Naci <sofiane.naci@arm.com>
4448
22fc7d3d
UB
4449 * config/arm/arm.md (attribute "insn"): Move multiplication and
4450 division attributes to...
09485a08
SN
4451 (attribute "type"): ... here. Remove mult.
4452 (attribute "mul32"): New attribute.
4453 (attribute "mul64"): Add umaal.
4454 (*arm_mulsi3): Update attributes.
4455 (*arm_mulsi3_v6): Likewise.
4456 (*thumb_mulsi3): Likewise.
4457 (*thumb_mulsi3_v6): Likewise.
4458 (*mulsi3_compare0): Likewise.
4459 (*mulsi3_compare0_v6): Likewise.
4460 (*mulsi_compare0_scratch): Likewise.
4461 (*mulsi_compare0_scratch_v6): Likewise.
4462 (*mulsi3addsi): Likewise.
4463 (*mulsi3addsi_v6): Likewise.
4464 (*mulsi3addsi_compare0): Likewise.
4465 (*mulsi3addsi_compare0_v6): Likewise.
4466 (*mulsi3addsi_compare0_scratch): Likewise.
4467 (*mulsi3addsi_compare0_scratch_v6): Likewise.
4468 (*mulsi3subsi): Likewise.
4469 (*mulsidi3adddi): Likewise.
4470 (*mulsi3addsi_v6): Likewise.
4471 (*mulsidi3adddi_v6): Likewise.
4472 (*mulsidi3_nov6): Likewise.
4473 (*mulsidi3_v6): Likewise.
4474 (*umulsidi3_nov6): Likewise.
4475 (*umulsidi3_v6): Likewise.
4476 (*umulsidi3adddi): Likewise.
4477 (*umulsidi3adddi_v6): Likewise.
4478 (*smulsi3_highpart_nov6): Likewise.
4479 (*smulsi3_highpart_v6): Likewise.
4480 (*umulsi3_highpart_nov6): Likewise.
4481 (*umulsi3_highpart_v6): Likewise.
4482 (mulhisi3): Likewise.
4483 (*mulhisi3tb): Likewise.
4484 (*mulhisi3bt): Likewise.
4485 (*mulhisi3tt): Likewise.
4486 (maddhisi4): Likewise.
4487 (*maddhisi4tb): Likewise.
4488 (*maddhisi4tt): Likewise.
4489 (maddhidi4): Likewise.
4490 (*maddhidi4tb): Likewise.
4491 (*maddhidi4tt): Likewise.
4492 (divsi3): Likewise.
4493 (udivsi3): Likewise.
4494 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
4495 (thumb2_mulsi_short_compare0): Likewise.
4496 (thumb2_mulsi_short_compare0_scratch): Likewise.
4497 * config/arm/arm1020e.md (1020mult1): Update attribute change.
4498 (1020mult2): Likewise.
4499 (1020mult3): Likewise.
4500 (1020mult4): Likewise.
4501 (1020mult5): Likewise.
4502 (1020mult6): Likewise.
22fc7d3d
UB
4503 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
4504 change.
09485a08
SN
4505 (cortex_a15_mult64): Likewise.
4506 (cortex_a15_sdiv): Likewise.
4507 (cortex_a15_udiv): Likewise.
4508 * config/arm/arm1026ejs.md (mult1): Update attribute change.
4509 (mult2): Likewise.
4510 (mult3): Likewise.
4511 (mult4): Likewise.
4512 (mult5): Likewise.
4513 (mult6): Likewise.
4514 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
4515 (pj4_ir_div): Likewise.
4516 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
4517 (11_mult2): Likewise.
4518 (11_mult3): Likewise.
4519 (11_mult4): Likewise.
4520 (11_mult5): Likewise.
4521 (11_mult6): Likewise.
4522 (11_mult7): Likewise.
4523 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
4524 (cortex_a8_mla): Likewise.
4525 (cortex_a8_mull): Likewise.
4526 (cortex_a8_smulwy): Likewise.
4527 (cortex_a8_smlald): Likewise.
4528 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
4529 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
4530 (cortex_r4_mul_3): Likewise.
4531 (cortex_r4_mla_4): Likewise.
4532 (cortex_r4_mla_3): Likewise.
4533 (cortex_r4_smlald): Likewise.
4534 (cortex_r4_mull): Likewise.
4535 (cortex_r4_sdiv): Likewise.
4536 (cortex_r4_udiv): Likewise.
4537 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
4538 (cortex_a7_idiv): Likewise.
4539 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
4540 (9_mult2): Likewise.
4541 (9_mult3): Likewise.
4542 (9_mult4): Likewise.
4543 (9_mult5): Likewise.
4544 (9_mult6): Likewise.
4545 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
4546 (cortex_a53_sdiv): Likewise.
4547 (cortex_a53_udiv): Likewise.
4548 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
4549 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
4550 (mp626_mult2): Likewise.
4551 (mp626_mult3): Likewise.
4552 (mp626_mult4): Likewise.
4553 * config/arm/fa526.md (526_mult1): Update attribute change.
4554 (526_mult2): Likewise.
4555 * config/arm/arm-generic.md (mult): Update attribute change.
4556 (mult_ldsched_strongarm): Likewise.
4557 (mult_ldsched): Likewise.
4558 (multi_cycle): Likewise.
4559 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
4560 * config/arm/fa606te.md (606te_mult1): Update attribute change.
4561 (606te_mult2): Likewise.
4562 (606te_mult3): Likewise.
4563 (606te_mult4): Likewise.
4564 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
4565 (cortex_a9_mac16): Likewise.
4566 (cortex_a9_multiply): Likewise.
4567 (cortex_a9_mac): Likewise.
4568 (cortex_a9_multiply_long): Likewise.
4569 * config/arm/fa626te.md (626te_mult1): Update attribute change.
4570 (626te_mult2): Likewise.
4571 (626te_mult3): Likewise.
4572 (626te_mult4): Likewise.
4573
8951f345
RB
45742013-06-18 Richard Biener <rguenther@suse.de>
4575
4576 PR lto/57334
4577 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
4578
d8101d05
AK
45792013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4580
4581 PR target/57609
4582 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4583 with next_active_insn.
4584
69eff9da
AM
45852013-06-18 Alan Modra <amodra@gmail.com>
4586
4587 * config/rs6000/rs6000.h (enum data_align): New.
4588 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
4589 (DATA_ABI_ALIGNMENT): Define.
4590 (CONSTANT_ALIGNMENT): Correct comment.
4591 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
4592 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
4593
8ed00d76
DM
45942013-06-17 David Malcolm <dmalcolm@redhat.com>
4595
4596 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
4597 ATTRIBUTE_UNUSED marking.
4598
3d33d06b
SN
45992013-06-17 Sofiane Naci <sofiane.naci@arm.com>
4600
4601 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
4602 alternative and update.
4603 (aarch64_dup_lanedi): Delete.
4604 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
4605 * config/aarch64/aarch64-simd-builtins.def: Update.
4606
ee03e71d
RB
46072013-06-17 Richard Biener <rguenther@suse.de>
4608
4609 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
4610 (lto_input_scc): Declare.
4611 (lto_input_tree_1): Likewise.
4612 (struct lto_stats_d): Add num_tree_bodies_output and
4613 num_pickle_refs_output.
4614 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
4615 (lto_read_tree_1): Split out from ...
4616 (lto_read_tree): ... this.
4617 (lto_input_scc): New function.
4618 (lto_input_tree_1): Split out from ...
4619 (lto_input_tree): ... this. Handle LTO_tree_scc.
4620 (lto_data_in_create): Create the streamer cache without hashes.
4621 * lto-streamer-out.c (create_output_block): Create the streamer
4622 cache with hashes when not doing WPA.
4623 (lto_write_tree_1): Split out from ...
4624 (lto_write_tree): ... this.
4625 (get_symbol_initial_value): New function.
4626 (lto_output_tree_1): Split out from ...
4627 (lto_output_tree): ... this. Write trees as series of SCCs
4628 using a DFS walk via DFS_write_tree.
4629 (struct sccs, struct scc_entry): New types.
4630 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
4631 (DFS_write_tree_body): New function.
4632 (DFS_write_tree): Likewise.
4633 (hash_tree): Likewise.
4634 (scc_entry_compare): Likewise.
4635 (hash_scc): Likewise.
4636 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
4637 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
4638 TREE_CHAIN as regular reference.
4639 (streamer_read_integer_cst): Remove.
4640 (streamer_get_pickled_tree): Adjust.
4641 * tree-streamer-out.c (streamer_write_chain): Disable streaming
4642 of DECL_EXTERNALs in BLOCK_VARS for now.
4643 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
4644 reference.
4645 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
4646 Add hash value argument and record that if hashes are recorded
4647 in the cache.
4648 (streamer_tree_cache_insert_1): Adjust.
4649 (streamer_tree_cache_insert): Likewise.
4650 (streamer_tree_cache_insert_at): Rename to ...
4651 (streamer_tree_cache_replace_tree): ... this and adjust.
4652 (streamer_tree_cache_append): Adjust.
4653 (record_common_node): Likewise.
4654 (streamer_tree_cache_create): Add argument whether to
4655 record hash values together with trees.
4656 (streamer_tree_cache_delete): Adjust.
4657 * tree-streamer.h (struct streamer_tree_cache_d): Add
4658 vector of hashes.
4659 (streamer_read_integer_cst): Remove.
4660 (streamer_tree_cache_insert): Adjust.
4661 (streamer_tree_cache_append): Likewise.
4662 (streamer_tree_cache_insert_at): Rename to ...
4663 (streamer_tree_cache_replace_tree): ... this and adjust.
4664 (streamer_tree_cache_create): Add argument whether to record hashes.
4665 (streamer_tree_cache_get): Rename to ...
4666 (streamer_tree_cache_get_tree): ... this.
4667 (streamer_tree_cache_get_hash): New function.
4668 * tree.c (cache_integer_cst): New function.
4669 * tree.h (cache_integer_cst): Declare.
4670 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
4671 * lto-symtab.c (lto_varpool_replace_node): Only release
4672 DECL_INITIAL of non-prevailing decls.
4673 * varpool.c (varpool_remove_initializer): Do not release
4674 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
4675
0de86a92
JU
46762013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4677
4678 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
4679 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
4680 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
4681 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
4682 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
4683 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
4684 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
4685 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
4686 instead of TARGET_64BIT.
4687 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
4688 Require ISA_HAS_<D>DIV.
4689
74240413
RS
46902013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
4691
4692 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
4693 (mips*-*-linux*): Move default with_llsc setting to where other
4694 defaults are set.
4695 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
4696 with_arch block.
4697 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
4698 Likewise. Remove default with_tune setting. Move default float
4699 setting to its own block. Handle with_llsc in the same block as above.
4700
393f9fed
JR
47012013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
4702
4703 PR rtl-optimization/57425
4704 PR rtl-optimization/57569
299a5f6a 4705 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR
4706 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
4707 Changed all callers.
4708 (canon_anti_dependence): New function.
4709 * cse.c (check_dependence): Use canon_anti_dependence.
4710 * cselib.c (cselib_invalidate_mem): Likewise.
4711 * rtl.h (canon_anti_dependence): Declare.
4712
b259d352
JU
47132013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
4714
4715 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
4716 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
4717 ".set mips3" for 64-bit targets.
4718
4de09b85
DC
47192013-06-15 Dehao Chen <dehao@google.com>
4720
4721 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
4722 * gimple-low.c (gimple_check_call_matching_types): Likewise.
4723 (gimple_check_call_args): Likewise.
4724 * value-prof.c (check_ic_target): Likewise.
4725 * ipa-inline.c (early_inliner): Likewise.
4726 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
4727 * cgraph.c (cgraph_create_edge_1): Likewise.
4728 (cgraph_make_edge_direct): Likewise.
4729
d1aa4a76
MM
47302013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
4731
4732 PR target/57615
4733 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
4734 rs6000_output_move_128bit to handle emitting quad memory
4735 operations. Set attribute length to 8 bytes.
4736
24244e4d
VP
47372013-06-14 Vidya Praveen <vidyapraveen@arm.com>
4738
4739 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
4740 New pattern.
4741 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
4742 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
4743 (aarch64_<su>mlsl<mode>): Likewise.
4744
f5273574
MS
47452013-06-14 Mike Stump <mikestump@comcast.net>
4746
4747 * Makefile.in (TARGET_H): Add insn-codes.h.
4748
ef7e7ab2
AM
47492013-06-14 Alan Modra <amodra@gmail.com>
4750
4751 PR middle-end/57134
4752 PR middle-end/57586
4753 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
4754 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
4755 bitfield expansion when EXPAND_MEMORY.
4756 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
4757
135faab6
MM
47582013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
4759
4760 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
4761 test for clearing quad memory on 32-bit later.
4762
948a1fd9
MG
47632013-06-13 Marc Glisse <marc.glisse@inria.fr>
4764
4765 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
4766 (fold_negate_expr): Likewise.
4767 (fold_real_zero_addition_p): Handle vectors.
4768 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
4769
0597fb9c
AM
47702013-06-14 Alan Modra <amodra@gmail.com>
4771
4772 * varasm.c (force_const_mem): Revert 2013-06-07 change.
4773
e5b962d0
JH
47742013-06-13 Jan Hubicka <jh@suse.cz>
4775
4776 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
4777 Local comdats are not externally visible.
4778 * symtab.c (dump_symtab_base): Dump externally visible.
4779 (verify_symtab_base): Verify back links in the symtab hash.
4780
a8355e51
BC
47812013-06-13 Bin Cheng <bin.cheng@arm.com>
4782
4783 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
4784 CONVERT_EXPR as equal nodes.
4785
8f7e6e33
BC
47862013-06-13 Bin Cheng <bin.cheng@arm.com>
4787
22fc7d3d 4788 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 4789
5d418483
MG
47902013-06-13 Marc Glisse <marc.glisse@inria.fr>
4791
4792 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
4793 Generalize to complex and vector.
4794 * tree.c (build_all_ones_cst): New function.
4795 * tree.h (build_all_ones_cst): Declare it.
4796
50751417
AM
47972013-06-13 Alan Modra <amodra@gmail.com>
4798
4799 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
4800 * config/rs6000/rs6000.md (signbittf2): New insn.
4801 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
4802 (abstf2_internal, cmptf_internal2): Likewise.
4803 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
4804
4b02c962
MM
48052013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4806 Pat Haugen <pthaugen@us.ibm.com>
4807 Peter Bergner <bergner@vnet.ibm.com>
4808
4809 * config/rs6000/rs6000.c (emit_load_locked): Add support for
4810 power8 byte, half-word, and quad-word atomic instructions.
4811 (emit_store_conditional): Likewise.
4812 (rs6000_expand_atomic_compare_and_swap): Likewise.
4813 (rs6000_expand_atomic_op): Likewise.
4814
4815 * config/rs6000/sync.md (larx): Add new modes for power8.
4816 (stcx): Likewise.
4817 (AINT): New mode iterator to include TImode as well as normal
4818 integer modes on power8.
4819 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
4820 that VSX registers are not considered. Use AINT mode iterator
4821 instead of INT1 to allow inclusion of quad word atomic operations
4822 on power8.
4823 (load_locked<mode>): Likewise.
4824 (store_conditional<mode>): Likewise.
4825 (atomic_compare_and_swap<mode>): Likewise.
4826 (atomic_exchange<mode>): Likewise.
4827 (atomic_nand<mode>): Likewise.
4828 (atomic_fetch_<fetchop_name><mode>): Likewise.
4829 (atomic_nand_fetch<mode>): Likewise.
4830 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
4831 each type.
4832 (ATOMIC): On power8, add QImode, HImode modes.
4833 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
4834 modes that promote to SImode.
4835 (load_lockedti): Convert TImode arguments to PTImode, so that we
4836 get a guaranteed even/odd register pair.
4837 (load_lockedpti): Likewise.
4838 (store_conditionalti): Likewise.
4839 (store_conditionalpti): Likewise.
4840
4841 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
4842 atomic load/store instructions.
4843 (HSI): Likewise.
4844
710b6cc1
RS
48452013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
4846
4847 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
4848 loads.
4849 (insn_count): New attribute, with most cases extracted from...
4850 (length): ...here. Redefine most cases in terms of insn_count.
4851 (single_insn): Delete.
4852 (can_delay): Use insn_count to check for single instructions.
4853 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
4854 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
4855 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
4856 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
4857 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
4858 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
4859 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
4860 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
4861 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
4862 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
4863 rather than "length".
4864 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
4865 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
4866 Use "insn_count" rather than "length".
4867 * config/mips/mips-dsp.md
4868 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
4869 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
4870 length attributes.
4871
1951f101
MG
48722013-06-12 Marc Glisse <marc.glisse@inria.fr>
4873
4874 PR tree-optimization/57361
4875 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
4876
8b033a8a
SN
48772013-06-12 Sofiane Naci <sofiane.naci@arm.com>
4878
22fc7d3d
UB
4879 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
4880 to split.
8b033a8a
SN
4881 (aarch64_simd_combine<mode>): New instruction expansion.
4882 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
4883 function prototype.
4884 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
4885 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
4886
ca0f62a8
JH
48872013-06-12 Jan Hubicka <jh@suse.cz>
4888
4889 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
4890 decl has when in streaming stage.
4891 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
4892 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
4893
cddd69eb
RS
48942013-06-12 Roland Stigge <stigge@antcom.de>
4895
4896 PR target/57578
4897 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
4898
bfdeda2c
JJ
48992013-06-12 Jakub Jelinek <jakub@redhat.com>
4900
4901 PR tree-optimization/57537
4902 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
4903 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
4904
a4fa02d1
RB
49052013-06-12 Richard Biener <rguenther@suse.de>
4906
4907 * data-streamer.h (streamer_write_char_stream): CSE
4908 obs->current_pointer.
4909 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
4910 streamer_write_char_stream manually and optimize the resulting loop.
4911 (streamer_write_hwi_stream): Likewise.
4912
bbf9ad07
JH
49132013-06-12 Jan Hubicka <jh@suse.cz>
4914
4915 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
4916 * cgraph.h (varpool_create_empty_node): Declare.
4917 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
4918 duplicated nodes.
4919 * symtab.c (symtab_unregister_node): Be lax about missin entries
4920 in node hash.
4921 (symtab_get_node): Update comment.
4922 * varpool.c (varpool_create_empty_node): Break out from ...
4923 (varpool_node_for_decl): ... here.
4924 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
4925
a53f90ad
EB
49262013-06-12 Eric Botcazou <ebotcazou@adacore.com>
4927
4928 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
4929 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
4930 part. Use straight-line flow at the end.
4931 <COMPONENT_REF>: Remove superfluous else.
4932 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
4933
946f9306
JJ
49342013-06-12 Jakub Jelinek <jakub@redhat.com>
4935
4936 PR target/56564
4937 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
4938 target hook even for !TREE_PUBLIC decls. If no resolution info
4939 is available, return false for common and external decls.
4940
a7fca6f0
KP
49412013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
4942
4943 * config/rl78/constraints.md (U): New constraint.
4944 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
4945 valloc attribute.
4946
c7d8f446
MM
49472013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
4948
4949 PR target/57589
4950 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
4951 to allow returning address to AT_PLATFORM name.
4952
08346abd
JH
49532013-06-11 Jan Hubicka <jh@suse.cz>
4954
4955 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
4956 * cgraph.h (symtab_node_base): Add weakref flag.
4957 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
4958 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
4959 (output_weakrefs): Use weakref flag.
4960 * fold-const.c (simple_operand_p): Handle WEAK.
4961 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
4962 * ipa.c (varpool_externally_visible_p): Drop weakref.
4963 (function_and_variable_visibility): Update comment; fix weakref
4964 sanity checks; do not clear DECL_WEAK on them.
4965 * lto-cgraph.c (lto_output_node): update.
4966 (lto_output_varpool_node): Update.
4967 (input_overwrite_node): Update.
4968 (input_node): Update.
4969 (input_varpool_node): Update.
4970 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
4971 (lto_symtab_merge_symbols): Add sanity check.
4972 (lto_symtab_prevailing_decl): Do not special case weakrefs.
4973 * passes.c (rest_of_decl_compilation): Set static flag, too.
4974 * symtab.c (dump_symtab_base): Dump weakref.
4975 (verify_symtab_base): Sanity check weakrefs.
4976 (symtab_make_decl_local): Remove duplicated code.
4977 (symtab_alias_ultimate_target): Simplify.
4978 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 4979
d9af3652
DD
49802013-06-11 DJ Delorie <dj@redhat.com>
4981
4982 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
4983 (rl78_unwind_word_mode): New.
4984
ffa66012
DM
49852013-06-11 David Malcolm <dmalcolm@redhat.com>
4986
4987 * final.c (debug_prefix_maps): Make static.
4988
783f2d35
DM
49892013-06-11 David Malcolm <dmalcolm@redhat.com>
4990
4991 * function.c (initial_trampoline): Remove stray copy.
4992
7af26def
SN
49932013-06-11 Sofiane Naci <sofiane.naci@arm.com>
4994
4995 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
4996
97756c0e
MJ
49972013-06-11 Martin Jambor <mjambor@suse.cz>
4998
4999 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5000 within bounds at the beginning of the function.
5001
2c7b8bf6
AM
50022013-06-11 Alan Modra <amodra@gmail.com>
5003
5004 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5005 reporting.
5006 (get_named_section): Don't NULL !DECL_P decl.
5007
77cea46e
IZ
50082013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5009
5010 * doc/invoke.texi (core-avx2): Document.
5011 (slm): Likewise.
5012 (atom): Updated with MOVBE.
5013
888f0920
RB
50142013-06-11 Richard Biener <rguenther@suse.de>
5015
22fc7d3d 5016 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 5017
6b438d58
AB
50182013-06-11 Anton Blanchard <anton@samba.org>
5019
5020 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5021 correct shift value in little-endian mode.
5022
3bc66938
JJ
50232013-06-11 Jakub Jelinek <jakub@redhat.com>
5024
5025 PR target/56564
5026 * varasm.c (get_variable_align): Move #endif to the right place.
5027
a5dc7f88
CC
50282013-06-10 Cary Coutant <ccoutant@google.com>
5029
5030 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5031 for hash so that hash table traversal order is deterministic.
5032
19be72ab
MM
50332013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5034 Pat Haugen <pthaugen@us.ibm.com>
5035 Peter Bergner <bergner@vnet.ibm.com>
5036
5037 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5038 of vectors in GPRS if they are direct moves or quad word load or
5039 store moves.
5040
5041 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5042 declaration.
5043 (direct_move_p): Likewise.
5044 (quad_load_store_p): Likewise.
5045
5046 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5047 classes into bins based on the physical register type.
5048 (reg_class_to_reg_type): Likewise.
5049 (IS_STD_REG_TYPE): Likewise.
5050 (IS_FP_VECT_REG_TYPE): Likewise.
5051 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5052 use direct move instructions.
5053 (reload_gpr_vsx): Likewise.
5054 (reload_vsx_gpr): Likewise.
5055 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5056 information that is a simplification of register classes. Also
5057 precalculate direct move reload helpers.
5058 (direct_move_p): New function to return true if the operation can
5059 be done as a direct move instruciton.
5060 (quad_load_store_p): New function to return true if the operation
5061 is a quad memory operation.
5062 (rs6000_legitimize_address): If quad memory, only allow register
5063 indirect for TImode addresses.
5064 (rs6000_legitimate_address_p): Likewise.
5065 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5066 (rs6000_reload_register_type): Likewise.
5067 (register_to_reg_type): Return register type.
5068 (rs6000_secondary_reload_simple_move): New helper function for
5069 secondary reload and secondary memory needed to identify anything
5070 that is a simple move, and does not need reloading.
5071 (rs6000_secondary_reload_direct_move): New helper function for
5072 secondary reload to identify cases that can be done with several
5073 instructions via the direct move instructions.
5074 (rs6000_secondary_reload_move): New helper function for secondary
5075 reload to identify moves between register types that can be done.
5076 (rs6000_secondary_reload): Add support for quad memory operations
5077 and for direct move.
5078 (rs6000_secondary_memory_needed): Likewise.
5079 (rs6000_debug_secondary_memory_needed): Change argument names.
5080 (rs6000_output_move_128bit): New function to return the move to
5081 use for 128-bit moves, including knowing about the various
5082 limitations of quad memory operations.
5083
5084 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5085 memory operations. call rs6000_output_move_128bit for the actual
5086 instruciton(s) to generate.
5087 (vsx_movti_64bit): Likewise.
5088
5089 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5090 (UNSPEC_P8V_MTVSRWZ): Likewise.
5091 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5092 (UNSPEC_P8V_MTVSRD): Likewise.
5093 (UNSPEC_P8V_XXPERMDI): Likewise.
5094 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5095 (UNSPEC_FUSION_GPR): Likewise.
5096 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 5097 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM
5098 (f32_sv): Likewise.
5099 (f32_dm): Likewise.
5100 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5101 loads and direct move instructions.
5102 (zero_extendsidi2_lfiwzx): Likewise.
5103 (extendsidi2_lfiwax): Likewise.
5104 (extendsidi2_nocell): Likewise.
5105 (floatsi<mode>2_lfiwax): Likewise.
5106 (lfiwax): Likewise.
5107 (floatunssi<mode>2_lfiwzx): Likewise.
5108 (lfiwzx): Likewise.
5109 (fix_trunc<mode>_stfiwx): Likewise.
5110 (fixuns_trunc<mode>_stfiwx): Likewise.
5111 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5112 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5113 (parity<mode>2_cmpb): Set length/type attr.
5114 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5115 for 'mr.' to fast_compare.
5116 (bpermd_<mode>): Change type attr to popcnt.
5117 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5118 (p8_mtvsrwz_1): Likewise.
5119 (p8_mtvsrwz_2): Likewise.
5120 (reload_fpr_from_gpr<mode>): Likewise.
5121 (p8_mtvsrd_1): Likewise.
5122 (p8_mtvsrd_2): Likewise.
5123 (p8_xxpermdi_<mode>): Likewise.
5124 (reload_vsx_from_gpr<mode>): Likewise.
5125 (reload_vsx_from_gprsf): Likewise.
5126 (p8_mfvsrd_3_<mode>): LIkewise.
5127 (reload_gpr_from_vsx<mode>): Likewise.
5128 (reload_gpr_from_vsxsf): Likewise.
5129 (p8_mfvsrd_4_disf): Likewise.
5130 (multi-word GPR splits): Do not split direct moves or quad memory
5131 operations.
5132
16876bdc
DM
51332013-06-10 David Malcolm <dmalcolm@redhat.com>
5134
5135 * tree-into-ssa.c (interesting_blocks): Make static.
5136
df8a1d28
JJ
51372013-06-10 Jakub Jelinek <jakub@redhat.com>
5138
5139 PR target/56564
5140 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5141 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5142 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5143 (get_variable_align): New function.
5144 (get_variable_section, emit_bss, emit_common,
5145 assemble_variable_contents, place_block_symbol): Use
5146 get_variable_align instead of DECL_ALIGN.
5147 (assemble_noswitch_variable): Add align argument, use it
5148 instead of DECL_ALIGN.
5149 (assemble_variable): Adjust caller. Use get_variable_align
5150 instead of DECL_ALIGN.
5151 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5152 caller.
5153 (DATA_ABI_ALIGNMENT): Define.
5154 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5155 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5156 opt is false, only return the psABI mandated alignment increase.
5157 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5158 (DATA_ABI_ALIGNMENT): ... this.
5159 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5160 (DATA_ABI_ALIGNMENT): ... this.
5161 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5162 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5163 (DATA_ABI_ALIGNMENT): ... this.
5164 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5165 * doc/tm.texi: Regenerated.
5166
57b29ca6
UB
51672013-06-10 Uros Bizjak <ubizjak@gmail.com>
5168
5169 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5170 cmp_code to construct REG_EQUAL note.
5171
3e56ed50
JJ
51722013-06-09 Jakub Jelinek <jakub@redhat.com>
5173
5174 PR target/57568
5175 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5176 that operands[2] doesn't overlap with operands[0].
5177
74fe2a1d
DE
51782013-06-09 David Edelsohn <dje.gcc@gmail.com>
5179 Jan Hubicka <jh@suse.cz>
5180
5181 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5182 hack to mark symbols as used.
5183
ed52a84e
VM
51842013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5185
5186 PR rtl-optimization/57559
5187 * lra-constraints.c (process_alt_operands): Don't discourage
5188 memory with known offset for offsetable memory constraint.
5189 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5190
299316ed
EB
51912013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5192
5193 * varasm.c (struct oc_local_state): Reorder fields.
5194 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5195 and adjust accordingly.
5196 (output_constructor): Reorder initialization code and adjust call to
5197 output_constructor_bitfield.
5198
88ac60d3
JH
51992013-06-07 Jan Hubicka <jh@suse.cz>
5200
5201 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5202
4cdce1a8
DM
52032013-06-07 David Malcolm <dmalcolm@redhat.com>
5204
5205 * tree-object-size.c (unknown): Make const.
5206
b5e0425c
AK
52072013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5208
5209 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5210 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5211 for last alternative in the cpu_facility attribute.
5212
79678d04
KT
52132013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5214
5215 PR target/56315
5216 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5217 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5218 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5219 * config/arm/constraints.md (Dg): New constraint.
5220 * config/arm/neon.md (xordi3_neon): Remove.
5221 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5222 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5223
b31ddbdb
KT
52242013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5225
5226 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5227 Clean up alternatives.
5228
8f90b7d4
AM
52292013-06-07 Alan Modra <amodra@gmail.com>
5230
5231 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5232 va_list_gpr_size.
5233
961ce119
AM
52342013-06-07 Alan Modra <amodra@gmail.com>
5235
5236 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5237
33e49835
KT
52382013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5239
5240 * config/arm/constraints.md (Df): New constraint.
5241 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5242 Correct length attribute for last two alternatives.
5243
acd17ae6
AM
52442013-06-07 Alan Modra <amodra@gmail.com>
5245
5246 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5247 override user -mfp-in-toc.
5248 (offsettable_ok_by_alignment): Consider just the current access
5249 rather than the whole object, unless BLKmode. Handle
5250 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5251 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5252 for -mcmodel=medium.
5253 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5254 override user -mfp-in-toc or -msum-in-toc. Default to
5255 -mno-fp-in-toc for -mcmodel=medium.
5256
73310b0e
DD
52572013-06-06 DJ Delorie <dj@redhat.com>
5258
5259 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5260 TARGET_VALID_POINTER_MODE.
5261
0bd62dca
MM
52622013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5263 Pat Haugen <pthaugen@us.ibm.com>
5264 Peter Bergner <bergner@vnet.ibm.com>
5265
5266 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5267 Document new power8 builtins.
5268
5269 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5270 condition code register, to allow 128-bit logical operations to be
5271 done in the VSX or GPR registers.
5272 (nor<mode>3): Use the canonical form for nor.
5273 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5274 vclz*, and vpopcnt* vector instructions.
5275 (nand<mode>3): Likewise.
5276 (orc<mode>3): Likewise.
5277 (clz<mode>2): LIkewise.
5278 (popcount<mode>2): Likewise.
5279
5280 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5281 that only the GPRs are recognized.
5282
5283 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5284 support for new power8 builtins.
5285
5286 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5287 builtin functions.
5288 (xscvdpspn): Likewise.
5289 (vclz): Likewise.
5290 (vclzb): Likewise.
5291 (vclzh): Likewise.
5292 (vclzw): Likewise.
5293 (vclzd): Likewise.
5294 (vpopcnt): Likewise.
5295 (vpopcntb): Likewise.
5296 (vpopcnth): Likewise.
5297 (vpopcntw): Likewise.
5298 (vpopcntd): Likewise.
5299 (vgbbd): Likewise.
5300 (vmrgew): Likewise.
5301 (vmrgow): Likewise.
5302 (eqv): Likewise.
5303 (eqv_v16qi3): Likewise.
5304 (eqv_v8hi3): Likewise.
5305 (eqv_v4si3): Likewise.
5306 (eqv_v2di3): Likewise.
5307 (eqv_v4sf3): Likewise.
5308 (eqv_v2df3): Likewise.
5309 (nand): Likewise.
5310 (nand_v16qi3): Likewise.
5311 (nand_v8hi3): Likewise.
5312 (nand_v4si3): Likewise.
5313 (nand_v2di3): Likewise.
5314 (nand_v4sf3): Likewise.
5315 (nand_v2df3): Likewise.
5316 (orc): Likewise.
5317 (orc_v16qi3): Likewise.
5318 (orc_v8hi3): Likewise.
5319 (orc_v4si3): Likewise.
5320 (orc_v2di3): Likewise.
5321 (orc_v4sf3): Likewise.
5322 (orc_v2df3): Likewise.
5323
5324 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
5325 allow power8 quad mode in 64-bit.
5326 (rs6000_builtin_vectorized_function): Add support to vectorize
5327 ISA 2.07 count leading zeros, population count builtins.
5328 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
5329 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
5330 (builtin_function_type): Add vgbbd builtin function which takes an
5331 unsigned argument.
5332 (altivec_expand_vec_perm_const): Add support for new power8 merge
5333 instructions.
5334
5335 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
5336 that does not include TImdoe for use with 32-bit.
5337 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
5338 instructions.
5339 (UNSPEC_VSX_CVDPSPN): Likewise.
5340 (vsx_xscvdpspn): Likewise.
5341 (vsx_xscvspdpn): Likewise.
5342 (vsx_xscvdpspn_scalar): Likewise.
5343 (vsx_xscvspdpn_directmove): Likewise.
5344 (vsx_and<mode>3): Split logical operations into 32-bit and
5345 64-bit. Add support to do logical operations on TImode as well as
5346 VSX vector types. Allow logical operations to be done in either
5347 VSX registers or in general purpose registers in 64-bit mode. Add
5348 splitters if GPRs were used. For AND, add clobber of CCmode to
5349 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
5350 encoding.
5351 (vsx_and<mode>3_32bit): Likewise.
5352 (vsx_and<mode>3_64bit): Likewise.
5353 (vsx_ior<mode>3): Likewise.
5354 (vsx_ior<mode>3_32bit): Likewise.
5355 (vsx_ior<mode>3_64bit): Likewise.
5356 (vsx_xor<mode>3): Likewise.
5357 (vsx_xor<mode>3_32bit): Likewise.
5358 (vsx_xor<mode>3_64bit): Likewise.
5359 (vsx_one_cmpl<mode>2): Likewise.
5360 (vsx_one_cmpl<mode>2_32bit): Likewise.
5361 (vsx_one_cmpl<mode>2_64bit): Likewise.
5362 (vsx_nor<mode>3): Likewise.
5363 (vsx_nor<mode>3_32bit): Likewise.
5364 (vsx_nor<mode>3_64bit): Likewise.
5365 (vsx_andc<mode>3): Likewise.
5366 (vsx_andc<mode>3_32bit): Likewise.
5367 (vsx_andc<mode>3_64bit): Likewise.
5368 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
5369 and xxlorc instructions.
5370 (vsx_eqv<mode>3_64bit): Likewise.
5371 (vsx_nand<mode>3_32bit): Likewise.
5372 (vsx_nand<mode>3_64bit): Likewise.
5373 (vsx_orc<mode>3_32bit): Likewise.
5374 (vsx_orc<mode>3_64bit): Likewise.
5375
5376 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
5377
5378 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
5379 instruction.
5380 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
5381 (p8_vmrgow): Likewise.
5382 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
5383 GPRs to be split under VSX.
5384 (p8v_clz<mode>2): Add power8 count leading zero support.
5385 (p8v_popcount<mode>2): Add power8 population count support.
5386 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
5387 support.
5388
5389 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
5390 instruction.
5391
5392 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
5393 builtin functions.
5394 (vec_nand): Likewise.
5395 (vec_vclz): Likewise.
5396 (vec_vclzb): Likewise.
5397 (vec_vclzd): Likewise.
5398 (vec_vclzh): Likewise.
5399 (vec_vclzw): Likewise.
5400 (vec_vgbbd): Likewise.
5401 (vec_vmrgew): Likewise.
5402 (vec_vmrgow): Likewise.
5403 (vec_vpopcnt): Likewise.
5404 (vec_vpopcntb): Likewise.
5405 (vec_vpopcntd): Likewise.
5406 (vec_vpopcnth): Likewise.
5407 (vec_vpopcntw): Likewise.
5408
37684c46
VM
54092013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5410
5411 PR rtl-optimization/57468
5412 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
5413 spilled pseudos.
5414
d89ae27c
VM
54152013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5416
5417 PR rtl-optimization/57459
5418 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
5419 type when setting live regs.
5420
3597e113
VM
54212013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5422
5423 * config/s390/s390.opt (mlra): New option.
5424 * config/s390/s390.c (s390_decompose_address): Check displacement
5425 for all registers for LRA.
5426 (s390_secondary_reload): Don't used secondary reloads for LRA.
5427 (s390_lra_p): New function.
5428 (TARGET_LRA_P): Define.
5429 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
5430 of attribute cpu_facility to zarch for the last alternative.
5431 (*cmpmem_short): Ditto.
5432
01037aeb
EB
54332013-06-06 Eric Botcazou <ebotcazou@adacore.com>
5434
5435 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
5436 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
5437 (arm_expand_prologue): Likewise.
5438
3371a64f
TJ
54392013-06-06 Teresa Johnson <tejohnson@google.com>
5440
5441 PR c++/53743
5442 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
5443 as this is now done by redirect_edge_and_branch_force.
5444 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
5445 barriers, and fix interaction with splitting.
5446 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
5447 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
5448 reflect changes made in the routine.
5449 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
5450 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
5451 since this is called in cfglayout mode, and replace partition fixup
5452 with assert as that is now done by force_nonfallthru_and_redirect.
5453 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
5454 already be marked with region crossing note.
5455 (insert_section_boundary_note): Make non-static, gate on flag
5456 has_bb_partition, rewrite to also check for multiple partitions.
5457 (rest_of_handle_reorder_blocks): Remove call to
5458 insert_section_boundary_note, now done later during free_cfg.
5459 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
5460 * bb-reorder.h (insert_section_boundary_note): Declare.
5461 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
5462 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
5463 invoke insert_section_boundary_note.
5464 (try_redirect_by_replacing_jump): Remove unnecessary
5465 check for region crossing note.
5466 (fixup_partition_crossing): New function.
5467 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
5468 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
5469 in non-cfglayout mode.
5470 (force_nonfallthru_and_redirect): Fixup partition boundaries,
5471 remove old code that tried to do this. Emit barrier correctly
5472 when we are in cfglayout mode.
5473 (last_bb_in_partition): New function.
5474 (rtl_split_edge): Correctly fixup partition boundaries.
5475 (commit_one_edge_insertion): Remove old code that tried to
5476 fixup region crossing edge since this is now handled in
5477 split_block, and set up insertion point correctly since
5478 block may now end in a jump.
5479 (verify_hot_cold_block_grouping): Guard against checking when not in
5480 linearized RTL mode.
5481 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
5482 notes.
5483 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
5484 rtl_verify_flow_info, so not called in cfglayout mode.
5485 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
5486 (fixup_reorder_chain): Remove old code that attempted to fixup region
5487 crossing note as this is now handled in force_nonfallthru_and_redirect.
5488 (duplicate_insn_chain): Don't duplicate switch section notes.
5489 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
5490 note.
5491 * basic-block.h (emit_barrier_after_bb): Declare.
5492
66071e10
KT
54932013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5494
5495 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
5496 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
5497 arm_usatsihi): Adjust alternatives for arm_restrict_it.
5498
0e26bf3d
KT
54992013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5500
5501 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
5502 where appropriate.
5503 * config/arm/ldmstm.md: Regenerate.
5504
12b4e7ef
KT
55052013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5506
5507 * config/arm/sync.md (atomic_loaddi_1):
5508 Disable predication for arm_restrict_it.
5509 (arm_load_exclusive<mode>): Likewise.
5510 (arm_load_exclusivesi): Likewise.
5511 (arm_load_exclusivedi): Likewise.
5512 (arm_load_acquire_exclusive<mode>): Likewise.
5513 (arm_load_acquire_exclusivesi): Likewise.
5514 (arm_load_acquire_exclusivedi): Likewise.
5515 (arm_store_exclusive<mode>): Likewise.
5516 (arm_store_exclusive<mode>): Likewise.
5517 (arm_store_release_exclusivedi): Likewise.
5518 (arm_store_release_exclusive<mode>): Likewise.
5519
15d16c8a
RB
55202013-06-06 Richard Biener <rguenther@suse.de>
5521
5522 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
5523 after LTO_null.
5524 (lto_tag_is_tree_code_p): Adjust.
5525 (lto_tag_is_gimple_code_p): Likewise.
5526 (lto_gimple_code_to_tag): Likewise.
5527 (lto_tag_to_gimple_code): Likewise.
5528 (lto_tree_code_to_tag): Likewise.
5529 (lto_tag_to_tree_code): Likewise.
5530 * data-streamer.h (streamer_write_hwi_in_range): Use
5531 uhwi streaming to stream the normalized range.
5532 (streamer_read_hwi_in_range): Likewise.
5533
17a7fc37
KT
55342013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5535
5536 * config/arm/arm.md (enabled_for_depr_it): New attribute.
5537 (predicable_short_it): Likewise.
5538 (predicated): Likewise.
5539 (enabled): Handle above.
5540 (define_cond_exec): Set predicated attribute to yes.
5541
b57ca59b
MS
55422013-06-05 Mike Stump <mikestump@comcast.net>
5543
5544 * gdbinit.in (__FUNCTION__): Add.
5545
c1e183a9
UB
55462013-06-05 Uros Bizjak <ubizjak@gmail.com>
5547
5548 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
5549 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
5550
0378bddb
JH
55512013-06-05 Jan Hubicka <jh@suse.cz>
5552
5553 * varasm.c (mark_decl_referenced): Revert the removal until targets
5554 are fixed.
5555
8f8a2057
DE
55562013-06-05 David Edelsohn <dje.gcc@gmail.com>
5557
5558 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
5559 instead of mark_decl_referenced.
5560
edb983b2
JH
55612013-06-05 Jan Hubicka <jh@suse.cz>
5562
5563 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
5564 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
5565 and symtab_used_from_object_file_p.
5566 (cgraph_make_node_local_1): Clear forced_by_abi.
5567 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
5568 * cgraph.h (symtab_node_base): Add forced_by_abi.
5569 (decide_is_variable_needed): Remove.
5570 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
5571 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
5572 (decide_is_symbol_needed): ... this one; handle symbols in general;
5573 always analyze virtuals; honnor forced_by_abi.
5574 (cgraph_finalize_function): Update.
5575 (varpool_finalize_decl): Update.
5576 (symbol_defined_and_needed): Remove.
5577 (analyze_functions): Update.
5578 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
5579 output_refs, input_overwrite_node): Handle forced_by_abi.
5580 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
5581 (address_taken_from_non_vtable_p): ... this one.
5582 (comdat_can_be_unshared_p_1): New function.
5583 (cgraph_comdat_can_be_unshared_p): Rename to ...
5584 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
5585 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
5586 (function_and_variable_visibility): Clear forced_by_abi as needed.
5587 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
5588 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH
5589 * symtab.c (dump_symtab_base): Dump forced_by_abi.
5590 * varpool.c (decide_is_variable_needed): Remove.
5591
9912dbe5
KT
55922013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5593
5594 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
5595 (arm_option_override): Override arm_restrict_it where appropriate.
5596 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
5597 * config/arm/arm.opt (mrestrict-it): New command-line option.
5598 * doc/invoke.texi: Document -mrestrict-it.
5599
34c136b6
DM
56002013-06-05 David Malcolm <dmalcolm@redhat.com>
5601
5602 * tsan.c (tsan_atomic_table): Make const.
5603
a367df53
RB
56042013-06-05 Richard Biener <rguenther@suse.de>
5605
5606 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
5607 index associated with the tree we are supposed to replace.
5608 * tree-streamer-out.c (pack_ts_base_value_fields): Output
5609 TREE_ASM_WRITTEN as zero for everything but SSA names.
5610
70d51a19
DM
56112013-06-05 David Malcolm <dmalcolm@redhat.com>
5612
5613 * tree-ssa-structalias.c (call_stmt_vars): Make static.
5614
b9bd2075
JH
56152013-06-04 Jan Hubicka <jh@suse.cz>
5616
5617 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
5618 (input_node, input_varpool_node): Handle correctly external same
5619 body aliases.
5620 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
5621 nodes at ltrans stage.
5622
aaae719d
JH
56232013-06-04 Jan Hubicka <jh@suse.cz>
5624
5625 * ipa-inline.c (update_caller_keys): Fix availability test.
5626 (update_callee_keys): Likewise.
5627 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
5628 to follow ELF standard.
5629
107eea2c
JU
56302013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
5631
5632 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
5633 (mips64r5900el-*-elf*): New configurations.
5634 * config/mips/mips-cpus.def (r5900): New processor.
5635 * config/mips/mips-tables.opt: Regenerate.
5636 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
5637 (mips_issue_rate): Handle PROCESSOR_R5900.
5638 (mips_reorg_process_insns): Force reorder mode for the R5900.
5639 * config/mips/mips.h (TARGET_MIPS5900): Define.
5640 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
5641 TARGET_MIPS5900.
5642 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
5643 TARGET_MIPS5900.
5644 * config/mips/mips.md (processor): Add r5900.
5645 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
5646
b7342d25
IB
56472013-06-04 Ian Bolton <ian.bolton@arm.com>
5648
5649 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
5650 into function to generate MOVI instruction.
c1e183a9 5651 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB
5652 (aarch64_preferred_simd_mode): Turn into wrapper.
5653 (aarch64_output_scalar_simd_mov_immediate): New function.
5654 * config/aarch64/aarch64-protos.h: Add prototype for above.
5655
81c2dfb9
IB
56562013-06-04 Ian Bolton <ian.bolton@arm.com>
5657
5658 * config/aarch64/aarch64.c (simd_immediate_info): Remove
5659 element_char member.
5660 (sizetochar): Return signed char.
5661 (aarch64_simd_valid_immediate): Remove elchar and other
5662 unnecessary variables.
5663 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
5664 Calculate element_char as required.
5665 * config/aarch64/aarch64-protos.h: Update and move prototype
5666 for aarch64_output_simd_mov_immediate.
5667 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
5668 Update arguments.
5669
48063b9d
IB
56702013-06-04 Ian Bolton <ian.bolton@arm.com>
5671
5672 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
5673 information completed by aarch64_simd_valid_immediate.
5674 (aarch64_legitimate_constant_p): Update arguments.
5675 (aarch64_simd_valid_immediate): Work with struct rather than many
5676 pointers.
5677 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
5678 (aarch64_simd_make_constant): Update arguments.
5679 (aarch64_output_simd_mov_immediate): Work with struct rather than
5680 many pointers. Output immediate directly rather than as operand.
5681 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5682 Update prototype.
5683 * config/aarch64/constraints.md (Dn): Update arguments.
5684
3ea63f60
IB
56852013-06-04 Ian Bolton <ian.bolton@arm.com>
5686
5687 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
5688 longer static.
5689 (aarch64_simd_immediate_valid_for_move): Remove.
5690 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
5691 (aarch64_simd_make_constant): Update call.
5692 (aarch64_output_simd_mov_immediate): Update call.
5693 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
5694 Add prototype.
5695 * config/aarch64/constraints.md (Dn): Update call.
5696
d8edd899
IB
56972013-06-04 Ian Bolton <ian.bolton@arm.com>
5698
5699 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
5700 return type to bool for prototype.
5701 (aarch64_legitimate_constant_p): Check for true instead of not -1.
5702 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
5703 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
5704
44b20bb8
CM
57052013-06-04 Catherine Moore <clm@codesourcery.com>
5706
5707 * config/mips/mips.opt (meva): New.
c1e183a9 5708 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM
5709 (ASM_SPEC): Handle -meva.
5710 * doc/invoke.texi (meva): Document.
5711
52befbd8
AM
57122013-06-04 Alan Modra <amodra@gmail.com>
5713
5714 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
5715 constant output.
5716
aadaf24e
KT
57172013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5718
5719 * rtl.def: Add extra fourth optional field to define_cond_exec.
5720 * gensupport.c (process_one_cond_exec): Process attributes from
5721 define_cond_exec.
5722 * doc/md.texi: Document fourth field in define_cond_exec.
5723
0cd9e9ee
EB
57242013-06-04 Eric Botcazou <ebotcazou@adacore.com>
5725
5726 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
5727 out the processing order as in store_bit_field_1.
5728
cb0f665d
JH
57292013-06-04 Jan Hubicka <jh@suse.cz>
5730
5731 PR middle-end/57500
5732 * cgraphunit.c (cgraph_process_same_body_aliases): Create
5733 non-VAR_DECL node if it does not exist yet.
5734
53984b9b
RS
57352013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
5736
5737 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
5738 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
5739 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
5740 target_cpu_default setting.
5741
5d318fd4
TJ
57422013-06-03 Teresa Johnson <tejohnson@google.com>
5743
5744 * dumpfile.c (opt_info_switch_p): Change -fopt-info
5745 default to -fopt-info=optimized instead of all.
5746 * doc/invoke.texi: Ditto.
5747 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
5748 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
5749 (execute_vect_slp): Emit BB vectorization success under
5750 MSG_OPTIMIZED_LOCATIONS.
5751 * tree-vect-slp.c (vect_slp_transform_bb): Change
5752 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
5753 * tree-vect-loop.c (vect_transform_loop): Ditto.
5754
ec3af349
JM
57552013-06-03 Jason Merrill <jason@redhat.com>
5756
5757 PR c++/57415
5758 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
5759 Use TARGET_EXPR for C++.
5760
1baf9159
JJ
57612013-06-03 Jakub Jelinek <jakub@redhat.com>
5762
5763 PR rtl-optimization/57268
0cd9e9ee 5764 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ
5765 if DEBUG_INSN_P (insn).
5766
5767 Reapply
5768 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
5769
5770 PR rtl-optimization/57268
5771 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5772 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5773
0f1d3965
YR
57742013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
5775
5776 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
5777 (ix86_avoid_lea_for_addr): Likewise.
5778 (exact_dependency_1): Likewise.
5779 (ix86_adjust_cost): Likewise.
5780 (swap_top_of_ready_list): Fix formatting and !reload_completed check
5781 removed.
5782 (do_reorder_for_imul): Fix typo, formatting and
5783 !reload_completed check removed.
5784 (ix86_sched_reorder): Fix typo and formatting.
5785 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
5786 list.
5787
3bc30db4
SN
57882013-06-03 Sofiane Naci <sofiane.naci@arm.com>
5789
5790 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
5791
19cc6fac
EB
57922013-06-03 Eric Botcazou <ebotcazou@adacore.com>
5793
5794 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
5795 <STRING_CST>: Likewise.
5796 <VECTOR_CST>: Likewise.
5797
c68a6e08
JW
57982013-06-01 Janus Weil <janus@gcc.gnu.org>
5799 Mikael Morin <mikael@gcc.gnu.org>
5800
5801 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
5802 * config.in: Regenerated.
5803 * configure: Regenerated.
5804
38e55e5c
JH
58052013-06-01 Jan Hubicka <jh@suse.cz>
5806
5807 PR middle-end/57366
5808 * cgraphunit.c (compile): When weakref is not supported,
5809 set up transparent aliases before final output pass.
5810 * varasm.c (assemble_alias): Do not try to do it here.
5811
eb51d2ff
JH
58122013-06-01 Jan Hubicka <jh@suse.cz>
5813
5814 PR middle-end/57467
5815 * passes.c (for_per_function): Skip unanalyzed functions.
5816
40a7fe1e
JH
58172013-06-01 Jan Hubicka <jh@suse.cz>
5818
5819 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
5820 (lto_symtab_merge_symbols_1): ... this one.
5821 (lto_symtab_merge_cgraph_nodes): Rename to ...
5822 (lto_symtab_merge_symbols): ... this one; simplify.
5823 * cgraph.c (same_body_aliases_done): Rename to ...
5824 (cpp_implicit_aliases_done): ... this one.
5825 (cgraph_create_function_alias): Update.
5826 (cgraph_same_body_alias): Update.
c1e183a9 5827 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 5828 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 5829 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH
5830 (cgraph_node): Remove same_body_alias.
5831 (varpool_node): Remove alias_of and extra_name_alias.
5832 (same_body_aliases_done): Rename to ..
5833 (cpp_implicit_aliases_done): ... this one.
5834 (symtab_alias_ultimate_target): Add default parameter.
5835 (symtab_resolve_alias): New function.
5836 (fixup_same_cpp_alias_visibility): Declare.
5837 (cgraph_function_node): Add default parameter.
5838 (cgraph_node_asm_name): Likewise.
5839 (cgraph_function_or_thunk_node): Add default parameter; do
5840 not ICE when it is NULL.
5841 (varpool_variable_node): Likewise.
5842 * tree-emutls.c (create_emultls_var): Update.
5843 (ipa_lower_emutls): Update.
5844 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
5845 (cgraph_reset_node): Reset alias info.
5846 (cgraph_finalize_function): Update.
5847 (fixup_same_cpp_alias_visibility): Move to symtab.c.
5848 (analyze_function): Simplify.
5849 (cgraph_process_same_body_aliases): Simplify.
5850 (analyze_functions): Fixup same body aliases.
5851 (handle_alias_pairs): Simplify.
5852 (assemble_thunk): Update.
5853 (assemble_thunks_and_aliases): Update.
5854 (output_weakrefs): Rewrite.
5855 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
5856 (lto_output_varpool_node): Likewise.
5857 (compute_ltrans_boundary): Remve assert.
5858 (get_alias_symbol): New functoin.
5859 (input_node): Rewrite alias handling.
5860 (input_varpool_node): Likewise.
5861 * ipa-pure-const.c (propagate_pure_const): Fix formating.
5862 * ipa.c (process_references): Handle weakrefs correctly.
5863 (symtab_remove_unreachable_nodes): Likewise.
5864 * trans-mem.c (get_cg_data): Update.
5865 (ipa_tm_create_version_alias): Update.
5866 (ipa_tm_execute): Update.
5867 * symtab.c (dump_symtab_base): Dump aliases.
5868 (verify_symtab_base): Verify aliases.
5869 (symtab_node_availability): New function.
5870 (symtab_alias_ultimate_target): Simplify.
5871 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
5872 handle all the fixup cases.
5873 (symtab_resolve_alias): New function.
5874 * passes.c (ipa_write_summaries): Handle weakrefs.
5875 * varpool.c (varpool_analyze_node): Simplify.
5876 (assemble_aliases): Update.
5877 (varpool_create_variable_alias): Simplify.
5878 (varpool_extra_name_alias): Simplify.
5879 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
5880 (lto_symtab_merge_symbols): ... this one.
5881
21759881
DT
58822013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
5883
5884 Revert
5885 PR rtl-optimization/57268
5886 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5887 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5888
daf4e940
TB
58892013-06-01 Tobias Burnus <burnus@net-b.de>
5890
5891 Partially reverted:
5892 2013-05-31 Tobias Burnus <burnus@net-b.de>
5893
5894 PR middle-end/57073
5895 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
5896 further up.
5897
c359d8d0
DT
58982013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
5899
5900 PR rtl-optimization/57268
5901 * sched-deps.c (sched_analyze_2): Flush dependence lists if
5902 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
5903
4172215d
EB
59042013-05-31 Eric Botcazou <ebotcazou@adacore.com>
5905
5906 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
5907 unordered comparison operators when -fno-trapping-math is in effect
5908 on the e500.
5909 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
5910 and implement unordered comparison operators properly on the e500.
5911
a8c50132
EB
59122013-05-31 Eric Botcazou <ebotcazou@adacore.com>
5913
5914 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
5915 for constant scalar integers.
5916 (simplify_relational_operation_1): Likewise.
5917
44626634
SB
59182013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
5919
5920 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
5921 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
5922 Fix comment.
5923
55805e54
YR
59242013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
5925 Igor Zamyatin <igor.zamyatin@intel.com>
5926
5927 Silvermont (SLM) architecture performance tuning.
5928 * config/i386/i386.h (enum ix86_tune_indices): Add
5929 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
5930 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
5931
5932 * config/i386/i386.c (initial_ix86_tune_features)
5933 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
5934 (ix86_lea_outperforms): Handle Silvermont tuning.
5935 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
5936 call.
5937 (ix86_use_lea_for_mov): Likewise.
5938 (ix86_avoid_lea_for_addr): Likewise.
5939 (ix86_lea_for_add_ok): Likewise.
5940 (exact_dependency_1): New function.
5941 (exact_store_load_dependency): Likewise.
5942 (ix86_adjust_cost): Handle Silvermont tuning.
5943 (do_reoder_for_imul): Likewise.
5944 (swap_top_of_ready_list): New function.
5945 (ix86_sched_reorder): Changed to handle Silvermont tuning.
5946
5947 * config/i386/i386.md (peepholes that split memory operand in fp
5948 converts): New.
5949
e19c9de2
MS
59502013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
5951
5952 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5953 Remove un-necessary braces.
5954
38e6c9a6
MS
59552013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
5956
5957 * config/aarch64/aarch64.c (aarch64_classify_symbol):
5958 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
5959
e3530904
TB
59602013-05-31 Tobias Burnus <burnus@net-b.de>
5961
5962 PR middle-end/57073
c1e183a9 5963 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 5964
b6af05a9
KT
59652013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5966
5967 PR target/56315
5968 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
5969 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
5970 * config/arm/neon.md (iordi3_neon): Remove.
5971 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
5972 * config/arm/predicates.md (imm_for_neon_logic_operand):
5973 Move to earlier in the file.
5974 (neon_logic_op2): Likewise.
5975 (arm_iordi_operand_neon): New predicate.
5976
f800c166
RB
59772013-05-31 Richard Biener <rguenther@suse.de>
5978
5979 PR tree-optimization/57478
5980 PR tree-optimization/57453
5981 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
5982 are life as well.
5983
e571fa59
KP
59842013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5985
5986 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
5987 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
5988
0fa6e0ef
TB
59892013-05-30 Tobias Burnus <burnus@net-b.de>
5990 Thomas Koenig <tkoenig@gcc.gnu.org>
5991
5992 PR middle-end/57073
5993 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
5994 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
5995
4688ddf5
SB
59962013-05-30 Steven Bosscher <steven@gcc.gnu.org>
5997
0fa6e0ef 5998 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 5999
3b9ceb4b
VM
60002013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6001
6002 * target.def (register_usage_leveling_p): New hook.
6003 * targhooks.c (default_register_usage_leveling_p): New.
6004 * targhooks.h (default_register_usage_leveling_p): New prototype.
6005 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6006 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6007 * doc/tm.texi: Update.
6008 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6009
65f9e789
IB
60102013-05-30 Ian Bolton <ian.bolton@arm.com>
6011
6012 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6013 (*insv_reg<mode>): New define_insn.
6014
ca035367
JR
60152013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6016
6017 PR rtl-optimization/57439
6018 * postreload.c (move2add_valid_value_p): Check that we have
6019 a zero subreg_regno_offset when accessing the register in
6020 the requested mode.
6021
0b871ccf
YR
60222013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6023 Igor Zamyatin <igor.zamyatin@intel.com>
6024
6025 Silvermont (SLM) architecture pipeline model, tuning and
6026 insn selection.
6027 * config.gcc: Add slm config options and target.
6028
6029 * config/i386/slm.md: New.
6030
6031 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6032
6033 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6034 PROCESSOR_SLM.
6035 (ix86_target_macros_internal): Likewise.
6036
6037 * gcc/config/i386/i386.c (slm_cost): New cost.
6038 (m_SLM): New macro flag.
6039 (initial_ix86_tune_features): Set m_SLM.
6040 (x86_accumulate_outgoing_args): Likewise.
6041 (x86_arch_always_fancy_math_387): Likewise.
6042 (processor_target_table): Add slm cost.
6043 (cpu_names): Add slm cpu name.
6044 (x86_option_override_internal): Set SLM ISA.
6045 (ix86_issue_rate): New case PROCESSOR_SLM.
6046 (ia32_multipass_dfa_lookahead): Likewise.
6047 (fold_builtin_cpu): Add slm.
6048
6049 * config/i386/i386.h (TARGET_SLM): New target macro.
6050 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6051 (processor_type): Add PROCESSOR_SLM.
6052
6053 * config/i386/i386.md (cpu): Add new value "slm".
6054 (slm.md): Include slm.md.
6055
24d5b097
XG
60562013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6057 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6058
6059 * config/arm/arm-protos.h: Add and update function protos.
6060 * config/arm/arm.c (use_simple_return_p): New added.
6061 (thumb2_expand_return): Check simple_return flag.
6062 * config/arm/arm.md: Add simple_return and conditional simple_return.
6063 * config/arm/iterators.md: Add iterator for return and simple_return.
6064
c1cccc15
ZC
60652013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6066
6067 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6068 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6069 (arm_emit_vfp_multi_reg_pop): Likewise.
6070 (thumb2_emit_ldrd_pop): Likewise.
6071 (arm_expand_epilogue): Add misc REG_CFA notes.
6072 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6073
3a4f280b
LC
60742013-05-29 Lawrence Crowl <crowl@google.com>
6075
6076 * config/arm/t-arm: Update for below.
6077
6078 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6079 Change type to hash_table. Update dependent calls and types.
6080
6081 * config/i386/t-cygming: Update for below.
6082
6083 * config/i386/t-interix: Update for below.
6084
6085 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6086 Change type to hash_table. Update dependent calls and types.
6087 (i386_find_on_wrapper_list::wrappers): Likewise.
6088
6089 * config/ia64/t-ia64: Update for below.
6090
6091 * config/ia64/ia64.c (bundle_state_table):
6092 Change type to hash_table. Update dependent calls and types.
6093
6094 * config/mips/mips.c (mips_reorg_process_insns::htab):
6095 Change type to hash_table. Update dependent calls and types.
6096
6097 * config/sol2.c (solaris_comdat_htab):
6098 Change type to hash_table. Update dependent calls and types.
6099
6100 * config/t-sol2: Update for above.
6101
1388a0e3
TJ
61022013-05-29 Teresa Johnson <tejohnson@google.com>
6103
6104 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6105 functions are not yet marked as defined.
6106
a5965b52
MM
61072013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6108 Pat Haugen <pthaugen@us.ibm.com>
6109 Peter Bergner <bergner@vnet.ibm.com>
6110
6111 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6112 instructions.
6113 (VEC_A): Likewise.
6114 (VEC_C): Likewise.
6115 (vrotl<mode>3): Likewise.
6116 (vashl<mode>3): Likewise.
6117 (vlshr<mode>3): Likewise.
6118 (vashr<mode>3): Likewise.
6119
6120 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6121 support for power8 V2DI builtins.
6122
6123 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6124 power8 V2DI builtins.
6125 (vupkhsw): Likewise.
6126 (vupklsw): Likewise.
6127 (vaddudm): Likewise.
6128 (vminsd): Likewise.
6129 (vmaxsd): Likewise.
6130 (vminud): Likewise.
6131 (vmaxud): Likewise.
6132 (vpkudum): Likewise.
6133 (vpksdss): Likewise.
6134 (vpkudus): Likewise.
6135 (vpksdus): Likewise.
6136 (vrld): Likewise.
6137 (vsld): Likewise.
6138 (vsrd): Likewise.
6139 (vsrad): Likewise.
6140 (vsubudm): Likewise.
6141 (vcmpequd): Likewise.
6142 (vcmpgtsd): Likewise.
6143 (vcmpgtud): Likewise.
6144 (vcmpequd_p): Likewise.
6145 (vcmpgtsd_p): Likewise.
6146 (vcmpgtud_p): Likewise.
6147 (vupkhsw): Likewise.
6148 (vupklsw): Likewise.
6149 (vaddudm): Likewise.
6150 (vmaxsd): Likewise.
6151 (vmaxud): Likewise.
6152 (vminsd): Likewise.
6153 (vminud): Likewise.
6154 (vpksdss): Likewise.
6155 (vpksdus): Likewise.
6156 (vpkudum): Likewise.
6157 (vpkudus): Likewise.
6158 (vrld): Likewise.
6159 (vsld): Likewise.
6160 (vsrad): Likewise.
6161 (vsrd): Likewise.
6162 (vsubudm): Likewise.
6163
6164 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6165 support for power8 V2DI instructions.
6166
6167 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6168 power8 V2DI instructions. Combine pack and unpack insns to use an
6169 iterator for each mode. Check whether a particular mode supports
6170 Altivec instructions instead of just checking TARGET_ALTIVEC.
6171 (UNSPEC_VPKUWUM): Likewise.
6172 (UNSPEC_VPKSHSS): Likewise.
6173 (UNSPEC_VPKSWSS): Likewise.
6174 (UNSPEC_VPKUHUS): Likewise.
6175 (UNSPEC_VPKSHUS): Likewise.
6176 (UNSPEC_VPKUWUS): Likewise.
6177 (UNSPEC_VPKSWUS): Likewise.
6178 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6179 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6180 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6181 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6182 (UNSPEC_VUPKHSB): Likewise.
6183 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6184 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6185 (UNSPEC_VUPKHSH): Likewise.
6186 (UNSPEC_VUPKLSB): Likewise.
6187 (UNSPEC_VUPKLSH): Likewise.
6188 (VI2): Likewise.
6189 (VI_char): Likewise.
6190 (VI_scalar): Likewise.
6191 (VI_unit): Likewise.
6192 (VP): Likewise.
6193 (VP_small): Likewise.
6194 (VP_small_lc): Likewise.
6195 (VU_char): Likewise.
6196 (add<mode>3): Likewise.
6197 (altivec_vaddcuw): Likewise.
6198 (altivec_vaddu<VI_char>s): Likewise.
6199 (altivec_vadds<VI_char>s): Likewise.
6200 (sub<mode>3): Likewise.
6201 (altivec_vsubcuw): Likewise.
6202 (altivec_vsubu<VI_char>s): Likewise.
6203 (altivec_vsubs<VI_char>s): Likewise.
6204 (altivec_vavgs<VI_char>): Likewise.
6205 (altivec_vcmpbfp): Likewise.
6206 (altivec_eq<mode>): Likewise.
6207 (altivec_gt<mode>): Likewise.
6208 (altivec_gtu<mode>): Likewise.
6209 (umax<mode>3): Likewise.
6210 (smax<mode>3): Likewise.
6211 (umin<mode>3): Likewise.
6212 (smin<mode>3): Likewise.
6213 (altivec_vpkuhum): Likewise.
6214 (altivec_vpkuwum): Likewise.
6215 (altivec_vpkshss): Likewise.
6216 (altivec_vpkswss): Likewise.
6217 (altivec_vpkuhus): Likewise.
6218 (altivec_vpkshus): Likewise.
6219 (altivec_vpkuwus): Likewise.
6220 (altivec_vpkswus): Likewise.
6221 (altivec_vpks<VI_char>ss): Likewise.
6222 (altivec_vpks<VI_char>us): Likewise.
6223 (altivec_vpku<VI_char>us): Likewise.
6224 (altivec_vpku<VI_char>um): Likewise.
6225 (altivec_vrl<VI_char>): Likewise.
6226 (altivec_vsl<VI_char>): Likewise.
6227 (altivec_vsr<VI_char>): Likewise.
6228 (altivec_vsra<VI_char>): Likewise.
6229 (altivec_vsldoi_<mode>): Likewise.
6230 (altivec_vupkhsb): Likewise.
6231 (altivec_vupkhs<VU_char>): Likewise.
6232 (altivec_vupkls<VU_char>): Likewise.
6233 (altivec_vupkhsh): Likewise.
6234 (altivec_vupklsb): Likewise.
6235 (altivec_vupklsh): Likewise.
6236 (altivec_vcmpequ<VI_char>_p): Likewise.
6237 (altivec_vcmpgts<VI_char>_p): Likewise.
6238 (altivec_vcmpgtu<VI_char>_p): Likewise.
6239 (abs<mode>2): Likewise.
6240 (vec_unpacks_hi_v16qi): Likewise.
6241 (vec_unpacks_hi_v8hi): Likewise.
6242 (vec_unpacks_lo_v16qi): Likewise.
6243 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6244 (vec_unpacks_lo_v8hi): Likewise.
6245 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6246 (vec_pack_trunc_v8h): Likewise.
6247 (vec_pack_trunc_v4si): Likewise.
6248 (vec_pack_trunc_<mode>): Likewise.
6249
6250 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6251 V2DI builtins.
6252 (vec_vmaxsd): Likewise.
6253 (vec_vmaxud): Likewise.
6254 (vec_vminsd): Likewise.
6255 (vec_vminud): Likewise.
6256 (vec_vpksdss): Likewise.
6257 (vec_vpksdus): Likewise.
6258 (vec_vpkudum): Likewise.
6259 (vec_vpkudus): Likewise.
6260 (vec_vrld): Likewise.
6261 (vec_vsld): Likewise.
6262 (vec_vsrad): Likewise.
6263 (vec_vsrd): Likewise.
6264 (vec_vsubudm): Likewise.
6265 (vec_vupkhsw): Likewise.
6266 (vec_vupklsw): Likewise.
6267
e70670cf
JH
62682013-05-29 Jan Hubicka <jh@suse.cz>
6269
6270 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6271 flags; reorder rest of fields in more consistent way.
6272 (varpool_node): Remove analyzed, finalized and alias.
6273 (cgraph_ndoe): Likewise.
6274 (symtab_alias_ultimate_target): New function.
6275 (cgraph_function_node): Move offline.
6276 (cgraph_reset_node): Declare.
6277 (cgraph_comdat_can_be_unshared_p): Remove.
6278 (varpool_remove_initializer): Declare.
6279 (varpool_first_defined_variable, varpool_next_defined_variable
6280 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6281 (cgraph_function_with_gimple_body_p): Update.
6282 (varpool_all_refs_explicit_p): Update.
6283 (symtab_alias_target): New function.
6284 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6285 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB
6286 (cgraph_function_or_thunk_node): Simplify using
6287 symtab_alias_ultimate_target.
e70670cf
JH
6288 (varpool_variable_node): Likewise.
6289 * cgraph.c (cgraph_create_function_alias): Update.
6290 (cgraph_add_thunk): Update.
6291 (cgraph_remove_node): Update.
6292 (dump_cgraph_node): Do not dump removed flags.
6293 (cgraph_function_body_availability): Update.
6294 (cgraph_propagate_frequency): Update.
6295 (verify_cgraph_node): Check sanity of local flag.
6296 (cgraph_function_node): Move here from cgraph.h; revamp for
6297 cgraph_function_or_thunk_node.
6298 * lto-symtab.c (lto_varpool_replace_node): Update.
6299 (lto_symtab_resolve_can_prevail_p): Update.
6300 (lto_symtab_merge_cgraph_nodes): Update.
6301 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6302 propagate_constants_accross_call, devirtualization_time_bonus,
6303 ipcp_propagate_stage): Update.
6304 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB
6305 * ipa-inline-transform.c (clone_inlined_nodes,
6306 preserve_function_body_p): Update.
e70670cf
JH
6307 * ipa-reference.c (propagate): Update.
6308 (write_node_summary_p): Update.
6309 * toplev.c (wrapup_global_declaration_2): Update.
6310 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6311 (analyze_function) ... this one.
6312 (cgraph_process_new_functions): Update.
6313 (cgraph_reset_node): Export.
6314 (cgraph_finalize_function): Update.
6315 (cgraph_add_new_function): Update.
6316 (process_function_and_variable_attributes): Update.
6317 (varpool_finalize_decl): Update.
6318 (symbol_finalized): Remove.
6319 (symbol_finalized_and_needed): Rename to ...
6320 (symbol_defined_and_needed): ... update.
6321 (cgraph_analyze_functions): Update.
6322 (handle_alias_pairs): Update.
6323 (mark_functions_to_output): Update.
6324 (assemble_thunk): Update.
6325 (output_in_order): Update.
6326 (output_weakrefs): Update.
6327 (finalize_compilation_unit): Update.
6328 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
6329 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
6330 input_node, input_varpool_node): Update.
6331 * dbxout.c (dbxout_expand_expr): Update.
6332 * cgraphclones.c (cgraph_clone_node): Update.
6333 (cgraph_copy_node_for_versioning): Update.
6334 (cgraph_materialize_clone): Update.
6335 (cgraph_materialize_all_clones): Update.
6336 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
6337 propagate_pure_const, propagate_nothrow): Update.
6338 * lto-streamer-out.c (lto_output, write_symbol): Update.
6339 * ipa-utils.c (ipa_reverse_postorder): Update.
6340 * ipa-inline.c (can_inline_edge_p): Update.
6341 (update_caller_keys, ipa_inline): Update.
6342 * dwarf2out.c (reference_to_unused,
6343 premark_types_used_by_global_vars_helper): Update.
6344 * tree-eh.c (tree_could_trap_p): Update.
6345 * ipa-split.c (consider_split, execute_split_functions): Update.
6346 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB
6347 has_addr_references_p): Update; move ahead in file for better
6348 readability.
e70670cf
JH
6349 (process_references): Simplify.
6350 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
6351 bodies are removed.
6352 (cgraph_comdat_can_be_unshared_p): Make static.
6353 (cgraph_externally_visible_p): Update.
6354 (varpool_externally_visible_p): Update.
6355 (function_and_variable_visibility): Update.
6356 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
6357 ipa_tm_mark_force_output_node): Update.
6358 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
6359 estimate_edge_devirt_benefit, inline_generate_summary,
6360 inline_write_summary): Update.
6361 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
6362 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB
6363 (ipa_print_node_params, ipa_prop_read_section,
6364 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH
6365 * varasm.c (mark_decl_referenced): Update.
6366 (assemble_alias, dump_tm_clone_pairs): Update.
6367 * tree-inline.c (copy_bb): Update.
6368 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
6369 Update.
6370 * symtab.c (dump_symtab_base): Print new flags.
6371 (verify_symtab_base): Verify new flags.
6372 (symtab_alias_ultimate_target): New function.
6373 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
6374 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
6375 Update.
c1e183a9
UB
6376 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
6377 Update.
e70670cf
JH
6378 * i386.c (ix86_get_function_versions_dispatcher,
6379 ix86_generate_version_dispatcher_body): Update.
6380 (fold_builtin_cpu): Use varpool_add_new_variable.
6381 * varpool.c (varpool_remove_initializer): Break out from ...
6382 (varpool_remove_node): ... this one.
6383 (dump_varpool_node, varpool_node_for_asm,
6384 cgraph_variable_initializer_availability, varpool_analyze_node,
6385 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB
6386 varpool_finalize_named_section_flags, varpool_create_variable_alias):
6387 Update.
e70670cf 6388
182802ad
JH
63892013-05-29 Jan Hubicka <jh@suse.cz>
6390
6391 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
6392
8190b609
ER
63932013-05-29 Easwaran Raman <eraman@google.com>
6394
6395 PR tree-optimization/57442
6396 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
6397 when control exits the main loop.
6398
69f5aa9b
SKS
63992013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
6400
6401 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
6402 and RX600.
c1e183a9 6403 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS
6404 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
6405 * rx/t-rx: Add rx100 under multi library matches option for nofpu
6406 option.
6407
4b847da9
BS
64082013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6409
6410 PR tree-optimization/57441
6411 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
6412 Don't limit size of incr_vec to number of candidates.
6413
4fc43c39
SE
64142013-05-29 Steve Ellcey <sellcey@imgtec.com>
6415
6416 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
6417 and mips16 directories.
c1e183a9 6418 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE
6419 (MULTILIB_DIRNAMES): Ditto.
6420 (MULTILIB_EXCEPTIONS): Add new exceptions.
6421 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
6422 (MULTILIB_DIRNAMES): Ditto.
6423 (MULTILIB_EXCEPTIONS): Add new exceptions.
6424
a5350ddc
CSS
64252012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6426 Marcus Shawcroft <marcus.shawcroft@arm.com>
6427
6428 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
6429 SYMBOL_TINY_ABSOLUTE.
6430 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
6431 SYMBOL_TINY_ABSOLUTE.
6432 (aarch64_expand_mov_immediate): Likewise.
6433 (aarch64_classify_symbol): Likewise.
6434 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
6435 Permit SYMBOL_TINY_ABSOLUTE.
6436 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
6437
17f4d4bf
CSS
64382013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6439 Marcus Shawcroft <marcus.shawcroft@arm.com>
6440
6441 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
6442 Refactor if/switch. Replace gcc_assert with if.
6443
c0186656
GG
64442013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
6445
6446 * config/i386/i386.c (initial_ix86_tune_features): Enable
6447 FP Reassociation for AMD bdver1 and bdver2.
6448
d20188f3
MJ
64492013-05-29 Martin Jambor <mjambor@suse.cz>
6450
6451 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
6452 and IMAGPART_EXPR do not occur within other handled_components.
6453
292cba13
RB
64542013-05-29 Richard Biener <rguenther@suse.de>
6455
6456 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
6457 access on whether the use is in the BB we currently try to
6458 vectorize.
6459 (vect_bb_vectorization_profitable_p): Pass the BB we currently
6460 vectorize to vect_bb_slp_scalar_cost.
6461
6eddf228
RB
64622013-05-29 Richard Biener <rguenther@suse.de>
6463
6464 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
6465 computing scalar cost offsetted by stmts that are kept live
6466 by scalar uses.
6467 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
6468 for computation of scalar cost.
6469
7df36117
SE
64702013-05-28 Steve Ellcey <sellcey@mips.com>
6471
6472 * config/mips/mips-cpus.def (mips32r2): Change processor type.
6473
36536d79
BI
64742013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6475
6476 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
6477 array notation built-in reduction functions.
6478 * doc/passes.texi (Passes): Added documentation about changes done
6479 for Cilk Plus.
6480 * doc/invoke.texi (C Dialect Options): Added documentation about
6481 the -fcilkplus flag.
6482 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
6483 (BUILTINS_DEF): Depend on cilkplus.def.
6484 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
6485 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
6486 * cilkplus.def: New file.
6487
b49eefa5
JR
64882013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
6489
6490 PR rtl-optimization/57439
6491 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
6492
9daf6dbb
ER
64932013-05-28 Easwaran Raman <eraman@google.com>
6494
6495 PR tree-optimization/57337
6496 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
6497 (find_insert_point): Correctly identify the insertion point
6498 when two statements with the same UID is compared.
6499
fbd7e877
RB
65002013-05-28 Richard Biener <rguenther@suse.de>
6501
6502 PR tree-optimization/56787
6503 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
6504 from the list of data references.
6505 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
6506 clobbers.
6507 (vect_analyze_loop_operations): Likewise.
6508 (vect_transform_loop): Remove clobbers.
6509
bbba1117
MJ
65102013-05-28 Martin Jambor <mjambor@suse.cz>
6511
6512 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
6513 and REALPART_EXPRs have scalar type.
6514
bd388c2a
RB
65152013-05-28 Richard Biener <rguenther@suse.de>
6516
6517 PR tree-optimization/57411
6518 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
6519 virtual operands.
6520 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
6521 virtual operand propagation.
6522
2f56a311
EB
65232013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6524
6525 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
6526 destination register for bmasksi_vis.
6527 (vector_init_bshuffle): Likewise.
6528 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
6529
5c3eacbb
EB
65302013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6531
6532 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
6533 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
6534 mode if the instruction isn't available in the original mode.
6535 * config/sparc/sparc.opt (mfix-ut699): New option.
6536 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
6537 (divdf3): Turn into expander.
6538 (divdf3_nofix): New insn.
6539 (divdf3_fix): Likewise.
6540 (divsf3): Disable if -mfix-ut699.
6541 (sqrtdf2): Turn into expander.
6542 (sqrtdf2_nofix): New insn.
6543 (sqrtdf2_fix): Likewise.
6544 (sqrtsf2): Disable if -mfix-ut699.
6545
a1756c0a
RB
65462013-05-27 Richard Biener <rguenther@suse.de>
6547
6548 PR middle-end/57412
6549 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
6550 block for the new loop.
6551
5a892248
RB
65522013-05-27 Richard Biener <rguenther@suse.de>
6553
6554 PR tree-optimization/57343
6555 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
6556 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
6557 (number_of_iterations_cond): Do not build the folded tree.
6558
d1de852b
RB
65592013-05-27 Richard Biener <rguenther@suse.de>
6560
6561 Revert
6562 PR middle-end/57381
6563 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6564 OEP_CONSTANT_ADDRESS_OF retained.
6565
6566 PR tree-optimization/57417
6567 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
6568 for unchanged base.
6569 (set_ssa_val_to): Compare addresses using
6570 get_addr_base_and_unit_offset.
6571
7894bc6b
JR
65722013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
6573
6574 PR rtl-optimization/56833
6575 * postreload.c (move2add_record_mode): New function.
6576 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
6577 (move2add_use_add2_insn): Use move2add_record_sym_value.
6578 (move2add_use_add3_insn): Likewise.
6579 (reload_cse_move2add): Use move2add_valid_value_p and
6580 move2add_record_mode. Invalidate call-clobbered and REG_INC
6581 affected regs by setting reg_mode to VOIDmode.
6582 (move2add_note_store): Don't pretend the inside of a SUBREG is
6583 the actual destination. Invalidate single/leading registers by
6584 setting reg_mode to VOIDmode.
6585 Use move2add_record_sym_value, move2add_valid_value_p and
6586 move2add_record_mode.
6587
b03be25f
RB
65882013-05-27 Richard Biener <rguenther@suse.de>
6589
6590 PR tree-optimization/57396
6591 * tree-affine.c (double_int_constant_multiple_p): Properly
6592 return false for val == 0 and div != 0.
6593
44e88db2
RS
65942013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6595
6596 * config/mips/mips.h: Use #elif in preprocessor conditions.
6597
3b859704
RS
65982013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6599
6600 PR target/53916
6601 * config/mips/constraints.md (kl): New constraint.
6602 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
6603 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
6604 constraint for operand 0. Split after CSE for MIPS16. Emit a move
6605 from LO for MIPS16.
6606 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
6607
c979d5f5
RS
66082013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
6609
6610 PR target/55777
6611 * config/mips/mips.c (mips_can_inline_p): New function.
6612 (TARGET_CAN_INLINE_P): Define.
6613
8e90de43
SB
66142013-05-25 Steven Bosscher <steven@gcc.gnu.org>
6615
6616 * sched-int.h (ds_t, dw_t): Make unsigned int.
6617 Fix documentation that describes how all the ds_t bits are used.
6618 Reserve the last bit for delayed-branch scheduling.
6619 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
6620 (BITS_PER_DEP_WEAK): Fix definition and documentation.
6621 (gen_dep_weak_1): Remove prototype.
6622 * sched-deps.c (get_dep_weak_1): Make static.
6623 * target.def (speculate_insn, needs_block_p, gen_spec_check,
6624 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
6625 * doc/tm.texi: Regenerate.
6626 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
6627
cb5cb194
SB
66282013-05-24 Steven Bosscher <steven@gcc.gnu.org>
6629
6630 PR debug/56950
6631 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
6632
5bd93ff6
NS
66332013-05-24 Nathan Sidwell <nathan@codesourcery.com>
6634 Sandra Loosemore <sandra@codesourcery.com>
6635
6636 * config.gcc (powerpc-*): Allow native for with-cpu.
6637
2343af65
JL
66382013-05-24 Jeff Law <law@redhat.com>
6639
6640 PR tree-optimization/57124
6641 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
6642 conversion feeding a condition if the range has an overflow
6643 if -fstrict-overflow. Add warnings for when we do make the
6644 transformation.
6645
3ad695b9
DC
66462013-05-24 Dehao Chen <dehao@google.com>
6647
04960246 6648 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC
6649 (locus_discrim_hasher): Likewise.
6650 (locus_discrim_hasher::hash): Likewise.
6651 (locus_discrim_hasher::equal): Likewise.
6652
82338059
MJ
66532013-05-24 Martin Jambor <mjambor@suse.cz>
6654
6655 PR tree-optimization/57294
6656 * cgraph.h (ipa_record_stmt_references): Declare.
6657 * cgraphbuild.c (ipa_record_stmt_references): New function.
6658 (build_cgraph_edges): Use ipa_record_stmt_references.
6659 (rebuild_cgraph_edges): Likewise.
6660 (cgraph_rebuild_references): Likewise.
6661 * ipa-prop.c (ipa_modify_call_arguments): Discard references
6662 associated with the old statement and build references from the
6663 newly built statements.
6664 * ipa-ref.c (ipa_remove_stmt_references): New function.
6665 * ipa-ref.h (ipa_remove_stmt_references): Declare.
6666
1ccd4874
VM
66672013-05-24 Vladimir Makarov <vmakarov@redhat.com>
6668
55805e54 6669 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 6670 mem-mem moves.
55805e54 6671 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 6672 too.
55805e54 6673 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM
6674 secondary memory moves.
6675 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
6676 reg set up in the current insn.
6677
25e25c73
DC
66782013-05-24 Dehao Chen <dehao@google.com>
6679
04960246 6680 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC
6681 hash function.
6682 (locus_descrim_hasher::equal): Likewise.
04960246 6683 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC
6684 (make_edges): Likewise.
6685 (next_discriminator_for_locus): Likewise.
6686 (same_line_p): Likewise.
6687 (assign_discriminators): Likewise.
6688 (make_cond_expr_edges): Likewise.
6689 (make_gimple_switch_edges): Likewise.
6690 (make_goto_expr_edges): Likewise.
6691 (make_gimple_asm_edges): Likewise.
6692
50d38551
IB
66932013-05-24 Ian Bolton <ian.bolton@arm.com>
6694
6695 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
6696 X format specifier to only display bottom 16 bits.
6697 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
6698 immediate to match for operand 2, since it will be masked.
6699
aea0101d
RB
67002013-05-24 Richard Biener <rguenther@suse.de>
6701
6702 PR tree-optimization/57287
6703 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
6704 all SSA names that occur in abnormal PHIs.
6705
634e03d3
AI
67062013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
6707
6708 PR tree-ssa/57385
6709 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
6710 that index is not negative.
6711
b17c024f
EB
67122013-05-24 Eric Botcazou <ebotcazou@adacore.com>
6713
6714 PR rtl-optimization/55177
6715 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
6716 (simplify_byte_swapping_operation): New.
6717 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
6718 (simplify_relational_operation_1): Deal with BSWAP.
6719
46aeac1b
RH
67202013-05-23 Richard Henderson <rth@redhat.com>
6721
6722 PR target/56742
6723 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
6724 (ix86_reorg): Call it.
6725
70cc1536
UB
67262013-05-23 Uros Bizjak <ubizjak@gmail.com>
6727
6728 PR target/57379
6729 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
6730 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
6731 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
6732
b3851501
CB
67332013-05-23 Christian Bruel <christian.bruel@st.com>
6734
6735 PR debug/57351
6736 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
6737
4e194912
CSS
67382013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6739 Marcus Shawcroft <marcus.shawcroft@arm.com>
6740
6741 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
6742 * config/aarch64/constraints.md (Usa): Remove.
6743 * doc/md.texi (AArch64 Usa): Remove.
6744
83f8c414
CSS
67452013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6746 Marcus Shawcroft <marcus.shawcroft@arm.com>
6747
6748 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
6749 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
6750 * config/aarch64/predicates.md (aarch64_const_address): Remove.
6751 (aarch64_mov_operand): Use aarch64_mov_operand_p.
6752
b5574232
VP
67532013-05-23 Vidya Praveen <vidyapraveen@arm.com>
6754
6755 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
6756 instruction (AdvSIMD).
6757 * config/aarch64/aarch64-builtins.c
6758 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
6759 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
6760
76c36cb1 67612013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ
6762
6763 PR middle-end/57347
6764 * tree.h (contains_bitfld_component_ref_p): Declare.
6765 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB
6766 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
6767 caller.
7d2fb524
MJ
6768 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
6769 not access a bit-field. Assert all final offsets are byte-aligned.
6770
ce521ae6
RB
67712013-05-23 Richard Biener <rguenther@suse.de>
6772
6773 PR tree-optimization/57380
6774 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
6775 least one invariant or re-used load.
6776 * passes.c (init_optimization_passes): Move pass_phiprop before
6777 pass_forwprop.
6778
75c7257f
JG
67792013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
6780
6781 * config/aarch64/aarch64-simd.md
6782 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
6783
af6d44b5
RB
67842013-05-23 Richard Biener <rguenther@suse.de>
6785
6786 PR middle-end/57381
6787 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6788 OEP_CONSTANT_ADDRESS_OF retained.
6789
bd3647bf
JJ
67902013-05-23 Jakub Jelinek <jakub@redhat.com>
6791
6792 PR middle-end/57344
70cc1536
UB
6793 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
6794 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 6795
a55757ea
RB
67962013-05-23 Richard Biener <rguenther@suse.de>
6797
6798 PR rtl-optimization/57341
6799 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
6800 instead of true_dependence.
6801
cfb00b41
DM
68022013-05-22 David Malcolm <dmalcolm@redhat.com>
6803
dd1c676f
DM
6804 * bb-reorder.c (branch_threshold): Make const.
6805 (exec_threshold): Ditto.
cfb00b41 6806
f62511da
MM
68072013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6808 Pat Haugen <pthaugen@us.ibm.com>
6809 Peter Bergner <bergner@vnet.ibm.com>
6810
6811 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
6812 documentation for the power8 crypto builtins.
6813
6814 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
6815
6816 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
6817 macros for defining power8 builtin functions.
6818 (BU_P8V_AV_2): Likewise.
6819 (BU_P8V_AV_P): Likewise.
6820 (BU_P8V_VSX_1): Likewise.
6821 (BU_P8V_OVERLOAD_1): Likewise.
6822 (BU_P8V_OVERLOAD_2): Likewise.
6823 (BU_CRYPTO_1): Likewise.
6824 (BU_CRYPTO_2): Likewise.
6825 (BU_CRYPTO_3): Likewise.
6826 (BU_CRYPTO_OVERLOAD_1): Likewise.
6827 (BU_CRYPTO_OVERLOAD_2): Likewise.
6828 (XSCVSPDP): Fix typo, point to the correct instruction.
6829 (VCIPHER): Add power8 crypto builtins.
6830 (VCIPHERLAST): Likewise.
6831 (VNCIPHER): Likewise.
6832 (VNCIPHERLAST): Likewise.
6833 (VPMSUMB): Likewise.
6834 (VPMSUMH): Likewise.
6835 (VPMSUMW): Likewise.
6836 (VPERMXOR_V2DI): Likewise.
6837 (VPERMXOR_V4SI: Likewise.
6838 (VPERMXOR_V8HI: Likewise.
6839 (VPERMXOR_V16QI: Likewise.
6840 (VSHASIGMAW): Likewise.
6841 (VSHASIGMAD): Likewise.
6842 (VPMSUM): Likewise.
6843 (VPERMXOR): Likewise.
6844 (VSHASIGMA): Likewise.
6845
6846 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6847 __CRYPTO__ if the crypto instructions are available.
6848 (altivec_overloaded_builtins): Add support for overloaded power8
6849 builtins.
6850
6851 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
6852 support for power8 crypto builtins.
6853 (builtin_function_type): Likewise.
6854 (altivec_init_builtins): Add support for builtins that take vector
6855 long long (V2DI) arguments.
6856
6857 * config/rs6000/crypto.md: New file, define power8 crypto
6858 instructions.
6859
68602013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
6861 Pat Haugen <pthaugen@us.ibm.com>
6862 Peter Bergner <bergner@vnet.ibm.com>
6863
6864 * doc/invoke.texi (Option Summary): Add power8 options.
6865 (RS/6000 and PowerPC Options): Likewise.
6866
6867 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
6868 constraints.md instead of rs6000.h. Reorder w* constraints. Add
6869 wm, wn, wr documentation.
6870
6871 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
6872 registers if direct move instructions are enabled.
6873 (wn): New constraint for no registers.
6874 (wq): New constraint for quad word even GPR registers.
6875 (wr): New constraint if 64-bit instructions are enabled.
6876 (wv): New constraint if power8 vector instructions are enabled.
6877 (wQ): New constraint for quad word memory locations.
6878
6879 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
6880 constraint for 0..15 for crypto instructions.
6881 (gpc_reg_operand): If VSX allow registers in VSX registers as well
6882 as GPR and floating point registers.
6883 (int_reg_operand): New predicate to match only GPR registers.
6884 (base_reg_operand): New predicate to match base registers.
6885 (quad_int_reg_operand): New predicate to match even GPR registers
6886 for quad memory operations.
6887 (vsx_reg_or_cint_operand): New predicate to allow vector logical
6888 operations in both GPR and VSX registers.
6889 (quad_memory_operand): New predicate for quad memory operations.
6890 (reg_or_indexed_operand): New predicate for direct move support.
6891
6892 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
6893 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
6894 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
6895 (POWERPC_MASKS): Add power8 options.
6896 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
6897 various options.
6898
6899 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6900 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
6901
6902 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
6903 (-mpower8-fusion): New power8 options.
6904 (-mpower8-fusion-sign): Likewise.
6905 (-mpower8-vector): Likewise.
6906 (-mcrypto): Likewise.
6907 (-mdirect-move): Likewise.
6908 (-mquad-memory): Likewise.
6909
6910 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
6911 power8.
6912 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
6913 registers.
70cc1536 6914 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM
6915 (rs6000_debug_vector_unit): Add p8_vector.
6916 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
6917 definitions. Also print fusion state.
6918 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
6919 (rs6000_builtin_mask_calculate): Add power8 builtin support.
6920 (rs6000_option_override_internal): Add support for power8.
6921 (rs6000_common_init_builtins): Add debugging for skipped builtins
6922 if -mdebug=builtin.
6923 (rs6000_adjust_cost): Add power8 support.
6924 (rs6000_issue_rate): Likewise.
6925 (insn_must_be_first_in_group): Likewise.
6926 (insn_must_be_last_in_group): Likewise.
6927 (force_new_group): Likewise.
6928 (rs6000_register_move_cost): Likewise.
6929 (rs6000_opt_masks): Likewise.
6930
6931 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
6932 power8 capable assembler, default to power7 options.
6933 (TARGET_DIRECT_MOVE): Likewise.
6934 (TARGET_CRYPTO): Likewise.
6935 (TARGET_P8_VECTOR): Likewise.
6936 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
6937 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
6938 (VECTOR_MEM_P8_VECTOR_P): Likewise.
6939 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
6940 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
6941 (TARGET_XSCVDPSPN): Likewise.
6942 (TARGET_XSCVSPDPN): Likewsie.
6943 (TARGET_SYNC_HI_QI): Likewise.
6944 (TARGET_SYNC_TI): Likewise.
6945 (MASK_CRYPTO): Likewise.
6946 (MASK_DIRECT_MOVE): Likewise.
6947 (MASK_P8_FUSION): Likewise.
6948 (MASK_P8_VECTOR): Likewise.
70cc1536
UB
6949 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
6950 temporary used by some of the direct move instructions to get two FP
6951 temporary registers does not force creation of a stack frame.
f62511da
MM
6952 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
6953 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
6954 that any VSX registers are tieable, even if they are also an
6955 Altivec vector mode.
6956 (r6000_reg_class_enum): Add wm, wr, wv constraints.
6957 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
6958 (RS6000_BTM_CRYPTO): Likewise.
6959 (RS6000_BTM_COMMON): Likewise.
6960
6961 * config/rs6000/rs6000.md (cpu attribute): Add power8.
6962 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
6963 (enum rs6000_vector): Add power8 vector support.
6964
73a1a707
RR
69652013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6966
6967 PR target/19599
6968 PR target/57340
6969 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
6970 (any_sibcall_could_use_r3): this and handle indirect calls.
6971 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
6972
d28073d4
BS
69732013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6974
6975 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
6976
bee0b10c
RB
69772013-05-22 Richard Biener <rguenther@suse.de>
6978
6979 PR middle-end/57349
6980 * profile.c (branch_prob): Do not split blocks that are
6981 abnormally receiving from ECF_RETURNS_TWICE functions.
6982
98409b51
RS
69832013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6984
6985 * recog.c (offsettable_address_addr_space_p): Fix calculation of
6986 address mode. Move pointer mode initialization to the same place.
6987
c0602ab8
MZ
69882013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6989
6990 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
6991 while it has any effect.
6992
4f7a634e
ER
69932013-05-21 Easwaran Raman <eraman@google.com>
6994
6995 PR tree-optimization/57322
9daf6dbb
ER
6996 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
6997 UID of the statement added to the BB to be 1.
4f7a634e 6998
a4ce1258
JJ
69992013-05-21 Jakub Jelinek <jakub@redhat.com>
7000
7001 PR tree-optimization/57331
70cc1536
UB
7002 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7003 of conversion from pointer type to integral type with integer.
a4ce1258 7004
1b14621a
MJ
70052013-05-21 Martin Jambor <mjambor@suse.cz>
7006
7007 PR lto/57289
7008 * ipa-prop.c (ipa_read_node_info): Process param_used and
7009 controlled_uses in the same order as when writing.
7010
e60661f0
MG
70112013-05-21 Magnus Granberg <baldrick@free.fr>
7012
7013 PR plugins/56754
ee49aa34 7014 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 7015
4a61be9a
RB
70162013-05-21 Richard Biener <rguenther@suse.de>
7017
7018 PR tree-optimization/57318
7019 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7020 estimate stmts with side-effects as likely eliminated.
7021
c52da5f7
RB
70222013-05-21 Richard Biener <rguenther@suse.de>
7023
7024 PR tree-optimization/57330
7025 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7026 preserve the call stmts fntype.
7027
7ec67e2a
RB
70282013-05-21 Richard Biener <rguenther@suse.de>
7029
7030 PR tree-optimization/57303
7031 * tree-ssa-sink.c (statement_sink_location): Improve killing
7032 stmt detection and properly handle self-assignments.
7033
b112d513
CB
70342013-05-21 Christian Bruel <christian.bruel@st.com>
7035
55805e54
YR
7036 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7037 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB
7038 registers. Set register size out of the PARALLEL loop.
7039
14c2ec26
OE
70402013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7041
7042 PR target/56547
7043 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7044 (*fmasf4, *fmasf4_media): New insns.
7045
da734fa1
RS
70462013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7047
7048 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7049 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7050 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7051 (mips_idiv_insns): Update the comments to say that the returned
7052 instruction counts are in units of BASE_INSN_LENGTH.
7053 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7054 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7055 using 2 rather than 4 as the length of indirect MIPS16 and
7056 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7057 length of a NOP. Don't divide MIPS16 lengths by 2.
7058 (mips16_split_long_branches): Assume a branch is long if the
7059 length is greater than 4 rather than 8.
7060 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7061 rather than multiplying them by 2. Multiply instruction counts
7062 by BASE_INSN_LENGTH rather than 4.
7063 (*jump_mips16, tls_get_tp_mips16_<mode>)
7064 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7065
13719e8b
RS
70662013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7067
7068 * config/mips/mips.md (extended_mips16): Remove branch case.
7069 (length): Remove duplicated extended_mips16 test.
7070
c3850d14
RS
70712013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7072
7073 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7074
8da2e059
RS
70752013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7076
7077 * recog.h (Recog_data): Rename to...
7078 (recog_data_d): ...this.
7079 (recog_data): Update accordingly.
7080 * recog.c (recog_data): Likewise.
7081 * reload.c (save_recog_data): Likewise.
7082 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7083 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7084
2df013f3
JB
70852013-05-17 Julian Brown <julian@codesourcery.com>
7086
7087 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7088 found in a REG_EQUAL note, invalidate it.
7089
371e77e3 70902013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER
7091
7092 * tree-ssa-reassoc.c (find_insert_point): New function.
7093 (insert_stmt_after): Likewise.
7094 (get_def_stmt): Likewise.
7095 (ensure_ops_are_available): Likewise.
7096 (not_dominated_by): Likewise.
7097 (rewrite_expr_tree): Do not move statements beyond what is
7098 necessary. Remove call to swap_ops_for_binary_stmt...
7099 (reassociate_bb): ... and move it here.
7100 (build_and_add_sum): Assign UIDs for new statements.
7101 (linearize_expr): Likewise.
7102 (do_reassoc): Renumber gimple statement UIDs.
7103
e01c7cca
JH
71042013-05-17 Jan Hubicka <jh@suse.cz>
7105
7106 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7107 weakrefs.
7108 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB
7109 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7110 weakrefs.
e01c7cca
JH
7111 (output_weakrefs): Update.
7112
c3272a92
PCC
71132013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7114 Martin Jambor <mjambor@suse.cz>
7115
7116 PR middle-end/57276
7117 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7118 value that corresponds to the given aggval is found in values vector.
7119
11c2aa39
UB
71202013-05-17 Uros Bizjak <ubizjak@gmail.com>
7121
7122 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7123 sse, sse2, sse3, ssse3 and sse4a flags to options.
7124
abecc8c6
DM
71252013-05-17 David Malcolm <dmalcolm@redhat.com>
7126
7127 * gengtype-state.c: (s_expr_writer): New class, to handle
7128 prettifying of output layout of s-expressions.
7129 (state_writer): New class, to write out gtype.state.
7130 (state_written_type_count): Move this variable into member data of
7131 state_writer.
7132 (s_expr_writer::s_expr_writer): New code: constructor for new class
7133 (state_writer::state_writer(): ditto
7134 (s_expr_writer::write_new_line): New function
7135 (s_expr_writer::write_any_indent): ditto
7136 (s_expr_writer::begin_s_expr): ditto
7137 (s_expr_writer::end_s_expr): ditto
7138 (write_state_fileloc): convert to method of state_writer...
7139 (state_writer:: write_state_fileloc): ...and use methods of
7140 s_expr_writer to write indentation into the gtype.state output file
7141 to visually represent the hierarchical structure of the list
7142 structures
7143 (write_state_fields): ditto, renaming to...
7144 (state_writer::write_state_fields)
7145 (write_state_a_string): ditto, renaming to...
7146 (state_writer::write_state_a_string)
7147 (write_state_string_option): ditto, renaming to...
7148 (state_writer::write_state_string_option)
7149 (write_state_type_option): ditto, renaming to...
7150 (state_writer::write_state_type_option)
7151 (write_state_nested_option): ditto, renaming to...
7152 (state_writer::write_state_nested_option)
7153 (write_state_option): ditto, renaming to...
7154 (state_writer::write_state_option)
7155 (write_state_options): ditto, renaming to...
7156 (state_writer::write_state_options)
7157 (write_state_lang_bitmap): ditto, renaming to...
7158 (state_writer::write_state_lang_bitmap)
7159 (write_state_version): ditto, renaming to...
7160 (state_writer::write_state_version)
7161 (write_state_scalar_type): ditto, renaming to...
7162 (state_writer::write_state_scalar_type)
7163 (write_state_string_type): ditto, renaming to...
7164 (state_writer::write_state_string_type)
7165 (write_state_undefined_type): ditto, renaming to...
7166 (state_writer::write_state_undefined_type)
7167 (write_state_struct_union_type): ditto, renaming to...
7168 (state_writer::write_state_struct_union_type)
7169 (write_state_struct_type): ditto, renaming to...
7170 (state_writer::write_state_struct_type)
7171 (write_state_user_struct_type): ditto, renaming to...
7172 (state_writer::write_state_user_struct_type)
7173 (write_state_lang_struct_type): ditto, renaming to...
7174 (state_writer::write_state_lang_struct_type)
7175 (write_state_param_struct_type): ditto, renaming to...
7176 (state_writer::write_state_param_struct_type)
7177 (write_state_pointer_type): ditto, renaming to...
7178 (state_writer::write_state_pointer_type)
7179 (write_state_array_type): ditto, renaming to...
7180 (state_writer::write_state_array_type)
7181 (write_state_gc_used): ditto, renaming to...
7182 (state_writer::write_state_gc_used)
7183 (write_state_common_type_content): ditto, renaming to...
7184 (state_writer::write_state_common_type_content)
7185 (write_state_type): ditto, renaming to...
7186 (state_writer::write_state_type)
7187 (write_state_pair_list): ditto, renaming to...
7188 (state_writer::write_state_pair_list)
7189 (write_state_pair): ditto, renaming to...
7190 (state_writer::write_state_pair)
7191 (write_state_typedefs): ditto, renaming to...
7192 (state_writer::write_state_typedefs)
7193 (write_state_structures): ditto, renaming to...
7194 (state_writer::write_state_structures)
7195 (write_state_param_structs): ditto, renaming to...
7196 (state_writer::write_state_param_structs)
7197 (write_state_variables): ditto, renaming to...
7198 (state_writer::write_state_variables)
7199 (write_state_srcdir): ditto, renaming to...
7200 (state_writer::write_state_srcdir)
7201 (write_state_files_list): ditto, renaming to...
7202 (state_writer::write_state_files_list)
7203 (write_state_languages): ditto, renaming to...
7204 (state_writer::write_state_languages)
7205 (write_state): create a state_writer instance and use it when
7206 writing out the state file
7207
d6545f29
MS
72082013-05-17 Mike Stump <mikestump@comcast.net>
7209
816a3f73 7210 PR rtl-optimization/57304
d6545f29
MS
7211 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7212 accessing DF_REF_REAL_LOC.
7213
38c821cf
JJ
72142013-05-17 Jakub Jelinek <jakub@redhat.com>
7215
7216 PR rtl-optimization/57281
7217 PR rtl-optimization/57300
7218 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7219 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7220 what the other splitter did if the registers are dead.
7221
2a293391
RB
72222013-05-17 Richard Biener <rguenther@suse.de>
7223
7224 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7225 MEM_REF offsets.
7226
17042d2b
JJ
72272013-05-17 Jakub Jelinek <jakub@redhat.com>
7228
7229 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7230 linking.
7231
5b115c1f
MP
72322013-05-17 Marek Polacek <polacek@redhat.com>
7233
11c2aa39
UB
7234 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7235 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 7236
52d84413
JJ
72372013-05-17 Jakub Jelinek <jakub@redhat.com>
7238
68119618
JJ
7239 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7240 vect_external_def oprnd1 with loop_vinfo, try to emit
7241 optional cast, negation and and stmts on the loop preheader
7242 edge instead of into the pattern def seq.
7243
52d84413
JJ
7244 PR tree-optimization/57051
7245 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7246 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7247
17b962bd
NC
72482013-05-16 Nick Clifton <nickc@redhat.com>
7249
7250 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7251 (rl78_is_naked_func): New function.
56aefbf7
UB
7252 (rl78_expand_prologue): Skip prologue generation for naked functions.
7253 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC
7254 * doc/extend.texi (naked): Add RL78 to the list of processors
7255 that supports this attribute.
7256
b1a0f84e
JL
72572013-05-16 Jeff Law <law@redhat.com>
7258
7259 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7260
af0e415b
UB
72612013-05-16 Uros Bizjak <ubizjak@gmail.com>
7262
7263 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7264 cache parameters using detect_caches_amd also for CYRIX,
7265 NSC and TM2 signatures.
7266
19db293a
UB
72672013-05-16 Uros Bizjak <ubizjak@gmail.com>
7268 Dzianis Kahanovich <mahatma@eu.by>
7269
7270 PR target/45359
7271 PR target/46396
7272 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7273 VIA/Centaur processors and determine their cache parameters
7274 using detect_caches_amd.
7275
251a41b9
TJ
72762013-05-16 Teresa Johnson <tejohnson@google.com>
7277
7278 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7279 (rtl_verify_edges): New function.
7280 (rtl_verify_bb_insns): Ditto.
7281 (rtl_verify_bb_pointers): Ditto.
7282 (rtl_verify_bb_insn_chain): Ditto.
7283 (rtl_verify_fallthru): Ditto.
7284 (rtl_verify_bb_layout): Ditto.
7285 (rtl_verify_flow_info_1): Outline checks into new functions.
7286 (rtl_verify_flow_info): Ditto.
7287
f14540b6
SE
72882013-05-16 Steve Ellcey <sellcey@imgtec.com>
7289
7290 * cfghooks.c (copy_bbs): Add update_dominance argument.
7291 * cfghooks.h (copy_bbs): Update prototype.
7292 * tree-cfg.c (gimple_duplicate_sese_region):
7293 Add update_dominance argument.
7294 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7295 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7296 gimple_duplicate_sese_region call.
7297 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7298 Update copy_bbs call.
7299 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7300 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7301
7e9a3abb
JJ
73022013-05-16 Jakub Jelinek <jakub@redhat.com>
7303
7304 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7305 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7306 vect_recog_rotate_pattern.
7307 (vect_recog_rotate_pattern): New function.
7308
427b248d
JM
73092013-05-16 Jason Merrill <jason@redhat.com>
7310
7311 * Makefile.in (LLINKER): New variable.
7312 (mostlyclean): Remove link mutex.
7313 * configure.ac: Handle --enable-link-mutex.
7314 * lock-and-run.sh: New script.
7315
b871e3d2
RR
73162013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7317
7318 PR target/19599
7319 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
7320 for NULL decl.
7321
ce858126
RO
73222013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7323
7324 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
7325
798d3d04
GY
73262013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
7327
7328 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
7329 * config/arm/arm.c (next_consecutive_mem): New function.
7330 (gen_movmem_ldrd_strd): Likewise.
7331 * config/arm/arm.md (movmemqi): Update condition and code.
7332 (unaligned_loaddi, unaligned_storedi): New patterns.
7333
0baddc45
RO
73342013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7335
7336 * config.gcc: Obsolete *-*-solaris2.9*.
7337 * doc/install.texi (Specific, *-*-solaris2*): Document it.
7338
aa9d5bcf
RB
73392013-05-16 Richard Biener <rguenther@suse.de>
7340
7341 * passes.c (init_optimization_passes): Move pass_parallelize_loops
7342 earlier, after GRAPHITE transforms and IV canonicalization.
7343
5a0f4dd3
JJ
73442013-05-16 Jakub Jelinek <jakub@redhat.com>
7345
7346 * omp-low.c (extract_omp_for_data): For collapsed loops,
7347 if at least one of the loops is known at compile time to
7348 iterate zero times, set count to 0.
7349 (expand_omp_regimplify_p): New function.
7350 (expand_omp_for_generic): For collapsed loops, if at least
7351 one of the loops isn't known to iterate at least once,
7352 add runtime check with setting count to 0.
7353 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7354 For unsigned types if it isn't known at compile time that
7355 the loop will iterate at least once, add runtime check to bypass
7356 the whole loop if initial condition isn't true.
7357
e3753785
NS
73582013-05-16 Nathan Sidwell <nathan@codesourcery.com>
7359
7360 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
7361
43bb4dd1
MG
73622013-05-16 Marc Glisse <marc.glisse@inria.fr>
7363
7364 PR middle-end/57286
7365 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
7366 transformations to avoid an infinite loop.
7367
3571dde6
MP
73682013-05-16 Marek Polacek <polacek@redhat.com>
7369
7370 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
7371
b8b3f0ca
LE
73722013-05-15 Leif Ekblad <leif@rdos.net>
7373
7374 * config/i386/i386.c (ix86_decompose_address): Use
7375 DEFAULT_TLS_SEG_REG to access TLS segment register.
7376 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
7377 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
7378 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
7379
e299a383
RS
73802013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
7381
7382 PR target/57260
7383 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
7384 sibling calls to functions that would normally be lazily bound,
7385 unless $gp is call-clobbered.
7386
e7208ea3
UB
73872013-05-15 Uros Bizjak <ubizjak@gmail.com>
7388
19db293a 7389 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 7390 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB
7391 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
7392 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB
7393 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
7394 of TARGET_3DNOW.
7395 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
7396
e7413f3d
AS
73972013-05-15 Andreas Schwab <schwab@suse.de>
7398
7399 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
7400 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
7401 third operand.
7402
ad4db775
TJ
74032013-05-15 Teresa Johnson <tejohnson@google.com>
7404
7405 * loop-unroll.c (report_unroll_peel): Check decision before
7406 emitting unroll/peel message.
7407
af205f67
TJ
74082013-05-15 Teresa Johnson <tejohnson@google.com>
7409
7410 * function.h (has_bb_partition): New rtl_data flag.
7411 (bb_reorder_complete): Ditto.
7412 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
7413 instead of flag_reorder_blocks_and_partition.
7414 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
7415 with some enhancements.
7416 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
7417 * bb-reorder.c (connect_traces): Check for has_bb_partition
7418 instead of flag_reorder_blocks_and_partition.
7419 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
7420 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
7421 verify_hot_cold_block_grouping.
7422 (partition_hot_cold_basic_blocks): Set has_bb_partition.
7423
9adcfa3c
RR
74242013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7425
7426 PR target/19599
7427 * config/arm/predicates.md (call_insn_operand): New predicate.
7428 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
7429 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
7430 if insn is not a tail call.
7431 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
7432 registers.
7433 * config/arm/arm.h (enum reg_class): New caller save register class.
7434 (REG_CLASS_NAMES): Likewise.
7435 (REG_CLASS_CONTENTS): Likewise.
7436 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
7437 without decls.
7438
ccb3ad87
RB
74392013-05-15 Richard Biener <rguenther@suse.de>
7440
7441 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
7442 of MSG_OPTIMIZED_LOCATIONS.
7443 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
7444 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
7445 message.
7446 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
7447 of MSG_OPTIMIZED_LOCATIONS.
7448 (execute_vect_slp): Likewise.
7449 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
7450 (vect_create_cond_for_alias_checks): Likewise.
7451 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
7452 (vect_recog_widen_mult_pattern): Likewise.
7453 (vect_recog_widen_sum_pattern): Likewise.
7454 (vect_recog_over_widening_pattern): Likewise.
7455 (vect_recog_widen_shift_pattern): Likewise.
7456 (vect_recog_vector_vector_shift_pattern): Likewise.
7457 (vect_recog_divmod_pattern): Likewise.
7458 (vect_recog_mixed_size_cond_pattern): Likewise.
7459 (vect_recog_bool_pattern): Likewise.
7460 (vect_pattern_recog_1): Likewise.
7461
48b1474e
MJ
74622013-05-15 Martin Jambor <mjambor@suse.cz>
7463
7464 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
7465 non-functions to builtin_unreachable.
7466 * ipa-inline-transform.c (inline_call): Do not assert estimates were
7467 correct when new direct edges were discovered.
7468
9de04252
MJ
74692013-05-15 Martin Jambor <mjambor@suse.cz>
7470
7471 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
7472 header, print symbol order instead of node uid, print more information
7473 about indirect edge targets.
7474 (ipa_make_edge_direct_to_target): Print symbol order instead of node
7475 uids.
7476 (ipa_make_edge_direct_to_target): Likewise.
7477 (remove_described_reference): Likewise.
7478 (propagate_controlled_uses): Likewise.
7479 (ipa_print_node_params): Also print symbol order.
7480 (ipcp_transform_function): Print symbol order instead of node uids.
7481 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
7482 (cgraph_get_create_real_symbol_node): Likewise.
7483 * ipa-cp.c (print_lattice): Likewise.
7484 (print_all_lattices): Likewise.
7485 (determine_versionability): Likewise.
7486 (initialize_node_lattices): Likewise.
7487 (estimate_local_effects): Likewise.
7488 (update_profiling_info): Likewise.
7489 (create_specialized_node): Likewise.
7490 (perhaps_add_new_callers): Likewise.
7491 (decide_about_value): Likewise.
7492 (decide_whether_version_node): Likewise.
7493 (identify_dead_nodes): Likewise.
7494 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
7495 (dump_inline_summary): Likewise.
7496 (estimate_node_size_and_time): Likewise.
7497 (inline_analyze_function): Likewise.
7498 * ipa-inline.c (report_inline_failed_reason): Likewise.
7499 (want_early_inline_function_p): Likewise.
7500 (edge_badness): Likewise.
7501 (update_edge_key): Likewise.
7502 (inline_small_functions): Likewise. Add dumping of order to two other
7503 dumps.
7504 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
7505 instead of node uids.
7506 (propagate_pure_const): Likewise.
7507 (propagate_pure_const): Likewise.
7508 * ipa-utils.c (dump_cgraph_node_set): Likewise.
7509 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
7510 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
7511 of node uids.
7512 * tree-pretty-print.c (dump_function_header): Likewise.
7513 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
7514 Print symbol order instead of node uids.
7515
1dd03b91
AK
75162013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7517
7518 * config/s390/s390.c (s390_register_move_cost): Don't impose the
7519 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
7520
fdf6a7b9
RB
75212013-05-15 Richard Biener <rguenther@suse.de>
7522
7523 PR tree-optimization/57275
7524 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
7525 return value for fail to do runtime alias checks for gather loads.
7526
2d6e4603
JH
75272013-05-15 Jan Hubicka <jh@suse.cz>
7528
7529 PR lto/57038
7530 PR lto/47375
e7208ea3
UB
7531 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
7532 weakrefs are not external.
7533 (lto_symtab_merge_decls): Fix thinko when dealing with
7534 non-lto_symtab decls.
2d6e4603
JH
7535 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
7536 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
7537 * varpool.c (dump_varpool_node): Dump more flags.
7538
83f44b39
GG
75392013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7540
7541 * config/i386/i386.c (processor_alias_table): Add instruction
7542 FSGSBASE for AMD bdver3 architecture.
7543
2e55d062
JJ
75442013-05-14 Jakub Jelinek <jakub@redhat.com>
7545
7546 * tree.c (warn_deprecated_use): Print file:line using locus color.
7547 * diagnostic.c (diagnostic_report_current_module): Print file:line
7548 and file:line:column using locus color.
7549
0bfdb81e
MS
75502013-05-14 Mike Stump <mikestump@comcast.net>
7551
7552 * gdbinit.in: Add __null.
7553
a508ef22
MS
75542013-05-14 Mike Stump <mikestump@comcast.net>
7555
7556 * recog.h: Rename struct recog_data to Recog_data.
7557 * recog.c: Likewise.
7558 * reload.c (can_reload_into): Likewise.
7559 * config/picochip/picochip.c: Likewise.
7560
e7180acb
MS
75612013-05-14 Mike Stump <mikestump@comcast.net>
7562
7563 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
7564
61216c26
SB
75652013-05-14 Steven Bosscher <steven@gcc.gnu.org>
7566
fd6beed4
SB
7567 * resource.h (struct resources): Remove unch_memory member.
7568 (CLEAR_RESOURCE): Don't clear unch_memory.
7569 * resource.c (mark_referenced_resources): Don't set it.
7570 (mark_set_resources): Likewise.
7571 (mark_target_live_regs): Don't clear it.
7572 (init_resource_info): Likewise.
7573 * reorg.c (resource_conflicts_p): Don't compare it.
7574 (redundant_insn): Don't set it.
7575
61216c26
SB
7576 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
7577 Remove prototypes.
7578 * emit-rtl.c (next_label): Remove unused function.
7579 (skip_consecutive_labels, link_cc0_insns): Move to ...
7580 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
7581 only place where these functions are used, and make them static.
7582
418d1b87
MG
75832013-05-14 Marc Glisse <marc.glisse@inria.fr>
7584
7585 * fold-const.c (fold_negate_expr): Handle vectors.
7586 (fold_truth_not_expr): Make it static.
7587 (fold_invert_truthvalue): New static function.
7588 (invert_truthvalue_loc): Handle vectors. Do not call
7589 fold_truth_not_expr directly.
7590 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
7591 <TRUTH_NOT_EXPR>: Do not cast to boolean.
7592 (fold_comparison): Handle vector constants.
7593 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
7594 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
7595 * tree.h (fold_truth_not_expr): Remove declaration.
7596
fc21784d
JG
75972013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
7598
7599 * config/aarch64/aarch64-simd.md
7600 (aarch64_vcond_internal<mode>): Rename to...
7601 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
7602 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
7603 float modes. Clarify all iterator modes.
7604 (vcond<mode><mode>): Use new name for vcond expanders.
7605 (vcond<v_cmp_result><mode>): Likewise.
7606 (vcondu<mode><mode>: Likewise.
7607 * config/aarch64/iterators.md (VDQF_COND): New.
7608
d4c52634
MG
76092013-05-14 Marc Glisse <marc.glisse@inria.fr>
7610
7611 PR bootstrap/57266
7612 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
7613 variable for the shift amount. Check that we shift by non-negative
7614 amounts.
7615
2b261262
CLT
76162013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
7617
7618 PR target/42017
7619 * config/arm/arm.h (EPILOGUE_USES): Only return true
7620 for LR_REGNUM after epilogue_completed.
7621
6e022d7b
JR
76222013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
7623
7624 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 7625 is error_mark_node.
6e022d7b 7626
56cf7859
RO
76272013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7628
7629 PR target/57261
7630 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
7631 and Solaris 11+/x86 with gld.
7632 * configure: Regenerate.
7633
75776c6d
JJ
76342013-05-14 Jakub Jelinek <jakub@redhat.com>
7635
7636 * expmed.c (expand_shift_1): Canonicalize rotates by
7637 constant bitsize / 2 to bitsize - 1.
7f998021 7638 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ
7639 case ROTATERT>: Likewise.
7640
7641 Revert:
7642 2013-05-10 Jakub Jelinek <jakub@redhat.com>
7643
7644 * config/i386/i386.md (rotateinv): New code attr.
7645 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7646 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7647 roll $31, %eax, etc.
7648
df35498a
RB
76492013-05-14 Richard Biener <rguenther@suse.de>
7650
7651 PR middle-end/57235
7652 * tree-eh.c (sink_clobbers): Give up for successors with
7653 multiple predecessors and no virtual uses.
7654
cc6e7ece
EB
76552013-05-14 Eric Botcazou <ebotcazou@adacore.com>
7656
7657 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
7658 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
7659
f778dd4d
JJ
76602013-05-14 Jakub Jelinek <jakub@redhat.com>
7661
7662 PR middle-end/57251
7663 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
7664 the case when both op0 and op1 have VOIDmode.
7665
bad4df9b
KP
76662013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7667
7668 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
7669 in multiply-accumulate mode.
7670
56f3e9ac
GW
76712013-05-13 Guozhi Wei <carrot@google.com>
7672
7673 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
7674
a3d7ab92
KT
76752013-05-13 Kai Tietz <ktietz@redhat.com>
7676
7677 PR target/56975
7678 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 7679 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 7680 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 7681 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT
7682 (ix86_expand_prologue): Likewise.
7683 (ix86_expand_split_stack_prologue): Likewise.
7684 (legitimate_pic_address_disp_p): Likewise.
7685 (legitimize_pic_address): Likewise.
7686 (legitimize_tls_address): Likewise.
7687 (legitimize_pe_coff_symbol): Likewise.
7688 (output_pic_addr_const): Likewise.
7689 (construct_plt_address): Likewise.
7690 (ix86_expand_call): Likewise.
7691 (x86_output_mi_thunk): Likewise.
7692 (x86_function_profiler): Likewise.
7693
c59b7e28
SN
76942013-05-13 Sofiane Naci <sofiane.naci@arm.com>
7695
7696 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
7697 similar switch cases.
7698 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
7699 (aarch64_simd_mov_to_<mode>low): Delete.
7700 (aarch64_simd_mov_to_<mode>high): Delete.
7701 (move_lo_quad_<mode>): Add w<-r alternative.
7702 (aarch64_simd_move_hi_quad_<mode>): Likewise.
7703 (aarch64_simd_mov_from_*): Update type attribute.
7704 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
7705 statement.
7706
5f28524a
JH
77072013-05-13 Jan Hubicka <jh@suse.cz>
7708
7709 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
7710 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB
7711 ix86_expand_epilogue, emit_i387_cw_initialization,
7712 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
7713 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 7714
ae6fa899
JJ
77152013-05-13 Jakub Jelinek <jakub@redhat.com>
7716
7717 PR tree-optimization/45216
7718 PR tree-optimization/57157
7719 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
7720 the (-Y) & (B - 1) variant if OP is |.
7721 * expmed.c (expand_shift_1): For rotations by const0_rtx just
7722 return shifted. Use (-op1) & (prec - 1) as other_amount
7723 instead of prec - op1.
7724
4502fe8d
MJ
77252013-05-13 Martin Jambor <mjambor@suse.cz>
7726
7727 PR middle-end/42371
7728 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
7729 (ipa_constant_data): New type.
7730 (ipa_jump_func): Use ipa_constant_data to hold information about
7731 constant jump functions.
7732 (ipa_get_jf_constant): Adjust to jump function type changes.
7733 (ipa_get_jf_constant_rdesc): New function.
7734 (ipa_param_descriptor): New field controlled_uses.
7735 (ipa_get_controlled_uses): New function.
7736 (ipa_set_controlled_uses): Likewise.
7737 * ipa-ref.h (ipa_find_reference): Declare.
7738 * ipa-prop.c (ipa_cst_ref_desc): New type.
7739 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
7740 changes.
7741 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
7742 New parameter cs. Adjust all callers.
7743 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
7744 (remove_described_reference): New function.
7745 (jfunc_rdesc_usable): Likewise.
7746 (try_make_edge_direct_simple_call): Decrement controlled use count,
7747 attempt to remove reference if it hits zero.
7748 (combine_controlled_uses_counters): New function.
7749 (propagate_controlled_uses): Likewise.
7750 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
7751 (ipa_edge_duplication_hook): Duplicate reference descriptions.
7752 (ipa_print_node_params): Print described use counter.
7753 (ipa_write_jump_function): Adjust to jump function type changes.
7754 (ipa_read_jump_function): New parameter CS, pass it to
7755 ipa_set_jf_constant. Adjust caller.
7756 (ipa_write_node_info): Stream controlled use count
7757 (ipa_read_node_info): Likewise.
7758 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
7759 asserting.
7760 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
7761 count. Remove cloning-added reference if it reaches zero.
7762 * ipa-ref.c (ipa_find_reference): New function.
7763
0864bfc2
GG
77642013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7765
e7208ea3 7766 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG
7767 alignment values for AMD BD and BT architectures.
7768
640bfeb2
MG
77692013-05-13 Marc Glisse <marc.glisse@inria.fr>
7770
7771 * tree-vect-generic.c (uniform_vector_p): Move ...
7772 * tree.c (uniform_vector_p): ... here.
7773 * tree.h (uniform_vector_p): Declare it.
7774 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
7775 into a scalar.
7776
3a60f32b
JJ
77772013-05-13 Jakub Jelinek <jakub@redhat.com>
7778
198fe1bf
JJ
7779 PR tree-optimization/57230
7780 * tree-ssa-strlen.c (handle_char_store): Record length for
7781 array store from STRING_CST.
7782
3a60f32b
JJ
7783 PR tree-optimization/57230
7784 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
7785 check.
7786
566be57c
JR
77872013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
7788
7789 * config/epiphany/epiphany.c (epiphany_init): Check size of
7790 NUM_MODES_FOR_MODE_SWITCHING.
7791 (epiphany_expand_prologue):
7792 Remove CONFIG_REGNUM initial value handling code.
7793 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
7794 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 7795 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR
7796 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
7797 Don't return 1 for FP_MODE_NONE.
7798 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
7799 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
7800 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
7801 * config/epiphany/epiphany.md (save_config): New pattern.
7802
0f2c2331
UB
78032013-05-12 Uros Bizjak <ubizjak@gmail.com>
7804
7805 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
7806
5b3f93c7
UB
78072013-05-10 Uros Bizjak <ubizjak@gmail.com>
7808
7809 * config/i386/i386.md (memory): Handle sseishft1.
7810 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
7811 (*vec_extractv2di_1): Ditto.
7812
1f873f0e
VM
78132013-05-10 Vladimir Makarov <vmakarov@redhat.com>
7814
7815 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
7816 saved registers.
7817
341427fa 78182013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH
7819
7820 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
7821 Add mthumb/march=armv7-a multilib.
7822 Add mthumb/march=armv7-r multilib.
7823 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
7824
9e69bdde
RC
78252013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
7826
7827 * config/v850/t-rtems: Add more multilibs.
7828
9ff09a22
RB
78292013-05-10 Richard Biener <rguenther@suse.de>
7830
7831 PR tree-optimization/57214
7832 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
7833 not propagate from SSA names that occur in abnormal PHI nodes.
7834
a5e0cd1d
MG
78352013-05-10 Marc Glisse <marc.glisse@inria.fr>
7836
7837 * stor-layout.c (element_precision): New function.
7838 * machmode.h (element_precision): Declare it.
7839 * tree.c (build_minus_one_cst): New function.
7840 (element_precision): Likewise.
7841 * tree.h (build_minus_one_cst): Declare new function.
7842 (element_precision): Likewise.
7843 * fold-const.c (operand_equal_p): Use element_precision.
7844 (fold_binary_loc): Handle vector types.
7845 * convert.c (convert_to_integer): Use element_precision.
7846 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
7847 separately.
7848
cb2558bc
RS
78492013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
7850
7851 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
7852 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
7853 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
7854 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7855 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
7856 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
7857 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
7858 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
7859 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
7860 (Uuw8): New constraints.
7861 (Usb4): Move into alphabetical order.
7862 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
7863 (sd8_operand, ub8_operand, uw8_operand): New predicates.
7864 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
7865 previously unnamed patterns.
7866 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
7867 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
7868 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
7869 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
7870 of set_attr_alternative/if_then_else. Use extended_mips16 instead
7871 of specific lengths.
7872
cb3b8d33
JJ
78732013-05-10 Jakub Jelinek <jakub@redhat.com>
7874
6f93c008
JJ
7875 * config/i386/i386.md (rotateinv): New code attr.
7876 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
7877 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
7878 roll $31, %eax, etc.
7879
cb3b8d33
JJ
7880 PR tree-optimization/45216
7881 PR tree-optimization/57157
7882 * tree-ssa-forwprop.c (simplify_rotate): New function.
7883 (ssa_forward_propagate_and_combine): Call it.
7884
afb119be
RB
78852013-05-10 Richard Biener <rguenther@suse.de>
7886
7887 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
7888 disable peeling when we version for aliasing.
7889 (vector_alignment_reachable_p): Honor explicit user alignment.
7890 (vect_supportable_dr_alignment): Likewise.
7891 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
7892 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
7893 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
7894 then peeling to arrange for the cost-model check to come first.
7895
01ae4861
AM
78962013-05-10 Alan Modra <amodra@gmail.com>
7897
7898 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
7899 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
7900 * configure: Regenerate.
7901
ebc9a431
AM
79022013-05-10 Alan Modra <amodra@gmail.com>
7903
7904 PR target/55033
7905 * varasm.c (default_elf_select_section): Move !DECL_P check..
7906 (get_named_section): ..to here before calling get_section_name.
7907 Adjust assertion.
7908 (default_section_type_flags): Add DECL_P check.
7909 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
7910 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
7911
d4bca93c
JR
79122013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
7913
7914 * config/epiphany/epiphany.c (epiphany_expand_prologue):
7915 When using gen_stack_adjust_str with a register offset, add a
7916 REG_FRAME_RELATED_EXPR note.
7917
60ca9a65
UB
79182013-05-09 Uros Bizjak <ubizjak@gmail.com>
7919
7920 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
7921 (*vec_extractv4si_zext_mem): Ditto.
7922 (*vec_extractv2di): Add 0->x and x->x alternatives.
7923 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
7924 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
7925
a3409c02
JM
79262013-05-09 Jason Merrill <jason@redhat.com>
7927
0138d6b2
JM
7928 N3639 C++1y VLA support
7929 * gimplify.c (gimplify_vla_decl): Don't touch an existing
7930 DECL_VALUE_EXPR.
7931
a3409c02
JM
7932 * tree.c (build_constructor_va): New.
7933 * tree.h: Declare it.
7934
66e6b990
MJ
79352013-05-09 Martin Jambor <mjambor@suse.cz>
7936
7937 PR lto/57084
7938 * gimple-fold.c (canonicalize_constructor_val): Call
7939 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
7940
64cfa6c0
JH
79412013-05-09 Jan Hubicka <jh@suse.cz>
7942 Richard Biener <rguenther@suse.de>
7943
7944 PR lto/54095
7945 * symtab.c (symtab_make_decl_local): Do not add private names.
7946
c3167b00
JH
79472013-05-09 Jan Hubicka <jh@suse.cz>
7948
7949 PR lto/54095
7950 * symtab.c (insert_to_assembler_name_hash): Handle clones.
7951 (unlink_from_assembler_name_hash): Likewise.
7952 (symtab_prevail_in_asm_name_hash, symtab_register_node,
7953 symtab_unregister_node, symtab_initialize_asm_name_hash,
7954 change_decl_assembler_name): Update.
7955
12dc6974
SN
79562013-05-09 Sofiane Naci <sofiane.naci@arm.com>
7957
7958 * config/aarch64/aarch64.md: New movtf split.
7959 (*movtf_aarch64): Update.
7960 (aarch64_movdi_tilow): Handle TF modes and rename to
7961 aarch64_movdi_<mode>low.
7962 (aarch64_movdi_tihigh): Handle TF modes and rename to
7963 aarch64_movdi_<mode>high
7964 (aarch64_movtihigh_di): Handle TF modes and rename to
7965 aarch64_mov<mode>high_di
7966 (aarch64_movtilow_di): Handle TF modes and rename to
7967 aarch64_mov<mode>low_di
7968 (aarch64_movtilow_tilow): Remove spurious whitespace.
7969 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
7970 splits.
7971 (aarch64_print_operand): Update.
7972
227eb343
AM
79732013-05-09 Alan Modra <amodra@gmail.com>
7974
7975 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
7976 powerpc64le.
7977 * configure: Regenerate.
7978
0b013847
UB
79792013-05-08 Uros Bizjak <ubizjak@gmail.com>
7980
7981 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
7982 splitter preparation statements.
7983 * config/i386/sse.md (*vec_extract* splitters): Ditto.
7984 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
7985 adjust_address_nv.
7986
1dc3d6e9
BS
79872013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7988
7989 * gimple-ssa-strength-reduction.c (count_candidates): Change
7990 return value to int.
7991 (analyze_candidates_and_replace): Change type of length to int.
7992
e61e7d28
UB
79932013-05-08 Uros Bizjak <ubizjak@gmail.com>
7994
7995 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
7996 (*vec_extract<mode>): Use VI12_128 mode iterator.
7997 (*vec_extract<mode>_mem): Ditto.
7998 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
7999 attribute.
8000
4fbfcf44
DN
80012013-05-08 Diego Novillo <dnovillo@google.com>
8002
8003 PR bootstrap/54659
8004
8005 Revert:
4fbfcf44
DN
8006 2012-08-17 Diego Novillo <dnovillo@google.com>
8007
e61e7d28
UB
8008 PR bootstrap/54281
8009 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8010 * config.in: Regenerate.
8011 * configure: Regenerate.
8012 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 8013
702d8703
JH
80142013-05-08 Jan Hubicka <jh@suse.cz>
8015
8016 PR lto/54095
8017 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8018 * cgraph.h (symtab_node_base): Add unique_name.
8019 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8020 input_overwrite_node, input_varpool_node): Stream unique_name.
8021 * cgraphclones.c (cgraph_create_virtual_clone,
8022 cgraph_function_versioning): Set unique_name.
8023 * ipa.c (function_and_variable_visibility): Set unique_name.
8024
8b28cf47
BS
80252013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8026
8027 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8028 (alloc_cand_and_find_basis): Restrict conditional candidate
8029 processing to CAND_MULTs.
8030
e86074fd
JH
80312013-05-08 Jan Hubicka <jh@suse.cz>
8032
8033 PR lto/54095
8034 lto-symtab.c (lto_symtab_symbol_p): New function.
8035 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8036 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8037 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8038 Skip static symbols.
8039
44398cbe
PC
80402013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8041
8042 PR tree-optimization/57200
8043 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8044 Only call inform if the preceding warning_at returns true.
8045
f6bc1c4a
HS
80462013-05-07 Han Shen <shenhan@google.com>
8047
8048 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8049 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8050 * common.opt (fstack-protector-strong): New option.
8051 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8052 * doc/invoke.texi (Optimization Options): Document
8053 "-fstack-protector-strong".
8054 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8055
4ffecb1f
SB
80562013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8057
8058 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8059
5a107a0f
VM
80602013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8061
8062 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8063 it is not lhs of the insn.
8064 (setup_reg_equiv): Remove insn having equiv memory which it is not
8065 lhs of the insn.
8066 * lra-constraints.c (process_address): Try to improve generation
8067 code for address base + disp.
8068 (lra_constraints): Make correct the code for checking insn setting
8069 up backward equivalence. Remove insn only if it is in the init
8070 insn list.
8071 * lra-eliminations.c (update_reg_eliminate): Change return value.
8072 (lra_eliminate): Use the result.
8073
3f5783ea
UB
80742013-05-07 Uros Bizjak <ubizjak@gmail.com>
8075
8076 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8077 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8078 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8079 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8080 register target operands.
8081 (*vec_extractv8hi_sse2): New pattern.
8082 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8083 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8084 (*vec_extract<mode>_mem): New insn and split pattern.
8085
8a5800b8
CL
80862013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8087
8088 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8089 (TARGET_ASAN_SHADOW_OFFSET): Define.
8090 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8091 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8092
7bf55a70
BS
80932013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8094
8095 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8096 (incr_vec_index): Return -1 if increment not found.
8097 (create_add_on_incoming_edge): Assert if increment not found.
8098 (record_increment): Limit number of increments recorded.
8099 (all_phi_incrs_profitable): Return false if an increment not found.
8100 (replace_profitable_candidates): Don't process increments that were
8101 not recorded.
8102 (analyze_candidates_and_replace): Limit size of incr_vec.
8103
3f8825c0
RB
81042013-05-07 Richard Biener <rguenther@suse.de>
8105
8106 * calls.c (special_function_p): setjmp-like functions are leaf.
8107 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8108 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8109
fd4842cd
SN
81102013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8111
8112 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8113 (aarch64_simd_mov<mode>): New expander.
8114 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8115 (aarch64_simd_mov_to_<mode>high): Likewise.
8116 (aarch64_simd_mov_from_<mode>low): Likewise.
8117 (aarch64_simd_mov_from_<mode>high): Likewise.
8118 (aarch64_dup_lane<mode>): Update.
8119 (aarch64_dup_lanedi): New instruction pattern.
8120 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8121 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8122
a7a7d10e
BS
81232013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8124
8125 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8126 (replace_mult_candidate): Remove unnecessary argument; remove
8127 unnecessary parameter from call to introduce_cast_before_cand.
8128 (replace_unconditional_candidate): Remove unnecessary parameter
8129 from call to replace_mult_candidate.
8130 (replace_conditional_candidate): Likewise.
8131 (insert_initializers): Use make_temp_ssa_name.
8132 (introduce_cast_before_cand): Remove unnecessary argument; use
8133 make_temp_ssa_name.
8134 (replace_one_candidate): Remove unnecessary argument; remove
8135 unnecessary parameter from calls to introduce_cast_before_cand.
8136 (replace_profitable_candidates): Remove unnecessary parameters
8137 from calls to replace_one_candidate.
8138
29105868
BS
81392013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8140
8141 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8142 phi def as possibly hiding a basis for a CAND_ADD whose operands
8143 have been commuted in the analysis.
8144 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8145
4095f9fa
N
81462013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8147
8148 * config/aarch64/aarch64.md
8149 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8150 shift value between 0-4.
8151
39e843e8
RB
81522013-05-07 Richard Biener <rguenther@suse.de>
8153
8154 * double-int.h (rshift): New overload.
8155 * double-int.c (rshift): New function.
8156 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8157 (create_reference_ops_from_ref): Remove.
8158 (vn_reference_insert): Use shared ops for constructing the
8159 reference and copy it.
8160
0a1a83cb
RB
81612013-05-07 Richard Biener <rguenther@suse.de>
8162
8163 PR middle-end/57190
8164 * tree-eh.c (sink_clobbers): Properly propagate
8165 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8166
43303d6f
JJ
81672013-05-07 Jakub Jelinek <jakub@redhat.com>
8168
ba7e83f8
JJ
8169 PR tree-optimization/57149
8170 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8171 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8172 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8173 uninit_undefined_value_p instead of ssa_undefined_value_p.
8174
43303d6f
JJ
8175 PR debug/57184
8176 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8177 for modifier == EXPAND_INITIALIZER.
8178
14523c25
AB
81792013-05-07 Anton Blanchard <anton@samba.org>
8180
8181 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8182 for powerpc64 little endian.
8183 * configure: Regenerate.
8184
cb7c8be9
GS
81852013-05-06 Graham Stott <grahams@btinternet.com>
8186
8187 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8188 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8189 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8190 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8191
81922013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS
8193
8194 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8195 codes which allow non-lvalues.
8196
9a0ee7b0
MG
81972013-05-06 Marc Glisse <marc.glisse@inria.fr>
8198
8199 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8200 components are all 1s.
8201 (integer_minus_onep): New function.
8202 * tree.h (integer_minus_onep): Declare it.
8203 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8204 integer_minus_onep instead of integer_all_onesp.
8205
f2c17ea9
OE
82062013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8207
8208 PR target/52933
8209 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8210 variations of these patterns.
8211
f75e6a51
UB
82122013-05-06 Uros Bizjak <ubizjak@gmail.com>
8213
8214 * config/i386/i386.md (isa): Add x64_sse4 member.
8215 (enabled): Handle x64_sse4.
8216 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8217 instruction for 64bit SSE4_1 targets. Update insn attributes.
8218 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8219 instruction for SSE4_1 targets. Update insn attributes.
8220 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8221 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8222 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8223 const_1 selector.
8224 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8225 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8226 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8227
a986d468
OE
82282013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8229
8230 PR target/57108
8231 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8232
382522cb
MK
82332013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8234
8235 * final.c (do_assembler_dialects): Don't handle curly braces and
8236 vertical bar escaped by % as dialect delimiters.
8237 (output_asm_insn): Print curly braces and vertical bar if escaped
8238 by % and ASSEMBLER_DIALECT defined.
8239 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8240 * doc/tm.texi: Regenerated.
8241
fb0d5c60
SB
82422013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8243
fb0d5c60
SB
8244 * config/mips/mips.c: Include tree-pass.h.
8245 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8246 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8247 (pass_mips_machine_reorg2): New machine specific pass.
8248 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8249 (mips_option_override): Register the new pass.
8250 * rtl.h (cleanup_barriers): Remove prototype.
8251 (dbr_schedule): Likewise.
8252 * jump.c (cleanup_barriers): Make static.
8253 * reorg.c (dbr_schedule): Likewise.
8254
aa06a978
RB
82552013-05-06 Richard Biener <rguenther@suse.de>
8256
8257 PR tree-optimization/57185
8258 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8259 SSA names properly.
8260
0b953bec
UB
82612013-05-06 Uros Bizjak <ubizjak@gmail.com>
8262
8263 PR target/57106
8264 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8265 between "first_arg" and "insn", not "last" and "insn".
8266
28708525
WS
82672013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8268
8269 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8270 (find_candidates_in_block): Re-enable slsr_process_phi.
8271 (create_phi_basis): Fix double counting of candidate adjustment.
8272
0107dca2
RB
82732013-05-06 Richard Biener <rguenther@suse.de>
8274
8275 PR middle-end/57147
8276 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8277 the edge is also fallthru, preserve it and just clear the
8278 abnormal flag.
8279 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8280 also complex, preserve that and just clear the fallthru flag.
8281 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8282 update virtual operands.
8283
470d4d13
AM
82842013-05-06 Alan Modra <amodra@gmail.com>
8285
8286 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8287 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8288 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8289 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8290 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8291 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8292
8f1dbf8d
AM
82932013-05-06 Alan Modra <amodra@gmail.com>
8294
8295 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8296 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8297 (DEFAULT_ASM_ENDIAN): Define.
8298 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8299 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8300 Update -K PIC clause from sysv4.h.
8301 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8302 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8303
54890767
AM
83042013-05-06 Alan Modra <amodra@gmail.com>
8305
8306 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8307 twice for little-endian.
8308 (ashrdi3_no_power, ashrdi3): Support little-endian.
8309
2353515d
OE
83102013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8311
8312 PR target/55303
8313 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
8314 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
8315 related expanders.
8316 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
8317 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
8318 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
8319 New predicates.
8320
33e67557
SB
83212013-05-05 Steven Bosscher <steven@gcc.gnu.org>
8322 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
8323
8324 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
8325 * config/pa/pa.opt: Make mbig-switch a no-op.
8326 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
8327 (CASE_VECTOR_MODE): Always return SImode.
8328 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
8329 for the !TARGET_BIG_SWITCH case.
8330 * config/pa/pa-linux.h: Likewise.
8331 * config/pa/pa-openbsd.h: Likewise.
8332 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
8333 * config/pa/pa.md (short_jump): Remove define_insn.
8334 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
8335 (casesi0): Remove define_insn.
8336 (type): Remove btable_branch.
8337 (pa_combine_type): Likewise.
8338 (in_nullified_branch_delay): Likewise.
8339 (in_call_delay): Likewise.
8340 (define_delay): Likewise.
8341 (define_insn_reservation "Z3"): Likewise.
8342 (define_insn_reservation "Z4"): Likewise.
8343 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
8344 (pa_adjust_insn_length): Remove adjustment for btable branches.
8345 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
8346 and mno-big-switch
8347
3095685e
UB
83482013-05-05 Uros Bizjak <ubizjak@gmail.com>
8349
8350 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8351 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
8352 Add m->r,x alternatives.
8353 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
8354 splitters using SWI48x mode iterator.
8355 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
8356 TARGET_64BIT. Add m->x alternative.
8357 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
8358 Add o->x alternative. Enable for TARGET_SSE.
8359 (sse_storeq): Remove expander.
8360 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
8361 with memory input operand.
8362 (*vec_extractv2di_1 splitter): New.
8363 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
8364 * config/i386/i386.md (ssevecmodelower): New mode attribute.
8365
4b36ae28
SB
83662013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8367
8368 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
8369 (INT_LOWPART): Delete.
8370 (extract_MB): Adjust.
8371 (extract_ME): Adjust.
8372 (print_operand): Adjust.
8373
da226db2
SB
83742013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8375
8376 * config/rs6000/predicates.md (reg_or_add_cint_operand,
8377 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8378 (reg_or_logical_cint_operand, easy_fp_constant,
8379 logical_const_operand): Delete "CONST_DOUBLE" case.
8380 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
8381 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 8382 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB
8383 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
8384 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
8385 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
8386 test.
8387 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
8388 CONST_DOUBLE DImode/VOIDmode case.
8389 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
8390 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8391 CONST_DOUBLE VOIDmode case.
8392 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8393 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
8394 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
8395 Delete CONST_DOUBLE case.
8396 (splitters for mov FMOVE64 const_double): Delete
8397 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8398 "HOST_BITS_PER_WIDE_INT >= 64" test.
8399 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
8400 case.
8401 (mov DI const_double): Delete.
8402
40de22d6
JJ
84032013-05-04 Jakub Jelinek <jakub@redhat.com>
8404
8405 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
8406 on op shows all bits zero in mode of a lowpart subreg, return zero.
8407
5ec6aff2
MM
84082013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8409
8410 PR target/57150
8411 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
8412 to save TFmode registers and DImode to save TImode registers for
8413 caller save operations.
8414 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
8415 mark being partially clobbered since they only use the first
8416 double word.
8417
8418 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
8419 and TDmode only use the upper 64-bits of each VSX register.
8420
2cefad90
BS
84212013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8422
8423 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
8424 (find_candidates_in_block): Disable slsr_process_phi.
8425
d6d7eee1
GW
84262013-05-03 Guozhi Wei <carrot@google.com>
8427
8428 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB
8429 constructor to ...
8430 (build_init_ctor): ... here.
d6d7eee1 8431
9b92d12b
BS
84322013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8433
8434 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
8435 (slsr_cand_d): Redefine def_phi.
8436 (stride_status, phi_adjust_status, count_phis_status): New enums.
8437 (find_phi_def): New.
8438 (find_basis_for_base_expr): New.
8439 (find_basis_for_candidate): Handle hidden bases.
8440 (alloc_cand_and_find_basis): Handle phi candidates.
8441 (slsr_process_phi): New.
8442 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
8443 (create_mul_imm_cand): Likewise.
8444 (create_add_ssa_cand): Exclude phi base candidates.
8445 (create_add_imm_cand): Likewise.
8446 (slsr_process_cast): Likewise.
8447 (slsr_process_copy): Likewise.
8448 (find_candidates_in_block): Handle phi candidates.
8449 (dump_candidate): Likewise.
8450 (unconditional_cands): Delete.
8451 (unconditional_cands_with_known_stride_p): Delete.
8452 (phi_dependent_cand_p): New.
8453 (cand_increment): Handle phi-dependent candidates.
8454 (replace_dependent): Delete.
8455 (replace_mult_candidate): New.
8456 (replace_unconditional_candidate): New.
8457 (incr_vec_index): Move to avoid forward reference.
8458 (create_add_on_incoming_edge): New.
8459 (create_phi_basis): New.
8460 (replace_dependents): Delete.
8461 (replace_conditional_candidate): New.
8462 (phi_add_costs): New.
8463 (replace_uncond_cands_and_profitable_phis): New.
8464 (record_increment): Handle phi adjustments.
8465 (record_phi_increments): New.
8466 (record_increments): Handle phi adjustments.
8467 (phi_incr_cost): New.
8468 (lowest_cost_path): Handle phis.
8469 (total_savings): Likewise.
8470 (analyze_increments): Likewise.
8471 (ncd_with_phi): New.
8472 (ncd_of_cand_and_phis): New.
8473 (nearest_common_dominator_for_cands): Handle phi increments.
8474 (all_phi_incrs_profitable): New.
8475 (replace_profitable_candidates): Handle phi-dependent candidates.
8476 (analyze_candidates_and_replace): Likewise.
8477
68f073d4
TJ
84782013-05-03 Teresa Johnson <tejohnson@google.com>
8479
8480 PR bootstrap/57154
8481 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
8482 do not exceed REG_BR_PROB_BASE.
8483
a4ee7cb9
JL
84842013-05-03 Jeff Law <law@redhat.com>
8485
ade67f70 8486 PR tree-optimization/57144
a4ee7cb9
JL
8487 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
8488 operand of the condition will bit into the new type when eliminating
8489 a cast feeding a condition.
8490
47954c4e
JJ
84912013-05-03 Jakub Jelinek <jakub@redhat.com>
8492
8493 PR rtl-optimization/57130
3095685e
UB
8494 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
8495 of COMPARE as in_code to the recursive call if needed.
47954c4e 8496
3c21604f
UB
84972013-05-03 Uros Bizjak <ubizjak@gmail.com>
8498
8499 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
8500 (enabled): Handle new members.
8501 * config/i386/sse.md (*vec_concatv2si): Merge from
8502 *vec_concatv2si_sse2 and vec_concatv2si_sse.
8503 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
8504
95c03b36
JR
85052013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
8506
8507 PR tree-optimization/57027
8508 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
8509 for fnms opportunity, check we got the prerequisite kind
8510 of tree / gimple before using accessor functions.
8511
07bfc9ec
RB
85122013-05-03 Richard Biener <rguenther@suse.de>
8513
8514 * double-int.h (lshift): New overload without precision
8515 and arith argument.
8516 (operator *=, operator +=, operator -=): Move ...
8517 * double-int.c (operator *=, operator +=, operator -=): ... here
8518 and implement more efficiently.
8519 (mul_double_with_sign): Remove.
8520 (lshift_double): Adjust to take unsinged shift argument, push
8521 dispatching code to callers.
8522 (mul_double_wide_with_sign): Add early out for callers that
8523 are not interested in high parts or overflow.
8524 (lshift): New function.
8525 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
8526 dispatch code here.
8527 (lrotate, rrotate): Use logical shifts.
8528 * expr.c (get_inner_reference): Use lshift.
8529 * fixed-value.c (do_fixed_divide): Likewise.
8530 * tree-dfa.c (get_ref_base_and_extent): Likewise.
8531 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
8532 (indirect_refs_may_alias_p): Likewise.
8533 (stmt_kills_ref_p_1): Likewise.
8534
7769bb64
VP
85352013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8536
8537 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
8538
1d0c8e5c
VP
85392013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8540
8541 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
8542 scalar form of FABD instruction.
8543
f15feaf9
VM
85442013-05-02 Vladimir Makarov <vmakarov@redhat.com>
8545
8546 * lra-constraints.c (process_alt_operands): Add checking alt
8547 number to choose the best alternative.
8548
d90e76d4
RB
85492013-05-02 Richard Biener <rguenther@suse.de>
8550
8551 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
8552 bitmap and its handling.
8553 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
8554
9f8e7a96
RB
85552013-05-02 Richard Biener <rguenther@suse.de>
8556
8557 PR middle-end/57140
8558 * tree-inline.c (copy_loops): Properly handle removed loops.
8559 (copy_cfg_body): Mark destination loops for fixup if source
8560 loops needed fixup.
8561
f3a81b39
GY
85622013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
8563
8564 PR target/56732
8565 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
8566 generating simple_return for naked functions.
8567
7b920a9a
MJ
85682013-05-02 Martin Jambor <mjambor@suse.cz>
8569
8570 PR middle-end/56988
8571 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
8572 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
8573 flags match.
8574 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
8575 ipa_agg_replacement_value structures.
8576 (known_aggs_to_agg_replacement_list): Likewise.
8577 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
8578 (read_agg_replacement_chain): Likewise.
8579 (ipcp_transform_function): Also check that by_ref flags match.
8580
2c41c19d
RB
85812013-05-02 Richard Biener <rguenther@suse.de>
8582
8583 * graphds.h (struct graph): Add obstack member.
8584 * graphds.c (new_graph): Initialize obstack and allocate
8585 vertices from it.
8586 (add_edge): Allocate edge from the obstack.
3c21604f 8587 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 8588
8b47039c
TJ
85892013-05-02 Teresa Johnson <tejohnson@google.com>
8590
8591 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
8592 divides.
8593 * cfg.c (update_bb_profile_for_threading): Ditto.
8594 * tree-inline.c (copy_bb): Ditto.
8595 (copy_edges_for_bb): Ditto.
8596 (initialize_cfun): Ditto.
8597 (copy_cfg_body): Ditto.
8598 (expand_call_inline): Ditto.
8599 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
8600 (estimate_node_size_and_time): Ditto.
8601 (inline_merge_summary): Ditto.
8602 * cgraphclones.c (cgraph_clone_edge): Ditto.
8603 (cgraph_clone_node): Ditto.
8604 * sched-rgn.c (compute_dom_prob_ps): Ditto.
8605 (compute_trg_info): Ditto.
8606
da65928c
IB
86072013-05-02 Ian Bolton <ian.bolton@arm.com>
8608
8609 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
8610 S reg when fp attribute set.
8611 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
8612
add3c965
IB
86132013-05-02 Ian Bolton <ian.bolton@arm.com>
8614
8615 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
8616 New pattern.
8617 (*and_one_cmplsi3_compare0_uxtw): Likewise.
8618 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
8619 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
8620
fdd43ac4
RB
86212013-05-02 Richard Biener <rguenther@suse.de>
8622
8623 * tree-scalar-evolution.c (scev_info_hasher): Remove.
8624 (struct instantiate_cache_entry): New type.
8625 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
8626 (struct instantiate_cache_type): New type.
8627 (set_instantiated_value, get_instantiated_value): Remove.
8628 (get_instantiated_value_entry): New function.
8629 (instantiate_scev_name): Use the new cache and adjust.
8630 (instantiate_scev_poly): Adjust.
8631 (instantiate_scev_binary): Likewise.
8632 (instantiate_array_ref): Likewise.
8633 (instantiate_scev_convert): Likewise.
8634 (instantiate_scev_not): Likewise.
8635 (instantiate_scev_3): Likewise.
8636 (instantiate_scev_2): Likewise.
8637 (instantiate_scev_r): Likewise.
8638 (instantiate_scev): Likewise.
8639 (resolve_mixers): Likewise.
8640
36ff9dfb
VM
86412013-05-01 Vladimir Makarov <vmakarov@redhat.com>
8642
8643 PR target/57091
8644 * lra-constraints.c (best_small_class_operands_num): Remove.
8645 (process_alt_operands): Remove small_class_operands_num. Take
8646 small classes operands into losers and only if the operand is not
8647 matched. Modify debugging output.
8648 (curr_insn_transform): Remove best_small_class_operands_num.
8649 Print insn name.
8650
36054fab
JG
86512013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8652
8653 * config/aarch64/aarch64-builtins.c
8654 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
8655 * config/aarch64/aarch64-simd-builtins.def
8656 (reduc_splus_): Add new modes.
8657 (reduc_uplus_): New.
8658 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
8659 (reduc_uplus_v4sf): Likewise.
8660 (reduc_splus_v4sf): Likewise.
8661 (aarch64_addv<mode>): Likewise.
8662 (reduc_uplus_<mode>): Likewise.
8663 (reduc_splus_<mode>): Likewise.
8664 (aarch64_addvv2di): Likewise.
8665 (reduc_uplus_v2di): Likewise.
8666 (reduc_splus_v2di): Likewise.
8667 (aarch64_addvv2si): Likewise.
8668 (reduc_uplus_v2si): Likewise.
8669 (reduc_splus_v2si): Likewise.
8670 (reduc_<sur>plus_<mode>): New.
8671 (reduc_<sur>plus_v2di): Likewise.
8672 (reduc_<sur>plus_v2si): Likewise.
8673 (reduc_<sur>plus_v4sf): Likewise.
8674 (aarch64_addpv4sf): Likewise.
8675 * config/aarch64/arm_neon.h
8676 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
8677 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
8678 add UNSPEC_SADDV, UNSPEC_UADDV.
8679 (SUADDV): New.
8680 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
8681
6dce23a8
JG
86822013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8683
8684 * config/aarch64/arm_neon.h
8685 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
8686
1598945b
JG
86872013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8688
8689 * config/aarch64/aarch64-builtins
8690 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
8691
998eaf97
JG
86922013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8693
8694 * config/aarch64/aarch64-simd-builtins.def
8695 (reduc_smax_): New.
8696 (reduc_smin_): Likewise.
8697 (reduc_umax_): Likewise.
8698 (reduc_umin_): Likewise.
8699 (reduc_smax_nan_): Likewise.
8700 (reduc_smin_nan_): Likewise.
8701 (fmax): Remove.
8702 (fmin): Likewise.
8703 (smax): Update for V2SF, V4SF and V2DF modes.
8704 (smin): Likewise.
8705 (smax_nan): New.
8706 (smin_nan): Likewise.
8707 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
8708 (<su><maxmin><mode>3): ...This, refactor.
8709 (s<maxmin><mode>3): New.
8710 (<maxmin_uns><mode>3): Likewise.
8711 (reduc_<maxmin_uns>_<mode>): Refactor.
8712 (reduc_<maxmin_uns>_v4sf): Likewise.
8713 (reduc_<maxmin_uns>_v2si): Likewise.
8714 (aarch64_<fmaxmin><mode>: Remove.
8715 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
8716 new builtin names.
8717 (vmin<q>_f<32,64>): Likewise.
8718 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8719 (FMAXMIN): New.
8720 (su): Add mappings for smax, smin, umax, umin.
8721 (maxmin): New.
8722 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
8723 (FMAXMIN): Rename as...
8724 (FMAXMIN_UNS): ...This.
8725 (maxminv): Remove.
8726 (fmaxminv): Likewise.
8727 (fmaxmin): Likewise.
8728 (maxmin_uns): New.
8729 (maxmin_uns_op): Likewise.
8730
bd11644e
JG
87312013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8732
8733 * config/aarch64/arm_neon.h
8734 (vac<ge, gt><sd>_f<32, 64>): Rename to...
8735 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
8736 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
8737
75dd5ace
JG
87382013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8739
8740 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
8741 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
8742
7c19979f
JG
87432013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8744
8745 * config/aarch64/aarch64-simd.md
8746 (vcond<mode>_internal): Handle special cases for constant masks.
8747 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
8748 (vcondu<mode><mode>): Likewise.
8749 (vcond<v_cmp_result><mode>): New.
8750
bb60efd9
JG
87512013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8752
8753 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
8754 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
8755 * config/aarch64/aarch64-simd-builtins.def
8756 (cmeq): Update to BUILTIN_VALLDI.
8757 (cmgt): Likewise.
8758 (cmge): Likewise.
8759 (cmle): Likewise.
8760 (cmlt): Likewise.
8761 * config/aarch64/arm_neon.h
8762 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
8763 to builtins or C as appropriate.
8764
889b9412
JG
87652013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
8766
8767 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
8768 (cmgeu): ...This.
8769 (cmhi): Rename to...
8770 (cmgtu): ...This.
8771 * config/aarch64/aarch64-simd.md
8772 (simd_mode): Add SF.
8773 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
8774 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
8775 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
8776 (cstore<mode>_neg): ...This.
8777 * config/aarch64/iterators.md
8778 (VALLF): new.
8779 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
8780 (COMPARISONS): New.
8781 (UCOMPARISONS): Likewise.
8782 (optab): Add missing comparisons.
8783 (n_optab): New.
8784 (cmp_1): Likewise.
8785 (cmp_2): Likewise.
8786 (CMP): Likewise.
8787 (cmp): Remove.
8788 (VCMP_S): Likewise.
8789 (VCMP_U): Likewise.
8790 (V_cmp_result): Add DF, SF modes.
8791 (v_cmp_result): Likewise.
8792 (v): Likewise.
8793 (vmtype): Likewise.
8794 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
8795
0a7dbb76
GY
87962013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
8797
8798 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
8799 define_insn to define_insn_and_split.
8800 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
8801 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
8802 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
8803 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
8804 (thumb2_negscc): Likewise.
8805
fb614ca6
GY
88062013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8807
8808 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
8809
9e64a0bf
GY
88102013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
8811
8812 * config/arm/thumb2.md: Remove trailing whitespaces.
8813
d6b28156
RS
88142013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8815
8816 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
8817 Use gen_int_mode rather than GEN_INT.
8818
f91674c3
L
88192013-04-30 H.J. Lu <hongjiu.lu@intel.com>
8820
b0dec607 8821 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L
8822 debug_gimple_stmt.
8823
3551257c
RB
88242013-04-30 Richard Biener <rguenther@suse.de>
8825
8826 PR middle-end/57122
3c21604f 8827 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 8828
4c1aff1c
RB
88292013-04-30 Richard Biener <rguenther@suse.de>
8830
8831 PR middle-end/57107
8832 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
8833
54b8379a
AB
88342013-04-30 Andrey Belevantsev <abel@ispras.ru>
8835
8836 PR rtl-optimization/56957
8837 PR rtl-optimization/57105
54b8379a
AB
8838 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
8839 variable. Use just INSN_UID for determining whether an insn
8840 should be only disconnected from the insn stream.
8841 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
8842
abc27962
JJ
88432013-04-30 Jakub Jelinek <jakub@redhat.com>
8844
8845 PR tree-optimization/57104
8846 * tsan.c (instrument_expr): Don't instrument accesses to
8847 DECL_HARD_REGISTER VAR_DECLs.
8848
0fc822d0
RB
88492013-04-30 Richard Biener <rguenther@suse.de>
8850
8851 * function.h (loops_for_fn): New inline function.
8852 (set_loops_for_fn): Likewise.
8853 * cfgloop.h (place_new_loop): Add struct function parameter.
8854 (get_loop): Likewise.
8855 (get_loops): Likewise.
8856 (number_of_loops): Likewise.
8857 (fel_next): Adjust.
8858 (fel_init): Likewise.
8859 * cfg.c (get_loop_copy): Adjust.
8860 * cfgloop.c (flow_loops_dump): Likewise.
8861 (record_loop_exits): Likewise.
8862 (verify_loop_structure): Likewise.
8863 * cfgloopanal.c (mark_irreducible_loops): Likewise.
8864 (estimate_reg_pressure_cost): Likewise.
8865 (mark_loop_exit_edges): Likewise.
8866 * cfgloopmanip.c (place_new_loop): Likewise.
8867 (add_loop): Likewise.
8868 (duplicate_loop): Likewise.
8869 * graph.c (draw_cfg_nodes): Likewise.
8870 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
8871 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
8872 (extract_affine_chrec): Likewise.
8873 (build_scop_iteration_domain): Likewise.
8874 * graphite.c (graphite_initialize): Likewise.
8875 * ira-build.c (create_loop_tree_nodes): Likewise.
8876 (more_one_region_p): Likewise.
8877 (rebuild_regno_allocno_maps): Likewise.
8878 (mark_loops_for_removal): Likewise.
8879 (mark_all_loops_for_removal): Likewise.
8880 (remove_unnecessary_regions): Likewise.
8881 (ira_build): Likewise.
8882 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
8883 * loop-init.c (fix_loop_structure): Likewise.
8884 (gate_rtl_move_loop_invariants): Likewise.
8885 (gate_rtl_unswitch): Likewise.
8886 (gate_rtl_unroll_and_peel_loops): Likewise.
8887 (rtl_doloop): Likewise.
8888 * lto-streamer-in.c (input_cfg): Likewise.
8889 * lto-streamer-out.c (output_cfg): Likewise.
8890 * modulo-sched.c (sms_schedule): Likewise.
8891 * predict.c (tree_estimate_probability): Likewise.
8892 (tree_estimate_probability_driver): Likewise.
8893 (estimate_loops): Likewise.
8894 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
8895 (move_sese_region_to_fn): Likewise.
8896 (debug_loop_num): Likewise.
8897 * tree-chrec.c (chrec_evaluate): Likewise.
8898 (hide_evolution_in_other_loops_than_loop): Likewise.
8899 (chrec_component_in_loop_num): Likewise.
8900 (reset_evolution_in_loop): Likewise.
8901 (evolution_function_is_invariant_rec_p): Likewise.
8902 * tree-if-conv.c (main_tree_if_conversion): Likewise.
8903 * tree-inline.c (copy_loops): Likewise.
8904 (copy_cfg_body): Likewise.
8905 (tree_function_versioning): Likewise.
8906 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
8907 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
8908 Likewise.
8909 (add_to_evolution_1): Likewise.
8910 (scev_const_prop): Likewise.
8911 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
8912 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
8913 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
8914 (tree_ssa_lim_initialize): Likewise.
8915 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
8916 (verify_loop_closed_ssa): Likewise.
8917 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
8918 (tree_ssa_loop_im): Likewise.
8919 (tree_ssa_loop_unswitch): Likewise.
8920 (tree_vectorize): Likewise.
8921 (check_data_deps): Likewise.
8922 (tree_ssa_loop_ivcanon): Likewise.
8923 (tree_ssa_loop_bounds): Likewise.
8924 (tree_complete_unroll): Likewise.
8925 (tree_complete_unroll_inner): Likewise.
8926 (tree_parallelize_loops): Likewise.
8927 (tree_ssa_loop_prefetch): Likewise.
8928 (tree_ssa_loop_ivopts): Likewise.
8929 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
8930 * tree-vectorizer.c (vectorize_loops): Likewise.
8931
37953bd3
MF
89322013-04-29 Mike Frysinger <vapier@gentoo.org>
8933
8934 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
8935 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
8936 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
8937 with EABI_LINK_SPEC.
8938
f9ed28db
UB
89392013-04-29 Uros Bizjak <ubizjak@gmail.com>
8940
8941 PR target/44578
8942 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
8943 alternative.
8944
deca73f5
VM
89452013-04-29 Vladimir Makarov <vmakarov@redhat.com>
8946
8947 PR target/57097
37953bd3 8948 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM
8949 using memory for pseudos. Print cost dump for alternatives.
8950 Modify cost values for conflicts with early clobbers.
8951 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
8952
fdca7d03
UB
89532013-04-29 Uros Bizjak <ubizjak@gmail.com>
8954
8955 PR target/57098
8956 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
8957
0b064172
IB
89582013-04-29 Ian Bolton <ian.bolton@arm.com>
8959
8960 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
8961 from/to S register.
8962 (movdi_aarch64): Support LDR/STR from/to D register.
8963
473cec55
IB
89642013-04-29 Ian Bolton <ian.bolton@arm.com>
8965
8966 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
8967 or higher by default.
8968
a6f30e66
RB
89692013-04-29 Richard Biener <rguenther@suse.de>
8970
8971 PR middle-end/57075
8972 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
8973 even if not adding abnormal edges for calls that can make
8974 abnormal gotos.
8975
0c2b2040
RB
89762013-04-29 Richard Biener <rguenther@suse.de>
8977
8978 PR middle-end/57103
8979 * tree-cfg.c (move_stmt_op): Fix condition under which to update
8980 TREE_BLOCK.
8981 (move_stmt_r): Remove redundant checking.
8982
f41f80f9
TJ
89832013-04-29 Teresa Johnson <tejohnson@google.com>
8984
8985 PR bootstrap/57077
8986 * basic-block.h (apply_scale): New function.
8987 (apply_probability): Use apply_scale.
8988 * gimple-streamer-in.c (input_bb): Ditto.
8989 * lto-streamer-in.c (input_cfg): Ditto.
8990 * lto-cgraph.c (merge_profile_summaries): Ditto.
8991 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 8992 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ
8993 (copy_edges_for_bb): Ditto.
8994 (copy_cfg_body): Ditto.
8995
315bbd2e
TV
89962013-04-29 Tom de Vries <tom@codesourcery.com>
8997
8998 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
8999 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9000 (tail_merge_optimize): Handle current_loops == NULL.
9001
ebbd90d8
JL
90022013-04-26 Jeff Law <law@redhat.com>
9003
9004 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9005 (simplify_cond_using_ranges): Generalize code to simplify
9006 COND_EXPRs where one argument is a constant and the other
9007 is an SSA_NAME created by an integral type conversion.
9008
8b9b57eb
KT
90092013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9010
9011 * config/arm/arm.md (store_minmaxsi): Use only when
9012 optimize_insn_for_size_p.
9013
9498e5dc
CB
90142013-04-29 Christian Bruel <christian.bruel@st.com>
9015
9016 PR target/57108
9017 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9018
6093bc06
RB
90192013-04-29 Richard Biener <rguenther@suse.de>
9020
9021 PR middle-end/57089
fdca7d03
UB
9022 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9023 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB
9024 (expand_omp_for_generic): Properly add loops.
9025 (expand_omp_for_static_nochunk): Likewise.
9026 (expand_omp_for_static_chunk): Likewise.
9027 (expand_omp_for): For the degenerate case fixup loops.
9028 (expand_omp_sections): Fix default bb placement in loops.
9029 (expand_omp_atomic_pipeline): Properly add loops.
9030
84aacbfd
KT
90312013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9032
9033 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9034
8dee4479
TV
90352013-04-29 Tom de Vries <tom@codesourcery.com>
9036
9037 * tree-ssa-tail-merge.c: Update header comment.
9038
47934dc4
JG
90392013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9040
9041 * config/aarch64/arm_neon.h
9042 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9043 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9044 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9045 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9046 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9047 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9048 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9049 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9050
384be29f
JG
90512013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9052
9053 * config/aarch64/aarch64-simd.md
9054 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9055 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9056 fix_trunc, fixuns_trunc.
9057 (ftrunc<VDQF:mode>2): New.
9058 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9059 (fix_trunc_optab): New.
9060
0386b123
JG
90612013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9062
9063 * config/aarch64/aarch64-builtins.c
9064 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9065 iceilf, lround, iroundf.
9066
00fcb892
UB
90672013-04-29 Uros Bizjak <ubizjak@gmail.com>
9068
9069 PR target/54349
9070 * config/i386/i386.h (enum ix86_tune_indices)
9071 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9072 New, split from X86_TUNE_INTER_UNIT_MOVES.
9073 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9074 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9075 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9076 (TARGET_INTER_UNIT_MOVES): Remove.
9077 * config/i386/i386.c (initial_ix86_tune_features): Update.
9078 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9079 (ix86_expand_convert_uns_didf_sse): Use
9080 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9081 (ix86_expand_vector_init_one_nonzero): Ditto.
9082 (ix86_expand_vector_init_interleave): Ditto.
9083 (inline_secondary_memory_needed): Return true for moves from SSE class
9084 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9085 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9086 * config/i386/constraints.md (Yi, Ym): Depend on
9087 TARGET_INTER_UNIT_MOVES_TO_VEC.
9088 (Yj, Yn): New constraints.
9089 * config/i386/i386.md (*movdi_internal): Change constraints of
9090 operand 1 from Yi to Yj and from Ym to Yn.
9091 (*movsi_internal): Ditto.
9092 (*movdf_internal): Ditto.
9093 (*movsf_internal): Ditto.
9094 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9095 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9096 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9097 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9098 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9099 * config/i386/sse.md (movdi_to_sse): Ditto.
9100 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9101 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9102 TARGET_INTER_UNIT_MOVES.
9103 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9104 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9105 instead of TARGET_INTER_UNIT_MOVES.
9106 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9107 operand 1 from Yi to Yj and from Ym to Yn.
9108
4c871069
JG
91092013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9110
9111 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9112 (float_truncate_hi_): Likewise.
9113 (float_extend_lo_): Likewise.
9114 (float_truncate_lo_): Likewise.
9115 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9116 (aarch64_float_extend_lo_v2df): Likewise.
9117 (vec_unpacks_hi_v4sf): Likewise.
9118 (aarch64_float_truncate_lo_v2sf): Likewise.
9119 (aarch64_float_truncate_hi_v4sf): Likewise.
9120 (vec_pack_trunc_v2df): Likewise.
9121 (vec_pack_trunc_df): Likewise.
9122
1709ff9b
JG
91232013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9124
9125 * config/aarch64/aarch64-builtins.c
9126 (aarch64_fold_builtin): Fold float conversions.
9127 * config/aarch64/aarch64-simd-builtins.def
9128 (floatv2si, floatv4si, floatv2di): New.
9129 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9130 * config/aarch64/aarch64-simd.md
9131 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9132 * config/aarch64/iterators.md (FLOATUORS): New.
9133 (optab): Add float, floatuns.
9134 (su_optab): Likewise.
9135
ce966824
JG
91362013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9137
9138 * config/aarch64/aarch64-builtins.c
9139 (aarch64_builtin_vectorized_function): Use new names for
9140 fcvt builtins.
9141 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9142 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9143 (fcvtzu): Split as...
9144 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9145 (fcvtas): Split as...
9146 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9147 (fcvtau): Split as...
9148 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9149 (fcvtps): Split as...
9150 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9151 (fcvtpu): Split as...
9152 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9153 (fcvtms): Split as...
9154 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9155 (fcvtmu): Split as...
9156 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9157 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9158 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9159 (lfrintnusf, lfrintnudf): Likewise.
9160 * config/aarch64/aarch64-simd.md
9161 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9162 define_insn.
9163 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9164 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9165 (fcvt_pattern): Likewise.
9166
b9de24fe
JG
91672013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9168
9169 * config/aarch64/aarch64-simd.md
9170 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9171 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9172
77a205be
JG
91732013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9174
9175 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9176 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9177 (vrnd<a,m,n,p>_f32): Implement using builtins.
9178 (vrnd<i,x><q>_f<32, 64>): New.
9179
0659ce6f
JG
91802013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9181
9182 * config/aarch64/aarch64-builtins.c
9183 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9184 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9185 (frintz): Rename to...
9186 (btrunc): ...this.
9187 (frintp): Rename to...
9188 (ceil): ...this.
9189 (frintm): Rename to...
9190 (floor): ...this.
9191 (frinti): Rename to...
9192 (nearbyint): ...this.
9193 (frintx): Rename to...
9194 (rint): ...this.
9195 (frinta): Rename to...
9196 (round): ...this.
9197 * config/aarch64/aarch64-simd.md
9198 (aarch64_frint<frint_suffix><mode>): Delete.
9199 (<frint_pattern><mode>2): Convert to insn.
9200 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9201 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9202 (frint_pattern): Likewise.
9203 (frint_suffix): Likewise.
9204
ea78906a
RB
92052013-04-29 Richard Biener <rguenther@suse.de>
9206
9207 PR tree-optimization/57081
9208 * loop-init.c: Include tree-flow.h.
9209 (loop_optimizer_finalize): Free number of iteration estimates.
9210 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9211
baee1763
JJ
92122013-04-29 Jakub Jelinek <jakub@redhat.com>
9213
94dc5332
JJ
9214 PR tree-optimization/57083
9215 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9216 non-singleton shift count range, zero extend low_bound for uns case.
9217
baee1763
JJ
9218 * config/i386/predicates.md (general_vector_operand): New predicate.
9219 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9220 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9221 if they aren't nonimmediate operands. If their original values
9222 satisfy const_vector_equal_evenodd_p, don't shift them.
9223 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9224 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9225 if not nonimmediate_operand.
9226 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9227 instead of register_operand.
9228 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9229
a9073727 92302013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB
9231
9232 * stor-layout.c (finalize_size_functions): Allocate a structure and
9233 reset cfun before dumping the functions.
9234
ba8011e6
JJ
92352013-04-27 Jakub Jelinek <jakub@redhat.com>
9236
d6fde69e
JJ
9237 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9238
ba8011e6
JJ
9239 PR target/56866
9240 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9241 use xop_pmacsdqh if uns_p.
9242 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9243 the immediate rotate count.
9244
2c62cbaa
VM
92452013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9246
9247 * rtl.h (struct rtx_def): Add comment for field jump.
9248 (LRA_SUBREG_P): New macro.
9249 * recog.c (register_operand): Check LRA_SUBREG_P.
9250 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9251 stack frame.
9252 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9253 (lra_final_code_change): Skip subreg change for operators.
9254 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9255 if there are no operand changes.
9256 * lra-constraints.c (curr_insn_set): New.
9257 (match_reload): Set LRA_SUBREG_P.
9258 (emit_spill_move): Ditto.
9259 (check_and_process_move): Use curr_insn_set. Process only single
9260 set insns. Don't initialize sec_mem_p and change_p.
9261 (simplify_operand_subreg): Use LRA_SUBREG_P.
9262 (reg_in_class_p): New function.
9263 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9264 of #ifdef. Add code to remove cycling.
9265 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9266 non-null disp. Reload inner instead of disp when base and index
9267 are null. Try to put lo_sum into register.
9268 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9269 (check_and_process_move): Move code for move cost check to
9270 simple_move_p. Remove equiv_substitution.
9271 (simple_move_p): New function.
9272 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9273 curr_insn_set. Call check_and_process_move only for single set
9274 insns. Use the new function. Move call of check_and_process_move
9275 after operand equiv substitution and address process.
9276
e7d764f3
JJ
92772013-04-26 Jakub Jelinek <jakub@redhat.com>
9278
9279 PR go/57045
9280 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9281 with nonlocal goto receivers or returns twice calls, ignore
9282 unininitialized values from abnormal edges to nl goto receiver
9283 or returns twice call.
9284
41e10689
JJ
92852013-04-26 Jakub Jelinek <jakub@redhat.com>
9286
9287 PR tree-optimization/57051
9288 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9289 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9290 bitsize.
9291
d7ed20db
RB
92922013-04-26 Richard Biener <rguenther@suse.de>
9293
9294 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9295 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9296 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9297 (fixup_loop_arrays_after_move): New function.
9298 (move_sese_region_to_fn): Properly outline the loop tree parts
9299 of the SESE region.
9300
df93505e
UB
93012013-04-26 Uros Bizjak <ubizjak@gmail.com>
9302
9303 * config/i386/i386.md (type, unit): Fix long lines.
9304
dd366ec3
RB
93052013-04-26 Richard Biener <rguenther@suse.de>
9306
9307 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9308 (lto-streamer-out.o): Likewise.
9309 * cfgloop.c (init_loops_structure): Export, add struct function
9310 argument and adjust.
9311 (flow_loops_find): Adjust.
9312 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9313 (init_loops_structure): Declare.
9314 * lto-streamer-in.c: Include cfgloop.h.
9315 (input_cfg): Input the loop tree.
9316 * lto-streamer-out.c: Include cfgloop.h.
9317 (output_cfg): Output the loop tree.
9318 (output_struct_function_base): Do not drop PROP_loops.
9319
a9e0d843
RB
93202013-03-26 Richard Biener <rguenther@suse.de>
9321
9322 * tree-cfg.c (execute_build_cfg): Build the loop tree.
9323 (pass_build_cfg): Provide PROP_loops.
9324 (move_sese_region_to_fn): Remove loops that are outlined into fn
9325 for now.
9326 * tree-inline.c: Include cfgloop.h.
9327 (initialize_cfun): Do not drop PROP_loops.
9328 (copy_loops): New function.
9329 (copy_cfg_body): Copy loop structure.
9330 (tree_function_versioning): Initialize destination loop tree.
9331 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
9332 (pass_parallelize_loops): Do IL verification.
9333 * loop-init.c (loop_optimizer_init): Fixup loops if required.
9334 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
9335 the CFG make sure we fixup loops as well.
9336 * tree-ssa-tail-merge.c: Include cfgloop.h.
9337 (replace_block_by): When merging loop latches mark loops for fixup.
9338 * lto-streamer-out.c (output_struct_function_base): Drop
9339 PROP_loops for now.
9340 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
9341 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
9342 * ipa-split.c: Include cfgloop.h.
9343 (split_function): Add the new return block to the loop tree root.
9344 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
9345 whether we have removed the forwarder block.
9346 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
9347 * cfgloop.h (place_new_loop): Declare.
9348 * cfgloopmanip.c (place_new_loop): Export.
9349 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
9350 (tree-switch-conversion.o): Likewise.
9351 (tree-complex.o): Likewise.
9352 (tree-inline.o): Likewise.
9353 (tree-ssa-tailmerge.o): Likewise.
9354 (ipa-split.o): Likewise.
9355 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
9356 (tree-ssa-copy.o): Likewise.
9357 * tree-switch-conversion.c: Include cfgloop.h
9358 (process_switch): If we emit a bit-test cascade, schedule loops
9359 for fixup.
9360 * tree-complex.c: Include cfgloop.h.
9361 (expand_complex_div_wide): Properly add new basic-blocks to loops.
9362 * asan.c: Include cfgloop.h.
9363 (create_cond_insert_point): Properly add new basic-blocks to
9364 loops, schedule loop fixup.
9365 * cfgloop.c (verify_loop_structure): Check that looks are not
9366 marked for fixup.
9367 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
9368 to loops.
9369 (expand_omp_for_generic): Likewise.
9370 (expand_omp_sections): Likewise.
9371 (expand_omp_atomic_pipeline): Schedule loops for fixup.
9372 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
9373 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
9374 is initialized, not when loops are present.
9375 * tree-parloops.c (parallelize_loops): Remove checking here.
9376 * passes.c (init_optimization_passes): Schedule a copy-propagation
9377 pass before complete unrolling of inner loops.
9378
e78e8a0b
JJ
93792013-04-26 Jakub Jelinek <jakub@redhat.com>
9380
a2e836b2
JJ
9381 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
9382 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
9383 (colorize_init): Add argument to _WIN32 version.
9384 * toplev.c: Include diagnostic-color.h.
9385 (process_options): Default to -fdiagnostics-color=auto if
9386 GCC_COLORS env var is in the environment.
9387 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB
9388 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
9389 env var is in the environment, the default is auto rather than never.
a2e836b2 9390
e78e8a0b
JJ
9391 * diagnostic.h (file_name_as_prefix): Add context argument.
9392 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
9393 the string as locus.
9394 * langhooks.c (lhd_print_error_function): Adjust caller.
9395
013e5ef9
LC
93962013-04-25 Lawrence Crowl <crowl@google.com>
9397
9398 * var-tracking.c (shared_hash_def::htab):
9399 Change type to hash_table. Update dependent calls and types.
9400
4a8fb1a1
LC
94012013-04-25 Lawrence Crowl <crowl@google.com>
9402
9403 * Makefile.in: Update as needed below.
9404
9405 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
9406 Move declaration to after the type's method definitons.
9407
9408 * attribs.c (htab_t scoped_attributes::attribute_hash):
9409 Change type to hash_table. Update dependent calls and types.
9410
9411 * bitmap.c (htab_t bitmap_desc_hash):
9412 Change type to hash_table. Update dependent calls and types.
9413
9414 * cselib.c (htab_t cselib_hash_table):
9415 Change type to hash_table. Update dependent calls and types.
9416
9417 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
9418 (hash_string_slot_node): Move implementation into lto-streamer.h
9419 struct string_slot_hasher.
9420 (eq_string_slot_node): Likewise.
9421
9422 * data-streamer-out.c: Update output_block::string_hash_table
9423 dependent calls and types.
9424
9425 * dwarf2cfi.c (htab_t trace_index):
9426 Change type to hash_table. Update dependent calls and types.
9427
9428 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
9429 Change type to hash_table. Update dependent calls and types.
9430 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
9431 (htab_t optimize_external_refs::map): Likewise.
9432 (htab_t output_comp_unit::extern_map): Likewise.
9433 (htab_t output_comdat_type_unit::extern_map): Likewise.
9434 (htab_t output_macinfo::macinfo_htab): Likewise.
9435 (htab_t optimize_location_lists::htab): Likewise.
9436 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
9437
9438 * except.c (htab_t ehspec_hash_type):
9439 Change type to hash_table. Update dependent calls and types.
9440 (assign_filter_values::ttypes): Likewise.
9441 (assign_filter_values::ehspec): Likewise.
9442 (sjlj_assign_call_site_values::ar_hash): Likewise.
9443 (convert_to_eh_region_ranges::ar_hash): Likewise.
9444
9445 * gcse.c (htab_t pre_ldst_table):
9446 Change type to hash_table. Update dependent calls and types.
9447
9448 * ggc-common.c (htab_t saving_htab):
9449 Change type to hash_table. Update dependent calls and types.
9450 (htab_t loc_hash): Likewise.
9451 (htab_t ptr_hash): Likewise.
9452 (call_count): Rename ggc_call_count.
9453 (call_alloc): Rename ggc_call_alloc.
9454 (loc_descriptor): Rename make_loc_descriptor.
9455 (add_statistics): Rename ggc_add_statistics.
9456
9457 * ggc-common.c (saving_htab):
9458 Change type to hash_table. Update dependent calls and types.
9459
9460 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
9461 (push_gimplify_context): Likewise.
9462 (pop_gimplify_context): Likewise.
9463 (struct gimple_temp_hash_elt): Added.
9464 (struct gimplify_hasher): Likewise.
9465 (struct gimplify_ctx.temp_htab):
9466 Change type to hash_table. Update dependent calls and types.
9467
9468 * gimple-fold.c: Include gimplify-ctx.h.
9469
9470 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
9471 Change type to hash_table. Update dependent calls and types.
9472 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
9473 avoid potential global name collision.
9474
9475 * gimplify.c: Include gimplify-ctx.h.
9476 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
9477 (htab_t gimplify_ctx::temp_htab):
9478 Update dependent calls and types for new type hash_table.
9479 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
9480 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
9481
9482 * gimplify-ctx.h: New.
9483 (struct gimple_temp_hash_elt): Move from gimplify.c.
9484 (class gimplify_hasher): New.
9485 (struct gimplify_ctx): Move from gimple.h.
9486 (htab_t gimplify_ctx::temp_htab):
9487 Change type to hash_table. Update dependent calls and types.
9488
9489 * graphite-clast-to-gimple.c: Include graphite-htab.h.
9490 (htab_t ivs_params::newivs_index):
9491 Change type to hash_table. Update dependent calls and types.
9492 (htab_t ivs_params::params_index): Likewise.
9493 (htab_t print_generated_program::params_index): Likewise.
9494 (htab_t gloog::newivs_index): Likewise.
9495 (htab_t gloog::params_index): Likewise.
9496
9497 * graphite.c: Include graphite-htab.h.
9498 4htab_t graphite_transform_loops::bb_pbb_mapping):
9499 Change type to hash_table. Update dependent calls and types.
9500
9501 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
9502 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
9503 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
9504
9505 * graphite-dependences.c: Include graphite-htab.h.
9506 (loop_is_parallel_p): Change hash table type of parameter.
9507
9508 * graphite-htab.h: New.
9509 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
9510 (extern find_pbb_via_hash): Move from graphite-poly.h.
9511 (extern loop_is_parallel_p): Move from graphite-poly.h.
9512 (extern get_loop_body_pbbs): Move from graphite-poly.h.
9513
9514 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
9515 (extern loop_is_parallel_p): Move to graphite-htab.h.
9516 (extern get_loop_body_pbbs): Move to graphite-htab.h.
9517
9518 * haifa-sched.c (htab_t delay_htab):
9519 Change type to hash_table. Update dependent calls and types.
9520 (htab_t delay_htab_i2): Likewise.
9521
9522 * ira-color.c (htab_t allocno_hard_regs_htab):
9523 Change type to hash_table. Update dependent calls and types.
9524
9525 * ira-costs.c (htab_t cost_classes_htab):
9526 Change type to hash_table. Update dependent calls and types.
9527
9528 * loop-invariant.c (htab_t merge_identical_invariants::eq):
9529 Change type to hash_table. Update dependent calls and types.
9530
9531 * loop-iv.c (htab_t bivs):
9532 Change type to hash_table. Update dependent calls and types.
9533
9534 * loop-unroll.c (htab_t opt_info::insns_to_split):
9535 Change type to hash_table. Update dependent calls and types.
9536 (htab_t opt_info::insns_with_var_to_expand): Likewise.
9537
9538 * lto-streamer.h (struct string_slot): Move from data-streamer.h
9539 (struct string_slot_hasher): New.
9540 (htab_t output_block::string_hash_table):
9541 Change type to hash_table. Update dependent calls and types.
9542
9543 * lto-streamer-in.c (freeing_string_slot_hasher): New.
9544 (htab_t file_name_hash_table):
9545 Change type to hash_table. Update dependent calls and types.
9546
9547 * lto-streamer-out.c: Update output_block::string_hash_table dependent
9548 calls and types.
9549
9550 * lto-streamer.c (htab_t tree_htab):
9551 Change type to hash_table. Update dependent calls and types.
9552
9553 * omp-low.c: Include gimplify-ctx.h.
9554
9555 * passes.c (htab_t name_to_pass_map):
9556 Change type to hash_table. Update dependent calls and types.
9557 (pass_traverse): Rename to passes_pass_traverse.
9558
9559 * plugin.c (htab_t event_tab):
9560 Change type to hash_table. Update dependent calls and types.
9561
9562 * postreload-gcse.c (htab_t expr_table):
9563 Change type to hash_table. Update dependent calls and types.
9564 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
9565
9566 * sese.c (debug_rename_map_1): Make extern.
9567 (htab_t copy_bb_and_scalar_dependences::rename_map):
9568 Change type to hash_table. Update dependent calls and types.
9569
9570 * sese.h (extern debug_rename_map): Move to .c file.
9571
9572 * store-motion.c (htab_t store_motion_mems_table):
9573 Change type to hash_table. Update dependent calls and types.
9574
9575 * trans-mem.c (htab_t tm_new_mem_hash):
9576 Change type to hash_table. Update dependent calls and types.
9577
9578 * tree-browser.c (htab_t TB_up_ht):
9579 Change type to hash_table. Update dependent calls and types.
9580
9581 * tree-cfg.c (htab_t discriminator_per_locus):
9582 Change type to hash_table. Update dependent calls and types.
9583
9584 * tree-complex.c: Include tree-hasher.h
9585 (htab_t complex_variable_components):
9586 Change type to hash_table. Update dependent calls and types.
9587
9588 * tree-eh.c (htab_t finally_tree):
9589 Change type to hash_table. Update dependent calls and types.
9590
9591 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
9592 struct int_tree_hasher.
9593 (extern int_tree_map_eq): Likewise.
9594 (uid_decl_map_hash): Removed.
9595 (extern decl_tree_map_eq): Likewise.
9596
9597 * tree-hasher.h: New.
9598 (struct int_tree_hasher): New.
9599 (typedef int_tree_htab_type): New.
9600
9601 * tree-inline.c: Include gimplify-ctx.h.
9602
9603 * tree-mudflap.c: Include gimplify-ctx.h.
9604
9605 * tree-parloops.c: Include tree-hasher.h.
9606 (htab_t eliminate_local_variables_stmt::decl_address):
9607 Change type to hash_table. Update dependent calls and types.
9608 (htab_t separate_decls_in_region::decl_copies): Likewise.
9609
9610 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
9611 Change type to hash_table. Update dependent calls and types.
9612
9613 * tree-sra.c (candidates):
9614 Change type to hash_table. Update dependent calls and types.
9615
9616 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
9617 in tree-flow.h.
9618 (int_tree_map_hash): Likewise.
9619
9620 * tree-ssa-dom.c (htab_t avail_exprs):
9621 Change type to hash_table. Update dependent calls and types.
9622
9623 * tree-ssa-live.c (var_map_base_init::tree_to_index):
9624 Change type to hash_table. Update dependent calls and types.
9625
9626 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
9627 Change type to hash_table. Update dependent calls and types.
9628
9629 * tree-ssa-phiopt.c (seen_ssa_names):
9630 Change type to hash_table. Update dependent calls and types.
9631
9632 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
9633 Change type to hash_table. Update dependent calls and types.
9634
9635 * tree-ssa-uncprop.c (equiv):
9636 Change type to hash_table. Update dependent calls and types.
9637
c5a44004
JJ
96382013-04-25 Jakub Jelinek <jakub@redhat.com>
9639
9640 PR rtl-optimization/57003
9641 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
9642 call note_stores with kill_clobbered_value callback again after
9643 killing regs_invalidated_by_call.
9644
09962a4a
JG
96452013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9646
9647 * config/aarch64/aarch64-simd.md
9648 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
9649 (aarch64_simd_bsl<mode>): Likewise.
9650 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
9651
ea28bb0b
MP
96522013-04-25 Marek Polacek <polacek@redhat.com>
9653
9654 PR tree-optimization/57066
3c21604f 9655 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 9656
96659611
JG
96572013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
9658
9659 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
9660
9697e620
JG
96612013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9662
9663 * config/aarch64/aarch64-builtins.c
9664 (aarch64_fold_builtin): New.
9665 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
9666 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
9667 * config/aarch64/aarch64-simd-builtins.def (abs): New.
9668 * config/aarch64/arm_neon.h
9669 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
9670
0ac198d3
JG
96712013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9672 Tejas Belagod <tejas.belagod@arm.com>
9673
9674 * config/aarch64/aarch64-builtins.c
9675 (aarch64_gimple_fold_builtin): New.
9676 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
9677 * config/aarch64/aarch64-simd-builtins.def (addv): New.
9678 * config/aarch64/aarch64-simd.md (addpv4sf): New.
9679 (addvv4sf): Update.
9680 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
9681
58cff58c
N
96822013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9683
df93505e 9684 * config/aarch64/aarch64.md
58cff58c
N
9685 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
9686
7e0228bf
N
96872013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9688
9689 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
9690 (*ngcsi_uxtw): New pattern.
9691
5819f96f 96922013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 9693 Julian Brown <julian@codesourcery.com>
5819f96f
KT
9694
9695 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
9696 (TB_DREG): Add T_V4HF.
9697 (v4hf_UP): New macro.
9698 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 9699 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
9700 Handle initialisation of V4HF. Adjust initialisation of reinterpret
9701 built-ins.
df93505e 9702 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
9703 (arm_vector_mode_supported_p): Handle V4HF.
9704 (arm_mangle_map): Handle V4HFmode.
9705 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
9706 * config/arm/arm_neon_builtins.def: Add entries for
9707 vcvtv4hfv4sf, vcvtv4sfv4hf.
9708 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
9709 (neon_vcvtv4hfv4sf): Likewise.
9710 * config/arm/neon-gen.ml: Handle half-precision floating point
9711 features.
9712 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
9713 * config/arm/arm_neon.h: Regenerate.
9714 * config/arm/neon.ml (type elts): Add F16.
9715 (type vectype): Add T_float16x4, T_floatHF.
9716 (type vecmode): Add V4HF.
9717 (type features): Add Requires_FP_bit feature.
9718 (elt_width): Handle F16.
9719 (elt_class): Likewise.
9720 (elt_of_class_width): Likewise.
9721 (mode_of_elt): Refactor.
9722 (type_for_elt): Handle F16, fix error messages.
9723 (vectype_size): Handle T_float16x4.
9724 (vcvt_sh): New function.
9725 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
9726 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
9727 (string_of_mode): Handle V4HF.
9728 * doc/arm-neon-intrinsics.texi: Regenerate.
9729
1ef395e4
JG
97302013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
9731
9732 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
9733 format specifier in 'X' case.
9734
41c34e94
AM
97352013-04-25 Alan Modra <amodra@gmail.com>
9736
9737 PR target/57052
9738 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
9739 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
9740 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
9741 Repeat for many other rotate/shift and mask patterns using subregs.
9742 Name lshiftrt insns.
9743 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
9744 on WORDS_BIG_ENDIAN.
9745
b9a7eb5d
AM
97462013-04-25 Alan Modra <amodra@gmail.com>
9747
9748 * config.gcc: Support little-endian powerpc-linux targets.
9749 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
9750 (LINK_OS_LINUX_SPEC): Define.
9751 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
9752 Preserve MASK_LITTLE_ENDIAN.
9753 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
9754 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
9755 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
9756 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
9757 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
9758 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
9759 Correct fp word order for little-endian. Don't shift toc entries
9760 smaller than a word for little-endian.
9761 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
9762 (bswapdi2 splits): Correct low-part subreg for little-endian.
9763 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
9764 low/high where such is correct only for be.
9765 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
9766 little-endian for -mcall-aixdesc.
9767
87f73374
AM
97682013-04-25 Alan Modra <amodra@gmail.com>
9769
9770 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9771 replace_equiv_address_nv.
9772
cabf91cd
AM
97732013-04-25 Alan Modra <amodra@gmail.com>
9774
9775 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
9776
0ae24cc8
VM
97772013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9778
9779 Revert:
9780 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9781 * rtl.h (struct rtx_def): ...
cabf91cd 9782
77bce07c
VM
97832013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9784
9785 PR rtl-optimizations/57046
9786 * lra-constraints (split_reg): Set up lra_risky_transformations_p
9787 for multi-reg splits.
9788
0db63e7f
L
97892013-04-24 H.J. Lu <hongjiu.lu@intel.com>
9790
9791 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
9792
3362b6b6
SA
97932013-04-24 Sterling Augustine <saugustine@google.com>
9794
9795 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
9796 (comp_dir_string, debug_str_dwo_section): New.
9797 (DEBUG_STR_DWO_SECTION): Rename to ...
9798 (DEBUG_DWO_STR_SECTION): ... this.
9799 (DEBUG_NORM_STR_SECTION): Delete.
9800 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
9801 (DEBUG_STR_DWO_SECTION_FLAGS): New.
9802 (find_AT_string): Move most logic to ...
9803 (find_AT_string_in_table): ... here. New.
9804 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
9805 add_skeleton_AT_string. Delete logic.
9806 (output_skeleton_debug_sections): Remove call to
9807 add_top_level_skeleton_die_attrs.
9808 (add_comp_dir_attribute): Move logic to comp_dir_string.
9809 (dwarf2out_init): Initialize debug_str_dwo_section.
9810 (output_indirect_string): Call find_string_form.
9811 (output_indirect_strings): Rewrite.
9812 (prune_unused_types): Empty skeleton_debug_str_hash.
9813 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
9814 (dwarf2out_finish): Call output_indirect_strings.
9815
e93e18e9
PC
98162013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9817
9818 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
9819
f6ce35ac
VM
98202013-04-24 Vladimir Makarov <vmakarov@redhat.com>
9821
cabf91cd 9822 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM
9823 (LRA_SUBREG_P): New macro.
9824 * recog.c (register_operand): Check LRA_SUBREG_P.
9825 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9826 stack frame.
9827 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9828 (lra_final_code_change): Skip subreg change for operators.
9829 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9830 if there are no operand changes.
9831 * lra-constraints.c (curr_insn_set): New.
9832 (match_reload): Set LRA_SUBREG_P.
9833 (emit_spill_move): Ditto.
9834 (check_and_process_move): Use curr_insn_set. Process only single
9835 set insns. Don't initialize sec_mem_p and change_p.
9836 (simplify_operand_subreg): Use LRA_SUBREG_P.
9837 (reg_in_class_p): New function.
9838 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9839 of #ifdef. Add code to remove cycling.
9840 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9841 non-null disp. Reload inner instead of disp when base and index
9842 are null. Try to put lo_sum into register.
9843 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 9844 (check_and_process_move): Move code for move cost check to
f6ce35ac 9845 simple_move_p. Remove equiv_substitution.
cabf91cd 9846 (simple_move_p): New function.
f6ce35ac
VM
9847 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9848 curr_insn_set. Call check_and_process_move only for single set
9849 insns. Use the new function. Move call of check_and_process_move
9850 after operand equiv substitution and address process.
9851
38047d90
JG
98522013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
9853
9854 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
9855 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
9856 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
9857
13f39b2e
PC
98582013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
9859
9860 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
9861
97c116dc
MP
98622013-04-24 Marek Polacek <polacek@redhat.com>
9863
9864 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
9865 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
9866 (select_loops_exit_conditions): Likewise.
9867 (number_of_iterations_for_all_loops): Likewise.
9868 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
9869 (scev_analysis): Likewise.
9870
83082391
CM
98712013-04-02 Catherine Moore <clm@codesourcery.com>
9872 Chao-ying Fu <fu@mips.com>
9873
cabf91cd
AM
9874 * config/mips/micromips.md (jraddiusp): New pattern.
9875 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
9876 instruction if possible.
83082391 9877
19e34aa2
AM
98782013-04-24 Alan Modra <amodra@gmail.com>
9879
9880 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
9881
fdb6603c
JB
98822013-04-24 Julian Brown <julian@codesourcery.com>
9883 Chung-Lin Tang <cltang@codesourcery.com>
9884
9885 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB
9886 dependency behavior in enumeration type DIE generation. Add TODO note
9887 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 9888
bf190e8d
LC
98892013-04-23 Lawrence Crowl <crowl@google.com>
9890
9891 * Makefile.in: Update as needed below.
9892
9893 * hash-table.h (class hash_table):
9894 Correct many methods with parameter types compare_type to the correct
9895 value_type. (Correct code was unlikely to notice the change.)
9896 (hash_table::elements_with_deleted) New.
9897 (class hashtable::iterator): New.
9898 (hashtable::begin()): New.
9899 (hashtable::end()): New.
9900 (FOR_EACH_HASH_TABLE_ELEMENT): New.
9901
9902 * statistics.c (statistics_hashes):
9903 Change type to hash_table. Update dependent calls and types.
9904
9905 * tree-into-ssa.c (var_infos):
9906 Change type to hash_table. Update dependent calls and types.
9907
9908 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
9909 Change type to hash_table. Update dependent calls and types.
9910
9911 * tree-ssa-loop-im.c (struct mem_ref.refs):
9912 Change type to hash_table. Update dependent calls and types.
9913
9914 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
9915 Change type to hash_table. Update dependent calls and types.
9916
9917 * tree-ssa-sccvn.c (vn_tables_s::nary):
9918 Change type to hash_table. Update dependent calls and types.
9919 (vn_tables_s::phis): Likewise.
9920 (vn_tables_s::references): Likewise.
9921
9922 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
9923 (vn_reference_eq): Update parameter and return types.
9924
9925 * tree-ssa-structalias.c (pointer_equiv_class_table):
9926 Change type to hash_table. Update dependent calls and types.
9927 (location_equiv_class_table): Likewise.
9928
9929 * tree-vect-data-refs.c: Consequential changes for making
9930 peeling a hash_table.
9931
9932 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
9933 (destroy_loop_vec_info): Dependent hash_table update.
9934
9935 * tree-vectorizer.h (peeling_htab):
9936 Change type to hash_table. Update dependent calls and types.
9937
d70a81dd
SC
99382013-04-23 Shiva Chen <shiva0217@gmail.com>
9939
cabf91cd
AM
9940 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
9941 to check the register content is equal or not.
9942 * lra-constraints.c (match_reload): Use lra_assign_reg_val
9943 to assign register content record.
9944 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 9945 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM
9946 * lra-int.h (struct lra_reg): Add offset member.
9947 (lra_reg_val_equal_p): New static inline function.
9948 (lra_update_reg_val_offset): New static inline function.
9949 (lra_assign_reg_val): New static inline function.
9950 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
9951 to assign register content record.
9952 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 9953
b894a1f3
CM
99542013-04-23 Catherine Moore <clm@codesourcery.com>
9955
9956 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
9957 operands. Record compression.
9958
ecd14de9
XDL
99592013-04-23 Xinliang David Li <davidxl@google.com>
9960
9961 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
9962
92e776e9
RB
99632013-04-23 Richard Biener <rguenther@suse.de>
9964
9965 PR middle-end/57036
9966 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
9967 parameter, only add abnormal goto edges from the copied body
9968 if the call could perform abnormal gotos.
9969 (copy_cfg_body): Adjust.
9970
a15ee567
SN
99712013-04-23 Sofiane Naci <sofiane.naci@arm.com>
9972
9973 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
9974
08c52234
AS
99752013-04-23 Andreas Schwab <schwab@linux-m68k.org>
9976
9977 * coretypes.h (gimple_stmt_iterator): Add struct to make
9978 compatible with C.
9979
999c1171
RB
99802013-04-23 Richard Biener <rguenther@suse.de>
9981
9982 PR tree-optimization/57026
9983 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
9984 from SSA names occuring in abnormal PHI nodes.
9985
53669259
AK
99862013-04-22 Andi Kleen <ak@linux.intel.com>
9987
9988 * lto/lto.c (print_lto_report_1): Fix LTO report names.
9989
1a0ad150
AK
99902013-04-22 Andi Kleen <ak@linux.intel.com>
9991
9992 * lto/lto.c (print_lto_report_1): Declare early.
9993 (read_cgraph_and_symbols): Call print_lto_report_1 early.
9994
057f8f20
AK
99952013-04-22 Andi Kleen <ak@linux.intel.com>
9996
9997 * common.opt (-flto-report-wpa): Add.
9998 * doc/invoke.texi (-flto-report-wpa): Add.
9999 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10000 (lto_main): dito.
10001
473b1e05
XDL
100022013-04-22 Xinliang David Li <davidxl@google.com>
10003
10004 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10005 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10006 * Makefile.in: New dependency
10007
cabf91cd 10008 David Daney <ddaney.cavm@gmail.com>
b1485a33 10009
cabf91cd
AM
10010 * configure.ac (gcc_cv_as_micromips_support): Use the
10011 --fatal-warnings option.
10012 * configure: Regenerate.
b1485a33 10013
829d0168
MP
100142013-04-22 Marek Polacek <polacek@redhat.com>
10015
10016 PR sanitizer/56990
10017 * tsan.c (instrument_expr): Don't instrument expression
10018 in case its size is zero.
10019
6d9b7208
UB
100202013-04-22 Uros Bizjak <ubizjak@gmail.com>
10021
10022 PR target/57032
10023 Revert:
10024 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10025
10026 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10027
ea679d55
JG
100282013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10029
10030 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10031 (gimple_stmt_iterator): New typedef.
10032 * gimple.h (gimple_stmt_iterator): Rename to...
10033 (gimple_stmt_iterator_d): ... This.
10034 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10035 trees be valid for GIMPLE and GENERIC.
10036 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10037 * gimple-fold.c (gimple_fold_call): Call target hook
10038 gimple_fold_builtin.
10039 * hooks.c (hook_bool_gsiptr_false): New.
10040 * hooks.h (hook_bool_gsiptr_false): New.
10041 * target.def (fold_stmt): New.
10042 * doc/tm.texi: Regenerate.
10043
88a581da
VM
100442013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10045
10046 PR target/57018
10047 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10048 a set sp if no stack realignment.
10049
92be22dc
NC
100502013-04-22 Nick Clifton <nickc@redhat.com>
10051
10052 * config.gcc (tilegx-linux): Extend extra_objs rather than
10053 overwriting it.
10054 (tilepro-linux): Likewise.
10055
0ddec79f
JG
100562013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10057
10058 * config/aarch64/aarch64-builtins.c
10059 (CF): Remove.
10060 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10061 (VAR<1-12>): Add MAP parameter.
10062 (BUILTIN_*): Likewise.
10063 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10064 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10065 (aarch64_ushl_n<mode>): Likewise.
10066 (aarch64_sshr_n<mode>): Likewise.
10067 (aarch64_ushr_n<mode>): Likewise.
10068 (aarch64_<maxmin><mode>): Likewise.
10069 (aarch64_sqrt<mode>): Likewise.
10070 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10071 (vshr<q>_n_*): Likewise.
10072
0050faf8
JG
100732013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10074
10075 * config/aarch64/aarch64-builtins.c
10076 (aarch64_simd_builtin_type_mode): Handle SF types.
10077 (sf_UP): Define.
10078 (BUILTIN_GPF): Define.
10079 (aarch64_init_simd_builtins): Handle SF types.
10080 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10081 (frecps): Likewise.
10082 (frecpx): Likewise.
10083 * config/aarch64/aarch64-simd.md
10084 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10085 (aarch64_frecpe<mode>): New.
10086 (aarch64_frecps<mode>): Likewise.
10087 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10088 (v8type): Add frecp<esx>.
10089 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10090 (aarch64_frecps<mode>): Likewise.
10091 * config/aarch64/iterators.md (FRECP): New.
10092 (frecp_suffix): Likewise.
10093 * config/aarch64/arm_neon.h
10094 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10095
0fad3dbc 100962013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB
10097
10098 PR target/56995
10099 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10100 (REG_CLASS_NAMES): Idem.
10101 (REG_CLASS_CONTENTS): Idem.
10102 (REGCLASS_HAS_FP_REG): Idem.
10103 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10104 (sh_conditional_register_usage): Idem.
10105
3e8a33f9
JL
101062013-04-21 Jeff Law <law@redhat.com>
10107
10108 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10109 (ssa_forward_propagate_and_combine): Use it.
10110
f38e1b0a
VM
101112013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10112
10113 * lra.c: Update the flow chart diagram.
10114
682303da
VM
101152013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10116
10117 PR rtl-optimization/56847
10118 * lra-constraints.c (process_alt_operands): Discourage alternative
10119 with non-matche doffsettable memory constraint fro memory with
10120 known offset.
10121
f6b64c35
RB
101222013-04-19 Richard Biener <rguenther@suse.de>
10123
10124 PR tree-optimization/56982
10125 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10126 function.
10127 * gimplify.c (gimplify_call_expr): Notice special calls.
10128 (gimplify_modify_expr): Likewise.
10129 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10130 abnormal control flow receivers.
10131 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10132 in the same way as cfun->has_nonlocal_labels.
10133 (gimple_purge_dead_abnormal_call_edges): Likewise.
10134 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10135 receivers start a basic-block.
10136
01d8bf07
RB
101372013-04-19 Richard Biener <rguenther@suse.de>
10138
10139 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10140 member ...
10141 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10142 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10143 (SLP_TREE_LOAD_PERMUTATION): Add.
10144 (vect_transform_slp_perm_load): Adjust prototype.
10145 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10146 (vect_free_slp_instance): Likewise.
10147 (vect_create_new_slp_node): Likewise.
10148 (vect_supported_slp_permutation_p): Remove.
10149 (vect_slp_rearrange_stmts): Adjust.
10150 (vect_supported_load_permutation_p): Likewise. Inline
10151 vect_supported_slp_permutation_p here.
10152 (vect_analyze_slp_instance): Compute load permutations per
10153 slp node instead of per instance.
10154 (vect_get_slp_defs): Adjust.
10155 (vect_transform_slp_perm_load): Likewise.
10156 (vect_schedule_slp_instance): Remove redundant code.
10157 (vect_schedule_slp): Remove hack for PR56270, add it ...
10158 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10159 CSE loads for SLP. Adjust.
10160
ede22fc3
GY
101612013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10162
10163 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10164 spelling in two comments.
10165
67bc84fb
GY
101662013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10167
10168 PR target/56797
10169 * config/arm/arm.c (load_multiple_sequence): Require SP
10170 as base register for loads if SP is in the register list.
10171
e248d83f
MJ
101722013-04-19 Martin Jambor <mjambor@suse.cz>
10173
10174 PR tree-optimization/56718
10175 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10176 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10177 and made public. Adjusted all callers.
10178 (ipa_intraprocedural_devirtualization): New function.
10179 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10180 (ipa_intraprocedural_devirtualization): Likewise.
10181 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10182
4891e8f8
RB
101832013-04-19 Richard Biener <rguenther@suse.de>
10184
10185 PR tree-optimization/57000
10186 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10187
dad89f7c
TG
101882013-04-19 Terry Guo <terry.guo@arm.com>
10189
10190 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10191 Replace with ...
10192 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10193 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10194 (cortex_m4_fmacs): Use new reservations.
10195 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10196
72ea0d47
VM
101972013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10198
f1e6512c 10199 PR rtl-optimization/56999
72ea0d47
VM
10200 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10201 related code.
10202 (lra_coalesce): Remove split_origin_bitmap and related code.
10203 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10204 ranges if necessary.
10205
780a5b71
UB
102062013-04-18 Uros Bizjak <ubizjak@gmail.com>
10207
10208 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10209 New array.
10210 (ix86_expand_call): Remove clobbered_registers array and use
10211 x86_64_ms_sysv_extra_clobbered_registers instead.
10212 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10213 Declare here.
10214 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10215 predicate.
10216 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10217 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10218 (*call_value_rex64_ms_sysv): Ditto.
10219
6f5a366a
CC
102202013-04-18 Cary Coutant <ccoutant@google.com>
10221
10222 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10223 parent instead of die_mark.
10224
475b8f37
DN
102252013-04-18 Diego Novillo <dnovillo@google.com>
10226
10227 * gimple.c (create_gimple_tmp): New.
10228 (get_expr_type): New.
10229 (build_assign): New.
10230 (build_type_cast): New.
10231 * gimple.h (enum ssa_mode): Define.
10232 (gimple_seq_set_location): New.
10233 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 10234 to use build_assign and build_type_cast.
475b8f37 10235
08940f33
RB
102362013-04-18 Richard Biener <rguenther@suse.de>
10237
10238 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10239 handle negative step. Remove redundant checks.
10240 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10241 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10242 for negative step and grouped loads fail to vectorize.
10243
0e0f87d4
SB
102442013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10245
10246 * emit-rtl.c (reset_insn_used_flags): New function.
10247 (reset_all_used_flags): Use it.
10248 (verify_insn_sharing): New function.
10249 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10250
4c445590
JJ
102512013-04-18 Jakub Jelinek <jakub@redhat.com>
10252
10253 PR tree-optimization/56984
10254 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10255 and (x >> M) >= N don't register any assertion if N << M is the
10256 minimum value.
10257
6873ecab
SB
102582013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10259
10260 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10261 do not delete_insn insns that have not yet been emitted, only
10262 unlink them with remove_insn.
10263 * df-scan.c (df_insn_delete): Revert r197492.
10264
3ccb989e
SB
102652013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10266
10267 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10268 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10269
102702013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY
10271
10272 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10273 define_insn_and_split.
10274 (and_scc,ior_scc,negscc): Likewise.
10275 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10276
3ccb989e 102772013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY
10278
10279 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10280 can benefit from using a sequence of LDRD instructions in epilogue
10281 instead of a single LDM instruction.
10282
6d10a203
MLI
102832013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10284
10285 PR 45688
10286 * doc/extend.texi: Fix typo.
10287
6983e6b5
RB
102882013-04-17 Richard Biener <rguenther@suse.de>
10289
10290 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10291 (vect_build_slp_tree): ... here.
10292 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10293 match. Remove special-casing of mismatched complex loads.
10294 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10295 re-try the match with swapped commutative operands.
10296 (vect_supported_load_permutation_p): Remove special-casing of
10297 mismatched complex loads.
10298 (vect_analyze_slp_instance): Adjust.
10299
ef23e6a2
RB
103002013-04-17 Richard Biener <rguenther@suse.de>
10301
10302 PR rtl-optimization/56921
10303 * cfgloop.h (struct loop): Add simple_loop_desc member.
10304 (struct niter_desc): Mark with GTY(()).
10305 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10306 * loop-iv.c (get_simple_loop_desc): Likewise.
10307 (free_simple_loop_desc): Likewise.
10308
10309 Revert
10310 2013-04-16 Richard Biener <rguenther@suse.de>
10311
10312 PR rtl-optimization/56921
10313 * loop-init.c (pass_rtl_move_loop_invariants): Add
10314 TODO_do_not_ggc_collect to todo_flags_finish.
10315 (pass_rtl_unswitch): Same.
10316 (pass_rtl_unroll_and_peel_loops): Same.
10317 (pass_rtl_doloop): Same.
10318
fc6f94f5
EB
103192013-04-17 Eric Botcazou <ebotcazou@adacore.com>
10320
10321 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
10322 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
10323 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
10324 references.
10325 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
10326 * tree-streamer.c (record_common_node): Adjust reference in comment.
10327
10a88311
TG
103282013-04-17 Terry Guo <terry.guo@arm.com>
10329
10330 * config/arm/cortex-m4.md: Add a new bypass.
10331
6d9b7208 103322013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N
10333
10334 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
10335 New pattern.
10336 (*subs_<optab><mode>_multp2): New pattern.
10337 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10338 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10339
6d9b7208 103402013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N
10341
10342 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
10343 (*subs_mul_imm_<mode>): New pattern.
10344
18a6701e
DE
103452013-04-16 David Edelsohn <dje.gcc@gmail.com>
10346
10347 PR target/56948
10348 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
10349 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
10350 (vsx_movti_32bit): Change j->wa to O->wa.
10351
07c37b2f
RB
103522013-04-16 Richard Biener <rguenther@suse.de>
10353
10354 PR rtl-optimization/56921
10355 * loop-init.c (pass_rtl_move_loop_invariants): Add
10356 TODO_do_not_ggc_collect to todo_flags_finish.
10357 (pass_rtl_unswitch): Same.
10358 (pass_rtl_unroll_and_peel_loops): Same.
10359 (pass_rtl_doloop): Same.
10360
0e0f87d4 103612013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY
10362
10363 * config/arm/arm.c (emit_multi_reg_push): New declaration
10364 for an existing function.
10365 (arm_emit_strd_push): New function.
10366 (arm_expand_prologue): Used here.
10367 (arm_emit_ldrd_pop): New function.
10368 (arm_expand_epilogue): Used here.
10369 (arm_get_frame_offsets): Update condition.
10370 (arm_emit_multi_reg_pop): Add a special case for load of a single
10371 register with writeback.
10372
5e8e2af4
UB
103732013-04-16 Uros Bizjak <ubizjak@gmail.com>
10374
10375 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
10376 description.
10377
9fd9ccf7
RB
103782013-04-16 Richard Biener <rguenther@suse.de>
10379
10380 PR tree-optimization/56756
10381 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
10382 (first_mem_ref_loc): New.
10383 (execute_sm): Place the load temporarily before a previous
10384 access instead of in the latch edge to ensure its SSA dependencies
10385 are defined at points dominating the load.
10386
96fba521
SB
103872013-04-16 Steven Bosscher <steven@gcc.gnu.org>
10388
4c8af858
SB
10389 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
10390 correct fix by moving header and footer insn to the footer of
10391 the merged basic block. Clear BB_END of the merged-away block.
10392
96fba521
SB
10393 PR middle-end/43631
10394 * emit-rtl.c (make_note_raw): New function.
10395 (link_insn_into_chain): New static inline function.
10396 (add_insn): Use it.
10397 (add_insn_before, add_insn_after): Factor insn chain linking code...
10398 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
10399 using link_insn_into_chain.
10400 (note_outside_basic_block_p): New helper function for emit_note_after
10401 and emit_note_before.
10402 (emit_note_after): Use nobb variant of add_insn_after if the note
10403 should not be contained in a basic block.
10404 (emit_note_before): Use nobb variant of add_insn_before if the note
10405 should not be contained in a basic block.
10406 (emit_note_copy): Use make_note_raw.
10407 (emit_note): Likewise.
10408 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
10409 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
10410 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
10411 the moved barrier the tail of the basic block it follows.
10412 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
10413
7b8265ba
JJ
104142013-04-15 Jakub Jelinek <jakub@redhat.com>
10415
10416 PR tree-optimization/56962
10417 * gimple-ssa-strength-reduction.c (record_increment): Only set
10418 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
10419 either rhs1 or rhs2 is equal to c->base_expr.
10420
5185d248
RB
104212013-04-15 Richard Biener <rguenther@suse.de>
10422
10423 PR tree-optimization/56933
10424 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
10425 member.
10426 (GROUP_READ_WRITE_DEPENDENCE): Remove.
10427 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
10428 * tree-vect-data-refs.c (vect_analyze_group_access): Move
10429 dependence check ...
10430 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
10431 ... here.
10432 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
10433 GROUP_READ_WRITE_DEPENDENCE.
10434
a24243a0
AK
104352013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10436
10437 * emit-rtl.c (reset_all_used_flags): New function.
10438 (verify_rtl_sharing): Call reset_all_used_flags before and after
10439 performing the checks.
10440
1c50eada
KT
104412013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10442
10443 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
10444 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
10445 * config/arm/constraints.md (De): New constraint.
10446 * config/arm/neon.md (anddi3_neon): Delete.
10447 (neon_vand<mode>): Expand to standard anddi3 pattern.
10448 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
10449 Move earlier in the file.
10450 (neon_inv_logic_op2): Likewise.
10451 (arm_anddi_operand_neon): New predicate.
10452
e927b6ad
RO
104532013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10454
10455 * configure.ac (gcc_cv_ld_as_needed): Set
10456 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
10457 Use -z ignore, -z record on *-*-solaris2*.
10458 (HAVE_LD_AS_NEEDED): Update comment.
10459 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
10460 * configure: Regenerate.
10461 * config.in: Regenerate.
10462 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
10463 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
10464 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
10465 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
10466 equivalents. Fix markup.
10467 * doc/tm.texi: Regenerate.
10468
e0ea8797
AH
104692013-04-15 Andrew Hsieh <andrewhsieh.google.com>
10470
10471 * config/i386/i386.opt: New option mstack-protector-guard=.
10472 * config/i386/i386-opts.h: Add enum stack_protector_guard.
10473 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
10474 TARGET_SSP_TLS_GUARD.
10475 * config/i386/i386.c (ix86_option_override_internal): Set
10476 ix86_stack_protector_guard.
10477 * config/i386/i386.md (stack_protect_set): Enable for
10478 TARGET_SSP_TLS_GUARD only.
10479 (stack_protect_set_<mode>): Ditto.
10480 (stack_protect_test): Ditto.
10481 (stack_protect_test_<mode>): Ditto.
10482 * doc/invoke.texi (i386 Option): Document.
10483
811b72f9
EB
104842013-04-15 Eric Botcazou <ebotcazou@adacore.com>
10485
10486 PR target/56890
10487 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
10488 (S_MODES): Set H_MODE bit.
10489 (SF_MODES): Set only S_MODE and SF_MODE bits.
10490 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
10491 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
10492 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
10493 <MODE_FLOAT>: Likewise.
10494
5529fdd6
JY
104952013-04-15 Joey Ye <joey.ye@arm.com>
10496
10497 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
10498
517b1da2
JY
104992013-04-15 Joey Ye <joey.ye@arm.com>
10500
10501 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
10502 for real far jump.
10503 (thumb_far_jump_used_p): Count instruction size and set
10504 far_jump_used.
10505
01007ae0
EB
105062013-04-14 Eric Botcazou <ebotcazou@adacore.com>
10507
10508 * reorg.c (fill_simple_delay_slots): Reindent block of code.
10509 * resource.c (mark_target_live_regs): Reformat conditional block.
10510
c46f6580
SB
105112013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10512
10513 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
10514 notes, they are emitted only just before final.
10515 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
10516
1f397f45
SB
105172013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10518
10519 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
10520 * cfgrtl.c (delete_insn): Call it here instead.
10521 * lra-spills.c (lra_final_code_change): Use delete_insn.
10522 * haifa-sched.c (sched_remove_insn): Likewise.
10523 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
10524 returning to the nop pool.
10525 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
10526 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
10527
58a51369
SB
105282013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10529
10530 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
10531 * doc/tm.texi: Regenerated.
10532
33159866
UB
105332013-04-12 Uros Bizjak <ubizjak@gmail.com>
10534
10535 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
10536 QImode checks.
10537
226e378f
SB
105382013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10539
10540 * df-core.c (df_find_def): Compare register numbers.
10541 (df_find_use): Likewise.
10542
fafb9b18
VM
105432013-04-12 Vladimir Makarov <vmakarov@redhat.com>
10544
10545 PR target/56903
10546 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
10547 lra_in_progress for return.
10548
9a946fd6
GY
105492013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10550
10551 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
10552 define_insn into define_insn_and_split and emit movsicc patterns.
10553
41b83758
GY
105542013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10555
10556 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
10557
d3afd9aa
RB
105582013-04-12 Richard Biener <rguenther@suse.de>
10559
10560 * tree-pass.h (TODO_do_not_ggc_collect): New.
10561 * passes.c (execute_one_ipa_transform_pass): Honor
10562 TODO_do_not_ggc_collect.
10563 (execute_one_pass): Likewise.
10564
10565 Revert
10566 2013-04-10 Richard Biener <rguenther@suse.de>
10567
10568 * passes.c (init_optimization_passes): Remove reload pass.
10569 * ira.c (do_reload): Merge into ...
10570 (ira): ... this.
10571 (rest_of_handle_reload): Remove.
10572 (pass_reload): Likewise.
10573 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10574 of reload for vzeroupper pass placement.
10575
06f9b387
JJ
105762013-04-12 Jakub Jelinek <jakub@redhat.com>
10577
10578 PR tree-optimization/56918
10579 PR tree-optimization/56920
10580 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
10581 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
10582 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
10583 use wide_mul_with_sign method.
10584
953094d2
RB
105852013-04-12 Richard Biener <rguenther@suse.de>
10586
10587 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
10588 not be considered a gimple constant.
10589
bb506982
MG
105902013-04-12 Marc Glisse <marc.glisse@inria.fr>
10591
10592 * fold-const.c (const_binop): Handle vector shifts by a scalar.
10593 (fold_binary_loc): Call const_binop also for mixed vector-scalar
10594 operations.
10595
4b84d650
JJ
105962013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
10597 Jakub Jelinek <jakub@redhat.com>
10598
10599 * opts.c: Include diagnostic-color.h.
10600 (common_handle_option): Handle OPT_fdiagnostics_color_.
10601 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
10602 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
10603 (diagnostic-color.o): New.
10604 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
10605 (diagnostic_color_rule): New enum.
10606 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
10607 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
10608 the location string.
10609 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
10610 either NULL, or color kind.
10611 * diagnostic-color.c: New file.
10612 * diagnostic-color.h: New file.
10613 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
10614 arguments.
10615 * doc/invoke.texi (-fdiagnostics-color): Document.
10616 * pretty-print.h (pp_show_color): Define.
10617 (struct pretty_print_info): Add show_color field.
10618 * diagnostic.c: Include diagnostic-color.h.
10619 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
10620 macros. Colorize error:, warning: etc. strings and also the location
10621 string.
10622 (diagnostic_show_locus): Colorize the caret line.
10623 * pretty-print.c: Include diagnostic-color.h.
10624 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
10625 inside of %< %> quotes or quoted through q format modifier.
10626
067a1e71
AK
106272013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10628
33159866 10629 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 10630
33159866 106312013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N
10632
10633 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
10634 code in CC_NZ mode.
10635 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
10636 pattern.
10637
7b55f98f
MP
106382013-04-11 Marek Polacek <polacek@redhat.com>
10639
10640 PR tree-optimization/48184
33159866 10641 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 10642
966b587e
EB
106432013-04-11 Eric Botcazou <ebotcazou@adacore.com>
10644
10645 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
10646 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
10647 (skip_simple_arithmetic): Tidy up.
10648 * tree.h (skip_simple_constant_arithmetic): Declare.
10649
33159866 106502013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N
10651
10652 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
10653
1d42c1ec
RB
106542013-04-11 Richard Biener <rguenther@suse.de>
10655
10656 * tree-vect-loop.c (get_initial_def_for_induction): Properly
10657 generate vector constants.
10658
4ba5ea11
RB
106592013-04-11 Richard Biener <rguenther@suse.de>
10660
10661 PR tree-optimization/56878
10662 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
10663 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
10664 New function.
10665 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
10666 Prefer to align the DR with the most invariant base address.
10667
f0defe58
SKS
106682013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10669
10670 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
10671 comment.
10672
d07458be
JG
106732013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
10674
10675 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
10676 floating-point vector comparisons against 0.
10677
146b8692
JJ
106782013-04-11 Jakub Jelinek <jakub@redhat.com>
10679
10680 PR tree-optimization/56899
10681 * fold-const.c (extract_muldiv_1): Apply distributive law
10682 only if TYPE_OVERFLOW_WRAPS (ctype).
10683
b8578ff7
BC
106842013-04-11 Bin Cheng <bin.cheng@arm.com>
10685
10686 PR target/56124
10687 * ira-costs.c (scan_one_insn): Check whether the source rtx of
10688 loading has side effect.
10689
0ea8a6f9
SB
106902013-04-10 Steven Bosscher <steven@gcc.gnu.org>
10691
10692 * config/sparc/sparc.c: Include tree-pass.h.
10693 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
10694 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
10695 head of file. Change return type. Split off gate function.
10696 (sparc_gate_work_around_errata): New function.
10697 (pass_work_around_errata): New pass definition.
10698 (insert_pass_work_around_errata) New pass insert definition to
10699 insert pass_work_around_errata just after delayed-branch scheduling.
10700 (sparc_option_override): Insert the pass.
10701 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
10702
42e37616
DM
107032013-04-10 David S. Miller <davem@davemloft.net>
10704
89deeb3b
DM
10705 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
10706 or -mcpu=hypersparc.
10707
42e37616
DM
10708 * target.def (cstore_mode): New hook.
10709 * target.h: Include insn-codes.h
10710 * targhooks.c: Likewise.
10711 (default_cstore_mode): New function.
10712 * targhooks.h: Declare it.
10713 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
10714 * doc/tm.texi: Rebuild.
10715 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
10716 target hook, rather than inspecting the insn_data.
10717 * config/sparc/sparc.c (sparc_cstore_mode): New function.
10718 (TARGET_CSTORE_MODE): Redefine.
10719 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
10720 result patterns.
10721 * config/sparc/predicates.md (cstore_result_operand): New special
10722 predicate.
10723 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
10724 Use it for operand 0.
10725 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
10726 (*snesi_special): Likewise.
10727 (*snesi_zero): Likewise.
10728 (*seqsi_zero): Likewise.
10729 (*sltu_insn): Likewise.
10730 (*sgeu_insn): Likewise.
10731 (*seqdi_special): Make operand 0 and comparison operation be of
10732 DImode.
10733 (*snedi_special): Likewise.
10734 (*snedi_special_vis3): Likewise.
10735 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
10736 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
10737 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
10738 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
10739 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
10740 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
10741 (*sltu_extend_sp64): Likewise.
10742 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
10743 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
10744 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
10745 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
10746 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
10747
95ca411e
YZ
107482013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
10749
10750 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
10751 (aarch64_start_file): Use the new function.
10752
6782438d 107532013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 10754 Jason Merrill <jason@redhat.com>
6782438d
SKS
10755
10756 * common.opt: Add -gdwarf.
10757 * opts.c (common_handle_option): Handle it.
10758 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
10759
bb313b93
RB
107602013-04-10 Richard Biener <rguenther@suse.de>
10761
10762 * passes.c (execute_todo): Do not call ggc_collect conditional here.
10763 (execute_one_ipa_transform_pass): But unconditionally here.
10764 (execute_one_pass): And here.
10765 (init_optimization_passes): Remove reload pass.
10766 * tree-pass.h (TODO_ggc_collect): Remove.
10767 (pass_reload): Likewise.
10768 * ira.c (do_reload): Merge into ...
10769 (ira): ... this.
10770 (rest_of_handle_reload): Remove.
10771 (pass_reload): Likewise.
10772 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10773 of reload for vzeroupper pass placement.
10774 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
10775 and todo_flags_finish of all passes.
10776
793d9a16
RB
107772013-04-10 Richard Biener <rguenther@suse.de>
10778
10779 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
10780 first_const_oprnd field, rename first_def_type to first_op_type.
10781 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
10782 (vect_get_and_check_slp_defs): Always use the type of the
10783 operand. Allow mixed vect_external_def, vect_constant_def types.
10784 (vect_get_constant_vectors): Handle mixed vect_external_def,
10785 vect_constant_def types.
10786
ee8a9b7b
JR
107872013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
10788
10789 PR tree-optimization/55524
10790 * tree-ssa-math-opts.c
10791 (convert_mult_to_fma): Don't use an fms construct
10792 when we don't have an fms operation, but fnma, and it looks
10793 likely that we'll be able to use the latter.
10794
6957a6f6
ZZ
107952013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
10796
10797 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
10798 function.
10799 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
10800 inline fail caused by overwritable functions.
10801
34ab4a5b
CJW
108022013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
10803
10804 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
10805 unnecessary bits in the constant power of two case.
10806
abf9bfbc
RB
108072013-04-10 Richard Biener <rguenther@suse.de>
10808
10809 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
10810 broken code swapping operands.
10811 (vect_build_slp_tree): Do not compute load permutations here.
10812 (vect_analyze_slp_instance): Compute load permutations here,
10813 after building the SLP tree.
10814
f408477e
CB
108152013-04-09 Christian Bruel <christian.bruel@st.com>
10816
10817 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
10818 of next/prev_real_insn.
10819
5ac42672
JH
108202013-04-09 Jan Hubicka <jh@suse.cz>
10821
abf9bfbc
RB
10822 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10823 Drop aliased parameter.
5ac42672
JH
10824 (function_and_variable_visibility): Do not handle alias pairs.
10825 * cgraph.c (varpool_externally_visible_p): Update prototype.
10826 * varpool.c (varpool_add_new_variable): Update.
10827
5017f1d2
KT
108282013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10829
10830 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
10831
48eecbee
SB
108322013-04-09 Steven Bosscher <steven@gcc.gnu.org>
10833
a949cf1c
SB
10834 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
10835
48eecbee
SB
10836 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
10837
75ef8e3d
MP
108382013-04-09 Marek Polacek <polacek@redhat.com>
10839
10840 PR tree-optimization/48762
33159866 10841 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 10842
23847df4
RB
108432013-04-09 Richard Biener <rguenther@suse.de>
10844
10845 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
10846 dealing with cost.
10847 (vect_build_slp_tree): Likewise.
10848 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
10849 calculating the cost of a SLP instance.
10850 (vect_analyze_slp_instance): Use it from here, after building
10851 the SLP tree.
10852
ea3a0fde
JJ
108532013-04-09 Jakub Jelinek <jakub@redhat.com>
10854
10855 PR middle-end/56883
10856 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
10857 expand_omp_for_static_chunk): Use simple_p = true in
10858 force_gimple_operand_gsi calls when assigning to addressable decls.
10859
a32dfe9d
JL
108602013-04-09 Jeff Law <law@redhat.com>
10861
10862 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
10863 when the boolean was created by converting a wider object which
10864 had a boolean range.
10865
d755c7ef
RB
108662013-04-09 Richard Biener <rguenther@suse.de>
10867
10868 * tree-vectorizer.h (slp_void_p): Remove.
10869 (slp_tree): Typedef before _slp_tree declaration.
10870 (struct _slp_tree): Use a vector of slp_tree as children.
10871 (vect_get_place_in_interleaving_chain): Remove.
10872 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
10873 Move ...
10874 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
10875 and make static.
10876 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
10877 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
10878 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
10879 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
10880 Use slp_node instead of slp_void_p and adjust.
10881
3d741091
RB
108822013-04-09 Richard Biener <rguenther@suse.de>
10883
10884 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
10885 work that is not necessary.
10886
39307ba7
JJ
108872013-04-09 Jakub Jelinek <jakub@redhat.com>
10888
10889 PR tree-optimization/56854
10890 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
10891 forward into clobber stmts if it would change MEM_REF lhs into
10892 non-MEM_REF.
10893
343881fd
MK
108942013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
10895
10896 * tree.c (type_hash_lookup, type_hash_add): Make static.
10897 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
10898
3922658a
RB
108992013-04-09 Richard Biener <rguenther@suse.de>
10900
10901 * tree.h (unsave_expr_now): Remove.
10902 * tree-inline.c (mark_local_for_remap_r): Remove.
10903 (unsave_expr_1): Likewise.
10904 (unsave_r): Likewise.
10905 (unsave_expr_now): Likewise.
10906 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
10907 (propagate_tree_value): Likewise.
10908
9fb6b620
SB
109092013-04-08 Steven Bosscher <steven@gcc.gnu.org>
10910
10911 * doc/rtl.texi (sequence): Rewrite documentation to match the
10912 current use of SEQUENCE rtl objects.
10913 * rtl.def (SEQUENCE): Likewise.
10914
10915 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
10916 Update documentation.
10917 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
10918 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
10919
10920 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
10921
8ddb5a29
TJ
109222013-04-08 Teresa Johnson <tejohnson@google.com>
10923
10924 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
10925 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 10926 methods.
8ddb5a29
TJ
10927 (estimate_edge_size_and_time): Add comment to suggest using rounding
10928 methods.
10929 (estimate_node_size_and_time): Ditto.
10930 (remap_edge_change_prob): Use helper rounding divide methods.
10931 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
10932 (gimple_mod_pow2_value_transform): Ditto.
10933 (gimple_mod_subtract_transform): Ditto.
10934 (gimple_ic_transform): Ditto.
10935 (gimple_stringops_transform): Ditto.
10936 * stmt.c (conditional_probability): Ditto.
10937 (emit_case_dispatch_table): Ditto.
10938 * lto-cgraph.c (merge_profile_summaries): Ditto.
10939 * tree-optimize.c (execute_fixup_cfg): Ditto.
10940 * cfgcleanup.c (try_forward_edges): Ditto.
10941 * cfgloopmanip.c (scale_loop_profile): Ditto.
10942 (loopify): Ditto.
10943 (duplicate_loop_to_header_edge): Ditto.
10944 (lv_adjust_loop_entry_edge): Ditto.
10945 * tree-vect-loop.c (vect_transform_loop): Ditto.
10946 * profile.c (compute_branch_probabilities): Ditto.
10947 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
10948 * lto-streamer-in.c (input_cfg): Ditto.
10949 * gimple-streamer-in.c (input_bb): Ditto.
10950 * ipa-cp.c (update_profiling_info): Ditto.
10951 (update_specialized_profile): Ditto.
10952 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
10953 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 10954 rounding methods.
8ddb5a29
TJ
10955 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10956 (compute_trg_info): Ditto.
10957 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
10958 (purge_dead_edges): Ditto.
10959 * loop-unswitch.c (unswitch_loop): Ditto.
10960 * cgraphclones.c (cgraph_clone_edge): Ditto.
10961 (cgraph_clone_node): Ditto.
10962 * tree-inline.c (copy_bb): Ditto.
10963 (copy_edges_for_bb): Ditto.
10964 (initialize_cfun): Ditto.
10965 (copy_cfg_body): Ditto.
10966 (expand_call_inline): Ditto.
10967
661e6bd7
KT
109682013-04-08 Kai Tietz <ktietz@redhat.com>
10969
10970 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
10971 TARGET_CYGWIN64 by TARGET_64BIT.
10972
105766f3
JR
109732013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
10974
10975 * config/epiphany/epiphany.md (GPR_1): New constant.
10976 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
10977 * config/epiphany/epiphany.c (gen_compare_reg):
10978 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
10979 is already in place.
10980 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
10981 Don't require being called during rtl expansion; If y operlaps r0,
10982 return 0.
10983 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
10984 (epiphany_expand_epilogue): Likewise.
10985
8afab237
JR
10986 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
10987 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 10988 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 10989
fa7707d6
JR
10990 * config/epiphany/constraints.md (CnL): New constraint.
10991 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
10992 * config/epiphany/predicates.md (add_operand): Allow 1024.
10993
5138e18d
JR
10994 * config/epiphany/epiphany.md (logical_op): New code iterator.
10995 (op_mnc): New code attribute.
10996 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
10997 (mov_f+1, mov_f+2): New peephole2 patterns.
10998
2ccc703d
JR
10999 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11000 (cstoresi4): Also allow re-use of zero result when doing a NE
11001 comparison to a non-zero operand.
aefb0819 11002 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 11003
093ac0a5
JR
11004 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11005 Use gen_addsi3_i / gen_subsi3_i.
11006
f223bb13
JJ
110072013-04-08 Jakub Jelinek <jakub@redhat.com>
11008
11009 PR c++/34949
11010 PR c++/50243
11011 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11012 contain anything but clobbers, at most one __builtin_stack_restore,
11013 optionally debug stmts and final resx, and if it has at least one
11014 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11015 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11016 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11017 which isn't defaut definition, remove them.
11018 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11019 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11020 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11021 with MEM_REF LHS with SSA_NAME address.
11022
4481581f
JL
110232013-04-08 Jeff Law <law@redhat.com>
11024
11025 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 11026
451dabda
RB
110272013-04-08 Richard Biener <rguenther@suse.de>
11028
11029 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11030 extra newline.
11031 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11032 determined vector type.
11033 (vect_analyze_data_refs): Likewise.
11034 (vect_get_new_vect_var): Adjust.
11035 (vect_create_destination_var): Preserve SSA name versions.
11036 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11037 not dump anything here.
11038
3b088b47
JR
110392013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11040
11041 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11042 Add member lr_slot_known.
11043 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11044 if necessary.
11045 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11046 Remove code that sets lr_slot_offset according to what a previous
11047 version of epiphany_emit_save_restore used to do.
11048 (epiphany_emit_save_restore): When doing an lr save or restore,
11049 set/verify lr_slot_known and lr_slot_offset.
11050
d8484d41
XQ
110512013-04-08 Xinyu Qi <xyqi@marvell.com>
11052
33159866
UB
11053 PR target/54338
11054 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ
11055 in ALL_REGS.
11056
481be1c4
RB
110572013-04-08 Richard Biener <rguenther@suse.de>
11058
11059 * alias.c (find_base_term): Fix thinko in previous change.
11060
401f3a81
JJ
110612013-04-08 Jakub Jelinek <jakub@redhat.com>
11062
11063 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11064 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11065 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11066 if possible to compute val.
11067 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11068 For QImode integers don't require anything about precision. Use
11069 const_with_all_bytes_same to find out if the constant doesn't have
11070 repeated bytes in it.
11071
03ed99a8
AK
110722013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11073
11074 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11075 within mode size.
11076
781b2e62
MP
110772013-04-08 Marek Polacek <polacek@redhat.com>
11078
11079 PR rtl-optimization/48182
11080 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11081 value to 1.
11082
27e430a2
JDA
110832013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11084
11085 PR target/55487
11086 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11087 nuses, make sure we have a label.
11088
4902aa64
BS
110892013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11090
11091 PR target/56843
11092 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11093 (rs6000_emit_swdiv_low_precision): Remove.
11094 (rs6000_emit_swdiv): Rewrite to handle between one and four
11095 iterations of Newton-Raphson generally; modify required number of
11096 iterations for some cases.
11097 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11098
7bca81dc
SB
110992013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11100
11101 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11102 set-but-unused variable.
11103
11104 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11105 basic blocks of released function bodies garbage-collectable.
11106
11107 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11108 (struct rtl_opt_pass): Add TODO_df_finish.
11109
11110 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11111
4542a38a
GY
111122013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11113
11114 * config/arm/constraints.md (q): New constraint.
11115 * config/arm/ldrdstrd.md: New file.
11116 * config/arm/arm.md (ldrdstrd.md) New include.
11117 (arm_movdi): Use "q" instead of "r" constraint
11118 for double-word memory access.
11119 (movdf_soft_insn): Likewise.
11120 * config/arm/vfp.md (movdi_vfp): Likewise.
11121 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 11122 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY
11123 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11124 (mem_ok_for_ldrd_strd): Likewise.
11125 (output_move_double): Update assertion.
11126
2385b218
GY
111272013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11128
11129 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11130
75fe1cb5
GY
111312013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11132
11133 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11134 define_insn_and_split.
11135 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11136
dd660e8e
GY
111372013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11138
11139 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11140 define_insn_and_split.
33159866 11141 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY
11142 (shiftsi3_compare): New pattern.
11143 (rrx): New pattern.
11144 * config/arm/unspecs.md (UNSPEC_RRX): New.
11145
045e472c
GY
111462013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11147
11148 * config/arm/arm.md (negdi_extendsidi): New pattern.
11149 (negdi_zero_extendsidi): Likewise.
11150
3f3bf1a8
GY
111512013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11152
11153 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11154 define_insn_and_split.
11155 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11156 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11157
b0b49556
GY
111582013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11159
11160 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11161 define_insn_and_split.
11162 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11163 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11164
d633dd84
GY
111652013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11166
11167 * config/arm/arm.md (subsi3_carryin): New pattern.
11168 (subsi3_carryin_const): Likewise.
11169 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11170 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11171
f4499066
GY
111722013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11173
11174 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11175
ceef6fd9
GY
111762013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11177
11178 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 11179 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 11180
ddbdd8a7
KT
111812013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11182
11183 * config/arm/arm.c (arm_expand_builtin): Change fcode
11184 type to unsigned int.
11185
8456d78a
RR
111862013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11187
11188 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11189
526c230f
ILT
111902013-04-04 Ian Lance Taylor <iant@google.com>
11191
33159866
UB
11192 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11193 language standard.
526c230f 11194
3cfbe04d
SB
111952013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11196
11197 PR middle-end/56729
11198 * df-scan.c (df_insn_delete): Disable failing assert.
11199
dfa3f8d0
KT
112002013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11201
11202 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11203 New function prototype.
11204 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11205 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11206 (arm_builtin_vectorized_function): New function.
11207
f7837758
KT
112082013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11209
11210 * config/arm/arm_neon_builtins.def: New file.
11211 * config/arm/arm.c (neon_builtin_data): Move contents to
11212 arm_neon_builtins.def.
11213 (enum arm_builtins): Include neon builtin definitions.
11214 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 11215 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 11216
39c1b6db
MP
112172013-04-04 Marek Polacek <polacek@redhat.com>
11218
11219 PR tree-optimization/48186
11220 * predict.c (maybe_hot_frequency_p): Return false if
11221 HOT_BB_FREQUENCY_FRACTION is 0.
11222 (cgraph_maybe_hot_edge_p): Likewise.
11223
314f64eb
RB
112242013-04-04 Richard Biener <rguenther@suse.de>
11225
11226 PR tree-optimization/56826
11227 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11228 more accurately.
11229
51a905b2
RB
112302013-04-04 Richard Biener <rguenther@suse.de>
11231
11232 PR tree-optimization/56213
11233 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 11234 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 11235
f1bf4f3a
RB
112362013-04-04 Richard Biener <rguenther@suse.de>
11237
11238 PR tree-optimization/56837
11239 * tree-loop-distribution.c (classify_partition): For non-zero
11240 values require that the value has the same precision as its
11241 mode to be useful as memset value.
11242
0bca7ded
NC
112432013-04-03 Nick Clifton <nickc@redhat.com>
11244
33159866 11245 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC
11246 (fmssf4): Use fmsf.s on E3V5 architectures.
11247 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11248 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11249
b4019227
JL
112502013-04-03 Jeff Law <law@redhat.com>
11251
11252 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11253 (lra-eliminations.o): Likewise.
11254
f57ddb5b
TJ
112552013-04-03 Teresa Johnson <tejohnson@google.com>
11256
11257 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 11258 compute_working_sets here from profile.c.
f57ddb5b 11259 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB
11260 (gcov_working_set_t): Moved typedef here from basic-block.h
11261 (compute_working_set): Declare.
f57ddb5b
TJ
11262 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11263 (get_working_sets): Renamed from compute_working_set,
33159866 11264 replace most of body with call to new compute_working_sets.
f57ddb5b 11265 (get_exec_counts): Replace call to compute_working_sets
33159866
UB
11266 to get_working_sets.
11267 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 11268 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 11269 to get_working_sets.
f57ddb5b
TJ
11270 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11271 * gcov-dump.c (dump_working_sets): New function.
11272
5e0919f1
KZ
112732013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11274
11275 * hwint.c (sext_hwi, zext_hwi): New functions.
11276 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11277 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11278 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11279 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11280 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11281 (sext_hwi, zext_hwi): New functions.
0bca7ded 11282
be672e08
JL
112832013-04-03 Jeff Law <law@redhat.com>
11284
11285 PR tree-optimization/56799
33159866
UB
11286 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11287 back test for widening conversion erroneously dropped in prior change.
be672e08 11288
9d821fa5
KT
112892013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11290
11291 PR target/56809
11292 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11293 instead of next_real_insn.
11294
71c581e7
MP
112952013-04-03 Marek Polacek <polacek@redhat.com>
11296
11297 PR sanitizer/55702
33159866 11298 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 11299
4a32ef80
KT
113002013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11301
11302 PR target/56809
11303 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11304 next_real_insn.
11305 (thumb1_output_casesi): Likewise.
11306 (thumb2_output_casesi): Likewise.
11307
1b2253d4
RB
113082013-04-03 Richard Biener <rguenther@suse.de>
11309
11310 PR tree-optimization/56817
11311 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11312 Split out ...
11313 (tree_unroll_loops_completely_1): ... new function to manually
11314 walk the loop tree, properly defering outer loops of unrolled
11315 loops to later iterations.
11316
38000232
MG
113172013-04-03 Marc Glisse <marc.glisse@inria.fr>
11318
11319 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
11320 (vectorizable_load): Likewise.
11321 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
11322 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
11323
3fa35298
MG
113242013-04-03 Marc Glisse <marc.glisse@inria.fr>
11325
11326 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
11327 BIT_FIELD_REF.
11328
b3d45ff0
UW
113292013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11330
11331 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
11332
ec9202a8
BC
113332013-04-03 Bin Cheng <bin.cheng@arm.com>
11334
11335 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
11336
6805bd36
MG
113372013-04-03 Marc Glisse <marc.glisse@inria.fr>
11338
11339 PR tree-optimization/56790
33159866
UB
11340 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
11341 folding.
6805bd36 11342
da694a77
MG
113432013-04-03 Marc Glisse <marc.glisse@inria.fr>
11344
11345 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
11346 Handle VEC_MERGE.
11347 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
11348 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
11349 equal arguments.
11350
4e7d7b3d
JJ
113512013-04-03 Jakub Jelinek <jakub@redhat.com>
11352
11353 PR c/19449
11354 * tree.h (force_folding_builtin_constant_p): New decl.
11355 * builtins.c (force_folding_builtin_constant_p): New variable.
11356 (fold_builtin_constant_p): Fold immediately also if
11357 force_folding_builtin_constant_p.
11358
e6c9d234
RB
113592013-04-03 Richard Biener <rguenther@suse.de>
11360
11361 PR tree-optimization/56812
11362 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
11363 DRs of the same interleaving chain are independent.
11364
984d07dd
JM
113652013-04-02 Jason Merrill <jason@redhat.com>
11366
11367 * gdbinit.in (pbb): Use debug fn.
11368
622849c9
LC
113692013-04-02 Lawrence Crowl <crowl@google.com>
11370
11371 * sese.h (struct ivtype_map_elt_s): Remove unused.
11372 (extern debug_ivtype_map): Remove unused.
11373 (extern eq_ivtype_map_elts): Remove unused.
11374 * sese.c (debug_ivtype_map): Removed unused.
11375 (debug_ivtype_map_1): Removed unused.
11376 (debug_ivtype_elt): Remove unused.
11377 (eq_ivtype_map_elts): Remove unused.
11378
11379
82c0e1a0
KT
113802013-04-02 Kai Tietz <ktietz@redhat.com>
11381
11382 PR target/52790
11383 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
11384 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB
11385 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
11386 function.
82c0e1a0
KT
11387 (legitimize_pe_coff_symbol): Likewise.
11388 (is_imported_p): New helper-function.
11389 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
11390 for Windows x64 targets.
11391 (ix86_expand_prologue): Optimize for pe-coff targets.
11392 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
11393 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
11394 medium/large code-model.
11395 (legitimize_pic_address): Likewise.
11396 (legitimize_tls_address): Likewise.
11397 (ix86_expand_call): Likewise.
11398 (x86_output_mi_thunk): Likewise.
11399 (get_dllimport_decl): Add new beimport argument.
11400 (construct_plt_address): Don't assert for x64 pe-coff targets.
11401 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
11402 targets.
11403 (SYMBOL_FLAG_STUBVAR): New macro.
11404 (SYMBOL_REF_STUBVAR_P): Likewise.
11405 * config/i386/winnt.c (stub_list): New structure.
11406 (stub_head): New local variable.
11407 (i386_pe_record_stub): New function.
11408 (i386_pe_file_end): Emit refptr-stubs.
11409
5d751b0c
JJ
114102013-04-02 Jakub Jelinek <jakub@redhat.com>
11411
09bb4c99
JJ
11412 PR rtl-optimization/56745
11413 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
11414 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
11415
a1d8947a
JJ
11416 PR c++/34949
11417 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
11418 and both of them are MEM_REFs, just compare first argument for
11419 equality and attempt to deal even with differing offsets.
11420
5d751b0c
JJ
11421 PR c++/34949
11422 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
11423 of gimple_clobber_p to be MEM_REF.
11424 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
11425 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
11426 after gimplification.
11427 * asan.c (get_mem_ref_of_assignment): Don't instrument
11428 gimple_clobber_p stmts.
11429 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
11430 gimple_clobber_p stmt if they have MEM_REF lhs and
11431 are dead because of another gimple_clobber_p stmt.
11432 * tree-ssa-live.c (clear_unused_block_pointer): Treat
11433 gimple_clobber_p stmts like debug stmts.
11434 (remove_unused_locals): Remove clobbers with MEM_REF lhs
11435 that refer to unused VAR_DECLs or uninitialized values.
11436 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
11437 gimple_clobber_p stmts if they refer to removed parameters.
11438 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
11439 formatting.
11440
e594716a
UB
114412013-04-02 Uros Bizjak <ubizjak@gmail.com>
11442
11443 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
11444 using SWI48 mode attribute.
11445
7a80735b
WM
114462013-04-02 Wei Mi <wmi@google.com>
11447
11448 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
11449 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
11450 *<rotate_insn><mode>3_mask in i386.md.
11451
f423a9e4
AI
114522013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
11453
11454 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
11455
90eb75f2
RB
114562013-04-02 Richard Biener <rguenther@suse.de>
11457
11458 PR tree-optimization/56778
11459 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
11460 Runtime alias tests are not supported for gather loads.
11461 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
11462 stmts referenced from SSA operands before updating SSA form.
11463
d8c69a92
IC
114642013-04-02 Ian Caulfield <ian.caulfield@arm.com>
11465 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11466
11467 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB
11468 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
11469 * config/arm/cortex-a53.md: New file.
11470 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
11471 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
11472 * config/arm/arm.c (arm_issue_rate): Likewise.
11473 * config/arm/arm-tune.md: Regenerate
11474 * config/arm/arm-tables.opt: Regenerate.
11475 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 11476
239eb04c
ZC
114772013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
11478
11479 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
11480 non-static link.
11481
c902d3c8
SN
114822013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11483
d8c69a92
IC
11484 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
11485 scalar load/store operations using B/H registers.
c902d3c8
SN
11486 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
11487
051d0e2f
SN
114882013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11489
11490 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
11491 scalar move.
11492 * config/aarch64/aarch64.c
11493 (aarch64_simd_scalar_immediate_valid_for_move): New.
11494 * config/aarch64/aarch64-protos.h
11495 (aarch64_simd_scalar_immediate_valid_for_move): New.
11496 * config/aarch64/constraints.md (Dh, Dq): New.
11497 * config/aarch64/iterators.md (hq): New.
11498
0ee1e3d9
EB
114992013-04-02 Eric Botcazou <ebotcazou@adacore.com>
11500
11501 * reorg.c (get_branch_condition): Deal with conditional returns.
11502 (fill_simple_delay_slots): Remove dead code dealing with jumps.
11503
136fb3f7
WM
115042013-04-01 Wei Mi <wmi@google.com>
11505
11506 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
11507 Truncate operand 2 using %b asm operand modifier.
11508 (*<shift_insn><mode>3_mask): Ditto.
11509 (*<rotate_insn><mode>3_mask): Ditto.
11510
6388c738
SB
115112013-04-01 Steven Bosscher <steven@gcc.gnu.org>
11512
11513 PR middle-end/56798
11514 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
11515
f7a4d826
KK
115162013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
11517
11518 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
11519 of next_real_insn.
11520 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
11521
4b943a49
LC
115222013-03-30 Lawrence Crowl <crowl@google.com>
11523
11524 * dse.c (clear_alias_sets): Remove never set.
11525 (disqualified_clear_alias_sets): Remove never set.
11526 (clear_alias_mode_pool): Remove never set.
11527 (dse_step0): Remove condition that is never true.
11528 (canon_address): Remove condition that is never true.
11529 (dse_step7): Remove condition that is never true.
11530 (rest_of_handle_dse): Remove condition that is never true.
11531 (rest_of_handle_dse::did_global): Remove never read from above.
11532 (dse_step2_spill): Remove never called from above.
11533 (dse_step5_spill): Remove never called from above.
11534
39718607
SB
115352013-03-30 Steven Bosscher <steven@gcc.gnu.org>
11536
da5c6bde
SB
11537 * doc/md.texi (Standard Names) <casesi>: Update documentation for
11538 JUMP_TABLE_DATA changes.
11539 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
11540 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
11541 (Insns) <jump_table_data>: New entry.
11542 * doc/tm.texi: Regenerate.
11543
39718607
SB
11544 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
11545
11546 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
11547 for table jump at the end of a basic block using tablejump_p.
11548 * targhooks.c (default_invalid_within_doloop): Likewise.
11549 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
11550 target hook implementation that is identical to the default hook.
11551 (rs6000_invalid_within_doloop): Remove.
11552
11553 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
11554 unused variable from tablejump_p call.
11555
11556 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
11557 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
11558 (INSN_DELETED_P): Likewise.
11559 (emit_jump_table_data): New prototype.
11560 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
11561 after 4th as unused.
11562 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
11563 * sched-vis.c (print_insn): Likewise.
11564 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
11565 insn for compatibility with back ends that use next_active_insn to
11566 identify jump table data.
11567 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
11568 (remove_insn): Likewise.
11569 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
11570 to be emitted.
11571 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
11572 (emit_jump_table_data): New function.
11573
11574 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
11575 basic block, a JUMP_TABLE_DATA never is.
11576 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
11577 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
11578 off from code handling real insns.
11579 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
11580 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
11581 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
11582 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
11583 is not a NONDEBUG_INSN_P.
11584 * ira-costs.c (scan_one_insn): Likewise.
11585 * jump.c (mark_all_labels): Likewise.
11586 (mark_jump_label_1): Likewise.
11587 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11588 * lra.c (get_insn_freq): Expect all insns reaching here to be in
11589 a basic block.
11590 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
11591 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
11592 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
11593 JUMP_TABLE_DATA_P insns.
11594 (calculate_elim_costs_all_insns): Likewise.
11595 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
11596 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
11597 (delete_output_reload): Code style fixups.
11598 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
11599 insn flags on this non-insn.
11600 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
11601 as scheduling barriers, for pre-change compatibility.
11602 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
11603 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
11604
11605 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
11606 redundant JUMP_TABLE_DATA_P test.
11607 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
11608 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11609 (frv_for_each_packet): Likewise.
11610 * config/i386/i386.c (min_insn_size): Likewise.
11611 (ix86_avoid_jump_mispredicts): Likewise.
11612 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11613 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
11614 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11615 (mips16_insn_length): Robustify.
11616 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
11617 (mips16_split_long_branches): Likewise.
11618 * config/pa/pa.c (pa_combine_instructions): Likewise.
11619 * config/rs6000/rs6000.c (get_next_active_insn): Treat
11620 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
11621 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
11622 as contributing to pool range lengths.
11623 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
11624 Remove redundant JUMP_TABLE_DATA_P test.
11625 (sh_loop_align): Likewise.
11626 (split_branches): Likewise.
11627 (sh_insn_length_adjustment): Likewise.
11628 * config/spu/spu.c (get_branch_target): Likewise.
11629
0208f7da
JH
116302013-03-29 Jan Hubicka <jh@suse.cz>
11631
11632 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
11633 gcov streaming; stream hot bb threshold to ltrans.
11634 * predict.c (get_hot_bb_threshold): Break out from ....
11635 (maybe_hot_count_p): ... here.
11636 (set_hot_bb_threshold): New function.
11637 * lto-section-in.c (lto_section_name): Add profile.
11638 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
11639 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
11640 and data-streamer.h
11641 (histogram_entry): New structure.
11642 (histogram, histogram_pool): New global vars.
11643 (histogram_hash): New structure.
11644 (histogram_hash::hash): New method.
11645 (histogram_hash::equal): Likewise.
11646 (account_time_size): New function.
11647 (cmp_counts): New function.
11648 (dump_histogram): New function.
11649 (ipa_profile_generate_summary): New function.
11650 (ipa_profile_write_summary): New function.
11651 (ipa_profile_read_summary): New function.
11652 (ipa_profile): Decide on threshold.
e594716a
UB
11653 (pass_ipa_profile): Add ipa_profile_write_summary and
11654 ipa_profile_read_summary.
0208f7da
JH
11655 * Makefile.in (ipa.o): Update dependencies.
11656 * lto-streamer.h (LTO_section_ipa_profile): New section.
11657
5a6ccc94
GDR
116582013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
11659
11660 * tree.h (VAR_P): New.
11661
39385fa6
PC
116622013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
11663
11664 PR lto/56777
11665 * doc/invoke.texi ([-fwhole-program]): Fix typo.
11666
34f0d87a
SB
116672013-03-29 Steven Bosscher <steven@gcc.gnu.org>
11668
11669 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
11670 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11671 (control_flow_insn_p): Likewise.
11672 * cfgrtl.c (duplicate_insn_chain): Likewise.
11673 * final.c (get_attr_length_1): Likewise.
11674 (shorten_branches): Likewise.
11675 (final_scan_insn): Likewise.
11676 * function.c (instantiate_virtual_regs): Likewise.
11677 * gcse.c (insert_insn_end_basic_block): Likewise.
11678 * ira-costs.c (scan_one_insn): Likewise.
11679 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
11680 * lra.c (check_rtl): Likewise.
11681 * reload1.c (elimination_costs_in_insn): Likewise.
11682 * reorg.c (follow_jumps): Likewise.
11683
11684 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
11685 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
11686 (thumb_far_jump_used_p): Likewise.
11687 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
11688 (workaround_speculation): Likewise.
11689 (add_sched_insns_for_speculation): Likewise.
11690 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
11691 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
11692 (frv_for_each_packet): Likewise.
11693 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
11694 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
11695 (final_emit_insn_group_barriers): Likewise.
11696 * config/m32r/m32r.c (m32r_is_insn): Likewise.
11697 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
11698 (mips16_insn_length): Likewise.
11699 * config/pa/pa.c (pa_reorg): Likewise.
11700 (pa_combine_instructions): Likewise.
11701 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
11702 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
11703 (sh_reorg): Likewise.
11704 (split_branches): Likewise.
11705 * config/spu/spu.c (get_branch_target): Likewise.
11706
11707 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
11708 JUMP_TABLE_DATA_P.
11709
4ac761b0
KY
117102013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
11711
39385fa6 11712 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY
11713 Fix declaration name.
11714
58a49189
LC
117152013-03-28 Lawrence Crowl <crowl@google.com>
11716
11717 * graphds.h (struct graph.indicies): Remove unused.
11718 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
11719 (SCOP_ORIGINAL_PDDRS): Remove unused.
11720 * sese.h (extern insert_loop_close_phis): Removed unused.
11721 (extern insert_guard_phis): Removed unused.
11722 (extern ivtype_map_elt_info): Removed unused.
11723 (new_ivtype_map_elt): Removed unused.
11724 * sese.c (ivtype_map_elt_info): Removed unused.
11725
7b3b6ae4
LC
117262013-03-28 Lawrence Crowl <crowl@google.com>
11727
11728 * Makefile.in: Add several missing include dependences.
11729 (DUMPFILE_H): New.
11730 (test-dump.o): New. This object is not added to any executable,
11731 but is present for ad-hoc testing.
11732 * bitmap.c
11733 (debug (const bitmap_head_def &)): New.
11734 (debug (const bitmap_head_def *)): New.
11735 * bitmap.h
11736 (extern debug (const bitmap_head_def &)): New.
11737 (extern debug (const bitmap_head_def *)): New.
11738 * cfg.c
11739 (debug (edge_def &)): New.
11740 (debug (edge_def *)): New.
11741 * cfghooks.c
11742 (debug (basic_block_def &)): New.
11743 (debug (basic_block_def *)): New.
11744 * dumpfile.h
11745 (dump_node (const_tree, int, FILE *)): Correct source file.
11746 * dwarf2out.c
11747 (debug (die_struct &)): New.
11748 (debug (die_struct *)): New.
11749 * dwarf2out.h
11750 (extern debug (die_struct &)): New.
11751 (extern debug (die_struct *)): New.
11752 * gimple-pretty-print.c
11753 (debug (gimple_statement_d &)): New.
11754 (debug (gimple_statement_d *)): New.
11755 * gimple-pretty-print.h
11756 (extern debug (gimple_statement_d &)): New.
11757 (extern debug (gimple_statement_d *)): New.
11758 * ira-build.c
11759 (debug (ira_allocno_copy &)): New.
11760 (debug (ira_allocno_copy *)): New.
11761 (debug (ira_allocno &)): New.
11762 (debug (ira_allocno *)): New.
11763 * ira-int.h
11764 (extern debug (ira_allocno_copy &)): New.
11765 (extern debug (ira_allocno_copy *)): New.
11766 (extern debug (ira_allocno &)): New.
11767 (extern debug (ira_allocno *)): New.
11768 * ira-lives.c
11769 (debug (live_range &)): New.
11770 (debug (live_range *)): New.
11771 * lra-int.h
11772 (debug (lra_live_range &)): New.
11773 (debug (lra_live_range *)): New.
11774 * lra-lives.c
11775 (debug (lra_live_range &)): New.
11776 (debug (lra_live_range *)): New.
11777 * omega.c
11778 (debug (omega_pb_d &)): New.
11779 (debug (omega_pb_d *)): New.
11780 * omega.h
11781 (extern debug (omega_pb_d &)): New.
11782 (extern debug (omega_pb_d *)): New.
11783 * print-rtl.c
11784 (debug (const rtx_def &)): New.
11785 (debug (const rtx_def *)): New.
11786 * print-tree.c
11787 (debug_tree (tree): Move within file.
11788 (debug_raw (const tree_node &)): New.
11789 (debug_raw (const tree_node *)): New.
11790 (dump_tree_via_hooks (const tree_node *, int)): New.
11791 (debug (const tree_node &)): New.
11792 (debug (const tree_node *)): New.
11793 (debug_verbose (const tree_node &)): New.
11794 (debug_verbose (const tree_node *)): New.
11795 (debug_head (const tree_node &)): New.
11796 (debug_head (const tree_node *)): New.
11797 (debug_body (const tree_node &)): New.
11798 (debug_body (const tree_node *)): New.
11799 (debug_vec_tree (tree): Move and reimplement in terms of dump.
11800 (debug (vec<tree, va_gc> &)): New.
11801 (debug (vec<tree, va_gc> *)): New.
11802 * rtl.h
11803 (extern debug (const rtx_def &)): New.
11804 (extern debug (const rtx_def *)): New.
11805 * sbitmap.c
11806 (debug_raw (simple_bitmap_def &)): New.
11807 (debug_raw (simple_bitmap_def *)): New.
11808 (debug (simple_bitmap_def &)): New.
11809 (debug (simple_bitmap_def *)): New.
11810 * sbitmap.h
11811 (extern debug (simple_bitmap_def &)): New.
11812 (extern debug (simple_bitmap_def *)): New.
11813 (extern debug_raw (simple_bitmap_def &)): New.
11814 (extern debug_raw (simple_bitmap_def *)): New.
11815 * sel-sched-dump.c
11816 (debug (vinsn_def &)): New.
11817 (debug (vinsn_def *)): New.
11818 (debug_verbose (vinsn_def &)): New.
11819 (debug_verbose (vinsn_def *)): New.
11820 (debug (expr_def &)): New.
11821 (debug (expr_def *)): New.
11822 (debug_verbose (expr_def &)): New.
11823 (debug_verbose (expr_def *)): New.
11824 (debug (vec<rtx> &)): New.
11825 (debug (vec<rtx> *)): New.
11826 * sel-sched-dump.h
11827 (extern debug (vinsn_def &)): New.
11828 (extern debug (vinsn_def *)): New.
11829 (extern debug_verbose (vinsn_def &)): New.
11830 (extern debug_verbose (vinsn_def *)): New.
11831 (extern debug (expr_def &)): New.
11832 (extern debug (expr_def *)): New.
11833 (extern debug_verbose (expr_def &)): New.
11834 (extern debug_verbose (expr_def *)): New.
11835 (extern debug (vec<rtx> &)): New.
11836 (extern debug (vec<rtx> *)): New.
11837 * sel-sched-ir.h
11838 (_list_iter_cond_expr): Make inline instead of static.
11839 * sreal.c
11840 (debug (sreal &)): New.
11841 (debug (sreal *)): New.
11842 * sreal.h
11843 (extern debug (sreal &)): New.
11844 (extern debug (sreal *)): New.
11845 * tree.h
11846 (extern debug_raw (const tree_node &)): New.
11847 (extern debug_raw (const tree_node *)): New.
11848 (extern debug (const tree_node &)): New.
11849 (extern debug (const tree_node *)): New.
11850 (extern debug_verbose (const tree_node &)): New.
11851 (extern debug_verbose (const tree_node *)): New.
11852 (extern debug_head (const tree_node &)): New.
11853 (extern debug_head (const tree_node *)): New.
11854 (extern debug_body (const tree_node &)): New.
11855 (extern debug_body (const tree_node *)): New.
11856 (extern debug (vec<tree, va_gc> &)): New.
11857 (extern debug (vec<tree, va_gc> *)): New.
11858 * tree-cfg.c
11859 (debug (struct loop &)): New.
11860 (debug (struct loop *)): New.
11861 (debug_verbose (struct loop &)): New.
11862 (debug_verbose (struct loop *)): New.
11863 * tree-dump.c: Add header dependence.
11864 * tree-flow.h
11865 (extern debug (struct loop &)): New.
11866 (extern debug (struct loop *)): New.
11867 (extern debug_verbose (struct loop &)): New.
11868 (extern debug_verbose (struct loop *)): New.
11869 * tree-data-ref.c
11870 (debug (data_reference &)): New.
11871 (debug (data_reference *)): New.
11872 (debug (vec<data_reference_p> &)): New.
11873 (debug (vec<data_reference_p> *)): New.
11874 (debug (vec<ddr_p> &)): New.
11875 (debug (vec<ddr_p> *)): New.
11876 * tree-data-ref.h
11877 (extern debug (data_reference &)): New.
11878 (extern debug (data_reference *)): New.
11879 (extern debug (vec<data_reference_p> &)): New.
11880 (extern debug (vec<data_reference_p> *)): New.
11881 (extern debug (vec<ddr_p> &)): New.
11882 (extern debug (vec<ddr_p> *)): New.
11883 * tree-ssa-alias.c
11884 (debug (pt_solution &)): New.
11885 (debug (pt_solution *)): New.
11886 * tree-ssa-alias.h
11887 (extern debug (pt_solution &)): New.
11888 (extern debug (pt_solution *)): New.
11889 * tree-ssa-alias.c
11890 (debug (_var_map &)): New.
11891 (debug (_var_map *)): New.
11892 (debug (tree_live_info_d &)): New.
11893 (debug (tree_live_info_d *)): New.
11894 * tree-ssa-alias.h
11895 (extern debug (_var_map &)): New.
11896 (extern debug (_var_map *)): New.
11897 (extern debug (tree_live_info_d &)): New.
11898 (extern debug (tree_live_info_d *)): New.
11899
be77e1e5
JH
119002013-03-28 Jan Hubicka <jh@suse.cz>
11901
11902 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
11903
777e6976
IB
119042013-03-28 Ian Bolton <ian.bolton@arm.com>
11905
11906 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
11907 record only when desired or required.
11908
9cd347ae
UB
119092013-03-28 Uros Bizjak <ubizjak@gmail.com>
11910
11911 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
11912 *vec_extractv2di_1_rex64. Use x64 isa attribute.
11913
33159866 119142013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N
11915
11916 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
11917 (*andsi3_compare0_uxtw): New pattern.
11918 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
11919 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
11920
89ab31c1
JH
119212013-03-28 Jan Hubicka <jh@suse.cz>
11922
11923 * data-streamer-in.c (streamer_read_gcov_count): New function.
11924 * gimple-streamer-out.c: Include value-prof.h.
11925 (output_gimple_stmt): Output histogram.
11926 (output_bb): Use streamer_write_gcov_count.
11927 * value-prof.c: Include data-streamer.h
11928 (dump_histogram_value): Add HIST_TYPE_MAX.
11929 (stream_out_histogram_value): New function.
11930 (stream_in_histogram_value): New function.
11931 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
11932 (stream_out_histogram_value, stream_in_histogram_value): Declare.
11933 * data-streamer-out.c (streamer_write_gcov_count): New function.
11934 (streamer_write_gcov_count_stream): New function.
11935 * lto-cgraph.c (lto_output_edge): Update counter streaming.
11936 (lto_output_node): Likewise.
11937 (input_node, input_edge): Likewise.
11938 * lto-streamer-out.c (output_cfg): Update streaming.
11939 * lto-streamer-in.c (input_cfg): Likewise.
11940 * data-streamer.h (streamer_write_gcov_count,
11941 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
11942 * gimple-streamer-in.c: Include value-prof.h
11943 (input_gimple_stmt): Input histograms.
11944 (input_bb): Update profile streaming.
11945
e594716a 119462013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 11947
33159866 11948 * genmodes.c (emit_max_int): New function.
8697be17 11949 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC
11950 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
11951 Added doc.
8697be17 11952 * machmode.def: Fixed comment.
89ab31c1 11953
e594716a 119542013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ
11955
11956 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ
11957 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
11958
74946978
MP
119592013-03-28 Marek Polacek <polacek@redhat.com>
11960 Richard Biener <rguenther@suse.de>
11961
11962 PR tree-optimization/56695
11963 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
11964 build signed result of a vector comparison.
11965 * tree-cfg.c (verify_gimple_comparison): Check that a result
11966 of a vector comparison has signed type.
11967
a64b9c26
RB
119682013-03-28 Richard Biener <rguenther@suse.de>
11969
11970 PR tree-optimization/37021
11971 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
11972 do not restrict gaps between groups.
11973 * tree-vect-stmts.c (vectorizable_load): Properly account for
11974 a gap between groups.
11975
a9dc2a2f
EB
119762013-03-28 Eric Botcazou <ebotcazou@adacore.com>
11977
11978 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
11979 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
11980 is not enabled.
11981
53cb97f9
GP
119822013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
11983
11984 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
11985 * doc/extend.texi (Named Address Spaces): Ditto.
11986 (Variable Attributes): Ditto.
11987
b802ae5c
KT
119882013-03-27 Kai Tietz <ktietz@redhat.com>
11989
eddae10a
KT
11990 * config.build: Add support for cygwin x64 target.
11991 * config.gcc: Likewise.
11992 * config.host: Likewise.
11993 * configure.ac: Likewise
11994 * configure: Regenerated.
b802ae5c 11995
371e77e3 119962013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT
11997
11998 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
11999 * config/i386/t-cygwin-w64: New file.
12000 * config/i386/cygwin-w64.h: New file.
12001 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12002 and add support for x64-cygwin target.
12003 (CPP_SPEC): Likewise.
12004 (CXX_WRAP_SPEC_LIST): Undefine before define.
12005 (LIBGCJ_SONAME): Use 15 as version.
12006
f49b33cb
RB
120072013-03-27 Richard Biener <rguenther@suse.de>
12008
12009 PR tree-optimization/56716
12010 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12011 dumping for ref nodes.
12012
b37a6ce5
MJ
120132013-03-27 Martin Jambor <mjambor@suse.cz>
12014
12015 PR tree-optimization/55334
12016 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12017 restricted pointers to arrays.
12018
9469b9b2
GDR
120192013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12020
12021 * Makefile.in (.SUFFIXES): Add .cc.
12022 (.c.o): Apply same recipe for implicit rule .cc.o.
12023
7d24f650
RB
120242013-03-27 Richard Biener <rguenther@suse.de>
12025
12026 PR tree-optimization/37021
12027 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12028 REALPART/IMAGPART_EXPRs around the supported refs.
12029 * tree-ssa-structalias.c (find_func_aliases): Assume that
12030 floating-point values are not used to transfer pointers.
12031
2f251a05
AI
120322013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12033
39385fa6
PC
12034 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12035 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12036 * doc/tm.texi: Regenerate.
12037 * targhooks.h (default_has_ifunc_p): New.
12038 * targhooks.c (default_has_ifunc_p): Ditto.
12039 * config/linux-protos.h: New file.
12040 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
2f251a05
AI
12041 this hook for linux which disables support of indirect functions in
12042 android.
39385fa6
PC
12043 * config/linux-android.c: New file.
12044 * config/t-linux-android.c: Ditto.
12045 * config.gcc: Added new object file linux-android.o.
12046 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 12047 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
39385fa6 12048 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI
12049 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12050 doesn't support indirect functions.
12051 * configure: Regenerate.
12052
78b4e425
BC
120532013-03-27 Bin Cheng <bin.cheng@arm.com>
12054
12055 PR target/56102
12056 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12057 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12058 mult-word mode.
12059
a5ba7b92
AK
120602013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12061
12062 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12063
be624986
TG
120642013-03-27 Terry Guo <terry.guo@arm.com>
12065
12066 * config/arm/arm-cores.def: Added core cortex-r7.
12067 * config/arm/arm-tune.md: Regenerated.
12068 * config/arm/arm-tables.opt: Regenerated.
12069 * doc/invoke.texi: Added entry for core cortex-r7.
12070
0a514f47
WL
120712013-03-27 Walter Lee <walt@tilera.com>
12072
12073 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12074 double-decrement of next_scratch_regno.
12075
9332b0d2
WL
120762013-03-27 Walter Lee <walt@tilera.com>
12077
801d9b2a 12078 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL
12079 input operands.
12080 (insn_v1mulus): Ditto.
12081 (insn_v2muls): Ditto.
12082
f54ea5dd
WL
120832013-03-27 Walter Lee <walt@tilera.com>
12084
e594716a 12085 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL
12086 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12087
bf60f4ca
WL
120882013-03-27 Walter Lee <walt@tilera.com>
12089
e594716a 12090 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL
12091 (*sibcall_value): Ditto.
12092
e3b51eeb
WL
120932013-03-27 Walter Lee <walt@tilera.com>
12094
12095 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12096 (insn_mnz_v8qi): ... this ...
12097 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12098 vector equivalent.
12099 (insn_v<n>mnz): Replaced by ...
12100 (insn_v1mnz): ... this ...
12101 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12102 equivalent.
12103 (insn_mz_<mode>): Replaced by ...
12104 (insn_mz_v8qi): ... this ...
12105 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12106 vector equivalent.
12107 (insn_v<n>mz): Replaced by ...
12108 (insn_v1mz): ... this ...
12109 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12110 equivalent.
12111
4fc7b145
EB
121122013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12113
12114 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12115
6f33817e
RM
121162013-03-26 Roland McGrath <mcgrathr@google.com>
12117
12118 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12119 than fprintf with a non-constant, non-format string.
12120
93a4145b
UB
121212013-03-26 Uros Bizjak <ubizjak@gmail.com>
12122
12123 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12124 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12125 operand 0 predicate.
12126 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12127 attribute. Use general_x64nomem_operand as operand 1 predicate.
12128 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12129 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12130 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12131 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12132 (mov<mode>_insv_1): Remove expander. Merge insn with
12133 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12134 Use general_x64nomem_operand as operand 1 predicate.
12135 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12136 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12137 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12138 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12139 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12140 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12141 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12142 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12143 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12144 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12145 (general_x64nomem_operand): Ditto.
12146
c6a9ed5a 121472013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH
12148
12149 * config/rtems.opt: Add -pthread option.
12150
31b0a960
RB
121512013-03-26 Richard Biener <rguenther@suse.de>
12152
93a4145b
UB
12153 * alias.c (find_base_term): Avoid redundant and not used recursion.
12154 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB
12155 (true_dependence_1): Compute and pass base terms to base_alias_check.
12156 (write_dependence_p): Likewise.
12157 (may_alias_p): Likewise.
12158
79517551
SN
121592013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12160
12161 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12162 PC-relative load in SI modes and above only.
12163
a76213b9
XQ
121642013-03-26 Xinyu Qi <xyqi@marvell.com>
12165
12166 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB
12167 * config/arm/iwmmxt.md (WCGR0): Update.
12168 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 12169
37ff9355
UB
121702013-03-26 Uros Bizjak <ubizjak@gmail.com>
12171
12172 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12173 Use x64 and nox64 isa attributes.
12174
30b0317c
RB
121752013-03-26 Richard Biener <rguenther@suse.de>
12176
12177 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12178 alignment computations and rely on get_object_alignment_1
12179 for the !TYPE_P case.
12180 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12181
cbcdb140
WL
121822013-03-26 Walter Lee <walt@tilera.com>
12183
12184 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12185 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12186
dd884525 121872013-03-25 Jeff Law <law@redhat.com>
be672e08 12188
33159866
UB
12189 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12190 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL
12191 wrong version of prior patch.
12192
0051d3ec
WL
121932013-03-25 Walter Lee <walt@tilera.com>
12194
12195 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12196 TILEGX_INSN_SHUFFLEBYTES1.
12197 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12198 shufflebytes1.
12199 (tilegx_builtins): Ditto.
12200 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12201
065a3605
WL
122022013-03-25 Walter Lee <walt@tilera.com>
12203
12204 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL
12205 (floatunssisf2): New pattern.
12206 (floatsidf2): New pattern.
12207 (floatunssidf2): New pattern.
065a3605 12208
5b2a3c82
WL
122092013-03-25 Walter Lee <walt@tilera.com>
12210
12211 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12212 tests for constraint J, K, N, P.
12213
192ea533
WL
122142013-03-25 Walter Lee <walt@tilera.com>
12215
12216 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12217 Use indirect/pcrel encoding.
12218 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12219 Ditto.
12220
855e0d0b
SE
122212013-03-25 Steve Ellcey <sellcey@mips.com>
12222
12223 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12224 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12225 * config/mips/mips.c (mips_option_override): Set IMADD default.
12226 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12227 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12228 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12229 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12230 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12231
39e45653
JL
122322013-03-25 Jeff Law <law@redhat.com>
12233
12234 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12235 slightly to avoid creating and folding useless trees. Simplify
12236 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12237
7482c470
UB
122382013-03-25 Uros Bizjak <ubizjak@gmail.com>
12239
12240 * config/i386/i386.md (*zero_extendsidi2): Merge with
12241 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12242 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12243 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12244 targets. Clarify comment.
12245
4a53743e
MJ
122462013-03-25 Martin Jambor <mjambor@suse.cz>
12247
12248 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12249 pass-through jump functions differently.
12250 (ipa_read_jump_function): Likewise. Also use setter functions to set
12251 up jump functions.
12252
162712de
MJ
122532013-03-25 Martin Jambor <mjambor@suse.cz>
12254
12255 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12256 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12257 process it.
12258 (ipa_get_indirect_edge_target): New function.
12259 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12260 ipa_get_indirect_edge_target. Update all callers.
12261 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12262 ipa_get_indirect_edge_target_1 instead of calling
12263 ipa_get_indirect_edge_target.
12264 (create_specialized_node): Pass aggvlas to
12265 ipcp_discover_new_direct_edges.
12266
a5a4c20a
KT
122672013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12268
12269 * config/arm/arm.md (f_sels, f_seld): New types.
12270 (*cmov<mode>): New pattern.
12271 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12272 predicate.
12273
f992b9fc
KT
122742013-03-25 Kai Tietz <ktietz@redhat.com>
12275
12276 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12277 POSIX-printf for mingw-hosted builds.
12278
c16fd676
RB
122792013-03-25 Richard Biener <rguenther@suse.de>
12280
12281 PR middle-end/56694
12282 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12283 must-not-throw stmt location.
12284
5ad29f12
KT
122852013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12286
12287 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12288 Emit load-acquire versions when acq is true.
12289 (arm_emit_store_exclusive): Add rel parameter.
12290 Emit store-release versions when rel is true.
12291 (arm_split_compare_and_swap): Use acquire-release instructions
12292 instead.
12293 of barriers when appropriate.
12294 (arm_split_atomic_op): Likewise.
12295 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12296 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12297 (VUNSPEC_SLX): Likewise.
12298 (VUNSPEC_LDA): Likewise.
12299 (VUNSPEC_STL): Likewise.
12300 * config/arm/sync.md (atomic_load<mode>): New pattern.
12301 (atomic_store<mode>): Likewise.
12302 (arm_load_acquire_exclusive<mode>): Likewise.
12303 (arm_load_acquire_exclusivesi): Likewise.
12304 (arm_load_acquire_exclusivedi): Likewise.
12305 (arm_store_release_exclusive<mode>): Likewise.
12306
03a7dddb
CM
123072013-03-25 Catherine Moore <clm@codesourcery.com>
12308
12309 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12310 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12311 * config/mip/predicates.md (lwsp_swsp_operand,
12312 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12313 sb16_operand, db4_operand, db7_operand, ib3_operand,
12314 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
12315 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
12316 andi16_operand): New predicates.
12317 * config/mips/mips.md (compression): New attribute.
12318 (enabled): New attribute.
12319 (length): Consider compression in computing length.
12320 (shift_compression): New code attribute.
12321 (*add<mode>3): New operands. Record compression.
12322 (sub<mode>3): Likewise.
12323 (one_cmpl<mode>2): Likewise.
12324 (*and<mode>3): Likewise.
12325 (*ior<mode>3): Likewise.
12326 (unnamed pattern for xor): Likewise.
12327 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
12328 (*<optab><mode>3): Likewise.
12329 (*mov<mode>_internal: Likewise.
12330 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
12331 (mips_unsigned_immediate_p): New.
12332 (umips_lwsp_swsp_address_p): New.
12333 (m16_based_address_p): New.
12334 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
12335 (mips_unsigned_immediate_p): New prototype.
12336 (lwsp_swsp_address_p): New prototype.
12337 (m16_based_address_p): New prototype.
12338 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
12339 (mips_signed_immediate_p): New function.
12340 (m16_based_address_p): New function.
12341 (lwsp_swsp_address_p): New function.
12342 (mips_print_operand_punctuation): Recognize short delay slot insns
12343 for microMIPS.add<mode>3"
12344
f35c297f 123452013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 12346
f35c297f
KT
12347 PR target/56720
12348 * config/arm/iterators.md (v_cmp_result): New mode attribute.
12349 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
12350
051b9446
RB
123512013-03-25 Richard Biener <rguenther@suse.de>
12352
12353 PR tree-optimization/56689
12354 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
12355 any edge.
12356
374001cb
RB
123572013-03-25 Richard Biener <rguenther@suse.de>
12358
12359 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
12360 of bitmap.
12361 (memory_references): Likewise.
12362 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
12363 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
12364 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
12365 (gather_mem_refs_in_loops): Fold into ...
12366 (analyze_memory_references): ... this. Move initialization
12367 to tree_ssa_lim_initialize.
12368 (fill_always_executed_in): Rename to ...
12369 (fill_always_executed_in_1): ... this.
12370 (fill_always_executed_in): Move contains_call computation to
12371 this new function from ...
12372 (tree_ssa_lim_initialize): ... here.
12373 (tree_ssa_lim): Call fill_always_executed_in.
12374
57895947
EB
123752013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12376
12377 * postreload.c (reload_combine): Fix code detecting returns.
12378
1e1b18c1
EB
123792013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12380
12381 * function.c (emit_use_return_register_into_block): On cc0 targets,
12382 do not emit the sequence between cc0 setter and user.
12383
9216baf1
KT
123842013-03-25 Kai Tietz <ktietz@redhat.com>
12385
1e1b18c1
EB
12386 * config/i386/predicates.md (local_symbolic_operand): Interpret
12387 dll-imported symbols as none-local.
9216baf1 12388
76421b44
RB
123892013-03-25 Richard Biener <rguenther@suse.de>
12390
12391 * tree-ssa-loop-im.c (struct depend): Remove.
12392 (struct lim_aux_data): Make depends a vec of gimples.
12393 (free_lim_aux_data): Adjust.
12394 (add_dependency): Likewise.
12395 (set_level): Likewise.
12396
d154bfa2
RB
123972013-03-25 Richard Biener <rguenther@suse.de>
12398
12399 PR middle-end/56434
12400 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
12401 the pointer returned by calls with ECF_MALLOC set.
12402
2cd9804e
UB
124032013-03-24 Uros Bizjak <ubizjak@gmail.com>
12404
acef5fe0
UB
12405 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
12406
124072013-03-24 Uros Bizjak <ubizjak@gmail.com>
12408
12409 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB
12410 using MMXMODE mode iterator.
12411 (*move<mode>_internal): Merge with *movv2sf_internal and
12412 *movv2sf_internal_rex64 using MMXMODE mode iterator.
12413
225ccc68
SB
124142013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12415
85c0f02d
SB
12416 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
12417 (record_last_mem_set_info): Likewise.
12418
225ccc68
SB
12419 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
12420 of XNEWVEC followed by memset.
12421 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
12422
b64925dc
SB
124232013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12424
12425 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
12426 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
12427 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
12428 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
12429 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
12430 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
12431 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
12432 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
12433 BARRIER_P instead of GET_CODE.
12434
b9aaf52e
EB
124352013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12436
12437 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
12438 inaccuracy in the probing code.
12439
12440 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
12441 (ctrapdi4): Likewise.
12442
66de4d7c
EB
124432013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12444
12445 * calls.c (expand_call): Add missing guard to code handling return
12446 of non-BLKmode structures in MSB.
12447 * function.c (expand_function_end): Likewise.
12448
4f25c8fb
EB
124492013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12450
12451 * combine.c (try_combine): Adjust comment. Do not add the set of
12452 insn #0 if the destination indirectly is set or dies in insn #2.
12453 Tidy up code to distribute a new note.
12454
23b7850d
UB
124552013-03-22 Uros Bizjak <ubizjak@gmail.com>
12456
12457 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
12458 also for alternatives 16 and 17.
12459
bed852cf
UB
124602013-03-22 Uros Bizjak <ubizjak@gmail.com>
12461
12462 * config/i386/sse.md (*mov<mode>_internal): Merge with
12463 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
12464 Emit insn template depending on type attribute. Use
12465 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12466 movd instead of movq mnemonic for interunit moves. Rewrite mode
12467 attribute calculation. Remove unit attribute calculation.
12468 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
12469 Set prefix_data16 attribute for DImode ssemov types.
12470 Use Ym instead of y for SSE-MMX conversion alternatives.
12471 Reorder operand constraints.
12472
1e8a7937
SB
124732013-03-22 Steven Bosscher <steven@gcc.gnu.org>
12474
12475 * df.h (df_insn_delete): Adjust prototype.
12476 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
12477 and let it decide whether mark the basic block dirty.
12478 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
12479 * df-scan.c (df_insn_info_delete): New helper function, split
12480 off from df_insn_delete.
12481 (df_scan_free_bb_info): Use it.
12482 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
12483 Likewise.
12484 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
12485 that the insn is actually an insn and it has a non-NULL basic block.
12486 Do not mark basic block dirty if only deleting a DEBUG_INSN.
12487
e14d094c
RB
124882013-03-22 Richard Biener <rguenther@suse.de>
12489
12490 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
12491 dep_ref members.
12492 (mem_ref_alloc): Do not allocate them.
12493 (refs_independent_p): Do not query or maintain a cache.
12494
e9cf7316
RB
124952013-03-22 Richard Biener <rguenther@suse.de>
12496
12497 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
12498 (gather_mem_refs_in_loops): Do not compute it.
12499 (analyze_memory_references): Do not allocate it.
12500 (tree_ssa_lim_finalize): Do not free it.
12501 (for_all_locs_in_loop): Do not query all_refs_in_loop.
12502
f046e81b
RB
125032013-03-22 Richard Biener <rguenther@suse.de>
12504
12505 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 12506
a05c0ddf
IB
125072013-03-22 Ian Bolton <ian.bolton@arm.com>
12508
12509 * config/aarch64/aarch64.c (aarch64_print_operand): New
12510 format specifier for printing a constant in hex.
12511 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
12512 format specifier for printing second operand.
12513
c00217fc
RB
125142013-03-22 Richard Biener <rguenther@suse.de>
12515
12516 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
12517 bitmaps.
12518 (gather_mem_refs_in_loops): Perform store accumulation here.
12519 (create_vop_ref_mapping_loop): Remove.
12520 (create_vop_ref_mapping): Likewise.
12521 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 12522 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB
12523 (record_indep_loop): Remove.
12524 (record_dep_loop): New function.
12525 (ref_indep_loop_p_1): Adjust to only walk over references
12526 in the loop, not its subloops.
12527 (ref_indep_loop_p): Rename to ...
12528 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
12529 maintaining a more fine-grained cache.
12530 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
12531 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
12532
15d19bf8
RB
125332013-03-22 Richard Biener <rguenther@suse.de>
12534
12535 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
12536 (struct mem_ref): Make accesses_in_loop a vec of a vec of
12537 aggregate mem_ref_loc.
12538 (free_mem_ref_locs): Inline into ...
12539 (memref_free): ... this and adjust.
12540 (mem_ref_alloc): Adjust.
12541 (mem_ref_locs_alloc): Remove.
12542 (record_mem_ref_loc): Adjust.
12543 (get_all_locs_in_loop): Rewrite into ...
12544 (for_all_locs_in_loop): ... this iterator.
12545 (rewrite_mem_ref_loc): New functor.
12546 (rewrite_mem_refs): Use for_all_locs_in_loop.
12547 (sm_set_flag_if_changed): New functor.
12548 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
12549 (ref_always_accessed): New functor.
12550 (ref_always_accessed_p): Use for_all_locs_in_loop.
12551
6f37411d
MG
125522013-03-21 Marc Glisse <marc.glisse@inria.fr>
12553
12554 * tree-pass.h (PROP_gimple_lvec): New.
12555 * passes.c (dump_properties): Handle PROP_gimple_lvec.
12556 (init_optimization_passes): Move pass_lower_vector.
12557 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
12558 PROP_gimple_lvec.
12559 (pass_lower_vector): Provide PROP_gimple_lvec.
12560 (pass_lower_vector_ssa): Likewise.
12561 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
12562
511ef689
MW
125632013-03-21 Mark Wielaard <mjw@redhat.com>
12564
12565 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
12566
cc1df30b
UB
125672013-03-21 Uros Bizjak <ubizjak@gmail.com>
12568
12569 * config/i386/i386.md (*movdi_internal): Disparage slightly
12570 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
12571 conversion alternatives.
12572
c845cfe1
JJ
125732013-03-21 Jakub Jelinek <jakub@redhat.com>
12574
37e99116
JJ
12575 PR middle-end/48087
12576 * diagnostic.def (DK_WERROR): New kind.
12577 * diagnostic.h (werrorcount): Define.
12578 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
12579 promoted to DK_ERROR, increment DK_WERROR counter instead of
12580 DK_ERROR counter.
12581 * toplev.c (toplev_main): Call print_ignored_options even if
12582 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
12583 even if just werrorcount is non-zero.
12584
c845cfe1
JJ
12585 PR debug/55608
12586 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
12587 on failure.
12588 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
12589 (string_cst_pool_decl): New function.
12590 (optimize_one_addr_into_implicit_ptr): New function.
12591 (resolve_addr_in_expr): Optimize DWARF location expression
12592 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
12593 which doesn't live in memory, but has DW_AT_location or
12594 DW_AT_const_value, or refers to a string literal, into
12595 DW_OP_GNU_implicit_pointer.
12596 (optimize_location_into_implicit_ptr): New function.
12597 (resolve_addr): If removing DW_AT_location of a variable because
12598 it was DW_OP_addr of address of the variable, but the variable doesn't
12599 live in memory, try to emit const value attribute for the initializer.
12600
08e0cda6
MG
126012013-03-21 Marc Glisse <marc.glisse@inria.fr>
12602
12603 * tree.h (VECTOR_TYPE_P): New macro.
12604 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
12605 TYPE_MODE): Use it.
12606 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
12607 VEC_COND_EXPR cannot be lvalues.
12608 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
12609
d08633b4
MG
126102013-03-21 Marc Glisse <marc.glisse@inria.fr>
12611
12612 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
12613 Restrict the transformation to equal modes.
12614
e6647190
RB
126152013-03-21 Richard Biener <rguenther@suse.de>
12616
12617 PR tree-optimization/39326
12618 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
12619 (MEM_ANALYZABLE): Adjust.
12620 (record_mem_ref_loc): Move bitmap ops ...
12621 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
12622 unanalyzable refs, do not record locations for it.
12623 (analyze_memory_references): Allocate ref zero as shared
12624 unanalyzable ref.
12625 (refs_independent_p): Do not test for unanalyzed mems here.
12626 (ref_indep_loop_p_1): Special-case disambiguation against
12627 the unanalyzed ref.
cc1df30b 12628 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 12629
65074f54
CL
126302013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
12631
12632 * config/arm/arm-protos.h (tune_params): Add
12633 prefer_neon_for_64bits field.
12634 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
12635 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
12636 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
12637 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
12638 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
12639 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
12640 (arm_option_override): Handle -mneon-for-64bits new option.
12641 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
12642 (prefer_neon_for_64bits): Declare new variable.
12643 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 12644 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL
12645 (arch_enabled): Handle new arch types. Remove support for onlya8
12646 and nota8.
12647 (one_cmpldi2): Use new arch names.
bc5faa5b 12648 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL
12649 * config/arm/arm.opt (mneon-for-64bits): Add option.
12650 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
12651 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
12652 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
12653 of onlya8.
12654 * doc/invoke.texi (-mneon-for-64bits): Document.
12655
5a2d2a79
RB
126562013-03-21 Richard Biener <rguenther@suse.de>
12657
12658 PR tree-optimization/39326
12659 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
12660 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 12661 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 12662
5abe1e05
RB
126632013-03-21 Richard Biener <rguenther@suse.de>
12664
12665 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
12666 (vect_insert_into_interleaving_chain): Likewise.
12667 (vect_drs_dependent_in_basic_block): Inline ...
12668 (vect_slp_analyze_data_ref_dependence): ... here. New function,
12669 split out from ...
12670 (vect_analyze_data_ref_dependence): ... here. Simplify.
12671 (vect_check_interleaving): Simplify.
12672 (vect_analyze_data_ref_dependences): Likewise. Split out ...
12673 (vect_slp_analyze_data_ref_dependences): ... this new function.
12674 (dr_group_sort_cmp): New function.
12675 (vect_analyze_data_ref_accesses): Compute data-reference groups
12676 here instead of in vect_analyze_data_ref_dependence. Use
12677 a more efficient algorithm.
12678 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
12679 vect_slp_analyze_data_ref_dependences. Call
12680 vect_analyze_data_ref_accesses earlier.
12681 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
12682 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
12683 (vect_slp_analyze_data_ref_dependences): New prototype.
12684
cad1735b
RB
126852013-03-21 Richard Biener <rguenther@suse.de>
12686
12687 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
12688 ref is stored in the loop.
12689 (find_refs_for_sm): Walk only over all stores.
12690 (store_motion_loop): Allocate from lim_bitmap_obstack.
12691 (store_motion): Likewise.
12692
141310ef
RB
126932013-03-21 Richard Biener <rguenther@suse.de>
12694
12695 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
12696 Update virtual SSA form.
12697
5022315a
RO
126982013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12699
12700 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
12701 * configure: Regenerate.
12702 * config.in: Regenerate.
12703 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
12704 if !HAVE_LD_EH_FRAME_CIEV3.
12705
50d4421c
RB
127062013-03-21 Richard Biener <rguenther@suse.de>
12707
12708 * tree-cfg.c (verify_expr_no_block): New function.
12709 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
12710 nor DECL_VALUE_EXPR have locations with associated blocks.
12711 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
12712 (clear_unused_block_pointer): Remove code dealing with
12713 blocks in DECL_DEBUG_EXPR locations.
12714
839b422f
RB
127152013-03-21 Richard Biener <rguenther@suse.de>
12716
12717 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
12718 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
12719 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
12720 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
12721 instead of DECL_DEBUG_EXPR_IS_FROM.
12722 * gimplify.c (gimplify_modify_expr): Likewise.
12723 * tree-cfg.c (verify_expr_location_1): Likewise.
12724 * tree-complex.c (create_one_component_var): Likewise.
12725 * tree-sra.c (create_access_replacement): Likewise.
12726 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
12727 (clear_unused_block_pointer): Likewise.
12728 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
12729 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
12730 * var-tracking.c (var_debug_decl): Likewise.
12731 (track_expr_p): Likewise.
12732 * tree-inline.c (add_local_variables): Likewise. Set
12733 DECL_HAS_DEBUG_EXPR_P after copying it.
12734 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
12735 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
12736
fe04878d
UB
127372013-03-21 Uros Bizjak <ubizjak@gmail.com>
12738
12739 PR bootstrap/56656
12740 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
12741 * configure: Regenerate.
12742 * config.in: Regenerate.
12743 * config/i386/i386.md (*movdf_internal): Use
12744 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12745 movd instead of movq mnemonic for interunit moves.
12746 (*movdi_internal): Ditto.
12747
bd059b26 127482013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N
12749
12750 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
12751 (abd<mode>_3): New pattern.
12752 (aba<mode>_3): New pattern.
12753 (fabd<mode>_3): New pattern.
12754
bd059b26 127552013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N
12756
12757 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
12758 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
12759 occurrence of REGISTER_PREFIX as its empty string.
12760
4f1f78b9
JL
127612013-03-20 Jeff Law <law@redhat.com>
12762
12763 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
12764 addititional equivalences for equality comparisons between an SSA_NAME
12765 and a constant where the SSA_NAME was set from a widening conversion.
12766
327a1118
WL
127672013-03-20 Walter Lee <walt@tilera.com>
12768
12769 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
12770
813e0036
UB
127712013-03-20 Uros Bizjak <ubizjak@gmail.com>
12772
12773 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
12774 depending on type attribute.
12775 (*movti_internal): Ditto.
12776 (*movtf_internal): Ditto.
12777 (*movxf_internal): Ditto.
12778 (*movdf_internal): Ditto.
12779 (*movsf_internal): Ditto.
12780
7cf34aae
UB
127812013-03-20 Uros Bizjak <ubizjak@gmail.com>
12782
12783 * config/i386/i386.md (*movti_internal): Set prefix attribute to
12784 maybe_vex for sselog1 and ssemov types.
12785 (*movdi_internal): Reorder operand constraints.
12786 (*movsi_internal): Ditto. Set prefix attribute to
12787 maybe_vex for sselog1 and ssemov types.
12788 (*movtf_internal): Set prefix attribute to maybe_vex
12789 for sselog1 and ssemov types.
12790 (*movdf_internal): Ditto. Set prefix_data16 attribute for
12791 DImode ssemov types. Reorder operand constraints.
12792 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
12793 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
12794 attribute for SImode ssemov types. Reorder operand constraints.
12795
19321415
MJ
127962013-03-20 Martin Jambor <mjambor@suse.cz>
12797
12798 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
12799 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
12800
c6d5ff83
MM
128012013-03-20 Pat Haugen <pthaugen@us.ibm.com>
12802
12803 * config/rs6000/predicates.md (indexed_address, update_address_mem
12804 update_indexed_address_mem): New predicates.
12805 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
12806 attribute for load/store instructions.
12807 * config/rs6000/dfp.md (movsd_store): Likewise.
12808 (movsd_load): Likewise.
12809 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
12810 (unnamed HI->DI extend define_insn): Likewise.
12811 (unnamed SI->DI extend define_insn): Likewise.
12812 (unnamed QI->SI extend define_insn): Likewise.
12813 (unnamed QI->HI extend define_insn): Likewise.
12814 (unnamed HI->SI extend define_insn): Likewise.
12815 (unnamed HI->SI extend define_insn): Likewise.
12816 (extendsfdf2_fpr): Likewise.
12817 (movsi_internal1): Likewise.
12818 (movsi_internal1_single): Likewise.
12819 (movhi_internal): Likewise.
12820 (movqi_internal): Likewise.
12821 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
12822 attribute for load/store instructions.
12823 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
12824 instructions.
12825 (mov<mode>_softfloat): Likewise.
12826 (mov<mode>_hardfloat32): Likewise.
12827 (mov<mode>_hardfloat64): Likewise.
12828 (mov<mode>_softfloat64): Likewise.
12829 (movdi_internal32): Likewise.
12830 (movdi_internal64): Likewise.
12831 (probe_stack_<mode>): Likewise.
12832
128332013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
12834
12835 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
12836 floating point, and decimal floating point to reload iterator.
12837
12838 * config/rs6000/constraints.md (wl constraint): New constraints to
12839 return FLOAT_REGS if certain options are used to reduce the number
12840 of separate patterns that exist in the file.
12841 (wx constraint): Likewise.
12842 (wz constraint): Likewise.
12843
12844 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
12845 -mdebug=reg, print wg, wl, wx, and wz constraints.
12846 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
12847 Initialize the reload functions for 64-bit binary/decimal floating
12848 point types.
12849 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
12850 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
12851 create the buffer on the stack to overcome not having a 32-bit
12852 load and store.
12853 (rs6000_emit_move): Likewise.
12854 (rs6000_secondary_memory_needed_rtx): Likewise.
12855 (rs6000_alloc_sdmode_stack_slot): Likewise.
12856 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
12857 via xxlxor, just like DFmode 0.0.
12858
12859 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
12860 define as 1 if we are running on a power7 or newer.
12861 (enum r6000_reg_class_enum): Add new constraints.
12862
12863 * config/rs6000/dfp.md (movsd): Delete, combine with binary
12864 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
12865 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 12866 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM
12867 (movsd splitter): Likewise.
12868 (movsd_hardfloat): Likewise.
12869 (movsd_softfloat): Likewise.
12870
12871 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
12872 binary and decimal floating point moves.
12873 (fmove_ok): New attributes to combine binary and decimal floating
12874 point moves, and to combine power6x (mfpgpr) moves along normal
12875 floating moves.
12876 (real_value_to_target): Likewise.
12877 (f32_lr): Likewise.
12878 (f32_lm): Likewise.
12879 (f32_li): Likewise.
12880 (f32_sr): Likewise.
12881 (f32_sm): Likewise.
12882 (f32_si): Likewise.
12883 (movsf): Combine binary and decimal floating point moves. Combine
12884 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 12885 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM
12886 (mov<mode> for SFmode/SDmode); Likewise.
12887 (SFmode/SDmode splitters): Likewise.
12888 (movsf_hardfloat): Likewise.
12889 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
12890 (movsf_softfloat): Likewise.
12891 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
12892
12893 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
12894 wx and wz constraints.
12895
12896 * config/rs6000/constraints.md (wg constraint): New constraint to
12897 return FLOAT_REGS if -mmfpgpr (power6x) was used.
12898
12899 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
12900 constraint.
12901
12902 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
12903 -mdebug=reg, print wg, wl, wx, and wz constraints.
12904 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
12905 Initialize the reload functions for 64-bit binary/decimal floating
12906 point types.
12907 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
12908 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
12909 create the buffer on the stack to overcome not having a 32-bit
12910 load and store.
12911 (rs6000_emit_move): Likewise.
12912 (rs6000_secondary_memory_needed_rtx): Likewise.
12913 (rs6000_alloc_sdmode_stack_slot): Likewise.
12914 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
12915 via xxlxor, just like DFmode 0.0.
12916
c6d5ff83
MM
12917 * config/rs6000/dfp.md (movdd): Delete, combine with binary
12918 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
12919 with other moves by using conditional constraits (wg). Use LFIWZX
12920 and STFIWX for loading SDmode on power7.
12921 (movdd splitters): Likewise.
12922 (movdd_hardfloat32): Likewise.
12923 (movdd_softfloat32): Likewise.
12924 (movdd_hardfloat64_mfpgpr): Likewise.
12925 (movdd_hardfloat64): Likewise.
12926 (movdd_softfloat64): Likewise.
12927
12928 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
12929 64-bit binary and decimal floating point moves.
12930 (FMOVE64X): Likewise.
12931 (movdf): Combine 64-bit binary and decimal floating point moves.
12932 Combine power6x (mfpgpr) moves with other moves by using
12933 conditional constraits (wg).
12934 (mov<mode> for DFmode/DDmode): Likewise.
12935 (DFmode/DDmode splitters): Likewise.
12936 (movdf_hardfloat32): Likewise.
12937 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
12938 (movdf_softfloat32): Likewise.
12939 (movdf_hardfloat64_mfpgpr): Likewise.
12940 (movdf_hardfloat64): Likewise.
12941 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
12942 (movdf_softfloat64): Likewise.
12943 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
12944 (reload_<mode>_load): Move to later in the file so they aren't in
12945 the middle of the floating point move insns.
12946 (reload_<mode>_store): Likewise.
12947
12948 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
12949 constraint.
12950
12951 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
12952 constraint if -mdebug=reg.
bd059b26
UB
12953 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
12954 Enable using dd reload support if needed.
c6d5ff83
MM
12955
12956 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
12957 binary and decimal floating point moves in rs6000.md.
12958 (movtd_internal): Likewise.
12959
12960 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
12961 decimal floating point moves.
12962 (movtf): Likewise.
12963 (movtf_internal): Likewise.
12964 (mov<mode>_internal, TDmode/TFmode): Likewise.
12965 (movtf_softfloat): Likewise.
12966 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
12967
12968 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
12969 movdi_internal64, using wg constraint for move direct operations.
12970 (movdi_internal64): Likewise.
12971
12972 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
12973 MODES_TIEABLE_P for selected modes. Print the numerical value of
12974 the various virtual registers. Use GPR/FPR first/last values,
12975 instead of hard coding the register numbers. Print which modes
12976 have reload functions registered.
bd059b26
UB
12977 (rs6000_option_override_internal): If -mdebug=reg, trace the options
12978 settings before/after setting cpu, target and subtarget settings.
12979 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
12980 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM
12981 (rs6000_secondary_reload_fail): Likewise.
12982 (rs6000_secondary_reload_inner): Likewise.
12983
12984 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
12985 macros for first/last GPR and FPR registers.
12986 (LAST_GPR_REGNO): Likewise.
12987 (FIRST_FPR_REGNO): Likewise.
12988 (LAST_FPR_REGNO): Likewise.
12989
12990 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
12991 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
12992 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
12993 (vcond<mode><mode>): Likewise.
12994 (vcondu<mode><mode>): Likewise.
12995 (vector_gtu<mode>): Likewise.
12996 (vector_gte<mode>): Likewise.
12997 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 12998 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM
12999 (ior<mode>3): Likewise.
13000 (and<mode>3): Likewise.
13001 (one_cmpl<mode>2): Likewise.
13002 (nor<mode>3): Likewise.
13003 (andc<mode>3): Likewise.
13004
13005 * config/rs6000/constraints.md (wt constraint): New constraint
13006 that returns VSX_REGS if TImode is allowed in VSX registers.
13007
13008 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13009 constant under VSX.
13010
13011 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13012 similar to TImode, but it is restricted to being in the GPRs.
13013
13014 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13015 TImode to occupy a single VSX register.
13016
13017 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13018 -mvsx-timode for power7/power8.
13019 (power7 cpu): Likewise.
13020 (power8 cpu): Likewise.
13021
13022 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13023 sure that TFmode/TDmode take up two registers if they are ever
13024 allowed in the upper VSX registers.
13025 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13026 registers.
13027 (rs6000_init_hard_regno_mode_ok): Likewise.
13028 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13029 constraint. Print if LRA is turned on.
13030 (rs6000_option_override_internal): Give an error if -mvsx-timode
13031 and VSX is not enabled.
13032 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13033 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13034 to reg+offset addressing. Use PTImode when checking offset
13035 addresses for validity.
13036 (reg_offset_addressing_ok_p): Likewise.
13037 (rs6000_legitimate_offset_address_p): Likewise.
13038 (rs6000_legitimize_address): Likewise.
13039 (rs6000_legitimize_reload_address): Likewise.
13040 (rs6000_legitimate_address_p): Likewise.
13041 (rs6000_eliminate_indexed_memrefs): Likewise.
13042 (rs6000_emit_move): Likewise.
13043 (rs6000_secondary_reload): Likewise.
13044 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13045 reloads to fpr registers to continue to use reg+offset addressing,
13046 but 64-bit reloads to altivec registers need reg+reg addressing.
13047 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13048 it. Treat LO_SUM like a PLUS operation.
13049 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 13050 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM
13051 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13052 registers to share a register with a smaller sized type, since VSX
13053 puts scalars in the upper 64-bits.
13054 (print_operand): Add support for PTImode.
13055 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13056 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13057 registers, but don't have arithmetic support.
13058 (rs6000_memory_move_cost): Add test for VSX.
13059 (rs6000_opt_masks): Add -mvsx-timode.
13060
13061 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13062 for TImode.
13063 (VSs): Likewise.
13064 (VSr): Use wt constraint for TImode.
13065 (VSv): Drop TImode support.
13066 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13067 (vsx_movti_64bit): Likewise.
13068 (vsx_movti_32bit): Likewise.
13069 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13070 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13071 one '?' on the appropriate output constraint. Do not allow TImode
13072 logical operations on 32-bit systems.
13073 (vsx_ior<mode>3): Likewise.
13074 (vsx_xor<mode>3): Likewise.
13075 (vsx_one_cmpl<mode>2): Likewise.
13076 (vsx_nor<mode>3): Likewise.
13077 (vsx_andc<mode>3): Likewise.
13078 (vsx_concat_<mode>): Likewise.
13079 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13080
13081 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13082 OPTION_MASK_VSX_TIMODE.
13083 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13084 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13085
13086 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13087 (TI2 iterator): New iterator for TImode, PTImode.
13088 (wd mode attribute): Add values for vector types.
bd059b26
UB
13089 (movti_string): Replace TI move operations with operations for TImode
13090 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM
13091 (mov<mode>_string, TImode/PTImode): Likewise.
13092 (movti_ppc64): Likewise.
13093 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13094 (TI mode splitters): Likewise.
13095
13096 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13097 constraint.
13098
1fc5eced
MG
130992013-03-20 Marc Glisse <marc.glisse@inria.fr>
13100
13101 PR tree-optimization/56355
13102 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13103 Also handle integers with undefined overflow.
13104
22c4c869
CM
131052013-03-20 Catherine Moore <clm@codesourcery.com>
13106 Maciej W. Rozycki <macro@codesourcery.com>
13107 Tom de Vries <tom@codesourcery.com>
13108 Nathan Sidwell <nathan@codesourcery.com>
13109 Iain Sandoe <iain@codesourcery.com>
13110 Nathan Froyd <froydnj@codesourcery.com>
13111 Chao-ying Fu <fu@mips.com>
13112
13113 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 13114 Document new function attributes.
22c4c869
CM
13115 * doc/invoke.texi (minterlink-compressed, mmicromips,
13116 m14k, m14ke, m14kec): Document new options.
13117 (minterlink-mips16): Update documentation.
13118 * doc/md.texi (ZC, ZD): Document new constraints.
13119 * configure.ac (gcc_cv_as_micromips): Check if linker
13120 supports the .set micromips directive.
13121 * configure: Regenerate.
13122 * config.in: Regenerate.
13123 * config/mips/mips-tables.opt: Regenerate.
13124 * config/mips/micromips.md: New file.
13125 * constraints.md (ZC, ZD): New constraints.
13126 * config/mips/predicates.md (movep_src_register): New predicate.
13127 (movep_src_operand): New predicate.
13128 (non_volatile_mem_operand): New predicate.
13129 * config/mips/mips.md (multimem): New type.
13130 (length): Differentiate between 17-bit and 18-bit branch offsets.
13131 (MOVEP1, MOVEP2): New mode iterator.
33159866 13132 (mov_<load>l): Use ZC constraint.
22c4c869
CM
13133 (mov_<load>r): Likewise.
13134 (mov_<store>l): Likewise.
13135 (mov_<store>r): Likewise.
13136 (*branch_equality<mode>_inverted): Add microMIPS support.
13137 (*branch_equality<mode>): Likewise.
13138 (*jump_absolute): Likewise.
13139 (indirect_jump_<mode>): Likewise.
13140 (tablejump_<mode>): Likewise.
13141 (<optab>_internal): Likewise.
13142 (sibcall_internal): Likewise.
13143 (sibcall_value_internal): Likewise.
13144 (prefetch): Use constraint ZD.
13145 * config/mips/mips.opt (minterlink-compressed): New option.
13146 (minterlink-mips16): Now an alias for minterlink-compressed.
13147 (mmicromips): New option.
13148 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13149 (compare_and_swap_12): Likewise.
13150 (sync_add<mode>): Likewise.
13151 (sync_<optab>_12): Likewise.
13152 (sync_old_<optab>_12): Likewise.
13153 (sync_new_<optab>_12): Likewise.
13154 (sync_nand_12): Likewise.
13155 (sync_old_nand_12): Likewise.
13156 (sync_new_nand_12): Likewise.
13157 (sync_sub<mode>): Likewise.
13158 (sync_old_add<mode>): Likewise.
13159 (sync_old_sub<mode>): Likewise.
13160 (sync_new_add<mode>): Likewise.
13161 (sync_new_sub<mode>): Likewise.
13162 (sync_<optab><mode>): Likewise.
13163 (sync_old_<optab><mode>): Likewise.
13164 (sync_new_<optab><mode>): Likewise.
13165 (sync_nand<mode>): Likewise.
13166 (sync_old_nand<mode>): Likewise.
13167 (sync_new_nand<mode>): Likewise.
13168 (sync_lock_test_and_set<mode>): Likewise.
13169 (test_and_set_12): Likewise.
13170 (atomic_compare_and_swap<mode>): Likewise.
13171 (atomic_exchange<mode>_llsc): Likewise.
13172 (atomic_fetch_add<mode>_llsc): Likewise.
13173 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13174 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13175 (umips_save_restore_pattern_p): Likewise.
13176 (umips_load_store_pair_p): Likewise.
13177 (umips_output_load_store_pair): Likewise.
13178 (umips_movep_target_p): Likewise.
13179 (umips_12bit_offset_address_p): Likewise.
13180 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13181 (mips_base_mips16): Rename this...
13182 (mips_base_compression_flags): ...to this. Update all uses.
13183 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13184 (mips_mips16_decl_p): Delete.
13185 (mips_nomips16_decl_p): Delete.
33159866 13186 (mips_get_compress_on_flags): New function.
22c4c869
CM
13187 (mips_get_compress_off_flags): New function.
13188 (mips_get_compress_mode): New function.
13189 (mips_get_compress_on_name): New function.
13190 (mips_get_compress_off_name): New function.
13191 (mips_insert_attributes): Support multiple compression types.
13192 (mips_merge_decl_attributes): Likewise.
13193 (umips_12bit_offset_address_p): New function.
13194 (mips_start_function_definition): Emit .set micromips directive.
13195 (mips_call_may_need_jalx_p): New function.
13196 (mips_function_ok_for_sibcall): Add microMIPS support.
13197 (mips_print_operand_punctuation): Support short delay slots and
13198 compact jumps.
13199 (umips_swm_mask, umips_swm_encoding): New.
13200 (umips_build_save_restore): New function.
13201 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13202 (was_mips16_p): Remove.
13203 (old_compression_mode): New.
13204 (mips_set_compression_mode): New function.
13205 (mips_set_current_function): Add microMIPS support.
13206 (mips_option_override): Likewise.
13207 (umips_save_restore_pattern_p): New function.
13208 (umips_output_save_restore): New function.
13209 (umips_load_store_pair_p_1): New function.
13210 (umips_load_store_pair_p): New function.
13211 (umips_output_load_store_pair_1): New function.
13212 (umips_output_load_store_pair): New function.
13213 (umips_movep_target_p) New function.
13214 (mips_prepare_pch_save): Add microMIPS support.
13215 * config/mips/mips.h (TARGET_COMPRESSION): New.
13216 (TARGET_CPU_CPP_BUILTINS): Update macro
13217 to use new compression flags and to support microMIPS.
13218 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13219 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13220 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13221 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13222 (ASM_SPEC): Support mmicromips and mno-micromips.
13223 (M16STORE_REG_P): New macro.
13224 (MIPS_CALL): Support TARGET_MICROMIPS.
13225 (MICROMIPS_J): New macro.
13226 (mips_base_mips16): Rename this...
13227 (mips_base_compression_flags): ...to this.
13228 (UMIPS_12BIT_OFFSET_P): New macro.
13229 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13230 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB
132312013-03-20 Richard Biener <rguenther@suse.de>
13232
13233 PR tree-optimization/56661
13234 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13235 the result does not have to be distinct.
13236
54714c68
RB
132372013-03-20 Richard Biener <rguenther@suse.de>
13238
13239 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13240 remap_gimple_op_r.
13241
cca1130d
BS
132422013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13243 Steven Bosscher <steven@gcc.gnu.org>
13244
13245 PR rtl-optimization/56605
13246 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13247
2a930377
UB
132482013-03-20 Uros Bizjak <ubizjak@gmail.com>
13249
13250 PR bootstrap/56656
13251 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13252 that require movd instead of movq.
13253
d6d305fe
RB
132542013-03-20 Richard Biener <rguenther@suse.de>
13255
13256 * tree-ssa-structalias.c (struct variable_info): Add pointer
13257 to the first field of an aggregate with sub-vars. Make
13258 this and the pointer to the next subfield its ID.
13259 (vi_next): New function.
13260 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13261 storedanything_id, integer_id): Increment by one.
13262 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13263 get_call_clobber_vi): Adjust.
13264 (solution_set_expand): Simplify and speedup.
13265 (solution_set_add): Inline into ...
13266 (set_union_with_increment): ... this. Adjust accordingly.
13267 (do_sd_constraint): Likewise.
13268 (do_ds_constraint): Likewise.
13269 (do_complex_constraint): Simplify.
13270 (build_pred_graph): Adjust.
13271 (solve_graph): Likewise. Simplify and speedup.
13272 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13273 get_constraint_for_component_ref, get_constraint_for_1,
13274 first_vi_for_offset, first_or_preceding_vi_for_offset,
13275 create_function_info_for, create_variable_info_for_1,
13276 create_variable_info_for, intra_create_variable_infos): Adjust.
13277 (init_base_vars): Push NULL for ID zero.
13278 (compute_points_to_sets): Adjust.
13279
a271b42d
RB
132802013-03-20 Richard Biener <rguenther@suse.de>
13281
13282 * cfgloop.c (verify_loop_structure): Streamline and avoid
13283 ICEing on corrupt loop tree.
13284 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13285 loop tree.
13286
ebd65954
RB
132872013-03-20 Richard Biener <rguenther@suse.de>
13288
13289 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13290 check whether an SSA update is needed.
13291
4547b7ee
RS
132922013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13293
13294 * config/mips/constraints.md (T): Rename to...
13295 (Yf): ...this.
13296 (U): Rename to...
13297 (Yd): ...this.
13298 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13299 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13300
0b8cdc58
IB
133012013-03-19 Ian Bolton <ian.bolton@arm.com>
13302
13303 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13304 (*subsi3_carryin_uxtw): Likewise.
13305
5977a10d
IB
133062013-03-19 Ian Bolton <ian.bolton@arm.com>
13307
13308 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13309 (*rorsi3_insn_uxtw): Likewise.
13310
bd83ff2c
IB
133112013-03-19 Ian Bolton <ian.bolton@arm.com>
13312
13313 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
13314 (*extrsi5_insn_uxtw): Likewise.
13315
ba65123c
RB
133162013-03-19 Richard Biener <rguenther@suse.de>
13317
13318 PR tree-optimization/56273
13319 * passes.c (init_optimization_passes): Move second VRP after DOM.
13320
2eac0476
UB
133212013-03-19 Uros Bizjak <ubizjak@gmail.com>
13322
13323 * config/i386/i386.md (*movti_internal): Merge from
13324 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
13325 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
13326 nox64 isa attributes.
13327
6e55eda7
RB
133282013-03-18 Richard Biener <rguenther@suse.de>
13329
13330 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
13331 (unite): Likewise.
13332 (merge_node_constraints): Likewise.
13333 (build_succ_graph): Likewise.
13334 (valid_graph_edge): Inline into single caller.
13335 (unify_nodes): Likewise. Use bitmap_set_bit return value
13336 and cache varinfo.
13337 (scc_visit): Fix formatting and variable use.
13338 (do_sd_constraint): Use gcc_checking_assert.
13339 (do_ds_constraint): Likewise.
13340 (do_complex_constraint): Likewise.
13341 (condense_visit): Likewise. Cleanup.
13342 (dump_pred_graph): New function.
13343 (perform_var_substitution): Dump the pred-graph before
13344 variable substitution.
13345 (find_equivalent_node): Use gcc_checking_assert.
13346 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
13347
4bdd44c4
RB
133482013-03-18 Richard Biener <rguenther@suse.de>
13349
13350 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13351 Remove cond_expr_stmt_list argument and do not gimplify the
13352 built expression.
13353 (vect_loop_versioning): Adjust.
13354 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
13355 Cleanup to use less temporaries.
13356 (vect_create_data_ref_ptr): Cleanup.
13357
38c56a5b
JJ
133582013-03-18 Jakub Jelinek <jakub@redhat.com>
13359
13360 PR tree-optimization/56635
13361 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
13362 require types_compatible_p types.
13363
20b2e6a0
NC
133642013-03-18 Nick Clifton <nickc@redhat.com>
13365
a6178a25
NC
13366 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
13367 spurious backslash.
13368
20b2e6a0
NC
13369 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
13370 Add missing line to comment describing function.
13371
92608d0e
RB
133722013-03-18 Richard Biener <rguenther@suse.de>
13373
13374 PR tree-optimization/56210
13375 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
13376 Handle string / character search functions.
13377 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
13378
31348d52
RB
133792013-03-18 Richard Biener <rguenther@suse.de>
13380
13381 PR middle-end/56483
13382 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
13383 and implement properly.
13384 * gimple.h (gimple_cond_single_var_p): Remove.
13385
fcac74a1
RB
133862013-03-18 Richard Biener <rguenther@suse.de>
13387
13388 * tree-data-ref.h (find_data_references_in_loop): Declare.
13389 * tree-data-ref.c (get_references_in_stmt): Use a stack
13390 vector pre-allocated in the callers.
13391 (find_data_references_in_stmt): Adjust.
13392 (graphite_find_data_references_in_stmt): Likewise.
13393 (create_rdg_vertices): Likewise.
13394 (find_data_references_in_loop): Export.
13395 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
13396 Compute dependences here...
13397 (vect_analyze_data_refs): ...not here. When we encounter
13398 a non-vectorizable data reference in basic-block vectorization
13399 truncate the data reference vector. Do not bother to
13400 fixup data-dependence information for gather loads.
13401 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
13402 of data references, as reported.
13403
0d5a1b56
RB
134042013-03-18 Richard Biener <rguenther@suse.de>
13405
13406 PR tree-optimization/3713
13407 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
13408 has_constants and expr.
13409 (stmt_has_constants): Properly valueize SSA names when deciding
13410 whether the stmt has constants.
13411
789c34e3
RB
134122013-03-18 Richard Biener <rguenther@suse.de>
13413
13414 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
13415 whole function when there is nothing to do.
13416 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
13417 * tree-vectorizer.c (vectorize_loops): Update virtual and
13418 loop-closed SSA once.
13419 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
13420
076b4605
RB
134212013-03-18 Richard Biener <rguenther@suse.de>
13422
13423 PR middle-end/56113
13424 * domwalk.c (bb_postorder): New global static.
13425 (cmp_bb_postorder): New function.
13426 (walk_dominator_tree): Replace scheme imposing an order for
13427 visiting dominator sons by one sorting them at the time they
13428 are pushed on the stack.
13429
bdb01696
RB
134302013-03-18 Richard Biener <rguenther@suse.de>
13431
13432 PR tree-optimization/39326
13433 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
13434 (struct mem_ref): Replace mem member with ao_ref typed member.
13435 (MEM_ANALYZABLE): Adjust.
13436 (memref_eq): Likewise.
13437 (mem_ref_alloc): Likewise.
13438 (gather_mem_refs_stmt): Likewise.
13439 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
13440 (execute_sm_if_changed_flag_set): Adjust.
13441 (execute_sm): Likewise.
13442 (ref_always_accessed_p): Likewise.
13443 (refs_independent_p): Likewise.
13444 (can_sm_ref_p): Likewise.
13445
12d2dc5e
JJ
134462013-03-18 Jakub Jelinek <jakub@redhat.com>
13447
13448 PR c/56566
13449 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
13450 return 1 even for !unsignedp.
13451
286fb677
UB
134522013-03-17 Uros Bizjak <ubizjak@gmail.com>
13453
13454 * config/i386/i386.md (isa): Add x64 and nox64.
13455 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
13456 (*pushtf): Enable *roF alternative for x64 isa only.
13457 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
13458 mode attribute of integer alternatives to DImode for TARGET_64BIT.
13459 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
13460 (*movtf_internal): Merge from *movtf_internal_rex64 and
13461 *movtf_internal_sse. Use x64 and nox64 isa attributes.
13462 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
13463 nox64 isa attributes.
13464 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
13465 nox64 isa attributes.
13466 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
13467
88b97037
UB
134682013-03-17 Uros Bizjak <ubizjak@gmail.com>
13469
13470 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13471
9f4f1735
JJ
134722013-03-17 Jakub Jelinek <jakub@redhat.com>
13473
13474 PR target/56640
13475 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
13476 class names. Remove trailing comma after "ALL_REGS".
13477
48f4a6fa
JH
134782013-03-16 Jan Hubicka <jh@suse.cz>
13479
13480 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
13481 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
13482 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
13483 of cgraph_get_create_node.
13484 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
13485
98e81123
JM
134862013-03-16 Jason Merrill <jason@redhat.com>
13487
13488 PR debug/49090
13489 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
13490 with DW_AT_default_value.
13491
8a298c04
JJ
134922013-03-16 Jakub Jelinek <jakub@redhat.com>
13493
13494 * BASE-VER: Set to 4.9.0.
13495
4323afa0
AK
134962013-03-14 Andi Kleen <ak@linux.intel.com>
13497
13498 PR target/56619
13499 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
13500 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
13501 Document _x* TSX intrinsics.
13502
b3c0d469
JJ
135032013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
13504 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI
13505
13506 * configure.ac: Add MicroBlaze TLS support detection.
13507 * configure: Regenerate.
88b97037
UB
13508 * config/microblaze/microblaze-protos.h
13509 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
13510 symbol_mentioned_p, label_mentioned_p): Add prototypes.
13511 * config/microblaze/microblaze.c (microblaze_address_type): Add
13512 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI
13513 (microblaze_address_info): Add tls_reloc.
13514 (TARGET_HAVE_TLS): Define.
13515 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
13516 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB
13517 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
13518 load_tls_operand, microblaze_call_tls_get_addr,
13519 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI
13520 (microblaze_classify_unspec): Handle UNSPEC_TLS.
13521 (get_base_reg): Use microblaze_tls_symbol_p.
13522 (microblaze_classify_address): Handle TLS.
88b97037
UB
13523 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
13524 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI
13525 (microblaze_legitimize_address): Handle TLS.
13526 (microblaze_address_insns): Handle ADDRESS_TLS.
13527 (pic_address_needs_scratch): Handle TLS.
13528 (print_operand_address): Handle TLS.
13529 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
13530 (microblaze_expand_move): Handle TLS.
88b97037
UB
13531 (microblaze_legitimate_constant_p): Check
13532 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI
13533 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
13534 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
13535 (PIC_OFFSET_TABLE_REGNUM): Set.
13536 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
13537 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
13538 (addsi3, movsi_internal2, movdf_internal): Update constraints
13539 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB
13540 (move_operand): Redefine as move_src_operand,
13541 check microblaze_tls_referenced_p.
8cc9a5a5 13542
d803a491
IB
135432013-03-14 Ian Bolton <ian.bolton@arm.com>
13544
13545 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
13546 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
13547
a8504f22
IB
135482013-03-14 Ian Bolton <ian.bolton@arm.com>
13549
13550 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
13551 CC mode for AND.
13552
df2dfaea
JJ
135532013-03-14 Jakub Jelinek <jakub@redhat.com>
13554
fbd28bc3
JJ
13555 PR tree-optimization/53265
13556 * common.opt (Waggressive-loop-optimizations): New option.
13557 * tree-ssa-loop-niter.c: Include tree-pass.h.
13558 (do_warn_aggressive_loop_optimizations): New function.
13559 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
13560 if number_of_latch_executions returned constant.
13561 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
13562 early. If number_of_latch_executions returned constant, set
13563 nb_iterations_upper_bound back to it.
13564 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
13565 field.
13566 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
13567 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
13568
df2dfaea
JJ
13569 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
13570 (MULTILIB_OSDIRNAMES): Set.
13571 * genmultilib: If defaultosdirname doesn't start with :: , set
13572 defaultosdirname2 instead, clear it and emit two . multilib_raw
13573 entries instead of just one.
13574
ee0d2b68
KK
135752013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
13576
13577 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13578 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
13579 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13580 (SUBTARGET_OVERRIDE_OPTIONS): New.
13581
decc676e
OE
135822013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
13583
13584 PR target/49880
13585 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
13586 (musermode): Convert to Var(TARGET_USERMODE).
13587 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
13588 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
13589 * config/sh/sh.c (sh_option_override): Use
13590 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
13591 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
13592 condition.
13593 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
13594 TARGET_SH4.
13595 (udivsi3_i4_single, divsi3_i4_single): Use
13596 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
13597
f4b719c7
DK
135982013-03-13 Dave Korn <dave.korn.cygwin@....>
13599
13600 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
13601 default setting.
13602
c40eced0
RB
136032013-03-13 Richard Biener <rguenther@suse.de>
13604
13605 PR tree-optimization/56608
13606 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
13607 calls when vectorizing basic-blocks.
13608
1bfa5973
JJ
136092013-03-13 Jakub Jelinek <jakub@redhat.com>
13610
13611 PR plugins/45078
13612 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
13613 tm_file.
13614
6fcf5434
JJ
136152013-03-12 Jakub Jelinek <jakub@redhat.com>
13616
13617 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
13618
4f38fa8c
JH
136192013-03-11 Jan Hubicka <jh@suse.cz>
13620
13621 PR lto/56557
13622 * lto-streamer-out.c (output_symbol_p): Skip references from
13623 constructors of external variables.
13624
c5c5ba89
JH
136252013-03-11 Jan Hubicka <jh@suse.cz>
13626
13627 PR middle-end/56571
13628 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
13629 from pseudos.
13630 * emit-rtl.c (verify_rtx_sharing): Likewise.
13631 (copy_insn_1): Likewise.
13632 * rtl.c (copy_rtx): Likewise.
13633
c2a939b1
GJL
136342013-03-11 Georg-Johann Lay <avr@gjlay.de>
13635
13636 PR target/56591
13637 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
13638 output_operand_lossage message.
88b97037 13639
3c3279fb
RE
136402013-03-11 Richard Earnshaw <rearnsha@arm.com>
13641
13642 PR target/56470
13643 * arm.c (shift_op): Validate RTL pattern on the fly.
13644 (arm_print_operand, case 'S'): Don't use shift_operator to validate
13645 the RTL.
13646
aef5ef9d
JDA
136472013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13648
13649 PR target/56347
13650 * config/pa/pa.md (call_value): Check for calls to powf and direct to
13651 new call patterns that clobber %fr12.
13652 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
13653 split and postreload patterns.
13654 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
13655 registers %fr12 and %fr12R as call used.
13656
456610d3
SB
136572013-03-09 Steven Bosscher <steven@gcc.gnu.org>
13658
13659 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
13660 (canon_address, record_store, replace_read, check_mem_read_rtx,
13661 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
13662 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
13663 rest_of_handle_dse): Likewise.
13664
4b1baac8
RS
136652013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
13666
13667 PR middle-end/56524
13668 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
13669 Add base_optabs.
13670 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
13671 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
13672 (save_optabs_if_changed): Replace with...
13673 (init_tree_optimization_optabs): ...this.
13674 * optabs.c (save_optabs_if_changed): Rename to...
13675 (init_tree_optimization_optabs): ...this. Take the optimization node
13676 as argument. Do nothing if the base optabs are already correct.
13677 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
13678 to recompute optabs.
13679 * function.h (function): Remove optabs field.
13680 * function.c (invoke_set_current_function_hook): Call
13681 init_tree_optimization_optabs. Use the result to initialize
13682 this_fn_optabs.
13683
b7a78683
AH
136842013-02-27 Aldy Hernandez <aldyh@redhat.com>
13685
13686 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
13687 if GTMA_HAS_NO_INSTRUMENTATION.
13688 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
13689 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
13690 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
13691 * gimple-pretty-print.c (dump_gimple_transaction): Handle
13692 GTMA_HAS_NO_INSTRUMENTATION.
13693
6384c29b
JJ
136942013-03-08 Jakub Jelinek <jakub@redhat.com>
13695
13696 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
13697 libasan_preinit.o.
13698
ecd4f20a
MP
136992013-03-08 Marek Polacek <polacek@redhat.com>
13700 Jakub Jelinek <jakub@redhat.com>
13701
13702 PR tree-optimization/56478
13703 * predict.c (is_comparison_with_loop_invariant_p): Change the
13704 type of loop_step to tree.
13705 (predict_loops): Adjust.
13706 (predict_iv_comparison): Perform the computations on double_ints.
13707
64366d35
RB
137082013-03-08 Richard Biener <rguenther@suse.de>
13709
13710 PR tree-optimization/56570
13711 * tree-cfg.c (verify_expr_location_1): Verify locations for
13712 DECL_DEBUG_EXPR.
13713 * tree-sra.c (create_access_replacement): Strip locations
13714 from DECL_DEBUG_EXPRs.
13715
a9d5a059
RB
137162013-03-08 Richard Biener <rguenther@suse.de>
13717
13718 * tree-inline.c (expand_call_inline): Do not associate
13719 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
13720 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
13721
b7aa4e9a
RB
137222013-03-08 Richard Biener <rguenther@suse.de>
13723
13724 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
13725 or block changes with -Og. Fix for location / block encoding
13726 changes and PHI arguments with locations.
13727
c4c2f9fa
SB
137282013-03-07 Steven Bosscher <steven@gcc.gnu.org>
13729
13730 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
13731 for all counters.
13732 (struct output_info): Likewise.
13733 (register_overhead): Remove bad gcc_assert.
13734 (bitmap_find_bit): If there is only a single bitmap element, do not
13735 count a miss as a search.
13736 (print_statistics): Update for counter type changes.
13737 (dump_bitmap_statistics): Likewise. Print headers such that they
13738 are properly lined up with the printed counters.
13739
5bf6606a
JJ
137402013-03-07 Jakub Jelinek <jakub@redhat.com>
13741
13742 PR tree-optimization/56559
13743 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
13744 check that it has only a single use.
13745
2c653d46
RB
137462013-03-07 Richard Biener <rguenther@suse.de>
13747
13748 * doc/invoke.texi (fwhole-program): Discourage use in combination
13749 with -flto.
13750
a72d8780
JJ
137512013-03-06 Jakub Jelinek <jakub@redhat.com>
13752
01a454df
JJ
13753 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
13754
c1781047
JJ
13755 PR tree-optimization/56539
13756 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
13757 instead of GSI_CONTINUE_LINKING as last argument to
13758 force_gimple_operand_gsi. Adjust function comment.
13759
9772c47a
JJ
13760 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
13761 aarch64-cores.def.
13762
a72d8780
JJ
13763 PR middle-end/56548
13764 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
13765 promoted mode, convert the result back to the original mode.
13766
fa5556de
RB
137672013-03-06 Richard Biener <rguenther@suse.de>
13768
13769 PR middle-end/56294
13770 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
13771 (insert_updated_phi_nodes_compare_uids): New function.
13772 (update_ssa): Sort symbols_to_rename after UID before
13773 traversing it to insert PHI nodes.
13774
010403d1
RB
137752013-03-06 Richard Biener <rguenther@suse.de>
13776
13777 PR middle-end/50494
13778 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
13779 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
13780
13781 Revert
13782 2013-02-13 Richard Biener <rguenther@suse.de>
13783
13784 PR lto/50494
13785 * varasm.c (output_constant_def_1): Get the decl representing
13786 the constant as argument.
13787 (output_constant_def): Wrap output_constant_def_1.
13788 (make_decl_rtl): Use output_constant_def_1 with the decl
13789 representing the constant.
13790 (build_constant_desc): Optionally re-use a decl already
13791 representing the constant.
13792 (tree_output_constant_def): Adjust.
13793
3c27ce4c
JY
137942013-03-06 Joey Ye <joey.ye@arm.com>
13795
13796 PR lto/50293
13797 * gcc.c (convert_white_space): New function.
13798 (main): Handles white space in function name.
13799
8f6d1c86
OE
138002013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
13801
13802 PR target/56529
13803 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
13804 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
13805 to SH_DIV_CALL_TABLE for TARGET_SH2.
13806 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
13807 list.
13808 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
13809 call-table options.
13810
798a209f
SA
138112013-03-05 Sterling Augustine <saugustine@google.com>
13812 Cary Coutant <ccoutant@google.com>
13813
13814 PR debug/55364
13815 * dwarf2out.c (resolve_addr): Don't call
13816 remove_loc_list_addr_table_entries a second time for the same
13817 expression.
13818
6cfa417f 138192013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 13820
6cfa417f
JJ
13821 PR debug/56510
13822 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
13823 (avoid_complex_debug_insns): New function.
13824 (expand_debug_locations): Call it.
13825
13826 PR rtl-optimization/56484
13827 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
13828 lifetimes of hard registers on small register class machines.
13829
138302013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
13831
13832 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 13833 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 13834 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH
13835 fast_interrupt.
13836 (microblaze_fast_interrupt_function_p): New function.
13837 (microblaze_is_interrupt_handler): Rename to
13838 microblaze_is_interrupt_variant and add fast_interrupt check.
13839 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
13840 (save_restore_insns): Likewise.
13841 (compute_frame_size): Likewise.
13842 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
13843 (microblaze_globalize_label): Likewise.
6cfa417f
JJ
13844 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
13845 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH
13846 microblaze_is_interrupt_variant.
13847
731aefac
KT
138482013-03-05 Kai Tietz <ktietz@redhat.com>
13849
13850 * sdbout.c (sdbout_one_type): Switch to current function's section
13851 supporting cold/hot.
13852
a72d8780 138532013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH
13854
13855 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
13856 -mxl-reorder.
13857
334e71e8
JJ
138582013-03-05 Jakub Jelinek <jakub@redhat.com>
13859
0b50e654
JJ
13860 PR middle-end/56461
13861 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
13862 if VALGRIND_GET_VBITS is defined, temporarily make object
13863 memory all defined, and restore previous valgrind addressability
13864 and definability afterwards. Free this_object at the end.
13865
4ccf8f43
JJ
13866 PR middle-end/56461
13867 * lra.c (lra): Call lra_clear_live_ranges if live_p,
13868 right before calling lra_create_live_ranges, also call it
13869 when clearing live_p. Only call lra_clear_live_ranges
13870 at the end if live_p.
13871
334e71e8
JJ
13872 PR middle-end/56461
13873 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
13874
9ca966ca
RB
138752013-03-05 Richard Biener <rguenther@suse.de>
13876
13877 PR tree-optimization/56521
13878 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
13879 value-id.
13880
d273b176
SB
138812013-03-05 Steven Bosscher <steven@gcc.gnu.org>
13882
13883 PR c++/55135
13884 * except.h (remove_unreachable_eh_regions): New prototype.
13885 * except.c (remove_eh_handler_splicer): New function, split out
13886 of remove_eh_handler.
13887 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
13888 warning about running it on many EH regions one at a time.
13889 (remove_unreachable_eh_regions_worker): New function, walk the
13890 EH tree in depth-first order and remove non-marked regions.
13891 (remove_unreachable_eh_regions): New function.
13892 * tree-eh.c (mark_reachable_handlers): New function, split out
13893 from remove_unreachable_handlers.
13894 (remove_unreachable_handlers): Use mark_reachable_handlers and
13895 remove_unreachable_eh_regions.
13896 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
13897 and remove_unreachable_eh_regions.
13898
5e7f1aef
RB
138992013-03-05 Richard Biener <rguenther@suse.de>
13900
13901 PR middle-end/56525
13902 * loop-init.c (fix_loop_structure): Remove loops in two stages,
13903 not freeing them until the end.
13904
f276b762
AK
139052013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13906
13907 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
13908
12a43ab8
RB
139092013-03-05 Richard Biener <rguenther@suse.de>
13910
13911 PR tree-optimization/56270
13912 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
13913 of loads after scheduling an SLP instance.
13914
93675444
JJ
139152013-03-05 Jakub Jelinek <jakub@redhat.com>
13916
db4138e3
JJ
13917 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
13918 tic6x.exp.
13919 (check_gcc_parallelize): Run guality.exp as a separate job from
13920 vect.exp with unsorted.exp and $(dg_target_exps) separately from
13921 struct-layout-1.exp with stackalign.exp.
13922
dd3d1ec0
JJ
13923 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
13924
b8d381a3
JJ
13925 PR middle-end/56461
13926 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
13927 load_index sbitmap even if some bit in it isn't set.
13928
b4f9786b
JJ
13929 PR middle-end/56461
13930 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
13931 (discover_iteration_bound_by_body_walk): Change queues to
13932 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
13933 spelling in comment. Call safe_push on queues[bound_index] directly.
13934 Release queues[queue_index] in every iteration unconditionally.
13935 Release bounds vector.
13936
93675444
JJ
13937 PR middle-end/56461
13938 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
13939 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
13940 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
13941 inner_phis vector.
13942
3e492e9c
RB
139432013-03-05 Richard Biener <rguenther@suse.de>
13944
13945 PR lto/56515
13946 * tree-inline.c (remap_blocks_to_null): New function.
13947 (expand_call_inline): When expanding a call stmt without
13948 an associated block inline remap all callee blocks to NULL.
13949
a5d1569a
JJ
139502013-03-05 Jakub Jelinek <jakub@redhat.com>
13951
86efb5cd
JJ
13952 PR rtl-optimization/56494
13953 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
13954 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
13955 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
13956
85f5dbea
JJ
13957 PR middle-end/56461
13958 * sel-sched-ir.c (free_sched_pools): Release
13959 succs_info_pool.stack[succs_info_pool.max_top] vectors too
13960 if succs_info_pool.max_top isn't -1.
13961
a5d1569a
JJ
13962 PR bootstrap/56509
13963 * opts.c (opts_obstack, opts_concat): Moved to...
13964 * opts-common.c (opts_obstack, opts_concat): ... here.
13965
4432aa6c
JJ
139662013-03-04 Jakub Jelinek <jakub@redhat.com>
13967
13968 PR middle-end/56461
13969 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
13970
f3a1fb91
MJ
139712013-03-04 Martin Jambor <mjambor@suse.cz>
13972
13973 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
13974 all appropriate places.
13975
22110e6c
EB
139762013-01-04 Eric Botcazou <ebotcazou@adacore.com>
13977
13978 PR tree-optimization/56424
13979 * ipa-split.c (split_function): Do not set the RSO flag if result is
13980 not by reference and its type is a register type.
13981
a72d8780 139822013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 13983
88b97037 13984 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 13985 (microblaze_legitimate_pic_operand): Likewise
88b97037 13986 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH
13987 new function microblaze_legitimate_pic_operand
13988 * config/microblaze/microblaze-protos.h
13989 (microblaze_legitimate_pic_operand): Declare.
13990
a72d8780 139912013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 13992
a72d8780 13993 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 13994 New predicate for supported rtx code types.
a72d8780 13995 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI
13996 call_insn_simple_operand predicate.
13997
541d9ac8
JJ
139982013-03-04 Jakub Jelinek <jakub@redhat.com>
13999
5eb010bc
JJ
14000 PR middle-end/56461
14001 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14002 partitions.ordered_remove.
14003
30862efc
JJ
14004 PR middle-end/56461
14005 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14006 vec_oprnds0.create (1) for modifier == NONE.
14007
8930f723
JJ
14008 PR middle-end/56461
14009 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14010 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14011 vec_oprnds1 right before pushing anything to it for
14012 scalar_shift_arg.
14013
541d9ac8
JJ
14014 PR middle-end/56461
14015 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14016 set nbbs to 0 instead of having separate code path.
14017 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14018 instead of false as last argument if returning NULL.
14019
7aa7f2e3
SL
140202013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14021
14022 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14023 the attribute is now called "target" instead of "option".
14024 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14025 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14026 attribute/pragma name for TARGET_OPTION_VALID_P and
14027 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14028 * doc/tm.texi: Regenerated.
14029
8930f723 140302013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 14031
8930f723 14032 * config/microblaze/microblaze.c:
cb8a1637 14033 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ
14034 * config/microblaze/microblaze.h: Add -mxl-reorder to
14035 DRIVER_SELF_SPECS.
14036 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14037 instructions emitted if TARGET_REORDER.
88b97037
UB
14038 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14039 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 14040 separately.
cb8a1637 14041
4dc1d68c
XDL
140422013-03-01 Xinliang David Li <davidxl@google.com>
14043
14044 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14045 walk length.
14046
689eaba3
JJ
140472013-03-01 Jakub Jelinek <jakub@redhat.com>
14048
9d676bf6
JJ
14049 PR middle-end/56461
14050 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14051 vector even when returning true. Fix up function comment formatting.
14052
caff7edf
JJ
14053 PR middle-end/56461
14054 * ira-build.c (ira_loop_nodes_count): New variable.
14055 (create_loop_tree_nodes): Initialize it.
14056 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14057
b6b9227d
JJ
14058 PR middle-end/56461
14059 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14060 method on dr_chain and result_chain.
14061 * tree-vect-stmts.c (vectorizable_store): Only call
14062 result_chain.create if j == 0.
14063
689eaba3
JJ
14064 PR middle-end/56461
14065 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14066 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14067 before overwriting it.
14068
40bf31ed
TB
140692013-03-01 Tobias Burnus <burnus@net-b.de>
14070
14071 * doc/extended.texi (C Extensions): Change order in @menu
14072 to match @node.
14073 (Other MIPS Built-in Functions): Move last MIPS entry before
14074 "picoChip Built-in Functions".
14075 (SH Built-in Functions): Move after RX Built-in Functions.
14076 * doc/gcc.texi (Introduction): Change order in @menu
14077 to match @node.
14078 * doc/md.texi (Constraints): Ditto.
14079 * gty.texi (Type Information): Ditto.
14080 (User-provided marking routines for template types): Make
14081 subsection.
14082 * doc/invoke.texi (AArch64 Options): Move before
14083 "Adapteva Epiphany Options".
14084
e664c61c
KS
140852013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14086 Jakub Jelinek <jakub@redhat.com>
14087
14088 PR sanitizer/56454
14089 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14090 no_address_safety_analysis attribute.
14091 * doc/extend.texi (no_address_safety_attribute): Rename to
14092 no_sanitize_address attribute, mention no_address_safety_analysis
14093 attribute as deprecated alias.
14094
37b5ec8f
JJ
140952013-02-28 Jakub Jelinek <jakub@redhat.com>
14096
14097 PR middle-end/56461
14098 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14099 type to vec<vec<tree> > *.
14100 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14101 to be vec<tree> instead of vec<tree> *, set vec_defs
14102 to vNULL and call vec_defs.create (number_of_vects), adjust other
14103 uses of vec_defs.
14104 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14105 vectorizable_condition): Adjust vect_get_slp_defs callers.
14106
ba96cdfb
JG
141072013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14108
14109 * config/aarch64/aarch64.c
14110 (aarch64_float_const_representable): Remove unused variable.
14111
6f549691
JG
141122013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14113
14114 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14115
af55e82d
JG
141162013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14117
14118 * config/aarch64/aarch64-builtins.c
14119 (aarch64_init_simd_builtins): Make static.
14120
1df3f464
JG
141212013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14122
14123 * config/aarch64/aarch64.c
14124 (aarch64_simd_make_constant): Make static.
14125
f8f42513
MJ
141262013-02-28 Martin Jambor <mjambor@suse.cz>
14127
14128 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14129 with no initialization to the RHS of debug statements.
14130
b48b3fc4
MJ
141312013-02-28 Martin Jambor <mjambor@suse.cz>
14132
14133 PR tree-optimization/56294
14134 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14135 Adjust dumping.
14136 (get_access_replacement): Do not call create_access_replacement.
14137 Assert a replacement exists.
14138 (get_repl_default_def_ssa_name): Create the replacement declaration
14139 itself.
14140
c3ae224c
RR
141412013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14142
14143 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14144 final_end_function.
14145
45fa0eef
MP
141462013-02-28 Marek Polacek <polacek@redhat.com>
14147
14148 PR rtl-optimization/56466
14149 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14150 if we're changing a loop.
14151 (peel_loops_completely): Likewise.
14152
502c067d
PC
141532013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14154
14155 PR c++/55813
14156 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14157
ff24afc8
GJL
141582013-02-28 Georg-Johann Lay <avr@gjlay.de>
14159
14160 PR target/56445
14161 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14162 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14163 INTX_FTYPE_FX, FX_FTYPE_INTX.
14164 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14165
c7b1fc1b
GJL
141662013-02-28 Georg-Johann Lay <avr@gjlay.de>
14167
14168 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14169 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14170 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14171 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14172 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14173 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14174 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14175 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14176 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14177 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14178 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14179 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14180 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14181 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14182 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14183 (avrxmega6): Increase max flash segments from 5 to 6.
14184 * config/avr/t-multilib: Regenerate.
14185 * config/avr/avr-tables.opt: Regenerate.
14186 * doc/avr-mmcu.texi: Regenerate.
14187
4a0e3cfe
GJL
141882013-02-28 Georg-Johann Lay <avr@gjlay.de>
14189
14190 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14191 (avr_device_to_arch): Rename to avr_device_to_ld.
14192 (avr_device_to_as): New prototype.
14193 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14194 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14195 * config/avr/driver-avr.c (avr_device_to_as): New.
14196 (avr_device_to_arch): Rename to avr_device_to_ld.
14197
97785e52
JJ
141982013-02-27 Jakub Jelinek <jakub@redhat.com>
14199
3f292312
JJ
14200 PR middle-end/56461
14201 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14202 method on dr_chain and result_chain.
14203
a344216b
JJ
14204 PR middle-end/56461
14205 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14206 pointer_set_destroy on not_executed_last_iteration.
14207
f121ad02 14208 PR middle-end/56461
88b97037 14209 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 14210
307f83a3
JJ
14211 PR middle-end/56461
14212 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14213 FOR_EACH_DEFINED_FUNCTION when freeing state.
14214
e19624ee
JJ
14215 PR middle-end/56461
14216 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14217 pool_free.
14218 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14219 overwriting it.
14220
90e709fd
JJ
14221 PR middle-end/56461
14222 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14223 known_aggs[i].items and release known_aggs vector.
14224
97785e52
JJ
14225 PR middle-end/56461
14226 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14227
142282013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 14229
97785e52
JJ
14230 * config/microblaze/microblaze.c (microblaze_emit_compare):
14231 Use xor for EQ/NE comparisions.
14232 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14233 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 14234
6fa5e0ed
JJ
142352013-02-27 Jakub Jelinek <jakub@redhat.com>
14236
14237 PR middle-end/56461
14238 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14239 vector.
14240 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14241 vec_safe_push, always update *slot.
14242 (redirect_edge_var_map_clear): Use vec_free.
14243 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14244 (free_var_map_entry): Use vec_free.
14245 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14246 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14247
436a956a
AB
142482013-02-27 Andrey Belevantsev <abel@ispras.ru>
14249
14250 PR middle-end/45472
436a956a
AB
14251 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14252 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB
14253 Reorder tests for speculativeness in the logical and operator.
14254
f5c2caca
JJ
142552013-02-27 Jakub Jelinek <jakub@redhat.com>
14256
0fcb564b
JJ
14257 * incpath.c (add_standard_paths): Use reconcat instead of concat
14258 where appropriate and avoid leaking memory.
14259
dc357798
JJ
14260 * opts.h: Include obstack.h.
14261 (opts_concat): New prototype.
14262 (opts_obstack): New declaration.
14263 * opts.c (opts_concat): New function.
14264 (opts_obstack): New variable.
14265 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14266 (finish_options): Use opts_concat instead of concat
14267 and XOBNEWVEC instead of XNEWVEC.
14268 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14269 generate_option): Likewise.
14270 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14271 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14272
f5c2caca
JJ
14273 PR target/56455
14274 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14275 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14276
d0163673
JJ
142772013-02-26 Jakub Jelinek <jakub@redhat.com>
14278
14279 PR middle-end/56461
14280 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14281
b5ad2b8e
JR
142822013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14283
14284 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14285 (arm_block_move_unaligned_straight): Likewise.
14286 (arm_adjust_block_mem): Likewise.
14287
9b639e2c
JR
142882013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14289
14290 PR target/48901
14291 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14292 temp, cond and label.
14293 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14294
e13a0ccb
JR
14295 PR target/52500
14296 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14297 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14298
fbe4f171
JR
14299 PR target/52501
14300 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14301 of prologue/epilogue functions.
14302
ae006543
JR
14303 PR target/52550
14304 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14305 Remove unused variable cfa_offset.
14306 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14307
c81369fa
JR
14308 PR target/54639
14309 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14310 type promotion to unsigned.
14311
f8a8fea7
JR
14312 PR target/54640
14313 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
14314 for HOST_WIDE_INT of 32 bit / same size as int.
14315 (arm_block_move_unaligned_straight): Likewise.
14316 (arm_adjust_block_mem): Likewise.
14317
f8be5169
JR
14318 PR target/54662
14319 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
14320 ALL_CFLAGS.
14321
f1ad3354
MP
143222013-02-26 Marek Polacek <polacek@redhat.com>
14323
14324 PR tree-optimization/56426
88b97037 14325 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 14326
a6af0f42
RB
143272013-02-26 Richard Biener <rguenther@suse.de>
14328
14329 PR target/56444
14330 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
14331 unused variable loops.
14332
aca43c6c
JJ
143332013-02-26 Jakub Jelinek <jakub@redhat.com>
14334
cecbe5d9
JJ
14335 PR tree-optimization/56448
14336 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
14337 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
14338 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
14339 later operands of the references, or even first operand for
14340 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
14341
aca43c6c
JJ
14342 PR tree-optimization/56443
14343 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
14344 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
14345 to type_for_mode langhook.
14346
53e2e141
MT
143472013-02-25 Matt Turner <mattst88@gmail.com>
14348
14349 * doc/invoke.texi: Document r4700.
14350
259ee451
RB
143512013-02-25 Richard Biener <rguenther@suse.de>
14352
14353 PR tree-optimization/56175
14354 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
14355 split out from ...
14356 (simplify_bitwise_binary): ... here. Also guard the conversion
14357 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
14358
6d65e8f1
CM
143592013-02-25 Catherine Moore <clm@codesourcery.com>
14360
14361 Revert:
14362 2013-02-24 Catherine Moore <clm@codesourcery.com>
14363 Maciej W. Rozycki <macro@codesourcery.com>
14364 Tom de Vries <tom@codesourcery.com>
a72d8780 14365 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM
14366 Iain Sandoe <iain@codesourcery.com>
14367 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 14368 Chao-ying Fu <fu@mips.com>
88b97037 14369
6d65e8f1 14370 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 14371 Document new function attributes.
6d65e8f1
CM
14372 * doc/invoke.texi (minterlink-compressed, mmicromips,
14373 m14k, m14ke, m14kec): Document new options.
14374 (minterlink-mips16): Update documentation.
14375 * doc/md.texi (ZC, ZD): Document new constraints.
14376 * configure.ac (gcc_cv_as_micromips): Check if linker
14377 supports the .set micromips directive.
14378 * configure: Regenerate.
14379 * config.in: Regenerate.
14380 * config/mips/mips-tables.opt: Regenerate.
14381 * config/mips/micromips.md: New file.
14382 * constraints.md (ZC, AD): New constraints.
14383 * config/mips/predicates.md (movep_src_register): New predicate.
14384 (movep_src_operand): New predicate.
14385 (non_volatile_mem_operand): New predicate.
14386 * config/mips/mips.md (multimem): New type.
14387 (length): Differentiate between 17-bit and 18-bit branch offsets.
14388 (MOVEP1, MOVEP2): New mode iterator.
14389 (mov_<load>l): Use ZC constraint.
14390 (mov_<load>r): Likewise.
14391 (mov_<store>l): Likewise.
14392 (mov_<store>r): Likewise.
14393 (*branch_equality<mode>_inverted): Add microMIPS support.
14394 (*branch_equality<mode>): Likewise.
14395 (*jump_absolute): Likewise.
14396 (indirect_jump_<mode>): Likewise.
14397 (tablejump_<mode>): Likewise.
14398 (<optab>_internal): Likewise.
14399 (sibcall_internal): Likewise.
14400 (sibcall_value_internal): Likewise.
14401 (prefetch): Use constraint ZD.
14402 * config/mips/mips.opt (minterlink-compressed): New option.
14403 (minterlink-mips16): Now an alias for minterlink-compressed.
14404 (mmicromips): New option.
14405 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14406 (compare_and_swap_12): Likewise.
14407 (sync_add<mode>): Likewise.
14408 (sync_<optab>_12): Likewise.
14409 (sync_old_<optab>_12): Likewise.
14410 (sync_new_<optab>_12): Likewise.
14411 (sync_nand_12): Likewise.
14412 (sync_old_nand_12): Likewise.
14413 (sync_new_nand_12): Likewise.
14414 (sync_sub<mode>): Likewise.
14415 (sync_old_add<mode>): Likewise.
14416 (sync_old_sub<mode>): Likewise.
14417 (sync_new_add<mode>): Likewise.
14418 (sync_new_sub<mode>): Likewise.
14419 (sync_<optab><mode>): Likewise.
14420 (sync_old_<optab><mode>): Likewise.
14421 (sync_new_<optab><mode>): Likewise.
14422 (sync_nand<mode>): Likewise.
14423 (sync_old_nand<mode>): Likewise.
14424 (sync_new_nand<mode>): Likewise.
14425 (sync_lock_test_and_set<mode>): Likewise.
14426 (test_and_set_12): Likewise.
14427 (atomic_compare_and_swap<mode>): Likewise.
14428 (atomic_exchange<mode>_llsc): Likewise.
14429 (atomic_fetch_add<mode>_llsc): Likewise.
14430 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14431 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14432 (umips_save_restore_pattern_p): Likewise.
14433 (umips_load_store_pair_p): Likewise.
14434 (umips_output_load_store_pair): Likewise.
14435 (umips_movep_target_p): Likewise.
14436 (umips_12bit_offset_address_p): Likewise.
14437 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14438 (mips_base_mips16): Rename this...
14439 (mips_base_compression_flags): ...to this. Update all uses.
14440 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14441 (mips_mips16_decl_p): Delete.
14442 (mips_nomips16_decl_p): Delete.
14443 (mips_get_compress_on_flags): New function.
14444 (mips_get_compress_off_flags): New function.
14445 (mips_get_compress_mode): New function.
14446 (mips_get_compress_on_name): New function.
14447 (mips_get_compress_off_name): New function.
14448 (mips_insert_attributes): Support multiple compression types.
14449 (mips_merge_decl_attributes): Likewise.
14450 (umips_12bit_offset_address_p): New function.
14451 (mips_start_function_definition): Emit .set micromips directive.
14452 (mips_call_may_need_jalx_p): New function.
14453 (mips_function_ok_for_sibcall): Add microMIPS support.
14454 (mips_print_operand_punctuation): Support short delay slots and
14455 compact jumps.
14456 (umips_swm_mask, umips_swm_encoding): New.
14457 (umips_build_save_restore): New function.
14458 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14459 (was_mips16_p): Remove.
14460 (old_compression_mode): New.
14461 (mips_set_compression_mode): New function.
14462 (mips_set_current_function): Add microMIPS support.
14463 (mips_option_override): Likewise.
14464 (umips_save_restore_pattern_p): New function.
14465 (umips_output_save_restore): New function.
14466 (umips_load_store_pair_p_1): New function.
14467 (umips_load_store_pair_p): New function.
14468 (umips_output_load_store_pair_1): New function.
14469 (umips_output_load_store_pair): New function.
14470 (umips_movep_target_p) New function.
14471 (mips_prepare_pch_save): Add microMIPS support.
14472 * config/mips/mips.h (TARGET_COMPRESSION): New.
14473 (TARGET_CPU_CPP_BUILTINS): Update macro
14474 to use new compression flags and to support microMIPS.
14475 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14476 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14477 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14478 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14479 (ASM_SPEC): Support mmicromips and mno-micromips.
14480 (M16STORE_REG_P): New macro.
14481 (MIPS_CALL): Support TARGET_MICROMIPS.
14482 (MICROMIPS_J): New macro.
14483 (mips_base_mips16): Rename this...
14484 (mips_base_compression_flags): ...to this.
14485 (UMIPS_12BIT_OFFSET_P): New macro.
14486 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14487 (MULTILIB_DIRNAMES): Likewise.
14488
5e5df392
TV
144892013-02-25 Tom de Vries <tom@codesourcery.com>
14490
14491 PR rtl-optimization/56131
14492 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
14493 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
14494 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
14495
3fdb53c1
TB
144962013-02-25 Tobias Burnus <burnus@net-b.de>
14497
14498 * doc/invoke.texi (-fsanitize=): Move from optimization
14499 to debugging options.
14500
ed358aea
AB
145012013-02-25 Andrey Belevantsev <abel@ispras.ru>
14502
14503 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
14504
f45e9053
AB
145052013-02-25 Andrey Belevantsev <abel@ispras.ru>
14506 Alexander Monakov <amonakov@ispras.ru>
14507
14508 PR middle-end/56077
14509 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
14510 flush pending lists also on non-jumps. Adjust comment.
14511
6941b508
CM
145122013-02-24 Catherine Moore <clm@codesourcery.com>
14513 Maciej W. Rozycki <macro@codesourcery.com>
14514 Tom de Vries <tom@codesourcery.com>
a72d8780 14515 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM
14516 Iain Sandoe <iain@codesourcery.com>
14517 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 14518 Chao-ying Fu <fu@mips.com>
6941b508
CM
14519
14520 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 14521 Document new function attributes.
6941b508
CM
14522 * doc/invoke.texi (minterlink-compressed, mmicromips,
14523 m14k, m14ke, m14kec): Document new options.
14524 (minterlink-mips16): Update documentation.
14525 * doc/md.texi (ZC, ZD): Document new constraints.
14526 * configure.ac (gcc_cv_as_micromips): Check if linker
14527 supports the .set micromips directive.
14528 * configure: Regenerate.
14529 * config.in: Regenerate.
14530 * config/mips/mips-tables.opt: Regenerate.
14531 * config/mips/micromips.md: New file.
14532 * constraints.md (ZC, AD): New constraints.
14533 * config/mips/predicates.md (movep_src_register): New predicate.
14534 (movep_src_operand): New predicate.
14535 (non_volatile_mem_operand): New predicate.
14536 * config/mips/mips.md (multimem): New type.
14537 (length): Differentiate between 17-bit and 18-bit branch offsets.
14538 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 14539 (mov_<load>l): Use ZC constraint.
6941b508
CM
14540 (mov_<load>r): Likewise.
14541 (mov_<store>l): Likewise.
14542 (mov_<store>r): Likewise.
14543 (*branch_equality<mode>_inverted): Add microMIPS support.
14544 (*branch_equality<mode>): Likewise.
14545 (*jump_absolute): Likewise.
14546 (indirect_jump_<mode>): Likewise.
14547 (tablejump_<mode>): Likewise.
14548 (<optab>_internal): Likewise.
14549 (sibcall_internal): Likewise.
14550 (sibcall_value_internal): Likewise.
14551 (prefetch): Use constraint ZD.
14552 * config/mips/mips.opt (minterlink-compressed): New option.
14553 (minterlink-mips16): Now an alias for minterlink-compressed.
14554 (mmicromips): New option.
14555 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14556 (compare_and_swap_12): Likewise.
14557 (sync_add<mode>): Likewise.
14558 (sync_<optab>_12): Likewise.
14559 (sync_old_<optab>_12): Likewise.
14560 (sync_new_<optab>_12): Likewise.
14561 (sync_nand_12): Likewise.
14562 (sync_old_nand_12): Likewise.
14563 (sync_new_nand_12): Likewise.
14564 (sync_sub<mode>): Likewise.
14565 (sync_old_add<mode>): Likewise.
14566 (sync_old_sub<mode>): Likewise.
14567 (sync_new_add<mode>): Likewise.
14568 (sync_new_sub<mode>): Likewise.
14569 (sync_<optab><mode>): Likewise.
14570 (sync_old_<optab><mode>): Likewise.
14571 (sync_new_<optab><mode>): Likewise.
14572 (sync_nand<mode>): Likewise.
14573 (sync_old_nand<mode>): Likewise.
14574 (sync_new_nand<mode>): Likewise.
14575 (sync_lock_test_and_set<mode>): Likewise.
14576 (test_and_set_12): Likewise.
14577 (atomic_compare_and_swap<mode>): Likewise.
14578 (atomic_exchange<mode>_llsc): Likewise.
14579 (atomic_fetch_add<mode>_llsc): Likewise.
14580 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14581 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14582 (umips_save_restore_pattern_p): Likewise.
14583 (umips_load_store_pair_p): Likewise.
14584 (umips_output_load_store_pair): Likewise.
14585 (umips_movep_target_p): Likewise.
14586 (umips_12bit_offset_address_p): Likewise.
14587 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14588 (mips_base_mips16): Rename this...
14589 (mips_base_compression_flags): ...to this. Update all uses.
14590 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14591 (mips_mips16_decl_p): Delete.
14592 (mips_nomips16_decl_p): Delete.
3fdb53c1 14593 (mips_get_compress_on_flags): New function.
6941b508
CM
14594 (mips_get_compress_off_flags): New function.
14595 (mips_get_compress_mode): New function.
14596 (mips_get_compress_on_name): New function.
14597 (mips_get_compress_off_name): New function.
14598 (mips_insert_attributes): Support multiple compression types.
14599 (mips_merge_decl_attributes): Likewise.
14600 (umips_12bit_offset_address_p): New function.
14601 (mips_start_function_definition): Emit .set micromips directive.
14602 (mips_call_may_need_jalx_p): New function.
14603 (mips_function_ok_for_sibcall): Add microMIPS support.
14604 (mips_print_operand_punctuation): Support short delay slots and
14605 compact jumps.
14606 (umips_swm_mask, umips_swm_encoding): New.
14607 (umips_build_save_restore): New function.
14608 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14609 (was_mips16_p): Remove.
14610 (old_compression_mode): New.
14611 (mips_set_compression_mode): New function.
14612 (mips_set_current_function): Add microMIPS support.
14613 (mips_option_override): Likewise.
14614 (umips_save_restore_pattern_p): New function.
14615 (umips_output_save_restore): New function.
14616 (umips_load_store_pair_p_1): New function.
14617 (umips_load_store_pair_p): New function.
14618 (umips_output_load_store_pair_1): New function.
14619 (umips_output_load_store_pair): New function.
14620 (umips_movep_target_p) New function.
14621 (mips_prepare_pch_save): Add microMIPS support.
14622 * config/mips/mips.h (TARGET_COMPRESSION): New.
14623 (TARGET_CPU_CPP_BUILTINS): Update macro
14624 to use new compression flags and to support microMIPS.
14625 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14626 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14627 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14628 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14629 (ASM_SPEC): Support mmicromips and mno-micromips.
14630 (M16STORE_REG_P): New macro.
14631 (MIPS_CALL): Support TARGET_MICROMIPS.
14632 (MICROMIPS_J): New macro.
14633 (mips_base_mips16): Rename this...
14634 (mips_base_compression_flags): ...to this.
14635 (UMIPS_12BIT_OFFSET_P): New macro.
14636 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14637 (MULTILIB_DIRNAMES): Likewise.
14638
cdebbc6d
JJ
146392013-02-24 Jakub Jelinek <jakub@redhat.com>
14640
14641 PR target/52555
14642 * target-globals.c (save_target_globals): For init_reg_sets and
14643 target_reinit remporarily set this_fn_optabs to this_target_optabs.
14644
18c63565
JG
146452013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
14646
14647 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
14648 * config/aarch64/t-aarch64
14649 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
14650
2194f7a2
VM
146512013-02-22 Vladimir Makarov <vmakarov@redhat.com>
14652
14653 PR inline-asm/56148
14654 * lra-constraints.c (process_alt_operands): Reload operand
14655 conflicting with earlier clobber only if no more other conflicting
14656 operands.
14657
7d613735
JJ
146582013-02-22 Jakub Jelinek <jakub@redhat.com>
14659
14660 PR sanitizer/56393
14661 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
14662 if not linking a shared library.
14663
ac8d93a7
SL
146642013-02-22 Seth LaForge <sethml@google.com>
14665
14666 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
14667
e0237780
GY
146682013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
14669
3fdb53c1
TB
14670 * config/arm/arm.md (split for extendsidi): Update condition.
14671 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
14672 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
14673 (qhs_zextenddi_cstr): Likewise.
e0237780 14674
d7fde18c
JJ
146752013-02-21 Jakub Jelinek <jakub@redhat.com>
14676
be63b77d
JJ
14677 PR middle-end/56420
14678 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
14679 avoid signed wrapping.
14680 (expand_mult): Handle properly multiplication by
14681 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
14682 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
14683 in the compiler if coeff is HOST_WIDE_INT_MIN.
14684 (expand_divmod): Don't make ext_op1 static, change it's type to
14685 uhwi. Avoid undefined behavior in -INTVAL (op1).
14686
d7fde18c
JJ
14687 PR rtl-optimization/50339
14688 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
14689 field.
14690 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
14691 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
14692 into splitting_ashiftrt field.
14693 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
14694 ASHIFTRT.
14695 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
14696 choices.
14697
6aad4455
AH
146982013-02-20 Aldy Hernandez <aldyh@redhat.com>
14699
14700 PR middle-end/56108
14701 * trans-mem.c (execute_tm_mark): Do not expand transactions that
14702 are sure to go irrevocable.
14703
38fe784d
HPN
147042013-02-21 Hans-Peter Nilsson <hp@axis.com>
14705
14706 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
14707 scalars are valid operands.
14708
0fd44da3
MJ
147092013-02-21 Martin Jambor <mjambor@suse.cz>
14710
14711 PR tree-optimization/56310
14712 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
14713 only matching indices and non-negative final offsets.
14714 (intersect_aggregates_with_edge): Pass src_idx to
14715 agg_replacements_to_vector. Pass src_idx insstead of index to
14716 intersect_with_agg_replacements.
14717
7a92038b
MJ
147182013-02-21 Martin Jambor <mjambor@suse.cz>
14719
14720 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
14721 instead of hard-wired defaults.
14722
c0da9c37
MR
147232013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
14724
14725 * doc/invoke.texi (MIPS Options): Update documentation of the
14726 floating-point multiply-accumulate instruction restrictions.
14727
d247ea0c 147282013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS
14729
14730 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
14731 asan_shadow_offset on x86_64 linux.
14732
22deefcb
RB
147332013-02-21 Richard Biener <rguenther@suse.de>
14734
14735 PR tree-optimization/56415
14736 Revert
14737 2013-02-11 Richard Biener <rguenther@suse.de>
14738
14739 PR tree-optimization/56273
14740 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
14741 first VRP run.
14742
7bcc6e75
JJ
147432013-02-21 Jakub Jelinek <jakub@redhat.com>
14744
a0ad148f
JJ
14745 PR bootstrap/56258
14746 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
14747 instead of @itemx.
14748
7bcc6e75
JJ
14749 PR inline-asm/56405
14750 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
14751 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
14752
a0a7b611
JH
147532013-02-20 Jan Hubicka <jh@suse.cz>
14754
14755 PR tree-optimization/56265
88b97037
UB
14756 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
14757 when target is referenced for first time.
a0a7b611 14758
c0e50f72
RB
147592013-02-20 Richard Biener <rguenther@suse.de>
14760
14761 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
14762 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
14763 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
14764 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
14765 not return anything.
14766 (rename_ssa_copies): Do not remove unused locals.
14767 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 14768 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB
14769 * passes.c (execute_function_todo): Do not schedule unused locals
14770 removal if cleanup_tree_cfg did something.
14771 * tree-ssa-live.c (remove_unused_locals): Dump statistics
14772 about the number of removed locals.
14773
a52ca739
RB
147742013-02-20 Richard Biener <rguenther@suse.de>
14775
14776 PR tree-optimization/56398
88b97037 14777 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 14778
ec9c9d1b
MJ
147792013-02-20 Martin Jambor <mjambor@suse.cz>
14780
14781 PR tree-optimization/55334
14782 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
14783 restricted pointers to arrays.
14784
e91c8ed6 147852013-02-20 Richard Biener <rguenther@suse.de>
88b97037 14786 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB
14787
14788 PR tree-optimization/56396
14789 * tree-ssa-ccp.c (n_const_val): New static variable.
14790 (get_value): Return NULL for SSA names we don't have a lattice
14791 entry for.
14792 (ccp_initialize): Initialize n_const_val.
14793 * tree-ssa-copy.c (n_copy_of): New static variable.
14794 (init_copy_prop): Initialize n_copy_of.
14795 (get_value): Return NULL_TREE for SSA names we don't have a
14796 lattice entry for.
14797
3d916479
MJ
147982013-02-20 Martin Jambor <mjambor@suse.cz>
14799
14800 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
14801
71a86758
RB
148022013-02-20 Richard Biener <rguenther@suse.de>
14803
14804 * genpreds.c (write_lookup_constraint): Do not compare first
14805 letter of the constraint again.
14806
79836a12
RB
148072013-02-20 Richard Biener <rguenther@suse.de>
14808
14809 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
14810 and ceil_log2.
14811 (get_use_iv_cost): Terminate hashtable walk when coming across
14812 an empty entry.
14813
bbe4fb2c
IZ
148142013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
14815
14816 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
14817 reassociation for avx2 targets.
14818
a72d8780 148192012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 14820
c77f83d5
EI
14821 * config/microblaze/microblaze.c: microblaze_has_clz = 0
14822 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 14823 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI
14824 version and TARGET_PATTERN_COMPARE check
14825 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 14826
a72d8780 148272012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 14828
a72d8780 14829 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI
14830 function before branching.
14831
242387fa
AB
148322012-02-19 Andrey Belevantsev <abel@ispras.ru>
14833
14834 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
14835 DUMP_INSN_RTX_UID.
14836 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
14837
b4979ab9
AB
148382012-02-19 Andrey Belevantsev <abel@ispras.ru>
14839
14840 PR middle-end/55889
b4979ab9
AB
14841 * sel-sched.c: Include ira.h.
14842 (implicit_clobber_conflict_p): New function.
14843 (moveup_expr): Use it.
88b97037 14844 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 14845
24d63016
RB
148462013-02-19 Richard Biener <rguenther@suse.de>
14847
14848 PR tree-optimization/56384
14849 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
14850 (vn_hash_type): Split out from ...
14851 (vn_hash_constant_with_type): ... here.
14852 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
14853 (vn_phi_eq): Compare types from vn_phi_s structure.
14854 (vn_phi_lookup): Populate vn_phi_s type.
14855 (vn_phi_insert): Likewise.
14856
a475fd3d
JJ
148572013-02-19 Jakub Jelinek <jakub@redhat.com>
14858
47cc28f5
JJ
14859 PR tree-optimization/56350
14860 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
14861 if haven't found reduction or nested cycle operand, rather than
14862 asserting we must find it.
14863
a475fd3d
JJ
14864 PR tree-optimization/56381
14865 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
14866 to fold_build3.
14867
148682013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH
14869 Jakub Jelinek <jakub@redhat.com>
14870
14871 PR target/52555
14872 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
14873 (swap_optab_enable): Same.
14874 (init_all_optabs): Use argument instead of global.
88b97037 14875 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH
14876 * expr.h (init_all_optabs): Add argument to prototype.
14877 (TREE_OPTIMIZATION_OPTABS): New.
14878 (save_optabs_if_changed): Protoize.
14879 * optabs.h: Declare this_fn_optabs.
14880 * optabs.c (save_optabs_if_changed): New.
14881 Declare this_fn_optabs.
14882 (init_optabs): Add argument to init_all_optabs() call.
14883 * function.c (invoke_set_current_function_hook): Handle per
14884 function optabs.
14885 * function.h (struct function): New field optabs.
14886 * config/mips/mips.c (mips_set_mips16_mode): Handle when
14887 optimization_current_node has changed.
14888 * target-globals.h (save_target_globals_default_opts): Protoize.
14889 * target-globals.c (save_target_globals_default_opts): New.
14890
3f587ca3
JDA
148912013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14892
14893 PR target/56347
14894 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
14895 registers %fr12 and %fr12R as call used.
ceaca33e
JDA
14896
14897 PR target/56214
14898 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB
14899 and HImode, require all displacements to be an integer multiple of
14900 their mode size.
ceaca33e
JDA
14901 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
14902 only allow QImode and HImode when reload is in progress and strict is
14903 true. Likewise for symbolic addresses. Use base14_operand to check
14904 displacements in REG+BASE addresses.
14905
fe0b4796
RB
149062013-02-18 Richard Biener <rguenther@suse.de>
14907
14908 PR tree-optimization/56366
14909 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14910 handle sign-conversion of outer-loop initial induction value.
14911
6aaf596b
RB
149122013-02-18 Richard Biener <rguenther@suse.de>
14913
73db8ff1 14914 PR middle-end/56349
6aaf596b
RB
14915 * cfghooks.c (merge_blocks): If we merge a latch into another
14916 block adjust references to it.
14917 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
14918 (verify_loop_structure): Verify that a recorded latch is in fact
14919 a latch.
14920
5e97dfb6
RB
149212013-02-18 Richard Biener <rguenther@suse.de>
14922
14923 PR tree-optimization/56321
14924 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
14925 order SSA name release and virtual operand unlinking.
14926
825527e8
EI
149272013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14928
14929 * config/microblaze/microblaze.md (save_stack_block): Define.
14930 (restore_stack_block): Likewise.
14931
debd11d9
EI
149322013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14933
14934 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
14935 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
14936 * config/microblaze/microblaze.c (microblaze_option_override):
14937 Bail out early for PIC modes when target does not support PIC.
14938
8ec77be0
EI
149392013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14940
88b97037 14941 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI
14942 Replace with a microblaze version.
14943 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 14944 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI
14945 microblaze.
14946
b41288b3
JJ
149472013-02-16 Jakub Jelinek <jakub@redhat.com>
14948 Dodji Seketeli <dodji@redhat.com>
14949
14950 PR asan/56330
88b97037 14951 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ
14952 (instrument_mem_region_access): Do not forget to always put
14953 instrumentation of the of 'base' and 'base + len' in a "if (len !=
14954 0) statement, even for cases where either 'base' or 'base + len'
14955 are not instrumented -- because they have been previously
14956 instrumented. Simplify the logic by putting all the statements
14957 instrument 'base + len' inside a sequence, and then insert that
14958 sequence right before the current insertion point. Then, to
14959 instrument 'base + len', just get an iterator on that statement.
14960 And do not forget to update the pointer to iterator the function
14961 received as argument.
14962
47918951
VM
149632013-02-15 Vladimir Makarov <vmakarov@redhat.com>
14964
14965 PR rtl-optimization/56348
14966 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
14967
3fb7c699
SB
149682013-02-15 Steven Bosscher <steven@gcc.gnu.org>
14969
14970 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
14971 (clean_graph_dump_file): Pass base to start_graph_dump.
14972
576fe41a
RH
149732013-02-14 Richard Henderson <rth@redhat.com>
14974
14975 PR target/55941
14976 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
14977
bb3accfa
SB
149782013-02-14 Steven Bosscher <steven@gcc.gnu.org>
14979
14980 * collect2-aix.h: Define F_LOADONLY.
14981
03143140
RB
149822013-02-14 Richard Biener <rguenther@suse.de>
14983
14984 PR lto/50494
14985 * varasm.c (output_constant_def_1): Get the decl representing
14986 the constant as argument.
14987 (output_constant_def): Wrap output_constant_def_1.
14988 (make_decl_rtl): Use output_constant_def_1 with the decl
14989 representing the constant.
14990 (build_constant_desc): Optionally re-use a decl already
14991 representing the constant.
14992 (tree_output_constant_def): Adjust.
14993
8fb06726
DS
149942013-02-14 Dodji Seketeli <dodji@redhat.com>
14995
14996 Fix an asan crash
14997 * asan.c (instrument_builtin_call): Really put the length of the
14998 second source argument into src1_len.
14999
c4bfe8bf
JJ
150002013-02-13 Jakub Jelinek <jakub@redhat.com>
15001
15002 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15003 argument. If it is false, don't create edge from then_bb to
15004 fallthru_bb.
15005 (insert_if_then_before_iter): Pass true to it.
15006 (build_check_stmt): Pass false to it.
15007 (transform_statements): Flush hash table only on extended basic
15008 block boundaries, rather than at the beginning of every bb.
15009 Don't flush hash table on nonfreeing_call_p calls.
15010 * tree-flow.h (nonfreeing_call_p): New prototype.
15011 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15012
7afe2801
DM
150132013-02-13 David S. Miller <davem@davemloft.net>
15014
15015 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15016
70cc3288
VM
150172013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15018
15019 PR target/56184
15020 * ira.c (max_regno_before_ira): Move from ...
15021 (ira): ... here.
15022 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15023 vec_safe_length.
15024
6422242b
JJ
150252013-02-13 Jakub Jelinek <jakub@redhat.com>
15026
15027 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15028
976a81ee
RB
150292013-02-13 Richard Biener <rguenther@suse.de>
15030
15031 PR lto/56295
15032 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15033 globals in MEM_REFs.
15034
c1874a87
RB
150352013-02-13 Richard Biener <rguenther@suse.de>
15036
15037 * loop-init.c (loop_optimizer_init): Clear loop state when
15038 re-initializing preserved loops.
15039 * loop-unswitch.c (unswitch_single_loop): Return whether
15040 we unswitched the loop. Do not verify loop state here.
88b97037 15041 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 15042
b4ab7d34
KS
150432013-02-13 Kostya Serebryany <kcc@google.com>
15044
88b97037
UB
15045 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15046 on x86_64 linux.
b4ab7d34
KS
15047 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15048
bdcbe80c
DS
150492013-02-12 Dodji Seketeli <dodji@redhat.com>
15050
15051 Avoid instrumenting duplicated memory access in the same basic block
15052 * Makefile.in (asan.o): Add new dependency on hash-table.h
15053 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15054 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15055 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15056 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15057 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15058 (get_mem_ref_of_assignment): New functions.
15059 (get_mem_refs_of_builtin_call): Extract from
15060 instrument_builtin_call and tweak a little bit to make it fit with
15061 the new signature.
15062 (instrument_builtin_call): Use the new
15063 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15064 of is_gimple_builtin_call.
15065 (instrument_derefs, instrument_mem_region_access): Insert the
15066 instrumented memory reference into the hash table.
15067 (maybe_instrument_assignment): Renamed instrument_assignment into
15068 this, and change it to advance the iterator when instrumentation
15069 actually happened and return true in that case. This makes it
15070 homogeneous with maybe_instrument_assignment, and thus give a
15071 chance to callers to be more 'regular'.
15072 (transform_statements): Clear the memory reference hash table
15073 whenever we enter a new BB, when we cross a function call, or when
15074 we are done transforming statements. Use
15075 maybe_instrument_assignment instead of instrumentation. No more
15076 need to special case maybe_instrument_assignment and advance the
15077 iterator after calling it; it's now handled just like
15078 maybe_instrument_call. Update comment.
15079
4861a1f7
RB
150802013-02-13 Richard Biener <rguenther@suse.de>
15081
15082 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15083 Fix loop discovery code.
15084
1c86bd80
VM
150852013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15086
15087 PR inline-asm/56148
15088 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 15089 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM
15090 the operand is not reloaded. Prefer to reload conflicting operand
15091 if earlyclobber and matching operands are the same.
15092
7cbda518
RB
150932013-02-12 Richard Biener <rguenther@suse.de>
15094
15095 PR lto/56297
15096 * lto-streamer-out.c (write_symbol): Do not output symbols
15097 for hard register variables.
15098
e68a4ef6
GJL
150992013-02-12 Georg-Johann Lay <avr@gjlay.de>
15100
15101 PR target/54222
15102 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15103 (umulsidi3_insn, mulsidi3_insn): New insns.
15104
a72d8780 151052013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL
15106
15107 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15108 (struct tune_params): Add vec_costs field.
15109 * config/arm/arm.c (arm_builtin_vectorization_cost)
15110 (arm_add_stmt_cost): New functions.
15111 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15112 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15113 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15114 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15115 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15116 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15117 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15118
43320568
RB
151192013-02-12 Richard Biener <rguenther@suse.de>
15120
15121 PR lto/56295
15122 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15123 decls again if possible.
15124
a011aa39
RB
151252013-02-12 Richard Biener <rguenther@suse.de>
15126
15127 PR middle-end/56288
15128 * tree-ssa.c (verify_ssa_name): Fix check, move
15129 SSA_NAME_IN_FREE_LIST check up.
15130
6da26889
JJ
151312013-02-12 Jakub Jelinek <jakub@redhat.com>
15132 Steven Bosscher <steven@gcc.gnu.org>
15133
15134 PR rtl-optimization/56151
15135 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15136 equal to op0 or op1, and last_insn pattern is CODE operation
15137 with MEM dest and one of the operands matches that MEM.
15138
f80e0faf
ST
151392013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15140
15141 * doc/extend.texi: Document Function Multiversioning and "default"
15142 parameter string to target attribute.
15143 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15144 target attribute parameter is "default".
15145 (ix86_compare_version_priority): Remove checks for target attribute.
15146 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15147 Remove check for target attribute equal to NULL. Add assert.
15148 (ix86_generate_version_dispatcher_body): Change error to sorry.
15149
6c59ffd1
IS
151502013-02-11 Iain Sandoe <iain@codesourcery.com>
15151 Jack Howarth <howarth@bromo.med.uc.edu>
15152 Patrick Marlier <patrick.marlier@gmail.com>
15153
15154 PR libitm/55693
15155 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15156 define ENDFILE_SPEC as TM_DESTRUCTOR.
15157 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15158
a72d8780 151592013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP
15160 Jack Howarth <howarth@bromo.med.uc.edu>
15161 Jakub Jelinek <jakub@redhat.com>
15162
15163 PR sanitizer/55617
15164 * config/darwin.c (cdtor_record): Rename ctor_record.
15165 (sort_cdtor_records): Rename sort_ctor_records.
15166 (finalize_dtors): New routine to sort destructors by
15167 priority before use in assemble_integer.
15168 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15169
b63fe007
UB
151702013-02-11 Uros Bizjak <ubizjak@gmail.com>
15171
15172 PR rtl-optimization/56275
15173 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15174 offset is non-negative and less than cmode size before
15175 calling simplify_subreg.
15176
8e89b5b5
RB
151772013-02-11 Richard Biener <rguenther@suse.de>
15178
15179 PR tree-optimization/56264
15180 * cfgloop.h (fix_loop_structure): Adjust prototype.
15181 * loop-init.c (fix_loop_structure): Return the number of
15182 newly discovered loops.
15183 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15184 are discovered, do a full loop-closed SSA rewrite.
15185
b4a4b56d
RB
151862013-02-11 Richard Biener <rguenther@suse.de>
15187
15188 PR tree-optimization/56273
15189 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15190 first VRP run.
15191 (check_array_ref): Fix missing newline in dumps.
15192 (search_for_addr_array): Likewise.
15193
0c885229
DE
151942013-02-09 David Edelsohn <dje.gcc@gmail.com>
15195
15196 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15197
59ac9a55
JJ
151982013-02-09 Jakub Jelinek <jakub@redhat.com>
15199
15200 PR target/56256
15201 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15202
25bb0bb5
VM
152032013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15204
15205 PR rtl-optimization/56246
0c885229 15206 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM
15207 reload pseudo.
15208 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15209 constraints are satisfied.
15210
a698cc03
JL
152112013-02-08 Jeff Law <law@redhat.com>
15212
15213 PR debug/53948
15214 * emit-rtl.c (reg_is_parm_p): New function.
15215 * regs.h (reg_is_parm_p): New prototype.
15216 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15217 callee-clobbered registers.
15218
e1122ddd
MM
152192013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15220
15221 PR target/56043
15222 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15223 If there is no implicit builtin declaration, just return NULL.
15224
19c5f6e6
UB
152252013-02-08 Uros Bizjak <ubizjak@gmail.com>
15226
15227 * config/i386/sse.md (FMAMODEM): New mode iterator.
15228 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15229 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15230
2480f2ca 152312013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 15232
2480f2ca
UB
15233 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15234 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15235 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15236
152372013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15238
15239 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15240 (microblaze*-*-elf): Likewise.
15241 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15242 LINK_SPEC.
15243 * config/microblaze/microblaze-c.c: Add builtin defines for
15244 _LITTLE_ENDIAN and _BIG_ENDIAN.
15245 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15246 add to TARGET_DEFAULT flags.
76ef61fb 15247 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB
15248 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15249 * config/microblaze/microblaze.md: Update extendsidi2 and
15250 movdi_internal instructions to use low-order / high-order reg
15251 print_operands.
15252 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15253 options and inversemask / mask of LITTLE_ENDIAN.
15254 * config/microblaze/t-microblaze: Expand multilib options to
15255 include mlittle-endian (le) and update exceptions patterns.
15256
600a5961
JJ
152572013-02-08 Jakub Jelinek <jakub@redhat.com>
15258
5df81313
JJ
15259 PR rtl-optimization/56195
15260 * lra-constraints.c (get_reload_reg): Don't reuse regs
15261 if they have smaller mode than requested, if they have
15262 wider mode than requested, try to return a SUBREG.
15263
600a5961
JJ
15264 PR tree-optimization/56250
15265 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15266 if type is unsigned and code isn't MULT_EXPR.
15267
ff544649
GJL
152682013-02-08 Georg-Johann Lay <avr@gjlay.de>
15269
15270 PR tree-optimization/56064
15271 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15272 bits according to mode.
15273 * fixed-value.h (fixed_from_double_int)
15274 (const_fixed_from_double_int): Adjust comments.
15275
e45cde98
RB
152762013-02-08 Richard Biener <rguenther@suse.de>
15277
15278 PR lto/56231
15279 * lto-streamer.h (struct data_in): Remove current_file, current_line
15280 and current_col members.
15281 * lto-streamer-out.c (lto_output_location): Stream changed bits
15282 en-block for efficiency.
15283 * lto-streamer-in.c (clear_line_info): Remove.
15284 (lto_input_location): Cache current file, line and column
15285 globally via local statics. Read changed bits en-block.
15286 (input_function): Do not call clear_line_info.
15287 (lto_read_body): Likewise.
15288 (lto_input_toplevel_asms): Likewise.
15289
c1ca73d8
MM
152902013-02-08 Michael Matz <matz@suse.de>
15291
15292 PR tree-optimization/52448
15293 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15294 (nt_call_phase): New static.
15295 (add_or_mark_expr): Only mark accesses with newer phase than any
15296 call seen.
15297 (nonfreeing_call_p): New.
15298 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15299 (nt_fini_block): Keep blocks marked as visited.
15300 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15301
57548aa2
RB
153022013-02-08 Richard Biener <rguenther@suse.de>
15303
15304 * ira.c (ira): Free broken dominator information.
15305
8e10366f
UB
153062013-02-08 Uros Bizjak <ubizjak@gmail.com>
15307
15308 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15309
f64fb0fa
MP
153102013-02-08 Marek Polacek <polacek@redhat.com>
15311
8e10366f 15312 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 15313
0375167b
RB
153142013-02-08 Richard Biener <rguenther@suse.de>
15315
15316 PR middle-end/56181
15317 * cfgloop.h (flow_loops_find): Adjust.
15318 (bb_loop_header_p): Declare.
15319 * cfgloop.c (bb_loop_header_p): New function split out from ...
15320 (flow_loops_find): ... here. Adjust function signature,
15321 support incremental loop structure update.
15322 (verify_loop_structure): Cleanup. Verify a loop is a loop.
15323 * cfgloopmanip.c (fix_loop_structure): Move ...
15324 * loop-init.c (fix_loop_structure): ... here.
15325 (apply_loop_flags): Split out from ...
15326 (loop_optimizer_init): ... here.
15327 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
15328 in incremental mode, only remove dead loops here.
15329
85d768f3
GJL
153302013-02-08 Georg-Johann Lay <avr@gjlay.de>
15331
15332 PR target/54222
15333 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
15334 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
15335 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
15336 (*round<mode>3.libgcc): New insns for fixed-modes.
15337 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
15338 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
15339 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
15340 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
15341 implementations. Define to __builtin_avr_absFX,
15342 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
15343 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
15344 __builtin_avr_countlsFX, respectively.
15345 * config/avr/avr-c.c (target.h): Include it.
15346 (enum avr_builtin_id): New enum.
15347 (avr_resolve_overloaded_builtin): New static function.
15348 (avr_register_target_pragmas): Use it to set
15349 targetm.resolve_overloaded_builtin.
15350 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
15351 tree nodes used by DEF_BUILTIN.
15352 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
15353 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
15354 <AVR_BUILTIN_xxBITS>: Same.
15355
661bc682
RB
153562013-02-08 Richard Biener <rguenther@suse.de>
15357
15358 * cfgloop.c (verify_loop_structure): Properly handle
15359 a loop exiting to another loop header.
15360 * ira-int.h (ira_loops): Remove.
15361 * ira.c (ira_loops): Remove.
15362 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
15363 (do_reload): Use loop_optimizer_finalize.
15364 * ira-build.c (create_loop_tree_nodes): Use get_loops and
15365 number_of_loops to access the loop tree.
15366 (more_one_region_p): Likewise.
15367 (finish_loop_tree_nodes): Likewise.
15368 (rebuild_regno_allocno_maps): Likewise.
15369 (mark_loops_for_removal): Likewise.
15370 (mark_all_loops_for_removal): Likewise.
15371 (remove_unnecessary_regions): Likewise.
15372 (ira_build): Likewise.
15373 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15374
0d5049b2
RB
153752013-02-08 Richard Biener <rguenther@suse.de>
15376
15377 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
15378 * ipa-pure-const.c (analyze_function): Avoid calling
15379 mark_irreducible_loops twice.
8e10366f 15380 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 15381
499f32e8
DM
153822013-02-07 David S. Miller <davem@davemloft.net>
15383
15384 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
15385 on 'reg'.
15386 * var-tracking.c (vt_add_function_parameter): Test the presence of
15387 HAVE_window_save properly and do not remap argument registers when
15388 we have a leaf function.
15389
6edc3e32
UB
153902013-02-07 Uros Bizjak <ubizjak@gmail.com>
15391
15392 PR bootstrap/56227
15393 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
15394 instead of "ll".
15395 * config/i386/i386.c (ix86_print_operand): Ditto.
15396
5306401f
VM
153972013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15398
6edc3e32 15399 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 15400
027ece11
VM
154012013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15402
15403 PR rtl-optimization/56225
15404 * lra-constraints.c (process_alt_operands): Check that reload hard
15405 reg can hold value for strict_low_part.
15406
f980dfdb
JJ
154072013-02-07 Jakub Jelinek <jakub@redhat.com>
15408
15409 PR debug/56154
15410 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
15411 dwarf2out_end_function.
15412 (in_first_function_p, maybe_at_text_label_p,
15413 first_loclabel_num_not_at_text_label): New variables.
15414 (dwarf2out_var_location): In the first function find out
6edc3e32 15415 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ
15416 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
15417 functions.
15418
73dd3123
EB
154192013-02-07 Eric Botcazou <ebotcazou@adacore.com>
15420
15421 PR rtl-optimization/56178
15422 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
15423 SUBREG of a register. Tidy up related block of code.
15424 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
15425 note if the source is a register or a SUBREG of a register.
15426
e3936f47
JJ
154272013-02-07 Jakub Jelinek <jakub@redhat.com>
15428
15429 PR target/56228
15430 * config/rs6000/rs6000.md (ptrm): New mode attr.
15431 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
15432 call_value_indirect_aix<pttrsize>,
15433 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
15434 m in constraints.
15435
d96d674b
MH
154362013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
15437
15438 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
15439 if -bnortl. Convert to strcmp and strncmp.
15440
921f2dee
AM
154412013-02-07 Alan Modra <amodra@gmail.com>
15442
15443 PR target/54009
15444 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
15445 addresses won't wrap when offsetting.
15446 (rs6000_secondary_reload): Provide secondary reloads needed for
15447 wrapping LO_SUM addresses.
15448
d09c7dba
TS
154492013-02-06 Thomas Schwinge <thomas@codesourcery.com>
15450
15451 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
15452 MACH, just __MACH__.
15453
a44bbd48
RB
154542013-02-06 Richard Biener <rguenther@suse.de>
15455
15456 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
15457 instead of calling fix_loop_structure.
15458
6d840d99
JJ
154592013-02-06 Jakub Jelinek <jakub@redhat.com>
15460
15461 PR middle-end/56217
15462 * omp-low.c (use_pointer_for_field): Return false if
15463 lower_send_shared_vars doesn't generate any copy-out code.
15464
0f33baa9
TV
154652013-02-06 Tom de Vries <tom@codesourcery.com>
15466
15467 PR rtl-optimization/56131
15468 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
15469 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
15470 of the label is NULL. Add comment.
15471
11f1e3ab
JJ
154722013-02-05 Jakub Jelinek <jakub@redhat.com>
15473
31502f9f
JJ
15474 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
15475
a74db9bd
JJ
15476 PR sanitizer/55374
15477 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
15478 (STATIC_LIBTSAN_LIBS): Likewise.
15479 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
15480 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
15481 is defined, don't add anything else beyond that.
15482 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
15483 (LINK_COMMAND_SPEC): Use them.
15484
11f1e3ab
JJ
15485 PR tree-optimization/56205
15486 * tree-stdarg.c (check_all_va_list_escapes): Return true if
15487 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
15488 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
15489
14ac6aa2
RB
154902013-02-05 Richard Biener <rguenther@suse.de>
15491
15492 PR tree-optimization/53342
15493 PR tree-optimization/53185
15494 * tree-vectorizer.h (vect_check_strided_load): Remove.
15495 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
15496 not disallow peeling for vectorized strided loads.
15497 (vect_check_strided_load): Make static and simplify.
15498 (vect_analyze_data_refs): Adjust.
15499 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
15500 correctly when vectorizing strided loads.
15501
6f22445a
RB
155022013-02-05 Richard Biener <rguenther@suse.de>
15503
15504 * doc/install.texi: Refer to ISL, not PPL.
15505
39f9719e
JH
155062013-02-05 Jan Hubicka <jh@suse.cz>
15507
ec4224ac
JH
15508 PR tree-optimization/55789
15509 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
15510
155112013-02-05 Jan Hubicka <jh@suse.cz>
15512
15513 PR tree-optimization/55789
39f9719e
JH
15514 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
15515 the dead call anyway.
15516
956d3b33
EB
155172013-02-05 Eric Botcazou <ebotcazou@adacore.com>
15518
15519 PR sanitizer/55374
15520 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
15521
155222013-02-04 Alexander Potapenko <glider@google.com>
15523 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP
15524 Jakub Jelinek <jakub@redhat.com>
15525
15526 PR sanitizer/55617
15527 * config/darwin.c (sort_ctor_records): Stabilized qsort
15528 on constructor priority by using original position.
15529 (finalize_ctors): New routine to sort constructors by
15530 priority before use in assemble_integer.
15531 (machopic_asm_out_constructor): Use finalize_ctors if needed.
15532
7ac3af38
JJ
155332013-02-04 Jakub Jelinek <jakub@redhat.com>
15534
15535 PR libstdc++/54314
15536 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
15537 about visibility on artificial decls.
15538 * config/sol2.c (solaris_assemble_visibility): Likewise.
15539
152689dc
KT
155402013-02-04 Kai Tietz <ktietz@redhat.com>
15541
15542 PR target/56186
15543 * config/i386/i386.c (function_value_ms_64): Add additional valtype
15544 argument and improve checking of return-argument types for 16-byte
15545 modes.
15546 (ix86_function_value_1): Add additional valtype argument on call
15547 of function_value_64.
15548 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
15549 handling infunction_value_64 function.
15550
2b5987b5
MGD
155512013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
15552
7ac3af38 15553 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 15554
14597080
RB
155552013-02-04 Richard Biener <rguenther@suse.de>
15556
15557 PR tree-optimization/56188
15558 * tree-ssa-structalias.c (label_visit): Consider case with
15559 initially non-empty points-to set.
15560 (perform_var_substitution): Dump node mapping and clean up.
15561
ed73881e
RG
155622013-02-04 Richard Guenther <rguenther@suse.de>
15563
15564 PR lto/56168
15565 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
15566 node prevail as last resort.
15567 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 15568 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 15569
139a0707
RB
155702013-02-04 Richard Biener <rguenther@suse.de>
15571
15572 PR tree-optimization/56113
15573 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
15574 Merge into ...
15575 (equiv_class_lookup_or_add): ... this.
15576 (label_visit): Adjust and fix error in previous patch.
15577 (perform_var_substitution): Adjust.
15578
50fe8924
OE
155792013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
15580
15581 * config/sh/divtab.c: Fix formatting and comments throughout the file.
15582 * config/sh/sh4-300.md: Likewise.
15583 * config/sh/sh4a.md: Likewise.
15584 * config/sh/constraints.md: Likewise.
15585 * config/sh/sh.md: Likewise.
15586 * config/sh/netbsd-elf.h: Likewise.
15587 * config/sh/predicates.md: Likewise.
15588 * config/sh/sh-protos.h: Likewise.
15589 * config/sh/ushmedia.h: Likewise.
15590 * config/sh/linux.h: Likewise.
15591 * config/sh/sh.c: Likewise.
15592 * config/sh/superh.h: Likewise.
15593 * config/sh/elf.h: Likewise.
15594 * config/sh/sh4.md: Likewise.
15595 * config/sh/sh.h: Likewise.
15596
1a04ac2b
JDA
155972013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15598
15599 * config/pa/constraints.md: Adjust unused letters. Change "T"
15600 constraint to match_test floating_point_store_memory_operand().
15601 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
15602 (base14_operand): New.
15603 (floating_point_store_memory_operand): New.
15604 (integer_store_memory_operand): Revise to use base14_operand and
15605 reg_plus_base_memory_operand.
15606 (move_dest_operand): Allow symbolic_memory_operands.
15607 (symbolic_memory_operand): Check for LO_SOM.
15608 (symbolic_operand): Change default case to break.
15609 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
15610 CONST_DOUBLE values to be reloaded by putting them into memory when
15611 the destination is a floating point register.
15612 (movdf): Remove code to handle CONST_DOUBLE.
15613 (movsf): Likewise.
15614 (reload_indf_r1): New.
15615 (reload_insf_r1): New.
15616 Consistently use "Q" and "T" constraints with integer and floating
15617 point move instructions, respectively.
15618 (movdi): Remove FAIL.
15619 Change predicate for source operand unamed DImode move from
15620 general_operand to move_src_operand.
15621 (umulsidi3): Change predicate for destination operand to
15622 register_operand.
15623 Likewise for similar unamed patterns.
15624 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
15625 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
15626 (hppa_legitimize_address): Simplify mask calculation.
15627 (pa_emit_move_sequence): Revised handling of secondary reloads from
15628 REG+D addresses for floating point loads and stores. Directly handle
15629 loading CONST0_RTX (mode) to a floating point register.
15630 (pa_secondary_reload): Handle reloading DF and SFmode constant values
15631 to floating point registers. Don't restrict secondary reloads to
15632 floating point registers to integer modes. Revise some comments and
15633 cleanup some code.
15634 (TARGET_LEGITIMATE_ADDRESS_P): Define.
15635 (pa_legitimate_address_p): New.
15636 (pa_legitimize_reload_address): New.
15637 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
15638 (STRICT_REG_OK_FOR_BASE_P): New.
15639 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
15640 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
15641
3bf65591
DE
156422013-02-03 David Edelsohn <dje.gcc@gmail.com>
15643 Andrew Dixie <andrewd@gentrack.com>
15644
15645 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
15646 flag set.
15647
c0a8a3e6
RS
156482013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
15649
15650 * expmed.c (extract_bit_field_1): Pass the full width of the
15651 structure to get_best_reg_extraction_insn.
15652
99113dff
DE
156532013-02-01 David Edelsohn <dje.gcc@gmail.com>
15654
15655 PR target/54601
15656 * configure.ac (use_cxa_atexit): Add AIX.
15657 * configure: Regenerate.
15658
15659 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
15660
ddd84654
JJ
156612013-02-01 Jakub Jelinek <jakub@redhat.com>
15662
15663 PR debug/54793
15664 * final.c (need_profile_function): New variable.
15665 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
15666 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
15667 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
15668 notes, targetm.asm_out.function_prologue doesn't emit anything,
15669 HAVE_prologue and profiler should be emitted before prologue,
15670 set need_profile_function instead of emitting it.
15671 (final_scan_insn): If need_profile_function, emit
15672 profile_function on the first NOTE_INSN_BASIC_BLOCK or
15673 NOTE_INSN_FUNCTION_BEG note.
15674
156752013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH
15676
15677 * config/rs6000/rs6000.md (smulditi3): New.
15678 (umulditi3): New.
15679
ff2a9d88
RH
15680 * config/alpha/alpha.md (umulditi3): New.
15681
14d52b90
DE
156822013-02-01 David Edelsohn <dje.gcc@gmail.com>
15683
15684 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
15685 (ASM_OUTPUT_ALIGNED_LOCAL): New.
15686
8c7ca45c
RB
156872013-02-01 Richard Biener <rguenther@suse.de>
15688
15689 PR tree-optimization/56113
15690 * tree-ssa-structalias.c (label_visit): Reduce work for
15691 single-predecessor nodes.
15692
9f419393
EB
156932013-02-01 Eric Botcazou <ebotcazou@adacore.com>
15694
15695 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
15696 range isn't testing for zero.
15697
c5f4be84
SB
156982013-01-31 Steven Bosscher <steven@gcc.gnu.org>
15699
15700 PR middle-end/56113
15701 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
15702
dbdbd982
NC
157032013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
15704 Nick Clifton <nickc@redhat.com>
15705
15706 * config/v850/constraints.md (Q): Define as a memory constraint.
15707 * config/v850/predicates.md (label_ref_operand): New predicate.
15708 (e3v5_shift_operand): New predicate.
15709 (ior_operator): New predicate.
15710 * config/v850/t-v850: Add e3v5 multilib.
15711 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
15712 (v850_gen_movdi): Prototype.
15713 * config/v850/v850.c: Add support for e3v5 architecture.
15714 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
15715 TARGET_V850E_UP.
15716 (construct_save_jarl): Add e3v5 long JARL support.
15717 (v850_adjust_insn_length): New function. Adjust length of call
15718 insns when using e3v5 instructions.
15719 (v850_gen_movdi): New function: Generate instructions to move a
15720 DImode value.
15721 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
15722 (CPP_SPEC): Define __v850e3v5__ as appropriate.
15723 (TARGET_USE_FPU): Enable for e3v5.
15724 (CONST_OK_FOR_W): New macro.
15725 (ADJUST_INSN_LENGTH): Define.
15726 * config/v850/v850.md (UNSPEC_LOOP): Define.
15727 (attr cpu): Add v850e3v5.
15728 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
15729 (movdi): New pattern.
15730 (movdi_internal): New pattern.
15731 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
15732 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
15733 (cstoresf4): Likewise.
15734 (cstoredf4): Likewise.
15735 (insv): New pattern.
15736 (rotlso3_a): New pattern.
15737 (rotlsi3_b): New pattern
15738 (rotlsi3_v850e3v5): New pattern.
15739 (doloop_begin): New pattern.
15740 (fix_loop_counter): New pattern.
15741 (doloop_end): New pattern.
15742 (branch_normal): Add e3v5 long branch support.
15743 (branch_invert): Likewise.
15744 (branch_z_normal): Likewise.
15745 (branch_z_invert): Likewise.
15746 (branch_nz_normal): Likewise.
15747 (branch_nz_invert): Likewise.
15748 (call_internal_short): Add e3v5 register-indirect JARL support.
15749 (call_internal_long): Likewise.
15750 (call_value_internal_short): Likewise.
15751 (call_value_internal_long): Likewise.
15752 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
15753 (mloop): New option.
15754 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 15755 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 15756
73861a41
PK
157572013-01-31 Paul Koning <ni1d@arrl.net>
15758
15759 PR debug/55059
15760 PR debug/54508
15761 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
15762 children if parent is a class.
15763 (prune_unused_types_prune): Don't add DW_AT_declaration.
15764
e44978dc
RB
157652013-01-31 Richard Biener <rguenther@suse.de>
15766
15767 PR tree-optimization/56157
15768 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
15769 match up operand with SLP child.
15770
90b10dec
JM
157712013-01-31 Jason Merrill <jason@redhat.com>
15772
04d2dadd 15773 PR debug/54410
90b10dec
JM
15774 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
15775 parameters the first time.
15776 (gen_scheduled_generic_parms_dies): Check completeness here.
15777
6e616110
RB
157782013-01-31 Richard Biener <rguenther@suse.de>
15779
15780 PR middle-end/53073
15781 * common.opt (faggressive-loop-optimizations): New flag,
15782 enabled by default.
15783 * doc/invoke.texi (faggressive-loop-optimizations): Document.
15784 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
15785 infer_loop_bounds_from_undefined by it.
15786
636f59cf
RB
157872013-01-31 Richard Biener <rguenther@suse.de>
15788
15789 PR tree-optimization/56150
15790 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
15791 visit virtual operands.
15792 (find_uses_to_rename_bb): Likewise.
15793
157942013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB
15795
15796 PR tree-optimization/56150
15797 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
15798 mixed store non-store stmts.
15799
32887460
JJ
158002013-01-30 Jakub Jelinek <jakub@redhat.com>
15801
e60e09a0
JJ
15802 PR sanitizer/55374
15803 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
15804 LIBASAN_EARLY_SPEC is defined.
15805 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
15806 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
15807 before %o.
15808 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
15809
32887460
JJ
15810 PR c++/55742
15811 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
15812 invalid args instead of ICEing on it.
15813 (ix86_valid_target_attribute_tree): Return error_mark_node if
15814 ix86_valid_target_attribute_inner_p failed.
15815 (ix86_valid_target_attribute_p): Return false only if
15816 ix86_valid_target_attribute_tree returned error_mark_node. Allow
15817 target("default") attribute.
15818 (sorted_attr_string): Change argument from const char * to tree,
15819 merge in all target attribute arguments rather than just one.
15820 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
15821 instead of free. Avoid using strcat.
15822 (ix86_mangle_function_version_assembler_name): Mangle
15823 target("default") as if no target attribute is present. Adjust
15824 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
15825 instead of xmalloc and XDELETEVEC instead of free.
15826 (ix86_function_versions): Don't return true if one of the decls
15827 doesn't have target attribute. If they don't and one of the decls
15828 is DECL_FUNCTION_VERSIONED, report an error. Adjust
15829 sorted_attr_string caller. Use XDELETEVEC instead of free.
15830 (ix86_supports_function_versions): Remove.
15831 (make_name): Fix up formatting.
15832 (make_dispatcher_decl): Remove resolver_name and its initialization.
15833 Avoid leaking memory.
15834 (is_function_default_version): Return true if there is
15835 target("default") attribute rather than no target attribute at all.
15836 (make_resolver_func): Avoid leaking memory.
15837 (ix86_generate_version_dispatcher_body): Likewise.
15838 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
15839 * target.def (supports_function_versions): Remove.
15840 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
15841 * doc/tm.texi: Regenerated.
15842
73cca0cc
VM
158432013-01-30 Vladimir Makarov <vmakarov@redhat.com>
15844
15845 PR rtl-optimization/56144
15846 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
15847 for values with side effects.
15848
11452e7b
RB
158492013-01-30 Richard Biener <rguenther@suse.de>
15850
15851 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
15852 (sparseset_pop): Likewise.
15853 * cfganal.c (compute_idf): Likewise. Increase work-stack size
15854 to be able to use quick_push in the worker loop.
15855
01cb1ef5
MP
158562013-01-30 Marek Polacek <polacek@redhat.com>
15857
15858 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
15859
e892936e
RB
158602013-01-30 Richard Biener <rguenther@suse.de>
15861
15862 PR lto/56147
6edc3e32 15863 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 15864
cc06c01d
GJL
158652013-01-30 Georg-Johann Lay <avr@gjlay.de>
15866
15867 PR tree-optimization/56064
15868 * fixed-value.c (fixed_from_double_int): New function.
15869 * fixed-value.h (fixed_from_double_int): New prototype.
15870 (const_fixed_from_double_int): New static inline function.
15871 * fold-const.c (native_interpret_fixed): New static function.
15872 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
15873 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
15874 (native_encode_fixed): New static function.
15875 (native_encode_expr) <FIXED_CST>: Use it.
15876 (native_interpret_int): Move double_int worker code to...
15877 * double-int.c (double_int::from_buffer): ...this new static method.
15878 * double-int.h (double_int::from_buffer): Prototype it.
15879
d394a308
RB
158802013-01-30 Richard Biener <rguenther@suse.de>
15881
15882 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
15883 New pointer-map and obstack.
15884 (init_alias_vars): Allocate pointer-map and obstack.
15885 (delete_points_to_sets): Free them.
15886 (find_what_var_points_to): Cache result.
15887 (find_what_p_points_to): Adjust for changed interface of
15888 find_what_var_points_to.
15889 (compute_points_to_sets): Likewise.
15890 (ipa_pta_execute): Likewise.
15891
20804d96
RO
158922013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15893
15894 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
15895 * configure: Regenerate.
15896 * config.in: Regenerate.
15897 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
15898 #nobits/#progbits if supported.
15899
a7ad88a2
OE
159002013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
15901
15902 PR target/56121
15903 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
15904 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
15905 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
15906
7af79f92
GY
159072013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15908
91bfca59
OE
15909 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
15910 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 15911
753bcf7b
GY
159122013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15913
91bfca59
OE
15914 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
15915 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 15916
8cbc2ea8
GY
159172013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15918
91bfca59
OE
15919 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
15920 declaration.
15921 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
15922 * config/arm/cortex-a7.md: New bypasses using
15923 arm_mac_accumulator_is_result.
8cbc2ea8 15924
697a3325
GY
159252013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15926
91bfca59 15927 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 15928 (cortex_a7_neon_mla): Likewise.
91bfca59
OE
15929 (cortex_a7_fpfmad): New reservation.
15930 (cortex_a7_fpmacs): Use ffmas and update required units.
15931 (cortex_a7_fpmuld): Update required units and latency.
15932 (cortex_a7_fpmacd): Likewise.
15933 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
15934 (cortex_a7_neon). Likewise.
15935 (bypass) Update participating units.
697a3325 15936
29637783
GY
159372013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
15938
91bfca59
OE
15939 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
15940 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
15941 from fmac to ffma.
15942 * config/arm/vfp11.md (vfp_farith): Use ffmas.
15943 (vfp_fmul): Use ffmad.
15944 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
15945 (cortex_r4_fmacd): Use ffmad.
15946 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
15947 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
15948 (cortex_a9_fmacd): Use ffmad.
15949 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
15950 (cortex_a8_vfp_macd): Use ffmad.
15951 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
15952 (cortex_a5_fpmacd): Use ffmad.
15953 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
15954 (cortex_a15_vfp_macd): Use ffmad.
15955 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 15956
2ee8a2d5
JM
159572013-01-29 Jason Merrill <jason@redhat.com>
15958
15959 PR libstdc++/54314
15960 * varasm.c (default_assemble_visibility): Don't warn about
15961 visibility on artificial decls.
15962
78d087bc
RB
159632013-01-29 Richard Biener <rguenther@suse.de>
15964
15965 PR tree-optimization/56113
15966 * tree-ssa-structalias.c (equiv_class_lookup): Also return
15967 the bitmap leader.
15968 (label_visit): Free duplicate bitmaps and record the leader instead.
15969 (perform_var_substitution): Adjust.
15970
83ba4d6f
RB
159712013-01-29 Richard Biener <rguenther@suse.de>
15972
15973 PR tree-optimization/55270
15974 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
15975 the CFG, schedule loops for fixup.
15976
66dfe13f
NC
159772013-01-29 Nick Clifton <nickc@redhat.com>
15978
15979 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
15980 SP_REG.
15981
5a579c3b
LE
159822013-01-28 Leif Ekblad <leif@rdos.net>
15983
15984 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
15985 * config/i386/i386.h (TARGET_RDOS): New macro.
15986 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
15987 * config/i386/i386.c (ix86_option_override_internal): For 64bit
15988 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
15989 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
15990 DEFAULT_LARGE_SECTION_THRESHOLD.
15991 * config/i386/i386.md (R14_REG, R15_REG): New constants.
15992 * config/i386/rdos.h: New file.
15993 * config/i386/rdos64.h: New file.
15994
51e44392
BS
159952013-01-28 Bernd Schmidt <bernds@codesourcery.com>
15996
15997 PR other/54814
15998 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
15999 TEST_HARD_REG_BIT.
16000
db1fb332
JJ
160012013-01-28 Jakub Jelinek <jakub@redhat.com>
16002
16003 PR rtl-optimization/56117
16004 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16005 call cselib_lookup_from_insn on the MEM before calling
16006 add_insn_mem_dependence.
16007
16917761
RB
160082013-01-28 Richard Biener <rguenther@suse.de>
16009
16010 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16011 to a stmt that didn't have one.
16012 (copy_phis_for_bb): Likewise for PHI arguments.
16013 (copy_debug_stmt): Likewise for debug stmts.
16014
b9fc0497
RB
160152013-01-28 Richard Biener <rguenther@suse.de>
16016
16017 PR tree-optimization/56034
6edc3e32 16018 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB
16019 (partition_builtin_p): Adjust.
16020 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16021 it is the last partition.
16022 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16023 up the vertex for the definition.
16024 (classify_partition): Classify whether a partition is a
16025 PKIND_REDUCTION, thus has uses outside of the loop.
16026 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16027 Merge all PKIND_REDUCTION partitions into the last partition.
16028 (tree_loop_distribution): Seed partitions from reductions as well.
16029
aa710d25
JJ
160302013-01-28 Jakub Jelinek <jakub@redhat.com>
16031
0bfbca58
JJ
16032 PR tree-optimization/56125
16033 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16034 pow(x,c) into sqrt(x) * powi(x, n/2) or
16035 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16036 optimizing for size.
16037 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16038 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16039 integer.
16040
aa710d25
JJ
16041 PR tree-optimization/56094
16042 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16043 to UNKNOWN_LOCATION while gimplifying expr.
16044
77dc5297
UB
160452013-01-27 Uros Bizjak <ubizjak@gmail.com>
16046
16047 PR target/56114
16048 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16049 operand 0 in movabs insn template for -masm=intel asm alternative.
16050 (*movabs<mode>_2): Ditto for operand 1.
16051
0bfbca58 160522013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH
16053
16054 PR target/54663
16055 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16056 of microblaze-c.o
16057
0bfbca58 160582013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI
16059
16060 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16061 tm_file.
16062
0bfbca58 160632013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N
16064
16065 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16066 Undef to avoid warning.
16067
478f60f9
MH
160682013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16069
16070 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16071 * configure: Regenerate.
16072
d7fa6ee2
JJ
160732013-01-25 Jakub Jelinek <jakub@redhat.com>
16074
16075 PR tree-optimization/56098
16076 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16077 for stmts with volatile ops.
16078 (cond_store_replacement): Don't optimize if assign has volatile ops.
16079 (cond_if_else_store_replacement_1): Don't optimize if either
16080 then_assign or else_assign have volatile ops.
16081 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16082 volatile ops.
16083
f8fe87bd
GJL
160842013-01-25 Georg-Johann Lay <avr@gjlay.de>
16085
16086 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16087
28398d0d
GJL
160882013-01-25 Georg-Johann Lay <avr@gjlay.de>
16089
16090 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16091 missing ':' in asm example.
16092
b7d7d917
TB
160932013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16094
16095 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16096 entries into lane and laneq entries.
77dc5297
UB
16097 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16098 Remove AdvSIMD scalar modes.
b7d7d917
TB
16099 (aarch64_sq<r>dmulh_laneq<mode>): New.
16100 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16101 modes.
16102 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16103 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16104 * config/aarch64/iterators.md (VCOND): New.
16105 (VCONQ): New.
16106
556f9906
GJL
161072013-01-25 Georg-Johann Lay <avr@gjlay.de>
16108
16109 PR target/54222
16110 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16111 Add NULL LIBNAME argument to existing definitions.
16112 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16113 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16114 * config/avr/avr.c (DEF_BUILTIN): Same.
16115 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16116 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16117 implementation is available (DECL_ASSEMBLER_NAME is set).
16118 (avr_fold_absfx): New static function.
16119 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16120 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16121 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16122 AVR_BUILTIN_ABSLLK.
16123 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16124 (abshk, absk, abslk, absllk): Provide as static inline functions.
16125
1f546bbb
MP
161262013-01-25 Marek Polacek <polacek@redhat.com>
16127
16128 PR tree-optimization/56035
16129 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16130
7e184bd7
UB
161312012-01-24 Uros Bizjak <ubizjak@gmail.com>
16132
16133 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16134 (*movtf_internal_rex64): Add (!o,C) alternative
16135 (*movxf_internal_rex64): Ditto.
16136 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16137
f5ba49ac
SM
161382013-01-24 Shenghou Ma <minux.ma@gmail.com>
16139
16140 * doc/invoke.texi: fix typo.
16141 * doc/objc.texi: fix typo.
16142
a6343728
RS
161432013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16144
16145 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16146 for the first two alternatives.
16147
cd030c07
DN
161482013-01-24 Diego Novillo <dnovillo@google.com>
16149
77dc5297 16150 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN
16151 (ggc-zone.o): Remove.
16152 * configure.ac: Remove option --with-gc.
16153 * configure: Re-generate.
16154 * doc/install.texi: Remove documentation for --with-gc.
16155 * gengtype.c (write_enum_defn): Remove. Update all users.
16156 (write_Types_process_field): Remove generation of gt_e_* argument.
16157 (output_type_enum): Remove. Update all users.
16158 (write_enum_defn): Remove. Update all users.
16159 (enum alloc_zone): Remove. Update all users.
77dc5297 16160 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN
16161 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16162 Update all callers.
16163 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 16164 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN
16165 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16166 Update all users.
aaf1e810 16167 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN
16168 (struct alloc_zone): Remove.
16169 (ggc_internal_alloc_zone_stat): Remove.
16170 (ggc_internal_cleared_alloc_zone_stat): Remove.
16171 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB
16172 (ggc_pch_count_object): Remove last argument. Update all users.
16173 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN
16174 (struct alloc_zone): Remove.
16175 * ggc-zone.c: Remove.
77dc5297 16176 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN
16177 (struct alloc_zone): Remove.
16178 (ggc_alloc_typed_stat): Remove.
16179 (ggc_alloc_typed): Remove.
16180 (ggc_splay_alloc): Remove first argument.
16181 (rtl_zone): Remove. Update all users.
16182 (tree_zone): Remove. Update all users.
16183 (tree_id_zone): Remove. Update all users.
16184 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 16185 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 16186 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 16187 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 16188
a861ffa4
GJL
161892013-01-24 Georg-Johann Lay <avr@gjlay.de>
16190
16191 * config/avr/avr.c (avr_out_fract): Make register numbers that
16192 might be outside of source operand signed.
16193
593c0ddd
UB
161942013-01-24 Uros Bizjak <ubizjak@gmail.com>
16195
16196 * config/i386/constraints.md (Yf): New constraint.
16197 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16198 of f constraint to conditionaly disable x87 register preferences.
16199 (*movdf_internal): Ditto.
16200 (*movsf_internal): Ditto.
16201
e86c0101
SB
162022013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16203
16204 PR inline-asm/55934
16205 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16206 that have operands with impossible constraints.
16207 Add a FIXME for a speed-up opportunity.
16208 * lra-constraints.c (process_alt_operands): Verify that a class
16209 selected from constraints on asms is valid for the operand mode.
16210 (curr_insn_transform): Remove incorrect comment.
16211
f6fee35f
DE
162122013-01-23 David Edelsohn <dje.gcc@gmail.com>
16213
16214 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16215 TOC operand is a valid symbol ref in the constant pool.
16216
162172013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 16218
aaf1e810 16219 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 16220
dc62d7d1
GJL
162212013-01-23 Georg-Johann Lay <avr@gjlay.de>
16222
16223 PR target/54222
16224 * config/avr/stdfix.h: New file.
16225 * t-avr (stdfix-gcc.h): New rule to build it.
16226 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16227
2660d12d
KS
162282013-01-23 Kostya Serebryany <kcc@google.com>
16229
77dc5297
UB
16230 * config/darwin.h: remove dependency on
16231 CoreFoundation (asan on Mac OS).
2660d12d 16232
a70418fc
JJ
162332013-01-23 Jakub Jelinek <jakub@redhat.com>
16234
16235 PR target/49069
16236 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16237 instead of cmpdi_operand for first comparison operand.
16238 Don't assert that comparison operands aren't both constants.
16239
47876a2a
JW
162402013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16241
16242 * doc/install.texi (Downloading the Source): Update references to
16243 downloading separate components.
16244
162452013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW
16246
16247 * doc/extend.texi (__int128): Improve grammar.
16248
47876a2a 162492013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB
16250
16251 PR target/56028
16252 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16253 alternative to (o,r).
16254 (*movdi_internal_rex64): Remove (!o,n) alternative.
16255 (DImode immediate->memory splitter): Remove.
16256 (DImode immediate->memory peephole2): Remove.
16257 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16258 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16259 alternative to (!o,*r).
16260 (*movtf_internal_sse): New pattern.
16261 (*movxf_internal_rex64): New pattern.
16262 (*movxf_internal): Disable for TARGET_64BIT.
16263 (*movdf_internal_rex64): Remove (!o,F) alternative.
16264
3a984f10
JJ
162652013-01-22 Jakub Jelinek <jakub@redhat.com>
16266
502498d5
JJ
16267 PR middle-end/56074
16268 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16269 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16270 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16271 stmt locations where LOCATION_LOCUS of the stmt location is
16272 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16273
3a984f10
JJ
16274 PR target/55686
16275 * config/i386/i386.md (UNSPEC_STOS): New.
16276 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16277 *strsetqi_1): Add UNSPEC_STOS.
16278
fa817f7f
PC
162792013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16280
16281 PR c++/56067
16282 * doc/invoke.texi: Remove left over -Wsynth example.
16283
8f498c1b
JJ
162842013-01-21 Jakub Jelinek <jakub@redhat.com>
16285
16286 PR tree-optimization/56051
16287 * fold-const.c (fold_binary_loc): Don't fold
16288 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16289 a narrowing conversion, or widening conversion from signed
16290 to unsigned.
16291
47876a2a 162922013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB
16293
16294 PR rtl-optimization/56023
16295 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16296 dependent on debug instruction.
16297
5b9db1bc
MJ
162982013-01-21 Martin Jambor <mjambor@suse.cz>
16299
16300 PR middle-end/56022
16301 * function.c (allocate_struct_function): Call
16302 invoke_set_current_function_hook earlier.
16303
e8bb7d68
JJ
163042013-01-21 Jakub Jelinek <jakub@redhat.com>
16305
16306 * reload1.c (init_reload): Only initialize reload_obstack
16307 during the first call.
16308
616a4e32
MP
163092013-01-21 Marek Polacek <polacek@redhat.com>
16310
16311 * cfgloop.c (verify_loop_structure): Fix up grammar.
16312
4401981b
YHH
163132013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
16314
16315 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
16316 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
16317
8e87740b
RR
163182013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16319
16320 PR target/56058
16321 * config/arm/marvell-pj4.md: Update copyright year.
16322 Fix up use of alu to alu_reg and simple_alu_imm.
16323
47876a2a 163242013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB
16325
16326 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
16327
89d56d79
VM
163282013-01-20 Vladimir Makarov <vmakarov@redhat.com>
16329
16330 PR target/55433
16331 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 16332 insn for secondary memory move when memory mode should be different.
89d56d79 16333
fe603553
JDA
163342013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16335
16336 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
16337 atomic_storedi_1): New patterns.
16338
01284895
VK
163392013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16340
16341 btver2 pipeline descriptions.
16342 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
16343 descriptions.
16344 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 16345 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK
16346 type attributes.
16347 * config/i386/btver2.md: New file describing btver2 pipelines.
16348
5630e3e1
JL
163492013-01-19 Andrew Pinski <apinski@cavium.com>
16350
16351 PR tree-optimization/52631
16352 * tree-ssa-sccvn (visit_use): Before looking up the original
16353 statement, try looking up the simplified expression.
16354
650ae806
AG
163552013-01-19 Anthony Green <green@moxielogic.com>
16356
16357 * config/moxie/moxie.c (moxie_expand_prologue): Set
16358 current_function_static_stack_size.
16359
e300ec2d
JJ
163602013-01-18 Jakub Jelinek <jakub@redhat.com>
16361
16362 PR tree-optimization/56029
16363 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
16364 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
16365
a3d7af04
SS
163662013-01-18 Sharad Singhai <singhai@google.com>
16367
16368 PR tree-optimization/55995
16369 * dumpfile.c (dump_loc): Print location only if available.
16370 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
16371
66aa7879
VM
163722013-01-18 Vladimir Makarov <vmakarov@redhat.com>
16373
16374 PR target/55433
16375 * lra-constraints.c (curr_insn_transform): Reuse original insn for
16376 secondary memory move.
16377 (inherit_reload_reg): Use rclass instead of cl for
16378 check_secondary_memory_needed_p.
16379
3f0fee7b
JJ
163802013-01-18 Jakub Jelinek <jakub@redhat.com>
16381
16382 PR middle-end/56015
16383 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 16384 the case where writing real complex part of target modifies op1.
3f0fee7b 16385
70c67693
JG
163862013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
16387
16388 * config/aarch64/aarch64-simd.md
16389 (aarch64_vcond_internal<mode>): Handle unordered cases.
16390 * config/aarch64/iterators.md (v_cmp_result): New.
16391
df8de9b3
YHH
163922013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
16393 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16394
16395 * config/arm/marvell-pj4.md: New file.
16396 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
16397 * config/arm/arm.md (generic_sched): Add marvell_pj4.
16398 (generic_vfp): Likewise.
16399 * config/arm/arm-cores.def: Add marvell-pj4.
16400 * config/arm/arm-tune.md: Regenerate.
16401 * config/arm/arm-tables.opt: Regenerate.
16402 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
16403 * doc/invoke.texi: Document marvell-pj4.
16404
be30c356
TB
164052013-01-18 Tejas Belagod <tejas.belagod@arm.com>
16406
16407 * config/aarch64/arm_neon.h: Map scalar types to standard types.
16408
0bfbca58 164092013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO
16410
16411 PR debug/54114
16412 PR debug/54402
16413 PR debug/49888
16414 * var-tracking.c (negative_power_of_two_p): New.
16415 (global_get_addr_cache, local_get_addr_cache): New.
16416 (get_addr_from_global_cache, get_addr_from_local_cache): New.
16417 (vt_canonicalize_addr): Rewrite using the above. Adjust the
16418 heading comment.
16419 (vt_stack_offset_p): Remove.
16420 (vt_canon_true_dep): Always canonicalize loc's address.
16421 (clobber_overlapping_mems): Make sure we have a MEM.
16422 (local_get_addr_clear_given_value): New.
16423 (val_reset): Clear local cached entries.
16424 (compute_bb_dataflow): Create and release the local cache.
16425 Disable duplicate MEMs clobbering.
16426 (emit_notes_in_bb): Clobber MEMs likewise.
16427 (vt_emit_notes): Create and release the local cache.
16428 (vt_initialize, vt_finalize): Create and release the global
16429 cache, respectively.
1f6bc337 16430 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 16431
0bfbca58 164322013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO
16433
16434 PR libmudflap/53359
16435 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
16436 not found in the symtab.
16437
0bfbca58 164382013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 16439
c350ba53 16440 PR debug/56006
3aa03517
AO
16441 PR rtl-optimization/55547
16442 PR rtl-optimization/53827
16443 PR debug/53671
16444 PR debug/49888
16445 * alias.c (offset_overlap_p): New, factored out of...
16446 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
16447 the conservative special case for symbolic constants. Don't
16448 adjust zero sizes on alignment.
16449
c664546f
JL
164502013-01-18 Bernd Schmidt <bernds@codesourcery.com>
16451
16452 PR rtl-optimization/52573
16453 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
16454 REG_UNUSED for the same register.
16455
1bd3f750
MP
164562013-01-17 Richard Biener <rguenther@suse.de>
16457 Marek Polacek <polacek@redhat.com>
16458
16459 PR rtl-optimization/55833
16460 * loop-unswitch.c (unswitch_loops): Move loop verification...
16461 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
16462 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
16463 Set it to true when we're removing a loop from hierarchy tree in
16464 an irreducible region.
16465 (fix_bb_placements): Adjust caller.
16466 (fix_loop_placements): Likewise.
16467
e52a8b71
GJL
164682013-01-17 Georg-Johann Lay <avr@gjlay.de>
16469
16470 * config/avr/builtins.def (DEF_BUILTIN): Factor out
16471 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
16472 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
16473 Remove ID. Adjust comments.
16474 * config/avr/avr-c.c (avr_builtin_name): Remove.
16475 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
16476 * config/avr/avr.c (avr_tolower): New static function.
16477 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
16478 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
16479 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
16480 default expansion.
16481
8386a7ea
JH
164822013-01-17 Jan Hubicka <jh@suse.cz>
16483
610fb637 16484 PR tree-optimization/55273
8386a7ea
JH
16485 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
16486
47876a2a 164872013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB
16488
16489 PR target/55981
16490 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
16491 store through atomic_store<mode>_1.
16492 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
16493
8222c37e
MJ
164942013-01-17 Martin Jambor <mjambor@suse.cz>
16495
16496 PR tree-optimizations/55264
16497 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
16498 for virtual methods.
16499 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
16500 virtual methods before inlining is over.
16501 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
16502 virtual functions.
16503 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
16504 non-virtual.
16505
79f01c76
VM
165062013-01-16 Vladimir Makarov <vmakarov@redhat.com>
16507
16508 PR rtl-optimization/56005
16509 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
16510 pending reads for prefetch.
16511
d0b6bb1b
IB
165122013-01-16 Ian Bolton <ian.bolton@arm.com>
16513
aaf1e810 16514 * config/aarch64/aarch64.md
d0b6bb1b
IB
16515 (*cstoresi_neg_uxtw): New pattern.
16516 (*cmovsi_insn_uxtw): New pattern.
16517 (*<optab>si3_uxtw): New pattern.
16518 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
16519 (*<optab>si3_insn_uxtw): New pattern.
16520 (*bswapsi2_uxtw): New pattern.
16521
cb9cf03b
RB
165222013-01-16 Richard Biener <rguenther@suse.de>
16523
16524 * tree-inline.c (tree_function_versioning): Remove set but
16525 never used variable.
16526
2cfc56b9
RB
165272013-01-16 Richard Biener <rguenther@suse.de>
16528
16529 PR tree-optimization/55964
16530 * tree-flow.h (rename_variables_in_loop): Remove.
16531 (rename_variables_in_bb): Likewise.
16532 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
16533 (copy_loop_before): Adjust and delete update-ssa status.
16534 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 16535 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB
16536 (rename_variables_in_loop): Remove.
16537 (slpeel_update_phis_for_duplicate_loop): Likewise.
16538 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
16539 use available cfg machinery instead of duplicating it.
16540 Update PHI nodes and perform poor-mans SSA update here.
16541 (slpeel_tree_peel_loop_to_edge): Adjust.
16542
c25a0c60
RB
165432013-01-16 Richard Biener <rguenther@suse.de>
16544
16545 PR tree-optimization/54767
16546 PR tree-optimization/53465
16547 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
16548 (vrp_visit_phi_node): For PHI arguments coming via backedges
16549 drop all symbolical range information.
16550 (execute_vrp): Compute backedges.
16551
04b535af
RB
165522013-01-16 Richard Biener <rguenther@suse.de>
16553
16554 * doc/install.texi: Update CLooG and ISL requirements to
16555 0.18.0 and 0.11.1.
16556
8b0a1e0b
CB
165572013-01-16 Christian Bruel <christian.bruel@st.com>
16558
16559 PR target/55301
16560 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
16561 (broken_move): Handle UNSPECV_SP_SWITCH_B.
16562 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
16563
165642013-01-16 DJ Delorie <dj@redhat.com>
16565
16566 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
16567 (UNSPECV_SP_SWITCH_E): New.
16568 (sp_switch_1): Change to an unspec.
16569 (sp_switch_2): Change to an unspec. Don't use post-inc when we
16570 replace $r15.
16571
165722013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB
16573
16574 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
16575 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
16576 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
16577 (expand_mem_thread_fence): Ditto.
16578 (expand_mem_signal_fence): Ditto.
16579 (expand_atomic_load): Ditto.
16580 (expand_atomic_store): Ditto.
16581
0bfbca58 165822013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO
16583
16584 PR rtl-optimization/55547
16585 PR rtl-optimization/53827
16586 PR debug/53671
16587 PR debug/49888
16588 * alias.c (memrefs_conflict_p): Set sizes to negative after
16589 AND adjustments.
16590
305e3ac1
JJ
165912013-01-15 Jakub Jelinek <jakub@redhat.com>
16592
16593 PR target/55940
16594 * function.c (thread_prologue_and_epilogue_insns): Always
16595 add crtl->drap_reg to set_up_by_prologue.set, even if
16596 stack_realign_drap is false.
16597
f78ac4f2
JBG
165982013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16599
16600 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
16601 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
16602 *call): Fix indention.
16603
a78a8cc4
TV
166042013-01-15 Tom de Vries <tom@codesourcery.com>
16605
16606 PR target/55876
16607 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
16608 Update comment.
16609
0e80383f
VM
166102013-01-15 Vladimir Makarov <vmakarov@redhat.com>
16611
305e3ac1 16612 PR rtl-optimization/55153
0e80383f
VM
16613 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
16614
207b5956
MJ
166152013-01-15 Martin Jambor <mjambor@suse.cz>
16616
16617 PR tree-optimization/55920
16618 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
16619 accesses as grp_to_be_debug_replaced.
16620
a7818b54
JJ
166212013-01-15 Jakub Jelinek <jakub@redhat.com>
16622
16623 PR tree-optimization/55920
16624 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
16625 there is non-useless type conversion needed from debug rhs to lhs,
16626 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
16627
b0fe107e
JM
166282013-01-15 Joseph Myers <joseph@codesourcery.com>
16629 Mikael Pettersson <mikpe@it.uu.se>
16630
16631 PR target/43961
16632 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
16633 Thumb.
16634 (ASM_OUTPUT_CASE_LABEL): Remove.
16635 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
16636 * final.c (shorten_branches): Update alignment of labels before
16637 jump tables if CASE_VECTOR_SHORTEN_MODE.
16638
34ab62ee
RB
166392013-01-15 Richard Biener <rguenther@suse.de>
16640
16641 PR bootstrap/55961
16642 * system.h: Do not include gmp.h for building host tools.
16643
783a3a05
RB
166442013-01-15 Richard Biener <rguenther@suse.de>
16645
16646 PR middle-end/55882
16647 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
16648 account for bitpos when computing alignment.
16649
3a579e09
VY
166502013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
16651
16652 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
16653 (ix86_target_macros_internal): Likewise.
16654
16655 * config/i386/i386.c (m_CORE2I7): Removed.
16656 (m_CORE_HASWELL): New macro.
16657 (m_CORE_ALL): Likewise.
16658 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
16659 (initial_ix86_arch_features): Likewise.
16660 (processor_target_table): Initializations for Core avx2.
16661 (cpu_names): New names "core-avx2".
16662 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
16663 PROCESSOR_CORE_HASWELL.
16664 (ix86_issue_rate): New case.
16665 (ia32_multipass_dfa_lookahead): Likewise.
16666 (ix86_sched_init_global): Likewise.
16667
16668 * config/i386/i386.h (TARGET_HASWELL): New macro.
16669 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
16670 (processor_type): New PROCESSOR_HASWELL.
16671
955f5a07
JJ
166722013-01-15 Jakub Jelinek <jakub@redhat.com>
16673
ff784829
JJ
16674 PR tree-optimization/55955
16675 * tree-vect-loop.c (vectorizable_reduction): Give up early on
16676 *SHIFT_EXPR and *ROTATE_EXPR codes.
16677
955f5a07
JJ
16678 PR tree-optimization/48766
16679 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
16680 -ftrapv disable -fwrapv.
16681
aeb8b4e9
GJL
166822013-01-14 Georg-Johann Lay <avr@gjlay.de>
16683
16684 PR target/55974
16685 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
16686 etc. to 1 and not to __flash.
16687 Use LL suffix for __INT24_MAX__ with -mint8.
16688 Use ULL suffix for __UINT24_MAX__ with -mint8.
16689
1c494c6a
GJL
166902013-01-14 Georg-Johann Lay <avr@gjlay.de>
16691
16692 * config/avr/avr-arch.h
16693 (struct base_arch_s): Use typedef avr_arch_t instead.
16694 (struct arch_info_s): Use typedef avr_arch_info_t instead.
16695 (struct mcu_type_s): Use typedef avr_mcu_t instead.
16696 * config/avr/avr.c: Same.
16697 * config/avr/avr-devices.c: Same.
16698 * config/avr/driver-avr.c: Same.
16699 * config/avr/gen-avr-mmcu-texi.c: Same.
16700 * config/avr/avr-mcus.def: Adjust comment.
16701
a50344cb
TB
167022013-01-14 Tejas Belagod <tejas.belagod@arm.com>
16703
88e784e6
UB
16704 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
16705 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 16706
47876a2a 167072013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB
16708 Andi Kleen <ak@linux.intel.com>
16709
16710 PR target/55948
16711 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
16712 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
16713 memmodel flag.
16714
00892272
GJL
167152013-01-14 Georg-Johann Lay <avr@gjlay.de>
16716
16717 * config/avr/avr-stdint.h: Remove trailing blanks.
16718 * config/avr/avr-log.h: Same.
16719 * config/avr/avr-arch.h: Same.
16720 * config/avr/avr-devices.c: Same.
16721 * config/avr/avr-dimode.md: Same.
16722 * config/avr/predicates.md: Same.
16723 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 16724
00892272
GJL
16725 * config/avr/avr-protos.h: Same. And:
16726 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
16727 (init_cumulative_args): Rename to avr_init_cumulative_args.
16728 (expand_prologue): Rename to avr_expand_prologue.
16729 (expand_epilogue): Rename to avr_expand_epilogue.
16730 (adjust_insn_length): Rename to avr_adjust_insn_length.
16731 (notice_update_cc): Rename to avr_notice_update_cc.
16732 (final_prescan_insn): Rename to avr_final_prescan_insn.
16733 * config/avr/avr.c: Same.
16734 * config/avr/avr.h: Same.
16735 * config/avr/avr.md: Remove trailing blanks.
16736 (prologue): Use avr_expand_prologue.
16737 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
16738
7bb01996
RB
167392013-01-14 Richard Biener <rguenther@suse.de>
16740
16741 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
16742 verify_location, collect_subblocks): New functions.
16743 (verify_gimple_in_cfg): Verify that locations only reference
16744 BLOCKs in the functions BLOCK tree.
16745
2724573f
RB
167462013-01-14 Richard Biener <rguenther@suse.de>
16747
16748 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
16749 PHI argument.
16750 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
16751 unshare reference.
16752 (insert_out_of_ssa_copy_on_edge): Likewise.
16753 (rewrite_close_phi_out_of_ssa): Likewise.
16754 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
16755 debug expressions.
16756 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
16757 propagated constants.
16758 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
16759 can not be shared.
16760
9a0bbab6
GJL
167612013-01-14 Georg-Johann Lay <avr@gjlay.de>
16762
16763 * config/avr/avr-modes.def: Add GPL copyright notice.
16764
45805f17
UB
167652013-01-13 Uros Bizjak <ubizjak@gmail.com>
16766
16767 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
16768 MEMMODEL_MASK to determine memory model.
16769 (atomic_store<mode>): Ditto from operands[2].
16770 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
16771
9d60be38
JJ
167722013-01-13 Jakub Jelinek <jakub@redhat.com>
16773
16774 PR fortran/55935
45805f17 16775 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ
16776 (fold_gimple_assign): Don't call unshare_expr here.
16777 (fold_ctor_reference): Call unshare_expr.
16778
e7f49d92
TG
167792013-01-13 Terry Guo <terry.guo@arm.com>
16780
aaf1e810
EB
16781 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
16782 * doc/fragments.texi: Document MULTILIB_REUSE.
16783 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 16784 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 16785 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG
16786 (tmpmultilib4): Ditto.
16787 (multilib_reuse): New multilib argument.
16788
fbd03a27
RS
167892013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
16790
16791 * Makefile.in: Update copyright.
16792
b3681f13
TV
167932013-01-12 Tom de Vries <tom@codesourcery.com>
16794
16795 PR middle-end/55890
16796 * calls.c (expand_call): Check if arg_nr is valid.
16797
3f287e4b
MM
167982013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
16799
16800 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
16801 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM
16802 documentation. Add missing '__' in front of
16803 __builtin_ia32_packssdw256.
3f287e4b 16804
1abcd5eb
AK
168052013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16806
16807 PR target/55719
16808 * config/s390/s390.c (s390_preferred_reload_class): Do not return
16809 NO_REGS for larl operands.
16810 (s390_reload_larl_operand): Use s390_load_address instead of
16811 emit_move_insn.
16812
980d0812
RB
168132013-01-11 Richard Biener <rguenther@suse.de>
16814
16815 * tree-cfg.c (verify_node_sharing_1): Split out from ...
16816 (verify_node_sharing): ... here.
16817 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
16818
7cb9fd07
EB
168192013-01-11 Eric Botcazou <ebotcazou@adacore.com>
16820
16821 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
16822 Substitute TREECHECKING.
16823 * configure: Regenerate.
16824 * Makefile.in (TREECHECKING): New.
16825
47876a2a 168262013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG
16827
16828 PR tree-optimization/44061
16829 * tree-vrp.c (extract_range_basic): Compute zero as
16830 value-range for __builtin_constant_p of function parameters.
16831
d1e082c2
RS
168322013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
16833
45805f17 16834 Update copyright years.
d1e082c2 16835
f85021b0
VM
168362013-01-10 Vladimir Makarov <vmakarov@redhat.com>
16837
9d60be38 16838 PR rtl-optimization/55672
0160442c
VM
16839 * lra-eliminations.c (mark_not_eliminable): Permit addition with
16840 const to be eliminable.
f85021b0 16841
7a8b1ec4
DE
168422013-01-10 David Edelsohn <dje.gcc@gmail.com>
16843
16844 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
16845 * configure: Regenerate.
16846
ad2e5b71
RB
168472013-01-10 Richard Biener <rguenther@suse.de>
16848
ddf9322d 16849 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 16850
ffc5b2bb
RB
168512013-01-10 Richard Biener <rguenther@suse.de>
16852
16853 PR bootstrap/55792
16854 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
16855 locations for virtual PHI arguments.
16856 (rewrite_update_phi_arguments): Likewise.
16857
e1f674e4
JS
168582013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
16859
16860 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
16861 on to assembler.
16862
a57fc743
JJ
168632013-01-10 Jakub Jelinek <jakub@redhat.com>
16864
16865 PR tree-optimization/55921
16866 * tree-complex.c (expand_complex_asm): New function.
16867 (expand_complex_operations_1): Call it for GIMPLE_ASM.
16868
0ff4390d
AK
168692013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16870
16871 PR target/55718
16872 * config/s390/s390.c (s390_symref_operand_p)
16873 (s390_loadrelative_operand_p): Merge the two functions.
16874 (s390_check_qrst_address, print_operand_address): Add parameters
16875 to s390_loadrelative_operand_p invokation.
16876 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
16877 (s390_reload_larl_operand, s390_secondary_reload): Use
16878 s390_loadrelative_operand_p instead of s390_symref_operand_p.
16879 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
16880
6f557e0e
MS
168812013-01-09 Mike Stump <mikestump@comcast.net>
16882
16883 * dse.c (record_store): Remove unnecessary assert.
16884
7770c9e9
JH
168852013-01-09 Jan Hubicka <jh@suse.cz>
16886
16887 PR tree-optimization/55569
16888 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
16889 * cfgloop.h (scale_loop_profile): Likewise.
16890
a19b1432
JH
168912013-01-09 Jan Hubicka <jh@suse.cz>
16892
16893 PR lto/45375
ddf9322d
UB
16894 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
16895 functions.
a19b1432
JH
16896 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
16897
9a002da8
RS
168982013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
16899
16900 PR middle-end/55114
16901 * expr.h (maybe_emit_group_store): Declare.
16902 * expr.c (maybe_emit_group_store): New function.
16903 * builtins.c (expand_builtin_int_roundingfn): Call it.
16904 (expand_builtin_int_roundingfn_2): Likewise.
16905
511dcace
VM
169062013-01-09 Vladimir Makarov <vmakarov@redhat.com>
16907
e1f2b729 16908 PR rtl-optimization/55829
511dcace
VM
16909 * lra-constraints.c (match_reload): Add code for absent output.
16910 (curr_insn_transform): Add code for reloads of matched inputs
16911 without output.
16912
7b0fe4f4
UB
169132013-01-09 Uros Bizjak <ubizjak@gmail.com>
16914
16915 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
16916 attribute of movddup insn to DF.
16917 (*vec_interleave_lowv2df): Ditto.
16918 (vec_dupv2df): Ditto.
16919
870ca331
JH
169202013-01-09 Jan Hubicka <jh@suse.cz>
16921
16922 PR tree-optimiation/55875
16923 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
16924 EVERY_ITERATION parameter.
7b0fe4f4 16925 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH
16926 (idx_infer_loop_bounds): Similarly here.
16927 (n_of_executions_at_most): Simplify
16928 to only test for cases where statement is dominated by the
7b0fe4f4 16929 particular bound; handle correctly the "postdominance" test.
870ca331
JH
16930 (scev_probably_wraps_p): Use max loop iterations info
16931 as a global bound first.
16932
6f575fe4 169332013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD
16934 Nick Clifton <nickc@redhat.com>
16935
16936 * config/v850/v850.md (cbranchsf4): New pattern.
16937 (cstoresf4): New pattern.
16938 (cbranchdf4): New pattern.
16939 (cstoredf4): New pattern.
16940 (movsicc): Disallow floating point comparisons.
16941 (cmpsf_le_insn): Fix order of operators.
16942 (cmpsf_lt_insn): Likewise.
16943 (cmpsf_eq_insn): Likewise.
16944 (cmpdf_le_insn): Likewise.
16945 (cmpdf_lt_insn): Likewise.
16946 (cmpdf_eq_insn): Likewise.
16947 (cmpsf_ge_insn): Use LE comparison.
16948 (cmpdf_ge_insn): Likewise.
16949 (cmpsf_gt_insn): Use LT comparison.
16950 (cmpdf_gt_insn): Likewise.
16951 (cmpsf_ne_insn): Delete pattern.
16952 (cmpdf_ne_insn): Delete pattern.
16953 * config/v850/v850.c (v850_gen_float_compare): Use
16954 gen_cmpdf_eq_insn for NE comparison.
16955 (v850_float_z_comparison_operator)
16956 (v850_float_nz_comparison_operator): Move from here ...
16957 * config/v850/predicates.md: ... to here. Move GT and GE
16958 comparisons into v850_float_z_comparison_operator.
16959 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
16960 Delete prototype.
16961 (v850_float_nz_comparison_operator): Likewise.
16962
f0d54148
JDA
169632013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16964
16965 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
16966 with calls to gen_insvsi/gen_insvdi.
16967
8f01beca
VK
169682013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16969
16970 * config/i386/i386.c (initial_ix86_tune_features): Set up
16971 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
16972
2aa579ad
JJ
169732013-01-09 Steven Bosscher <steven@gcc.gnu.org>
16974 Jakub Jelinek <jakub@redhat.com>
16975
16976 PR tree-optimization/48189
16977 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
16978 If nitercst is 0, don't predict the exit edge.
16979
6edc3e32 169802013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N
16981
16982 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
16983 in asm_fprintf with reg_names.
16984 (aarch64_print_operand_address): Likewise.
16985 (aarch64_return_addr): Likewise.
16986 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
16987
f6f94d94
JDA
169882013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16989
16990 * config/pa/pa.h (VAL_U6_BITS_P): Define.
16991 (INT_U6_BITS): Likewise.
16992 * config/pa/predicates.md (uint6_operand): New predicate.
16993 (shift5_operand, shift6_operand): Likewise.
16994 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
16995 arith32_operand.
16996 (lshrdi3): Use shift6_operand.
16997 (shrpsi4, shrpdi4): New insn patterns.
16998 (extzv): Delete expander.
16999 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17000 predicates in unamed zero extract patterns. Tighten common constraint.
17001 (extv): Delete expander.
17002 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17003 predicates in unamed sign extract patterns. Tighten common constraint.
17004 (insv): Delete expander.
17005 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17006 predicates in unamed insert patterns. Tighten common constraint.
17007 Change uint32_operand predicate to uint6_operand predicate in unamed
17008 DImode pattern to insert constant values of type 1...1xxxx.
17009
36b72910
JH
170102013-01-04 Jan Hubicka <jh@suse.cz>
17011
17012 PR tree-optimization/55823
7b0fe4f4
UB
17013 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17014 issue.
36b72910 17015
47876a2a 170162013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ
17017 Uros Bizjak <ubizjak@gmail.com>
17018
17019 PR rtl-optimization/55845
17020 * df-problems.c (can_move_insns_across): Stop scanning at
17021 volatile_insn_p source instruction or give up if
17022 across_from .. across_to range contains any volatile_insn_p
17023 instructions.
17024
4369c11e
TB
170252013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17026
7b0fe4f4
UB
17027 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17028 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17029 Declare.
4369c11e 17030 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 17031 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 17032
aeb7e7c1
JJ
170332013-01-08 Jakub Jelinek <jakub@redhat.com>
17034
17035 PR fortran/55341
17036 * asan.c (asan_clear_shadow): New function.
17037 (asan_emit_stack_protection): Use it.
17038
a02ad1aa
TB
170392013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17040
17041 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17042 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17043 with tab instead of space.
17044
f90d87f5
NC
170452013-01-08 Nick Clifton <nickc@redhat.com>
17046
17047 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17048 register bank 0 at the start of an interrupt handler.
83ffd964
NC
17049 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17050 MDBH registers.
f90d87f5 17051
385eb93d
JG
170522013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17053
17054 * config/aarch64/aarch64-simd.md
17055 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17056 (aarch64_simd_bsl): Likewise.
17057 (aarch64_vcond_internal<mode>): Likewise.
17058 (vcond<mode><mode>): Likewise.
17059 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17060 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17061
4dcd1054
JG
170622013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17063
17064 * config/aarch64/aarch64-builtins.c
17065 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17066
4267a4a6
MJ
170672013-01-08 Martin Jambor <mjambor@suse.cz>
17068
17069 PR debug/55579
17070 * tree-sra.c (analyze_access_subtree): Return true also after
17071 potentially creating a debug-only replacement.
17072
5f4e6de3
JJ
170732013-01-08 Jakub Jelinek <jakub@redhat.com>
17074
3138f224
JJ
17075 PR middle-end/55890
17076 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17077
5f4e6de3
JJ
17078 PR tree-optimization/54120
17079 * tree-vrp.c (range_fits_type_p): Don't allow
17080 src_precision < precision from signed vr to unsigned_p
17081 if vr->min or vr->max is negative.
17082 (simplify_float_conversion_using_ranges): Test can_float_p
17083 against CODE_FOR_nothing.
17084
f3ef18ff
JJ
170852013-01-08 Jakub Jelinek <jakub@redhat.com>
17086 Richard Biener <rguenther@suse.de>
17087
17088 PR middle-end/55851
17089 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17090 types instead of just INTEGER_TYPE types.
17091
25c210f9
MK
170922013-01-07 Mark Kettenis <kettenis@openbsd.org>
17093
17094 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17095 TF_SIZE): Define.
f90d87f5 17096
81826a7b
SE
170972013-01-07 Steve Ellcey <sellcey@mips.com>
17098
17099 PR target/42661
17100 * config/mips/mips.opt: Change mad to mmad to match documentation.
17101
35678514
GJL
171022013-01-07 Georg-Johann Lay <avr@gjlay.de>
17103
17104 PR target/55897
17105 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17106 .progmemx.data now.
17107
e5669488
GJL
171082013-01-07 Georg-Johann Lay <avr@gjlay.de>
17109
17110 PR target/55897
17111 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17112 (avr_addrspace_t): Add .section_name field.
17113 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17114 array size.
17115 (avr_addrspace): Same. Initialize .section_name. Remove last
17116 NULL entry. Put __memx into .progmemx.data.
17117 (progmem_section_prefix): Remove.
17118 (avr_asm_init_sections): No need to initialize progmem_section.
17119 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17120 section name prefix.
17121 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17122 retrieve the progmem section.
17123 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17124 boundary to run over avr_addrspace[].
17125 (avr_register_target_pragmas): Ditto.
17126
656e6f37
JJ
171272013-01-06 Jakub Jelinek <jakub@redhat.com>
17128
17129 * varasm.c (output_constant_def_contents): For asan_protect_global
17130 protected strings, adjust DECL_ALIGN if needed, before testing for
17131 anchored symbols.
17132 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17133 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17134 normal decls.
17135 (output_object_block): For asan protected decls, emit asan padding
17136 after their contents.
17137 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17138 (asan_finish_file): Test it here instead.
17139
6edc3e32
UB
171402013-01-07 Nick Clifton <nickc@redhat.com>
17141 Matthias Klose <doko@debian.org>
17142 Doug Kwan <dougkwan@google.com>
17143 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC
17144
17145 PR driver/55470
17146 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17147
17148 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17149
17150 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17151
7b0fe4f4 17152 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC
17153
17154 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17155
2a095093
GJL
171562013-01-07 Georg-Johann Lay <avr@gjlay.de>
17157
f3b503f4 17158 PR target/54461
2a095093
GJL
17159 * doc/install.texi (Cross-Compiler-Specific Options): Document
17160 --with-avrlibc.
17161
383f9b34
TB
171622013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17163
17164 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17165 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17166 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17167 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17168 vqmovun_high_s64): Fix source operand number and update copyright.
17169
3626621a
RB
171702013-01-07 Richard Biener <rguenther@suse.de>
17171
17172 PR middle-end/55890
17173 * gimple.h (gimple_call_builtin_p): New overload.
17174 * gimple.c (validate_call): New function.
17175 (gimple_call_builtin_p): Likewise.
17176 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17177 Use gimple_call_builtin_p.
17178 (find_func_clobbers): Likewise.
17179 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17180 (strlen_optimize_stmt): Likewise.
17181
8b2ea410
JG
171822013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17183
17184 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17185 (vld1q_dup_*): Likewise.
17186 (vld1_*): Likewise.
17187 (vld1q_*): Likewise.
17188 (vld1_lane_*): Likewise.
17189 (vld1q_lane_*): Likewise.
17190
9713d329
RB
171912013-01-07 Richard Biener <rguenther@suse.de>
17192
17193 * lto-streamer.h (LTO_minor_version): Bump to 2.
17194
3520f7cc
JG
171952013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17196
17197 * config/aarch64/aarch64-protos.h
17198 (aarch64_const_double_zero_rtx_p): Rename to...
17199 (aarch64_float_const_zero_rtx_p): ...this.
17200 (aarch64_float_const_representable_p): New.
17201 (aarch64_output_simd_mov_immediate): Likewise.
17202 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17203 move immediate case.
17204 * config/aarch64/aarch64.c
17205 (aarch64_const_double_zero_rtx_p): Rename to...
17206 (aarch64_float_const_zero_rtx_p): ...this.
17207 (aarch64_print_operand): Allow printing of new constants.
17208 (aarch64_valid_floating_const): New.
17209 (aarch64_legitimate_constant_p): Check for valid floating-point
17210 constants.
17211 (aarch64_simd_valid_immediate): Likewise.
17212 (aarch64_vect_float_const_representable_p): New.
17213 (aarch64_float_const_representable_p): Likewise.
17214 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17215 (aarch64_output_simd_mov_immediate): New.
17216 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17217 (*movdf_aarch64): Likewise.
17218 * config/aarch64/constraints.md (Ufc): New.
17219 (Y): call aarch64_float_const_zero_rtx.
17220 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17221
e167c04d
RB
172222013-01-07 Richard Biener <rguenther@suse.de>
17223
17224 PR tree-optimization/55888
17225 PR tree-optimization/55862
17226 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17227 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17228 not if it is contained therein.
17229
0139aaab
GJL
172302013-01-07 Georg-Johann Lay <avr@gjlay.de>
17231
17232 * config/avr/t-avr: Typo.
17233
4a176b23
GJL
172342013-01-07 Georg-Johann Lay <avr@gjlay.de>
17235
17236 PR55243
17237 * config/avr/t-avr: Don't automatically rebuild
17238 $(srcdir)/config/avr/t-multilib
17239 $(srcdir)/config/avr/avr-tables.opt
17240 $(srcdir)/doc/avr-mmcu.texi
17241 (avr-mcus): New phony target to build them on request.
17242 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17243 * avr/avr-mcus.def: Adjust comments.
17244
c7afdc98
UB
172452013-01-07 Uros Bizjak <ubizjak@gmail.com>
17246
17247 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17248
1ab05c31
RS
172492013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17250
17251 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17252
488e3acc
RS
172532013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17254
17255 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17256
a89599d2
DE
172572013-01-05 David Edelsohn <dje.gcc@gmail.com>
17258
17259 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17260 to generate profiling.
17261 * config/rs6000/aix64.h (LIB_SPEC): Same.
17262
70f09188
AP
172632013-01-04 Andrew Pinski <apinski@cavium.com>
17264
17265 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17266 New function.
17267 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17268
918d445f
UB
172692013-01-04 Uros Bizjak <ubizjak@gmail.com>
17270
17271 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17272 unconditionally.
17273 (ix86_expand_move): Ditto.
17274 (ix86_zero_extend_to_Pmode): Ditto.
17275 (ix86_expand_call): Ditto.
17276 (ix86_expand_special_args_builtin): Ditto.
17277 (ix86_expand_builtin): Ditto.
17278
361618ec
RB
172792013-01-04 Richard Biener <rguenther@suse.de>
17280
17281 PR tree-optimization/55862
17282 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17283 translating them through PHI nodes.
17284
87eab554
MJ
172852013-01-04 Martin Jambor <mjambor@suse.cz>
17286
17287 PR tree-optimization/55755
17288 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17289 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17290
33879b9f
MP
172912013-01-04 Marek Polacek <polacek@redhat.com>
17292
17293 PR middle-end/55859
17294 * opts.c (default_options_optimization): Clarify error message.
17295
3068819a
RB
172962013-01-04 Richard Biener <rguenther@suse.de>
17297
17298 PR middle-end/55863
17299 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17300 reassociation.
17301
0e4ae794
JDA
173022013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17303
17304 PR target/53789
17305 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA
17306 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17307 references.
0e4ae794 17308
65c88cef
RH
173092013-01-03 Richard Henderson <rth@redhat.com>
17310
17311 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17312 after eliminating TLS symbols.
17313
8ac16127
MG
173142013-01-03 Marc Glisse <marc.glisse@inria.fr>
17315
17316 PR bootstrap/50167
17317 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
17318 * graphite-poly.c (debug_gmp_value): Likewise.
17319
bb664f09
UB
173202013-01-03 Uros Bizjak <ubizjak@gmail.com>
17321
17322 PR target/55712
17323 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
17324 selected code model, define __code_mode_small__, __code_model_medium__,
17325 __code_model_large__, __code_model_32__ or __code_model_kernel__.
17326 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
17327 xchg temporary register with %k. Declare temporary register as
17328 early clobbered.
17329 [__x86_64__]: For medium and large code models, preserve %rbx register.
17330
bcf1ef00
RB
173312013-01-03 Richard Biener <rguenther@suse.de>
17332
0506634a 17333 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB
17334 (dump_subscript): Adjust.
17335 (finalize_ddr_dependent): Do not dump redundant info.
17336 (analyze_siv_subscript): Adjust.
17337 (subscript_dependence_tester): Likewise.
17338 (compute_affine_dependence): Likewise.
17339
59fd17e3
RB
173402013-01-03 Richard Biener <rguenther@suse.de>
17341
17342 Revert
17343 2013-01-03 Richard Biener <rguenther@suse.de>
17344
17345 PR tree-optimization/55857
17346 * tree-vect-stmts.c (vectorizable_load): Do not setup
17347 re-alignment for invariant loads.
17348
17349 2013-01-02 Richard Biener <rguenther@suse.de>
17350
17351 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 17352 invariant load do not generate a vector load from the scalar location.
59fd17e3 17353
595c2679
RB
173542013-01-03 Richard Biener <rguenther@suse.de>
17355
17356 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
17357 for not vectorizing.
17358 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
17359 not build INDIRECT_REFs, call get_name once only.
17360 (vect_create_data_ref_ptr): Likewise. Dump base object kind
17361 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
17362
90bb7d7a
RB
173632013-01-03 Richard Biener <rguenther@suse.de>
17364
17365 PR tree-optimization/55857
17366 * tree-vect-stmts.c (vectorizable_load): Do not setup
17367 re-alignment for invariant loads.
17368
f09b77ca
RB
173692013-01-03 Richard Biener <rguenther@suse.de>
17370
17371 PR lto/55848
17372 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
17373 prefer a built-in decl.
17374
3461a16e
JJ
173752013-01-03 Jakub Jelinek <jakub@redhat.com>
17376
df8e2b4f
JJ
17377 * gcc.c (process_command): Update copyright notice dates.
17378 * gcov.c (print_version): Likewise.
17379 * gcov-dump.c (print_version): Likewise.
17380
3461a16e
JJ
17381 PR rtl-optimization/55838
17382 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
17383 iv0.step, iv1.step and step.
17384
8b5546d6
JJ
173852013-01-03 Jakub Jelinek <jakub@redhat.com>
17386 Marc Glisse <marc.glisse@inria.fr>
17387
17388 PR tree-optimization/55832
17389 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
17390 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
17391 integer_{one,zero}_node.
17392
8ab1d2e9
JJ
173932013-01-03 Jakub Jelinek <jakub@redhat.com>
17394
17395 PR debug/54402
17396 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
17397 * var-tracking.c (reverse_op): Don't add reverse ops to
17398 VALUEs that have already
0506634a 17399 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 17400
5b9ad1d4
GP
174012013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
17402
17403 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
17404
e25a6711
TJ
174052013-01-02 Teresa Johnson <tejohnson@google.com>
17406
17407 * dumpfile.c (dump_loc): Print filename with location.
17408 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
17409 new location_t parameter to emit complete unroll message with
17410 new dump framework.
17411 (canonicalize_loop_induction_variables): Compute loops location
17412 and pass to try_unroll_loop_completely.
17413 * loop-unroll.c (report_unroll_peel): New function.
17414 (peel_loops_completely): Use new dump format with location
17415 for main dumpfile message, and invoke report_unroll_peel on success.
17416 (decide_unrolling_and_peeling): Ditto.
17417 (decide_peel_once_rolling): Remove old dumpfile message subsumed
17418 by report_unroll_peel.
17419 (decide_peel_completely): Ditto.
17420 (decide_unroll_constant_iterations): Ditto.
17421 (decide_unroll_runtime_iterations): Ditto.
17422 (decide_peel_simple): Ditto.
17423 (decide_unroll_stupid): Ditto.
17424 * cfgloop.c (get_loop_location): New function.
17425 * cfgloop.h (get_loop_location): Declare.
17426
77878621
ST
174272013-01-02 Sriraman Tallam <tmsriram@google.com>
17428
17429 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
17430 NULL.
17431
9e65d03e
JDA
174322013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17433
17434 PR middle-end/55198
17435 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
17436 BLKmode objects when EXPAND_MEMORY is specified.
17437
6a7da30f
ST
174382013-01-02 Sriraman Tallam <tmsriram@google.com>
17439
17440 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
17441 in loop predicate.
17442 (fold_builtin_cpu): Do not share cpu model decls across statements.
17443
e78167f2
JM
174442013-01-02 Jason Merrill <jason@redhat.com>
17445
17446 PR c++/55804
17447 * tree.c (build_array_type_1): Revert earlier change.
17448
8c075fb4
YZ
174492013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
17450
17451 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
17452 "cortex-a57".
17453 * config/aarch64/aarch64-tune.md: Re-generate.
17454
0682ed3e
RB
174552013-01-02 Richard Biener <rguenther@suse.de>
17456
17457 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 17458 invariant load do not generate a vector load from the scalar location.
0682ed3e 17459
b8f6a302
RB
174602013-01-02 Richard Biener <rguenther@suse.de>
17461
17462 PR bootstrap/55784
17463 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
17464 * configure: Regenerate.
17465
04b80dbb
RS
174662013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17467
17468 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
17469 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
17470 (expand_builtin_int_roundingfn_2): Keep the original target around
17471 for the fallback case.
17472
635b0b0c
RS
174732013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17474
17475 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
17476 to be clear for sign changes.
17477
42d57399
JH
174782013-01-01 Jan Hubicka <jh@suse.cz>
17479
17480 * ipa-inline-analysis.c: Fix formatting.
17481
5bb6669d
JJ
174822013-01-01 Jakub Jelinek <jakub@redhat.com>
17483
17484 PR tree-optimization/55831
17485 * tree-vect-loop.c (get_initial_def_for_induction): Use
17486 gsi_after_labels instead of gsi_start_bb.
ad41bd84 17487\f
86a2db33 17488Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM
17489
17490Copying and distribution of this file, with or without modification,
17491are permitted in any medium without royalty provided the copyright
17492notice and this notice are preserved.
This page took 6.947638 seconds and 5 git commands to generate.