]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
projects / thirdparty / gcc.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
re PR target/57735 (ICE with -mtune=xscale (error: could not split insn) when buildin...
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
088c5368
RS		 12013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2
3 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
4 of an rtx/bitpos pair.
5 (store_fixed_bit_field): Update accordingly.
6
69db2d57
RS		 72013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8
9 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
10 GEN_INT.
11 * builtins.c (expand_errno_check): Likewise.
12 * dwarf2cfi.c (init_return_column_size): Likewise.
13 * except.c (sjlj_mark_call_sites): Likewise.
14 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
15 * lra-constraints.c (emit_inc): Likewise.
16 * ree.c (combine_set_extension): Likewise.
17 * regmove.c (fixup_match_2): Likewise.
18 * reload1.c (inc_for_reload): Likewise.
19
69a59f0f
RS		 202013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
21
22 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
23 (force_to_mode, simplify_shift_const_1, simplify_comparison):
24 Use gen_int_mode with the mode of the associated simplify_* call.
25 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
26 * expmed.c (expand_shift_1): Likewise.
27 * function.c (instantiate_virtual_regs_in_insn): Likewise.
28 * loop-iv.c (iv_number_of_iterations): Likewise.
29 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
30 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
31
2f1cd2eb
RS		 322013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
33
34 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
35 of the associated expand_* call.
36 (asan_emit_stack_protection): Likewise.
37 * builtins.c (round_trampoline_addr): Likewise.
38 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
39 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
40 (emit_store_flag): Likewise.
41 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
42 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
43 Likewise.
44 * function.c (instantiate_virtual_regs_in_insn): Likewise.
45 * ifcvt.c (noce_try_store_flag_constants): Likewise.
46 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
47 * modulo-sched.c (generate_prolog_epilog): Likewise.
48 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
49 (expand_ctz, expand_ffs, expand_unop): Likewise.
50
4789c0ce
RS		 512013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
52
53 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
54 of the associated gen_rtx_* call.
55 * caller-save.c (init_caller_save): Likewise.
56 * combine.c (find_split_point, make_extraction): Likewise.
57 (make_compound_operation): Likewise.
58 * dwarf2out.c (mem_loc_descriptor): Likewise.
59 * explow.c (plus_constant, probe_stack_range): Likewise.
60 * expmed.c (expand_mult_const): Likewise.
61 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
62 * reload1.c (init_reload): Likewise.
63 * valtrack.c (cleanup_auto_inc_dec): Likewise.
64 * var-tracking.c (adjust_mems): Likewise.
65 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
66 rather than gen_rtx_fmt_ee.
67
a811f7cb
JH		 682013-09-09 Jan Hubicka <jh@suse.cz>
69
70 PR middle-end/58294
71 * value-prof.c (gimple_ic): Copy also abnormal edges.
72
dcad1dd3
RS		 732013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
74
75 * asan.c (asan_shadow_cst): Use gen_int_mode.
76
daf5c770
JH		 772013-09-08 Jan Hubicka <jh@suse.cz>
78
79 * ipa-profile.c: Add toplevel comment.
80 (ipa_propagate_frequency_1): Be more conservative when profile is read.
81 (contains_hot_call_p): New function.
82 (ipa_propagate_frequency): Set frequencies based on counts when
83 profile is read.
84 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
85 profile; do not tamper with profile after inlining if it is read.
86
a12cd2db
JH		 872013-09-08 Jan Hubicka <jh@suse.cz>
88
89 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
90 speculative edges.
91
856a1588
JH		 922013-09-08 Jan Hubicka <jh@suse.cz>
93
94 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
95 summary generation.
96
4f4b0b73
JL		 972013-09-08 Jeff Law <law@redhat.com>
98
99 PR bootstrap/58340
100 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
101 of 'found'.
102
cc4029ee
AK		 1032013-09-08 Andi Kleen <ak@linux.intel.com>
104
105 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
106
7b395ddd
JH		 1072013-09-08 Jan Hubicka <jh@suse.cz>
108
109 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
110
c34c46dd
RS		 1112013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
112
113 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
114 for non-debug insns.
115 * lra.c (new_insn_reg): Take the containing insn as a parameter.
116 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
117 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
118 accordingly.
119
3462aa02
JH		 1202013-09-08 Jan Hubicka <jh@suse.cz>
121
122 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
123 targets and devirtualize to BUILT_IN_UNREACHABLE.
124 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
125 * ipa.c (walk_polymorphic_call_targets): New function.
126 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
127 functions; use the new timevar.
128 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
129 was removed from the program.
130 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
131 not consider it in the walk when its vtable is dead.
132 (possible_polymorphic_call_targets_1): Pass anonymous flag to
133 record_binfo.
134 (devirt_variable_node_removal_hook): New function.
135 (possible_polymorphic_call_targets): Also register
136 devirt_variable_node_removal_hook.
137 (ipa_devirt): Do not do non-speculative devirtualization.
138 (gate_ipa_devirt): One execute if devirtualizing speculatively.
139
26e5b0fd
JH		 1402013-09-08 Jan Hubicka <jh@suse.cz>
141
142 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
143 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
144 varpool_remove_variable_insertion_hook): Declare.
145 * varpool.c (varpool_node_hook_list): New structure.
146 (first_varpool_node_removal_hook,
147 first_varpool_variable_insertion_hook): New variables.
148 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
149 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
150 varpool_remove_variable_insertion_hook,
151 varpool_call_variable_insertion_hooks): New functions.
152 (varpool_remove_node): Use it.
153
97f1f314 1542013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 155
97f1f314
PC		 156 PR c++/54941
157 * diagnostic.c (diagnostic_build_prefix): When s.file is
158 "<built-in>" don't output line and column numbers.
6df2ca09 159
bcb650cb
JH		 1602013-09-06 Jan Hubicka <jh@suse.cz>
161
162 * cgraphunit.c (expand_thunk): Get body before touching arguments.
163 * lto-streamer-out.c: Stream thunks, too.
164 * lto-streamer-in.c (input_function): Pop cfun here
165 (lto_read_body): Instead of here.
166
87f3fea8
CT		 1672013-09-06 Caroline Tice <cmtice@google.com>
168
169 * doc/install.texi: Add documentation for the --enable-vtable-verify
170 and the --disable-libvtv configure options.
171
5e8d9b45
JL		 1722013-09-06 Jeff Law <law@redhat.com>
173
174 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
175 edge implied equivalences into successor phis.
176
3a001aff
JR		 1772013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
178
179 * resource.c (mark_referenced_resources): Handle COND_EXEC.
180
1812013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ		 182
183 * resource.c (mark_target_live_regs): Compute resources taking
184 into account if a call is predicated or not.
185
2c641f8b
EB		 1862013-09-06 Eric Botcazou <ebotcazou@adacore.com>
187
188 * toplev.c (output_stack_usage): Be prepared for suffixes created by
189 the compiler in the function names.
190
a81b0a3d
JH		 1912013-09-06 Jan Hubicka <jh@suse.cz>
192
193 PR middle-end/58094
194 * ipa-inline.c (has_caller_p): New function.
195 (want_inline_function_to_all_callers_p): Use it.
196 (sum_callers, inline_to_all_callers): Break out from ...
197 (ipa_inline): ... here.
198
68e4ca7e
JH		 1992013-09-06 Jan Hubicka <jh@suse.cz>
200
2c641f8b
EB		 201 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
202 only when AVX is enabled.
68e4ca7e 203
3c73e2fe
JG		 2042013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
205
206 * config/aarch64/aarch64.md
207 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
208 is fpsimd_<load/store>2.
209 (load_pair<mode>): Likewise.
210 (store_pair<mode>): Likewise.
211
8c48eecd
JG		 2122013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
213
214 * config/arm/types.md (type): Add "mrs" type.
215 * config/aarch64/aarch64.md
216 (aarch64_load_tp_hard): Make type "mrs".
217 * config/arm/arm.md
218 (load_tp_hard): Make type "mrs".
219 * config/arm/cortex-a15.md: Update with new attributes.
220 * config/arm/cortex-a5.md: Update with new attributes.
221 * config/arm/cortex-a53.md: Update with new attributes.
222 * config/arm/cortex-a7.md: Update with new attributes.
223 * config/arm/cortex-a8.md: Update with new attributes.
224 * config/arm/cortex-a9.md: Update with new attributes.
225 * config/arm/cortex-m4.md: Update with new attributes.
226 * config/arm/cortex-r4.md: Update with new attributes.
227 * config/arm/fa526.md: Update with new attributes.
228 * config/arm/fa606te.md: Update with new attributes.
229 * config/arm/fa626te.md: Update with new attributes.
230 * config/arm/fa726te.md: Update with new attributes.
231
9ea187a5
JG		 2322013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
233
234 * config/aarch64/aarch64.md
235 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
236 (*movtf_aarch64): Likewise.
237 * config/arm/arm.md
238 (thumb1_movdi_insn): Use "multiple" for type where more than one
239 instruction is used for a move.
240 (*arm32_movhf): Likewise.
241 (*thumb_movdf_insn): Likewise.
242
292b89b3
JG		 2432013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
244
245 * config/arm/types.md (type): Rename fcpys to fmov.
246 * config/arm/vfp.md
247 (*arm_movsi_vfp): Rename type fcpys as fmov.
248 (*thumb2_movsi_vfp): Likewise
249 (*movhf_vfp_neon): Likewise
250 (*movhf_vfp): Likewise
251 (*movsf_vfp): Likewise
252 (*thumb2_movsf_vfp): Likewise
253 (*movsfcc_vfp): Likewise
254 (*thumb2_movsfcc_vfp): Likewise
255 * config/aarch64/aarch64-simd.md
256 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
257 * config/aarch64/aarch64.md
258 (*movsi_aarch64): Replace type mov_reg with fmovs.
259 (*movdi_aarch64): Likewise
260 (*movsf_aarch64): Likewise
261 (*movdf_aarch64): Likewise
262 * config/arm/arm.c
263 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
264 * config/arm/iwmmxt.md
265 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
266 * config/arm/arm1020e.md: Update with new attributes.
267 * config/arm/cortex-a15-neon.md: Update with new attributes.
268 * config/arm/cortex-a5.md: Update with new attributes.
269 * config/arm/cortex-a53.md: Update with new attributes.
270 * config/arm/cortex-a7.md: Update with new attributes.
271 * config/arm/cortex-a8-neon.md: Update with new attributes.
272 * config/arm/cortex-a9.md: Update with new attributes.
273 * config/arm/cortex-m4-fpu.md: Update with new attributes.
274 * config/arm/cortex-r4f.md: Update with new attributes.
275 * config/arm/marvell-pj4.md: Update with new attributes.
276 * config/arm/vfp11.md: Update with new attributes.
277
6a24a83e
JG		 2782013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
279
280 * config/aarch64/aarch64.md
281 (*madd<mode>): Fix type attribute.
282 (*maddsi_uxtw): Likewise.
283 (*msub<mode>): Likewise.
284 (*msubsi_uxtw): Likewise.
285 (<su_optab>maddsidi4): Likewise.
286 (<su_optab>msubsidi4): Likewise.
287
b86923f0
JG		 2882013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
289
290 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
291 * config/arm/arm.md (core_cycles): Remove fdiv.
292 * config/arm/vfp.md:
293 (*sqrtsf2_vfp): Update for attribute changes.
294 (*sqrtdf2_vfp): Likewise.
295 * config/aarch64/aarch64.md:
296 (sqrt<mode>2): Update for attribute changes.
297 * config/arm/arm1020e.md: Update with new attributes.
298 * config/arm/cortex-a15-neon.md: Update with new attributes.
299 * config/arm/cortex-a5.md: Update with new attributes.
300 * config/arm/cortex-a53.md: Update with new attributes.
301 * config/arm/cortex-a7.md: Update with new attributes.
302 * config/arm/cortex-a8-neon.md: Update with new attributes.
303 * config/arm/cortex-a9.md: Update with new attributes.
304 * config/arm/cortex-m4-fpu.md: Update with new attributes.
305 * config/arm/cortex-r4f.md: Update with new attributes.
306 * config/arm/marvell-pj4.md: Update with new attributes.
307 * config/arm/vfp11.md: Update with new attributes.
308
7b49c9e1
JG		 3092013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
310
311 * config/arm/types.md
312 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
313 * config/aarch64/aarch64.md
314 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
315 new attributes.
316 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
317 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
318 (float<GPI:mode><GPF:mode>2): Likewise.
319 * config/arm/vfp.md
320 (*truncsisf2_vfp): Update with new attributes.
321 (*truncsidf2_vfp): Likewise.
322 (fixuns_truncsfsi2): Likewise.
323 (fixuns_truncdfsi2): Likewise.
324 (*floatsisf2_vfp): Likewise.
325 (*floatsidf2_vfp): Likewise.
326 (floatunssisf2): Likewise.
327 (floatunssidf2): Likewise.
328 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
329 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
330 * config/arm/arm1020e.md: Update with new attributes.
331 * config/arm/cortex-a15-neon.md: Update with new attributes.
332 * config/arm/cortex-a5.md: Update with new attributes.
333 * config/arm/cortex-a53.md: Update with new attributes.
334 * config/arm/cortex-a7.md: Update with new attributes.
335 * config/arm/cortex-a8-neon.md: Update with new attributes.
336 * config/arm/cortex-a9.md: Update with new attributes.
337 * config/arm/cortex-m4-fpu.md: Update with new attributes.
338 * config/arm/cortex-r4f.md: Update with new attributes.
339 * config/arm/marvell-pj4.md: Update with new attributes.
340 * config/arm/vfp11.md: Update with new attributes.
341
e7df8af8
JG		 3422013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
343
344 * config/aarch64/arm_neon.h
345 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
346 (vqtbx<1,2,3,4><q>_s8): Likewise.
347
594726e4
JG		 3482013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
349
350 * config/arm/types.md: Add "no_insn", "multiple" and "untyped"
351 types.
352 * config/arm/arm-fixed.md: Add type attribute to all insn
353 patterns.
354 (add<mode>3): Add type attribute.
355 (add<mode>3): Likewise.
356 (usadd<mode>3): Likewise.
357 (ssadd<mode>3): Likewise.
358 (sub<mode>3): Likewise.
359 (sub<mode>3): Likewise.
360 (ussub<mode>3): Likewise.
361 (sssub<mode>3): Likewise.
362 (ssmulsa3): Likewise.
363 (usmulusa3): Likewise.
364 (arm_usatsihi): Likewise.
365 * config/arm/vfp.md
366 (*movdi_vfp): Add types for all instructions.
367 (*movdi_vfp_cortexa8): Likewise.
368 (*movhf_vfp_neon): Likewise.
369 (*movhf_vfp): Likewise.
370 (*movdf_vfp): Likewise.
371 (*thumb2_movdf_vfp): Likewise.
372 (*thumb2_movdfcc_vfp): Likewise.
373 * config/arm/arm.md: Add type attribute to all insn patterns.
374 (*thumb1_adddi3): Add type attribute.
375 (*arm_adddi3): Likewise.
376 (*adddi_sesidi_di): Likewise.
377 (*adddi_zesidi_di): Likewise.
378 (*thumb1_addsi3): Likewise.
379 (addsi3_compare0): Likewise.
380 (*addsi3_compare0_scratch): Likewise.
381 (*compare_negsi_si): Likewise.
382 (cmpsi2_addneg): Likewise.
383 (*addsi3_carryin_<optab>): Likewise.
384 (*addsi3_carryin_alt2_<optab>): Likewise.
385 (*addsi3_carryin_clobercc_<optab>): Likewise.
386 (*subsi3_carryin): Likewise.
387 (*subsi3_carryin_const): Likewise.
388 (*subsi3_carryin_compare): Likewise.
389 (*subsi3_carryin_compare_const): Likewise.
390 (*arm_subdi3): Likewise.
391 (*thumb_subdi3): Likewise.
392 (*subdi_di_zesidi): Likewise.
393 (*subdi_di_sesidi): Likewise.
394 (*subdi_zesidi_di): Likewise.
395 (*subdi_sesidi_di): Likewise.
396 (*subdi_zesidi_ze): Likewise.
397 (thumb1_subsi3_insn): Likewise.
398 (*arm_subsi3_insn): Likewise.
399 (*anddi3_insn): Likewise.
400 (*anddi_zesidi_di): Likewise.
401 (*anddi_sesdi_di): Likewise.
402 (*ne_zeroextracts): Likewise.
403 (*ne_zeroextracts): Likewise.
404 (*ite_ne_zeroextr): Likewise.
405 (*ite_ne_zeroextr): Likewise.
406 (*anddi_notdi_di): Likewise.
407 (*anddi_notzesidi): Likewise.
408 (*anddi_notsesidi): Likewise.
409 (andsi_notsi_si): Likewise.
410 (thumb1_bicsi3): Likewise.
411 (*iordi3_insn): Likewise.
412 (*iordi_zesidi_di): Likewise.
413 (*iordi_sesidi_di): Likewise.
414 (*thumb1_iorsi3_insn): Likewise.
415 (*xordi3_insn): Likewise.
416 (*xordi_zesidi_di): Likewise.
417 (*xordi_sesidi_di): Likewise.
418 (*arm_xorsi3): Likewise.
419 (*andsi_iorsi3_no): Likewise.
420 (*smax_0): Likewise.
421 (*smax_m1): Likewise.
422 (*arm_smax_insn): Likewise.
423 (*smin_0): Likewise.
424 (*arm_smin_insn): Likewise.
425 (*arm_umaxsi3): Likewise.
426 (*arm_uminsi3): Likewise.
427 (*minmax_arithsi): Likewise.
428 (*minmax_arithsi_): Likewise.
429 (*satsi_<SAT:code>): Likewise.
430 (arm_ashldi3_1bit): Likewise.
431 (arm_ashrdi3_1bit): Likewise.
432 (arm_lshrdi3_1bit): Likewise.
433 (*arm_negdi2): Likewise.
434 (*thumb1_negdi2): Likewise.
435 (*arm_negsi2): Likewise.
436 (*thumb1_negsi2): Likewise.
437 (*negdi_extendsid): Likewise.
438 (*negdi_zero_extend): Likewise.
439 (*arm_abssi2): Likewise.
440 (*thumb1_abssi2): Likewise.
441 (*arm_neg_abssi2): Likewise.
442 (*thumb1_neg_abss): Likewise.
443 (one_cmpldi2): Likewise.
444 (extend<mode>di2): Likewise.
445 (*compareqi_eq0): Likewise.
446 (*arm_extendhisi2addsi): Likewise.
447 (*arm_movdi): Likewise.
448 (*thumb1_movdi_insn): Likewise.
449 (*arm_movt): Likewise.
450 (*thumb1_movsi_insn): Likewise.
451 (pic_add_dot_plus_four): Likewise.
452 (pic_add_dot_plus_eight): Likewise.
453 (tls_load_dot_plus_eight): Likewise.
454 (*thumb1_movhi_insn): Likewise.
455 (*thumb1_movsf_insn): Likewise.
456 (*movdf_soft_insn): Likewise.
457 (*thumb_movdf_insn): Likewise.
458 (cbranchsi4_insn): Likewise.
459 (cbranchsi4_scratch): Likewise.
460 (*negated_cbranchsi4): Likewise.
461 (*tbit_cbranch): Likewise.
462 (*tlobits_cbranch): Likewise.
463 (*tstsi3_cbranch): Likewise.
464 (*cbranchne_decr1): Likewise.
465 (*addsi3_cbranch): Likewise.
466 (*addsi3_cbranch_scratch): Likewise.
467 (*arm_cmpdi_insn): Likewise.
468 (*arm_cmpdi_unsig): Likewise.
469 (*arm_cmpdi_zero): Likewise.
470 (*thumb_cmpdi_zero): Likewise.
471 (*deleted_compare): Likewise.
472 (*mov_scc): Likewise.
473 (*mov_negscc): Likewise.
474 (*mov_notscc): Likewise.
475 (*cstoresi_eq0_thumb1_insn): Likewise.
476 (cstoresi_nltu_thumb1): Likewise.
477 (cstoresi_ltu_thu): Likewise.
478 (thumb1_addsi3_addgeu): Likewise.
479 (*arm_jump): Likewise.
480 (*thumb_jump): Likewise.
481 (*check_arch2): Likewise.
482 (arm_casesi_internal): Likewise.
483 (thumb1_casesi_dispatch): Likewise.
484 (*arm_indirect_jump): Likewise.
485 (*thumb1_indirect_jump): Likewise.
486 (nop): Likewise.
487 (*and_scc): Likewise.
488 (*ior_scc): Likewise.
489 (*compare_scc): Likewise.
490 (*cond_move): Likewise.
491 (*cond_arith): Likewise.
492 (*cond_sub): Likewise.
493 (*cmp_ite0): Likewise.
494 (*cmp_ite1): Likewise.
495 (*cmp_and): Likewise.
496 (*cmp_ior): Likewise.
497 (*ior_scc_scc): Likewise.
498 (*ior_scc_scc_cmp): Likewise.
499 (*and_scc_scc): Likewise.
500 (*and_scc_scc_cmp): Likewise.
501 (*and_scc_scc_nod): Likewise.
502 (*negscc): Likewise.
503 (movcond_addsi): Likewise.
504 (movcond): Likewise.
505 (*ifcompare_plus_move): Likewise.
506 (*if_plus_move): Likewise.
507 (*ifcompare_move_plus): Likewise.
508 (*if_move_plus): Likewise.
509 (*ifcompare_arith_arith): Likewise.
510 (*if_arith_arith): Likewise.
511 (*ifcompare_arith_move): Likewise.
512 (*if_arith_move): Likewise.
513 (*ifcompare_move_arith): Likewise.
514 (*if_move_arith): Likewise.
515 (*ifcompare_move_not): Likewise.
516 (*if_move_not): Likewise.
517 (*ifcompare_not_move): Likewise.
518 (*if_not_move): Likewise.
519 (*ifcompare_shift_move): Likewise.
520 (*if_shift_move): Likewise.
521 (*ifcompare_move_shift): Likewise.
522 (*if_move_shift): Likewise.
523 (*ifcompare_shift_shift): Likewise.
524 (*ifcompare_not_arith): Likewise.
525 (*ifcompare_arith_not): Likewise.
526 (*if_arith_not): Likewise.
527 (*ifcompare_neg_move): Likewise.
528 (*if_neg_move): Likewise.
529 (*ifcompare_move_neg): Likewise.
530 (*if_move_neg): Likewise.
531 (prologue_thumb1_interwork): Likewise.
532 (*cond_move_not): Likewise.
533 (*sign_extract_onebit): Likewise.
534 (*not_signextract_onebit): Likewise.
535 (stack_tie): Likewise.
536 (align_4): Likewise.
537 (align_8): Likewise.
538 (consttable_end): Likewise.
539 (consttable_1): Likewise.
540 (consttable_2): Likewise.
541 (consttable_4): Likewise.
542 (consttable_8): Likewise.
543 (consttable_16): Likewise.
544 (*thumb1_tablejump): Likewise.
545 (prefetch): Likewise.
546 (force_register_use): Likewise.
547 (thumb_eh_return): Likewise.
548 (load_tp_hard): Likewise.
549 (load_tp_soft): Likewise.
550 (tlscall): Likewise.
551 (*arm_movtas_ze): Likewise.
552 (*arm_rev): Likewise.
553 (*arm_revsh): Likewise.
554 (*arm_rev16): Likewise.
555 * config/arm/thumb2.md
556 (*thumb2_smaxsi3): Likewise.
557 (*thumb2_sminsi3): Likewise.
558 (*thumb32_umaxsi3): Likewise.
559 (*thumb2_uminsi3): Likewise.
560 (*thumb2_negdi2): Likewise.
561 (*thumb2_abssi2): Likewise.
562 (*thumb2_neg_abss): Likewise.
563 (*thumb2_movsi_insn): Likewise.
564 (tls_load_dot_plus_four): Likewise.
565 (*thumb2_movhi_insn): Likewise.
566 (*thumb2_mov_scc): Likewise.
567 (*thumb2_mov_negs): Likewise.
568 (*thumb2_mov_negs): Likewise.
569 (*thumb2_mov_nots): Likewise.
570 (*thumb2_mov_nots): Likewise.
571 (*thumb2_movsicc_): Likewise.
572 (*thumb2_movsfcc_soft_insn): Likewise.
573 (*thumb2_indirect_jump): Likewise.
574 (*thumb2_and_scc): Likewise.
575 (*thumb2_ior_scc): Likewise.
576 (*thumb2_ior_scc_strict_it): Likewise.
577 (*thumb2_cond_move): Likewise.
578 (*thumb2_cond_arith): Likewise.
579 (*thumb2_cond_ari): Likewise.
580 (*thumb2_cond_sub): Likewise.
581 (*thumb2_negscc): Likewise.
582 (*thumb2_movcond): Likewise.
583 (thumb2_casesi_internal): Likewise.
584 (thumb2_casesi_internal_pic): Likewise.
585 (*thumb2_alusi3_short): Likewise.
586 (*thumb2_mov<mode>_shortim): Likewise.
587 (*thumb2_addsi_short): Likewise.
588 (*thumb2_subsi_short): Likewise.
589 (thumb2_addsi3_compare0): Likewise.
590 (*thumb2_cbz): Likewise.
591 (*thumb2_cbnz): Likewise.
592 (*thumb2_one_cmplsi2_short): Likewise.
593 (*thumb2_negsi2_short): Likewise.
594 (*orsi_notsi_si): Likewise.
595 * config/arm/arm1020e.md: Update with new attributes.
596 * config/arm/arm1026ejs.md: Update with new attributes.
597 * config/arm/arm1136jfs.md: Update with new attributes.
598 * config/arm/arm926ejs.md: Update with new attributes.
599 * config/arm/cortex-a15.md: Update with new attributes.
600 * config/arm/cortex-a5.md: Update with new attributes.
601 * config/arm/cortex-a53.md: Update with new attributes.
602 * config/arm/cortex-a7.md: Update with new attributes.
603 * config/arm/cortex-a8.md: Update with new attributes.
604 * config/arm/cortex-a9.md: Update with new attributes.
605 * config/arm/cortex-m4.md: Update with new attributes.
606 * config/arm/cortex-r4.md: Update with new attributes.
607 * config/arm/fa526.md: Update with new attributes.
608 * config/arm/fa606te.md: Update with new attributes.
609 * config/arm/fa626te.md: Update with new attributes.
610 * config/arm/fa726te.md: Update with new attributes.
611
1c83b673
JG		 6122013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
613
614 * config/aarch64/aarch64-simd.md
615 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
616 <vwx> iterator to ensure correct register choice.
617 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
618 (aarch64_sqdmull_n<mode>): Likewise.
619 (aarch64_sqdmull2_n<mode>_internal): Likewise.
620 * config/aarch64/arm_neon.h
621 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
622 (vml<as><q>_n_<su>16): Likewise.
623 (vml<as>l_high_lane<q>_<su>16): Likewise.
624 (vml<as>l_high_n_<su>16): Likewise.
625 (vml<as>l_lane<q>_<su>16): Likewise.
626 (vml<as>l_n_<su>16): Likewise.
627 (vmul<q>_lane<q>_<su>16): Likewise.
628 (vmul<q>_n_<su>16): Likewise.
629 (vmull_lane<q>_<su>16): Likewise.
630 (vmull_n_<su>16): Likewise.
631 (vmull_high_lane<q>_<su>16): Likewise.
632 (vmull_high_n_<su>16): Likewise.
633 (vqrdmulh<q>_n_s16): Likewise.
634
bb1ae543
TB		 6352013-09-06 Tejas Belagod <tejas.belagod@arm.com>
636
637 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
638 have the correct lane parameter.
639
f23c0742
RB		 6402013-09-06 Richard Biener <rguenther@suse.de>
641
642 * cfganal.c (control_dependences::~control_dependences):
643 Properly free all of the vector.
644
c1d49770
KY		 6452013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
646
647 PR target/58269
648 * config/i386/i386.c (ix86_conditional_register_usage):
649 Proper initialize extended SSE registers.
650
6da47f52
JH		 6512013-09-06 Jan Hubicka <jh@suse.cz>
652
653 PR tree-optimization/58311
654 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
655
6562013-09-06 Jan Hubicka <jh@suse.cz>
657
658 * Makefile.in (tree-sra.o): Update dependencies.
659 * tree-sra.c: Include ipa-utils.h
660 (scan_function): Use recursive_call_p.
661 (has_caller_p): New function.
662 (cgraph_for_node_and_aliases): Count also callers of aliases.
663
fc11f321
JH		 6642013-09-06 Jan Hubicka <jh@suse.cz>
665
666 PR middle-end/58094
667 * cgraph.h (symtab_semantically_equivalent_p): Declare.
668 * tree-tailcall.c: Include ipa-utils.h.
669 (find_tail_calls): Use it.
670 * ipa-pure-const.c (check_call): Likewise.
671 * ipa-utils.c (recursive_call_p): New function.
672 * ipa-utils.h (recursive_call_p): Dclare.
673 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
674 (symtab_semantically_equivalent_p): New function.
675 * Makefile.in (tree-tailcall.o): Update dependencies.
676
9a6af450
EB		 6772013-09-06 Eric Botcazou <ebotcazou@adacore.com>
678
679 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
680 non-inlinable part.
681
35f32ad4
RB		 6822013-09-06 Richard Biener <rguenther@suse.de>
683
684 * lto-streamer.h (lto_global_var_decls): Remove.
685 * Makefile.in (OBJS): Remove lto-symtab.o.
686 (lto-symtab.o): Remove.
687 (GTFILES): Remove lto-symtab.c
688 * lto-symtab.c: Move to lto/
689
d12a76f3
AK		 6902013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
691
692 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
693 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
694 (UNSPEC_FPINT_RINT): New constant definitions.
695 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
696 definition with 2 attributes.
697 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
698 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
699 definitions.
700
7012013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
702
703 * config/s390/s390.md: Add "bcr_flush" value to mnemonic
704 attribute.
705 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
706 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
707 attribute to "z196_alone".
708 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
709 "zEC12_simple".
710
c8e9d8c3
RB		 7112013-09-06 Richard Biener <rguenther@suse.de>
712
713 * basic-block.h (class control_dependences): New.
714 * tree-ssa-dce.c (control_dependence_map): Remove.
715 (cd): New global.
716 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
717 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
718 find_pdom, find_control_dependence, find_all_control_dependences):
719 Move to cfganal.c.
720 (mark_control_dependent_edges_necessary, find_obviously_necessary_stmts,
721 propagate_necessity, tree_dce_init, tree_dce_done,
722 perform_tree_ssa_dce): Adjust.
723 * cfganal.c (set_control_dependence_map_bit,
724 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
725 find_all_control_dependences): Move from tree-ssa-dce.c and
726 implement as methods of control_dependences class.
727 (control_dependences::control_dependences): New.
728 (control_dependences::~control_dependences): Likewise.
729 (control_dependences::get_edges_dependent_on): Likewise.
730 (control_dependences::get_edge): Likewise.
731
7d2ba471
JH		 7322013-09-04 Jan Hubicka <jh@suse.cz>
733
734 * tree.c (types_same_for_odr): Drop overactive check.
735 * ipa-devirt.c (hash_type_name): Likewise.
736
e18412fc
JH		 7372013-09-04 Jan Hubicka <jh@suse.cz>
738
739 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
740 (analyze_functions): ... here.
741
d352b245
JH		 7422013-09-04 Jan Hubicka <jh@suse.cz>
743
744 PR middle-end/58201
745 * cgraphunit.c (analyze_functions): Clear AUX fields
746 after processing; initialize assembler name has.
747
770da076
JL		 7482013-09-05 Jeff Law <law@redhat.com>
749
750 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
751 from thread_around_empty_block. Record threading path into PATH.
752 Recurse if threading through the initial block is successful.
753 (thread_across_edge): Corresponding changes to slightly simplify.
754
fe6f68e2
JG		 7552013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
756
757 * config/aarch64/aarch64.md
758 (type): Remove frecpe, frecps, frecpx.
759 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
760 fix to be a TARGET_SIMD instruction.
761 (aarch64_frecps): Remove.
762 * config/aarch64/aarch64-simd.md
763 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
764 (aarch64_frecps<mode>): Handle all float/vector of float modes.
765
6e4150e1
JG		 7662013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
767 Sofiane Naci <sofiane.naci@arm.com>
768
769 * config/arm/types.md (define_attr "type"):
770 Expand "arlo_imm"
771 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
772 Expand "arlo_reg"
773 into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm", "alu_ext",
774 "alu_reg", "alus_ext", "alus_reg", "bfm", "csel", "logic_reg",
775 "logics_reg", "rev".
776 Expand "arlo_shift"
777 into "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
778 "logics_shift_imm".
779 Expand "arlo_shift_reg"
780 into "alu_shift_reg", "alus_shift_reg", "logic_shift_reg",
781 "logics_shift_reg".
782 Expand "clz" into "clz, "rbit".
783 Rename "shift" to "shift_imm".
784 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
785 changes.
786 Update for attribute changes all occurrences of arlo_* and
787 shift* types.
788 * config/arm/arm-fixed.md: Update for attribute changes
789 all occurrences of arlo_* types.
790 * config/arm/thumb2.md: Update for attribute changes all occurrences
791 of arlo_* types.
792 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
793 (cortexa7_older_only): Likewise.
794 (cortexa7_younger): Likewise.
795 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
796 (1020alu_shift_op): Likewise.
797 (1020alu_shift_reg_op): Likewise.
798 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
799 (alu_shift_op): Likewise.
800 (alu_shift_reg_op): Likewise.
801 * config/arm/arm1136jfs.md (11_alu_op): Update for
802 attribute changes.
803 (11_alu_shift_op): Likewise.
804 (11_alu_shift_reg_op): Likewise.
805 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
806 (9_alu_shift_reg_op): Likewise.
807 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
808 attribute changes.
809 (cortex_a15_alu_shift): Likewise.
810 (cortex_a15_alu_shift_reg): Likewise.
811 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
812 attribute changes.
813 (cortex_a5_alu_shift): Likewise.
814 * config/arm/cortex-a53.md
815 (cortex_a53_alu): Update for attribute changes.
816 (cortex_a53_alu_shift): Likewise.
817 * config/arm/cortex-a7.md
818 (cortex_a7_alu_imm): Update for attribute changes.
819 (cortex_a7_alu_reg): Likewise.
820 (cortex_a7_alu_shift): Likewise.
821 * config/arm/cortex-a8.md
822 (cortex_a8_alu): Update for attribute changes.
823 (cortex_a8_alu_shift): Likewise.
824 (cortex_a8_alu_shift_reg): Likewise.
825 * config/arm/cortex-a9.md
826 (cortex_a9_dp): Update for attribute changes.
827 (cortex_a9_dp_shift): Likewise.
828 * config/arm/cortex-m4.md
829 (cortex_m4_alu): Update for attribute changes.
830 * config/arm/cortex-r4.md
831 (cortex_r4_alu): Update for attribute changes.
832 (cortex_r4_mov): Likewise.
833 (cortex_r4_alu_shift_reg): Likewise.
834 * config/arm/fa526.md
835 (526_alu_op): Update for attribute changes.
836 (526_alu_shift_op): Likewise.
837 * config/arm/fa606te.md
838 (606te_alu_op): Update for attribute changes.
839 * config/arm/fa626te.md
840 (626te_alu_op): Update for attribute changes.
841 (626te_alu_shift_op): Likewise.
842 * config/arm/fa726te.md
843 (726te_alu_op): Update for attribute changes.
844 (726te_alu_shift_op): Likewise.
845 (726te_alu_shift_reg_op): Likewise.
846 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
847 (mp626_alu_shift_op): Likewise.
848 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
849 (pj4_alu_conds): Likewise.
850 (pj4_shift): Likewise.
851 (pj4_shift_conds): Likewise.
852 (pj4_alu_shift): Likewise.
853 (pj4_alu_shift_conds): Likewise.
854 * config/aarch64/aarch64.md: Update for attribute change
855 all occurrences of arlo_* and shift* types.
856
e6523306
MS		 8572013-09-05 Mike Stump <mikestump@comcast.net>
858
859 * tree.h: Move documentation for tree_function_decl to tree-core.h
860 with the declaration.
861
80a18161
PB		 8622013-09-05 Peter Bergner <bergner@vnet.ibm.com>
863
864 PR target/58139
865 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
866 looking for widest mode.
867
6994430a
EB		 8682013-09-05 Eric Botcazou <ebotcazou@adacore.com>
869
870 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
871
e67f39f7
RB		 8722013-09-05 Richard Biener <rguenther@suse.de>
873
874 PR tree-optimization/58137
875 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
876 Do not create vectors of pointers.
877 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
878 types for the components of the vector initializer.
879 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
880 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
881
568cda29
MJ		 8822013-09-05 Martin Jambor <mjambor@suse.cz>
883
884 * ipa-prop.c (remove_described_reference): Accept missing references,
885 return false if that hppens, otherwise return true.
886 (cgraph_node_for_jfunc): New function.
887 (try_decrement_rdesc_refcount): Likewise.
888 (try_make_edge_direct_simple_call): Use them.
889 (ipa_edge_removal_hook): Remove references from rdescs.
890 (ipa_edge_duplication_hook): Clone rdescs and their references
891 when the new edge has the same caller as the old one.
892 * cgraph.c (cgraph_resolve_speculation): Remove speculative
893 reference before removing any edges.
894
5a200acb
RE		 8952013-09-05 Richard Earnshaw <rearnsha@arm.com>
896
897 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
898 initial store.
899 * thumb2.md (thumb2_storewb_parisi): New pattern.
900
5922847b
YZ		 9012013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
902
903 * config/aarch64/aarch64-option-extensions.def: Add
904 AARCH64_OPT_EXTENSION of 'crc'.
905 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
906 (AARCH64_ISA_CRC): Ditto.
907 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
908 description of the CRC extension.
909
996746aa
AI		 9102013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
911
912 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
913 * config/rs6000/linux.h: Ditto.
914 * alpha/linux.h: Ditto.
915 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
916 no_c99_libc_has_function.
917 * config/c6x/uclinux-elf.h: Ditto.
918 * config/lm32/uclinux-elf.h: Ditto.
919 * config/m68k/uclinux.h: Ditto.
920 * config/moxie/uclinux.h: Ditto.
921 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
922 (crisv32-*-linux*, cris-*-linux*): Ditto.
923 * config/bfin/bfin.c: Include "tm_p.h".
924
c370d99f
RB		 9252013-09-05 Richard Biener <rguenther@suse.de>
926
927 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
928 check for a definition without a basic-block.
929
003bb7f3
JG		 9302013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
931 Sofiane Naci <sofiane.naci@arm.com>
932
933 * config/aarch64/aarch64.md
934 (*movti_aarch64): Rename r_2_f and f_2_r.
935 (*movsf_aarch64): Likewise.
936 (*movdf_aarch64): Likewise.
937 (*movtf_aarch64): Likewise.
938 (aarch64_movdi_<mode>low): Likewise.
939 (aarch64_movdi_<mode>high): Likewise.
940 (aarch64_mov<mode>high_di): Likewise.
941 (aarch64_mov<mode>low_di): Likewise.
942 (aarch64_movtilow_tilow): Likewise.
943 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
944 values to config/arm/types.md
945 (attribute "conds"): Update for attribute change.
946 (anddi3_insn): Likewise.
947 (iordi3_insn): Likewise.
948 (xordi3_insn): Likewise.
949 (one_cmpldi2): Likewise.
950 * config/arm/types.md (type): Add Neon types.
951 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
952 use "type" attribute.
953 (movmisalign<mode>_neon_store): Likewise.
954 (movmisalign<mode>_neon_load): Likewise.
955 (vec_set<mode>_internal): Likewise.
956 (vec_setv2di_internal): Likewise.
957 (vec_extract<mode>): Likewise.
958 (vec_extractv2di): Likewise.
959 (add<mode>3_neon): Likewise.
960 (adddi3_neon): Likewise.
961 (sub<mode>3_neon): Likewise.
962 (subdi3_neon): Likewise.
963 (mul<mode>3_neon): Likewise.
964 (mul<mode>3add<mode>_neon): Likewise.
965 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
966 (fma<VCVTF:mode>4)): Likewise.
967 (fma<VCVTF:mode>4_intrinsic): Likewise.
968 (fmsub<VCVTF:mode>4)): Likewise.
969 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
970 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
971 (ior<mode>3): Likewise.
972 (and<mode>3): Likewise.
973 (anddi3_neon): Likewise.
974 (orn<mode>3_neon): Likewise.
975 (orndi3_neon): Likewise.
976 (bic<mode>3_neon): Likewise.
977 (bicdi3_neon): Likewise.
978 (xor<mode>3): Likewise.
979 (one_cmpl<mode>2): Likewise.
980 (abs<mode>2): Likewise.
981 (neg<mode>2): Likewise.
982 (umin<mode>3_neon): Likewise.
983 (umax<mode>3_neon): Likewise.
984 (smin<mode>3_neon): Likewise.
985 (smax<mode>3_neon): Likewise.
986 (vashl<mode>3): Likewise.
987 (vashr<mode>3_imm): Likewise.
988 (vlshr<mode>3_imm): Likewise.
989 (ashl<mode>3_signed): Likewise.
990 (ashl<mode>3_unsigned): Likewise.
991 (neon_load_count): Likewise.
992 (ashldi3_neon_noclobber): Likewise.
993 (signed_shift_di3_neon): Likewise.
994 (unsigned_shift_di3_neon): Likewise.
995 (ashrdi3_neon_imm_noclobber): Likewise.
996 (lshrdi3_neon_imm_noclobber): Likewise.
997 (widen_ssum<mode>3): Likewise.
998 (widen_usum<mode>3): Likewise.
999 (quad_halves_<code>v4si): Likewise.
1000 (quad_halves_<code>v4sf): Likewise.
1001 (quad_halves_<code>v8hi): Likewise.
1002 (quad_halves_<code>v16qi): Likewise.
1003 (reduc_splus_v2di): Likewise.
1004 (neon_vpadd_internal<mode>): Likewise.
1005 (neon_vpsmin<mode>): Likewise.
1006 (neon_vpsmax<mode>): Likewise.
1007 (neon_vpumin<mode>): Likewise.
1008 (neon_vpumax<mode>): Likewise.
1009 (ss_add<mode>_neon): Likewise.
1010 (us_add<mode>_neon): Likewise.
1011 (ss_sub<mode>_neon): Likewise.
1012 (us_sub<mode>_neon): Likewise.
1013 (neon_vadd<mode>_unspec): Likewise.
1014 (neon_vaddl<mode>): Likewise.
1015 (neon_vaddw<mode>): Likewise.
1016 (neon_vhadd<mode>): Likewise.
1017 (neon_vqadd<mode>): Likewise.
1018 (neon_vaddhn<mode>): Likewise.
1019 (neon_vmul<mode>): Likewise.
1020 (neon_vmla<mode>): Likewise.
1021 (neon_vmlal<mode>): Likewise.
1022 (neon_vmls<mode>): Likewise.
1023 (neon_vmlsl<mode>): Likewise.
1024 (neon_vqdmulh<mode>): Likewise.
1025 (neon_vqdmlal<mode>): Likewise.
1026 (neon_vqdmlsl<mode>): Likewise.
1027 (neon_vmull<mode>): Likewise.
1028 (neon_vqdmull<mode>): Likewise.
1029 (neon_vsub<mode>_unspec): Likewise.
1030 (neon_vsubl<mode>): Likewise.
1031 (neon_vsubw<mode>): Likewise.
1032 (neon_vqsub<mode>): Likewise.
1033 (neon_vhsub<mode>): Likewise.
1034 (neon_vsubhn<mode>): Likewise.
1035 (neon_vceq<mode>): Likewise.
1036 (neon_vcge<mode>): Likewise.
1037 (neon_vcgeu<mode>): Likewise.
1038 (neon_vcgt<mode>): Likewise.
1039 (neon_vcgtu<mode>): Likewise.
1040 (neon_vcle<mode>): Likewise.
1041 (neon_vclt<mode>): Likewise.
1042 (neon_vcage<mode>): Likewise.
1043 (neon_vcagt<mode>): Likewise.
1044 (neon_vtst<mode>): Likewise.
1045 (neon_vabd<mode>): Likewise.
1046 (neon_vabdl<mode>): Likewise.
1047 (neon_vaba<mode>): Likewise.
1048 (neon_vabal<mode>): Likewise.
1049 (neon_vmax<mode>): Likewise.
1050 (neon_vmin<mode>): Likewise.
1051 (neon_vpaddl<mode>): Likewise.
1052 (neon_vpadal<mode>): Likewise.
1053 (neon_vpmax<mode>): Likewise.
1054 (neon_vpmin<mode>): Likewise.
1055 (neon_vrecps<mode>): Likewise.
1056 (neon_vrsqrts<mode>): Likewise.
1057 (neon_vqabs<mode>): Likewise.
1058 (neon_vqneg<mode>): Likewise.
1059 (neon_vcls<mode>): Likewise.
1060 (clz<mode>2): Likewise.
1061 (popcount<mode>2): Likewise.
1062 (neon_vrecpe): Likewise.
1063 (neon_vrsqrte): Likewise.
1064 (neon_vget_lane<mode>_sext_internal): Likewise.
1065 (neon_vget_lane<mode>_zext_internal): Likewise.
1066 (neon_vdup_n<mode>): Likewise.
1067 (neon_vdup_nv2di): Likewise.
1068 (neon_vdpu_lane<mode>_internal): Likewise.
1069 (neon_vswp<mode>): Likewise.
1070 (float<mode><V_cvtto>2): Likewise.
1071 (floatuns<mode><V_cvtto>2): Likewise.
1072 (fix_trunc<mode><V_cvtto>)2): Likewise
1073 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1074 (neon_vcvt<mode>): Likewise.
1075 (neon_vcvtv4sfv4hf): Likewise.
1076 (neon_vcvtv4hfv4sf): Likewise.
1077 (neon_vcvt_n<mode>): Likewise.
1078 (neon_vmovn<mode>): Likewise.
1079 (neon_vqmovn<mode>): Likewise.
1080 (neon_vqmovun<mode>): Likewise.
1081 (neon_vmovl<mode>): Likewise.
1082 (neon_vmul_lane<mode>): Likewise.
1083 (neon_vmull_lane<mode>): Likewise.
1084 (neon_vqdmull_lane<mode>): Likewise.
1085 (neon_vqdmulh_lane<mode>): Likewise.
1086 (neon_vmla_lane<mode>): Likewise.
1087 (neon_vmlal_lane<mode>): Likewise.
1088 (neon_vqdmlal_lane<mode>): Likewise.
1089 (neon_vmls_lane<mode>): Likewise.
1090 (neon_vmlsl_lane<mode>): Likewise.
1091 (neon_vqdmlsl_lane<mode>): Likewise.
1092 (neon_vext<mode>): Likewise.
1093 (neon_vrev64<mode>): Likewise.
1094 (neon_vrev32<mode>): Likewise.
1095 (neon_vrev16<mode>): Likewise.
1096 (neon_vbsl<mode>_internal): Likewise.
1097 (neon_vshl<mode>): Likewise.
1098 (neon_vqshl<mode>): Likewise.
1099 (neon_vshr_n<mode>): Likewise.
1100 (neon_vshrn_n<mode>): Likewise.
1101 (neon_vqshrn_n<mode>): Likewise.
1102 (neon_vqshrun_n<mode>): Likewise.
1103 (neon_vshl_n<mode>): Likewise.
1104 (neon_vqshl_n<mode>): Likewise.
1105 (neon_vqshlu_n<mode>): Likewise.
1106 (neon_vshll_n<mode>): Likewise.
1107 (neon_vsra_n<mode>): Likewise.
1108 (neon_vsri_n<mode>): Likewise.
1109 (neon_vsli_n<mode>): Likewise.
1110 (neon_vtbl1v8qi): Likewise.
1111 (neon_vtbl2v8qi): Likewise.
1112 (neon_vtbl3v8qi): Likewise.
1113 (neon_vtbl4v8qi): Likewise.
1114 (neon_vtbx1v8qi): Likewise.
1115 (neon_vtbx2v8qi): Likewise.
1116 (neon_vtbx3v8qi): Likewise.
1117 (neon_vtbx4v8qi): Likewise.
1118 (neon_vtrn<mode>_internal): Likewise.
1119 (neon_vzip<mode>_internal): Likewise.
1120 (neon_vuzp<mode>_internal): Likewise.
1121 (neon_vld1<mode>): Likewise.
1122 (neon_vld1_lane<mode>): Likewise.
1123 (neon_vld1_dup<mode>): Likewise.
1124 (neon_vld1_dupv2di): Likewise.
1125 (neon_vst1<mode>): Likewise.
1126 (neon_vst1_lane<mode>): Likewise.
1127 (neon_vld2<mode>): Likewise.
1128 (neon_vld2_lane<mode>): Likewise.
1129 (neon_vld2_dup<mode>): Likewise.
1130 (neon_vst2<mode>): Likewise.
1131 (neon_vst2_lane<mode>): Likewise.
1132 (neon_vld3<mode>): Likewise.
1133 (neon_vld3qa<mode>): Likewise.
1134 (neon_vld3qb<mode>): Likewise.
1135 (neon_vld3_lane<mode>): Likewise.
1136 (neon_vld3_dup<mode>): Likewise.
1137 (neon_vst3<mode>): Likewise.
1138 (neon_vst3qa<mode>): Likewise.
1139 (neon_vst3qb<mode>): Likewise.
1140 (neon_vst3_lane<mode>): Likewise.
1141 (neon_vld4<mode>): Likewise.
1142 (neon_vld4qa<mode>): Likewise.
1143 (neon_vld4qb<mode>): Likewise.
1144 (neon_vld4_lane<mode>): Likewise.
1145 (neon_vld4_dup<mode>): Likewise.
1146 (neon_vst4<mode>): Likewise.
1147 (neon_vst4qa<mode>): Likewise.
1148 (neon_vst4qb<mode>): Likewise.
1149 (neon_vst4_lane<mode>): Likewise.
1150 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1151 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1152 (neon_vec_<US>mult_lo_<mode>): Likewise.
1153 (neon_vec_<US>mult_hi_<mode>): Likewise.
1154 (neon_vec_<US>shiftl_<mode>): Likewise.
1155 (neon_unpack<US>_<mode>): Likewise.
1156 (neon_vec_<US>mult_<mode>): Likewise.
1157 (vec_pack_trunc_<mode>): Likewise.
1158 (neon_vec_pack_trunk_<mode>): Likewise.
1159 (neon_vabd<mode>_2): Likewise.
1160 (neon_vabd<mode>_3): Likewise.
1161 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1162 (thumb2_movsi_vfp): Likewise.
1163 (movdi_vfp): Likewise.
1164 (movdi_vfp_cortexa8): Likewise.
1165 (movhf_vfp_neon): Likewise.
1166 (movhf_vfp): Likewiwse.
1167 (movsf_vfp): Likewiwse.
1168 (thumb2_movsf_vfp): Likewiwse.
1169 (movdf_vfp): Likewise.
1170 (thumb2_movdf_vfp): Likewise.
1171 (movsfcc_vfp): Likewise.
1172 (thumb2_movsfcc_vfp): Likewise.
1173 (movdfcc_vfp): Likewise.
1174 (thumb2_movdfcc_vfp): Likewise.
1175 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1176 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1177 (v10_v2c): Likewise.
1178 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1179 attribute change.
1180 (cortex_a15_neon_int_2): Likewise.
1181 (cortex_a15_neon_int_3): Likewise.
1182 (cortex_a15_neon_int_4): Likewise.
1183 (cortex_a15_neon_int_5): Likewise.
1184 (cortex_a15_neon_vqneg_vqabs): Likewise.
1185 (cortex_a15_neon_vmov): Likewise.
1186 (cortex_a15_neon_vaba): Likewise.
1187 (cortex_a15_neon_vaba_qqq): Likewise.
1188 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1189 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1190 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1191 scalar_64_32_long_scalar): Likewise.
1192 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1193 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1194 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1195 lotype_qdd_64_32_long): Likewise.
1196 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1197 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1198 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1199 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1200 (cortex_a15_neon_shift_1): Likewise.
1201 (cortex_a15_neon_shift_2): Likewise.
1202 (cortex_a15_neon_shift_3): Likewise.
1203 (cortex_a15_neon_vshl_ddd): Likewise.
1204 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1205 (cortex_a15_neon_vsra_vrsra): Likewise.
1206 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1207 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1208 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1209 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1210 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1211 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1212 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1213 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1214 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1215 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1216 (cortex_a15_neon_bp_simple): Likewise.
1217 (cortex_a15_neon_bp_2cycle): Likewise.
1218 (cortex_a15_neon_bp_3cycle): Likewise.
1219 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1220 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1221 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1222 (cortex_a15_neon_vld2_4_regs): Likewise.
1223 (cortex_a15_neon_vld3_vld4): Likewise.
1224 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1225 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1226 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1227 (cortex_a15_neon_vst3_vst4): Likewise.
1228 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1229 (cortex_a15_neon_vld3_vld4_lane" 10
1230 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1231 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1232 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1233 (cortex_a15_neon_ldm_2): Likewise.0
1234 (cortex_a15_neon_stm_2): Likewise.
1235 (cortex_a15_neon_mcr): Likewise.
1236 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1237 (cortex_a15_neon_mrc): Likewise.
1238 (cortex_a15_neon_mrrc): Likewise.
1239 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1240 change.
1241 (cortex_a15_alu_shift): Likewise.
1242 (cortex_a15_alu_shift_reg): Likewise.
1243 (cortex_a15_mult32): Likewise.
1244 (cortex_a15_mult64): Likewise.
1245 (cortex_a15_block): Likewise.
1246 (cortex_a15_branch): Likewise.
1247 (cortex_a15_load1): Likewise.
1248 (cortex_a15_load3): Likewise.
1249 (cortex_a15_store1): Likewise.
1250 (cortex_a15_store3): Likewise.
1251 (cortex_a15_call): Likewise.
1252 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute
1253 change.
1254 (cortex_a5_f2r): Likewise.
1255 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1256 change.
1257 (cortex_a53_f2r): Likewise.
1258 * config/arm/cortex-a7.md
1259 (cortex_a7_branch): Update for attribute change.
1260 (cortex_a7_call): Likewise.
1261 (cortex_a7_alu_imm): Likewise.
1262 (cortex_a7_alu_reg): Likewise.
1263 (cortex_a7_alu_shift): Likewise.
1264 (cortex_a7_mul): Likewise.
1265 (cortex_a7_load1): Likewise.
1266 (cortex_a7_store1): Likewise.
1267 (cortex_a7_load2): Likewise.
1268 (cortex_a7_store2): Likewise.
1269 (cortex_a7_load3): Likewise.
1270 (cortex_a7_store3): Likewise.
1271 (cortex_a7_load4): Likewise.
1272 (cortex_a7_store4): Likewise.
1273 (cortex_a7_fpalu): Likewise.
1274 (cortex_a7_fconst): Likewise.
1275 (cortex_a7_fpmuls): Likewise.
1276 (cortex_a7_neon_mul): Likewise.
1277 (cortex_a7_fpmacs): Likewise.
1278 (cortex_a7_neon_mla: Likewise.
1279 (cortex_a7_fpmuld: Likewise.
1280 (cortex_a7_fpmacd: Likewise.
1281 (cortex_a7_fpfmad: Likewise.
1282 (cortex_a7_fdivs: Likewise.
1283 (cortex_a7_fdivd: Likewise.
1284 (cortex_a7_r2f: Likewise.
1285 (cortex_a7_f2r: Likewise.
1286 (cortex_a7_f_flags: Likewise.
1287 (cortex_a7_f_loads: Likewise.
1288 (cortex_a7_f_loadd: Likewise.
1289 (cortex_a7_f_stores: Likewise.
1290 (cortex_a7_f_stored: Likewise.
1291 (cortex_a7_neon): Likewise.
1292 * config/arm/cortex-a8-neon.md
1293 (cortex_a8_neon_mrc): Update for attribute change.
1294 (cortex_a8_neon_mrrc): Likewise.
1295 (cortex_a8_neon_int_1): Likewise.
1296 (cortex_a8_neon_int_2): Likewise.
1297 (cortex_a8_neon_int_3): Likewise.
1298 (cortex_a8_neon_int_4): Likewise.
1299 (cortex_a8_neon_int_5): Likewise.
1300 (cortex_a8_neon_vqneg_vqabs): Likewise.
1301 (cortex_a8_neon_vmov): Likewise.
1302 (cortex_a8_neon_vaba): Likewise.
1303 (cortex_a8_neon_vaba_qqq): Likewise.
1304 (cortex_a8_neon_vsma): Likewise.
1305 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1306 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1307 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1308 Likewise.
1309 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1310 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1311 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1312 long_scalar_qdd_64_32_long): Likewise.
1313 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1314 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1315 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1316 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1317 (cortex_a8_neon_shift_1): Likewise.
1318 (cortex_a8_neon_shift_2): Likewise.
1319 (cortex_a8_neon_shift_3): Likewise.
1320 (cortex_a8_neon_vshl_ddd): Likewise.
1321 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1322 (cortex_a8_neon_vsra_vrsra): Likewise.
1323 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1324 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1325 (cortex_a8_neon_fp_vsum): Likewise.
1326 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1327 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1328 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1329 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1330 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1331 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1332 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1333 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1334 (cortex_a8_neon_bp_simple): Likewise.
1335 (cortex_a8_neon_bp_2cycle): Likewise.
1336 (cortex_a8_neon_bp_3cycle): Likewise.
1337 (cortex_a8_neon_ldr): Likewise.
1338 (cortex_a8_neon_str): Likewise.
1339 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1340 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1341 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1342 (cortex_a8_neon_vld2_4_regs): Likewise.
1343 (cortex_a8_neon_vld3_vld4): Likewise.
1344 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1345 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1346 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1347 (cortex_a8_neon_vst3_vst4): Likewise.
1348 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1349 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1350 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1351 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1352 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1353 (cortex_a8_neon_mcr): Likewise.
1354 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1355 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
1356 change.
1357 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
1358 change.
1359 (ca9_neon_mrrc): Likewise.
1360 (cortex_a9_neon_int_1): Likewise.
1361 (cortex_a9_neon_int_2): Likewise.
1362 (cortex_a9_neon_int_3): Likewise.
1363 (cortex_a9_neon_int_4): Likewise.
1364 (cortex_a9_neon_int_5): Likewise.
1365 (cortex_a9_neon_vqneg_vqabs): Likewise.
1366 (cortex_a9_neon_vmov): Likewise.
1367 (cortex_a9_neon_vaba): Likewise.
1368 (cortex_a9_neon_vaba_qqq): Likewise.
1369 (cortex_a9_neon_vsma): Likewise.
1370 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1371 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1372 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1373 Likewise.
1374 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1375 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1376 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1377 long_scalar_qdd_64_32_long): Likewise.
1378 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1379 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1380 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1381 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1382 (cortex_a9_neon_shift_1): Likewise.
1383 (cortex_a9_neon_shift_2): Likewise.
1384 (cortex_a9_neon_shift_3): Likewise.
1385 (cortex_a9_neon_vshl_ddd): Likewise.
1386 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1387 (cortex_a9_neon_vsra_vrsra): Likewise.
1388 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1389 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1390 (cortex_a9_neon_fp_vsum): Likewise.
1391 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1392 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1393 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1394 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1395 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1396 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1397 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1398 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1399 (cortex_a9_neon_bp_simple): Likewise.
1400 (cortex_a9_neon_bp_2cycle): Likewise.
1401 (cortex_a9_neon_bp_3cycle): Likewise.
1402 (cortex_a9_neon_ldr): Likewise.
1403 (cortex_a9_neon_str): Likewise.
1404 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1405 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1406 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1407 (cortex_a9_neon_vld2_4_regs): Likewise.
1408 (cortex_a9_neon_vld3_vld4): Likewise.
1409 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1410 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1411 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1412 (cortex_a9_neon_vst3_vst4): Likewise.
1413 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1414 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1415 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1416 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1417 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1418 (cortex_a9_neon_mcr): Likewise.
1419 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1420 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
1421 (cortex_a9_fps): Likewise.
1422 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
1423 change.
1424 (cortex_m4_fmuls): Likewise.
1425 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
1426 change.
1427 (cortex_r4_mrc): Likewise.
1428 * config/arm/iterators.md: Update comment referring to neon_type.
1429 * config/arm/iwmmxt.md
1430 (iwmmxt_arm_movdi): Update for attribute change.
1431 (iwmmxt_movsi_insn): Likewise.
1432 * config/arm/marvell-pj4.md
1433 (pj4_vfp_to_core): Update for attribute change.
1434 (pj4_core_to_vfp): Likewise.
1435 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
1436 attribute change.
1437 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
1438 (vfp_fstore): Likewise.
1439 * doc/md.texi: Change references to neon_type to refer to type.
1440
fbe0dc78
DS		 14412013-09-04 Dodji Seketeli <dodji@redhat.com>
1442
1443 * tree.h (DECL_BUILT_IN): Fix typo in comment.
1444
c167bc5b
DE		 14452013-09-04 David Edelsohn <dje.gcc@gmail.com>
1446
1447 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
1448 lglobl if not weak.
1449
035cb59f
ER		 14502013-09-04 Easwaran Raman <eraman@google.com>
1451
1452 PR middle-end/57370
1453 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
1454 (build_and_add_sum): Use it.
1455 (appears_later_in_bb): Simplify code.
1456
71191083
TJ		 14572013-09-04 Teresa Johnson <tejohnson@google.com>
1458
1459 * dumpfile.c (dump_finish): Don't close stderr/stdout.
1460
0c488213
JG		 14612013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
1462
1463 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
1464
61a74079
JH		 14652013-09-04 Jan Hubicka <jh@suse.cz>
1466
1467 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
1468 * ipa-devirt.c: Include diganostic.h
1469 (odr_type_d): Add types and types_set.
1470 (hash_type_name): Work for types with vtables during LTO.
1471 (odr_hasher::remove): Fix comment; destroy types_set.
1472 (add_type_duplicate): New function,
1473 (get_odr_type): Use it.
1474 (dump_type_inheritance_graph): Dump type duplicates.
1475 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
1476 graph.
1477 * tree.c (types_same_for_odr): Give exact answers on types with
1478 virtual tables.
1479
24ba8195
DS		 14802013-09-04 Dodji Seketeli <dodji@redhat.com>
1481
1482 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
1483 explaining their differences.
1484
0a04869e
SKS		 14852013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
1486
1487 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
1488
3b18bc42
JL		 14892013-09-03 Jeff Law <law@redhat.com>
1490
1491 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
1492 when not threading through a joiner block. Pass joiner/no joiner
1493 state to register_jump_thread.
1494 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
1495 state from argument rather than implying on path length.
1496 Dump the entire jump thread path into debugging dump.
1497 * tree-flow.h (register_jump_thread): Update prototype.
1498
ae84e151
XDL		 14992013-08-29 Xinliang David Li <davidxl@google.com>
1500
1501 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1502 Remove a trivial gcc_assert.
1503
c716e67f
XDL		 15042013-08-29 Xinliang David Li <davidxl@google.com>
1505
1506 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
1507 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
1508 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1509 Delay base decl alignment adjustment.
1510 * tree-vectorizer.c (vect_destroy_datarefs): New function.
1511 * tree-vectorizer.h: New data structure.
1512 (set_dr_misalignment): New function.
1513 (dr_misalignment): Ditto.
1514 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
1515 (vectorizable_load): Ditto.
1516 (ensure_base_align): New function.
1517 (vectorize_loops): Add dbg_cnt support.
1518 (execute_vect_slp): Ditto.
1519 * dbgcnt.def: New debug counter.
1520 * Makefile: New dependency.
1521
db8800bc
MI		 15222013-09-03 Meador Inge <meadori@codesourcery.com>
1523
1524 Revert:
1525
1526 2013-08-30 Meador Inge <meadori@codesourcery.com>
1527
1528 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1529
0d9b4f5c
DE		 15302013-09-03 David Edelsohn <dje.gcc@gmail.com>
1531
1532 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
1533 function descriptor.
1534
29e810b9
RB		 15352013-09-03 Richard Biener <rguenther@suse.de>
1536
1537 * tree-affine.c (add_elt_to_tree): Fix association issue,
1538 avoid useless converts and make sure to always return a
1539 properly typed result.
1540
dbfc2894
RB		 15412013-09-03 Richard Biener <rguenther@suse.de>
1542
1543 PR middle-end/57656
1544 * fold-const.c (negate_expr_p): Fix division case.
1545 (negate_expr): Likewise.
1546
fa3bf4ce
RB		 15472013-09-03 Richard Biener <rguenther@suse.de>
1548
1549 PR lto/58285
1550 * tree-streamer-out.c: Include tm.h.
1551 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
1552
a64bbb3f
JH		 15532013-09-03 Jan Hubicka <jh@suse.cz>
1554
1555 * tree-profile.c (tree_profiling): Cleanup CFG when done.
1556
93ed07e2
AM		 15572013-09-03 Alan Modra <amodra@gmail.com>
1558
1559 * config.gcc (powerpc*-*-linux*): Add support for little-endian
1560 multilibs to big-endian target and vice versa.
1561 * config/rs6000/t-linux64: Use := assignment on all vars.
1562 (MULTILIB_EXTRA_OPTS): Remove fPIC.
1563 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
1564 * config/rs6000/t-linux64le: New file.
1565 * config/rs6000/t-linux64bele: New file.
1566 * config/rs6000/t-linux64lebe: New file.
1567
07d6cd64
JH		 15682013-09-02 Jan Hubicka <jh@suse.cz>
1569
1570 * ipa-inline-transform.c (inline_transform): Do not
1571 optimize_inline_calls when not optimizing.
1572
8e4c9a10
JH		 15732013-09-02 Jan Hubicka <jh@suse.cz>
1574
1575 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
1576 duplicated nodes for assembler names.
1577 * symtab.c (symtab_unregister_node): Do not attempt to unlink
1578 hard registers from assembler name hash.
1579
c91061e6
JH		 15802013-09-02 Jan Hubicka <jh@suse.cz>
1581
1582 * ipa-split.c (execute_split_functions): Split externally visible
1583 functions called once.
1584
2fd0985c
MJ		 15852013-09-02 Martin Jambor <mjambor@suse.cz>
1586
1587 PR ipa/58106
1588 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
1589 linked list. When finding the correct duplicate, also consider also
1590 the caller in additon to its inlined_to node.
1591
91bd4114
JG		 15922013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
1593
1594 * config/aarch64/aarch64-simd-builtins.def
1595 (dup_lane_scalar): Remove.
1596 * config/aarch64/aarch64-simd.md
1597 (aarch64_simd_dup): Add 'w->w' alternative.
1598 (aarch64_dup_lane<mode>): Allow for VALL.
1599 (aarch64_dup_lane_scalar<mode>): Remove.
1600 (aarch64_dup_lane_<vswap_width_name><mode>): New.
1601 (aarch64_get_lane_signed<mode>): Add w->w altenative.
1602 (aarch64_get_lane_unsigned<mode>): Likewise.
1603 (aarch64_get_lane<mode>): Likewise.
1604 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
1605 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
1606 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
1607 (VCON): Change container of V2SF.
1608 (vswap_width_name): Likewise.
1609 * config/aarch64/arm_neon.h
1610 (__aarch64_vdup_lane_any): New.
1611 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1612 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
1613 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1614
d617d2d8
EB		 16152013-09-02 Eric Botcazou <ebotcazou@adacore.com>
1616
1617 PR middle-end/56382
1618 * expr.c (emit_move_complex): Do not move complex FP values as parts if
1619 the source or the destination is a single hard register.
1620
63fdb7be
RB		 16212013-09-02 Richard Biener <rguenther@suse.de>
1622
1623 PR middle-end/57511
1624 * tree-scalar-evolution.c (instantiate_scev_name): Allow
1625 non-linear SCEVs.
1626
78de2333
RB		 16272013-09-02 Richard Biener <rguenther@suse.de>
1628
1629 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
1630 arithmetic to sizetype.
1631
85ff4ec6
BC		 16322013-09-02 Bin Cheng <bin.cheng@arm.com>
1633
1634 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
1635 Find auto-increment use both before and after candidate.
1636
576016fe
MP		 16372013-09-02 Marek Polacek <polacek@redhat.com>
1638
1639 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
1640
08f835dc
JH		 16412013-09-01 Jan Hubicka <jh@suse.cz>
1642
1643 * Makefile.in: Add ipa-profile.o
1644 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
1645 * cgraph.c (struct cgraph_propagate_frequency_data,
1646 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
1647 ipa-profile.c; replace cgraph_ by ipa_ prefix.
1648 * cgraph.h (cgraph_propagate_frequency): Remove.
1649 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
1650 (inline_update_callee_summaries): Update.
1651 * ipa-profile.c: New file.
1652 * ipa-utils.h (ipa_propagate_frequency): Declare.
1653 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
1654 data-streamer.h, value-prof.h
1655 (symtab_remove_unreachable_nodes): Update profile.
1656 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
1657 account_time_size, cmp_counts, dump_histogram,
1658 ipa_profile_generate_summary, ipa_profile_write_summary,
1659 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
1660 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
1661 Move to ipa-profile.c
1662
cec8583c
JDA		 16632013-09-01 John David Anglin <danglin@gcc.gnu.org>
1664
1665 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
1666
bbc9396b
JH		 16672013-09-01 Jan Hubicka <jh@suse.cz>
1668
1669 * common.opt (fdevirtualize-speculatively): New function.
1670 * invoke.texi (fdevirtualize-speculatively): Document.
1671 * ipa-devirt.c: Include ipa-inline.h
1672 (likely_target_p): New function.
1673 (ipa_devirt): New function.
1674 (gate_ipa_devirt): New function.
1675 (pass_data_ipa_devirt): New static var.
1676 (pass_ipa_devirt): Likewise.
1677 (make_pass_ipa_devirt): New function.
1678 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
1679 (common_handle_option): Disable devirtualization when
1680 value range profiling is available.
1681 * passes.def (pass_ipa_devirt): Add.
1682 * timever.def (TV_IPA_DEVIRT): New timevar.
1683 * tree-pass.h (make_pass_ipa_devirt):
1684
0cea211e
IS		 16852013-09-01 Iain Sandoe <iain@codesourcery.com>
1686
1687 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
1688 include sanitize(undefined).
1689
5e351e96
DN		 16902013-08-31 Diego Novillo <dnovillo@google.com>
1691
1692 * Makefile.in (TREE_CORE_H): Define.
1693 (TREE_H): Use.
1694 (GTFILES): Add tree-core.h.
1695 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
1696 size the array.
1697 * tree-core.h: New file.
1698 Move all data structures, enum, typedefs, global
1699 declarations and constants from ...
1700 * tree.h: ... here.
1701
63bf9a90
JH		 17022013-08-31 Jan Hubicka <jh@suse.cz>
1703
1704 * bulitins.c (expand_builtin): Do not early exit for gcov
1705 instrumented functions.
1706
5c0abd6a
MP		 17072013-08-31 Marek Polacek <polacek@redhat.com>
1708
1709 * ubsan.c: Include tm_p.h.
1710
4843f032
JH		 17112013-08-31 Jan Hubicka <jh@suse.cz>
1712
8356c89c
JH		 1713 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
1714 warning.
1715
4843f032
JH		 1716 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
1717 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 1718 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH		 1719 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
1720 (ipa_merge_profiles): New function.
1721 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
1722 (lto_input_function_body): Likewise.
5e581212
JH		 1723 * ipa-utils.h (ipa_merge_profiles): Declare.
1724 * lto-streamer.h (lto_input_function_body): Update prototype.
1725 (emit_label_in_global_context_p): Remove.
1726 * lto-symtab.c: Include ipa-utils.h
1727 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 1728
57292ce9
JH		 17292013-08-31 Jan Hubicka <jh@suse.cz>
1730
1731 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
1732
b9cbfeeb
JH		 17332013-08-31 Jan Hubicka <jh@suse.cz>
1734
1735 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
1736
41dedebd
UB		 17372013-08-31 Uros Bizjak <ubizjak@gmail.com>
1738
1739 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
1740 "cmp" RTX before signed_comparison_operator check to account
1741 for "code" changes.
1742
06d65050
JH		 17432013-08-30 Jan Hubicka <jh@suse.cz>
1744
41dedebd 1745 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH		 1746 (detect_type_change_1): Rename to ...
1747 (detect_type_change): ... this one; early return on non-polymorphic
1748 types.
41dedebd 1749 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH		 1750 use of detect_type_change.
1751 (compute_complex_assign_jump_func): Add param_type parameter;
1752 update use of detect_type_change_ssa.
1753 (compute_complex_ancestor_jump_func): Likewise.
1754 (ipa_get_callee_param_type): New function.
1755 (ipa_compute_jump_functions_for_edge): Compute parameter type;
1756 update calls to the jump function computation functions.
1757
600b5b1d
TJ		 17582013-08-30 Teresa Johnson <tejohnson@google.com>
1759 Steven Bosscher <steven@gcc.gnu.org>
1760
1761 * cfgrtl.c (fixup_new_cold_bb): New routine.
1762 (commit_edge_insertions): Invoke fixup_partitions.
1763 (find_partition_fixes): New routine.
1764 (fixup_partitions): Ditto.
1765 (verify_hot_cold_block_grouping): Update comments.
1766 (rtl_verify_edges): Invoke find_partition_fixes.
1767 (rtl_verify_bb_pointers): Update comments.
1768 (rtl_verify_bb_layout): Ditto.
1769 * basic-block.h (probably_never_executed_edge_p): Declare.
1770 (fixup_partitions): Ditto.
1771 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
1772 * bb-reorder.c (sanitize_hot_paths): New function.
1773 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
1774 sanitize_hot_paths.
1775 * predict.c (probably_never_executed_edge_p): New routine.
1776 * cfg.c (check_bb_profile): Add partition insanity warnings.
1777
0a6ea5c9
MI		 17782013-08-30 Meador Inge <meadori@codesourcery.com>
1779
1780 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1781
de5a5fa1
MP		 17822013-08-30 Marek Polacek <polacek@redhat.com>
1783
1784 * Makefile.in (ubsan.o): Add.
1785 (c-family/c-ubsan.o): Add.
1786 (builtins.o): Add ubsan.h dependency.
1787 * ubsan.h: New file.
1788 * ubsan.c: New file.
1789 * common.opt: Add -fsanitize=undefined option.
1790 (flag_sanitize): Add variable.
1791 (fsanitize=): Add option. Add Driver.
1792 (fsanitize=thread): Remove option.
1793 (fsanitize=address): Likewise.
1794 (static-libubsan): New option.
1795 * doc/invoke.texi: Document the new flag and -static-libubsan.
1796 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
1797 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
1798 * builtin-attrs.def (ATTR_COLD): Define.
1799 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1800 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
1801 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
1802 * flag-types.h (sanitize_code): New enum.
1803 * opts.c (common_handle_option): Parse command line arguments
1804 of -fsanitize=. Add -fsanitize=unreachable option.
1805 * varasm.c (get_variable_section): Adjust.
1806 (assemble_noswitch_variable): Likewise.
1807 (assemble_variable): Likewise.
1808 (output_constant_def_contents): Likewise.
1809 (categorize_decl_for_section): Likewise.
1810 (place_block_symbol): Likewise.
1811 (output_object_block): Likewise.
1812 * builtins.def: Likewise.
1813 * toplev.c (compile_file): Likewise.
1814 (process_options): Likewise.
1815 * cppbuiltin.c: Likewise.
1816 * tsan.c (tsan_pass): Likewise.
1817 (tsan_gate): Likewise.
1818 (tsan_gate_O0): Likewise.
1819 * cfgexpand.c (partition_stack_vars): Likewise.
1820 (expand_stack_vars): Likewise.
1821 (defer_stack_allocation): Likewise.
1822 (expand_used_vars): Likewise.
1823 * cfgcleanup.c (old_insns_match_p): Likewise.
1824 * asan.c (asan_finish_file): Likewise.
1825 (asan_instrument): Likewise.
1826 (gate_asan): Likewise.
1827 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
1828 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1829 (asan_global_struct): Use pointer_sized_int_node instead
1830 calling build_nonstandard_integer_type.
1831 (initialize_sanitizer_builtins): Likewise.
1832 (asan_finish_file): Likewise.
1833 * gcc.c: Document %{%:function(args):X}.
1834 (static_spec_functions): Add sanitize.
1835 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
1836 store funcval != NULL there.
1837 (do_spec_1): Adjust handle_spec_function caller.
1838 (handle_braces): Allow %:function(args) as condition.
1839 (sanitize_spec_function): New function.
1840 (ADD_STATIC_LIBUBSAN_LIBS): Define.
1841 (LIBUBSAN_SPEC): Likewise.
1842 (LIBUBSAN_EARLY_SPEC): Likewise.
1843 (SANITIZER_SPEC): Handle libubsan.
1844 (SANITIZER_EARLY_SPEC): Likewise.
1845 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
1846 instead of fsanitize=address.
1847 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
1848 instead of fsanitize=address*.
1849 * builtins.c: Include ubsan.h.
1850 (fold_builtin_0): Instrument __builtin_unreachable.
1851 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
1852 instead of flag_asan.
1853 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
1854 (pointer_sized_int_node): Define.
41dedebd 1855 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 1856
f07f30cf
MS		 18572013-08-30 Mike Stump <mikestump@comcast.net>
1858
1859 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
1860 with RE patterns.
1861
8a41354f
JH		 18622013-08-29 Jan Hubicka <jh@suse.cz>
1863
1864 * cgraph.c (cgraph_function_body_availability): Handle weakref
1865 correctly.
1866 * passes.def: Remove pass_fixup_cfg.
1867 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
1868 track when we need to remove functions.
1869 (gate_ipa_inline): Execute inlining always; add comment why.
1870 (pass_data_ipa_inline): Remove TODO_remove_functions.
1871 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
1872 do not produce summaries.
1873 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
1874 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
1875 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
1876 constant pool and vtable.
1877
dc8d7a0f
TB		 18782013-08-30 Tejas Belagod <tejas.belagod@arm.com>
1879
41dedebd
UB		 1880 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
1881 New arm_neon.h's internal macros to specify 64-bit constants.
1882 Avoid using stdint.h's macros.
dc8d7a0f 1883
ac7eacd2
JR		 18842013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
1885
1886 * recog.c (verify_changes): Verify that changes[i].old is non-zero
1887 before applying REG_P.
1888
8b29fd4e
JJ		 18892013-08-30 Jakub Jelinek <jakub@redhat.com>
1890
1891 PR tree-optimization/58277
1892 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
1893 after seeing too many stmts with vdef in between dombb and current
1894 bb, invalidate everything.
1895
26d75703
RB		 18962013-08-30 Richard Biener <rguenther@suse.de>
1897
1898 * fold-const.c (fold_single_bit_test): Fix overflow test.
1899
f7e088e7
EB		 19002013-08-30 Eric Botcazou <ebotcazou@adacore.com>
1901
1902 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
1903 and which can live in a register, always retrieve the value on entry.
1904 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
1905 passed by invisible reference specially.
1906 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
1907 (vt_add_function_parameter): Correctly deal with a parameter passed by
1908 invisible reference.
1909
19102013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH		 1911
1912 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
1913
6e8dad05
RB		 19142013-08-30 Richard Biener <rguenther@suse.de>
1915
1916 PR tree-optimization/58228
1917 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
1918 allow invariant loads in nested loop vectorization.
1919
062ef2c8
RB		 19202013-08-30 Richard Biener <rguenther@suse.de>
1921
1922 PR tree-optimization/58223
1923 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
1924 (has_anti_or_output_dependence): ... this and adjust to also
1925 look for output dependences.
1926 (mark_nodes_having_upstream_mem_writes): Adjust.
1927 (rdg_flag_uses): Likewise.
1928
7a764c60
RB		 19292013-08-30 Richard Biener <rguenther@suse.de>
1930
1931 PR tree-optimization/58010
1932 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
1933 assert that we have a loop-closed PHI.
1934
82e9d642
JH		 19352013-08-29 Jan Hubicka <jh@suse.cz>
1936
1937 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
1938 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
1939 * lto-section-in.c (lto_free_function_in_decl_state): New function.
1940 (lto_free_function_in_decl_state_for_node): New function.
1941
9cc1fb4b
XDL		 19422013-08-29 Xinliang David Li <davidxl@google.com>
1943
41dedebd 1944 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL		 1945 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
1946 Emit alignment peeling message with default -fopt-info.
1947 (vect_loop_versioning): Emit loop version info message.
41dedebd 1948 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL		 1949 (execute_vect_slp): Ditto.
1950
440a5082
EB		 19512013-08-29 Eric Botcazou <ebotcazou@adacore.com>
1952
1953 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
1954 of the clone from the DECL_NAME of the original function.
1955
3fa3690d
OE		 19562013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
1957
1958 * passes.c (register_pass): Add overload.
1959 * tree-pass.h (register_pass): Forward declare it. Add comment.
1960
0170f33c
JH		 19612013-08-29 Jan Hubicka <jh@suse.cz>
1962
41dedebd
UB		 1963 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
1964 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
1965 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
1966 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 1967 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
1968 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 1969 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 1970 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
1971
103ff0d6
TJ		 19722013-08-29 Teresa Johnson <tejohnson@google.com>
1973
1974 * dumpfile.c (dump_loc): Output column number.
1975 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
1976 * doc/invoke.texi: Document optall -fopt-info flag.
1977 * profile.c (read_profile_edge_counts): Use new dump framework.
1978 (compute_branch_probabilities): Ditto.
1979 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
1980 when pass not in any opt group.
1981 * pass_manager.h (pass_manager::get_pass_profile): New method.
1982 * value-prof.c (check_counter): Use new dump framework.
1983 (check_ic_target): Ditto.
1984 * coverage.c (get_coverage_counts): Ditto.
1985 (coverage_init): Setup new dump framework.
1986
301bbc16
RB		 19872013-08-29 Richard Biener <rguenther@suse.de>
1988
1989 PR tree-optimization/58246
1990 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
1991 handle the dominance check inside a basic-block.
1992
1b275000
RB		 19932013-08-29 Richard Biener <rguenther@suse.de>
1994
1995 PR middle-end/57287
1996 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
1997 of default defs that appear in abnormal PHI nodes.
1998
6b1184ba
RB		 19992013-08-29 Richard Biener <rguenther@suse.de>
2000
2001 PR tree-optimization/57685
2002 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2003 single-use operands to avoid exponential complexity.
2004
1b08b734
DC		 20052013-08-28 Dehao Chen <dehao@google.com>
2006
2007 * ipa-inline.c (edge_badness): Fix integer underflow.
2008
48a3fa69
UB		 20092013-08-28 Uros Bizjak <ubizjak@gmail.com>
2010
2011 * gtm-builtins.def (_ITM_free): Declare leaf.
2012
215f73e6
JJ		 20132013-08-28 Jakub Jelinek <jakub@redhat.com>
2014
2015 PR target/58067
2016 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2017 (*tls_local_dynamic_base_64_largepic): Likewise.
2018 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2019 Remove predicate from call operand.
2020 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2021 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2022
34554d1a
JL		 20232013-08-28 Jeff Law <law@redhat.com>
2024
2025 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2026 checks for the number of predecessors and successors allowed.
2027 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2028 which require copying a joiner block if there is a request which
2029 is a subpath that requires no joiner block copying.
2030
c01c111b
JH		 20312013-08-28 Jan Hubicka <jh@suse.cz>
2032
2033 * lto-streamer-out.c (DFS_write_tree_body): Drop
2034 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2035 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2036 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2037 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2038 Do not read DECL_ERROR_ISSUED.
2039 (unpack_ts_decl_with_vis_value_fields): Do not read
2040 DECL_DEFER_OUTPUT.
48a3fa69
UB		 2041 (lto_input_ts_binfo_tree_pointers): Do not read
2042 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH		 2043 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2044 write DECL_ERROR_ISSUED..
2045 (pack_ts_decl_with_vis_value_fields): Do not write
2046 DECL_DEFER_OUTPUT.
2047 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 2048 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH		 2049 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2050 * tree.h (tree_decl_common): Update comment.
2051 (DECL_ERROR_ISSUED): Remove.
2052
4b128ece
JJ		 20532013-08-28 Jakub Jelinek <jakub@redhat.com>
2054
2055 PR middle-end/58257
2056 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2057
12211b99 20582013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH		 2059
2060 * builtins.def (free): Declare leaf.
2061
8dce4dbc
DM		 20622013-08-27 David Malcolm <dmalcolm@redhat.com>
2063
2064 * gdbhooks.py: New.
2065 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2066 * configure: Regenerate.
2067
b8f6e610
MJ		 20682013-08-27 Martin Jambor <mjambor@suse.cz>
2069
2070 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2071 (ipa_ancestor_jf_data): Likewise.
2072 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2073 (ipa_get_jf_pass_through_type_preserved): New function.
2074 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2075 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 2076 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ		 2077 (ipa_get_jf_ancestor_result): Likewise.
2078 (propagate_vals_accross_pass_through): Use
2079 ipa_get_jf_pass_through_result to do all the value mappings.
2080 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2081 type_preserved flag.
2082 (ipa_set_jf_cst_copy): New function.
2083 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2084 (ipa_set_jf_arith_pass_through): Likewise.
2085 (ipa_set_ancestor_jf): Likewise.
2086 (compute_complex_assign_jump_func): Set type_preserved instead of
2087 punting.
2088 (ipa_compute_jump_functions_for_edge): Likewise.
2089 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2090 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 2091 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ		 2092 (ipa_write_jump_function): Stream the type_preserved flags.
2093 (ipa_read_jump_function): Likewise.
2094
74bf76ed
JJ		 20952013-08-27 Jakub Jelinek <jakub@redhat.com>
2096 Aldy Hernandez <aldyh@redhat.com>
2097
2098 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
2099 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
2100 * function.h (struct function): Add has_force_vect_loops and
2101 has_simduid_loops.
12211b99 2102 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ		 2103 * gimple.c (gimple_build_omp_critical): Add KIND argument and
2104 handle it.
2105 * gimple.def: Update CLAUSES comments.
2106 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
2107 (gimple_build_omp_for): Add argument to prototype.
2108 (gimple_omp_for_kind): New.
2109 (gimple_omp_for_set_kind): New.
2110 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
2111 GOVD_DATA_SHARE_CLASS.
2112 (enum omp_region_type): Add ORT_SIMD.
2113 (gimple_add_tmp_var): Handle ORT_SIMD.
2114 (gimplify_var_or_parm_decl): Same.
2115 (is_gimple_stmt): Same.
2116 (omp_firstprivatize_variable): Same.
2117 (omp_add_variable): Only use splay_tree_insert if lookup failed.
2118 (omp_notice_variable): Handle ORT_SIMD.
12211b99 2119 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ		 2120 (omp_check_private): Handle ORT_SIMD.
2121 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2122 OMP_CLAUSE_SAFELEN.
2123 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
2124 Handle OMP_CLAUSE_LASTPRIVATE.
2125 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2126 OMP_CLAUSE_SAFELEN.
2127 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
2128 (gimplify_expr): Handle OMP_SIMD.
2129 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
2130 (expand_GOMP_SIMD_VF): New.
2131 (expand_GOMP_SIMD_LAST_LANE): New.
2132 * internal-fn.def (GOMP_SIMD_LANE): New.
2133 (GOMP_SIMD_VF): New.
2134 (GOMP_SIMD_LAST_LANE): New.
2135 * omp-low.c: Include target.h.
2136 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
2137 OMP_CLAUSE_SAFELEN.
2138 (check_omp_nesting_restrictions): Same.
2139 (omp_max_vf): New.
2140 (lower_rec_simd_input_clauses): New.
2141 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
2142 OMP_CLAUSE_LINEAR.
2143 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
2144 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
2145 (expand_omp_build_assign): New.
2146 (expand_omp_for_init_counts): New.
2147 (expand_omp_for_init_vars): New.
2148 (extract_omp_for_update_vars): New.
2149 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
2150 and rewrite accordingly.
2151 (expand_omp_simd): New.
2152 (expand_omp_for): Use expand_omp_simd.
2153 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2154 (lower_omp_for): Do not lower the body.
2155 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2156 in their own loops.
2157 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB		 2158 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2159 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ		 2160 (gate_tree_if_conversion): Similarly.
2161 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2162 gimple_build_omp_for.
2163 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2164 * tree-parloops (create_parallel_loop): Pass kind argument to
2165 gimple_build_omp_for.
2166 * tree-pretty-print.c (dump_omp_clause): Add cases for
2167 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2168 OMP_CLAUSE__SIMDUID_.
2169 (dump_generic_node): Handle OMP_SIMD.
2170 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2171 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2172 unroll OMP_SIMD loops here.
12211b99 2173 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 2174 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 2175 loop->safelen.
74bf76ed
JJ		 2176 (vect_analyze_data_refs): Handle simd loops.
2177 * tree-vect-loop.c (vectorizable_live_operation): Handle
2178 IFN_GOMP_SIMD*.
12211b99 2179 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ		 2180 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2181 (vectorizable_load): Same.
12211b99 2182 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ		 2183 (struct simduid_to_vf): New.
2184 (simduid_to_vf::hash): New.
2185 (simduid_to-vf::equal): New.
2186 (struct simd_array_to_simduid): New.
2187 (simd_array_to_simduid::hash): New.
2188 (simd_array_to_simduid::equal): New.
2189 (adjust_simduid_builtins): New.
2190 (struct note_simd_array_uses_struct): New.
2191 (note_simd_array_uses_cb): New.
2192 (note_simd_array_uses): New.
2193 (vectorize_loops): Handle simd hints and adjust simd builtins
2194 accordingly.
2195 * tree-vectorizer.h (struct _stmt_vec_info): Add
2196 simd_lane_access_p field.
2197 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2198 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2199 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2200 (omp_clause_code_name): Same.
2201 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2202 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2203 * tree.def (OMP_SIMD): New entry.
12211b99
UB		 2204 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2205 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ		 2206 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2207 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2208 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2209 (OMP_CLAUSE_LINEAR_STEP): New.
2210 (OMP_CLAUSE_SAFELEN_EXPR): New.
2211 (OMP_CLAUSE__SIMDUID__DECL): New.
2212 (find_omp_clause): New prototype.
2213
d0cf4e84
L		 22142013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2215
2216 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2217 Haswell processor detection.
2218
992592ec
CW		 22192013-08-27 Christian Widmer <shadow@umbrox.de>
2220
2221 PR target/57927
2222 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2223 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2224 AVX2 capable processors.
2225
9116eb22
TB		 22262013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2227
2228 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2229 of vget_low_* with implementations in terms of other intrinsics.
2230
1c5abb53
MG		 22312013-08-27 Marc Glisse <marc.glisse@inria.fr>
2232
2233 PR middle-end/57219
2234 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2235 values to -1, 0 and 1.
2236
2e100703
VP		 22372013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2238
2239 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2240 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2241 (<optab><mode>3_insn): Remove.
2242 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2243 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2244 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2245 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2246 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2247 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2248 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2249 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2250 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2251 (ror<mode>3_insn): Likewise.
2252 * config/aarch64/predicates.md (aarch64_simd_register): New.
2253
4ded8276
RB		 22542013-08-27 Richard Biener <rguenther@suse.de>
2255
2256 PR tree-optimization/57521
2257 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2258 one edge is non-critical.
2259 (find_phi_replacement_condition): Make sure to use a non-critical
2260 edge. Cleanup and remove old bug workarounds.
2261 (bb_postdominates_preds): Remove.
2262 (if_convertible_loop_p_1): Do not compute post-dominators.
2263 (combine_blocks): Do not free post-dominators.
2264 (main_tree_if_conversion): Likewise.
2265 (pass_data_if_conversion): Add TODO_verify_ssa.
2266
5aa11061
DD		 22672013-08-27 DJ Delorie <dj@redhat.com>
2268
2269 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2270
fd91cfe3
YZ		 22712013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2272
2273 * function.c (assign_parm_find_data_types): Set passed_mode and
2274 nominal_mode to the TYPE_MODE of nominal_type for the built
2275 pointer type in case of the struct-pass-by-reference.
2276
907555ce
JR		 22772013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2278
2279 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2280 (UINT16_TYPE): Change default to "unsigned int".
2281
67518c93
JR		 2282 * config/avr/avr.opt (mfract-convert-truncate): New option.
2283 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2284 is set, round negative fractional integers according to n1169
2285 when converting to integer types.
2286
d6d989fc
JH		 22872013-08-26 Jan Hubicka <jh@suse.cz>
2288
2289 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2290 methods can not be called indirectly when their address is not taken.
2291
2aa3da06
JH		 22922013-08-26 Jan Hubicka <jh@suse.cz>
2293
12211b99
UB		 2294 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2295 ctor_for_folding.
2aa3da06 2296
0987ffe7
JH		 22972013-08-26 Jan Hubicka <jh@suse.cz>
2298
2299 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2300 can be unshared.
2301
befe8647
JR		 23022013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2303
2304 * reload.c (find_valid_class): Allow classes that do not include
2305 FIRST_PSEUDO_REGISTER - 1.
2306
a21e735e
JH		 23072013-08-26 Jan Hubicka <jh@suse.cz>
2308
2309 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2310 fix edge count/frequency when speculation failed; fix type check
2311 for the direct call.
2312
e067bd43
JH		 23132013-08-26 Jan Hubicka <jh@suse.cz>
2314
2315 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2316
e8aec975
JH		 23172013-08-26 Jan Hubicka <jh@suse.cz>
2318
2319 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2320 to be changed by edge redirection.
2321
d0b66480
JH		 23222013-08-26 Jan Hubicka <jh@suse.cz>
2323
12211b99
UB		 2324 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2325 formating; add sanity check.
d0b66480
JH		 2326 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2327 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2328 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2329 (resolve_noninline_speculation): Update callee keys, too.
2330
0f9aaac7
JH		 23312013-08-26 Jan Hubicka <jh@suse.cz>
2332
2333 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2334 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2335
4b37444e
JR		 23362013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2337
2338 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2339 into proper place.
2340
5339fc0c
UB		 23412013-08-26 Uros Bizjak <ubizjak@gmail.com>
2342
2343 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2344 (x86_64_elf_select_section): Ditto.
2345 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2346 arguments.
2347 (ix86_pass_by_reference): Ditto.
5339fc0c
UB		 2348 (output_set_got): Ditto.
2349 (ix86_unary_operator_ok): Ditto.
2350 (ix86_expand_builtin): Ditto.
2351
d5c3d3ef
JH		 23522013-08-23 Jan Hubicka <jh@suse.cz>
2353
5339fc0c 2354 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 2355
5b1e7435
JH		 23562013-08-23 Jan Hubicka <jh@suse.cz>
2357
2358 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
2359 (tree_decl_with_vis): Add FINAL field.
2360
ead69dac
JL		 23612013-08-23 Jeff Law <law@redhat.com>
2362
2363 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
2364 the debugging dump when the expression is fully redundant.
2365
025311c4
GDR		 23662013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2367
2368 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
2369 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
2370 * pretty-print.c (pp_formatted_text_data): Likewise.
2371 (pp_write_text_to_stream): Likewise.
2372 (pp_write_text_as_dot_label_to_stream): Likewise.
2373 (pp_append_r): Likewise.
2374 (pp_format): Likewise.
2375 (pp_flush): Likewise.
2376 (pp_clear_output_area): Likewise.
2377 (pp_append_text): Likewise.
2378 (pp_formatted_text): Likewise.
2379 (pp_remaining_character_count_for_line): Likewise.
2380 (pp_newline): Likewise.
2381 (pp_character): Likewise.
2382 (output_buffer::~output_buffer): Define.
2383 (pretty_printer::~pretty_printer): Destruct output buffer.
2384 * pretty-print.h (output_buffer::~output_buffer): Declare.
2385 (pretty_printer::~pretty_printer): Declare virtual.
2386
0cadbfaa
MG		 23872013-08-24 Marc Glisse <marc.glisse@inria.fr>
2388
2389 PR other/57324
2390 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
2391 HOST_WIDE_INT_M1U): New macros.
2392 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
2393 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
2394 unsigned -1 for lshift.
2395 * cse.c (cse_insn): Likewise.
2396 * double-int.c (rshift_double, lshift_double): Likewise.
2397 * builtins.c (fold_builtin_bitop): Likewise.
2398 * combine.c (force_to_mode): Likewise.
2399 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
2400 * simplify-rtx.c (simplify_const_unary_operation,
2401 simplify_const_binary_operation): Likewise.
2402 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
2403 check_va_list_escapes): Likewise.
2404 * rtlanal.c (nonzero_bits1): Likewise.
2405 * expmed.c (expand_smod_pow2): Likewise.
2406 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
2407
009e5353
JH		 24082013-08-23 Jan Hubicka <jh@suse.cz>
2409
2410 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
2411 as having address taken.
2412
c4be6568
JH		 24132013-08-23 Jan Hubicka <jh@suse.cz>
2414
64cbf23d
JH		 2415 * ipa-utils.h (method_class_type): Declare.
2416 * ipa-devirt.c (method_class_type): Export.
2417
c4be6568
JH		 2418 * cgraphunit.c (analyze_functions): Do basic devirtualization;
2419 do not walk base classes of anonymous types.
2420
97aba8e9
KK		 24212013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
2422
2423 PR rtl-optimization/58220
2424 PR regression/58221
2425 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
2426 handle SEQUENCE insns properly.
2427
0fc80001
GDR		 24282013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2429
2430 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
2431 definition.
2432 (pp_newline_and_indent): Likewise.
2433 (pp_separate_with): Likewise.
2434 * pretty-print.c (pp_newline_and_flush): Define.
2435 (pp_newline_and_indent): Likewise.
2436 (pp_separate_with): Likewise.
2437
520a5868
JJ		 24382013-08-23 Jakub Jelinek <jakub@redhat.com>
2439
2440 PR target/58218
2441 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
2442 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
2443
e500c62a
KY		 24442013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
2445
2446 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
2447 * gcc/config/i386/i386.md (*movti_internal): Use
2448 predicate to determine if EVEX is needed.
2449 (*movsi_internal): Ditto.
2450 (*movdf_internal): Ditto.
2451 (*movsf_internal): Ditto.
2452 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
2453
6b00d7dd
JJ		 24542013-08-23 Jakub Jelinek <jakub@redhat.com>
2455
2456 PR tree-optimization/58209
2457 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
2458 (find_tail_calls): Give up for pointer result types if m is non-NULL.
2459 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
2460 emit POINTER_PLUS_EXPR.
2461 (create_tailcall_accumulator): For pointer result type accumulate in
2462 sizetype type.
2463
4f2a9f90
PC		 24642013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
2465
2466 * configure.ac: Add backslashes missing from the last change.
2467 * configure: Regenerate.
2468
7bf4274e
JH		 24692013-08-22 Jan Hubicka <jh@susue.cz>
2470
4f2a9f90
PC		 2471 * ipa.c (function_and_variable_visibility): First remember function
2472 was global and then make it local.
7bf4274e 2473
1bea243d
JB		 24742013-08-22 Julian Brown <julian@codesourcery.com>
2475
2476 * configure.ac: Add aarch64 to list of arches which use "nop" in
2477 debug_line test.
2478 * configure: Regenerate.
2479
bcba7ecc
AK		 24802013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2481
2482 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
2483 gnu_libc_has_function.
2484 * config/s390/tpf.h: Likewise.
2485
4c495b0d
JH		 24862013-08-22 Jan Hubicka <jh@susue.cz>
2487
2488 * timevar.c (validate_phases): Add cast.
2489
5e302bca
JH		 24902013-08-22 Jan Hubicka <jh@susue.cz>
2491
2492 * timevar.c (validate_phases): Use size_t for memory.
2493 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
2494
da6ca2b5
GDR		 24952013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
2496
2497 * pretty-print.h (output_buffer::output_buffer): Declare.
2498 (pretty_printer::pretty_printer): Likewise.
2499 (pp_construct): Remove.
2500 * pretty-print.c (output_buffer::output_buffer): Define.
2501 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
2502 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
2503 (print_gimple_expr): Likewise.
2504 (print_gimple_seq): Likewise.
2505 (gimple_dump_bb): Likewise.
2506 * sched-vis.c (dump_value_slim): Likewise.
2507 (dump_insn_slim): Likewise.
2508 (dump_rtl_slim): Likewise.
2509 (str_pattern_slim): Likewise.
2510 * tree-mudflap.c (mf_varname_tree): Likewise.
2511 * graph.c (print_graph_cfg): Likewise.
2512 (start_graph_dump): Likewise.
2513 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
2514 placement-new.
2515 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
2516 pretty printer initialization.
2517 * coretypes.h (diagnostic_context): Remove superflous type alias
2518 declaration.
2519 (pretty_printer): Likewise. Declare directly as a class.
2520 (pretty_print_info): Remove declaration as class.
2521 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
2522 and pp_clear_output_area.
2523 (asan_add_global): Likewise.
2524
0e1474e5
JH		 25252013-08-22 Jan Hubicka <jh@suse.cz>
2526
2527 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
2528 * ipa-utils.h (update_type_inheritance_graph): Declare.
2529 (possible_polymorphic_call_target_p): Declare.
2530 (possible_polymorphic_call_target_p): New.
2531 * ipa-devirt.c: Update toplevel comments.
2532 (cached_polymorphic_call_targets): Move up.
2533 (odr_type_d): Move ID down.
2534 (polymorphic_type_binfo_p): Update comment.
2535 (odr_hasher::remove): Likewise;
2536 (get_odr_type): Set anonymous_namespace.
2537 (dump_odr_type): Dump it.
2538 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
2539 (maybe_record_node): Record node in cached_polymorphic_call_targets.
2540 (record_binfo): Add comment.
5339fc0c
UB		 2541 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
2542 built.
0e1474e5
JH		 2543 (devirt_node_removal_hook): Do not iCE when cache is freed.
2544 (possible_polymorphic_call_target_p): New predicate.
2545 (update_type_inheritance_graph): New function.
2546
3f97cb0b
AI		 25472013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
2548 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2549 Sergey Lega <sergey.s.lega@intel.com>
2550 Anna Tikhonova <anna.tikhonova@intel.com>
2551 Ilya Tocar <ilya.tocar@intel.com>
2552 Andrey Turetskiy <andrey.turetskiy@intel.com>
2553 Ilya Verbin <ilya.verbin@intel.com>
2554 Kirill Yukhin <kirill.yukhin@intel.com>
2555 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2556
2557 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
2558 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
2559 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
2560 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
2561 (OPTION_MASK_ISA_AVX2_UNSET): Update.
2562 (OPTION_MASK_ISA_AVX512F_UNSET): New.
2563 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
2564 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
2565 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
2566 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
2567 OPT_mavx512pf, OPT_mavx512er cases.
2568 * config/i386/constraints.md (v): New constraint.
2569 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
2570 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
2571 (bit_AVX512CD): New.
2572 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
2573 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
2574 * config/i386/i386-c.c (ix86_target_macros_internal):
2575 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
2576 __AVX512PF__.
2577 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
2578 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
2579 * config/i386/i386.c (regclass_map, dbx_register_map)
2580 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
2581 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
2582 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
2583 -mavx512pf options.
2584 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
2585 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
2586 -mavx512pf options. Fix formatting.
2587 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
2588 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
2589 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
2590 -mavx512cd, -mavx512pf options.
2591 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
2592 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
2593 (ix86_preferred_output_reload_class): Replace SSE_REGS with
2594 ALL_SSE_REGS.
2595 (ix86_hard_regno_mode_ok): Support 512-bit registers.
2596 (ix86_set_reg_reg_cost): Ditto.
2597 (x86_order_regs_for_local_alloc): Ditto.
2598 (MAX_VECT_LEN): Extend to 64-byte.
2599 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
2600 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
2601 (TARGET_AVX512ER, TARGET_AVX512CD): New.
2602 (BIGGEST_ALIGNMENT): Extend to 512-bits.
2603 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
2604 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
2605 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
2606 (SSE_REG_MODE_P): Support new modes.
2607 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
2608 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
2609 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
2610 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
2611 (REG_CLASS_CONTENTS): Add new registers.
2612 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
2613 (EXT_REX_SSE_REGNO_P): New.
2614 (HI_REGISTER_NAMES): Add new registers.
2615 * config/i386/i386.md: Define constants for new registers.
2616 (mode): Add new 512-bit modes.
2617 (prefix): Support evex prefix.
2618 (isa): Support avx512f, noavx512f, fma_avx512f.
2619 (ssemodesuffix): Add new 512-bit modes.
2620 (movxi): New.
2621 (*movxi_internal_avx512f): Ditto.
2622 (*movdi_internal): Replace constraint "x" with the new constraint "v".
2623 Support MODE_XI.
2624 (*movsi_internal): Likewise.
2625 (*movdf_internal): Likewise.
2626 (*movsf_internal): Likewise.
2627 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
2628 (<code><mode>3): Likewise.
5339fc0c
UB		 2629 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
2630 New.
3f97cb0b
AI		 2631 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
2632 with the new constraint "v".
2633 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
2634 modes.
2635 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
2636 with the new constraint "v".
2637 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
2638 (<sse2>_storedqu<avxsizesuffix>): Likewise.
2639 (*<plusminus_insn><mode>3): Likewise.
2640 (<sse>_vm<plusminus_insn><mode>3): Likewise.
2641 (*mul<mode>3): Likewise.
2642 (<sse>_vmmul<mode>3): Likewise.
2643 (<sse>_div<mode>3): Likewise.
2644 (<sse>_vmdiv<mode>3): Likewise.
2645 (<sse>_sqrt<mode>2): Likewise.
2646 (<sse>_vmsqrt<mode>2): Likewise.
2647 (*<code><mode>3_finite): Likewise.
2648 (*<code><mode>3) <smaxmin>: Likewise.
2649 (<sse>_vm<code><mode>3): Likewise.
2650 (*<code><mode>3) <any_logic>: Likewise.
2651 (*fma_fmadd_<mode>): Likewise.
2652 (*fma_fmsub_<mode>): Likewise.
2653 (*fma_fnmadd_<mode>): Likewise.
2654 (*fma_fnmsub_<mode>): Likewise.
2655 (*fma_fmaddsub_<mode>): Likewise.
2656 (*fma_fmsubadd_<mode>): Likewise.
2657 (*fmai_fmadd_<mode>): Likewise.
2658 (*fmai_fmsub_<mode>): Likewise.
2659 (*fmai_fnmadd_<mode>): Likewise.
2660 (*fmai_fnmsub_<mode>): Likewise.
2661 (sse_cvtsi2ss): Likewise.
2662 (sse_cvtsi2ssq): Likewise.
2663 (sse_cvtss2si): Likewise.
2664 (sse_cvtss2si_2): Likewise.
2665 (sse_cvtss2siq): Likewise.
2666 (sse_cvtss2siq_2): Likewise.
2667 (sse_cvttss2si): Likewise.
2668 (sse_cvtss2siq_2): Likewise.
2669 (float<sseintvecmodelower><mode>2): Likewise.
2670 (sse2_cvtsd2si_2): Likewise.
2671 (sse2_cvtsd2siq_2): Likewise.
2672 (*<plusminus_insn><mode>3): Likewise.
2673 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
2674 (*<sse4_1_avx2>_mul<mode>3): Likewise.
2675 (ashr<mode>3): Likewise.
2676 (<shift_insn><mode>3): Likewise.
2677 (avx2_<code><mode>3): Likewise.
2678 (*avx2_<code><mode>3): Likewise.
2679 (*andnot<mode>3): Likewise.
2680 (*<code><mode>3) <any_logic>: Likewise.
2681 (abs<mode>2): Likewise.
2682 (avx2_permvar<mode>): Likewise.
2683 (avx2_perm<mode>_1): Likewise.
2684 (*avx_vpermilp<mode>): Likewise.
2685 (avx_vpermilvar<mode>3): Likewise.
2686 (avx2_ashrv<mode>): Likewise.
2687 (avx2_<shift_insn>v<mode>): Likewise.
2688 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
2689 -mavx512cd.
2690 * doc/rtl.texi: Document XImode.
2691
dd5e8423
JL		 26922013-08-21 Jeff Law <law@redhat.com>
2693
b9ebee5d
JL		 2694 * tree-flow.h (register_jump_thread): Pass vector of edges
2695 instead of each important edge.
2696 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
2697 thread path into a vector and pass that to register_jump_thread.
2698 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
2699 passed in edge vector to the current 3-edge form.
2700
dd5e8423
JL		 2701 Revert:
2702 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2703
9bb6628e 2704 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL		 2705 delete_unmarked_insns removed anything.
2706
445dc8df
JR		 27072013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
2708
2709 * reload.h (struct reg_equivs): Rename to ..
2710 (struct reg_equivs_s): .. this.
2711
e0df53dd
ML		 27122013-08-20 Martin Liska <marxin.liska@gmail.com>
2713
2714 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
2715
e7d1d3eb
RO		 27162013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2717
5339fc0c 2718 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 2719
5562e26e
JL		 27202013-08-21 Jeff Law <law@redhat.com>
2721
2722 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
2723 simplify assignments too. If the RHS collapses to a singleton
2724 range, then return the value for the range.
2725
c7ecdec6
KY		 27262013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
2727
2728 * config/i386/sse.md (V16): Rename to...
2729 (VMOVE): this.
2730 (mov<mode>): Update iterator name.
2731 (*mov<mode>_internal): Ditto.
2732 (push<mode>1): Ditto.
2733 (movmisalign<mode>): Ditto.
2734
bfa3b50a
JH		 27352013-08-20 Jan Hubicka <jh@suse.cz>
2736
2737 PR bootstrap/58186
2738 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
2739 entry for direct edges.
2740 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
2741
6cd4d135
DM		 27422013-08-20 David Malcolm <dmalcolm@redhat.com>
2743
2744 Revert my last two changes, r201865 and r201864:
2745
2746 Revert r201865:
2747 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2748
2749 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2750 instances can own GC refs.
2751
2752 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2753 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2754 (gcc::context::gt_pch_nx): Likewise.
2755 (gcc::context::gt_pch_nx): Likewise.
2756 * ggc.h (gt_ggc_mx <T>): New.
2757 (gt_pch_nx_with_op <T>): New.
2758 (gt_pch_nx <T>): New.
2759 * passes.c (opt_pass::gt_ggc_mx): New.
2760 (opt_pass::gt_pch_nx): New.
2761 (opt_pass::gt_pch_nx_with_op): New.
2762 (pass_manager::gt_ggc_mx): New.
2763 (pass_manager::gt_pch_nx): New.
2764 (pass_manager::gt_pch_nx_with_op): New.
2765 (pass_manager::operator new): Use
2766 ggc_internal_cleared_alloc_stat rather than xcalloc.
2767 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2768 (pass_manager::gt_ggc_mx): New.
2769 (pass_manager::gt_pch_nx): New.
2770 (pass_manager::gt_pch_nx_with_op): New.
2771 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2772 (opt_pass::operator new): New.
2773 (opt_pass::gt_ggc_mx): New.
2774 (opt_pass::gt_pch_nx): New.
2775 (opt_pass::gt_pch_nx_with_op): New.
2776
2777 Revert r201864:
2778 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2779
2780 * Makefile.in (GTFILES): Add context.h.
2781 * context.c (gcc::context::operator new): New.
2782 (gcc::context::gt_ggc_mx): New.
2783 (gcc::context::gt_pch_nx): New.
2784 (gcc::context::gt_pch_nx): New.
2785 * context.h (gcc::context): Add GTY((user)) marking.
2786 (gcc::context::operator new): New.
2787 (gcc::context::gt_ggc_mx): New.
2788 (gcc::context::gt_pch_nx): New.
2789 (gcc::context::gt_pch_nx): New.
2790 (g): Add GTY marking.
2791 (gt_ggc_mx (gcc::context *)): New.
2792 (gt_pch_nx (gcc::context *)): New.
2793 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2794 void *cookie)): New.
2795 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2796
a046136a
AM		 27972013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2798
2799 * dce.c (fini_dce): Call df_analyze again just in case
2800 delete_unmarked_insns removed anything.
2801
67598720
TJ		 28022013-08-20 Teresa Johnson <tejohnson@google.com>
2803
2804 PR rtl-optimizations/57451
2805 * final.c (reemit_insn_block_notes): Prevent lexical blocks
2806 from crossing split section boundaries.
2807
f8693faf
MGD		 28082013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
2809
2810 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
2811 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
2812 with MULTLIB_DEFAULTS.
2813
7d0b9a9c
NC		 28142013-08-20 Nick Clifton <nickc@redhat.com>
2815
2816 * target.def (narrow_volatile_bitfield): Note that the default
2817 value is false, not !TARGET_STRICT_ALIGN.
2818 * doc/tm.texi: Regenerate.
2819
8edb8dc8
PC		 28202013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
2821
2822 Fix LIB_SPEC for systems without libpthread.
2823
2824 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
2825 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
2826 for Android.
2827 * config/i386/linux-common.h: Likewise.
2828 * config/mips/linux-common.h: Likewise.
2829
12211b99 28302013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ		 2831
2832 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
2833 checks.
2834
e42c64cb
DM		 28352013-08-20 David Malcolm <dmalcolm@redhat.com>
2836
2837 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2838 instances can own GC refs.
2839
2840 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2841 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2842 (gcc::context::gt_pch_nx): Likewise.
2843 (gcc::context::gt_pch_nx): Likewise.
2844 * ggc.h (gt_ggc_mx <T>): New.
2845 (gt_pch_nx_with_op <T>): New.
2846 (gt_pch_nx <T>): New.
2847 * passes.c (opt_pass::gt_ggc_mx): New.
2848 (opt_pass::gt_pch_nx): New.
2849 (opt_pass::gt_pch_nx_with_op): New.
2850 (pass_manager::gt_ggc_mx): New.
2851 (pass_manager::gt_pch_nx): New.
2852 (pass_manager::gt_pch_nx_with_op): New.
2853 (pass_manager::operator new): Use
2854 ggc_internal_cleared_alloc_stat rather than xcalloc.
2855 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2856 (pass_manager::gt_ggc_mx): New.
2857 (pass_manager::gt_pch_nx): New.
2858 (pass_manager::gt_pch_nx_with_op): New.
2859 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2860 (opt_pass::operator new): New.
2861 (opt_pass::gt_ggc_mx): New.
2862 (opt_pass::gt_pch_nx): New.
2863 (opt_pass::gt_pch_nx_with_op): New.
2864
910c02a0
DM		 28652013-08-20 David Malcolm <dmalcolm@redhat.com>
2866
2867 * Makefile.in (GTFILES): Add context.h.
2868 * context.c (gcc::context::operator new): New.
2869 (gcc::context::gt_ggc_mx): New.
2870 (gcc::context::gt_pch_nx): New.
2871 (gcc::context::gt_pch_nx): New.
2872 * context.h (gcc::context): Add GTY((user)) marking.
2873 (gcc::context::operator new): New.
2874 (gcc::context::gt_ggc_mx): New.
2875 (gcc::context::gt_pch_nx): New.
2876 (gcc::context::gt_pch_nx): New.
2877 (g): Add GTY marking.
2878 (gt_ggc_mx (gcc::context *)): New.
2879 (gt_pch_nx (gcc::context *)): New.
2880 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2881 void *cookie)): New.
2882 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2883
af4b4236
AM		 28842013-08-20 Alan Modra <amodra@gmail.com>
2885
2886 PR target/57865
2887 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
2888 (rs6000_emit_epilogue): Likewise.
2889
12211b99 28902013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC		 2891
2892 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
2893
e2323f5b
PB		 28942013-08-19 Peter Bergner <bergner@vnet.ibm.com>
2895 Jakub Jelinek <jakub@redhat.com>
2896
2897 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
2898 (BUILT_IN_FABSD64): Likewise.
2899 (BUILT_IN_FABSD128): Likewise.
2900 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
2901 (fold_builtin_1): Likewise.
5339fc0c
UB		 2902 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
2903 destination and source operands.
e2323f5b
PB		 2904 (*abstd2_fpr): Likewise.
2905 (*nabstd2_fpr): Likewise.
2906
cfbf3ee8
RS		 29072013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
2908
2909 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
2910 JUMP_P and INSN_P.
2911
d3136aeb
AH		 29122013-08-19 Aldy Hernandez <aldyh@redhat.com>
2913
2914 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
2915 incomplete.
2916
d33d9e47
AI		 29172013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
2918
2919 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
2920 * builtins.c (default_libc_has_function): New.
2921 (gnu_libc_has_function): Ditto.
2922 (no_c99_libc_has_function): Ditto.
2923 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
2924 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
2925 (fold_builtin_sincos): Likewise.
2926 (fold_builtin_cexp): Likewise.
2927 * builtins.def (DEF_C94_BUILTIN): Likewise.
2928 (DEF_C99_BUILTIN): Likewise.
2929 (DEF_C99_C90RES_BUILTIN): Likewise.
2930 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
2931 definitions to using this define.
2932 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 2933 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI		 2934 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
2935 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
2936 * config/darwin.h: Ditto.
2937 * config/elfos.h: Ditto.
2938 * config/freebsd.h: Ditto.
2939 * config/i386/cygming.h: Ditto.
2940 * config/i386/djgpp.h: Ditto.
2941 * config/i386/i386-interix.h: Ditto.
2942 * config/microblaze/microblaze.h: Ditto.
2943 * config/mmix/mmix.h: Ditto.
2944 * config/gnu-user.h: Ditto.
2945 * config/ia64/hpux.h: Ditto.
2946 * config/pa/pa-hpux.h: Ditto.
2947 * config/pdp11/pdp11.h: Ditto.
2948 * config/picochip/picochip.h: Ditto.
2949 * config/linux.h: Ditto.
2950 * config/netbsd.h: Ditto.
2951 * config/openbsd.h: Ditto.
2952 * config/rs6000/aix43.h: Ditto.
2953 * config/rs6000/aix51.h: Ditto.
2954 * config/rs6000/aix52.h: Ditto.
2955 * config/rs6000/aix53.h: Ditto.
2956 * config/rs6000/aix61.h: Ditto.
2957 * config/rs6000/darwin.h: Ditto.
2958 * config/rs6000/linux.h: Ditto.
2959 * config/rs6000/linux64.h: Ditto.
2960 * config/s390/tpf.h: Ditto.
2961 * config/sol2-10.h: Ditto.
2962 * config/sol2.h: Ditto.
2963 * config/vms/vms.h: Ditto.
2964 * config/vxworks.h: Ditto.
2965 * config/linux-android.c (linux_android_libc_has_function):
2966 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
2967 * config/linux-protos.h (linux_android_libc_has_function):
2968 New declaration.
2969 * config/i386/i386.c (ix86_libc_has_function): New.
2970 * config/i386/i386-protos.h
2971 (ix86_libc_has_function): New declaration.
2972 * config/i386/i386.md
2973 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
2974 ("isinf<mode>2): Likewise.
2975 * convert.c (convert_to_integer): Using new target hook
2976 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
2977 TARGET_C99_FUNCTIONS.
2978 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
2979 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
2980 * coretypes.h (function_class): New enum for different
2981 classes of functions.
2982 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
2983 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
2984 (TARGET_HAS_SINCOS): Likewise.
2985 (TARGET_LIBC_HAS_FUNCTION): New.
2986 * doc/tm.texi: Regenerated.
2987 * targhooks.h (default_libc_has_function): New declaration.
2988 (no_c99_libc_has_function): Ditto.
2989 (gnu_libc_has_function): Ditto.
2990 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
2991 and TARGET_HAS_SINCOS.
2992
eefe9a99
JH		 29932013-08-18 Jan Hubicka <jh@suse.cz>
2994
2995 * Makeifle-in (ipa-devirt.o): New.
2996 (GTFILES): Add ipa-utils.h and ipa-devirt.c
2997 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
2998 (analyze_functions): Look into possible targets of polymorphic call.
2999 * dumpfile.c (dump_files): Add type-inheritance dump.
3000 * dumpfile.h (TDI_inheritance): New.
3001 * ipa-devirt.c: New file.
3002 * ipa-utils.h (odr_type_d): Forward declare.
3003 (odr_type): New type.
3004 (build_type_inheritance_graph): Declare.
3005 (possible_polymorphic_call_targets): Declare and introduce inline
3006 variant when only edge is pased.
3007 (dump_possible_polymorphic_call_targets): Likewise.
3008 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3009 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3010 (types_same_for_odr): ... here.
3011 * tree.h (type_in_anonymous_namespace_p): Declare.
3012
29b89442
JJ		 30132013-08-18 Jakub Jelinek <jakub@redhat.com>
3014
3015 PR tree-optimization/58006
3016 * tree-parloops.c (take_address_of): Don't ICE if get_name
3017 returns NULL.
3018 (eliminate_local_variables_stmt): Remove clobber stmts.
3019
4f219961
EB		 30202013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3021
3022 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3023 error message is issued for an alias to undefined symbol.
3024
1d5755ef
JH		 30252013-08-18 Jan Hubicka <jh@suse.cz>
3026
3027 * cgraph.c (cgraph_create_indirect_edge): Discover
3028 polymorphic calls and record basic info into indirect_info.
3029 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3030 devirtualization, ignore objc function calls.
3031 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3032 call with no parm index info.
3033 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3034 * tree.c (virtual_method_call_p): New function.
3035 * tree.h (virtual_method_call_p): Declare.
3036
5945bebf
JH		 30372013-08-16 Jan Hubicka <jh@suse.cz>
3038
3039 PR middle-end/58179
3040 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3041
47fa5d23
DE		 30422013-08-16 David Edelsohn <dje.gcc@gmail.com>
3043
3044 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3045 attribute.
3046
c74f54a0
DM		 30472013-08-16 David Malcolm <dmalcolm@redhat.com>
3048
3049 * gengtype.c (type_for_name): Add special-case support for
3050 locating types within the "gcc::" namespace.
3051 (open_base_files): Emit a "using namespace gcc" directive.
3052
f758f299
MM		 30532013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3054
3055 PR target/58160
3056 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3057 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3058
3059 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3060 array instead of each individual operand as a separate argument.
3061 (emit_fusion_gpr_load): Likewise.
3062 (expand_fusion_gpr_load): Add new function declaration.
3063
3064 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3065 signature to have the operands passed as an array, instead of as
3066 separate arguments. Allow ZERO_EXTEND to be in the memory
3067 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3068 depend on the register live/dead flags when peepholes are run.
3069 (expand_fusion_gpr_load): New function to be called from the
3070 peephole2 pass, to change the register that addis sets to be the
3071 target register.
3072 (emit_fusion_gpr_load): Change the calling signature to have the
3073 operands passed as an array, instead of as separate arguments.
3074 Allow ZERO_EXTEND to be in the memory address, and also
3075 SIGN_EXTEND if -mpower8-fusion-sign.
3076
3077 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3078 unspec enumeration.
3079 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3080 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 3081 depend on the register live/dead state when the peephole pass is done.
f758f299 3082
158f4e4f
DM		 30832013-08-16 David Malcolm <dmalcolm@redhat.com>
3084
3085 * gengtype.c (create_user_defined_type): Ensure that the kind
3086 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3087 declaration is seen before the GTY((user)) marking.
3088
22f8cea5
BE		 30892013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
3090
3091 PR target/58105
3092 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
3093
c49bdb2e
JH		 30942013-08-16 Jan Hubicka <jh@suse.cz>
3095
3096 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
3097 arugment expected_type.
3098 (gimple_fold_call): Use it.
3099 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
3100 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 3101 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH		 3102 (try_make_edge_direct_virtual_call): Likewise.
3103 * tree.c (obj_type_ref_class): New.
3104 * tree.h (obj_type_ref_class): Use it.
3105
4042dca9
GDR		 31062013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
3107
3108 * sched-vis.c (rtl_slim_pp_initialized): Remove.
3109 (rtl_slim_pp): Likewise.
3110 (init_rtl_slim_pretty_print): Likewise.
3111 (dump_value_slim): Don't call it. Use local pretty printer.
3112 (dump_insn_slim): Likewise.
3113 (dump_rtl_slim): Likewise.
3114 (str_pattern_slim): Likewise.
3115 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
3116 Simplify.
3117
7a460594
JJ		 31182013-08-16 Jakub Jelinek <jakub@redhat.com>
3119
639dc669
JJ		 3120 PR tree-optimization/58164
3121 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
3122 walk gimple_goto_dest of GIMPLE_GOTO.
3123
7a460594
JJ		 3124 PR tree-optimization/58165
3125 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
3126 bi_call must be the last stmt in a bb, don't split_block, instead
3127 use fallthru edge from it and give up if there is none.
3128 Release conds vector when returning early.
3129
e3c02cc7
XDL		 31302013-08-14 Xinliang David Li <davidxl@google.com>
3131
3132 * config/i386/i386.c (ix86_option_override_internal):
3133 Remove unused variable and field.
3134
6469da2c
BS		 31352013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3136
3137 PR target/57949
22f8cea5 3138 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS		 3139 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
3140 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
3141 and Linux, correct BLKmode alignment when 128-bit alignment is
3142 required and compatibility flag is not set.
22f8cea5
BE		 3143 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
3144 for zero-size arguments when compatibility flag is not set.
6469da2c 3145
4ca890e2
JJ		 31462013-08-14 Jakub Jelinek <jakub@redhat.com>
3147
3148 PR tree-optimization/58145
3149 * tree-sra.c (build_ref_for_offset): If prev_base has
3150 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
3151
daec1759 31522013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 3153
daec1759 3154 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 3155 Fix uninitialized variable error.
daec1759 3156
3ad20bd4
XDL		 31572013-08-14 Xinliang David Li <davidxl@google.com>
3158
3159 * config/i386/i386.opt: Define two new options.
3160 * config/i386/x86-tune.def: Add arch selector field in macros.
3161 * config/i386/i386.h: Adjust macro definition.
3162 * config/i386/i386.c (ix86_option_override_internal):
3163 Refactor the code.
3164 (parse_mtune_ctrl_str): New function.
3165 (set_ix86_tune_features): New function.
3166 (ix86_function_specific_restore): Call the new helper function.
3167
e839e2a9
AB		 31682013-08-14 Andrey Belevantsev <abel@ispras.ru>
3169
3170 PR rtl-optimization/57662
3171 * sel-sched.c (code_motion_process_successors): When the current insn
3172 is removed after the recursive traversal, break from the loop.
3173 Add comments and debug printouts.
3174
27ed665f
JJ		 31752013-08-14 Jakub Jelinek <jakub@redhat.com>
3176 Alexandre Oliva <aoliva@redhat.com>
3177
3178 PR target/58067
3179 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3180 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3181 there also UNSPEC_PLTOFF.
3182
61c1a609
MP		 31832013-08-14 Marek Polacek <polacek@redhat.com>
3184
3185 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3186 NUM_CONDITIONS bit positions.
3187
0e901c67
CC		 31882013-08-13 Cary Coutant <ccoutant@google.com>
3189
3190 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3191 (attr_checksum): Hash vector contents instead of pointer.
3192 (attr_checksum_ordered): Likewise.
3193
061eff6d
UB		 31942013-08-13 Uros Bizjak <ubizjak@gmail.com>
3195
3196 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3197 when Pmode != word_mode. Add length_address attribute.
3198 (sse3_monitor_<mode>): Merge from sse3_monitor and
3199 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3200 Pmode != word_mode. Update insn length attribute.
3201 * config/i386/i386.c (ix86_option_override_internal): Update
3202 ix86_gen_monitor selection for merged sse3_monitor insn.
3203
8ec9249b
JB		 32042013-08-13 Julian Brown <julian@codesourcery.com>
3205
3206 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3207 perform invalid legitimization on greater-than-word-size modes for
3208 TARGET_E500_DOUBLE.
3209
761a8eb7
VM		 32102013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3211
3212 * ira.c (setup_class_translate_array): Use aclass instead of cl
3213 for classes not fully covered by allocno classes.
3214
99c2bd54
JJ		 32152013-08-13 Jakub Jelinek <jakub@redhat.com>
3216
4029a5e0
JJ		 3217 PR tree-optimization/57661
3218 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3219 * tree-inline.c (tree_function_versioning): Initialize it.
3220 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3221 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3222 that is not being copied.
3223
99c2bd54
JJ		 3224 PR sanitizer/56417
3225 * asan.c (instrument_strlen_call): Fix typo in comment.
3226 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3227
b5f8f063
SE		 32282013-08-13 Steve Ellcey <sellcey@mips.com>
3229
3230 * config/mips/mips.md (prefetch): Use lw instead of ld on
3231 loongson in 32bit mode.
3232
4de80584
NC		 32332013-08-13 Nick Clifton <nickc@redhat.com>
3234
3235 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3236
09ce3660
JH		 32372013-08-13 Jan Hubicka <jh@suse.cz>
3238
3239 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3240 introduced edge; fix typo in sanity check.
3241 (cgraph_resolve_speculation): Export; improve diagnostic.
3242 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3243 speculation at type mismatch.
3244 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3245 (cgraph_resolve_speculation): Declare.
3246 (symtab_can_be_discarded): New function.
3247 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3248 * ipa-inline-transform.c (speculation_removed): New global var.
3249 (clone_inlined_nodes): See if speculation can be removed.
3250 (inline_call): If speculations was removed, we growths may not match.
3251 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3252 (speculation_useful_p): New function.
3253 (resolve_noninline_speculation): New function.
3254 (inline_small_functions): Resolve useless speculations.
3255 * ipa-inline.h (speculation_useful_p): Declare
3256 * ipa.c (can_replace_by_local_alias): Simplify.
3257 (ipa_profile): Produce speculative calls in non-lto, too;
3258 add simple cost model; produce local aliases.
3259
f971dc24
DM		 32602013-08-13 David Malcolm <dmalcolm@redhat.com>
3261
3262 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 3263 PASS_MANAGER_H.
f971dc24 3264
2e507b9a
PC		 32652013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3266
3267 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3268
c97fd4b8
UB		 32692013-08-12 Uros Bizjak <ubizjak@gmail.com>
3270
3271 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3272 iterator instead of X87MODEF.
3273
32742013-08-12 Perez Read <netfirewall@gmail.com>
3275
3276 PR target/58132
3277 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3278 operand 0 for intel asm alternative.
3279 (*movabs<mode>_2): Ditto for operand 1.
3280
f4f4204c
JG		 32812013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3282
3283 * config/aarch64/arm_none.h
3284 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3285
133b1a8e
NC		 32862013-08-12 Nick Clifton <nickc@redhat.com>
3287
3288 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3289
15dd8b3a
YR		 32902013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3291
3292 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3293 expand for QI/HImode operand to produce more effictive code for
3294 unsigned char(short) --> float(double) conversion.
3295
94109a6a
AM		 32962013-08-12 Alexander Monakov <amonakov@ispras.ru>
3297
3298 * doc/invoke.texi: Mention that -ftls-model does not force the final
3299 model.
3300
4e115102
MP		 33012013-08-12 Marek Polacek <polacek@redhat.com>
3302 Marc Glisse <marc.glisse@inria.fr>
3303
3304 PR tree-optimization/57980
3305 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3306 when creating -1 constant.
3307
748d71f3
JH		 33082013-08-10 Jan Hubicka <jh@suse.cz>
3309
1f97e193 3310 Workaround binutils PR14342.
748d71f3
JH		 3311 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3312 (gimple_init_edge_profiler): Likewise.
3313 (gimple_gen_ic_func_profiler): Likewise.
3314
5979aa54
JH		 33152013-08-09 Jan Hubicka <jh@suse.cz>
3316
3317 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3318
ad83025e
XDL		 33192013-08-09 Xinliang David Li <davidxl@google.com>
3320
3321 * config/i386/stringop.def: New file.
3322 * config/i386/stringop.opt: New file.
3323 * config/i386/i386-opts.h: Include stringopt.def.
3324 * config/i386/i386.opt: Include stringopt.opt.
3325 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 3326 Override default size based stringop inline strategies with options.
ad83025e
XDL		 3327 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3328 New function.
3329
e59df5fd
JH		 33302013-08-09 Jan Hubicka <jh@suse.cz>
3331
c97fd4b8 3332 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 3333
634ab819
JH		 33342013-08-09 Jan Hubicka <jh@suse.cz>
3335
3336 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3337 CGRAPH_FREQ_MAX.
3338 (dump_cgraph_node): Dump profile-id.
3339 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3340 and common_target_probability.
3341 * lto-cgraph.c (lto_output_edge): Stream common targets.
3342 (lto_output_node): Stream profile ids.
3343 (input_node): Stream profile ids.
3344 (input_edge): Stream common targets.
3345 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3346 * ipa.c: Include value-prof.h
3347 (ipa_profile_generate_summary): Turn indirect call statement histograms
3348 into common targets.
3349 (ipa_profile): Turn common targets into speculative edges.
3350
2fa3d31b
JH		 33512013-08-09 Jan Hubicka <jh@suse.cz>
3352
3353 * cgraph.h (cgraph_node): Add profile_id.
3354 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 3355 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH		 3356 (del_node_map): Update.
3357 (find_func_by_funcdef_no): Replace by ...
3358 (find_func_by_profile_id): ... this one.
3359 (gimple_ic_transform): Do not remove useful histograms when
3360 speculation is not done; dump info when indirect call removal
3361 can happen at LTO.
3362 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
3363 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
3364 (__gcov_indirect_call_profiler_v2): .. this one.
3365 * profile.h (init_node_map): Update.
3366 * coverage.c (coverage_compute_profile_id): New function.
3367 * coverage.h (coverage_compute_profile_id): Declare.
3368 * tree-profile.c (init_ic_make_global_vars): Make
3369 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
3370 (gimple_init_edge_profiler): Update prototype of
3371 __gcov_indirect_call_profiler.
3372 (gimple_gen_ic_func_profiler): Simplify.
3373 (tree_profiling): Use init_node_map
3374
042ae7d2
JH		 33752013-08-09 Jan Hubicka <jh@suse.cz>
3376
c97fd4b8
UB		 3377 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
3378 non-speculative refs.
042ae7d2
JH		 3379 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
3380 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
3381 (cgraph_set_call_stmt): Likewise.
3382 (cgraph_create_edge_1): Fix release checking compilatoin;
3383 clear lto_stmt_uid.
3384 (cgraph_free_edge): Free indirect info.
3385 (cgraph_turn_edge_to_speculative): New function.
3386 (cgraph_speculative_call_info): New function.
3387 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 3388 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH		 3389 (dump_cgraph_node): Dump speculation.
3390 (verify_edge_count_and_frequency): Accept speculative edges.
3391 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
3392 (verify_cgraph_node): Handle speculation.
3393 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
3394 (cgraph_set_call_stmt): Update prototype.
3395 (cgraph_make_edge_direct): Update prototype.
3396 (cgraph_speculative_call_info): Declare.
3397 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
3398 to change; update call of ipa_find_references.
3399 * ipa-ref.c (ipa_record_reference): Fix return value; clear
3400 lto_stmt_uid and speculative flags.
3401 (ipa_dump_references): Dump speculation.
3402 (ipa_clone_references): Clone speculative flag.
3403 (ipa_clone_referring): Likewise.
3404 (ipa_clone_ref): New function.
3405 (ipa_find_reference): Look into lto_stmt_uids
3406 (ipa_clear_stmts_in_references): Do not clear speculative calls.
3407 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
3408 (ipa_find_reference): Update declaration.
3409 (ipa_clone_ref): Declare.
3410 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
3411 stream speculative flag.
3412 (lto_output_ref): Stream statements uids and speculation.
3413 (input_ref): Likewise.
3414 (input_edge): Stream speuclation.
3415 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
3416 (cgraph_set_call_stmt_including_clones): Handle speculation.
3417 * ipa-inline.c (heap_edge_removal_hook): New function.
3418 (inline_small_functions): Register it.
3419 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
3420 also initialize refs.
3421 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
3422 edge to change.
3423 (try_make_edge_direct_simple_call): Likewise.
3424 (try_make_edge_direct_simple_call): Likewise.
3425 (update_indirect_edges_after_inlining): Likewise.
3426 (remove_described_reference): Look proper lto_stmt_uid.
3427 (propagate_controlled_uses): Likewise.
3428 (propagate_controlled_uses): Liekwise.
3429 * tree-inline.c (copy_bb): Copy speculative edges.
3430 (redirect_all_calls): New function.
c97fd4b8
UB		 3431 (copy_cfg_body): Do redirection after loop info is updated.
3432 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 3433
aa1e10cc
JH		 34342013-08-09 Jan Hubicka <jh@suse.cz>
3435
3436 * lto-streamer-out.c (output_function): Renumber PHIs.
3437 * lto-streamer-in.c (input_function): Likewise.
3438
66adb8eb
JG		 34392013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
3440
3441 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
3442 (get_lane_unsigned): Likewise.
3443 (dup_lane_scalar): Likewise.
3444 (get_lane): enable for VALL.
3445 * config/aarch64/aarch64-simd.md
3446 (aarch64_dup_lane_scalar<mode>): Remove.
3447 (aarch64_get_lane_signed<mode>): Likewise.
3448 (aarch64_get_lane_unsigned<mode>): Likewise.
3449 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
3450 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
3451 (aarch64_get_lane<mode>): Enable for all vector modes.
3452 (aarch64_get_lanedi): Remove misleading constraints.
3453 * config/aarch64/arm_neon.h
3454 (__aarch64_vget_lane_any): Define.
3455 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
3456 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
3457 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
3458 * config/aarch64/iterators.md (VDQQH): New.
3459 (VDQQHS): Likewise.
3460 (vwcore): Likewise.
3461
1f65ae7a
EB		 34622013-08-09 Eric Botcazou <ebotcazou@adacore.com>
3463
3464 * configure.ac: Add GAS check for LEON instructions on SPARC.
3465 * configure: Regenerate.
3466 * config.in: Likewise.
3467 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
3468 sparc*-*-* block.
3469 * config/sparc/sparc.opt (LEON, LEON3): New masks.
3470 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
3471 for LEON or LEON3.
3472 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
3473 (AS_LEON_FLAG): New macro.
3474 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
3475 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
3476 Deal with LEON and LEON3 for the memory model.
23a6cb78 3477 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB		 3478 (atomic_compare_and_swap<mode>_1): Likewise.
3479 (*atomic_compare_and_swap<mode>_1): Likewise.
3480
ff522f7f
ZC		 34812013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3482
3483 * config/arm/neon.md (vcond): Fix floating-point vector
3484 comparisons against 0.
3485
cb1cca12
VM		 34862013-08-08 Vladimir Makarov <vmakarov@redhat.com>
3487
3488 * lra-constraints.c (emit_spill_move): Remove assert.
3489 (process_alt_operands): Add more debugging
3490 output. Increase reject for spilling into memory. Decrease
3491 reject for reloading scratch.
3492 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
3493
a3719e31
SE		 34942013-08-08 Steve Ellcey <sellcey@mips.com>
3495
3496 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
3497 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
3498 micromips incompatible. Add nan2008.
3499 (MULTILIB_DIRNAMES): Add nan2008.
3500 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3501 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
3502 and micromips incompatible. Add nan2008.
3503 (MULTILIB_DIRNAMES): Add nan2008.
3504 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3505
f40423e2
RS		 35062013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
3507
3508 PR rtl-optimization/58079
3509 * combine.c (combine_simplify_rtx): Avoid using SUBST if
3510 simplify_comparison has widened a comparison with an integer.
3511
869b9125
KT		 35122013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3513
3514 * config/arm/neon.md (movmisalign<mode>): Disable when we
3515 don't allow unaligned accesses.
3516 (*movmisalign<mode>_neon_store): Likewise.
3517 (*movmisalign<mode>_neon_load): Likewise.
3518 (*movmisalign<mode>_neon_store): Likewise.
3519 (*movmisalign<mode>_neon_load): Likewise.
3520
71cafea9
JH		 35212013-08-08 Jan Hubicka <jh@suse.cz>
3522
3523 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
3524 (make_pass_rebuild_cgraph_edges): Also clear references.
3525 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
3526 * ipa-inline-transform.c (inline_transform): Remove all references
3527 after inlining.
c97fd4b8
UB		 3528 * cgraphunit.c (expand_function): Remove all references after
3529 expansion.
71cafea9
JH		 3530 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
3531 (ipa_find_reference): Rewrite to iterator.
3532 (remove_stmt_references): Likewise.
3533 (ipa_clear_stmts_in_references): New function.
3534 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB		 3535 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
3536 clear references.
71cafea9
JH		 3537 * ipa-split.c (split_function): Remove references in split function.
3538
27d2e612
RE		 35392013-08-08 Richard Earnshaw <rearnsha@arm.com>
3540
3541 PR target/57431
1f65ae7a 3542 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE		 3543 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
3544
5efc661c
RE		 35452013-08-08 Richard Earnshaw <rearnsha@arm.com>
3546
3547 PR target/56979
1f65ae7a 3548 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE		 3549 suggested mode for the assignment isn't compatible with the
3550 registers required.
3551
f276d31d
BE		 35522013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
3553
3554 PR target/58065
3555 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
3556
4b8bc035
XDL		 35572013-08-07 Xinliang David Li <davidxl@google.com>
3558
3559 * config/i386/i386.opt: New option -mtune-ctrl=.
3560 * config/i386/x86-tune.def: New file.
3561 * config/i386/i386.h: include x86-tune.def.
3562 * config/i386/i386.c (ix86_option_override_internal):
3563 Parsing -mtune-ctrl= option and set tune features.
3564
12211b99 35652013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE		 3566
3567 PR other/12081
03b0ee0a 3568 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE		 3569 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
3570 to gen_2arg_fn_t.
3571
5cf6635b
EB		 35722013-08-07 Eric Botcazou <ebotcazou@adacore.com>
3573
3574 * rtl.h (update_alignments): Declare.
3575 * final.c (grow_label_align): New function extracted from...
3576 (shorten_branches): ...here. Call it.
3577 (update_alignments): New function.
3578 * reorg.c (sibling_labels): New variable.
3579 (get_label_before): Add SIBLING parameter. If it is non-zero, push
3580 the new label along with it onto the sibling_labels vector.
3581 (fill_simple_delay_slots): Adjust call to get_label_before.
3582 (fill_slots_from_thread): Likewise.
3583 (relax_delay_slots): Likewise.
3584 (make_return_insns): Likewise.
3585 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
3586
3c8ca1ab
EB		 35872013-08-07 Eric Botcazou <ebotcazou@adacore.com>
3588
3589 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
3590 document its semantics.
3591 (diagnostic_report_diagnostic): Adjust accordingly.
3592
5ee5b32c
DM		 35932013-08-07 David Malcolm <dmalcolm@redhat.com>
3594
03b0ee0a 3595 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM		 3596 (sparc_option_override): ...and port to new C++ pass API.
3597 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
3598
abd566fa
PB		 35992013-08-07 Peter Bergner <bergner@vnet.ibm.com>
3600
3601 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
3602
2077db1b
CT		 36032013-08-06 Caroline Tice <cmtice@google.com>
3604
3605 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
3606 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
3607 * tree-pass.h: Add pass_vtable_verify.
3608 * varasm.c (assemble_variable): Add code to properly set the comdat
3609 section and name for the .vtable_map_vars section.
3610 (assemble_vtyv_preinit_initializer): New function.
3611 (default_sectin_type_flags): Make sure .vtable_map_vars section has
3612 LINK_ONCE flag.
3613 * output.h: Add function decl for assemble_vtv_preinit_initializer.
3614 * vtable-verify.c: New file.
3615 * vtable-verify.h: New file.
3616 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
3617 initialiation levels.
3618 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
3619 * passes.def: Insert pass_vtable_verify.
3620 * aclocal.m4: Reorder includes.
3621 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
3622 -fvtv-counts options.
3623 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
3624 as appropriate, if -fvtable-verify=... is used.
3625 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
3626 -fvtable-verify=... is used.
3627 * Makefile.in (OBJS): Add vtable-verify.o to list.
3628 (vtable-verify.o): Add new build rule.
3629 (GTFILES): Add vtable-verify.c to list.
3630 * common.opt (fvtable-verify=): New flag.
3631 (vtv_priority): Values for fvtable-verify= flag.
3632 (fvtv-counts): New flag.
3633 (fvtv-debug): New flag.
3634 * tree.h (save_vtable_map_decl): New extern function decl.
3635
03085d1c
DM		 36362013-08-07 David Malcolm <dmalcolm@redhat.com>
3637
3638 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
3639 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
3640 (pass_data_rl78_devirt): ...new pass_data instance and...
3641 (make_pass_rl78_devirt): ...new function.
3642 (rl78_asm_file_start): Port pass registration to new C++ API.
3643
05555c4a
DM		 36442013-08-07 David Malcolm <dmalcolm@redhat.com>
3645
3646 * coretypes.h (rtl_opt_pass): Add.
3647 (gcc::context): Add.
3648 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
3649 (epiphany_init): Port to new C++ pass API.
3650 (epiphany_optimize_mode_switching): Likewise.
3651 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
3652 (pass_manager::get_pass_mode_switching): New.
3653 (pass_manager::get_pass_peephole2): New.
3654 * mode-switching.c (pass_mode_switching): Add clone method.
3655 * recog.c (pass_peephole2): Add clone method.
3656 (pass_split_all_insns): Add clone method.
3657
c5a12e26
DM		 36582013-08-06 David Malcolm <dmalcolm@redhat.com>
3659
03b0ee0a
UB		 3660 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
3661 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 3662
a2e2a668
JH		 36632013-08-06 Jan Hubicka <jh@suse.cz>
3664
3665 * cgraph.c (cgraph_get_body): New function based on lto.c
3666 implementation.
3667 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB		 3668 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
3669 LTO paths.
a2e2a668
JH		 3670 * cgraphunit.c (expand_function): Get body prior expanding.
3671 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB		 3672 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
3673 really need.
a2e2a668
JH		 3674 * passes.c (do_per_function_toporder): Get body.
3675 * tree-inline.c (expand_call_inline): Get body prior inlining it.
3676 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
3677
15682f24
MJ		 36782013-08-06 Martin Jambor <mjambor@suse.cz>
3679
3680 PR fortran/57987
3681 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
3682 is not re-finalized. Rename second parameter to no_collect.
3683
78f6dd68
MJ		 36842013-08-06 Martin Jambor <mjambor@suse.cz>
3685
3686 PR middle-end/58041
3687 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
3688 MEM_REF has proper alignment information.
3689
12211b99 36902013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE		 3691
3692 PR other/12081
3693 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
3694 class insn_gen_fn.
3695 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
3696 rtx (*) (rtx, ...) with insn_gen_fn.
3697 * genoutput.c (output_insn_data): Cast gen_? function pointers to
3698 insn_gen_fn::stored_funcptr. Add initializer braces.
3699
8ac69a6c
DM		 37002013-08-05 David Malcolm <dmalcolm@redhat.com>
3701
3702 Rewrite how instances of passes are cloned to remove assumptions
3703 about their sizes (thus allowing pass subclasses to have
3704 additional data fields, albeit non-GC-managed ones at this point).
3705
3706 * passes.c (make_pass_instance): Now that passes have clone
3707 methods, rewrite this function to eliminate XNEW and memcpy
3708 calls that used hardcoded sizes. Since this function no longer
3709 creates pass instances, rename it to...
03b0ee0a
UB		 3710 (add_pass_instance): ...this. Document the old way that passes were
3711 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM		 3712 (next_pass_1): Add an initial_pass argument for use by
3713 add_pass_instance.
3714 (position_pass): When adding multiple instances of a pass, use
3715 the pass's clone method, rather than relying on the XNEW/memcpy
3716 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB		 3717 (pass_manager::pass_manager): When invoking next_pass_1, also supply
3718 the initial instance of the current pass within the pass manager.
8ac69a6c 3719
27a4cd48
DM		 37202013-08-05 David Malcolm <dmalcolm@redhat.com>
3721
3722 This is the automated part of the conversion of passes from C
3723 structs to C++ classes.
3724
3725 Patch autogenerated by refactor_passes.py from
3726 https://github.com/davidmalcolm/gcc-refactoring-scripts
3727 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
3728
3729 * asan.c (pass_asan): Convert from a global struct to a subclass of
3730 gimple_opt_pass along with...
3731 (pass_data_asan): ...new pass_data instance and...
3732 (make_pass_asan): ...new function.
3733 (pass_asan_O0): Convert from a global struct to a subclass of
3734 gimple_opt_pass along with...
3735 (pass_data_asan_O0): ...new pass_data instance and...
3736 (make_pass_asan_O0): ...new function.
3737 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
3738 subclass of rtl_opt_pass along with...
3739 (pass_data_inc_dec): ...new pass_data instance and...
3740 (make_pass_inc_dec): ...new function.
3741 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
3742 a subclass of rtl_opt_pass along with...
3743 (pass_data_reorder_blocks): ...new pass_data instance and...
3744 (make_pass_reorder_blocks): ...new function.
3745 (pass_duplicate_computed_gotos): Convert from a global struct to a
3746 subclass of rtl_opt_pass along with...
3747 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
3748 (make_pass_duplicate_computed_gotos): ...new function.
3749 (pass_partition_blocks): Convert from a global struct to a subclass of
3750 rtl_opt_pass along with...
3751 (pass_data_partition_blocks): ...new pass_data instance and...
3752 (make_pass_partition_blocks): ...new function.
3753 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
3754 struct to a subclass of rtl_opt_pass along with...
3755 (pass_data_branch_target_load_optimize1): ...new pass_data instance
3756 and...
3757 (make_pass_branch_target_load_optimize1): ...new function.
3758 (pass_branch_target_load_optimize2): Convert from a global struct to a
3759 subclass of rtl_opt_pass along with...
3760 (pass_data_branch_target_load_optimize2): ...new pass_data instance
3761 and...
3762 (make_pass_branch_target_load_optimize2): ...new function.
3763 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
3764 of rtl_opt_pass along with...
3765 (pass_data_jump): ...new pass_data instance and...
3766 (make_pass_jump): ...new function.
3767 (pass_jump2): Convert from a global struct to a subclass of
3768 rtl_opt_pass along with...
3769 (pass_data_jump2): ...new pass_data instance and...
3770 (make_pass_jump2): ...new function.
3771 * cfgexpand.c (pass_expand): Convert from a global struct to a
3772 subclass of rtl_opt_pass along with...
3773 (pass_data_expand): ...new pass_data instance and...
3774 (make_pass_expand): ...new function.
3775 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
3776 of rtl_opt_pass along with...
3777 (pass_data_free_cfg): ...new pass_data instance and...
3778 (make_pass_free_cfg): ...new function.
3779 (pass_into_cfg_layout_mode): Convert from a global struct to a
3780 subclass of rtl_opt_pass along with...
3781 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
3782 (make_pass_into_cfg_layout_mode): ...new function.
3783 (pass_outof_cfg_layout_mode): Convert from a global struct to a
3784 subclass of rtl_opt_pass along with...
3785 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
3786 (make_pass_outof_cfg_layout_mode): ...new function.
3787 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
3788 struct to a subclass of gimple_opt_pass along with...
3789 (pass_data_build_cgraph_edges): ...new pass_data instance and...
3790 (make_pass_build_cgraph_edges): ...new function.
3791 (pass_rebuild_cgraph_edges): Convert from a global struct to a
3792 subclass of gimple_opt_pass along with...
3793 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
3794 (make_pass_rebuild_cgraph_edges): ...new function.
3795 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
3796 subclass of gimple_opt_pass along with...
3797 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
3798 and...
3799 (make_pass_remove_cgraph_callee_edges): ...new function.
3800 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
3801 struct to a subclass of rtl_opt_pass along with...
3802 (pass_data_stack_adjustments): ...new pass_data instance and...
3803 (make_pass_stack_adjustments): ...new function.
3804 * combine.c (pass_combine): Convert from a global struct to a subclass
3805 of rtl_opt_pass along with...
3806 (pass_data_combine): ...new pass_data instance and...
3807 (make_pass_combine): ...new function.
3808 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
3809 global struct to a subclass of rtl_opt_pass along with...
3810 (pass_data_compare_elim_after_reload): ...new pass_data instance
3811 and...
3812 (make_pass_compare_elim_after_reload): ...new function.
3813 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
3814 of rtl_opt_pass along with...
3815 (pass_data_rtl_cprop): ...new pass_data instance and...
3816 (make_pass_rtl_cprop): ...new function.
3817 * cse.c (pass_cse): Convert from a global struct to a subclass of
3818 rtl_opt_pass along with...
3819 (pass_data_cse): ...new pass_data instance and...
3820 (make_pass_cse): ...new function.
3821 (pass_cse2): Convert from a global struct to a subclass of
3822 rtl_opt_pass along with...
3823 (pass_data_cse2): ...new pass_data instance and...
3824 (make_pass_cse2): ...new function.
3825 (pass_cse_after_global_opts): Convert from a global struct to a
3826 subclass of rtl_opt_pass along with...
3827 (pass_data_cse_after_global_opts): ...new pass_data instance and...
3828 (make_pass_cse_after_global_opts): ...new function.
3829 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
3830 of rtl_opt_pass along with...
3831 (pass_data_ud_rtl_dce): ...new pass_data instance and...
3832 (make_pass_ud_rtl_dce): ...new function.
3833 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
3834 rtl_opt_pass along with...
3835 (pass_data_fast_rtl_dce): ...new pass_data instance and...
3836 (make_pass_fast_rtl_dce): ...new function.
3837 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
3838 a subclass of rtl_opt_pass along with...
3839 (pass_data_df_initialize_opt): ...new pass_data instance and...
3840 (make_pass_df_initialize_opt): ...new function.
3841 (pass_df_initialize_no_opt): Convert from a global struct to a
3842 subclass of rtl_opt_pass along with...
3843 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
3844 (make_pass_df_initialize_no_opt): ...new function.
3845 (pass_df_finish): Convert from a global struct to a subclass of
3846 rtl_opt_pass along with...
3847 (pass_data_df_finish): ...new pass_data instance and...
3848 (make_pass_df_finish): ...new function.
3849 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
3850 rtl_opt_pass along with...
3851 (pass_data_rtl_dse1): ...new pass_data instance and...
3852 (make_pass_rtl_dse1): ...new function.
3853 (pass_rtl_dse2): Convert from a global struct to a subclass of
3854 rtl_opt_pass along with...
3855 (pass_data_rtl_dse2): ...new pass_data instance and...
3856 (make_pass_rtl_dse2): ...new function.
3857 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
3858 subclass of rtl_opt_pass along with...
3859 (pass_data_dwarf2_frame): ...new pass_data instance and...
3860 (make_pass_dwarf2_frame): ...new function.
3861 * except.c (pass_set_nothrow_function_flags): Convert from a global
3862 struct to a subclass of rtl_opt_pass along with...
3863 (pass_data_set_nothrow_function_flags): ...new pass_data instance
3864 and...
3865 (make_pass_set_nothrow_function_flags): ...new function.
3866 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
3867 subclass of rtl_opt_pass along with...
3868 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
3869 and...
3870 (make_pass_convert_to_eh_region_ranges): ...new function.
3871 * final.c (pass_compute_alignments): Convert from a global struct to a
3872 subclass of rtl_opt_pass along with...
3873 (pass_data_compute_alignments): ...new pass_data instance and...
3874 (make_pass_compute_alignments): ...new function.
3875 (pass_final): Convert from a global struct to a subclass of
3876 rtl_opt_pass along with...
3877 (pass_data_final): ...new pass_data instance and...
3878 (make_pass_final): ...new function.
3879 (pass_shorten_branches): Convert from a global struct to a subclass of
3880 rtl_opt_pass along with...
3881 (pass_data_shorten_branches): ...new pass_data instance and...
3882 (make_pass_shorten_branches): ...new function.
3883 (pass_clean_state): Convert from a global struct to a subclass of
3884 rtl_opt_pass along with...
3885 (pass_data_clean_state): ...new pass_data instance and...
3886 (make_pass_clean_state): ...new function.
3887 * function.c (pass_instantiate_virtual_regs): Convert from a global
3888 struct to a subclass of rtl_opt_pass along with...
3889 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
3890 (make_pass_instantiate_virtual_regs): ...new function.
3891 (pass_leaf_regs): Convert from a global struct to a subclass of
3892 rtl_opt_pass along with...
3893 (pass_data_leaf_regs): ...new pass_data instance and...
3894 (make_pass_leaf_regs): ...new function.
3895 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
3896 subclass of rtl_opt_pass along with...
3897 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
3898 and...
3899 (make_pass_thread_prologue_and_epilogue): ...new function.
3900 (pass_match_asm_constraints): Convert from a global struct to a
3901 subclass of rtl_opt_pass along with...
3902 (pass_data_match_asm_constraints): ...new pass_data instance and...
3903 (make_pass_match_asm_constraints): ...new function.
3904 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
3905 subclass of rtl_opt_pass along with...
3906 (pass_data_rtl_fwprop): ...new pass_data instance and...
3907 (make_pass_rtl_fwprop): ...new function.
3908 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
3909 rtl_opt_pass along with...
3910 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
3911 (make_pass_rtl_fwprop_addr): ...new function.
3912 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
3913 rtl_opt_pass along with...
3914 (pass_data_rtl_pre): ...new pass_data instance and...
3915 (make_pass_rtl_pre): ...new function.
3916 (pass_rtl_hoist): Convert from a global struct to a subclass of
3917 rtl_opt_pass along with...
3918 (pass_data_rtl_hoist): ...new pass_data instance and...
3919 (make_pass_rtl_hoist): ...new function.
3920 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
3921 subclass of gimple_opt_pass along with...
3922 (pass_data_lower_cf): ...new pass_data instance and...
3923 (make_pass_lower_cf): ...new function.
3924 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
3925 from a global struct to a subclass of gimple_opt_pass along with...
3926 (pass_data_strength_reduction): ...new pass_data instance and...
3927 (make_pass_strength_reduction): ...new function.
3928 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
3929 of rtl_opt_pass along with...
3930 (pass_data_rtl_ifcvt): ...new pass_data instance and...
3931 (make_pass_rtl_ifcvt): ...new function.
3932 (pass_if_after_combine): Convert from a global struct to a subclass of
3933 rtl_opt_pass along with...
3934 (pass_data_if_after_combine): ...new pass_data instance and...
3935 (make_pass_if_after_combine): ...new function.
3936 (pass_if_after_reload): Convert from a global struct to a subclass of
3937 rtl_opt_pass along with...
3938 (pass_data_if_after_reload): ...new pass_data instance and...
3939 (make_pass_if_after_reload): ...new function.
3940 * init-regs.c (pass_initialize_regs): Convert from a global struct to
3941 a subclass of rtl_opt_pass along with...
3942 (pass_data_initialize_regs): ...new pass_data instance and...
3943 (make_pass_initialize_regs): ...new function.
3944 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
3945 of ipa_opt_pass_d along with...
3946 (pass_data_ipa_cp): ...new pass_data instance and...
3947 (make_pass_ipa_cp): ...new function.
3948 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
3949 global struct to a subclass of gimple_opt_pass along with...
3950 (pass_data_inline_parameters): ...new pass_data instance and...
3951 (make_pass_inline_parameters): ...new function.
3952 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
3953 subclass of gimple_opt_pass along with...
3954 (pass_data_early_inline): ...new pass_data instance and...
3955 (make_pass_early_inline): ...new function.
3956 (pass_ipa_inline): Convert from a global struct to a subclass of
3957 ipa_opt_pass_d along with...
3958 (pass_data_ipa_inline): ...new pass_data instance and...
3959 (make_pass_ipa_inline): ...new function.
3960 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
3961 struct to a subclass of gimple_opt_pass along with...
3962 (pass_data_local_pure_const): ...new pass_data instance and...
3963 (make_pass_local_pure_const): ...new function.
3964 (pass_ipa_pure_const): Convert from a global struct to a subclass of
3965 ipa_opt_pass_d along with...
3966 (pass_data_ipa_pure_const): ...new pass_data instance and...
3967 (make_pass_ipa_pure_const): ...new function.
3968 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
3969 to a subclass of ipa_opt_pass_d along with...
3970 (pass_data_ipa_reference): ...new pass_data instance and...
3971 (make_pass_ipa_reference): ...new function.
3972 * ipa-split.c (pass_split_functions): Convert from a global struct to
3973 a subclass of gimple_opt_pass along with...
3974 (pass_data_split_functions): ...new pass_data instance and...
3975 (make_pass_split_functions): ...new function.
3976 (pass_feedback_split_functions): Convert from a global struct to a
3977 subclass of gimple_opt_pass along with...
3978 (pass_data_feedback_split_functions): ...new pass_data instance and...
3979 (make_pass_feedback_split_functions): ...new function.
3980 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
3981 global struct to a subclass of simple_ipa_opt_pass along with...
3982 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
3983 instance and...
3984 (make_pass_ipa_function_and_variable_visibility): ...new function.
3985 (pass_ipa_free_inline_summary): Convert from a global struct to a
3986 subclass of simple_ipa_opt_pass along with...
3987 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
3988 (make_pass_ipa_free_inline_summary): ...new function.
3989 (pass_ipa_whole_program_visibility): Convert from a global struct to a
3990 subclass of ipa_opt_pass_d along with...
3991 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
3992 and...
3993 (make_pass_ipa_whole_program_visibility): ...new function.
3994 (pass_ipa_profile): Convert from a global struct to a subclass of
3995 ipa_opt_pass_d along with...
3996 (pass_data_ipa_profile): ...new pass_data instance and...
3997 (make_pass_ipa_profile): ...new function.
3998 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
3999 ipa_opt_pass_d along with...
4000 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4001 (make_pass_ipa_cdtor_merge): ...new function.
4002 * ira.c (pass_ira): Convert from a global struct to a subclass of
4003 rtl_opt_pass along with...
4004 (pass_data_ira): ...new pass_data instance and...
4005 (make_pass_ira): ...new function.
4006 (pass_reload): Convert from a global struct to a subclass of
4007 rtl_opt_pass along with...
4008 (pass_data_reload): ...new pass_data instance and...
4009 (make_pass_reload): ...new function.
4010 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4011 subclass of rtl_opt_pass along with...
4012 (pass_data_cleanup_barriers): ...new pass_data instance and...
4013 (make_pass_cleanup_barriers): ...new function.
4014 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4015 of rtl_opt_pass along with...
4016 (pass_data_loop2): ...new pass_data instance and...
4017 (make_pass_loop2): ...new function.
4018 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4019 rtl_opt_pass along with...
4020 (pass_data_rtl_loop_init): ...new pass_data instance and...
4021 (make_pass_rtl_loop_init): ...new function.
4022 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4023 rtl_opt_pass along with...
4024 (pass_data_rtl_loop_done): ...new pass_data instance and...
4025 (make_pass_rtl_loop_done): ...new function.
4026 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4027 subclass of rtl_opt_pass along with...
4028 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4029 (make_pass_rtl_move_loop_invariants): ...new function.
4030 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4031 rtl_opt_pass along with...
4032 (pass_data_rtl_unswitch): ...new pass_data instance and...
4033 (make_pass_rtl_unswitch): ...new function.
4034 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4035 subclass of rtl_opt_pass along with...
4036 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4037 and...
4038 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4039 (pass_rtl_doloop): Convert from a global struct to a subclass of
4040 rtl_opt_pass along with...
4041 (pass_data_rtl_doloop): ...new pass_data instance and...
4042 (make_pass_rtl_doloop): ...new function.
4043 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4044 a subclass of rtl_opt_pass along with...
4045 (pass_data_lower_subreg): ...new pass_data instance and...
4046 (make_pass_lower_subreg): ...new function.
4047 (pass_lower_subreg2): Convert from a global struct to a subclass of
4048 rtl_opt_pass along with...
4049 (pass_data_lower_subreg2): ...new pass_data instance and...
4050 (make_pass_lower_subreg2): ...new function.
4051 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4052 struct to a subclass of ipa_opt_pass_d along with...
4053 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4054 (make_pass_ipa_lto_gimple_out): ...new function.
4055 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4056 of ipa_opt_pass_d along with...
4057 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4058 (make_pass_ipa_lto_finish_out): ...new function.
4059 * mode-switching.c (pass_mode_switching): Convert from a global struct
4060 to a subclass of rtl_opt_pass along with...
4061 (pass_data_mode_switching): ...new pass_data instance and...
4062 (make_pass_mode_switching): ...new function.
4063 * modulo-sched.c (pass_sms): Convert from a global struct to a
4064 subclass of rtl_opt_pass along with...
4065 (pass_data_sms): ...new pass_data instance and...
4066 (make_pass_sms): ...new function.
4067 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4068 subclass of gimple_opt_pass along with...
4069 (pass_data_expand_omp): ...new pass_data instance and...
4070 (make_pass_expand_omp): ...new function.
4071 (pass_lower_omp): Convert from a global struct to a subclass of
4072 gimple_opt_pass along with...
4073 (pass_data_lower_omp): ...new pass_data instance and...
4074 (make_pass_lower_omp): ...new function.
4075 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4076 of gimple_opt_pass along with...
4077 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4078 (make_pass_diagnose_omp_blocks): ...new function.
4079 * passes.c (pass_early_local_passes): Convert from a global struct to
4080 a subclass of simple_ipa_opt_pass along with...
4081 (pass_data_early_local_passes): ...new pass_data instance and...
4082 (make_pass_early_local_passes): ...new function.
4083 (pass_all_early_optimizations): Convert from a global struct to a
4084 subclass of gimple_opt_pass along with...
4085 (pass_data_all_early_optimizations): ...new pass_data instance and...
4086 (make_pass_all_early_optimizations): ...new function.
4087 (pass_all_optimizations): Convert from a global struct to a subclass
4088 of gimple_opt_pass along with...
4089 (pass_data_all_optimizations): ...new pass_data instance and...
4090 (make_pass_all_optimizations): ...new function.
4091 (pass_all_optimizations_g): Convert from a global struct to a subclass
4092 of gimple_opt_pass along with...
4093 (pass_data_all_optimizations_g): ...new pass_data instance and...
4094 (make_pass_all_optimizations_g): ...new function.
4095 (pass_rest_of_compilation): Convert from a global struct to a subclass
4096 of rtl_opt_pass along with...
4097 (pass_data_rest_of_compilation): ...new pass_data instance and...
4098 (make_pass_rest_of_compilation): ...new function.
4099 (pass_postreload): Convert from a global struct to a subclass of
4100 rtl_opt_pass along with...
4101 (pass_data_postreload): ...new pass_data instance and...
4102 (make_pass_postreload): ...new function.
4103 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
4104 subclass of rtl_opt_pass along with...
4105 (pass_data_gcse2): ...new pass_data instance and...
4106 (make_pass_gcse2): ...new function.
4107 * postreload.c (pass_postreload_cse): Convert from a global struct to
4108 a subclass of rtl_opt_pass along with...
4109 (pass_data_postreload_cse): ...new pass_data instance and...
4110 (make_pass_postreload_cse): ...new function.
4111 * predict.c (pass_profile): Convert from a global struct to a subclass
4112 of gimple_opt_pass along with...
4113 (pass_data_profile): ...new pass_data instance and...
4114 (make_pass_profile): ...new function.
4115 (pass_strip_predict_hints): Convert from a global struct to a subclass
4116 of gimple_opt_pass along with...
4117 (pass_data_strip_predict_hints): ...new pass_data instance and...
4118 (make_pass_strip_predict_hints): ...new function.
4119 * recog.c (pass_peephole2): Convert from a global struct to a subclass
4120 of rtl_opt_pass along with...
4121 (pass_data_peephole2): ...new pass_data instance and...
4122 (make_pass_peephole2): ...new function.
4123 (pass_split_all_insns): Convert from a global struct to a subclass of
4124 rtl_opt_pass along with...
4125 (pass_data_split_all_insns): ...new pass_data instance and...
4126 (make_pass_split_all_insns): ...new function.
4127 (pass_split_after_reload): Convert from a global struct to a subclass
4128 of rtl_opt_pass along with...
4129 (pass_data_split_after_reload): ...new pass_data instance and...
4130 (make_pass_split_after_reload): ...new function.
4131 (pass_split_before_regstack): Convert from a global struct to a
4132 subclass of rtl_opt_pass along with...
4133 (pass_data_split_before_regstack): ...new pass_data instance and...
4134 (make_pass_split_before_regstack): ...new function.
4135 (pass_split_before_sched2): Convert from a global struct to a subclass
4136 of rtl_opt_pass along with...
4137 (pass_data_split_before_sched2): ...new pass_data instance and...
4138 (make_pass_split_before_sched2): ...new function.
4139 (pass_split_for_shorten_branches): Convert from a global struct to a
4140 subclass of rtl_opt_pass along with...
4141 (pass_data_split_for_shorten_branches): ...new pass_data instance
4142 and...
4143 (make_pass_split_for_shorten_branches): ...new function.
4144 * ree.c (pass_ree): Convert from a global struct to a subclass of
4145 rtl_opt_pass along with...
4146 (pass_data_ree): ...new pass_data instance and...
4147 (make_pass_ree): ...new function.
4148 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
4149 subclass of rtl_opt_pass along with...
4150 (pass_data_stack_regs): ...new pass_data instance and...
4151 (make_pass_stack_regs): ...new function.
4152 (pass_stack_regs_run): Convert from a global struct to a subclass of
4153 rtl_opt_pass along with...
4154 (pass_data_stack_regs_run): ...new pass_data instance and...
4155 (make_pass_stack_regs_run): ...new function.
4156 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4157 subclass of rtl_opt_pass along with...
4158 (pass_data_cprop_hardreg): ...new pass_data instance and...
4159 (make_pass_cprop_hardreg): ...new function.
4160 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4161 subclass of rtl_opt_pass along with...
4162 (pass_data_reginfo_init): ...new pass_data instance and...
4163 (make_pass_reginfo_init): ...new function.
4164 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4165 of rtl_opt_pass along with...
4166 (pass_data_regmove): ...new pass_data instance and...
4167 (make_pass_regmove): ...new function.
4168 * regrename.c (pass_regrename): Convert from a global struct to a
4169 subclass of rtl_opt_pass along with...
4170 (pass_data_regrename): ...new pass_data instance and...
4171 (make_pass_regrename): ...new function.
4172 * reorg.c (pass_delay_slots): Convert from a global struct to a
4173 subclass of rtl_opt_pass along with...
4174 (pass_data_delay_slots): ...new pass_data instance and...
4175 (make_pass_delay_slots): ...new function.
4176 (pass_machine_reorg): Convert from a global struct to a subclass of
4177 rtl_opt_pass along with...
4178 (pass_data_machine_reorg): ...new pass_data instance and...
4179 (make_pass_machine_reorg): ...new function.
4180 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4181 of rtl_opt_pass along with...
4182 (pass_data_sched): ...new pass_data instance and...
4183 (make_pass_sched): ...new function.
4184 (pass_sched2): Convert from a global struct to a subclass of
4185 rtl_opt_pass along with...
4186 (pass_data_sched2): ...new pass_data instance and...
4187 (make_pass_sched2): ...new function.
4188 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4189 to a subclass of rtl_opt_pass along with...
4190 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4191 (make_pass_stack_ptr_mod): ...new function.
4192 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4193 to a subclass of rtl_opt_pass along with...
4194 (pass_data_rtl_store_motion): ...new pass_data instance and...
4195 (make_pass_rtl_store_motion): ...new function.
4196 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4197 of gimple_opt_pass along with...
4198 (pass_data_tracer): ...new pass_data instance and...
4199 (make_pass_tracer): ...new function.
4200 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4201 to a subclass of gimple_opt_pass along with...
4202 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4203 (make_pass_diagnose_tm_blocks): ...new function.
4204 (pass_lower_tm): Convert from a global struct to a subclass of
4205 gimple_opt_pass along with...
4206 (pass_data_lower_tm): ...new pass_data instance and...
4207 (make_pass_lower_tm): ...new function.
4208 (pass_tm_init): Convert from a global struct to a subclass of
4209 gimple_opt_pass along with...
4210 (pass_data_tm_init): ...new pass_data instance and...
4211 (make_pass_tm_init): ...new function.
4212 (pass_tm_mark): Convert from a global struct to a subclass of
4213 gimple_opt_pass along with...
4214 (pass_data_tm_mark): ...new pass_data instance and...
4215 (make_pass_tm_mark): ...new function.
4216 (pass_tm_edges): Convert from a global struct to a subclass of
4217 gimple_opt_pass along with...
4218 (pass_data_tm_edges): ...new pass_data instance and...
4219 (make_pass_tm_edges): ...new function.
4220 (pass_tm_memopt): Convert from a global struct to a subclass of
4221 gimple_opt_pass along with...
4222 (pass_data_tm_memopt): ...new pass_data instance and...
4223 (make_pass_tm_memopt): ...new function.
4224 (pass_ipa_tm): Convert from a global struct to a subclass of
4225 simple_ipa_opt_pass along with...
4226 (pass_data_ipa_tm): ...new pass_data instance and...
4227 (make_pass_ipa_tm): ...new function.
4228 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4229 subclass of gimple_opt_pass along with...
4230 (pass_data_call_cdce): ...new pass_data instance and...
4231 (make_pass_call_cdce): ...new function.
4232 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4233 subclass of gimple_opt_pass along with...
4234 (pass_data_build_cfg): ...new pass_data instance and...
4235 (make_pass_build_cfg): ...new function.
4236 (pass_split_crit_edges): Convert from a global struct to a subclass of
4237 gimple_opt_pass along with...
4238 (pass_data_split_crit_edges): ...new pass_data instance and...
4239 (make_pass_split_crit_edges): ...new function.
4240 (pass_warn_function_return): Convert from a global struct to a
4241 subclass of gimple_opt_pass along with...
4242 (pass_data_warn_function_return): ...new pass_data instance and...
4243 (make_pass_warn_function_return): ...new function.
4244 (pass_warn_function_noreturn): Convert from a global struct to a
4245 subclass of gimple_opt_pass along with...
4246 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4247 (make_pass_warn_function_noreturn): ...new function.
4248 (pass_warn_unused_result): Convert from a global struct to a subclass
4249 of gimple_opt_pass along with...
4250 (pass_data_warn_unused_result): ...new pass_data instance and...
4251 (make_pass_warn_unused_result): ...new function.
4252 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4253 a subclass of gimple_opt_pass along with...
4254 (pass_data_merge_phi): ...new pass_data instance and...
4255 (make_pass_merge_phi): ...new function.
4256 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4257 a subclass of gimple_opt_pass along with...
4258 (pass_data_lower_complex): ...new pass_data instance and...
4259 (make_pass_lower_complex): ...new function.
4260 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4261 gimple_opt_pass along with...
4262 (pass_data_lower_complex_O0): ...new pass_data instance and...
4263 (make_pass_lower_complex_O0): ...new function.
4264 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4265 subclass of gimple_opt_pass along with...
4266 (pass_data_lower_eh): ...new pass_data instance and...
4267 (make_pass_lower_eh): ...new function.
4268 (pass_refactor_eh): Convert from a global struct to a subclass of
4269 gimple_opt_pass along with...
4270 (pass_data_refactor_eh): ...new pass_data instance and...
4271 (make_pass_refactor_eh): ...new function.
4272 (pass_lower_resx): Convert from a global struct to a subclass of
4273 gimple_opt_pass along with...
4274 (pass_data_lower_resx): ...new pass_data instance and...
4275 (make_pass_lower_resx): ...new function.
4276 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4277 of gimple_opt_pass along with...
4278 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4279 (make_pass_lower_eh_dispatch): ...new function.
4280 (pass_cleanup_eh): Convert from a global struct to a subclass of
4281 gimple_opt_pass along with...
4282 (pass_data_cleanup_eh): ...new pass_data instance and...
4283 (make_pass_cleanup_eh): ...new function.
4284 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4285 to a subclass of simple_ipa_opt_pass along with...
4286 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4287 (make_pass_ipa_lower_emutls): ...new function.
4288 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4289 a subclass of gimple_opt_pass along with...
4290 (pass_data_if_conversion): ...new pass_data instance and...
4291 (make_pass_if_conversion): ...new function.
4292 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4293 subclass of gimple_opt_pass along with...
4294 (pass_data_build_ssa): ...new pass_data instance and...
4295 (make_pass_build_ssa): ...new function.
4296 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4297 global struct to a subclass of gimple_opt_pass along with...
4298 (pass_data_loop_distribution): ...new pass_data instance and...
4299 (make_pass_loop_distribution): ...new function.
4300 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4301 subclass of gimple_opt_pass along with...
4302 (pass_data_mudflap_1): ...new pass_data instance and...
4303 (make_pass_mudflap_1): ...new function.
4304 (pass_mudflap_2): Convert from a global struct to a subclass of
4305 gimple_opt_pass along with...
4306 (pass_data_mudflap_2): ...new pass_data instance and...
4307 (make_pass_mudflap_2): ...new function.
4308 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4309 subclass of gimple_opt_pass along with...
4310 (pass_data_mudflap_1): ...new pass_data instance and...
4311 (make_pass_mudflap_1): ...new function.
4312 (pass_mudflap_2): Convert from a global struct to a subclass of
4313 gimple_opt_pass along with...
4314 (pass_data_mudflap_2): ...new pass_data instance and...
4315 (make_pass_mudflap_2): ...new function.
4316 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4317 gimple_opt_pass along with...
4318 (pass_data_nrv): ...new pass_data instance and...
4319 (make_pass_nrv): ...new function.
4320 (pass_return_slot): Convert from a global struct to a subclass of
4321 gimple_opt_pass along with...
4322 (pass_data_return_slot): ...new pass_data instance and...
4323 (make_pass_return_slot): ...new function.
4324 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4325 to a subclass of gimple_opt_pass along with...
4326 (pass_data_object_sizes): ...new pass_data instance and...
4327 (make_pass_object_sizes): ...new function.
4328 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4329 global struct to a subclass of gimple_opt_pass along with...
4330 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4331 and...
4332 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4333 (pass_fixup_cfg): Convert from a global struct to a subclass of
4334 gimple_opt_pass along with...
4335 (pass_data_fixup_cfg): ...new pass_data instance and...
4336 (make_pass_fixup_cfg): ...new function.
4337 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4338 (make_pass_mudflap_1): ...new function.
4339 (pass_mudflap_2): Replace declaration with that of...
4340 (make_pass_mudflap_2): ...new function.
4341 (pass_asan): Replace declaration with that of...
4342 (make_pass_asan): ...new function.
4343 (pass_asan_O0): Replace declaration with that of...
4344 (make_pass_asan_O0): ...new function.
4345 (pass_tsan): Replace declaration with that of...
4346 (make_pass_tsan): ...new function.
4347 (pass_tsan_O0): Replace declaration with that of...
4348 (make_pass_tsan_O0): ...new function.
4349 (pass_lower_cf): Replace declaration with that of...
4350 (make_pass_lower_cf): ...new function.
4351 (pass_refactor_eh): Replace declaration with that of...
4352 (make_pass_refactor_eh): ...new function.
4353 (pass_lower_eh): Replace declaration with that of...
4354 (make_pass_lower_eh): ...new function.
4355 (pass_lower_eh_dispatch): Replace declaration with that of...
4356 (make_pass_lower_eh_dispatch): ...new function.
4357 (pass_lower_resx): Replace declaration with that of...
4358 (make_pass_lower_resx): ...new function.
4359 (pass_build_cfg): Replace declaration with that of...
4360 (make_pass_build_cfg): ...new function.
4361 (pass_early_tree_profile): Replace declaration with that of...
4362 (make_pass_early_tree_profile): ...new function.
4363 (pass_cleanup_eh): Replace declaration with that of...
4364 (make_pass_cleanup_eh): ...new function.
4365 (pass_sra): Replace declaration with that of...
4366 (make_pass_sra): ...new function.
4367 (pass_sra_early): Replace declaration with that of...
4368 (make_pass_sra_early): ...new function.
4369 (pass_early_ipa_sra): Replace declaration with that of...
4370 (make_pass_early_ipa_sra): ...new function.
4371 (pass_tail_recursion): Replace declaration with that of...
4372 (make_pass_tail_recursion): ...new function.
4373 (pass_tail_calls): Replace declaration with that of...
4374 (make_pass_tail_calls): ...new function.
4375 (pass_tree_loop): Replace declaration with that of...
4376 (make_pass_tree_loop): ...new function.
4377 (pass_tree_loop_init): Replace declaration with that of...
4378 (make_pass_tree_loop_init): ...new function.
4379 (pass_lim): Replace declaration with that of...
4380 (make_pass_lim): ...new function.
4381 (pass_tree_unswitch): Replace declaration with that of...
4382 (make_pass_tree_unswitch): ...new function.
4383 (pass_predcom): Replace declaration with that of...
4384 (make_pass_predcom): ...new function.
4385 (pass_iv_canon): Replace declaration with that of...
4386 (make_pass_iv_canon): ...new function.
4387 (pass_scev_cprop): Replace declaration with that of...
4388 (make_pass_scev_cprop): ...new function.
4389 (pass_empty_loop): Replace declaration with that of...
4390 (make_pass_empty_loop): ...new function.
4391 (pass_record_bounds): Replace declaration with that of...
4392 (make_pass_record_bounds): ...new function.
4393 (pass_graphite): Replace declaration with that of...
4394 (make_pass_graphite): ...new function.
4395 (pass_graphite_transforms): Replace declaration with that of...
4396 (make_pass_graphite_transforms): ...new function.
4397 (pass_if_conversion): Replace declaration with that of...
4398 (make_pass_if_conversion): ...new function.
4399 (pass_loop_distribution): Replace declaration with that of...
4400 (make_pass_loop_distribution): ...new function.
4401 (pass_vectorize): Replace declaration with that of...
4402 (make_pass_vectorize): ...new function.
4403 (pass_slp_vectorize): Replace declaration with that of...
4404 (make_pass_slp_vectorize): ...new function.
4405 (pass_complete_unroll): Replace declaration with that of...
4406 (make_pass_complete_unroll): ...new function.
4407 (pass_complete_unrolli): Replace declaration with that of...
4408 (make_pass_complete_unrolli): ...new function.
4409 (pass_parallelize_loops): Replace declaration with that of...
4410 (make_pass_parallelize_loops): ...new function.
4411 (pass_loop_prefetch): Replace declaration with that of...
4412 (make_pass_loop_prefetch): ...new function.
4413 (pass_iv_optimize): Replace declaration with that of...
4414 (make_pass_iv_optimize): ...new function.
4415 (pass_tree_loop_done): Replace declaration with that of...
4416 (make_pass_tree_loop_done): ...new function.
4417 (pass_ch): Replace declaration with that of...
4418 (make_pass_ch): ...new function.
4419 (pass_ccp): Replace declaration with that of...
4420 (make_pass_ccp): ...new function.
4421 (pass_phi_only_cprop): Replace declaration with that of...
4422 (make_pass_phi_only_cprop): ...new function.
4423 (pass_build_ssa): Replace declaration with that of...
4424 (make_pass_build_ssa): ...new function.
4425 (pass_build_alias): Replace declaration with that of...
4426 (make_pass_build_alias): ...new function.
4427 (pass_build_ealias): Replace declaration with that of...
4428 (make_pass_build_ealias): ...new function.
4429 (pass_dominator): Replace declaration with that of...
4430 (make_pass_dominator): ...new function.
4431 (pass_dce): Replace declaration with that of...
4432 (make_pass_dce): ...new function.
4433 (pass_dce_loop): Replace declaration with that of...
4434 (make_pass_dce_loop): ...new function.
4435 (pass_cd_dce): Replace declaration with that of...
4436 (make_pass_cd_dce): ...new function.
4437 (pass_call_cdce): Replace declaration with that of...
4438 (make_pass_call_cdce): ...new function.
4439 (pass_merge_phi): Replace declaration with that of...
4440 (make_pass_merge_phi): ...new function.
4441 (pass_split_crit_edges): Replace declaration with that of...
4442 (make_pass_split_crit_edges): ...new function.
4443 (pass_pre): Replace declaration with that of...
4444 (make_pass_pre): ...new function.
4445 (pass_profile): Replace declaration with that of...
4446 (make_pass_profile): ...new function.
4447 (pass_strip_predict_hints): Replace declaration with that of...
4448 (make_pass_strip_predict_hints): ...new function.
4449 (pass_lower_complex_O0): Replace declaration with that of...
4450 (make_pass_lower_complex_O0): ...new function.
4451 (pass_lower_complex): Replace declaration with that of...
4452 (make_pass_lower_complex): ...new function.
4453 (pass_lower_vector): Replace declaration with that of...
4454 (make_pass_lower_vector): ...new function.
4455 (pass_lower_vector_ssa): Replace declaration with that of...
4456 (make_pass_lower_vector_ssa): ...new function.
4457 (pass_lower_omp): Replace declaration with that of...
4458 (make_pass_lower_omp): ...new function.
4459 (pass_diagnose_omp_blocks): Replace declaration with that of...
4460 (make_pass_diagnose_omp_blocks): ...new function.
4461 (pass_expand_omp): Replace declaration with that of...
4462 (make_pass_expand_omp): ...new function.
4463 (pass_expand_omp_ssa): Replace declaration with that of...
4464 (make_pass_expand_omp_ssa): ...new function.
4465 (pass_object_sizes): Replace declaration with that of...
4466 (make_pass_object_sizes): ...new function.
4467 (pass_strlen): Replace declaration with that of...
4468 (make_pass_strlen): ...new function.
4469 (pass_fold_builtins): Replace declaration with that of...
4470 (make_pass_fold_builtins): ...new function.
4471 (pass_stdarg): Replace declaration with that of...
4472 (make_pass_stdarg): ...new function.
4473 (pass_early_warn_uninitialized): Replace declaration with that of...
4474 (make_pass_early_warn_uninitialized): ...new function.
4475 (pass_late_warn_uninitialized): Replace declaration with that of...
4476 (make_pass_late_warn_uninitialized): ...new function.
4477 (pass_cse_reciprocals): Replace declaration with that of...
4478 (make_pass_cse_reciprocals): ...new function.
4479 (pass_cse_sincos): Replace declaration with that of...
4480 (make_pass_cse_sincos): ...new function.
4481 (pass_optimize_bswap): Replace declaration with that of...
4482 (make_pass_optimize_bswap): ...new function.
4483 (pass_optimize_widening_mul): Replace declaration with that of...
4484 (make_pass_optimize_widening_mul): ...new function.
4485 (pass_warn_function_return): Replace declaration with that of...
4486 (make_pass_warn_function_return): ...new function.
4487 (pass_warn_function_noreturn): Replace declaration with that of...
4488 (make_pass_warn_function_noreturn): ...new function.
4489 (pass_cselim): Replace declaration with that of...
4490 (make_pass_cselim): ...new function.
4491 (pass_phiopt): Replace declaration with that of...
4492 (make_pass_phiopt): ...new function.
4493 (pass_forwprop): Replace declaration with that of...
4494 (make_pass_forwprop): ...new function.
4495 (pass_phiprop): Replace declaration with that of...
4496 (make_pass_phiprop): ...new function.
4497 (pass_tree_ifcombine): Replace declaration with that of...
4498 (make_pass_tree_ifcombine): ...new function.
4499 (pass_dse): Replace declaration with that of...
4500 (make_pass_dse): ...new function.
4501 (pass_nrv): Replace declaration with that of...
4502 (make_pass_nrv): ...new function.
4503 (pass_rename_ssa_copies): Replace declaration with that of...
4504 (make_pass_rename_ssa_copies): ...new function.
4505 (pass_sink_code): Replace declaration with that of...
4506 (make_pass_sink_code): ...new function.
4507 (pass_fre): Replace declaration with that of...
4508 (make_pass_fre): ...new function.
4509 (pass_check_data_deps): Replace declaration with that of...
4510 (make_pass_check_data_deps): ...new function.
4511 (pass_copy_prop): Replace declaration with that of...
4512 (make_pass_copy_prop): ...new function.
4513 (pass_vrp): Replace declaration with that of...
4514 (make_pass_vrp): ...new function.
4515 (pass_uncprop): Replace declaration with that of...
4516 (make_pass_uncprop): ...new function.
4517 (pass_return_slot): Replace declaration with that of...
4518 (make_pass_return_slot): ...new function.
4519 (pass_reassoc): Replace declaration with that of...
4520 (make_pass_reassoc): ...new function.
4521 (pass_rebuild_cgraph_edges): Replace declaration with that of...
4522 (make_pass_rebuild_cgraph_edges): ...new function.
4523 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
4524 (make_pass_remove_cgraph_callee_edges): ...new function.
4525 (pass_build_cgraph_edges): Replace declaration with that of...
4526 (make_pass_build_cgraph_edges): ...new function.
4527 (pass_local_pure_const): Replace declaration with that of...
4528 (make_pass_local_pure_const): ...new function.
4529 (pass_tracer): Replace declaration with that of...
4530 (make_pass_tracer): ...new function.
4531 (pass_warn_unused_result): Replace declaration with that of...
4532 (make_pass_warn_unused_result): ...new function.
4533 (pass_diagnose_tm_blocks): Replace declaration with that of...
4534 (make_pass_diagnose_tm_blocks): ...new function.
4535 (pass_lower_tm): Replace declaration with that of...
4536 (make_pass_lower_tm): ...new function.
4537 (pass_tm_init): Replace declaration with that of...
4538 (make_pass_tm_init): ...new function.
4539 (pass_tm_mark): Replace declaration with that of...
4540 (make_pass_tm_mark): ...new function.
4541 (pass_tm_memopt): Replace declaration with that of...
4542 (make_pass_tm_memopt): ...new function.
4543 (pass_tm_edges): Replace declaration with that of...
4544 (make_pass_tm_edges): ...new function.
4545 (pass_split_functions): Replace declaration with that of...
4546 (make_pass_split_functions): ...new function.
4547 (pass_feedback_split_functions): Replace declaration with that of...
4548 (make_pass_feedback_split_functions): ...new function.
4549 (pass_strength_reduction): Replace declaration with that of...
4550 (make_pass_strength_reduction): ...new function.
4551 (pass_ipa_lower_emutls): Replace declaration with that of...
4552 (make_pass_ipa_lower_emutls): ...new function.
4553 (pass_ipa_function_and_variable_visibility): Replace declaration with
4554 that of...
4555 (make_pass_ipa_function_and_variable_visibility): ...new function.
4556 (pass_ipa_tree_profile): Replace declaration with that of...
4557 (make_pass_ipa_tree_profile): ...new function.
4558 (pass_early_local_passes): Replace declaration with that of...
4559 (make_pass_early_local_passes): ...new function.
4560 (pass_ipa_whole_program_visibility): Replace declaration with that
4561 of...
4562 (make_pass_ipa_whole_program_visibility): ...new function.
4563 (pass_ipa_lto_gimple_out): Replace declaration with that of...
4564 (make_pass_ipa_lto_gimple_out): ...new function.
4565 (pass_ipa_increase_alignment): Replace declaration with that of...
4566 (make_pass_ipa_increase_alignment): ...new function.
4567 (pass_ipa_inline): Replace declaration with that of...
4568 (make_pass_ipa_inline): ...new function.
4569 (pass_ipa_free_lang_data): Replace declaration with that of...
4570 (make_pass_ipa_free_lang_data): ...new function.
4571 (pass_ipa_free_inline_summary): Replace declaration with that of...
4572 (make_pass_ipa_free_inline_summary): ...new function.
4573 (pass_ipa_cp): Replace declaration with that of...
4574 (make_pass_ipa_cp): ...new function.
4575 (pass_ipa_reference): Replace declaration with that of...
4576 (make_pass_ipa_reference): ...new function.
4577 (pass_ipa_pure_const): Replace declaration with that of...
4578 (make_pass_ipa_pure_const): ...new function.
4579 (pass_ipa_pta): Replace declaration with that of...
4580 (make_pass_ipa_pta): ...new function.
4581 (pass_ipa_lto_finish_out): Replace declaration with that of...
4582 (make_pass_ipa_lto_finish_out): ...new function.
4583 (pass_ipa_tm): Replace declaration with that of...
4584 (make_pass_ipa_tm): ...new function.
4585 (pass_ipa_profile): Replace declaration with that of...
4586 (make_pass_ipa_profile): ...new function.
4587 (pass_ipa_cdtor_merge): Replace declaration with that of...
4588 (make_pass_ipa_cdtor_merge): ...new function.
4589 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
4590 of...
4591 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4592 (pass_init_datastructures): Replace declaration with that of...
4593 (make_pass_init_datastructures): ...new function.
4594 (pass_fixup_cfg): Replace declaration with that of...
4595 (make_pass_fixup_cfg): ...new function.
4596 (pass_expand): Replace declaration with that of...
4597 (make_pass_expand): ...new function.
4598 (pass_instantiate_virtual_regs): Replace declaration with that of...
4599 (make_pass_instantiate_virtual_regs): ...new function.
4600 (pass_rtl_fwprop): Replace declaration with that of...
4601 (make_pass_rtl_fwprop): ...new function.
4602 (pass_rtl_fwprop_addr): Replace declaration with that of...
4603 (make_pass_rtl_fwprop_addr): ...new function.
4604 (pass_jump): Replace declaration with that of...
4605 (make_pass_jump): ...new function.
4606 (pass_jump2): Replace declaration with that of...
4607 (make_pass_jump2): ...new function.
4608 (pass_lower_subreg): Replace declaration with that of...
4609 (make_pass_lower_subreg): ...new function.
4610 (pass_cse): Replace declaration with that of...
4611 (make_pass_cse): ...new function.
4612 (pass_fast_rtl_dce): Replace declaration with that of...
4613 (make_pass_fast_rtl_dce): ...new function.
4614 (pass_ud_rtl_dce): Replace declaration with that of...
4615 (make_pass_ud_rtl_dce): ...new function.
4616 (pass_rtl_dce): Replace declaration with that of...
4617 (make_pass_rtl_dce): ...new function.
4618 (pass_rtl_dse1): Replace declaration with that of...
4619 (make_pass_rtl_dse1): ...new function.
4620 (pass_rtl_dse2): Replace declaration with that of...
4621 (make_pass_rtl_dse2): ...new function.
4622 (pass_rtl_dse3): Replace declaration with that of...
4623 (make_pass_rtl_dse3): ...new function.
4624 (pass_rtl_cprop): Replace declaration with that of...
4625 (make_pass_rtl_cprop): ...new function.
4626 (pass_rtl_pre): Replace declaration with that of...
4627 (make_pass_rtl_pre): ...new function.
4628 (pass_rtl_hoist): Replace declaration with that of...
4629 (make_pass_rtl_hoist): ...new function.
4630 (pass_rtl_store_motion): Replace declaration with that of...
4631 (make_pass_rtl_store_motion): ...new function.
4632 (pass_cse_after_global_opts): Replace declaration with that of...
4633 (make_pass_cse_after_global_opts): ...new function.
4634 (pass_rtl_ifcvt): Replace declaration with that of...
4635 (make_pass_rtl_ifcvt): ...new function.
4636 (pass_into_cfg_layout_mode): Replace declaration with that of...
4637 (make_pass_into_cfg_layout_mode): ...new function.
4638 (pass_outof_cfg_layout_mode): Replace declaration with that of...
4639 (make_pass_outof_cfg_layout_mode): ...new function.
4640 (pass_loop2): Replace declaration with that of...
4641 (make_pass_loop2): ...new function.
4642 (pass_rtl_loop_init): Replace declaration with that of...
4643 (make_pass_rtl_loop_init): ...new function.
4644 (pass_rtl_move_loop_invariants): Replace declaration with that of...
4645 (make_pass_rtl_move_loop_invariants): ...new function.
4646 (pass_rtl_unswitch): Replace declaration with that of...
4647 (make_pass_rtl_unswitch): ...new function.
4648 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
4649 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4650 (pass_rtl_doloop): Replace declaration with that of...
4651 (make_pass_rtl_doloop): ...new function.
4652 (pass_rtl_loop_done): Replace declaration with that of...
4653 (make_pass_rtl_loop_done): ...new function.
4654 (pass_web): Replace declaration with that of...
4655 (make_pass_web): ...new function.
4656 (pass_cse2): Replace declaration with that of...
4657 (make_pass_cse2): ...new function.
4658 (pass_df_initialize_opt): Replace declaration with that of...
4659 (make_pass_df_initialize_opt): ...new function.
4660 (pass_df_initialize_no_opt): Replace declaration with that of...
4661 (make_pass_df_initialize_no_opt): ...new function.
4662 (pass_reginfo_init): Replace declaration with that of...
4663 (make_pass_reginfo_init): ...new function.
4664 (pass_inc_dec): Replace declaration with that of...
4665 (make_pass_inc_dec): ...new function.
4666 (pass_stack_ptr_mod): Replace declaration with that of...
4667 (make_pass_stack_ptr_mod): ...new function.
4668 (pass_initialize_regs): Replace declaration with that of...
4669 (make_pass_initialize_regs): ...new function.
4670 (pass_combine): Replace declaration with that of...
4671 (make_pass_combine): ...new function.
4672 (pass_if_after_combine): Replace declaration with that of...
4673 (make_pass_if_after_combine): ...new function.
4674 (pass_ree): Replace declaration with that of...
4675 (make_pass_ree): ...new function.
4676 (pass_partition_blocks): Replace declaration with that of...
4677 (make_pass_partition_blocks): ...new function.
4678 (pass_match_asm_constraints): Replace declaration with that of...
4679 (make_pass_match_asm_constraints): ...new function.
4680 (pass_regmove): Replace declaration with that of...
4681 (make_pass_regmove): ...new function.
4682 (pass_split_all_insns): Replace declaration with that of...
4683 (make_pass_split_all_insns): ...new function.
4684 (pass_fast_rtl_byte_dce): Replace declaration with that of...
4685 (make_pass_fast_rtl_byte_dce): ...new function.
4686 (pass_lower_subreg2): Replace declaration with that of...
4687 (make_pass_lower_subreg2): ...new function.
4688 (pass_mode_switching): Replace declaration with that of...
4689 (make_pass_mode_switching): ...new function.
4690 (pass_sms): Replace declaration with that of...
4691 (make_pass_sms): ...new function.
4692 (pass_sched): Replace declaration with that of...
4693 (make_pass_sched): ...new function.
4694 (pass_ira): Replace declaration with that of...
4695 (make_pass_ira): ...new function.
4696 (pass_reload): Replace declaration with that of...
4697 (make_pass_reload): ...new function.
4698 (pass_clean_state): Replace declaration with that of...
4699 (make_pass_clean_state): ...new function.
4700 (pass_branch_prob): Replace declaration with that of...
4701 (make_pass_branch_prob): ...new function.
4702 (pass_value_profile_transformations): Replace declaration with that
4703 of...
4704 (make_pass_value_profile_transformations): ...new function.
4705 (pass_postreload_cse): Replace declaration with that of...
4706 (make_pass_postreload_cse): ...new function.
4707 (pass_gcse2): Replace declaration with that of...
4708 (make_pass_gcse2): ...new function.
4709 (pass_split_after_reload): Replace declaration with that of...
4710 (make_pass_split_after_reload): ...new function.
4711 (pass_branch_target_load_optimize1): Replace declaration with that
4712 of...
4713 (make_pass_branch_target_load_optimize1): ...new function.
4714 (pass_thread_prologue_and_epilogue): Replace declaration with that
4715 of...
4716 (make_pass_thread_prologue_and_epilogue): ...new function.
4717 (pass_stack_adjustments): Replace declaration with that of...
4718 (make_pass_stack_adjustments): ...new function.
4719 (pass_peephole2): Replace declaration with that of...
4720 (make_pass_peephole2): ...new function.
4721 (pass_if_after_reload): Replace declaration with that of...
4722 (make_pass_if_after_reload): ...new function.
4723 (pass_regrename): Replace declaration with that of...
4724 (make_pass_regrename): ...new function.
4725 (pass_cprop_hardreg): Replace declaration with that of...
4726 (make_pass_cprop_hardreg): ...new function.
4727 (pass_reorder_blocks): Replace declaration with that of...
4728 (make_pass_reorder_blocks): ...new function.
4729 (pass_branch_target_load_optimize2): Replace declaration with that
4730 of...
4731 (make_pass_branch_target_load_optimize2): ...new function.
4732 (pass_leaf_regs): Replace declaration with that of...
4733 (make_pass_leaf_regs): ...new function.
4734 (pass_split_before_sched2): Replace declaration with that of...
4735 (make_pass_split_before_sched2): ...new function.
4736 (pass_compare_elim_after_reload): Replace declaration with that of...
4737 (make_pass_compare_elim_after_reload): ...new function.
4738 (pass_sched2): Replace declaration with that of...
4739 (make_pass_sched2): ...new function.
4740 (pass_stack_regs): Replace declaration with that of...
4741 (make_pass_stack_regs): ...new function.
4742 (pass_stack_regs_run): Replace declaration with that of...
4743 (make_pass_stack_regs_run): ...new function.
4744 (pass_df_finish): Replace declaration with that of...
4745 (make_pass_df_finish): ...new function.
4746 (pass_compute_alignments): Replace declaration with that of...
4747 (make_pass_compute_alignments): ...new function.
4748 (pass_duplicate_computed_gotos): Replace declaration with that of...
4749 (make_pass_duplicate_computed_gotos): ...new function.
4750 (pass_variable_tracking): Replace declaration with that of...
4751 (make_pass_variable_tracking): ...new function.
4752 (pass_free_cfg): Replace declaration with that of...
4753 (make_pass_free_cfg): ...new function.
4754 (pass_machine_reorg): Replace declaration with that of...
4755 (make_pass_machine_reorg): ...new function.
4756 (pass_cleanup_barriers): Replace declaration with that of...
4757 (make_pass_cleanup_barriers): ...new function.
4758 (pass_delay_slots): Replace declaration with that of...
4759 (make_pass_delay_slots): ...new function.
4760 (pass_split_for_shorten_branches): Replace declaration with that of...
4761 (make_pass_split_for_shorten_branches): ...new function.
4762 (pass_split_before_regstack): Replace declaration with that of...
4763 (make_pass_split_before_regstack): ...new function.
4764 (pass_convert_to_eh_region_ranges): Replace declaration with that
4765 of...
4766 (make_pass_convert_to_eh_region_ranges): ...new function.
4767 (pass_shorten_branches): Replace declaration with that of...
4768 (make_pass_shorten_branches): ...new function.
4769 (pass_set_nothrow_function_flags): Replace declaration with that of...
4770 (make_pass_set_nothrow_function_flags): ...new function.
4771 (pass_dwarf2_frame): Replace declaration with that of...
4772 (make_pass_dwarf2_frame): ...new function.
4773 (pass_final): Replace declaration with that of...
4774 (make_pass_final): ...new function.
4775 (pass_rtl_seqabstr): Replace declaration with that of...
4776 (make_pass_rtl_seqabstr): ...new function.
4777 (pass_release_ssa_names): Replace declaration with that of...
4778 (make_pass_release_ssa_names): ...new function.
4779 (pass_early_inline): Replace declaration with that of...
4780 (make_pass_early_inline): ...new function.
4781 (pass_inline_parameters): Replace declaration with that of...
4782 (make_pass_inline_parameters): ...new function.
4783 (pass_update_address_taken): Replace declaration with that of...
4784 (make_pass_update_address_taken): ...new function.
4785 (pass_convert_switch): Replace declaration with that of...
4786 (make_pass_convert_switch): ...new function.
4787 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
4788 to a subclass of simple_ipa_opt_pass along with...
4789 (pass_data_ipa_tree_profile): ...new pass_data instance and...
4790 (make_pass_ipa_tree_profile): ...new function.
4791 * tree-sra.c (pass_sra_early): Convert from a global struct to a
4792 subclass of gimple_opt_pass along with...
4793 (pass_data_sra_early): ...new pass_data instance and...
4794 (make_pass_sra_early): ...new function.
4795 (pass_sra): Convert from a global struct to a subclass of
4796 gimple_opt_pass along with...
4797 (pass_data_sra): ...new pass_data instance and...
4798 (make_pass_sra): ...new function.
4799 (pass_early_ipa_sra): Convert from a global struct to a subclass of
4800 gimple_opt_pass along with...
4801 (pass_data_early_ipa_sra): ...new pass_data instance and...
4802 (make_pass_early_ipa_sra): ...new function.
4803 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
4804 subclass of gimple_opt_pass along with...
4805 (pass_data_ccp): ...new pass_data instance and...
4806 (make_pass_ccp): ...new function.
4807 (pass_fold_builtins): Convert from a global struct to a subclass of
4808 gimple_opt_pass along with...
4809 (pass_data_fold_builtins): ...new pass_data instance and...
4810 (make_pass_fold_builtins): ...new function.
4811 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
4812 subclass of gimple_opt_pass along with...
4813 (pass_data_copy_prop): ...new pass_data instance and...
4814 (make_pass_copy_prop): ...new function.
4815 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
4816 global struct to a subclass of gimple_opt_pass along with...
4817 (pass_data_rename_ssa_copies): ...new pass_data instance and...
4818 (make_pass_rename_ssa_copies): ...new function.
4819 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
4820 subclass of gimple_opt_pass along with...
4821 (pass_data_dce): ...new pass_data instance and...
4822 (make_pass_dce): ...new function.
4823 (pass_dce_loop): Convert from a global struct to a subclass of
4824 gimple_opt_pass along with...
4825 (pass_data_dce_loop): ...new pass_data instance and...
4826 (make_pass_dce_loop): ...new function.
4827 (pass_cd_dce): Convert from a global struct to a subclass of
4828 gimple_opt_pass along with...
4829 (pass_data_cd_dce): ...new pass_data instance and...
4830 (make_pass_cd_dce): ...new function.
4831 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
4832 subclass of gimple_opt_pass along with...
4833 (pass_data_dominator): ...new pass_data instance and...
4834 (make_pass_dominator): ...new function.
4835 (pass_phi_only_cprop): Convert from a global struct to a subclass of
4836 gimple_opt_pass along with...
4837 (pass_data_phi_only_cprop): ...new pass_data instance and...
4838 (make_pass_phi_only_cprop): ...new function.
4839 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
4840 subclass of gimple_opt_pass along with...
4841 (pass_data_dse): ...new pass_data instance and...
4842 (make_pass_dse): ...new function.
4843 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
4844 a subclass of gimple_opt_pass along with...
4845 (pass_data_forwprop): ...new pass_data instance and...
4846 (make_pass_forwprop): ...new function.
4847 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
4848 struct to a subclass of gimple_opt_pass along with...
4849 (pass_data_tree_ifcombine): ...new pass_data instance and...
4850 (make_pass_tree_ifcombine): ...new function.
4851 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
4852 subclass of gimple_opt_pass along with...
4853 (pass_data_ch): ...new pass_data instance and...
4854 (make_pass_ch): ...new function.
4855 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
4856 subclass of gimple_opt_pass along with...
4857 (pass_data_tree_loop): ...new pass_data instance and...
4858 (make_pass_tree_loop): ...new function.
4859 (pass_tree_loop_init): Convert from a global struct to a subclass of
4860 gimple_opt_pass along with...
4861 (pass_data_tree_loop_init): ...new pass_data instance and...
4862 (make_pass_tree_loop_init): ...new function.
4863 (pass_lim): Convert from a global struct to a subclass of
4864 gimple_opt_pass along with...
4865 (pass_data_lim): ...new pass_data instance and...
4866 (make_pass_lim): ...new function.
4867 (pass_tree_unswitch): Convert from a global struct to a subclass of
4868 gimple_opt_pass along with...
4869 (pass_data_tree_unswitch): ...new pass_data instance and...
4870 (make_pass_tree_unswitch): ...new function.
4871 (pass_predcom): Convert from a global struct to a subclass of
4872 gimple_opt_pass along with...
4873 (pass_data_predcom): ...new pass_data instance and...
4874 (make_pass_predcom): ...new function.
4875 (pass_vectorize): Convert from a global struct to a subclass of
4876 gimple_opt_pass along with...
4877 (pass_data_vectorize): ...new pass_data instance and...
4878 (make_pass_vectorize): ...new function.
4879 (pass_graphite): Convert from a global struct to a subclass of
4880 gimple_opt_pass along with...
4881 (pass_data_graphite): ...new pass_data instance and...
4882 (make_pass_graphite): ...new function.
4883 (pass_graphite_transforms): Convert from a global struct to a subclass
4884 of gimple_opt_pass along with...
4885 (pass_data_graphite_transforms): ...new pass_data instance and...
4886 (make_pass_graphite_transforms): ...new function.
4887 (pass_check_data_deps): Convert from a global struct to a subclass of
4888 gimple_opt_pass along with...
4889 (pass_data_check_data_deps): ...new pass_data instance and...
4890 (make_pass_check_data_deps): ...new function.
4891 (pass_iv_canon): Convert from a global struct to a subclass of
4892 gimple_opt_pass along with...
4893 (pass_data_iv_canon): ...new pass_data instance and...
4894 (make_pass_iv_canon): ...new function.
4895 (pass_scev_cprop): Convert from a global struct to a subclass of
4896 gimple_opt_pass along with...
4897 (pass_data_scev_cprop): ...new pass_data instance and...
4898 (make_pass_scev_cprop): ...new function.
4899 (pass_record_bounds): Convert from a global struct to a subclass of
4900 gimple_opt_pass along with...
4901 (pass_data_record_bounds): ...new pass_data instance and...
4902 (make_pass_record_bounds): ...new function.
4903 (pass_complete_unroll): Convert from a global struct to a subclass of
4904 gimple_opt_pass along with...
4905 (pass_data_complete_unroll): ...new pass_data instance and...
4906 (make_pass_complete_unroll): ...new function.
4907 (pass_complete_unrolli): Convert from a global struct to a subclass of
4908 gimple_opt_pass along with...
4909 (pass_data_complete_unrolli): ...new pass_data instance and...
4910 (make_pass_complete_unrolli): ...new function.
4911 (pass_parallelize_loops): Convert from a global struct to a subclass
4912 of gimple_opt_pass along with...
4913 (pass_data_parallelize_loops): ...new pass_data instance and...
4914 (make_pass_parallelize_loops): ...new function.
4915 (pass_loop_prefetch): Convert from a global struct to a subclass of
4916 gimple_opt_pass along with...
4917 (pass_data_loop_prefetch): ...new pass_data instance and...
4918 (make_pass_loop_prefetch): ...new function.
4919 (pass_iv_optimize): Convert from a global struct to a subclass of
4920 gimple_opt_pass along with...
4921 (pass_data_iv_optimize): ...new pass_data instance and...
4922 (make_pass_iv_optimize): ...new function.
4923 (pass_tree_loop_done): Convert from a global struct to a subclass of
4924 gimple_opt_pass along with...
4925 (pass_data_tree_loop_done): ...new pass_data instance and...
4926 (make_pass_tree_loop_done): ...new function.
4927 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
4928 struct to a subclass of gimple_opt_pass along with...
4929 (pass_data_cse_reciprocals): ...new pass_data instance and...
4930 (make_pass_cse_reciprocals): ...new function.
4931 (pass_cse_sincos): Convert from a global struct to a subclass of
4932 gimple_opt_pass along with...
4933 (pass_data_cse_sincos): ...new pass_data instance and...
4934 (make_pass_cse_sincos): ...new function.
4935 (pass_optimize_bswap): Convert from a global struct to a subclass of
4936 gimple_opt_pass along with...
4937 (pass_data_optimize_bswap): ...new pass_data instance and...
4938 (make_pass_optimize_bswap): ...new function.
4939 (pass_optimize_widening_mul): Convert from a global struct to a
4940 subclass of gimple_opt_pass along with...
4941 (pass_data_optimize_widening_mul): ...new pass_data instance and...
4942 (make_pass_optimize_widening_mul): ...new function.
4943 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
4944 subclass of gimple_opt_pass along with...
4945 (pass_data_phiopt): ...new pass_data instance and...
4946 (make_pass_phiopt): ...new function.
4947 (pass_cselim): Convert from a global struct to a subclass of
4948 gimple_opt_pass along with...
4949 (pass_data_cselim): ...new pass_data instance and...
4950 (make_pass_cselim): ...new function.
4951 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
4952 subclass of gimple_opt_pass along with...
4953 (pass_data_phiprop): ...new pass_data instance and...
4954 (make_pass_phiprop): ...new function.
4955 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
4956 subclass of gimple_opt_pass along with...
4957 (pass_data_pre): ...new pass_data instance and...
4958 (make_pass_pre): ...new function.
4959 (pass_fre): Convert from a global struct to a subclass of
4960 gimple_opt_pass along with...
4961 (pass_data_fre): ...new pass_data instance and...
4962 (make_pass_fre): ...new function.
4963 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
4964 subclass of gimple_opt_pass along with...
4965 (pass_data_reassoc): ...new pass_data instance and...
4966 (make_pass_reassoc): ...new function.
4967 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
4968 subclass of gimple_opt_pass along with...
4969 (pass_data_sink_code): ...new pass_data instance and...
4970 (make_pass_sink_code): ...new function.
4971 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
4972 subclass of gimple_opt_pass along with...
4973 (pass_data_strlen): ...new pass_data instance and...
4974 (make_pass_strlen): ...new function.
4975 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
4976 struct to a subclass of gimple_opt_pass along with...
4977 (pass_data_build_alias): ...new pass_data instance and...
4978 (make_pass_build_alias): ...new function.
4979 (pass_build_ealias): Convert from a global struct to a subclass of
4980 gimple_opt_pass along with...
4981 (pass_data_build_ealias): ...new pass_data instance and...
4982 (make_pass_build_ealias): ...new function.
4983 (pass_ipa_pta): Convert from a global struct to a subclass of
4984 simple_ipa_opt_pass along with...
4985 (pass_data_ipa_pta): ...new pass_data instance and...
4986 (make_pass_ipa_pta): ...new function.
4987 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
4988 subclass of gimple_opt_pass along with...
4989 (pass_data_uncprop): ...new pass_data instance and...
4990 (make_pass_uncprop): ...new function.
4991 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
4992 global struct to a subclass of gimple_opt_pass along with...
4993 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
4994 (make_pass_late_warn_uninitialized): ...new function.
4995 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
4996 to a subclass of gimple_opt_pass along with...
4997 (pass_data_init_datastructures): ...new pass_data instance and...
4998 (make_pass_init_datastructures): ...new function.
4999 (pass_early_warn_uninitialized): Convert from a global struct to a
5000 subclass of gimple_opt_pass along with...
5001 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5002 (make_pass_early_warn_uninitialized): ...new function.
5003 (pass_update_address_taken): Convert from a global struct to a
5004 subclass of gimple_opt_pass along with...
5005 (pass_data_update_address_taken): ...new pass_data instance and...
5006 (make_pass_update_address_taken): ...new function.
5007 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5008 struct to a subclass of gimple_opt_pass along with...
5009 (pass_data_release_ssa_names): ...new pass_data instance and...
5010 (make_pass_release_ssa_names): ...new function.
5011 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5012 subclass of gimple_opt_pass along with...
5013 (pass_data_stdarg): ...new pass_data instance and...
5014 (make_pass_stdarg): ...new function.
5015 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5016 global struct to a subclass of gimple_opt_pass along with...
5017 (pass_data_convert_switch): ...new pass_data instance and...
5018 (make_pass_convert_switch): ...new function.
5019 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5020 to a subclass of gimple_opt_pass along with...
5021 (pass_data_tail_recursion): ...new pass_data instance and...
5022 (make_pass_tail_recursion): ...new function.
5023 (pass_tail_calls): Convert from a global struct to a subclass of
5024 gimple_opt_pass along with...
5025 (pass_data_tail_calls): ...new pass_data instance and...
5026 (make_pass_tail_calls): ...new function.
5027 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5028 struct to a subclass of gimple_opt_pass along with...
5029 (pass_data_lower_vector): ...new pass_data instance and...
5030 (make_pass_lower_vector): ...new function.
5031 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5032 gimple_opt_pass along with...
5033 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5034 (make_pass_lower_vector_ssa): ...new function.
5035 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5036 to a subclass of gimple_opt_pass along with...
5037 (pass_data_slp_vectorize): ...new pass_data instance and...
5038 (make_pass_slp_vectorize): ...new function.
5039 (pass_ipa_increase_alignment): Convert from a global struct to a
5040 subclass of simple_ipa_opt_pass along with...
5041 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5042 (make_pass_ipa_increase_alignment): ...new function.
5043 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5044 gimple_opt_pass along with...
5045 (pass_data_vrp): ...new pass_data instance and...
5046 (make_pass_vrp): ...new function.
5047 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5048 subclass of simple_ipa_opt_pass along with...
5049 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5050 (make_pass_ipa_free_lang_data): ...new function.
5051 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5052 gimple_opt_pass along with...
5053 (pass_data_tsan): ...new pass_data instance and...
5054 (make_pass_tsan): ...new function.
5055 (pass_tsan_O0): Convert from a global struct to a subclass of
5056 gimple_opt_pass along with...
5057 (pass_data_tsan_O0): ...new pass_data instance and...
5058 (make_pass_tsan_O0): ...new function.
5059 * var-tracking.c (pass_variable_tracking): Convert from a global
5060 struct to a subclass of rtl_opt_pass along with...
5061 (pass_data_variable_tracking): ...new pass_data instance and...
5062 (make_pass_variable_tracking): ...new function.
5063 * web.c (pass_web): Convert from a global struct to a subclass of
5064 rtl_opt_pass along with...
5065 (pass_data_web): ...new pass_data instance and...
5066 (make_pass_web): ...new function.
5067 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5068 declaration with that of...
5069 (make_pass_mode_switch_use): ...new function.
5070 (pass_resolve_sw_modes): Replace declaration with that of...
5071 (make_pass_resolve_sw_modes): ...new function.
5072 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5073 from a global struct to a subclass of rtl_opt_pass along with...
5074 (pass_data_mode_switch_use): ...new pass_data instance and...
5075 (make_pass_mode_switch_use): ...new function.
5076 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5077 from a global struct to a subclass of rtl_opt_pass along with...
5078 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5079 (make_pass_resolve_sw_modes): ...new function.
5080 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5081 struct to a subclass of rtl_opt_pass along with...
5082 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5083 (make_pass_insert_vzeroupper): ...new function.
5084 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5085 global struct to a subclass of rtl_opt_pass along with...
5086 (pass_data_work_around_errata): ...new pass_data instance and...
5087 (make_pass_work_around_errata): ...new function.
5088 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5089 struct to a subclass of rtl_opt_pass along with...
5090 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
5091 (make_pass_mips_machine_reorg2): ...new function.
5092
6a389ed5
DM		 50932013-08-05 David Malcolm <dmalcolm@redhat.com>
5094
5095 * passes.c (pass_manager::operator new): New.
5096
f7695dbf
DM		 50972013-08-05 David Malcolm <dmalcolm@redhat.com>
5098
5099 Handwritten part of conversion of passes to C++ classes.
5100
5101 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
5102 (toplev.o): Add dep on PASS_MANAGER_H.
5103 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
5104 of early local pases to reflect this moving from a global to a
5105 member of gcc::pass_manager.
5106 (cgraph_add_new_function): Likewise.
5107 * lto-cgraph.c (lto_output_node): Update for conversion of
5108 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
5109 * passes.c (opt_pass::clone): New.
5110 (opt_pass::gate): New.
5111 (opt_pass::execute): New.
5112 (opt_pass::opt_pass): New.
5113 (pass_manager::execute_early_local_passes): New.
5114 (pass_manager::execute_pass_mode_switching): new.
5115 (finish_optimization_passes): Convert to...
5116 (pass_manager::finish_optimization_passes): ...this.
5117 (finish_optimization_passes): Update for conversion of passes to
5118 C++ classes.
5119 (register_dump_files_1): Use has_gate since we cannot portably
5120 check a vtable entry against NULL.
5121 (dump_one_pass): Likewise.
5122 (ipa_write_summaries_2): Likewise.
5123 (ipa_write_optimization_summaries_1): Likewise.
5124 (ipa_read_summaries_1): Likewise.
5125 (ipa_read_optimization_summaries_1): Likewise.
5126 (execute_ipa_stmt_fixups): Likewise.
5127 (pass_manager::pass_manager): Rewrite pass-creation, invoking
5128 pass-creation functions rather than wiring up globals, and
5129 storing the results in fields of pass_manager generated using
5130 pass-instances.def.
5131 (pass_manager::dump_profile_report): Update for conversion of
5132 passes to C++ classes.
5133 (pass_manager::execute_ipa_summary_passes): Likewise.
5134 (execute_one_ipa_transform_pass): Likewise.
5135 (execute_one_pass): Use has_gate and has_execute since we cannot
5136 portably check a vtable entry against NULL.
5137 * pass_manager.h (pass_manager::finish_optimization_passes): New.
5138 (pass_manager): Use pass-instances.def to add fields for the
5139 various pass instances.
5140 * toplev.c (finalize): Update for move of
5141 finish_optimization_passes to a method of gcc::pass_manager.
5142 * toplev.h (finish_optimization_passes): Move to method of class
5143 pass_manager.
5144 * tree-pass.h (struct pass_data): New.
03b0ee0a 5145 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM		 5146 (opt_pass::gate): Convert to virtual function.
5147 (opt_pass::~opt_pass): New.
5148 (opt_pass::clone): New.
5149 (opt_pass::execute): Convert to virtual function.
5150 (opt_pass::opt_pass): New.
5151 (opt_pass::ctxt_): new.
5152 (gimple_opt_pass): Convert to subclass of opt_pass.
5153 (gimple_opt_pass::gimple_opt_pass): New.
5154 (rtl_opt_pass): Convert to subclass of opt_pass.
5155 (rtl_opt_pass::rtl_opt_pass): New.
5156 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5157 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5158 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5159 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5160 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5161 invocation of pass_mode_switching to reflect this moving from a
5162 global to a member of gcc::pass_manager.
5163 (ix86_option_override): Rework how pass_insert_vzeroupper is
5164 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 5165 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 5166
d0b2f831
RE		 51672013-08-05 Richard Earnshaw <rearnsha@arm.com>
5168
5169 PR rtl-optimization/57708
5170 * recog.c (peep2_find_free_register): Validate all regs in a
5171 multi-reg mode.
5172
51a5c0c2
JH		 51732013-08-05 Jan Hubicka <jh@suse.cz>
5174
5175 PR lto/57602
03b0ee0a
UB		 5176 * cgraph.c (verify_cgraph_node): Accept local flags from other
5177 partitions.
51a5c0c2
JH		 5178 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5179 (function_and_variable_visibility): Likewise.
5180 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5181
65f0a120
GDR		 51822013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5183
5184 * graph.c (init_graph_slim_pretty_print): Remove.
5185 (print_graph_cfg): Do not call it. Use local pretty printer.
5186 (start_graph_dump): Likewise.
5187
b3f80694
GDR		 51882013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5189
5190 * gimple-pretty-print.c (buffer): Remove.
5191 (initialized): Likewise.
5192 (maybe_init_pretty_print): Likewise.
5193 (print_gimple_stmt): Do not call it. Use non-static local
5194 pretty_printer variable.
5195 (print_gimple_expr): Likewise.
5196 (print_gimple_seq): Likewise.
5197 (gimple_dump_bb): Likewise.
5198
11a877b3
GDR		 51992013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5200
5201 * asan.c (asan_pp): Remove.
5202 (asan_pp_initialized): Likewise.
5203 (asan_pp_initialize): Likewise.
5204 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5205 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5206 (asan_add_global): Likewise.
5207
b066401f
GDR		 52082013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5209
5210 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5211 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5212 * pretty-print.c (pp_formatted_text_data): Likewise.
5213 (pp_indent): Rename from pp_base_indent.
5214 (pp_format): Rename from pp_base_format.
5215 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5216 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5217 (pp_flush): Rename from pp_base_flush.
5218 (pp_set_line_maximum_length): Rename from
5219 pp_base_set_line_maximum_length.
5220 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5221 (pp_set_prefix): Rename from pp_base_set_prefix.
5222 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5223 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5224 (pp_append_text): Rename from pp_base_append_text.
5225 (pp_formatted_text): Rename from pp_base_formatted_text.
5226 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5227 (pp_remaining_character_count_for_line): Rename from
5228 pp_base_remaining_character_count_for_line.
5229 (pp_newline): Rename from pp_base_newline.
5230 (pp_character): Rename from pp_base_character.
5231 (pp_string): Rename from pp_base_string.
5232 (pp_maybe_space): Rename from pp_base_maybe_space.
5233 * asan.c (asan_pp_string): Adjust.
5234 (asan_emit_stack_protection): Likewise.
5235 (asan_add_global): Likewise.
5236 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5237 * tree-mudflap.c (mf_varname_tree): Likewise.
5238 * tree-pretty-print.c (pp_tree_identifier): Rename from
5239 pp_base_tree_identifier.
5240 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5241 Declare as function.
5242
137a1a27
GDR		 52432013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5244
5245 * pretty-print.h (pp_bar_bar): New.
5246 (pp_ampersand_ampersand): Likewise.
5247 (pp_less_equal): Likewise.
5248 (pp_greater_equal): Likewise.
5249 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5250 printer functions instead of pp_string or operators and punctuators.
5251 (dump_gimple_call): Likewise.
5252 (dump_gimple_omp_for): Likewise.
5253 (dump_gimple_transaction): Likewise.
5254 (dump_gimple_phi): Likewise.
5255 (pp_gimple_stmt_1): Likewise.
5256 * sched-vis.c (print_insn): Likewise.
5257 * tree-mudflap.c (mf_varname_tree): Likewise.
5258 * tree-pretty-print.c (dump_block_node): Likewise.
5259 (dump_generic_node): Likewise.
5260
815effe1
JH		 52612013-08-02 Jan Hubicka <jh@suse.cz>
5262
5263 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5264 boundaries.
5265 * lto-streamer-out.c (tree_is_indexable): Results decls and
5266 parm decls are not indexable.
5267 (DFS_write_tree_body): Do not follow args and results.
5268 (hash_tree): Likewise.
5269 (output_functions): Rearrange so struct function is needed
5270 only when real body is output; be able to also ouptut abstract
5271 functions; output DECL_ARGUMENTS and DECL_RESULT.
5272 (lto_output): When not in WPA, ale store abstract functions.
5273 (write_symbol): Do not care about RESULT_DECL.
5274 (output_symbol_p): Handle correctly sbtract decls.
5275 * lto-streamer-in.c (input_function): Rearrange so struct
5276 function can be NULL at entry; allow streaming of
5277 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5278 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5279 sanity check during LTO.
5280 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5281 RESULT_DECl and DECL_ARGUMENTS.
5282 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5283 Likewise.
5284
07838b13
GDR		 52852013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5286
5287 * pretty-print.h (pp_underscore): New.
5288 (pp_comma): Tidy.
5289 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5290 printer functions instead of pp_character.
5291 (dump_binary_rhs): Likewise.
5292 (dump_ternary_rhs): Likewise.
5293 (dump_gimple_call_args): Likewise.
5294 (pp_points_to_solution): Likewise.
5295 (dump_gimple_call): Likewise.
5296 (dump_gimple_switch): Likewise.
5297 (dump_gimple_cond): Likewise.
5298 (dump_gimple_bind): Likewise.
5299 (dump_gimple_try): Likewise.
5300 (dump_gimple_omp_for): Likewise.
5301 (dump_gimple_omp_continue): Likewise.
5302 (dump_gimple_omp_single): Likewise.
5303 (dump_gimple_omp_sections): Likewise.
5304 (dump_gimple_omp_block): Likewise.
5305 (dump_gimple_omp_critical): Likewise.
5306 (dump_gimple_transaction): Likewise.
5307 (dump_gimple_asm): Likewise.
5308 (dump_gimple_phi): Likewise.
5309 (dump_gimple_omp_parallel): Likewise.
5310 (dump_gimple_omp_task): Likewise.
5311 (dump_gimple_omp_atomic_load): Likewise.
5312 (dump_gimple_omp_atomic_store): Likewise.
5313 (dump_gimple_mem_ops): Likewise.
5314 (pp_gimple_stmt_1): Likewise.
5315 (pp_cfg_jump): Likewise.
5316 (dump_implicit_edges): Likewise.
5317 (gimple_dump_bb_for_graph): Likewise.
5318 * graph.c (draw_cfg_node): Likewise.
5319 * langhooks.c (lhd_print_error_function): Likewise.
5320 * sched-vis.c (print_exp): Likewise.
5321 (print_value): Likewise.
5322 (print_pattern): Likewise.
5323 (print_insn): Likewise.
5324 (rtl_dump_bb_for_graph): Likewise.
5325 * tree-pretty-print.c (dump_function_declaration): Likewise.
5326 (dump_array_domain): Likewise.
5327 (dump_omp_clause): Likewise.
5328 (dump_location): Likewise.
5329 (dump_generic_node): Likewise.
5330 (print_struct_decl): Likewise.
5331 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5332
bb0d2039
BS		 53332013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5334
5335 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5336 candidate table when replacing a candidate statement.
5337 (replace_rhs_if_not_dup): Likewise.
5338 (replace_one_candidate): Likewise.
5339
bc0ec027 53402013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 5341 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH		 5342
5343 * cgraphunit.c (add_new_function): Fix logic when adding from
5344 late IPA pass.
5345 (assemble_thunk): Rename to ...
03b0ee0a
UB		 5346 (expand_thunk); .. this one; export; get it working with
5347 general functions; make produced gimple valid.
bc0ec027
JH		 5348 * cgraph.h (expand_thunk): Declare.
5349
0e8853ee
JH		 53502013-08-02 Jan Hubicka <jh@suse.cz>
5351
03b0ee0a
UB		 5352 * ipa-cp.c (gather_context_independent_values): Use
5353 ipa_get_param_move_cost.
5354 (get_replacement_map): Remove PARAM; move parameter folding
5355 into tree-inline.c
0e8853ee
JH		 5356 (create_specialized_node): Update.
5357 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
5358 assert that we have gimple body; update move_cost.
5359 (count_formal_params): Assert that we have gimple body.
5360 (ipa_dump_param): New function.
5361 (ipa_alloc_node_params): Break out from ...
5362 (ipa_initialize_node_params): ... here.
5363 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
5364 (ipa_write_node_info): Stream move costs.
5365 (ipa_read_node_info): Read move costs.
5366 (ipa_update_after_lto_read): Do not recompute node params.
5367 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
5368 (ipa_get_param): Check we are not in WPA.
5369 (ipa_get_param_move_cost): New.
5370 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
5371 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
5372 parm numbers to be present.
5373
ecee672b
VM		 53742013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5375
5376 PR rtl-optimization/58048
5377 * lra-constraints.c (process_alt_operands): Don't check asm
5378 operand on register.
5379
13ec0527
EB		 53802013-08-02 Eric Botcazou <ebotcazou@adacore.com>
5381
5382 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
5383 the implied StoreLoad barrier for atomic operations if before.
5384
af15184a
JH		 53852013-08-02 Jan Hubicka <jh@suse.cz>
5386 Martin Liska <marxin.liska@gmail.com>
5387
03b0ee0a
UB		 5388 * cgraph.c (cgraph_function_body_availability): Do not check
5389 cgraph flags.
af15184a
JH		 5390 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
5391 symtab_node_availability): Declare.
5392 * ipa.c (can_replace_by_local_alias): New.
5393 (function_and_variable_visibility): Use it.
03b0ee0a
UB		 5394 * symtab.c (symtab_for_node_and_aliases,
5395 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 5396
01e54ef8
VM		 53972013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5398
5399 PR rtl-optimization/57963
03b0ee0a 5400 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM		 5401 (lra_constraints): Use them.
5402
e2fd7ca7
SN		 54032013-08-02 Sofiane Naci <sofiane.naci@arm.com>
5404
03b0ee0a
UB		 5405 * config/arm/types.md (define_attr "type"): Add "load_acq"
5406 and "store_rel".
e2fd7ca7
SN		 5407 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
5408 changes.
5409 (cortex_a53_store1): Likewise.
5410
0dcc26c3
JH		 54112013-08-01 Jan Hubicka <jh@suse.cz>
5412
03b0ee0a
UB		 5413 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
5414 partitions are not needed.
0dcc26c3 5415
0bd72901
UB		 54162013-08-01 Uros Bizjak <ubizjak@gmail.com>
5417
5418 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
5419 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
5420 MAYBE_NON_Q_CLASS_P where appropriate.
5421
12123452
JH		 54222013-08-01 Jan Hubicka <jh@suse.cz>
5423
5424 * cgraph.h (release_function_body): Declare.
5425 * tree.c (free_lang_data_in_decl): Free, parameters and return values
5426 of unused delcarations.
5427
ce852f9c
KT		 54282013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5429
03b0ee0a
UB		 5430 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
5431 RTL form when subtracting a constant.
ce852f9c 5432
c743b246
KT		 54332013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5434
5435 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
5436 Generate canonical plus rtx with negated immediate instead of minus
5437 where appropriate.
5438 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
5439
c0c123ef
JH		 54402013-08-01 Jan Hubicka <jh@suse.cz>
5441
5442 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
5443 (cgraph_release_function_body): Likewise.
5444 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5445 * cgraph.h (cgrpah_node): Rename abstract_and_needed
5446 to used_as_abstract_origin.
5447 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
5448 symbols used as abstract origins.
5449 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB		 5450 * ipa.c (symtab_remove_unreachable_nodes): Recompute
5451 used_as_abstract_origin.
c0c123ef 5452 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB		 5453 used_as_abstract_origin; be ready for DECL_RESULT and
5454 DECL_ARGUMENTS to be NULL.
c0c123ef 5455
0bd72901
UB		 5456 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
5457 for abstract functions.
5458 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
5459 real symbols.
c0c123ef 5460
1f029433
JH		 54612013-08-01 Jan Hubicka <jh@suse.cz>
5462
5463 * profile.c (compute_value_histograms): Fix thinko.
5464
c451f4d6
SN		 54652013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5466
5467 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
5468 aarch-common-protos.h to extra_headers.
5469 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5470 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
5471 * config/aarch64/t-aarch64 (aarch-common.o): Define.
5472
34b512f5
SN		 54732013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5474
5475 * config/aarch64/aarch64.md (define_attr "type"): Delete.
5476 Include "../arm/types.md". Define "type" attribute for all patterns.
5477 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
5478 attribute changes.
5479
d86e633a
MM		 54802013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
5481
5482 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
5483 to support power8 load fusion.
5484 (fusion_gpr_mem_load): Likewise.
5485
5486 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
5487
5488 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
5489 declarations for power8 load fusion.
5490 (emit_fusion_gpr_load): Likewise.
5491
5492 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
5493 tuning for power8, turn on fusion mode by default. Turn on sign
5494 extending fusion mode if normal fusion mode is on, and we are at
5495 -O2 or -O3.
5496 (fusion_gpr_load_p): New function, return true if we can fuse an
5497 addis instruction with a dependent load to a GPR.
5498 (emit_fusion_gpr_load): Emit the instructions for power8 load
5499 fusion to GPRs.
5500
0bd72901 5501 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM		 5502 (VSX load fusion peepholes): New peepholes to fuse together an
5503 addi instruction with a VSX load instruction.
5504
5505 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
5506 peepholes to fuse an addis instruction with a load to a GPR base
5507 register. If we are supporting sign extending fusions, convert
5508 sign extending loads to zero extending loads and add an explicit
5509 sign extension.
5510
c3f35647
SN		 55112013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5512
5513 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
5514 aarch-common-protos.h to extra_headers.
5515 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5516 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
5517 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 5518 (arm_no_early_alu_shift_dep): Likewise.
5519 (arm_no_early_alu_shift_value_dep): Likewise.
5520 (arm_no_early_mul_dep): Likewise.
5521 (arm_no_early_store_addr_dep): Likewise.
5522 (arm_mac_accumulator_is_mul_result): Likewise.
5523 (arm_mac_accumulator_is_result): Likewise.
c3f35647 5524 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB		 5525 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
5526 here to ...
c3f35647 5527 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 5528 (arm_no_early_alu_shift_dep): Likewise.
5529 (arm_no_early_alu_shift_value_dep): Likewise.
5530 (arm_no_early_mul_dep): Likewise.
5531 (arm_no_early_store_addr_dep): Likewise.
5532 (arm_mac_accumulator_is_mul_result): Likewise.
5533 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN		 5534 * config/arm/aarch-common-protos.h: ... here. New file.
5535 * config/arm/t-arm (aarch-common.o): Define.
5536
55372013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5538
5539 * config/arm/arm.md: Include new file "types.md".
5540 (define_attr "type"): Move from here to ...
5541 (define_attr "mul32"): Likewise.
5542 (define_attr "mul64"): Likewise.
5543 * config/arm/types.md: ... here. New file.
5544
228c1313
SH		 55452013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
5546
5547 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
5548 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
5549
d6e9046d 55502013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG		 5551
5552 * gen-pass-instances.awk: Fix offset of substr().
5553
a167b052
DM		 55542013-07-31 David Malcolm <dmalcolm@redhat.com>
5555
5556 * Makefile.in (pass-instances.def): New.
5557 (passes.o): Replace dependency on passes.def with one on
5558 pass-instances.def
5559
5560 * gen-pass-instances.awk: New.
5561
5562 * passes.c (pass_manager::pass_manager): Use pass-instances.def
5563 rather than passes.def, updating local definition of NEXT_PASS
5564 macro to add an extra NUM parameter (currently unused).
5565
315f8c0e
DM		 55662013-07-30 David Malcolm <dmalcolm@redhat.com>
5567
5568 * Makefile.in (PASS_MANAGER_H): New.
5569 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
5570 (passes.o): Likewise.
5571 (statistics.o): Likewise.
5572 (cgraphunit.o): Likewise.
5573 (context.o): Depend on PASS_MANAGER_H.
5574
5575 * pass_manager.h: New.
5576
5577 * cgraphunit.c (cgraph_add_new_function): Update for moves
5578 of globals to fields of pass_manager.
5579 (analyze_function): Likewise.
5580 (expand_function): Likewise.
5581 (ipa_passes): Likewise.
5582 (compile): Likewise.
5583
5584 * context.c (context::context): New.
5585 * context.h (context::context): New.
5586 (context::get_passes): New.
5587 (context::passes_): New.
5588
5589 * lto-cgraph.c (input_node): Update for moves of globals to
5590 fields of pass_manager.
5591
5592 * passes.c (all_passes): Remove, in favor of a field of the
5593 same name within the new class pass_manager.
5594 (all_small_ipa_passes): Likewise.
5595 (all_lowering_passes): Likewise.
5596 (all_regular_ipa_passes): Likewise.
5597 (all_late_ipa_passes): Likewise.
5598 (all_lto_gen_passes): Likewise.
5599 (passes_by_id): Likewise.
5600 (passes_by_id_size): Likewise.
5601 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
5602 the new class pass_manager.
5603 (set_pass_for_id): Convert to...
5604 (pass_manager::set_pass_for_id): ...method.
5605 (get_pass_for_id): Convert to...
5606 (pass_manager::get_pass_for_id): ...method.
5607 (register_one_dump_file): Move body of implementation into...
5608 (pass_manager::register_one_dump_file): ...here.
5609 (register_dump_files_1): Convert to...
5610 (pass_manager::register_dump_files_1): ...method.
5611 (register_dump_files): Convert to...
5612 (pass_manager::register_dump_files): ...method.
5613 (create_pass_tab): Update for moves of globals to fields of
5614 pass_manager.
5615 (dump_passes): Move body of implementation into...
5616 (pass_manager::dump_passes): ...here.
5617 (register_pass): Move body of implementation into...
5618 (pass_manager::register_pass): ...here.
5619 (init_optimization_passes): Convert into...
5620 (pass_manager::pass_manager): ...constructor for new
5621 pass_manager class, and initialize the pass_lists array.
5622 (check_profile_consistency): Update for moves of globals to
5623 fields of pass_manager.
5624 (dump_profile_report): Move body of implementation into...
5625 (pass_manager::dump_profile_report): ...here.
5626 (ipa_write_summaries_1): Update for moves of pass lists from
5627 being globals to fields of pass_manager.
5628 (ipa_write_optimization_summaries): Likewise.
5629 (ipa_read_summaries): Likewise.
5630 (ipa_read_optimization_summaries): Likewise.
5631 (execute_all_ipa_stmt_fixups): Likewise.
5632
5633 * statistics.c (statistics_fini): Update for moves of globals to
5634 fields of pass_manager.
5635
5636 * toplev.c (general_init): Replace call to
5637 init_optimization_passes with construction of the pass_manager
5638 instance.
5639
5640 * tree-pass.h (all_passes): Remove, in favor of a field of the
5641 same name within the new class pass_manager.
5642 (all_small_ipa_passes): Likewise.
5643 (all_lowering_passes): Likewise.
5644 (all_regular_ipa_passes): Likewise.
5645 (all_lto_gen_passes): Likewise.
5646 (all_late_ipa_passes): Likewise.
5647 (passes_by_id): Likewise.
5648 (passes_by_id_size): Likewise.
5649 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
5650 the new class pass_manager.
5651 (get_pass_for_id): Remove.
5652
02cab1c1
RE		 56532013-07-30 Richard Earnshaw <rearnsha@arm.com>
5654
5655 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
5656 configs.
5657
b2d3f886
RE		 56582013-07-30 Richard Earnshaw <rearnsha@arm.com>
5659
5660 * arm.md (mulhi3): New expand pattern.
5661
605e86fa
JH		 56622013-07-30 Jan Hubicka <jh@suse.cz>
5663 Martin Liska <marxin.liska@gmail.com>
5664
5665 * profile.c (compute_value_histograms): Do not ICE when
5666 there is mismatch only on some counters.
5667
14a87636
ZC		 56682013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5669
5670 PR rtl-optimization/57637
5671 * function.c (move_insn_for_shrink_wrap): Also check the
5672 GEN set of the LIVE problem for the liveness analysis
5673 if it exists, otherwise give up.
5674
0100cd3f
BS		 56752013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5676
5677 PR tree-optimization/57993
5678 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
5679 replaced statement in the candidate table.
5680 (phi_add_costs): Return infinite cost when the hidden basis does
5681 not dominate all phis on which the candidate is dependent.
5682 (replace_one_candidate): Record replaced statement in the
5683 candidate table.
5684
f0d811f7
JR		 56852013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
5686
5687 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
5688 (ashlv2si3): New expander.
5689 (*ashlv2si3_i): New define_insn_and_split.
5690 * predicates.md (float_operation): Allow patterns with three
5691 basic sub-patterns.
5692
bba33211
JR		 5693 PR rtl-optimization/58021
5694 * mode-switching.c (create_pre_exit): Always split off preceding
5695 insns if we are not at the basic block head.
5696
ff3f3951
MR		 56972013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
5698
5699 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
5700 (UCLIBC_DYNAMIC_LINKER): New macro.
5701 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
5702 `-mnan=2008'.
5703 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
5704 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
5705 `-mnan=2008'.
5706 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
5707 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
5708 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
5709 for SF and DF modes. Use ieee_quad_format for TF mode.
5710 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
5711 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
5712 (mips_option_override): Handle `-mnan=legacy'.
5713 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
5714 `-mabs=2008' and `-mnan=2008'.
5715 (OPTION_DEFAULT_SPECS): Add "nan" default.
5716 (ASM_SPEC): Handle `-mnan='.
5717 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
5718 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
5719 comment accordingly.
5720 (neg<mode>2): Likewise.
5721 * config/mips/mips.opt (mabs, mnan): New options.
5722 * doc/install.texi (Configuration): Document `--with-nan=' option.
5723 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
5724 `-mnan=' options.
5725 (MIPS Options): Document them.
5726 * config.gcc <mips*-*-*>: Handle `--with-nan='.
5727 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
5728 * configure: Regenerate.
5729 * config.in: Regenerate.
5730
576e2f37
UB		 57312013-07-29 Uros Bizjak <ubizjak@gmail.com>
5732
5733 * config/i386/i386.md (float post-reload splitters): Do not check
5734 for subregs of SSE registers.
5735
57362013-07-29 Uros Bizjak <ubizjak@gmail.com>
5737 H.J. Lu <hongjiu.lu@intel.com>
5738
5739 PR target/57954
5740 PR target/57988
5741 * config/i386/i386.md (post-reload splitter
5742 to avoid partial SSE reg dependency stalls): New pattern.
5743
a71f0749
DV		 57442013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
5745
5746 * config/s390/s390.md ("movcc"): Swap load and store instructions.
5747
4c97f1cc
JR		 57482013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
5749
5750 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
5751 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
5752
b0c733d8
CC		 57532013-07-26 Cary Coutant <ccoutant@google.com>
5754
5755 * dwarf2out.c (die_checksum_ordered): Don't include template
5756 instantiations in signature.
5757 (is_template_parameter): New function.
5758 (is_template_instantiation): New function.
5759 (generate_skeleton_bottom_up): Don't include template instantiations
5760 in type unit DIE.
5761 (generate_skeleton): Likewise.
5762 (break_out_comdat_types): Move recursive call to break out nested
5763 types earlier.
5764 (prune_unused_types_mark_generic_parms_dies): Call
5765 is_template_parameter.
5766
040d8a1c
IB		 57672013-07-26 Ian Bolton <ian.bolton@arm.com>
5768
5769 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
5770 uses vector registers.
89fdc743 5771 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 5772
85bd4ac6 57732013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 5774 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT		 5775
5776 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 5777 where appropriate.
85bd4ac6
KT		 5778 * config/arm/arm.md (movcond_addsi): New splitter.
5779
b6051207
SA		 57802013-07-25 Sterling Augustine <saugustine@google.com>
5781
5782 * dwarf2out.c (size_of_pubnames): Move code to...
5783 (include_pubname_in_output): ...here. New.
5784 (want_pubnames): Rearrange.
5785 (output_pubnames): Call include_pubname_in_output. Move assertion.
5786
41a7c215
CM		 57872013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5788
5789 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
5790
57912013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5792
5793 PR target/38836
5794 * doc/extend.texi: Remove obsolete builtins. Fix
5795 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
5796
49bde175
JH		 57972013-07-25 Jan Hubicka <jh@suse.cz>
5798
5799 * cgraph.c (release_function_body): Break out from ...
5800 (cgraph_release_function_body): ... this one; also release DECL_RESULT
5801 and DECL_ARGUMENTS.
5802 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
5803 old_tree in the map.
5804 (create_specialized_node): Update.
5805 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
5806 into index.
41a7c215
CM		 5807 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
5808 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH		 5809 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
5810 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
5811 DECL_RESULT.
5812
4870352d
KT		 58132013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5814
5815 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
5816 addsi3_carryin_alt2_<optab>): Correct output template.
5817
1d6aee1c
KT		 58182013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5819
5820 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
5821 Adjust for arm_restrict_it.
5822 Remove trailing whitespace.
5823
c7cafd75
MK		 58242013-07-25  Mark Kettenis  <kettenis@openbsd.org>
5825
017d38f5
MK		 5826 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
5827 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
5828
c7cafd75
MK		 5829 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
5830
00b7527b
VM		 58312013-07-25 Vladimir Makarov <vmakarov@redhat.com>
5832
5833 PR rtl-optimization/57960
5834 * lra-constraints.c (process_alt_operands): Use the right mode
5835 when checking strict_low.
5836
4f63dfc6
JH		 58372013-07-25 Jan Hubicka <jh@suse.cz>
5838
5839 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
5840 * cgraph.c (cgraph_remove_node): Do not release function body
5841 when in cgraph streaming.
5842 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM		 5843 in other partitions are not considered reachable; fix handling of
5844 clones.
4f63dfc6 5845
e6dcfa18
RR		 58462013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5847
5848 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
5849
0b93d3b6
RR		 58502013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5851
5852 PR target/19599
5853 PR target/57731
bb80c2eb 5854 PR target/57837
e6dcfa18 5855 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 5856 Ss with US. Adjust output for v5 and v4t.
41a7c215 5857 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 5858
0b93d3b6
RR		 5859 * config/arm/constraints.md ("Ss"): Rename to US.
5860
690688b3
TG		 58612013-07-25 Terry Guo <terry.guo@arm.com>
5862
5863 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
5864 shift_add/shift_sub0/shift_sub1 RTXs.
5865
a4ad093b 58662013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 5867 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS		 5868
5869 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
5870 (altivec_vpks<VI_char>ss): Likewise.
5871 (altivec_vpks<VI_char>us): Likewise.
5872 (altivec_vpku<VI_char>us): Likewise.
5873 (altivec_vpku<VI_char>um): Likewise.
5874
8c5005ce
DM		 58752013-07-24 David Malcolm <dmalcolm@redhat.com>
5876
5877 Introduce context class.
5878
5879 * Makefile.in (CONTEXT_H): New.
5880 (OBJS): Add context.o.
5881 (toplev.o): Add CONTEXT_H to dependencies.
5882 (context.o): New.
5883
41a7c215 5884 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM		 5885
5886 * context.c: New.
5887
5888 * context.h: New.
5889
ce4a9422
JR		 58902013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
5891
5892 PR rtl-optimization/57968
5893 * mode-switching.c (create_pre_exit): Allow instructions that
5894 don't set a return register to need a non-exit mode.
5895
493f4c9e 58962013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 5897 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 5898
b5860fd3
WS		 5899 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
5900 operands to vperm for little endian.
5901 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
5902 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 5903
73792b92 59042013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 5905 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS		 5906
5907 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
5908 two operands for little-endian.
5909
14c21302
SE		 59102013-07-23 Steve Ellcey <sellcey@mips.com>
5911
5912 * config/mips/mips.c (mips_case_values_threshold): New.
5913 (TARGET_CASE_VALUES_THRESHOLD): Define.
5914
a6056198 59152013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 5916 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS		 5917
5918 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
5919 selection of field for vector splat in little endian mode.
5920
dd7a40e1
MM		 59212013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
5922
5923 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
5924 expanders to rs6000.md.
5925 (ior<mode>3): Likewise.
5926 (and<mode>3): Likewise.
5927 (one_cmpl<mode>2): Likewise.
5928 (nor<mode>3): Likewise.
5929 (andc<mode>3): Likewise.
5930 (eqv<mode>3): Likewise.
5931 (nand<mode>3): Likewise.
5932 (orc<mode>3): Likewise.
5933
5934 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
5935 declaration.
5936
5937 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
5938 to split multi-word logical operations.
5939 (rs6000_split_logical_di): Likewise.
5940 (rs6000_split_logical): Likewise.
5941
5942 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
5943 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
5944 and allow TImode operations in 32-bit.
5945 (vsx_and<mode>3_64bit): Likewise.
5946 (vsx_ior<mode>3_32bit): Likewise.
5947 (vsx_ior<mode>3_64bit): Likewise.
5948 (vsx_xor<mode>3_32bit): Likewise.
5949 (vsx_xor<mode>3_64bit): Likewise.
5950 (vsx_one_cmpl<mode>2_32bit): Likewise.
5951 (vsx_one_cmpl<mode>2_64bit): Likewise.
5952 (vsx_nor<mode>3_32bit): Likewise.
5953 (vsx_nor<mode>3_64bit): Likewise.
5954 (vsx_andc<mode>3_32bit): Likewise.
5955 (vsx_andc<mode>3_64bit): Likewise.
5956 (vsx_eqv<mode>3_32bit): Likewise.
5957 (vsx_eqv<mode>3_64bit): Likewise.
5958 (vsx_nand<mode>3_32bit): Likewise.
5959 (vsx_nand<mode>3_64bit): Likewise.
5960 (vsx_orc<mode>3_32bit): Likewise.
5961 (vsx_orc<mode>3_64bit): Likewise.
5962
5963 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
5964 logical types in GPRs.
5965
5966 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
5967 logical insns to rs6000.md, and allow TImode operations in
5968 32-bit.
5969 (altivec_ior<mode>3): Likewise.
5970 (altivec_xor<mode>3): Likewise.
5971 (altivec_one_cmpl<mode>2): Likewise.
5972 (altivec_nor<mode>3): Likewise.
5973 (altivec_andc<mode>3): Likewise.
5974
5975 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
5976 attributes for moving the 128-bit logical operations into
5977 rs6000.md.
5978 (BOOL_REGS_OUTPUT): Likewise.
5979 (BOOL_REGS_OP1): Likewise.
5980 (BOOL_REGS_OP2): Likewise.
5981 (BOOL_REGS_UNARY): Likewise.
5982 (BOOL_REGS_AND_CR0): Likewise.
5983 (one_cmpl<mode>2): Add support for DI logical operations on
5984 32-bit, splitting the operations to 32-bit.
5985 (anddi3): Likewise.
5986 (iordi3): Likewise.
5987 (xordi3): Likewise.
5988 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
5989 changes to combine the 32/64-bit code, allow logical operations on
5990 TI mode in 32-bit, and to use similar match_operator patterns like
5991 scalar mode uses. Combine the Altivec and VSX code for logical
5992 operations, and move it here.
5993 (ior<mode>3, 128-bit types): Likewise.
5994 (xor<mode>3, 128-bit types): Likewise.
5995 (one_cmpl<mode>3, 128-bit types): Likewise.
5996 (nor<mode>3, 128-bit types): Likewise.
5997 (andc<mode>3, 128-bit types): Likewise.
5998 (eqv<mode>3, 128-bit types): Likewise.
5999 (nand<mode>3, 128-bit types): Likewise.
6000 (orc<mode>3, 128-bit types): Likewise.
6001 (and<mode>3_internal): Likewise.
6002 (bool<mode>3_internal): Likewise.
6003 (boolc<mode>3_internal1): Likewise.
6004 (boolc<mode>3_internal2): Likewise.
6005 (boolcc<mode>3_internal1): Likewise.
6006 (boolcc<mode>3_internal2): Likewise.
6007 (eqv<mode>3_internal1): Likewise.
6008 (eqv<mode>3_internal2): Likewise.
6009 (one_cmpl1<mode>3_internal): Likewise.
6010
12211b99 60112013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH		 6012
6013 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6014 Rename flag_stack_usage to flag_stack_usage_info.
6015
12211b99 60162013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 6017
a5f257fd
ME		 6018 * config/microblaze/sync.md: New file.
6019 * config/microblaze/microblaze.md: Include sync.md
6020 * config/microblaze/microblaze.c: Add print_operand 'y'.
6021 * config/microblaze/constraints.md: Add memory_contraint
6022 'Q' which is a single register.
450b0ebb 6023
49fba14f
EB		 60242013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6025
6026 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6027
c7b3b99f
PCC		 60282013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6029
6030 * reload.c (find_reloads): Exit loop once we find this operand
6031 cannot be reloaded somehow for this alternative.
6032
6033 * reload.c (find_reloads): Exit loop once we find a hard register.
6034
6035 * rtlanal.c (computed_jump_p): Exit loop once we find label
6036 reference is used.
6037
6038 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6039
6040 * cfgloopmanip.c (remove_path): Exit loop after setting
6041 irred_invalidated.
6042
6043 * gensupport.c (subst_dup): Avoid loop if code is not
6044 MATCH_DUP nor MATCH_OP_DUP.
6045
c8fbf1fa
NBJ		 60462013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6047
6048 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6049
9259db42
YZ		 60502013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6051
6052 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6053 true for SP_REGNUM if mode == ptr_mode.
6054 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6055 with value R0_REGNUM + 31.
6056
00edcfbe
YZ		 60572013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6058
6059 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6060 pad pointer-typed argument downward.
6061
43be9a95
YZ		 60622013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6063
6064 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6065 and __ILP32__ when the ILP32 model is in use.
6066
28514dda
YZ		 60672013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6068
6069 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6070 (aarch64_load_symref_appropriately): In the case of
6071 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6072 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6073 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6074 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6075 if 'mode' doesn't equal to 'ptr_mode'.
6076 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6077 'vcall_offset'; change to call aarch64_emit_move differently depending
6078 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6079 to calculate the upper bound of 'vcall_offset'.
6080 (aarch64_cannot_force_const_mem): Change to also return true if
6081 mode != ptr_mode.
6082 (aarch64_legitimize_reload_address): In the case of large
6083 displacements, add new local variable 'xmode' and an assertion
6084 based on it; change to use 'xmode' to generate the new rtx and
6085 reload.
6086 (aarch64_asm_trampoline_template): Change to generate the template
6087 differently depending on TARGET_ILP32 or not; change to use
6088 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6089 (aarch64_trampoline_size): Removed.
6090 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
6091 and replace immediate literals with it. Change to use 'ptr_mode'
6092 instead of 'DImode' and call convert_memory_address if the mode
6093 of 'fnaddr' doesn't equal to 'ptr_mode'.
6094 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
6095 to output symbol.
6096 (aarch64_elf_asm_destructor): Likewise.
6097 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
6098 on TARGET_ILP32 instead of aarch64_trampoline_size.
6099 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
6100 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
6101 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
6102 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
6103 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
6104 (storewb_pair<GPI:mode>_<P:mode>): ... this.
6105 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
6106 depending on the value of 'mode'.
6107 (add_losym_<mode>): New.
6108 (ldr_got_small_<mode>): New, based on ldr_got_small.
6109 (ldr_got_small): Remove.
6110 (ldr_got_small_sidi): New.
6111 * config/aarch64/iterators.md (P): New.
6112 (PTR): Change to 'ptr_mode' in the condition.
6113
17a819cb
YZ		 61142013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6115
6116 * config.gcc (aarch64*-*-*): Support --with-abi.
6117 (aarch64*-*-elf): Support --with-multilib-list.
6118 (aarch64*-*-linux*): Likewise.
6119 (supported_defaults): Add abi to aarch64*-*-*.
6120 * configure.ac: Mention AArch64 for --with-multilib-list.
6121 * configure: Re-generated.
6122 * config/aarch64/biarchilp32.h: New file.
6123 * config/aarch64/biarchlp64.h: New file.
6124 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
6125 (ABI_SPEC): Ditto.
6126 (MULTILIB_DEFAULTS): Ditto.
6127 (DRIVER_SELF_SPECS): Ditto.
6128 (ASM_SPEC): Update to also substitute -mabi.
6129 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
6130 file whose name depends on -mabi= and -mbig-endian.
6131 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
6132 TARGET_ILP32.
6133 (POINTER_SIZE): New define.
6134 (POINTERS_EXTEND_UNSIGNED): Ditto.
6135 (enum aarch64_abi_type): New enumeration tag.
6136 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
6137 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
6138 (TARGET_ILP32): New define.
6139 * config/aarch64/aarch64.opt (mabi): New.
6140 (aarch64_abi): New.
6141 (ilp32, lp64): New values for -mabi.
6142 * config/aarch64/t-aarch64 (comma): New define.
6143 (MULTILIB_OPTIONS): Ditto.
6144 (MULTILIB_DIRNAMES): Ditto.
6145 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
6146 * doc/invoke.texi: Document -mabi for AArch64.
6147
9f9cbdce
GJL		 61482013-07-23 Georg-Johann Lay <avr@gjlay.de>
6149
6150 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
6151
7061977b 61522013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 6153 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS		 6154
6155 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6156 endianness when selecting field to splat.
6157
75a41878
EC		 61582013-07-22 Eric Christopher <echristo@gmail.com>
6159
6160 * dwarf2out.c (die_odr_checksum): New function to use
6161 CHECKSUM_ macros and ULEB128 for DIE tag.
6162 (generate_type_signature): Use.
41a7c215 6163
38ae58ca
EB		 61642013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6165
6166 * config.gcc (sparc*-*-*): Accept leon3 processor.
6167 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6168 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6169 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6170 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6171 (mfix-ut699): Adjust comment.
6172 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6173 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6174 (CPP_CPU_SPEC): Likewise.
6175 (ASM_CPU_SPEC): Likewise.
6176 * config/sparc/sparc.c (leon3_cost): New constant.
6177 (sparc_option_override): Add leon3 support.
6178 (mem_ref): New function.
6179 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6180 (sparc_do_work_around_errata): Look into the instruction in the delay
6181 slot and adjust accordingly. Add fix for the data cache nullify issues
6182 of the UT699. Change insertion position for the NOP.
6183 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6184 (leon3_load): New reservation.
6185 (leon_store): Bump latency to 2.
6186 (grfpu): New automaton.
6187 (grfpu_alu): New unit.
6188 (grfpu_ds): Likewise.
6189 (leon_fp_alu): Adjust.
6190 (leon_fp_mult): Delete.
6191 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6192 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6193 * config/sparc/sparc.md (cpu): Add leon3.
6194 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6195 (swapsi): Likewise.
6196 (atomic_test_and_set): Likewise.
6197 (ldstub): Likewise.
6198
04dfc6df
JU		 61992013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6200
6201 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6202 default for R5900 targets.
6203 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6204 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6205 * config/mips/mips.c (mips_option_override): Report an error for
6206 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6207 for -march=r5900 -mhard-float.
6208
c7b3b99f 62092013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC		 6210
6211 * df-problems.c (can_move_insns_across): Exit loop once we
6212 find a non-fixed, non-global register.
6213
6214 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6215 setting can_throw.
6216
6217 * omega.c (omega_eliminate_red): Break after setting red_found.
6218 (omega_problem_has_red_equations): Similarly after setting found.
6219 (omega_query_variable): Similarly after setting coupled.
6220
9b6e6981
MP		 62212013-07-22 Marek Polacek <polacek@redhat.com>
6222
6223 * gimplify.c: Don't include gimple.h twice.
6224
5e5f7673
KT		 62252013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6226
6227 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6228 instead of TARGET_THUMB1.
6229 (Pz): New constraint.
6230 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6231 encodings.
6232 (compare_negsi_si): Likewise.
6233 (compare_addsi2_op0): Likewise.
6234 (compare_addsi2_op1): Likewise.
6235 (addsi3_carryin_<optab>): Likewise.
6236 (addsi3_carryin_alt2_<optab>): Likewise.
6237 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6238 for arm_restrict_it.
6239 (subsi3_carryin): Likewise.
6240 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6241 (minmax_arithsi): Disable for arm_restrict_it.
6242 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6243 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6244 (satsi_<SAT:code>_shift): Likewise.
6245 (arm_shiftsi3): Add alternative for 16-bit encoding.
6246 (arm32_movhf): Disable for arm_restrict_it.
6247 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6248 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6249
859abddd
SN		 62502013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6251
6252 * config/arm/arm.md (attribute "insn"): Delete.
6253 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6254 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6255 (not_shiftsi): Update for attribute change.
6256 (not_shiftsi_compare0): Likewise.
6257 (not_shiftsi_compare0_scratch): Likewise.
6258 (arm_one_cmplsi2): Likewise.
6259 (thumb1_one_cmplsi2): Likewise.
6260 (notsi_compare0): Likewise.
6261 (notsi_compare0_scratch): Likewise.
6262 (thumb1_movdi_insn): Likewise.
6263 (arm_movsi_insn): Likewise.
6264 (movhi_insn_arch4): Likewise.
6265 (movhi_bytes): Likewise.
6266 (arm_movqi_insn): Likewise.
6267 (thumb1_movqi_insn): Likewise.
6268 (arm32_movhf): Likewise.
6269 (thumb1_movhf): Likewise.
6270 (arm_movsf_soft_insn): Likewise.
6271 (thumb1_movsf_insn): Likewise.
6272 (thumb_movdf_insn): Likewise.
6273 (movsicc_insn): Likewise.
6274 (movsfcc_soft_insn): Likewise.
6275 (and_scc): Likewise.
6276 (cond_move): Likewise.
6277 (if_move_not): Likewise.
6278 (if_not_move): Likewise.
6279 (if_shift_move): Likewise.
6280 (if_move_shift): Likewise.
6281 (if_shift_shift): Likewise.
6282 (if_not_arith): Likewise.
6283 (if_arith_not): Likewise.
6284 (cond_move_not): Likewise.
6285 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6286 (neon_mov<mode>): Likewise.
6287 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6288 (thumb2_movsi_vfp): Likewise.
6289 (movsf_vfp): Likewise.
6290 (thumb2_movsf_vfp): Likewise.
371e77e3
UB		 6291 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6292 change.
859abddd
SN		 6293 (cortexa7_older_only): Likewise.
6294 (cortexa7_younger): Likewise.
6295 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6296 (1020alu_shift_op): Likewise.
6297 (1020alu_shift_reg_op): Likewise.
6298 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6299 (alu_shift_op): Likewise.
6300 (alu_shift_reg_op): Likewise.
6301 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6302 (11_alu_shift_op): Likewise.
6303 (11_alu_shift_reg_op): Likewise.
6304 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6305 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 6306 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6307 change.
859abddd
SN		 6308 (cortex_a15_alu_shift): Likewise.
6309 (cortex_a15_alu_shift_reg): Likewise.
6310 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6311 (cortex_a5_alu_shift): Likewise.
371e77e3
UB		 6312 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6313 change.
859abddd 6314 (cortex_a53_alu_shift): Likewise.
371e77e3
UB		 6315 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6316 change.
859abddd
SN		 6317 (cortex_a7_alu_reg): Likewise.
6318 (cortex_a7_alu_shift): Likewise.
6319 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6320 (cortex_a8_alu_shift): Likewise.
6321 (cortex_a8_alu_shift_reg): Likewise.
6322 (cortex_a8_mov): Likewise.
6323 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6324 (cortex_a9_dp_shift): Likewise.
6325 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6326 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6327 (cortex_r4_mov): Likewise.
6328 (cortex_r4_alu_shift): Likewise.
6329 (cortex_r4_alu_shift_reg): Likewise.
6330 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6331 (526_alu_shift_op): Likewise.
6332 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6333 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6334 (626te_alu_shift_op): Likewise.
6335 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6336 (726te_alu_op): Likewise.
6337 (726te_alu_shift_op): Likewise.
6338 (726te_alu_shift_reg_op): Likewise.
6339 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6340 (mp626_alu_shift_op): Likewise.
6341 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6342 (pj4_alu_e1_conds): Likewise.
6343 (pj4_alu): Likewise.
6344 (pj4_alu_conds): Likewise.
6345 (pj4_shift): Likewise.
6346 (pj4_shift_conds): Likewise.
6347 (pj4_alu_shift): Likewise.
6348 (pj4_alu_shift_conds): Likewise.
6349
95b97fac
KT		 63502013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6351
6352 * config/arm/predicates.md (shiftable_operator_strict_it):
6353 New predicate.
6354 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
6355 Disable cond_exec version for arm_restrict_it.
6356 (thumb2_smaxsi3): Convert to generate cond_exec.
6357 (thumb2_sminsi3): Likewise.
6358 (thumb32_umaxsi3): Likewise.
6359 (thumb2_uminsi3): Likewise.
6360 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
6361 (thumb2_neg_abssi2): Likewise.
6362 (thumb2_mov_scc): Add alternative for 16-bit encoding.
6363 (thumb2_movsicc_insn): Adjust alternatives.
6364 (thumb2_mov_negscc): Disable for arm_restrict_it.
6365 (thumb2_mov_negscc_strict_it): New pattern.
6366 (thumb2_mov_notscc_strict_it): New pattern.
6367 (thumb2_mov_notscc): Disable for arm_restrict_it.
6368 (thumb2_ior_scc): Likewise.
6369 (thumb2_ior_scc_strict_it): New pattern.
6370 (thumb2_cond_move): Adjust for arm_restrict_it.
6371 (thumb2_cond_arith): Disable for arm_restrict_it.
6372 (thumb2_cond_arith_strict_it): New pattern.
6373 (thumb2_cond_sub): Adjust for arm_restrict_it.
6374 (thumb2_movcond): Likewise.
6375 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
6376 (thumb2_zero_extendhisi2_v6): Likewise.
6377 (thumb2_zero_extendqisi2_v6): Likewise.
6378 (orsi_notsi_si): Likewise.
6379 (orsi_not_shiftsi_si): Likewise.
6380
d5a6ef82
GJL		 63812013-07-22 Georg-Johann Lay <avr@gjlay.de>
6382
6383 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
6384 instruction sequence is 1 byte shorter.
6385
d9e71154
UB		 63862013-07-22 Uros Bizjak <ubizjak@gmail.com>
6387
6388 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
6389 it is not needed after split.
6390
8c6381ec
IS		 63912013-07-20 Iain Sandoe <iain@codesourcery.com>
6392
6393 PR target/51784
6394 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
6395 second label for nonlocal goto receivers. Don't output pic base labels
6396 unless we're producing PIC; mark that action unreachable().
6397 (ix86_save_reg): If the function contains a nonlocal label, save the
6398 PIC base reg.
6399 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
371e77e3 6400 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS		 6401 (update_pic_label_number_if_needed): New.
6402 (machopic_output_function_base_name): Adjust for nonlocal receiver
6403 case.
6404 (machopic_should_output_picbase_label): New.
6405 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
6406 (nonlocal_goto_receiver): New insn and split.
6407
d05d0709
JG		 64082013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
6409
6410 * config/aarch64/aarch64-builtins.c
6411 (aarch64_fold_builtin): Fold abs in all modes.
6412 * config/aarch64/aarch64-simd-builtins.def
6413 (abs): Enable for all modes.
6414 * config/aarch64/arm_neon.h
6415 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
6416 (vabs_f64): Add missing intrinsic.
6417
ba590e63
IB		 64182013-07-19 Ian Bolton <ian.bolton@arm.com>
6419
6420 * config/aarch64/arm_neon.h (vabs_s64): New function
6421
02371798
GJL		 64222013-07-19 Georg-Johann Lay <avr@gjlay.de>
6423
6424 PR target/57516
6425 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
6426 * config/avr/avr.md (adjust_len): Add `round'.
6427 * config/avr/avr-protos.h (avr_out_round): New prototype.
6428 (avr_out_plus): Add `out_label' argument.
6429 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
6430 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
6431 Handle the case where `insn' is just a pattern.
6432 (avr_out_bitop): Handle the case where `insn' is just a pattern.
6433 (avr_out_round): New function.
6434 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
6435
12211b99 64362013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH		 6437
6438 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6439 Add check for flag_stack_usage to handle -fstack-usage support
6440
12211b99 64412013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH		 6442
6443 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
6444 interaction for new Power8 flags and VSX.
6445
df9dda2d
ST		 64462013-07-18 Sriraman Tallam <tmsriram@google.com>
6447
6448 PR middle-end/57698
6449 * tree-inline.c (expand_call_inline): Emit errors during
6450 early_inlining only if optimization is not turned on.
6451
0f9cdef4
DM		 64522013-07-18 David Malcolm <dmalcolm@redhat.com>
6453
6454 * passes.def: New.
6455
6456 * passes.c (init_optimization_passes): Move the construction of
6457 the pass hierarchy into a new passes.def file.
6458
6459 * Makefile.in (passes.o): Add dependency on passes.def.
6460
2efa4087
DM		 64612013-07-18 David Malcolm <dmalcolm@redhat.com>
6462
6463 * passes.c (init_optimization_passes): Introduce macros for
6464 constructing the tree of passes (INSERT_PASSES_AFTER,
6465 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
6466 TERMINATE_PASS_LIST).
6467
bc404e1b
VM		 64682013-07-18 Vladimir Makarov <vmakarov@redhat.com>
6469 Wei Mi <wmi@google.com>
6470
6471 PR rtl-optimization/57878
6472 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
6473 top.
6474 (reload_pseudo_compare_func): Check nregs first for reload
6475 pseudos.
6476
d675843f
DM		 64772013-07-18 David Malcolm <dmalcolm@redhat.com>
6478
6479 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
6480
c7b3b99f 64812013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC		 6482
6483 * read-rtl.c (validate_const_int): Once an invalid character is
6484 seen, quit the loop.
6485
6486 * gengtype.c (write_roots): Similarly once we find the "deletable"
6487 or "if_marked" option.
6488
52fceb44
SN		 64892013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6490
6491 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
6492 "xtab" and "sat". Move value "clz" from here to ...
6493 (attriubte "type"): ... here.
6494 (satsi_<SAT:code>): Delete "insn" attribute.
6495 (satsi_<SAT:code>_shift): Likewise.
6496 (arm_zero_extendqisi2addsi): Likewise.
6497 (arm_extendqisi2addsi): Likewise.
6498 (clzsi2): Update for attribute changes.
6499 (rbitsi2): Likewise.
371e77e3
UB		 6500 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
6501 attribute.
52fceb44
SN		 6502 (arm_usatsihi): Likewise.
6503 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6504
006bd006
SN		 65052013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6506
6507 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB		 6508 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
6509 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN		 6510 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
6511 in alphabetical order.
6512 (attribute "core_cycles"): Update for attribute changes.
6513 (arm_addsi3): Likewise.
6514 (addsi3_compare0): Likewise.
6515 (addsi3_compare0_scratch): Likewise.
6516 (addsi3_compare_op1): Likewise.
6517 (addsi3_compare_op2): Likewise.
6518 (compare_addsi2_op0): Likewise.
6519 (compare_addsi2_op1): Likewise.
6520 (addsi3_carryin_shift_<optab>): Likewise.
6521 (subsi3_carryin_shift): Likewise.
6522 (rsbsi3_carryin_shift): Likewise.
6523 (arm_subsi3_insn): Likewise.
6524 (subsi3_compare0): Likewise.
6525 (subsi3_compare): Likewise.
6526 (arm_andsi3_insn): Likewise.
6527 (thumb1_andsi3_insn): Likewise.
6528 (andsi3_compare0): Likewise.
6529 (andsi3_compare0_scratch): Likewise.
6530 (zeroextractsi_compare0_scratch
6531 (andsi_not_shiftsi_si): Likewise.
6532 (iorsi3_insn): Likewise.
6533 (iorsi3_compare0): Likewise.
6534 (iorsi3_compare0_scratch): Likewise.
6535 (arm_xorsi3): Likewise.
6536 (thumb1_xorsi3_insn): Likewise.
6537 (xorsi3_compare0): Likewise.
6538 (xorsi3_compare0_scratch): Likewise.
6539 (satsi_<SAT:code>_shift): Likewise.
6540 (rrx): Likewise.
6541 (arm_shiftsi3): Likewise.
6542 (shiftsi3_compare0): Likewise.
6543 (not_shiftsi): Likewise.
6544 (not_shiftsi_compare0): Likewise.
6545 (not_shiftsi_compare0_scratch): Likewise.
6546 (arm_one_cmplsi2): Likewise.
6547 (thumb_one_complsi2): Likewise.
6548 (notsi_compare0): Likewise.
6549 (notsi_compare0_scratch): Likewise.
6550 (thumb1_zero_extendhisi2): Likewise.
6551 (arm_zero_extendhisi2): Likewise.
6552 (arm_zero_extendhisi2_v6): Likewise.
6553 (arm_zero_extendhisi2addsi): Likewise.
6554 (thumb1_zero_extendqisi2): Likewise.
6555 (thumb1_zero_extendqisi2_v6): Likewise.
6556 (arm_zero_extendqisi2): Likewise.
6557 (arm_zero_extendqisi2_v6): Likewise.
6558 (arm_zero_extendqisi2addsi): Likewise.
6559 (thumb1_extendhisi2): Likewise.
6560 (arm_extendhisi2): Likewise.
6561 (arm_extendhisi2_v6): Likewise.
6562 (arm_extendqisi): Likewise.
6563 (arm_extendqisi_v6): Likewise.
6564 (arm_extendqisi2addsi): Likewise.
6565 (thumb1_extendqisi2): Likewise.
6566 (thumb1_movdi_insn): Likewise.
6567 (arm_movsi_insn): Likewise.
6568 (movsi_compare0): Likewise.
6569 (movhi_insn_arch4): Likewise.
6570 (movhi_bytes): Likewise.
6571 (arm_movqi_insn): Likewise.
6572 (thumb1_movqi_insn): Likewise.
6573 (arm32_movhf): Likewise.
6574 (thumb1_movhf): Likewise.
6575 (arm_movsf_soft_insn): Likewise.
6576 (thumb1_movsf_insn): Likewise.
6577 (movdf_soft_insn): Likewise.
6578 (thumb_movdf_insn): Likewise.
6579 (arm_cmpsi_insn): Likewise.
6580 (cmpsi_shiftsi): Likewise.
6581 (cmpsi_shiftsi_swp): Likewise.
6582 (arm_cmpsi_negshiftsi_si): Likewise.
6583 (movsicc_insn): Likewise.
6584 (movsfcc_soft_insn): Likewise.
6585 (arith_shiftsi): Likewise.
6586 (arith_shiftsi_compare0
6587 (arith_shiftsi_compare0_scratch
6588 (sub_shiftsi): Likewise.
6589 (sub_shiftsi_compare0
6590 (sub_shiftsi_compare0_scratch
6591 (and_scc): Likewise.
6592 (cond_move): Likewise.
6593 (if_plus_move): Likewise.
6594 (if_move_plus): Likewise.
6595 (if_move_not): Likewise.
6596 (if_not_move): Likewise.
6597 (if_shift_move): Likewise.
6598 (if_move_shift): Likewise.
6599 (if_shift_shift): Likewise.
6600 (if_not_arith): Likewise.
6601 (if_arith_not): Likewise.
6602 (cond_move_not): Likewise.
6603 (thumb1_ashlsi3): Set type attribute.
6604 (thumb1_ashrsi3): Likewise.
6605 (thumb1_lshrsi3): Likewise.
6606 (thumb1_rotrsi3): Likewise.
6607 (shiftsi3_compare0_scratch): Likewise.
6608 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
6609 (neon_mov<mode>): Likewise.
371e77e3
UB		 6610 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
6611 attribute changes.
006bd006
SN		 6612 (thumb2_movsi_insn): Likewise.
6613 (thumb2_cmpsi_neg_shiftsi): Likewise.
6614 (thumb2_extendqisi_v6): Likewise.
6615 (thumb2_zero_extendhisi2_v6): Likewise.
6616 (thumb2_zero_extendqisi2_v6): Likewise.
6617 (thumb2_shiftsi3_short): Likewise.
6618 (thumb2_addsi3_compare0_scratch): Likewise.
6619 (orsi_not_shiftsi_si): Likewise.
6620 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6621 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
6622 changes.
6623 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6624 (1020alu_shift_op): Likewise.
6625 (1020alu_shift_reg_op): Likewise.
6626 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6627 (alu_shift_op): Likewise.
6628 (alu_shift_reg_op): Likewise.
6629 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6630 (11_alu_shift_op): Likewise.
6631 (11_alu_shift_reg_op): Likewise.
6632 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6633 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 6634 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6635 changes.
006bd006
SN		 6636 (cortex_a15_alu_shift): Likewise.
6637 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB		 6638 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
6639 changes.
006bd006
SN		 6640 (cortex_a5_alu_shift): Likewise.
6641 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
6642 changes.
6643 (cortex_a53_alu_shift): Likewise.
6644 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6645 changes.
6646 (cortex_a7_alu_reg): Likewise.
6647 (cortex_a7_alu_shift): Likewise.
371e77e3
UB		 6648 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
6649 changes.
006bd006
SN		 6650 (cortex_a8_alu_shift): Likewise.
6651 (cortex_a8_alu_shift_reg): Likewise.
6652 (cortex_a8_mov): Likewise.
6653 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6654 (cortex_a9_dp_shift): Likewise.
371e77e3
UB		 6655 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
6656 changes.
6657 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
6658 changes.
006bd006
SN		 6659 (cortex_r4_mov): Likewise.
6660 (cortex_r4_alu_shift): Likewise.
6661 (cortex_r4_alu_shift_reg): Likewise.
6662 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6663 (526_alu_shift_op): Likewise.
6664 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6665 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6666 (626te_alu_shift_op): Likewise.
6667 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
6668 (726te_alu_op): Likewise.
6669 (726te_alu_shift_op): Likewise.
6670 (726te_alu_shift_reg_op): Likewise.
6671 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6672 (mp626_alu_shift_op): Likewise.
6673 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
6674 (pj4_alu_e1_conds): Likewise.
6675 (pj4_alu): Likewise.
6676 (pj4_alu_conds): Likewise.
6677 (pj4_shift): Likewise.
6678 (pj4_shift_conds): Likewise.
6679 (pj4_alu_shift): Likewise.
6680 (pj4_alu_shift_conds): Likewise.
371e77e3
UB		 6681 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6682 changes.
006bd006
SN		 6683 (cortexa7_older_only): Likewise.
6684 (cortexa7_younger): Likewise.
6685
651df1b2
DM		 66862013-07-18 David Malcolm <dmalcolm@redhat.com>
6687
6688 * ipa-pure-const.c (generate_summary): Rename to...
6689 (pure_const_generate_summary): ... this.
6690
ea717bd7
IS		 66912013-07-17 Iain Sandoe <iain@codesourcery.com>
6692
6693 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
6694
a1259a13
YR		 66952013-07-17 Yvan Roux <yvan.roux@linaro.org>
6696
6697 PR target/57909
6698 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
6699 usage in HI mode.
6700
167f68ed
AK		 67012013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6702
6703 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
6704 enabled without -march=zEC12.
6705 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
6706 flags to be set.
6707
f900a982
MR		 67082013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6709
6710 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
6711 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
6712 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
6713 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
6714 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
6715 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
6716 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
6717 ISA_HAS_FP4.
6718 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
6719 and ISA_HAS_NMADD3_NMSUB3.
6720 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
6721 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
6722 (nmsub4<mode>, nmsub3<mode>): Likewise.
6723 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
6724
d4ed27eb
MR		 67252013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6726
6727 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
6728 TARGET_MIPS5400 checking.
6729
7a5add18
PB		 67302013-07-16 Jakub Jelinek <jakub@redhat.com>
6731 Peter Bergner <bergner@vnet.ibm.com>
6732
6733 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
6734 registers in the comment.
6735 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
6736 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
6737 rather than FIRST_PSEUDO_REGISTERS.
6738
67392013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB		 6740
6741 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
6742 enable extra ISA flags with TARGET_HTM.
6743
d2ab0929
MR		 67442013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6745
6746 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
6747 Fix comment typos.
6748
839c74bc
CH		 67492013-07-15 Cong Hou <congh@google.com>
6750
6751 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
6752 in compare function for sorting.
6753
7a5add18 67542013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB		 6755
6756 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
6757 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
6758 * config/rs6000/rs6000.opt: Add -mhtm option.
6759 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
6760 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
6761 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6762 __HTM__ if the HTM instructions are available.
6763 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
6764 htm_spr_reg_operand): New define_predicates.
6765 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
6766 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
6767 Include htm.md.
6768 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
6769 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
6770 HTM builtin functions.
6771 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
6772 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
6773 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
6774 (rs6000_builtin_mask_calculate): Likewise.
6775 (rs6000_option_override_internal): Likewise.
6776 (bdesc_htm): Add new HTM builtin support.
6777 (htm_spr_num): New function.
6778 (htm_spr_regno): Likewise.
6779 (rs6000_htm_spr_icode): Likewise.
6780 (htm_expand_builtin): Likewise.
6781 (htm_init_builtins): Likewise.
6782 (rs6000_expand_builtin): Add support for HTM builtin functions.
6783 (rs6000_init_builtins): Likewise.
371e77e3
UB		 6784 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
6785 option.
0258b6e4
PB		 6786 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
6787 (TARGET_HTM, MASK_HTM): Define macros.
6788 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
6789 (FIXED_REGISTERS): Likewise.
6790 (CALL_USED_REGISTERS): Likewise.
6791 (CALL_REALLY_USED_REGISTERS): Likewise.
6792 (REG_ALLOC_ORDER): Likewise.
6793 (enum reg_class): Likewise.
6794 (REG_CLASS_NAMES): Likewise.
6795 (REG_CLASS_CONTENTS): Likewise.
6796 (REGISTER_NAMES): Likewise.
6797 (ADDITIONAL_REGISTER_NAMES): Likewise.
6798 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
6799 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
6800 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
6801 * config/rs6000/htm.md: New file.
6802 * config/rs6000/htmintrin.h: New file.
6803 * config/rs6000/htmxlintrin.h: New file.
6804
87dd8ab0
MS		 68052013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
6806
6807 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
6808 Define SYMBOL_TINY_GOT, update comment.
6809 * config/aarch64/aarch64.c
6810 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
6811 (aarch64_expand_mov_immediate): Likewise.
6812 (aarch64_print_operand): Likewise.
6813 (aarch64_classify_symbol): Likewise.
6814 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
6815 (ldr_got_tiny): Define.
6816
52d676b6
TG		 68172013-07-13 Tobias Grosser <tobias@grosser.es>
6818
6819 PR tree-optimization/54094
6820 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
6821 scheduling dimension for the parallelism check from the polyhedral
6822 information in the AST.
6823 * graphite-dependences.c (carries_deps): Do not assume the schedule is
6824 in 2D + 1 form.
6825
286e8fc1
JM		 68262013-07-13 Jason Merrill <jason@redhat.com>
6827
6828 * print-tree.c (debug_vec_tree): Use debug_raw.
6829 (debug_raw (vec<tree, va_gc> &)): New.
6830 (debug_raw (vec<tree, va_gc> *)): New.
6831 * tree.h: Declare them.
6832
e43257e8
BC		 68332013-07-13 Bin Cheng <bin.cheng@arm.com>
6834
6835 * ifcvt.c (ifcvt_after_combine): New static variable.
6836 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
6837 for size.
6838 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
6839 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
6840 rest_of_handle_if_after_reload): Pass new argument for if_convert.
6841
da582d46
MR		 68422013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
6843
6844 * config/mips/mips.c (mips_expand_call): Remove empty statement.
6845
18b0ea8f
MM		 68462013-07-12 Michael Matz <matz@suse.de>
6847
6848 PR middle-end/55771
6849 * convert.c (convert_to_real): Reject non-float inner types.
6850
e4f0f84d
TB		 68512013-07-12 Tejas Belagod <tejas.belagod@arm.com>
6852
6853 * config/aarch64/aarch64-protos.h
6854 (aarch64_simd_immediate_valid_for_move): Remove.
6855 * config/aarch64/aarch64.c (simd_immediate_info): New member.
6856 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
6857 cases.
6858 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
6859
76500022
SE		 68602013-07-11 Steve Ellcey <sellcey@mips.com>
6861
6862 * config/mips/mips.c (mips_conditional_register_usage): Do not
6863 use t[0-7] registers in MIPS16 mode when optimizing for size.
6864
78039734
ST		 68652013-07-11 Sriraman Tallam <tmsriram@google.com>
6866
6867 * config/i386/i386.c (dispatch_function_versions): Fix array
6868 indexing of function_version_info to match actual_versions.
6869
7b24b675
TJ		 68702013-07-11 Teresa Johnson <tejohnson@google.com>
6871
6872 * vec.h (struct va_gc): Move release out-of-line.
6873 (va_gc::release): Call ggc_free on released vec.
6874
85d4cbb8
UW		 68752013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6876
6877 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
6878 Require GOT register as additional operand in UNSPEC.
6879 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
6880 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6881 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6882 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
6883 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
6884 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
6885 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
6886
52b3f9e2
GJL		 68872013-07-11 Georg-Johann Lay <avr@gjlay.de>
6888
6889 PR target/57631
6890 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
6891 name seen by assembler/linker rather if available.
6892
3f3e266e
AS		 68932013-07-11 Andreas Schwab <schwab@suse.de>
6894
6895 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
6896
21231ca6
VM		 68972013-07-10 Vladimir Makarov <vmakarov@redhat.com>
6898
371e77e3 6899 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 6900
d331d374
JM		 69012013-07-10 Joseph Myers <joseph@codesourcery.com>
6902
673c2f63
JM		 6903 * doc/tm.texi.in: Move hook documentation to ....
6904 * target.def: ... here.
6905
d331d374
JM		 6906 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
6907 text on @hook line.
6908 * doc/tm.texi: Regenerate.
6909
dfeadaa0
PC		 69102013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
6911
6912 PR c++/57869
6913 * doc/invoke.texi: Document -Wconditionally-supported.
6914
bb6f2bac
GJL		 69152013-07-10 Georg-Johann Lay <avr@gjlay.de>
6916
6917 PR target/57844
6918 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
6919 of my_fp.
6920
5219b232
GJL		 69212013-07-10 Georg-Johann Lay <avr@gjlay.de>
6922
6923 PR target/57506
6924 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
6925 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
6926 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
6927 Remove duplicate devices.
6928 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
6929 * config/avr/t-multilib: Regenerate.
6930 * config/avr/avr-tables.opt: Regenerate.
6931 * doc/avr-mmcu.texi: Regenerate.
6932
283045c0
GJL		 69332013-07-10 Georg-Johann Lay <avr@gjlay.de>
6934
6935 PR target/56987
6936 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
6937
ceb2d59f 69382013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 6939
ceb2d59f
GS		 6940 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
6941 the cost of MULT when optimizing for size.
6942
055e0a99
JBG		 69432013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6944
6945 * config/cr16/cr16-protos.h: Don't include target.h.
6946
ef43e868
JM		 69472013-07-09 Joseph Myers <joseph@codesourcery.com>
6948
6949 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
6950 adjust register size for TDmode and TFmode for VSX registers.
6951
4a283fdf
KT		 69522013-07-08 Kai Tietz <ktietz@redhat.com>
6953
6954 PR target/56892
6955 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
6956 hook_bool_const_tree_true.
6957
2cf4c39e
AK		 69582013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6959
6960 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
6961 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
6962 * config/s390/s390.md: Define FPR*_REGNUM constants.
6963 Fix FPR2_REGNUM constant (18 -> 17).
6964 ("*trunc<BFP:mode><DFP_ALL:mode>2")
6965 ("*trunc<DFP_ALL:mode><BFP:mode>2")
6966 ("trunc<BFP:mode><DFP_ALL:mode>2")
6967 ("trunc<DFP_ALL:mode><BFP:mode>2")
6968 ("*extend<BFP:mode><DFP_ALL:mode>2")
6969 ("*extend<DFP_ALL:mode><BFP:mode>2")
6970 ("extend<BFP:mode><DFP_ALL:mode>2")
6971 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
6972 FPR4_REGNUM.
6973
e6ac0270 69742013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 6975
6a2fa4b2 6976 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 6977
b89b22fc
AK		 69782013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6979
6980 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
6981 and cfun_fpr_bit_p to cfun_fpr_save_p.
6982 (s390_frame_area, s390_register_info, s390_frame_info)
6983 (s390_emit_prologue, s390_emit_epilogue)
6984 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
6985 register numbers.
6986 * config/s390/s390.h: Define *_REGNUM macros for floating point
6987 register numbers.
6988
07f398aa
EB		 69892013-07-08 Eric Botcazou <ebotcazou@adacore.com>
6990
6991 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
6992
75fefa91
PCC		 69932013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
6994
6995 PR rtl-optimization/57786
6996 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
6997 and break out of the loop when it is set to false.
6998
046f1eee
JJ		 69992013-07-08 Jakub Jelinek <jakub@redhat.com>
7000
8140c065
JJ		 7001 PR target/57819
7002 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7003 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7004 (const_int 63)) 0)).
7005 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7006 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7007 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7008
046f1eee
JJ		 7009 PR rtl-optimization/57829
7010 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7011 mask bits outside of mode are just sign-extension from mode to HWI.
7012
930b700b
MZ		 70132013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7014
7015 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7016 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7017 adjust_address instead of change_address to keep info about alignment.
7018 (emit_strmov): Remove.
7019 (emit_memmov): New function.
7020 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7021 (expand_movmem_epilogue): Likewise and return updated rtx for
7022 destination.
7023 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7024 destination and source.
7025 (decide_alignment): Refactor, handle vector_loop.
7026 (ix86_expand_movmem): Likewise.
7027 (ix86_expand_setmem): Likewise.
7028 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7029
7b9d1bd8
UB		 70302013-07-07 Uros Bizjak <ubizjak@gmail.com>
7031
7032 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7033 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7034
eabb5f48
UB		 70352013-07-06 Uros Bizjak <ubizjak@gmail.com>
7036
7037 * config/i386/sse.md (sse_movlhps): Change alternative 3
7038 of operand 2 to "m".
7039
70402013-07-06 Uros Bizjak <ubizjak@gmail.com>
7041
7042 PR target/57807
7043 * config/i386/sse.md (iptr): New mode attribute.
7044 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7045 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7046 (<sse>_vmmul<mode>3): Ditto.
7047 (<sse>_vmdiv<mode>3): Ditto.
7048 (sse_vmrcpv4sf2): Ditto.
7049 (<sse>_vmsqrt<mode>2): Ditto.
7050 (sse_vmrsqrtv4sf2): Ditto.
7051 (<sse>_vm<code><mode>3): Ditto.
7052 (avx_vmcmp<mode>3): Ditto.
7053 (<sse>_vmmaskcmp<mode>3): Ditto.
7054 (<sse>_comi): Ditto.
7055 (<sse>_ucomi): Ditto.
7056 (*xop_vmfrcz_<mode>): Ditto.
7057 (*fmai_fmadd_<mode>): Ditto.
7058 (*fmai_fmsub_<mode>): Ditto.
7059 (*fmai_fnmadd_<mode>): Ditto.
7060 (*fmai_fnmsub_<mode>): Ditto.
7061 (*fma4i_vmfmadd_<mode>): Ditto.
7062 (*fma4i_vmfmsub_<mode>): Ditto.
7063 (*fma4i_vmfnmadd_<mode>): Ditto.
7064 (*fma4i_vmfnmsub_<mode>): Ditto.
7065 (*xop_vmfrcz_<mode>): Ditto.
7066 (sse_cvtps2pi): Ditto.
7067 (sse_cvttps2pi): Ditto.
7068 (sse_cvtss2si): Ditto.
7069 (sse_cvtss2si_2): Ditto.
7070 (sse_cvtss2siq_2): Ditto.
7071 (sse_cvttss2si): Ditto.
7072 (sse_cvttss2siq): Ditto.
7073 (sse_cvtsd2si): Ditto.
7074 (sse_cvtsd2si_2): Ditto.
7075 (sse_cvtsd2siq_2): Ditto.
7076 (sse_cvttsd2si): Ditto.
7077 (sse_cvttsd2siq): Ditto.
7078 (sse_cvtsd2ss): Ditto.
7079 (sse_cvtss2sd): Ditto.
7080 (avx2_pbroadcast<mode>): Ditto.
7081 (avx2_pbroadcast<mode>_1): Ditto.
7082 (*avx_vperm_broadcast_v4sf): Ditto.
7083
7084 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7085 (sse_movlhps): Ditto.
7086 (sse_storehps): Ditto.
7087 (sse_loadhps): Ditto.
7088 (sse_storelps): Ditto.
7089 (sse_loadlps): Ditto.
7090 (*vec_concatv4sf): Ditto.
7091 (*vec_interleave_highv2df): Ditto.
7092 (*vec_interleave_lowv2df): Ditto.
7093 (*vec_extractv2df_1_sse): Ditto.
7094 (*vec_extractv2df_0_sse): Ditto.
7095 (sse2_storelpd): Ditto.
7096 (sse2_loadlpd): Ditto.
7097 (sse2_movsd): Ditto.
7098 (*vec_concatv4si): Ditto.
7099 (vec_concatv2di): Ditto.
7100
7101 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
7102 for Intel asm dialect.
7103 (mmx_punpcklwd): Ditto.
7104 (mmx_punpckldq): Ditto.
7105
7106 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
7107 for intel assembler dialect.
7108
1f6eac90
JJ		 71092013-07-06 Jakub Jelinek <jakub@redhat.com>
7110
7111 PR target/29776
7112 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
7113 for BUILT_IN_C{LZ,LRSB}*.
7114 * tree.h (CASE_INT_FN): Add FN##IMAX case.
7115 * tree-vrp.c (extract_range_basic): Handle
7116 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
7117 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
7118 fall thru to code calling set_value*.
7119 * builtins.c (expand_builtin): Remove *IMAX cases.
7120 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
7121 if width is bigger than 2*HWI.
7122
2b778c9d
VM		 71232013-07-05 Vladimir Makarov <vmakarov@redhat.com>
7124
7125 PR rtl-optimization/55342
7126 * lra-int.h (lra_subreg_reload_pseudos): New.
7127 * lra.c: Add undoing optional reloads to the block diagram.
7128 (lra_subreg_reload_pseudos): New.
7129 (lra_optional_reload_pseudos): Change comments.
7130 (lra): Init and clear lra_subreg_reload_pseudos. Clear
7131 lra_optional_reload_pseudos after undo transformations.
7132 * lra-assigns.c (pseudo_prefix_title): New.
7133 (lra_setup_reg_renumber): Use it.
7134 (spill_for): Ditto. Check subreg reload pseudos too.
7135 (assign_by_spills): Consider subreg reload pseudos too.
7136 * lra-constraints.c (simplify_operand_subreg): Use
7137 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
7138 (curr_insn_transform): Recognize and do optional reloads.
7139 (undo_optional_reloads): New.
7140 (lra_undo_inheritance): Call undo_optional_reloads.
7141
207156e4
TQ		 71422013-07-05 Thomas Quinot <quinot@adacore.com>
7143
7144 * tree-complex.c (expand_complex_operations_1): Fix typo.
7145
8990e73a
TB		 71462013-07-04 Tejas Belagod <tejas.belagod@arm.com>
7147
7148 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
7149 (tune_params): New member 'const vec_costs'.
7150 * config/aarch64/aarch64.c (generic_vector_cost): New.
7151 (generic_tunings): New member 'generic_vector_cost'.
7152 (aarch64_builtin_vectorization_cost): New.
7153 (aarch64_add_stmt_cost): New.
7154 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7155 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7156
f7bc421d
JJ		 71572013-07-03 Jakub Jelinek <jakub@redhat.com>
7158
7159 PR target/57777
7160 * config/i386/predicates.md (vsib_address_operand): Disallow
7161 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7162
e90d1568
HPN		 71632013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7164
7165 PR middle-end/55030
7166 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7167 expand_builtin_setjmp_receiver.
7168 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7169 with NULL for the label parameter.
7170 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7171 the frame-pointer. Adjust comments.
7172 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7173 only if LABEL is non-NULL.
7174
cbd7413b
YZ		 71752013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7176
7177 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7178 (ARM_ABI_AAPCS64): Ditto.
7179 (arm_abi): Ditto.
7180 (ARM_DEFAULT_ABI): Ditto.
7181
dec11868
JG		 71822013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7183
7184 * config/aarch64/aarch64-builtins.c
7185 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7186 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7187 (st1): Likewise.
7188 * config/aarch64/aarch64-simd.md
7189 (aarch64_ld1<VALL:mode>): New.
7190 (aarch64_st1<VALL:mode>): Likewise.
7191 * config/aarch64/arm_neon.h
7192 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7193
786fb9b6
ST		 71942013-07-02 Sriraman Tallam <tmsriram@google.com>
7195
0ba6aff3
UB		 7196 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7197 target ISA is AVX.
786fb9b6
ST		 7198 (ix86_option_override_internal):Turn on all -mavx target flags by
7199 default as they are dependent on AVX anyway.
7200
4b043553
CC		 72012013-07-02 Cary Coutant <ccoutant@google.com>
7202
7203 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7204 deterministic hash.
7205 (loc_checksum_ordered): Likewise.
7206 (hash_loc_operands): Remove inline keyword.
7207
c4e87a13
JJ		 72082013-07-02 Jakub Jelinek <jakub@redhat.com>
7209
7210 PR tree-optimization/57741
7211 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7212 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7213 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7214 Allow REAL_CST step_exprs if flag_associative_math.
7215 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7216
26366d28
IB		 72172013-07-02 Ian Bolton <ian.bolton@arm.com>
7218
0ba6aff3 7219 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 7220
2879bb2b
IB		 72212013-07-02 Ian Bolton <ian.bolton@arm.com>
7222
7223 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7224
afc5e8a6
KT		 72252013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7226
7227 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7228 encoding.
7229 (iorsi3_insn): Likewise.
7230 (arm_xorsi3): Likewise.
7231
31f8442b
SN		 72322013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7233
7234 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7235 to ...
7236 (attribute "type"): ... here, and prefix with "wmmx_".
7237 (attribute "core_cycles"): Update for attribute changes.
7238 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7239 (tbcstv4hi): Likewise.
7240 (tbcstv2si): Likewise.
7241 (iwmmxt_iordi3): Likewise.
7242 (iwmmxt_xordi3): Likewise.
7243 (iwmmxt_anddi3): Likewise.
7244 (iwmmxt_nanddi3): Likewise.
7245 (iwmmxt_arm_movdi): Likewise.
7246 (iwmmxt_movsi_insn): Likewise.
7247 (mov<mode>_internal): Likewise.
7248 (and<mode>3_iwmmxt): Likewise.
7249 (ior<mode>3_iwmmxt): Likewise.
7250 (xor<mode>3_iwmmxt): Likewise.
7251 (add<mode>3_iwmmxt): Likewise.
7252 (ssaddv8qi3): Likewise.
7253 (ssaddv4hi3): Likewise.
7254 (ssaddv2si3): Likewise.
7255 (usaddv8qi3): Likewise.
7256 (usaddv4hi3): Likewise.
7257 (usaddv2si3): Likewise.
7258 (sub<mode>3_iwmmxt): Likewise.
7259 (sssubv8qi3): Likewise.
7260 (sssubv4hi3): Likewise.
7261 (sssubv2si3): Likewise.
7262 (ussubv8qi3): Likewise.
7263 (ussubv4hi3): Likewise.
7264 (ussubv2si3): Likewise.
7265 (mulv4hi3_iwmmxt): Likewise.
7266 (smulv4hi3_highpart): Likewise.
7267 (umulv4hi3_highpart): Likewise.
7268 (iwmmxt_wmacs): Likewise.
7269 (iwmmxt_wmacsz): Likewise.
7270 (iwmmxt_wmacu): Likewise.
7271 (iwmmxt_wmacuz): Likewise.
7272 (iwmmxt_clrdi): Likewise.
7273 (iwmmxt_clrv8qi): Likewise.
7274 (iwmmxt_clr4hi): Likewise.
7275 (iwmmxt_clr2si): Likewise.
7276 (iwmmxt_uavgrndv8qi3): Likewise.
7277 (iwmmxt_uavgrndv4hi3): Likewise.
7278 (iwmmxt_uavgv8qi3): Likewise.
7279 (iwmmxt_uavgv4hi3): Likewise.
7280 (iwmmxt_tinsrb): Likewise.
7281 (iwmmxt_tinsrh): Likewise.
7282 (iwmmxt_tinsrw): Likewise.
7283 (iwmmxt_textrmub): Likewise.
7284 (iwmmxt_textrmsb): Likewise.
7285 (iwmmxt_textrmuh): Likewise.
7286 (iwmmxt_textrmsh): Likewise.
7287 (iwmmxt_textrmw): Likewise.
7288 (iwmxxt_wshufh): Likewise.
7289 (eqv8qi3): Likewise.
7290 (eqv4hi3): Likewise.
7291 (eqv2si3): Likewise.
7292 (gtuv8qi3): Likewise.
7293 (gtuv4hi3): Likewise.
7294 (gtuv2si3): Likewise.
7295 (gtv8qi3): Likewise.
7296 (gtv4hi3): Likewise.
7297 (gtv2si3): Likewise.
7298 (smax<mode>3_iwmmxt): Likewise.
7299 (umax<mode>3_iwmmxt): Likewise.
7300 (smin<mode>3_iwmmxt): Likewise.
7301 (umin<mode>3_iwmmxt): Likewise.
7302 (iwmmxt_wpackhss): Likewise.
7303 (iwmmxt_wpackwss): Likewise.
7304 (iwmmxt_wpackdss): Likewise.
7305 (iwmmxt_wpackhus): Likewise.
7306 (iwmmxt_wpackwus): Likewise.
7307 (iwmmxt_wpackdus): Likewise.
7308 (iwmmxt_wunpckihb): Likewise.
7309 (iwmmxt_wunpckihh): Likewise.
7310 (iwmmxt_wunpckihw): Likewise.
7311 (iwmmxt_wunpckilb): Likewise.
7312 (iwmmxt_wunpckilh): Likewise.
7313 (iwmmxt_wunpckilw): Likewise.
7314 (iwmmxt_wunpckehub): Likewise.
7315 (iwmmxt_wunpckehuh): Likewise.
7316 (iwmmxt_wunpckehuw): Likewise.
7317 (iwmmxt_wunpckehsb): Likewise.
7318 (iwmmxt_wunpckehsh): Likewise.
7319 (iwmmxt_wunpckehsw): Likewise.
7320 (iwmmxt_wunpckelub): Likewise.
7321 (iwmmxt_wunpckeluh): Likewise.
7322 (iwmmxt_wunpckeluw): Likewise.
7323 (iwmmxt_wunpckelsb): Likewise.
7324 (iwmmxt_wunpckelsh): Likewise.
7325 (iwmmxt_wunpckelsw): Likewise.
7326 (ror<mode>3): Likewise.
7327 (ashr<mode>3_iwmmxt): Likewise.
7328 (lshr<mode>3_iwmmxt): Likewise.
7329 (ashl<mode>3_iwmmxt): Likewise.
7330 (ror<mode>3_di): Likewise.
7331 (ashr<mode>3_di): Likewise.
7332 (lshr<mode>3_di): Likewise.
7333 (ashl<mode>3_di): Likewise.
7334 (iwmmxt_wmadds): Likewise.
7335 (iwmmxt_wmaddu): Likewise.
7336 (iwmmxt_tmia): Likewise.
7337 (iwmmxt_tmiaph): Likewise.
7338 (iwmmxt_tmiabb): Likewise.
7339 (iwmmxt_tmiatb): Likewise.
7340 (iwmmxt_tmiabt): Likewise.
7341 (iwmmxt_tmiatt): Likewise.
7342 (iwmmxt_tmovmskb): Likewise.
7343 (iwmmxt_tmovmskh): Likewise.
7344 (iwmmxt_tmovmskw): Likewise.
7345 (iwmmxt_waccb): Likewise.
7346 (iwmmxt_wacch): Likewise.
7347 (iwmmxt_waccw): Likewise.
7348 (iwmmxt_waligni): Likewise.
7349 (iwmmxt_walignr): Likewise.
7350 (iwmmxt_walignr0): Likewise.
7351 (iwmmxt_walignr1): Likewise.
7352 (iwmmxt_walignr2): Likewise.
7353 (iwmmxt_walignr3): Likewise.
7354 (iwmmxt_wsadb): Likewise.
7355 (iwmmxt_wsadh): Likewise.
7356 (iwmmxt_wsadbz): Likewise.
7357 (iwmmxt_wsadhz): Likewise.
7358 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
7359 (iwmmxt_wabsdiffb): Likewise.
7360 (iwmmxt_wabsdiffh): Likewise.
7361 (iwmmxt_wabsdiffw): Likewise.
7362 (iwmmxt_waddsubhx): Likewise
7363 (iwmmxt_wsubaddhx): Likewise.
7364 (addc<mode>3): Likewise.
7365 (iwmmxt_avg4): Likewise.
7366 (iwmmxt_avg4r): Likewise.
7367 (iwmmxt_wmaddsx): Likewise.
7368 (iwmmxt_wmaddux): Likewise.
7369 (iwmmxt_wmaddsn): Likewise.
7370 (iwmmxt_wmaddun): Likewise.
7371 (iwmmxt_wmulwsm): Likewise.
7372 (iwmmxt_wmulwum): Likewise.
7373 (iwmmxt_wmulsmr): Likewise.
7374 (iwmmxt_wmulumr): Likewise.
7375 (iwmmxt_wmulwsmr): Likewise.
7376 (iwmmxt_wmulwumr): Likewise.
7377 (iwmmxt_wmulwl): Likewise.
7378 (iwmmxt_wqmulm): Likewise.
7379 (iwmmxt_wqmulwm): Likewise.
7380 (iwmmxt_wqmulmr): Likewise.
7381 (iwmmxt_wqmulwmr): Likewise.
7382 (iwmmxt_waddbhusm): Likewise.
7383 (iwmmxt_waddbhusl): Likewise.
7384 (iwmmxt_wqmiabb): Likewise.
7385 (iwmmxt_wqmiabt): Likewise.
7386 (iwmmxt_wqmiatb): Likewise.
7387 (iwmmxt_wqmiatt): Likewise.
7388 (iwmmxt_wqmiabbn): Likewise.
7389 (iwmmxt_wqmiabtn): Likewise.
7390 (iwmmxt_wqmiatbn): Likewise.
7391 (iwmmxt_wqmiattn): Likewise.
7392 (iwmmxt_wmiabb): Likewise.
7393 (iwmmxt_wmiabt): Likewise.
7394 (iwmmxt_wmiatb): Likewise.
7395 (iwmmxt_wmiatt): Likewise.
7396 (iwmmxt_wmiabbn): Likewise.
7397 (iwmmxt_wmiabtn): Likewise.
7398 (iwmmxt_wmiatbn): Likewise.
7399 (iwmmxt_wmiattn): Likewise.
7400 (iwmmxt_wmiawbb): Likewise.
7401 (iwmmxt_wmiawbt): Likewise.
7402 (iwmmxt_wmiawtb): Likewise.
7403 (iwmmxt_wmiawtt): Likewise.
7404 (iwmmxt_wmiawbbn): Likewise.
7405 (iwmmxt_wmiawbtn): Likewise.
7406 (iwmmxt_wmiawtbn): Likewise.
7407 (iwmmxt_wmiawttn): Likewise.
7408 (iwmmxt_wmerge): Likewise.
7409 (iwmmxt_tandc<mode>3): Likewise.
7410 (iwmmxt_torc<mode>3): Likewise.
7411 (iwmmxt_torvsc<mode>3): Likewise.
7412 (iwmmxt_textrc<mode>3): Likewise.
7413 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
7414 (wmmxt_pack): Likewise.
7415 (wmmxt_mult_c1): Likewise.
7416 (wmmxt_mult_c2): Likewise.
7417 (wmmxt_alu_c1): Likewise.
7418 (wmmxt_alu_c2): Likewise.
7419 (wmmxt_alu_c3): Likewise.
7420 (wmmxt_transfer_c1): Likewise.
7421 (wmmxt_transfer_c2): Likewise.
7422 (wmmxt_transfer_c3): Likewise.
7423 (marvell_f_iwmmxt_wstr): Likewise.
7424 (marvell_f_iwmmxt_wldr): Likewise.
7425
bdb7bf8a
YZ		 74262013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
7427
0ba6aff3 7428 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 7429
e6631d38
VM		 74302013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7431
7432 Revert:
7433 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7434 * lra-constraints.c (need_for_split_p): Check call used hard regs
7435 living through calls.
7436
7437 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
7438 call used regs for call insn.
7439
53c1275b
JJ		 74402013-06-28 Jakub Jelinek <jakub@redhat.com>
7441
7442 PR target/57736
0ba6aff3
UB		 7443 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
7444 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 7445
1141ed3f
BI		 74462013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7447
7448 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 7449
e32e4c4a
VM		 74502013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7451
7452 * lra-constraints.c (need_for_split_p): Check call used hard regs
7453 living through calls.
7454
f161bfd3
MM		 74552013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
7456
7457 PR target/57744
7458 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
7459 to tie with any other modes. Eliminate Altivec vector mode tests,
7460 since these are a subset of ALTIVEC or VSX vector modes. Simplify
7461 code, to return 0 if testing MODE2 for a condition, if we've
7462 already tested MODE1 for the same condition.
7463
7eda14e1
MS		 74642013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7465
7466 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
7467 layout.
7468
c822f852
MS		 74692013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7470
7471 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7472 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
7473
da4f13a4
MS		 74742013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7475
0ba6aff3
UB		 7476 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
7477 Define.
da4f13a4
MS		 7478 (aarch64_symbolic_constant_p): Remove.
7479 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
7480 static. Fix line length and white space.
7481 (aarch64_symbolic_constant_p): Remove.
7482 * config/aarch64/predicates.md (aarch64_valid_symref):
7483 Use aarch64_classify_symbol_expression.
7484
956a95a5
KT		 74852013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7486
7487 * config/arm/constraints.md (Ts): New constraint.
7488 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
7489 16-bit encodings.
7490 (compare_scc): Use "Ts" constraint for operand 0.
7491 (ior_scc_scc): Likewise.
7492 (and_scc_scc): Likewise.
7493 (and_scc_scc_nodom): Likewise.
7494 (ior_scc_scc_cmp): Likewise for operand 7.
7495 (and_scc_scc_cmp): Likewise.
7496 * config/arm/thumb2.md (thumb2_movsi_insn):
7497 Add alternatives for 16-bit encodings.
7498 (thumb2_movhi_insn): Likewise.
7499 (thumb2_movsicc_insn): Likewise.
7500 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
7501 (thumb2_negscc): Use "Ts" constraint.
7502 Move mvn instruction outside cond_exec block.
7503 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
7504 for 16-bit encodings.
7505
31dfce10
KT		 75062013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7507
7508 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
7509 encoding.
7510 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
7511 (mulsi3subsi): Likewise.
7512 (mulsidi3adddi): Likewise.
7513 (mulsidi3_v6): Likewise.
7514 (umulsidi3_v6): Likewise.
7515 (umulsidi3adddi_v6): Likewise.
7516 (smulsi3_highpart_v6): Likewise.
7517 (umulsi3_highpart_v6): Likewise.
7518 (mulhisi3tb): Likewise.
7519 (mulhisi3bt): Likewise.
7520 (mulhisi3tt): Likewise.
7521 (maddhisi4): Likewise.
7522 (maddhisi4tb): Likewise.
7523 (maddhisi4tt): Likewise.
7524 (maddhidi4): Likewise.
7525 (maddhidi4tb): Likewise.
7526 (maddhidi4tt): Likewise.
7527 (zeroextractsi_compare0_scratch): Likewise.
7528 (insv_zero): Likewise.
7529 (insv_t2): Likewise.
7530 (anddi_notzesidi_di): Likewise.
7531 (anddi_notsesidi_di): Likewise.
7532 (andsi_notsi_si): Likewise.
7533 (iordi_zesidi_di): Likewise.
7534 (xordi_zesidi_di): Likewise.
7535 (andsi_iorsi3_notsi): Likewise.
7536 (smax_0): Likewise.
7537 (smax_m1): Likewise.
7538 (smin_0): Likewise.
7539 (not_shiftsi): Likewise.
7540 (unaligned_loadsi): Likewise.
7541 (unaligned_loadhis): Likewise.
7542 (unaligned_loadhiu): Likewise.
7543 (unaligned_storesi): Likewise.
7544 (unaligned_storehi): Likewise.
7545 (extv_reg): Likewise.
7546 (extzv_t2): Likewise.
7547 (divsi3): Likewise.
7548 (udivsi3): Likewise.
7549 (arm_zero_extendhisi2addsi): Likewise.
7550 (arm_zero_extendqisi2addsi): Likewise.
7551 (compareqi_eq0): Likewise.
7552 (arm_extendhisi2_v6): Likewise.
7553 (arm_extendqisi2addsi): Likewise.
7554 (arm_movt): Likewise.
7555 (thumb2_ldrd): Likewise.
7556 (thumb2_ldrd_base): Likewise.
7557 (thumb2_ldrd_base_neg): Likewise.
7558 (thumb2_strd): Likewise.
7559 (thumb2_strd_base): Likewise.
7560 (thumb2_strd_base_neg): Likewise.
7561 (arm_negsi2): Add alternative for 16-bit encoding.
7562 (arm_one_cmplsi2): Likewise.
7563
1572e697
KT		 75642013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7565
7566 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
7567 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
7568 (movdfcc): Likewise.
7569 * config/arm/vfp.md (*thumb2_movsf_vfp):
7570 Disable predication for arm_restrict_it.
7571 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
7572 (*thumb2_movdfcc_vfp): Likewise.
7573 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
7574 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
7575 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
7576 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
7577 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
7578 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
7579 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
7580 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
7581 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
7582 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
7583 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
7584 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
7585 Disable predication for arm_restrict_it.
7586
ebf8f0ea
KY		 75872013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
7588
7589 * config/i386/bmiintrin.h (_bextr_u32): New.
7590 (_bextr_u64): Ditto.
7591
a290fcda
RS		 75922013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
7593
7594 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
7595 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
7596 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
7597 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
7598 * config/mips/n32-elf.h: ...this new file.
7599
ea169996
MG		 76002013-06-27 Marc Glisse <marc.glisse@inria.fr>
7601
7602 PR target/57224
7603 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
7604 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
7605
12211b99 76062013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM		 7607
7608 * config/mips/mips-tables.opt: Regenerate.
7609 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 7610 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 7611 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 7612
2444244e
JJ		 76132013-06-27 Jakub Jelinek <jakub@redhat.com>
7614
cd06d2a2
JJ		 7615 PR target/57623
7616 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
7617 constraints of operand 1 and 2.
7618
2444244e
JJ		 7619 PR target/57623
7620 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
7621 to match RTL canonicalization. Swap predicates and
7622 constraints of operand 1 and 2.
7623
2f259720
VM		 76242013-06-27 Vladimir Makarov <vmakarov@redhat.com>
7625
0ba6aff3
UB		 7626 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
7627 Process OP_INOUT regs for splitting too.
2f259720 7628
fef4d2b3
JJ		 76292013-06-27 Jakub Jelinek <jakub@redhat.com>
7630
7631 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
7632 decl before the loop, initialize to NULL.
7633 (vectorizable_load): Initialize ptr_incr to NULL.
7634
79ee9826
MJ		 76352013-06-27 Martin Jambor <mjambor@suse.cz>
7636
7637 PR lto/57208
7638 * ipa-ref.h (ipa_maybe_record_reference): Declare.
7639 * ipa-ref.c (ipa_maybe_record_reference): New function.
7640 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
7641 * ipa-cp.c (create_specialized_node): Record potential references from
7642 aggvals.
7643 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
7644
e18b4a81
YZ		 76452013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7646
7647 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
7648 parameter 'mode' of type 'enum machine_mode mode'; change to pass
7649 'mode' to force_reg.
7650 (aarch64_add_offset): Update calls to aarch64_force_temporary.
7651 (aarch64_expand_mov_immediate): Likewise.
7652
9c023bf0
YZ		 76532013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7654
7655 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
7656 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
7657
5a3fe9b6
AK		 76582013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7659
7660 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
7661 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7662 (struct machine_function): Add tbegin_p.
7663 (s390_canonicalize_comparison): Fold CC mode compares to
7664 conditional jump if possible.
7665 (s390_emit_jump): Return the emitted jump.
7666 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
7667 Handle CCRAWmode compares.
7668 (s390_option_override): Default to -mhtm if available.
7669 (s390_reg_clobbered_rtx): Handle floating point regs as well.
7670 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
7671 FPRs instead of df_regs_ever_live_p.
7672 (s390_optimize_nonescaping_tx): New function.
7673 (s390_init_frame_layout): Extend clobbered_regs array to cover
7674 FPRs as well.
7675 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
7676 (s390_expand_tbegin): New function.
7677 (enum s390_builtin): New enum definition.
7678 (code_for_builtin): New array definition.
7679 (s390_init_builtins): New function.
7680 (s390_expand_builtin): New function.
7681 (TARGET_INIT_BUILTINS): Define.
7682 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 7683 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK		 7684 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
7685 (s390_alc_comparison): Likewise.
7686 * config/s390/s390-modes.def: Add CCRAWmode.
7687 * config/s390/s390.h (processor_flags): Add PF_TX.
7688 (TARGET_CPU_HTM): Define macro.
7689 (TARGET_HTM): Define macro.
7690 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
7691 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
7692 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7693 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 7694 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK		 7695 (TBEGIN_MASK, TBEGINC_MASK): New constants.
7696 ("*cc_to_int"): Move up.
7697 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
7698 constants other than 0.
7699 ("*ccraw_to_int"): New insn and splitter definition.
7700 ("tbegin", "tbegin_nofloat", "tbegin_retry")
7701 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
7702 ("tx_assist"): New expander.
7703 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
7704 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
7705 * config/s390/s390.opt: Add -mhtm option.
7706 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
7707 * config/s390/htmxlintrin.h: New file.
7708 * config/s390/htmintrin.h: New file.
7709 * config/s390/s390intrin.h: New file.
7710 * doc/extend.texi: Document htm builtins.
7711 * config.gcc: Add the new header files to extra_headers.
7712
9e216629
TS		 77132013-06-26 Thomas Schwinge <thomas@codesourcery.com>
7714
7715 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 7716 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 7717
fd8c65e7 77182013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 7719 Pat Haugen <pthaugen@us.ibm.com>
7720 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM		 7721
7722 * config/rs6000/power8.md: New.
7723 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
7724 setting for power8 entry.
7725 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
7726 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
7727 test for Power4/Power5 only.
7728 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
7729 support.
7730 (force_new_group): Adjust comment.
7731 * config/rs6000/rs6000.md: Include power8.md.
7732
b24a2ce5
GY		 77332013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
7734
7735 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
7736 * config/arm/arm-protos.h (arm_max_conditional_execute): New
7737 declaration.
7738 (tune_params): Update comment.
7739 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
7740 (arm_max_conditional_execute): New function.
7741 (thumb2_final_prescan_insn): Use max_insn_skipped and
7742 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
7743
f6fdeb62
JJ		 77442013-06-25 Jakub Jelinek <jakub@redhat.com>
7745
7746 PR tree-optimization/57705
7747 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
7748 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 7749 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ		 7750 (get_initial_def_for_induction): Handle SSA_NAME IV step.
7751
c13bc3d9
MJ		 77522013-06-25 Martin Jambor <mjambor@suse.cz>
7753
7754 PR middle-end/57670
7755 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
7756 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
7757 calls in the dump.
7758 (ipa_note_param_call): Initialize member_ptr flag.
7759 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
7760 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
7761 (ipa_write_indirect_edge_info): Stream member_ptr flag.
7762 (ipa_read_indirect_edge_info): Likewise.
7763
8a845901
RB		 77642013-06-25 Richard Biener <rguenther@suse.de>
7765
7766 PR middle-end/56977
7767 * passes.c (init_optimization_passes): Move pass_fold_builtins
7768 and pass_dce earlier with -Og.
7769
33be0bec
EB		 77702013-06-25 Eric Botcazou <ebotcazou@adacore.com>
7771
7772 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
7773 <BIT_FIELD_REF>: Remove trailing TAB.
7774 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
7775 remove blank line.
7776
5fe8e757
MJ		 77772013-06-24 Martin Jambor <mjambor@suse.cz>
7778
7779 PR tree-optimization/57358
7780 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
7781 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
7782 (ipa_analyze_params_uses): Generate pessimistic info when true.
7783
44a60244
MJ		 77842013-06-24 Martin Jambor <mjambor@suse.cz>
7785
7786 PR tree-optimization/57539
7787 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
7788 global.inlined_to of the new node to it. All callers changed.
7789 * ipa-inline-transform.c (clone_inlined_nodes): New variable
7790 inlining_into, pass it to cgraph_clone_node.
7791 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
7792 ipa_free_edge_args_substructures.
7793 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
7794 rdesc linked list. Do not assert rdesc edges have inlined caller.
7795 Assert we have found an rdesc in the rdesc list.
7796
7c5848b8
RB		 77972013-06-24 Richard Biener <rguenther@suse.de>
7798
0ba6aff3 7799 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB		 7800 (pointer_set_lookup): Declare.
7801 (class pointer_map): New template class implementing a
7802 generic pointer to T map.
7803 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
7804 pointer_map<T>::contains, pointer_map<T>::insert,
7805 pointer_map<T>::traverse): New functions.
7806 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
7807 (pointer_set_lookup): New function.
7808 (pointer_set_contains): Use pointer_set_lookup.
7809 (pointer_set_insert): Likewise.
7810 (insert_aux): Remove.
7811 (struct pointer_map_t): Embed a pointer_set_t.
7812 (pointer_map_create): Adjust.
7813 (pointer_map_destroy): Likewise.
7814 (pointer_map_contains): Likewise.
7815 (pointer_map_insert): Likewise.
7816 (pointer_map_traverse): Likewise.
7817 * tree-streamer.h (struct streamer_tree_cache_d): Use a
7818 pointer_map<unsigned> instead of a pointer_map_t.
7819 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
7820 (streamer_tree_cache_lookup): Likewise.
7821 (streamer_tree_cache_create): Likewise.
7822 (streamer_tree_cache_delete): Likewise.
7823 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
7824 pointer_map<unsigned> instead of a pointer_map_t.
7825 (lto_init_tree_ref_encoder): Adjust.
7826 (lto_destroy_tree_ref_encoder): Likewise.
7827 * lto-section-out.c (lto_output_decl_index): Likewise.
7828 (lto_record_function_out_decl_state): Likewise.
7829 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
7830
3dbc97a9
RB		 78312013-06-24 Richard Biener <rguenther@suse.de>
7832
7833 PR tree-optimization/57488
7834 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
7835
8345965a
AM		 78362013-06-24 Alan Modra <amodra@gmail.com>
7837
7838 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
7839 (gen_easy_altivec_constant): Likewise.
7840 * config/rs6000/predicates.md (easy_vector_constant_add_self,
7841 easy_vector_constant_msb): Likewise.
7842
b7df379f
JJ		 78432013-06-23 Jakub Jelinek <jakub@redhat.com>
7844
7845 PR target/57688
7846 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
7847 add missing return true.
7848
12211b99 78492013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE		 7850
7851 PR target/52483
7852 * config/sh/predicates.md (general_extend_operand): Invoke
7853 general_movsrc_operand for memory operands.
7854 (general_movsrc_operand): Allow reg+reg addressing, do not use
7855 general_operand for memory operands.
7856
97db2bf7
ST		 78572013-06-23 Sriraman Tallam <tmsriram@google.com>
7858
7859 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
7860 when current target options does not apply.
7861 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
7862 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB		 7863 * config/i386/bmiintrin.h: Pass appropriate target
7864 attributes to header.
97db2bf7
ST		 7865 * config/i386/mmintrin.h: Ditto.
7866 * config/i386/nmmintrin.h: Ditto.
7867 * config/i386/avx2intrin.h: Ditto.
7868 * config/i386/fxsrintrin.h: Ditto.
7869 * config/i386/tbmintrin.h: Ditto.
7870 * config/i386/xsaveintrin.h: Ditto.
7871 * config/i386/f16cintrin.h: Ditto.
7872 * config/i386/xtestintrin.h: Ditto.
7873 * config/i386/xsaveoptintrin.h: Ditto.
7874 * config/i386/bmi2intrin.h: Ditto.
7875 * config/i386/lzcntintrin.h: Ditto.
7876 * config/i386/smmintrin.h: Ditto.
7877 * config/i386/wmmintrin.h: Ditto.
7878 * config/i386/x86intrin.h: Remove all header include guards.
7879 * config/i386/prfchwintrin.h: Ditto.
7880 * config/i386/pmmintrin.h: Ditto.
7881 * config/i386/tmmintrin.h: Ditto.
7882 * config/i386/xmmintrin.h: Ditto.
7883 * config/i386/popcntintrin.h: Ditto.
7884 * config/i386/rdseedintrin.h: Ditto.
7885 * config/i386/ammintrin.h: Ditto.
7886 * config/i386/emmintrin.h: Ditto.
7887 * config/i386/immintrin.h: Remove all header include guards.
7888 * config/i386/fma4intrin.h: Ditto.
7889 * config/i386/lwpintrin.h: Ditto.
7890 * config/i386/xopintrin.h: Ditto.
7891 * config/i386/ia32intrin.h: Ditto.
7892 * config/i386/avxintrin.h: Ditto.
7893 * config/i386/rtmintrin.h: Ditto.
7894 * config/i386/fmaintrin.h: Ditto.
7895 * config/i386/mm3dnow.h: Ditto.
7896
495e6879
ST		 78972013-06-22 Sriraman Tallam <tmsriram@google.com>
7898
7899 * common/config/i386/i386-common.c: Handle LZCNT.
7900
11e8f687
AK		 79012013-06-22 Andi Kleen <ak@linux.intel.com>
7902
7903 * doc/extend.texi: Use __atomic_store_n instead of
7904 __atomic_store in HLE example.
7905
12211b99 79062013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE		 7907
7908 * config/sh/sh.c: Remove <cstdlib> workaround.
7909
927f908b
AK		 79102013-06-21 Andi Kleen <ak@linux.intel.com>
7911
0ba6aff3 7912 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 7913
cca410b6
AK		 79142013-06-21 Andi Kleen <ak@linux.intel.com>
7915
7916 * doc/extend.texi: Document that __atomic_clear and
7917 __atomic_test_and_set should only be used with bool.
7918
d84db1ed
JH		 79192013-06-20 Jan Hubicka <jh@suse.cz>
7920
7921 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
7922 types_same_for_odr.
7923 * tree.c (decls_same_for_odr): New function.
7924 (same_for_edr): New function.
7925 (types_same_for_odr): New function.
7926 (get_binfo_at_offset): Use it.
7927 * tree.h (types_same_for_odr): Declare.
7928
12211b99 79292013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE		 7930 Jason Merrill <jason@redhat.com>
7931
7932 * system.h: Include <cstdlib> as well as <stdlib.h>.
7933
49b4ff63
UB		 79342013-06-20 Uros Bizjak <ubizjak@gmail.com>
7935
7936 PR target/57655
7937 * config/i386/i386.c (construct_container): Report error if
7938 long double is used with disabled x87 float returns.
7939
7a89ad00
JH		 79402013-06-20 Jan Hubicka <jh@suse.cz>
7941
7942 * lto-cgraph.c (input_symtab): Do not set cgraph state.
7943
12211b99 79442013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR		 7945
7946 PR rtl-optimization/57425
7947 PR rtl-optimization/57569
7948 * alias.c (write_dependence_p): Remove parameters mem_mode and
7949 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
7950 Changed all callers.
7951 (canon_anti_dependence): Get comments and semantics in sync.
7952 Add parameter mem_canonicalized. Changed all callers.
7953 * rtl.h (canon_anti_dependence): Update prototype.
7954
d16e9a99
RB		 79552013-06-20 Richard Biener <rguenther@suse.de>
7956
7957 * data-streamer-in.c (streamer_read_uhwi): Optimize single
7958 byte case, inline streamer_read_uchar and defer section
7959 overrun check.
7960
c3a9b91b
RB		 79612013-06-20 Richard Biener <rguenther@suse.de>
7962
7963 PR tree-optimization/57584
7964 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
7965 SSA names into the expanded expression that take part in
7966 abnormal coalescing.
7967
371e77e3 79682013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS		 7969
7970 * gcov.c (print_usage): Handle new option.
7971 (process_args): Ditto.
7972 (get_gcov_intermediate_filename): New function.
7973 (output_intermediate_file): New function.
7974 (output_gcov_file): New function
7975 (generate_results): Handle new option.
7976 (release_function): Relase demangled name.
7977 (read_graph_file): Handle demangled name.
7978 (output_lines): Ditto.
7979 * doc/gcov.texi: Document gcov intermediate format.
7980
11f5b71a
VM		 79812013-06-19 Vladimir Makarov <vmakarov@redhat.com>
7982
7983 PR bootstrap/57604
7984 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
7985 (lra_emit_add): Use the functions. Add comment about Y as an
7986 address segment.
7987
e1f9340b
DE		 79882013-06-19 David Edelsohn <dje.gcc@gmail.com>
7989
7990 PR driver/57652
7991 * collect2.c (collect_atexit): New.
7992 (collect_exit): Delete.
7993 (main): Register collect_atexit with atexit.
7994 (collect_wait): Change collect_exit to exit.
7995 (do_wait): Same.
7996 * collect2.h (collect_exit): Delete.
7997 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
7998
e3f9e0ac
WM		 79992013-06-19 Wei Mi <wmi@google.com>
8000
8001 PR rtl-optimization/57518
8002 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8003 if regno is used in paradoxical subreg.
8004 (update_equiv_regs): Check pdx_subregs[regno] before
8005 set a reg to be equivalent with a mem.
8006
a09f1a76
MK		 80072013-06-19 Matthias Klose <doko@ubuntu.com>
8008
8009 PR driver/57651
8010 * file-find.h (find_a_file): Add a mode parameter.
8011 * file-find.c (find_a_file): Likewise.
8012 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8013 with X_OK for the executables.
8014 * collect2.c (main): Call find_a_file with X_OK.
8015
7d18b0ad
SE		 80162013-06-19 Steve Ellcey <sellcey@mips.com>
8017
8018 PR target/56942
8019 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8020 Use NEXT_INSN instead of next_real_insn.
8021
6a6dac52
JH		 80222013-06-19 Jan Hubicka <jh@suse.cz>
8023
8024 * cgraph.h (const_value_known_p): Replace by ...
8025 (ctor_for_folding): .. this one.
8026 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8027 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8028 * expr.c (expand_expr_real_1): Likewise.
8029 (string_constant): Likewise.
8030 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8031 * ipa.c (process_references): Likewise.
8032 (symtab_remove_unreachable_nodes): Likewise.
8033 * ipa-inline-analysis.c (param_change_prob): Likewise.
8034 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8035 (get_base_constructor): Likwise.
8036 * varpool.c (varpool_remove_node): Likewise.
8037 (varpool_remove_initializer): LIkewise.
8038 (dump_varpool_node): LIkwise.
8039 (const_value_known_p): Rewrite to ...
8040 (ctor_for_folding): ... this one.
8041
216c12ab
JJ		 80422013-06-19 Jakub Jelinek <jakub@redhat.com>
8043
8044 PR driver/57651
8045 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8046 PERSONALITY in $PATH derived prefixes.
8047
d53e2f99
JL		 80482013-06-19 Jeff Law <law@redhat.com>
8049
ecdbf306
JL		 8050 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8051 in comment.
8052
d53e2f99
JL		 8053 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8054 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8055 booleans.
8056
89e7fc04
SN		 80572013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8058
8059 * config/arm/vfp.md: Move VFP instruction classification documentation
8060 to ...
8061 * config/arm/arm.md: ... here. Update instruction classification
8062 documentation.
8063
5e7f6aaa
RE		 80642013-06-19 Richard Earnshaw <rearnsha@arm.com>
8065
8066 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8067 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8068 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8069
5fdc1e5d
SB		 80702013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8071
8072 PR target/57609
8073 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8074 with NEXT_INSN. Use tablejump_p to check for jump table data
8075 insns.
8076
e952afba
PC		 80772013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8078
8079 PR c++/56544
8080 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8081 that now in C++ the value is correct per the C++ standards.
8082
3fd0303a
RB		 80832013-06-19 Richard Biener <rguenther@suse.de>
8084
8085 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8086 for global context.
8087
b04e4b08
AK		 80882013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8089
8090 Revert:
8091 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8092
8093 PR target/57609
8094 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8095 with next_active_insn.
8096
bef8491a
ST		 80972013-06-18 Sriraman Tallam <tmsriram@google.com>
8098
8099 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
8100 functions are inlined during failures to flag an error.
8101 * tree-inline.c (expand_call_inline): Allow the error to be flagged
8102 in early inline pass.
22fc7d3d 8103
c21b257d
L		 81042013-06-18 H.J. Lu <hongjiu.lu@intel.com>
8105
8106 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
8107 in comments.
8108
33255ae3
JB		 81092013-06-18 Julian Brown <julian@codesourcery.com>
8110
8111 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
8112 Permit virtual register pre-reload if !strict.
8113 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
8114 change.
8115 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
8116 prototype.
8117 * config/arm/neon.md (movmisalign<mode>): Use
8118 neon_perm_struct_or_reg_operand instead of
8119 neon_struct_or_register_operand.
8120 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
8121 neon_permissive_struct_operand instead of neon_struct_operand.
8122 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
8123 neon_vector_mem_operand.
8124 * config/arm/predicates.md (neon_struct_operand): Adjust call to
8125 neon_vector_mem_operand.
8126 (neon_permissive_struct_operand): New.
8127 (neon_struct_or_register_operand): Rename to...
8128 (neon_perm_struct_or_reg_operand): This. Adjust call to
8129 neon_vector_mem_operand.
8130
d579fcda
RB		 81312013-06-18 Richard Biener <rguenther@suse.de>
8132
8133 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
8134 * lto-streamer.h: Include pointer-set.h.
8135 (struct lto_decl_slot): Remove.
8136 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
8137 Remove next_index entry.
8138 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8139 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8140 (lto_init_tree_ref_encoder): Adjust.
8141 (lto_destroy_tree_ref_encoder): Likewise.
8142 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8143 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8144 (lto_output_decl_index): Adjust.
8145 (lto_new_out_decl_state): Likewise.
8146 (lto_record_function_out_decl_state): Likewise.
8147 * lto-streamer-out.c (copy_function): Likewise.
8148
31ee20ba
RB		 81492013-06-18 Richard Biener <rguenther@suse.de>
8150
8151 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
8152 * cgraphunit.c: Include cfgloop.h.
8153 (init_lowered_empty_function): Initialize the loop tree.
8154 (assemble_thunk): Insert new BBs into loops.
8155
ac0511f2
RB		 81562013-06-18 Richard Biener <rguenther@suse.de>
8157
8158 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8159 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8160 the map from cache entry to cache index optional.
8161 (streamer_tree_cache_replace_tree): Adjust accordingly.
8162 (streamer_tree_cache_append): Likewise.
8163 (streamer_tree_cache_delete): Likewise.
8164 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8165 streamer cache map from cache entry to cache index.
8166 * lto-streamer-out.c (create_output_block): Adjust.
8167
09485a08
SN		 81682013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8169
22fc7d3d
UB		 8170 * config/arm/arm.md (attribute "insn"): Move multiplication and
8171 division attributes to...
09485a08
SN		 8172 (attribute "type"): ... here. Remove mult.
8173 (attribute "mul32"): New attribute.
8174 (attribute "mul64"): Add umaal.
8175 (*arm_mulsi3): Update attributes.
8176 (*arm_mulsi3_v6): Likewise.
8177 (*thumb_mulsi3): Likewise.
8178 (*thumb_mulsi3_v6): Likewise.
8179 (*mulsi3_compare0): Likewise.
8180 (*mulsi3_compare0_v6): Likewise.
8181 (*mulsi_compare0_scratch): Likewise.
8182 (*mulsi_compare0_scratch_v6): Likewise.
8183 (*mulsi3addsi): Likewise.
8184 (*mulsi3addsi_v6): Likewise.
8185 (*mulsi3addsi_compare0): Likewise.
8186 (*mulsi3addsi_compare0_v6): Likewise.
8187 (*mulsi3addsi_compare0_scratch): Likewise.
8188 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8189 (*mulsi3subsi): Likewise.
8190 (*mulsidi3adddi): Likewise.
8191 (*mulsi3addsi_v6): Likewise.
8192 (*mulsidi3adddi_v6): Likewise.
8193 (*mulsidi3_nov6): Likewise.
8194 (*mulsidi3_v6): Likewise.
8195 (*umulsidi3_nov6): Likewise.
8196 (*umulsidi3_v6): Likewise.
8197 (*umulsidi3adddi): Likewise.
8198 (*umulsidi3adddi_v6): Likewise.
8199 (*smulsi3_highpart_nov6): Likewise.
8200 (*smulsi3_highpart_v6): Likewise.
8201 (*umulsi3_highpart_nov6): Likewise.
8202 (*umulsi3_highpart_v6): Likewise.
8203 (mulhisi3): Likewise.
8204 (*mulhisi3tb): Likewise.
8205 (*mulhisi3bt): Likewise.
8206 (*mulhisi3tt): Likewise.
8207 (maddhisi4): Likewise.
8208 (*maddhisi4tb): Likewise.
8209 (*maddhisi4tt): Likewise.
8210 (maddhidi4): Likewise.
8211 (*maddhidi4tb): Likewise.
8212 (*maddhidi4tt): Likewise.
8213 (divsi3): Likewise.
8214 (udivsi3): Likewise.
8215 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8216 (thumb2_mulsi_short_compare0): Likewise.
8217 (thumb2_mulsi_short_compare0_scratch): Likewise.
8218 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8219 (1020mult2): Likewise.
8220 (1020mult3): Likewise.
8221 (1020mult4): Likewise.
8222 (1020mult5): Likewise.
8223 (1020mult6): Likewise.
22fc7d3d
UB		 8224 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8225 change.
09485a08
SN		 8226 (cortex_a15_mult64): Likewise.
8227 (cortex_a15_sdiv): Likewise.
8228 (cortex_a15_udiv): Likewise.
8229 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8230 (mult2): Likewise.
8231 (mult3): Likewise.
8232 (mult4): Likewise.
8233 (mult5): Likewise.
8234 (mult6): Likewise.
8235 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8236 (pj4_ir_div): Likewise.
8237 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8238 (11_mult2): Likewise.
8239 (11_mult3): Likewise.
8240 (11_mult4): Likewise.
8241 (11_mult5): Likewise.
8242 (11_mult6): Likewise.
8243 (11_mult7): Likewise.
8244 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8245 (cortex_a8_mla): Likewise.
8246 (cortex_a8_mull): Likewise.
8247 (cortex_a8_smulwy): Likewise.
8248 (cortex_a8_smlald): Likewise.
8249 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8250 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8251 (cortex_r4_mul_3): Likewise.
8252 (cortex_r4_mla_4): Likewise.
8253 (cortex_r4_mla_3): Likewise.
8254 (cortex_r4_smlald): Likewise.
8255 (cortex_r4_mull): Likewise.
8256 (cortex_r4_sdiv): Likewise.
8257 (cortex_r4_udiv): Likewise.
8258 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8259 (cortex_a7_idiv): Likewise.
8260 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8261 (9_mult2): Likewise.
8262 (9_mult3): Likewise.
8263 (9_mult4): Likewise.
8264 (9_mult5): Likewise.
8265 (9_mult6): Likewise.
8266 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8267 (cortex_a53_sdiv): Likewise.
8268 (cortex_a53_udiv): Likewise.
8269 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8270 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8271 (mp626_mult2): Likewise.
8272 (mp626_mult3): Likewise.
8273 (mp626_mult4): Likewise.
8274 * config/arm/fa526.md (526_mult1): Update attribute change.
8275 (526_mult2): Likewise.
8276 * config/arm/arm-generic.md (mult): Update attribute change.
8277 (mult_ldsched_strongarm): Likewise.
8278 (mult_ldsched): Likewise.
8279 (multi_cycle): Likewise.
8280 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8281 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8282 (606te_mult2): Likewise.
8283 (606te_mult3): Likewise.
8284 (606te_mult4): Likewise.
8285 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8286 (cortex_a9_mac16): Likewise.
8287 (cortex_a9_multiply): Likewise.
8288 (cortex_a9_mac): Likewise.
8289 (cortex_a9_multiply_long): Likewise.
8290 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8291 (626te_mult2): Likewise.
8292 (626te_mult3): Likewise.
8293 (626te_mult4): Likewise.
8294
8951f345
RB		 82952013-06-18 Richard Biener <rguenther@suse.de>
8296
8297 PR lto/57334
8298 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8299
d8101d05
AK		 83002013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8301
8302 PR target/57609
8303 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8304 with next_active_insn.
8305
69eff9da
AM		 83062013-06-18 Alan Modra <amodra@gmail.com>
8307
8308 * config/rs6000/rs6000.h (enum data_align): New.
8309 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8310 (DATA_ABI_ALIGNMENT): Define.
8311 (CONSTANT_ALIGNMENT): Correct comment.
8312 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8313 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8314
8ed00d76
DM		 83152013-06-17 David Malcolm <dmalcolm@redhat.com>
8316
8317 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8318 ATTRIBUTE_UNUSED marking.
8319
3d33d06b
SN		 83202013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8321
8322 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8323 alternative and update.
8324 (aarch64_dup_lanedi): Delete.
8325 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8326 * config/aarch64/aarch64-simd-builtins.def: Update.
8327
ee03e71d
RB		 83282013-06-17 Richard Biener <rguenther@suse.de>
8329
8330 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8331 (lto_input_scc): Declare.
8332 (lto_input_tree_1): Likewise.
8333 (struct lto_stats_d): Add num_tree_bodies_output and
8334 num_pickle_refs_output.
8335 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8336 (lto_read_tree_1): Split out from ...
8337 (lto_read_tree): ... this.
8338 (lto_input_scc): New function.
8339 (lto_input_tree_1): Split out from ...
8340 (lto_input_tree): ... this. Handle LTO_tree_scc.
8341 (lto_data_in_create): Create the streamer cache without hashes.
8342 * lto-streamer-out.c (create_output_block): Create the streamer
8343 cache with hashes when not doing WPA.
8344 (lto_write_tree_1): Split out from ...
8345 (lto_write_tree): ... this.
8346 (get_symbol_initial_value): New function.
8347 (lto_output_tree_1): Split out from ...
8348 (lto_output_tree): ... this. Write trees as series of SCCs
8349 using a DFS walk via DFS_write_tree.
8350 (struct sccs, struct scc_entry): New types.
8351 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
8352 (DFS_write_tree_body): New function.
8353 (DFS_write_tree): Likewise.
8354 (hash_tree): Likewise.
8355 (scc_entry_compare): Likewise.
8356 (hash_scc): Likewise.
8357 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
8358 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
8359 TREE_CHAIN as regular reference.
8360 (streamer_read_integer_cst): Remove.
8361 (streamer_get_pickled_tree): Adjust.
8362 * tree-streamer-out.c (streamer_write_chain): Disable streaming
8363 of DECL_EXTERNALs in BLOCK_VARS for now.
8364 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
8365 reference.
8366 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
8367 Add hash value argument and record that if hashes are recorded
8368 in the cache.
8369 (streamer_tree_cache_insert_1): Adjust.
8370 (streamer_tree_cache_insert): Likewise.
8371 (streamer_tree_cache_insert_at): Rename to ...
8372 (streamer_tree_cache_replace_tree): ... this and adjust.
8373 (streamer_tree_cache_append): Adjust.
8374 (record_common_node): Likewise.
8375 (streamer_tree_cache_create): Add argument whether to
8376 record hash values together with trees.
8377 (streamer_tree_cache_delete): Adjust.
8378 * tree-streamer.h (struct streamer_tree_cache_d): Add
8379 vector of hashes.
8380 (streamer_read_integer_cst): Remove.
8381 (streamer_tree_cache_insert): Adjust.
8382 (streamer_tree_cache_append): Likewise.
8383 (streamer_tree_cache_insert_at): Rename to ...
8384 (streamer_tree_cache_replace_tree): ... this and adjust.
8385 (streamer_tree_cache_create): Add argument whether to record hashes.
8386 (streamer_tree_cache_get): Rename to ...
8387 (streamer_tree_cache_get_tree): ... this.
8388 (streamer_tree_cache_get_hash): New function.
8389 * tree.c (cache_integer_cst): New function.
8390 * tree.h (cache_integer_cst): Declare.
8391 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
8392 * lto-symtab.c (lto_varpool_replace_node): Only release
8393 DECL_INITIAL of non-prevailing decls.
8394 * varpool.c (varpool_remove_initializer): Do not release
8395 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
8396
0de86a92
JU		 83972013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8398
8399 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
8400 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
8401 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
8402 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
8403 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
8404 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
8405 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
8406 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
8407 instead of TARGET_64BIT.
8408 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
8409 Require ISA_HAS_<D>DIV.
8410
74240413
RS		 84112013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
8412
8413 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
8414 (mips*-*-linux*): Move default with_llsc setting to where other
8415 defaults are set.
8416 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
8417 with_arch block.
8418 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
8419 Likewise. Remove default with_tune setting. Move default float
8420 setting to its own block. Handle with_llsc in the same block as above.
8421
12211b99 84222013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR		 8423
8424 PR rtl-optimization/57425
8425 PR rtl-optimization/57569
299a5f6a 8426 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR		 8427 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
8428 Changed all callers.
8429 (canon_anti_dependence): New function.
8430 * cse.c (check_dependence): Use canon_anti_dependence.
8431 * cselib.c (cselib_invalidate_mem): Likewise.
8432 * rtl.h (canon_anti_dependence): Declare.
8433
b259d352
JU		 84342013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8435
8436 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
8437 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
8438 ".set mips3" for 64-bit targets.
8439
4de09b85
DC		 84402013-06-15 Dehao Chen <dehao@google.com>
8441
8442 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
8443 * gimple-low.c (gimple_check_call_matching_types): Likewise.
8444 (gimple_check_call_args): Likewise.
8445 * value-prof.c (check_ic_target): Likewise.
8446 * ipa-inline.c (early_inliner): Likewise.
8447 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
8448 * cgraph.c (cgraph_create_edge_1): Likewise.
8449 (cgraph_make_edge_direct): Likewise.
8450
d1aa4a76
MM		 84512013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
8452
8453 PR target/57615
8454 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
8455 rs6000_output_move_128bit to handle emitting quad memory
8456 operations. Set attribute length to 8 bytes.
8457
12211b99 84582013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP		 8459
8460 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
8461 New pattern.
8462 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
8463 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
8464 (aarch64_<su>mlsl<mode>): Likewise.
8465
f5273574
MS		 84662013-06-14 Mike Stump <mikestump@comcast.net>
8467
8468 * Makefile.in (TARGET_H): Add insn-codes.h.
8469
ef7e7ab2
AM		 84702013-06-14 Alan Modra <amodra@gmail.com>
8471
8472 PR middle-end/57134
8473 PR middle-end/57586
8474 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
8475 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
8476 bitfield expansion when EXPAND_MEMORY.
8477 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
8478
135faab6
MM		 84792013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
8480
8481 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
8482 test for clearing quad memory on 32-bit later.
8483
948a1fd9
MG		 84842013-06-13 Marc Glisse <marc.glisse@inria.fr>
8485
8486 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
8487 (fold_negate_expr): Likewise.
8488 (fold_real_zero_addition_p): Handle vectors.
8489 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
8490
0597fb9c
AM		 84912013-06-14 Alan Modra <amodra@gmail.com>
8492
8493 * varasm.c (force_const_mem): Revert 2013-06-07 change.
8494
e5b962d0
JH		 84952013-06-13 Jan Hubicka <jh@suse.cz>
8496
8497 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
8498 Local comdats are not externally visible.
8499 * symtab.c (dump_symtab_base): Dump externally visible.
8500 (verify_symtab_base): Verify back links in the symtab hash.
8501
a8355e51
BC		 85022013-06-13 Bin Cheng <bin.cheng@arm.com>
8503
8504 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
8505 CONVERT_EXPR as equal nodes.
8506
8f7e6e33
BC		 85072013-06-13 Bin Cheng <bin.cheng@arm.com>
8508
22fc7d3d 8509 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 8510
5d418483
MG		 85112013-06-13 Marc Glisse <marc.glisse@inria.fr>
8512
8513 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
8514 Generalize to complex and vector.
8515 * tree.c (build_all_ones_cst): New function.
8516 * tree.h (build_all_ones_cst): Declare it.
8517
50751417
AM		 85182013-06-13 Alan Modra <amodra@gmail.com>
8519
8520 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
8521 * config/rs6000/rs6000.md (signbittf2): New insn.
8522 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
8523 (abstf2_internal, cmptf_internal2): Likewise.
8524 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
8525
4b02c962 85262013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 8527 Pat Haugen <pthaugen@us.ibm.com>
8528 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM		 8529
8530 * config/rs6000/rs6000.c (emit_load_locked): Add support for
8531 power8 byte, half-word, and quad-word atomic instructions.
8532 (emit_store_conditional): Likewise.
8533 (rs6000_expand_atomic_compare_and_swap): Likewise.
8534 (rs6000_expand_atomic_op): Likewise.
8535
8536 * config/rs6000/sync.md (larx): Add new modes for power8.
8537 (stcx): Likewise.
8538 (AINT): New mode iterator to include TImode as well as normal
8539 integer modes on power8.
8540 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
8541 that VSX registers are not considered. Use AINT mode iterator
8542 instead of INT1 to allow inclusion of quad word atomic operations
8543 on power8.
8544 (load_locked<mode>): Likewise.
8545 (store_conditional<mode>): Likewise.
8546 (atomic_compare_and_swap<mode>): Likewise.
8547 (atomic_exchange<mode>): Likewise.
8548 (atomic_nand<mode>): Likewise.
8549 (atomic_fetch_<fetchop_name><mode>): Likewise.
8550 (atomic_nand_fetch<mode>): Likewise.
8551 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
8552 each type.
8553 (ATOMIC): On power8, add QImode, HImode modes.
8554 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
8555 modes that promote to SImode.
8556 (load_lockedti): Convert TImode arguments to PTImode, so that we
8557 get a guaranteed even/odd register pair.
8558 (load_lockedpti): Likewise.
8559 (store_conditionalti): Likewise.
8560 (store_conditionalpti): Likewise.
8561
8562 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
8563 atomic load/store instructions.
8564 (HSI): Likewise.
8565
710b6cc1
RS		 85662013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
8567
8568 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
8569 loads.
8570 (insn_count): New attribute, with most cases extracted from...
8571 (length): ...here. Redefine most cases in terms of insn_count.
8572 (single_insn): Delete.
8573 (can_delay): Use insn_count to check for single instructions.
8574 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
8575 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
8576 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
8577 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
8578 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
8579 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
8580 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
8581 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
8582 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
8583 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
8584 rather than "length".
8585 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
8586 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
8587 Use "insn_count" rather than "length".
8588 * config/mips/mips-dsp.md
8589 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
8590 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
8591 length attributes.
8592
1951f101
MG		 85932013-06-12 Marc Glisse <marc.glisse@inria.fr>
8594
8595 PR tree-optimization/57361
8596 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
8597
8b033a8a
SN		 85982013-06-12 Sofiane Naci <sofiane.naci@arm.com>
8599
22fc7d3d
UB		 8600 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
8601 to split.
8b033a8a
SN		 8602 (aarch64_simd_combine<mode>): New instruction expansion.
8603 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
8604 function prototype.
8605 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
8606 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
8607
ca0f62a8
JH		 86082013-06-12 Jan Hubicka <jh@suse.cz>
8609
8610 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
8611 decl has when in streaming stage.
8612 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
8613 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
8614
12211b99 86152013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS		 8616
8617 PR target/57578
8618 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
8619
bfdeda2c
JJ		 86202013-06-12 Jakub Jelinek <jakub@redhat.com>
8621
8622 PR tree-optimization/57537
8623 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
8624 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
8625
a4fa02d1
RB		 86262013-06-12 Richard Biener <rguenther@suse.de>
8627
8628 * data-streamer.h (streamer_write_char_stream): CSE
8629 obs->current_pointer.
8630 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
8631 streamer_write_char_stream manually and optimize the resulting loop.
8632 (streamer_write_hwi_stream): Likewise.
8633
bbf9ad07
JH		 86342013-06-12 Jan Hubicka <jh@suse.cz>
8635
8636 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
8637 * cgraph.h (varpool_create_empty_node): Declare.
8638 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
8639 duplicated nodes.
8640 * symtab.c (symtab_unregister_node): Be lax about missin entries
8641 in node hash.
8642 (symtab_get_node): Update comment.
8643 * varpool.c (varpool_create_empty_node): Break out from ...
8644 (varpool_node_for_decl): ... here.
8645 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
8646
a53f90ad
EB		 86472013-06-12 Eric Botcazou <ebotcazou@adacore.com>
8648
8649 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
8650 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
8651 part. Use straight-line flow at the end.
8652 <COMPONENT_REF>: Remove superfluous else.
8653 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
8654
946f9306
JJ		 86552013-06-12 Jakub Jelinek <jakub@redhat.com>
8656
8657 PR target/56564
8658 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
8659 target hook even for !TREE_PUBLIC decls. If no resolution info
8660 is available, return false for common and external decls.
8661
a7fca6f0
KP		 86622013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8663
8664 * config/rl78/constraints.md (U): New constraint.
8665 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
8666 valloc attribute.
8667
c7d8f446
MM		 86682013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8669
8670 PR target/57589
8671 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
8672 to allow returning address to AT_PLATFORM name.
8673
08346abd
JH		 86742013-06-11 Jan Hubicka <jh@suse.cz>
8675
8676 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
8677 * cgraph.h (symtab_node_base): Add weakref flag.
8678 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
8679 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
8680 (output_weakrefs): Use weakref flag.
8681 * fold-const.c (simple_operand_p): Handle WEAK.
8682 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
8683 * ipa.c (varpool_externally_visible_p): Drop weakref.
8684 (function_and_variable_visibility): Update comment; fix weakref
8685 sanity checks; do not clear DECL_WEAK on them.
8686 * lto-cgraph.c (lto_output_node): update.
8687 (lto_output_varpool_node): Update.
8688 (input_overwrite_node): Update.
8689 (input_node): Update.
8690 (input_varpool_node): Update.
8691 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
8692 (lto_symtab_merge_symbols): Add sanity check.
8693 (lto_symtab_prevailing_decl): Do not special case weakrefs.
8694 * passes.c (rest_of_decl_compilation): Set static flag, too.
8695 * symtab.c (dump_symtab_base): Dump weakref.
8696 (verify_symtab_base): Sanity check weakrefs.
8697 (symtab_make_decl_local): Remove duplicated code.
8698 (symtab_alias_ultimate_target): Simplify.
8699 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 8700
2e6a2f88
TV		 87012013-06-11 Tom de Vries <tom@codesourcery.com>
8702
8703 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
8704 sequence_vect == NULL.
8705
d9af3652
DD		 87062013-06-11 DJ Delorie <dj@redhat.com>
8707
8708 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
8709 (rl78_unwind_word_mode): New.
8710
ffa66012
DM		 87112013-06-11 David Malcolm <dmalcolm@redhat.com>
8712
8713 * final.c (debug_prefix_maps): Make static.
8714
783f2d35
DM		 87152013-06-11 David Malcolm <dmalcolm@redhat.com>
8716
8717 * function.c (initial_trampoline): Remove stray copy.
8718
7af26def
SN		 87192013-06-11 Sofiane Naci <sofiane.naci@arm.com>
8720
8721 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
8722
97756c0e
MJ		 87232013-06-11 Martin Jambor <mjambor@suse.cz>
8724
8725 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
8726 within bounds at the beginning of the function.
8727
2c7b8bf6
AM		 87282013-06-11 Alan Modra <amodra@gmail.com>
8729
8730 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
8731 reporting.
8732 (get_named_section): Don't NULL !DECL_P decl.
8733
77cea46e
IZ		 87342013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
8735
8736 * doc/invoke.texi (core-avx2): Document.
8737 (slm): Likewise.
8738 (atom): Updated with MOVBE.
8739
888f0920
RB		 87402013-06-11 Richard Biener <rguenther@suse.de>
8741
22fc7d3d 8742 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 8743
6b438d58
AB		 87442013-06-11 Anton Blanchard <anton@samba.org>
8745
8746 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
8747 correct shift value in little-endian mode.
8748
3bc66938
JJ		 87492013-06-11 Jakub Jelinek <jakub@redhat.com>
8750
8751 PR target/56564
8752 * varasm.c (get_variable_align): Move #endif to the right place.
8753
a5dc7f88
CC		 87542013-06-10 Cary Coutant <ccoutant@google.com>
8755
8756 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
8757 for hash so that hash table traversal order is deterministic.
8758
19be72ab 87592013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 8760 Pat Haugen <pthaugen@us.ibm.com>
8761 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM		 8762
8763 * config/rs6000/vector.md (GPR move splitter): Do not split moves
8764 of vectors in GPRS if they are direct moves or quad word load or
8765 store moves.
8766
8767 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
8768 declaration.
8769 (direct_move_p): Likewise.
8770 (quad_load_store_p): Likewise.
8771
8772 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
8773 classes into bins based on the physical register type.
8774 (reg_class_to_reg_type): Likewise.
8775 (IS_STD_REG_TYPE): Likewise.
8776 (IS_FP_VECT_REG_TYPE): Likewise.
8777 (reload_fpr_gpr): Arrays to determine what insn to use if we can
8778 use direct move instructions.
8779 (reload_gpr_vsx): Likewise.
8780 (reload_vsx_gpr): Likewise.
8781 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
8782 information that is a simplification of register classes. Also
8783 precalculate direct move reload helpers.
8784 (direct_move_p): New function to return true if the operation can
8785 be done as a direct move instruciton.
8786 (quad_load_store_p): New function to return true if the operation
8787 is a quad memory operation.
8788 (rs6000_legitimize_address): If quad memory, only allow register
8789 indirect for TImode addresses.
8790 (rs6000_legitimate_address_p): Likewise.
8791 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
8792 (rs6000_reload_register_type): Likewise.
8793 (register_to_reg_type): Return register type.
8794 (rs6000_secondary_reload_simple_move): New helper function for
8795 secondary reload and secondary memory needed to identify anything
8796 that is a simple move, and does not need reloading.
8797 (rs6000_secondary_reload_direct_move): New helper function for
8798 secondary reload to identify cases that can be done with several
8799 instructions via the direct move instructions.
8800 (rs6000_secondary_reload_move): New helper function for secondary
8801 reload to identify moves between register types that can be done.
8802 (rs6000_secondary_reload): Add support for quad memory operations
8803 and for direct move.
8804 (rs6000_secondary_memory_needed): Likewise.
8805 (rs6000_debug_secondary_memory_needed): Change argument names.
8806 (rs6000_output_move_128bit): New function to return the move to
8807 use for 128-bit moves, including knowing about the various
8808 limitations of quad memory operations.
8809
8810 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
8811 memory operations. call rs6000_output_move_128bit for the actual
8812 instruciton(s) to generate.
8813 (vsx_movti_64bit): Likewise.
8814
8815 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
8816 (UNSPEC_P8V_MTVSRWZ): Likewise.
8817 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
8818 (UNSPEC_P8V_MTVSRD): Likewise.
8819 (UNSPEC_P8V_XXPERMDI): Likewise.
8820 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
8821 (UNSPEC_FUSION_GPR): Likewise.
8822 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 8823 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM		 8824 (f32_sv): Likewise.
8825 (f32_dm): Likewise.
8826 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
8827 loads and direct move instructions.
8828 (zero_extendsidi2_lfiwzx): Likewise.
8829 (extendsidi2_lfiwax): Likewise.
8830 (extendsidi2_nocell): Likewise.
8831 (floatsi<mode>2_lfiwax): Likewise.
8832 (lfiwax): Likewise.
8833 (floatunssi<mode>2_lfiwzx): Likewise.
8834 (lfiwzx): Likewise.
8835 (fix_trunc<mode>_stfiwx): Likewise.
8836 (fixuns_trunc<mode>_stfiwx): Likewise.
8837 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
8838 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
8839 (parity<mode>2_cmpb): Set length/type attr.
8840 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
8841 for 'mr.' to fast_compare.
8842 (bpermd_<mode>): Change type attr to popcnt.
8843 (p8_fmrgow_<mode>): New insns for power8 direct move support.
8844 (p8_mtvsrwz_1): Likewise.
8845 (p8_mtvsrwz_2): Likewise.
8846 (reload_fpr_from_gpr<mode>): Likewise.
8847 (p8_mtvsrd_1): Likewise.
8848 (p8_mtvsrd_2): Likewise.
8849 (p8_xxpermdi_<mode>): Likewise.
8850 (reload_vsx_from_gpr<mode>): Likewise.
8851 (reload_vsx_from_gprsf): Likewise.
8852 (p8_mfvsrd_3_<mode>): LIkewise.
8853 (reload_gpr_from_vsx<mode>): Likewise.
8854 (reload_gpr_from_vsxsf): Likewise.
8855 (p8_mfvsrd_4_disf): Likewise.
8856 (multi-word GPR splits): Do not split direct moves or quad memory
8857 operations.
8858
16876bdc
DM		 88592013-06-10 David Malcolm <dmalcolm@redhat.com>
8860
8861 * tree-into-ssa.c (interesting_blocks): Make static.
8862
df8a1d28
JJ		 88632013-06-10 Jakub Jelinek <jakub@redhat.com>
8864
8865 PR target/56564
8866 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
8867 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
8868 Use DATA_ABI_ALIGNMENT for that case instead if defined.
8869 (get_variable_align): New function.
8870 (get_variable_section, emit_bss, emit_common,
8871 assemble_variable_contents, place_block_symbol): Use
8872 get_variable_align instead of DECL_ALIGN.
8873 (assemble_noswitch_variable): Add align argument, use it
8874 instead of DECL_ALIGN.
8875 (assemble_variable): Adjust caller. Use get_variable_align
8876 instead of DECL_ALIGN.
8877 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
8878 caller.
8879 (DATA_ABI_ALIGNMENT): Define.
8880 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
8881 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
8882 opt is false, only return the psABI mandated alignment increase.
8883 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
8884 (DATA_ABI_ALIGNMENT): ... this.
8885 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
8886 (DATA_ABI_ALIGNMENT): ... this.
8887 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
8888 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
8889 (DATA_ABI_ALIGNMENT): ... this.
8890 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
8891 * doc/tm.texi: Regenerated.
8892
57b29ca6
UB		 88932013-06-10 Uros Bizjak <ubizjak@gmail.com>
8894
8895 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
8896 cmp_code to construct REG_EQUAL note.
8897
3e56ed50
JJ		 88982013-06-09 Jakub Jelinek <jakub@redhat.com>
8899
8900 PR target/57568
8901 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
8902 that operands[2] doesn't overlap with operands[0].
8903
74fe2a1d
DE		 89042013-06-09 David Edelsohn <dje.gcc@gmail.com>
8905 Jan Hubicka <jh@suse.cz>
8906
8907 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
8908 hack to mark symbols as used.
8909
ed52a84e
VM		 89102013-06-08 Vladimir Makarov <vmakarov@redhat.com>
8911
8912 PR rtl-optimization/57559
8913 * lra-constraints.c (process_alt_operands): Don't discourage
8914 memory with known offset for offsetable memory constraint.
8915 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
8916
299316ed
EB		 89172013-06-08 Eric Botcazou <ebotcazou@adacore.com>
8918
8919 * varasm.c (struct oc_local_state): Reorder fields.
8920 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
8921 and adjust accordingly.
8922 (output_constructor): Reorder initialization code and adjust call to
8923 output_constructor_bitfield.
8924
88ac60d3
JH		 89252013-06-07 Jan Hubicka <jh@suse.cz>
8926
8927 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
8928
4cdce1a8
DM		 89292013-06-07 David Malcolm <dmalcolm@redhat.com>
8930
8931 * tree-object-size.c (unknown): Make const.
8932
b5e0425c
AK		 89332013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8934
8935 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
8936 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
8937 for last alternative in the cpu_facility attribute.
8938
79678d04
KT		 89392013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8940
8941 PR target/56315
8942 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
8943 (xordi3): Change operand 2 constraint to arm_xordi_operand.
8944 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
8945 * config/arm/constraints.md (Dg): New constraint.
8946 * config/arm/neon.md (xordi3_neon): Remove.
8947 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
8948 * config/arm/predicates.md (arm_xordi_operand): New predicate.
8949
b31ddbdb
KT		 89502013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8951
8952 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
8953 Clean up alternatives.
8954
8f90b7d4
AM		 89552013-06-07 Alan Modra <amodra@gmail.com>
8956
8957 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
8958 va_list_gpr_size.
8959
961ce119
AM		 89602013-06-07 Alan Modra <amodra@gmail.com>
8961
8962 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
8963
33e49835
KT		 89642013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8965
8966 * config/arm/constraints.md (Df): New constraint.
8967 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
8968 Correct length attribute for last two alternatives.
8969
acd17ae6
AM		 89702013-06-07 Alan Modra <amodra@gmail.com>
8971
8972 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8973 override user -mfp-in-toc.
8974 (offsettable_ok_by_alignment): Consider just the current access
8975 rather than the whole object, unless BLKmode. Handle
8976 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
8977 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
8978 for -mcmodel=medium.
8979 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
8980 override user -mfp-in-toc or -msum-in-toc. Default to
8981 -mno-fp-in-toc for -mcmodel=medium.
8982
73310b0e
DD		 89832013-06-06 DJ Delorie <dj@redhat.com>
8984
8985 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
8986 TARGET_VALID_POINTER_MODE.
8987
0bd62dca 89882013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 8989 Pat Haugen <pthaugen@us.ibm.com>
8990 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM		 8991
8992 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8993 Document new power8 builtins.
8994
8995 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
8996 condition code register, to allow 128-bit logical operations to be
8997 done in the VSX or GPR registers.
8998 (nor<mode>3): Use the canonical form for nor.
8999 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9000 vclz*, and vpopcnt* vector instructions.
9001 (nand<mode>3): Likewise.
9002 (orc<mode>3): Likewise.
9003 (clz<mode>2): LIkewise.
9004 (popcount<mode>2): Likewise.
9005
9006 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9007 that only the GPRs are recognized.
9008
9009 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9010 support for new power8 builtins.
9011
9012 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9013 builtin functions.
9014 (xscvdpspn): Likewise.
9015 (vclz): Likewise.
9016 (vclzb): Likewise.
9017 (vclzh): Likewise.
9018 (vclzw): Likewise.
9019 (vclzd): Likewise.
9020 (vpopcnt): Likewise.
9021 (vpopcntb): Likewise.
9022 (vpopcnth): Likewise.
9023 (vpopcntw): Likewise.
9024 (vpopcntd): Likewise.
9025 (vgbbd): Likewise.
9026 (vmrgew): Likewise.
9027 (vmrgow): Likewise.
9028 (eqv): Likewise.
9029 (eqv_v16qi3): Likewise.
9030 (eqv_v8hi3): Likewise.
9031 (eqv_v4si3): Likewise.
9032 (eqv_v2di3): Likewise.
9033 (eqv_v4sf3): Likewise.
9034 (eqv_v2df3): Likewise.
9035 (nand): Likewise.
9036 (nand_v16qi3): Likewise.
9037 (nand_v8hi3): Likewise.
9038 (nand_v4si3): Likewise.
9039 (nand_v2di3): Likewise.
9040 (nand_v4sf3): Likewise.
9041 (nand_v2df3): Likewise.
9042 (orc): Likewise.
9043 (orc_v16qi3): Likewise.
9044 (orc_v8hi3): Likewise.
9045 (orc_v4si3): Likewise.
9046 (orc_v2di3): Likewise.
9047 (orc_v4sf3): Likewise.
9048 (orc_v2df3): Likewise.
9049
9050 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9051 allow power8 quad mode in 64-bit.
9052 (rs6000_builtin_vectorized_function): Add support to vectorize
9053 ISA 2.07 count leading zeros, population count builtins.
9054 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9055 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9056 (builtin_function_type): Add vgbbd builtin function which takes an
9057 unsigned argument.
9058 (altivec_expand_vec_perm_const): Add support for new power8 merge
9059 instructions.
9060
9061 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9062 that does not include TImdoe for use with 32-bit.
9063 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9064 instructions.
9065 (UNSPEC_VSX_CVDPSPN): Likewise.
9066 (vsx_xscvdpspn): Likewise.
9067 (vsx_xscvspdpn): Likewise.
9068 (vsx_xscvdpspn_scalar): Likewise.
9069 (vsx_xscvspdpn_directmove): Likewise.
9070 (vsx_and<mode>3): Split logical operations into 32-bit and
9071 64-bit. Add support to do logical operations on TImode as well as
9072 VSX vector types. Allow logical operations to be done in either
9073 VSX registers or in general purpose registers in 64-bit mode. Add
9074 splitters if GPRs were used. For AND, add clobber of CCmode to
9075 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9076 encoding.
9077 (vsx_and<mode>3_32bit): Likewise.
9078 (vsx_and<mode>3_64bit): Likewise.
9079 (vsx_ior<mode>3): Likewise.
9080 (vsx_ior<mode>3_32bit): Likewise.
9081 (vsx_ior<mode>3_64bit): Likewise.
9082 (vsx_xor<mode>3): Likewise.
9083 (vsx_xor<mode>3_32bit): Likewise.
9084 (vsx_xor<mode>3_64bit): Likewise.
9085 (vsx_one_cmpl<mode>2): Likewise.
9086 (vsx_one_cmpl<mode>2_32bit): Likewise.
9087 (vsx_one_cmpl<mode>2_64bit): Likewise.
9088 (vsx_nor<mode>3): Likewise.
9089 (vsx_nor<mode>3_32bit): Likewise.
9090 (vsx_nor<mode>3_64bit): Likewise.
9091 (vsx_andc<mode>3): Likewise.
9092 (vsx_andc<mode>3_32bit): Likewise.
9093 (vsx_andc<mode>3_64bit): Likewise.
9094 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
9095 and xxlorc instructions.
9096 (vsx_eqv<mode>3_64bit): Likewise.
9097 (vsx_nand<mode>3_32bit): Likewise.
9098 (vsx_nand<mode>3_64bit): Likewise.
9099 (vsx_orc<mode>3_32bit): Likewise.
9100 (vsx_orc<mode>3_64bit): Likewise.
9101
9102 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
9103
9104 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
9105 instruction.
9106 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
9107 (p8_vmrgow): Likewise.
9108 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
9109 GPRs to be split under VSX.
9110 (p8v_clz<mode>2): Add power8 count leading zero support.
9111 (p8v_popcount<mode>2): Add power8 population count support.
9112 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
9113 support.
9114
9115 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
9116 instruction.
9117
9118 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
9119 builtin functions.
9120 (vec_nand): Likewise.
9121 (vec_vclz): Likewise.
9122 (vec_vclzb): Likewise.
9123 (vec_vclzd): Likewise.
9124 (vec_vclzh): Likewise.
9125 (vec_vclzw): Likewise.
9126 (vec_vgbbd): Likewise.
9127 (vec_vmrgew): Likewise.
9128 (vec_vmrgow): Likewise.
9129 (vec_vpopcnt): Likewise.
9130 (vec_vpopcntb): Likewise.
9131 (vec_vpopcntd): Likewise.
9132 (vec_vpopcnth): Likewise.
9133 (vec_vpopcntw): Likewise.
9134
37684c46
VM		 91352013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9136
9137 PR rtl-optimization/57468
9138 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
9139 spilled pseudos.
9140
d89ae27c
VM		 91412013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9142
9143 PR rtl-optimization/57459
9144 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
9145 type when setting live regs.
9146
3597e113
VM		 91472013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9148
9149 * config/s390/s390.opt (mlra): New option.
9150 * config/s390/s390.c (s390_decompose_address): Check displacement
9151 for all registers for LRA.
9152 (s390_secondary_reload): Don't used secondary reloads for LRA.
9153 (s390_lra_p): New function.
9154 (TARGET_LRA_P): Define.
9155 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
9156 of attribute cpu_facility to zarch for the last alternative.
9157 (*cmpmem_short): Ditto.
9158
01037aeb
EB		 91592013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9160
9161 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9162 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9163 (arm_expand_prologue): Likewise.
9164
3371a64f
TJ		 91652013-06-06 Teresa Johnson <tejohnson@google.com>
9166
9167 PR c++/53743
9168 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9169 as this is now done by redirect_edge_and_branch_force.
9170 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9171 barriers, and fix interaction with splitting.
9172 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9173 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9174 reflect changes made in the routine.
9175 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9176 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9177 since this is called in cfglayout mode, and replace partition fixup
9178 with assert as that is now done by force_nonfallthru_and_redirect.
9179 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9180 already be marked with region crossing note.
9181 (insert_section_boundary_note): Make non-static, gate on flag
9182 has_bb_partition, rewrite to also check for multiple partitions.
9183 (rest_of_handle_reorder_blocks): Remove call to
9184 insert_section_boundary_note, now done later during free_cfg.
9185 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9186 * bb-reorder.h (insert_section_boundary_note): Declare.
9187 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9188 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9189 invoke insert_section_boundary_note.
9190 (try_redirect_by_replacing_jump): Remove unnecessary
9191 check for region crossing note.
9192 (fixup_partition_crossing): New function.
9193 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9194 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9195 in non-cfglayout mode.
9196 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9197 remove old code that tried to do this. Emit barrier correctly
9198 when we are in cfglayout mode.
9199 (last_bb_in_partition): New function.
9200 (rtl_split_edge): Correctly fixup partition boundaries.
9201 (commit_one_edge_insertion): Remove old code that tried to
9202 fixup region crossing edge since this is now handled in
9203 split_block, and set up insertion point correctly since
9204 block may now end in a jump.
9205 (verify_hot_cold_block_grouping): Guard against checking when not in
9206 linearized RTL mode.
9207 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9208 notes.
9209 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9210 rtl_verify_flow_info, so not called in cfglayout mode.
9211 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9212 (fixup_reorder_chain): Remove old code that attempted to fixup region
9213 crossing note as this is now handled in force_nonfallthru_and_redirect.
9214 (duplicate_insn_chain): Don't duplicate switch section notes.
9215 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9216 note.
9217 * basic-block.h (emit_barrier_after_bb): Declare.
9218
66071e10
KT		 92192013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9220
9221 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9222 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9223 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9224
0e26bf3d
KT		 92252013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9226
9227 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9228 where appropriate.
9229 * config/arm/ldmstm.md: Regenerate.
9230
12b4e7ef
KT		 92312013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9232
9233 * config/arm/sync.md (atomic_loaddi_1):
9234 Disable predication for arm_restrict_it.
9235 (arm_load_exclusive<mode>): Likewise.
9236 (arm_load_exclusivesi): Likewise.
9237 (arm_load_exclusivedi): Likewise.
9238 (arm_load_acquire_exclusive<mode>): Likewise.
9239 (arm_load_acquire_exclusivesi): Likewise.
9240 (arm_load_acquire_exclusivedi): Likewise.
9241 (arm_store_exclusive<mode>): Likewise.
9242 (arm_store_exclusive<mode>): Likewise.
9243 (arm_store_release_exclusivedi): Likewise.
9244 (arm_store_release_exclusive<mode>): Likewise.
9245
15d16c8a
RB		 92462013-06-06 Richard Biener <rguenther@suse.de>
9247
9248 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9249 after LTO_null.
9250 (lto_tag_is_tree_code_p): Adjust.
9251 (lto_tag_is_gimple_code_p): Likewise.
9252 (lto_gimple_code_to_tag): Likewise.
9253 (lto_tag_to_gimple_code): Likewise.
9254 (lto_tree_code_to_tag): Likewise.
9255 (lto_tag_to_tree_code): Likewise.
9256 * data-streamer.h (streamer_write_hwi_in_range): Use
9257 uhwi streaming to stream the normalized range.
9258 (streamer_read_hwi_in_range): Likewise.
9259
17a7fc37
KT		 92602013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9261
9262 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9263 (predicable_short_it): Likewise.
9264 (predicated): Likewise.
9265 (enabled): Handle above.
9266 (define_cond_exec): Set predicated attribute to yes.
9267
b57ca59b
MS		 92682013-06-05 Mike Stump <mikestump@comcast.net>
9269
9270 * gdbinit.in (__FUNCTION__): Add.
9271
c1e183a9
UB		 92722013-06-05 Uros Bizjak <ubizjak@gmail.com>
9273
9274 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9275 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9276
0378bddb
JH		 92772013-06-05 Jan Hubicka <jh@suse.cz>
9278
9279 * varasm.c (mark_decl_referenced): Revert the removal until targets
9280 are fixed.
9281
8f8a2057
DE		 92822013-06-05 David Edelsohn <dje.gcc@gmail.com>
9283
9284 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9285 instead of mark_decl_referenced.
9286
edb983b2
JH		 92872013-06-05 Jan Hubicka <jh@suse.cz>
9288
9289 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9290 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9291 and symtab_used_from_object_file_p.
9292 (cgraph_make_node_local_1): Clear forced_by_abi.
9293 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9294 * cgraph.h (symtab_node_base): Add forced_by_abi.
9295 (decide_is_variable_needed): Remove.
9296 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9297 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9298 (decide_is_symbol_needed): ... this one; handle symbols in general;
9299 always analyze virtuals; honnor forced_by_abi.
9300 (cgraph_finalize_function): Update.
9301 (varpool_finalize_decl): Update.
9302 (symbol_defined_and_needed): Remove.
9303 (analyze_functions): Update.
9304 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9305 output_refs, input_overwrite_node): Handle forced_by_abi.
9306 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9307 (address_taken_from_non_vtable_p): ... this one.
9308 (comdat_can_be_unshared_p_1): New function.
9309 (cgraph_comdat_can_be_unshared_p): Rename to ...
9310 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9311 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9312 (function_and_variable_visibility): Clear forced_by_abi as needed.
9313 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9314 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH		 9315 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9316 * varpool.c (decide_is_variable_needed): Remove.
9317
9912dbe5
KT		 93182013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9319
9320 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9321 (arm_option_override): Override arm_restrict_it where appropriate.
9322 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9323 * config/arm/arm.opt (mrestrict-it): New command-line option.
9324 * doc/invoke.texi: Document -mrestrict-it.
9325
34c136b6
DM		 93262013-06-05 David Malcolm <dmalcolm@redhat.com>
9327
9328 * tsan.c (tsan_atomic_table): Make const.
9329
a367df53
RB		 93302013-06-05 Richard Biener <rguenther@suse.de>
9331
9332 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9333 index associated with the tree we are supposed to replace.
9334 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9335 TREE_ASM_WRITTEN as zero for everything but SSA names.
9336
70d51a19
DM		 93372013-06-05 David Malcolm <dmalcolm@redhat.com>
9338
9339 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9340
b9bd2075
JH		 93412013-06-04 Jan Hubicka <jh@suse.cz>
9342
9343 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9344 (input_node, input_varpool_node): Handle correctly external same
9345 body aliases.
9346 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9347 nodes at ltrans stage.
9348
aaae719d
JH		 93492013-06-04 Jan Hubicka <jh@suse.cz>
9350
9351 * ipa-inline.c (update_caller_keys): Fix availability test.
9352 (update_callee_keys): Likewise.
9353 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
9354 to follow ELF standard.
9355
107eea2c
JU		 93562013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
9357
9358 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
9359 (mips64r5900el-*-elf*): New configurations.
9360 * config/mips/mips-cpus.def (r5900): New processor.
9361 * config/mips/mips-tables.opt: Regenerate.
9362 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
9363 (mips_issue_rate): Handle PROCESSOR_R5900.
9364 (mips_reorg_process_insns): Force reorder mode for the R5900.
9365 * config/mips/mips.h (TARGET_MIPS5900): Define.
9366 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
9367 TARGET_MIPS5900.
9368 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
9369 TARGET_MIPS5900.
9370 * config/mips/mips.md (processor): Add r5900.
9371 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
9372
b7342d25
IB		 93732013-06-04 Ian Bolton <ian.bolton@arm.com>
9374
9375 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
9376 into function to generate MOVI instruction.
c1e183a9 9377 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB		 9378 (aarch64_preferred_simd_mode): Turn into wrapper.
9379 (aarch64_output_scalar_simd_mov_immediate): New function.
9380 * config/aarch64/aarch64-protos.h: Add prototype for above.
9381
81c2dfb9
IB		 93822013-06-04 Ian Bolton <ian.bolton@arm.com>
9383
9384 * config/aarch64/aarch64.c (simd_immediate_info): Remove
9385 element_char member.
9386 (sizetochar): Return signed char.
9387 (aarch64_simd_valid_immediate): Remove elchar and other
9388 unnecessary variables.
9389 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
9390 Calculate element_char as required.
9391 * config/aarch64/aarch64-protos.h: Update and move prototype
9392 for aarch64_output_simd_mov_immediate.
9393 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
9394 Update arguments.
9395
48063b9d
IB		 93962013-06-04 Ian Bolton <ian.bolton@arm.com>
9397
9398 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
9399 information completed by aarch64_simd_valid_immediate.
9400 (aarch64_legitimate_constant_p): Update arguments.
9401 (aarch64_simd_valid_immediate): Work with struct rather than many
9402 pointers.
9403 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
9404 (aarch64_simd_make_constant): Update arguments.
9405 (aarch64_output_simd_mov_immediate): Work with struct rather than
9406 many pointers. Output immediate directly rather than as operand.
9407 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9408 Update prototype.
9409 * config/aarch64/constraints.md (Dn): Update arguments.
9410
3ea63f60
IB		 94112013-06-04 Ian Bolton <ian.bolton@arm.com>
9412
9413 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
9414 longer static.
9415 (aarch64_simd_immediate_valid_for_move): Remove.
9416 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
9417 (aarch64_simd_make_constant): Update call.
9418 (aarch64_output_simd_mov_immediate): Update call.
9419 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9420 Add prototype.
9421 * config/aarch64/constraints.md (Dn): Update call.
9422
d8edd899
IB		 94232013-06-04 Ian Bolton <ian.bolton@arm.com>
9424
9425 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
9426 return type to bool for prototype.
9427 (aarch64_legitimate_constant_p): Check for true instead of not -1.
9428 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
9429 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
9430
44b20bb8
CM		 94312013-06-04 Catherine Moore <clm@codesourcery.com>
9432
9433 * config/mips/mips.opt (meva): New.
c1e183a9 9434 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM		 9435 (ASM_SPEC): Handle -meva.
9436 * doc/invoke.texi (meva): Document.
9437
52befbd8
AM		 94382013-06-04 Alan Modra <amodra@gmail.com>
9439
9440 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
9441 constant output.
9442
aadaf24e
KT		 94432013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9444
9445 * rtl.def: Add extra fourth optional field to define_cond_exec.
9446 * gensupport.c (process_one_cond_exec): Process attributes from
9447 define_cond_exec.
9448 * doc/md.texi: Document fourth field in define_cond_exec.
9449
0cd9e9ee
EB		 94502013-06-04 Eric Botcazou <ebotcazou@adacore.com>
9451
9452 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
9453 out the processing order as in store_bit_field_1.
9454
cb0f665d
JH		 94552013-06-04 Jan Hubicka <jh@suse.cz>
9456
9457 PR middle-end/57500
9458 * cgraphunit.c (cgraph_process_same_body_aliases): Create
9459 non-VAR_DECL node if it does not exist yet.
9460
53984b9b
RS		 94612013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
9462
9463 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
9464 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
9465 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
9466 target_cpu_default setting.
9467
5d318fd4
TJ		 94682013-06-03 Teresa Johnson <tejohnson@google.com>
9469
9470 * dumpfile.c (opt_info_switch_p): Change -fopt-info
9471 default to -fopt-info=optimized instead of all.
9472 * doc/invoke.texi: Ditto.
9473 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
9474 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
9475 (execute_vect_slp): Emit BB vectorization success under
9476 MSG_OPTIMIZED_LOCATIONS.
9477 * tree-vect-slp.c (vect_slp_transform_bb): Change
9478 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
9479 * tree-vect-loop.c (vect_transform_loop): Ditto.
9480
ec3af349
JM		 94812013-06-03 Jason Merrill <jason@redhat.com>
9482
9483 PR c++/57415
9484 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9485 Use TARGET_EXPR for C++.
9486
1baf9159
JJ		 94872013-06-03 Jakub Jelinek <jakub@redhat.com>
9488
9489 PR rtl-optimization/57268
0cd9e9ee 9490 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ		 9491 if DEBUG_INSN_P (insn).
9492
9493 Reapply
9494 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9495
9496 PR rtl-optimization/57268
9497 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9498 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9499
0f1d3965
YR		 95002013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9501
9502 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
9503 (ix86_avoid_lea_for_addr): Likewise.
9504 (exact_dependency_1): Likewise.
9505 (ix86_adjust_cost): Likewise.
9506 (swap_top_of_ready_list): Fix formatting and !reload_completed check
9507 removed.
9508 (do_reorder_for_imul): Fix typo, formatting and
9509 !reload_completed check removed.
9510 (ix86_sched_reorder): Fix typo and formatting.
9511 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
9512 list.
9513
3bc30db4
SN		 95142013-06-03 Sofiane Naci <sofiane.naci@arm.com>
9515
9516 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
9517
19cc6fac
EB		 95182013-06-03 Eric Botcazou <ebotcazou@adacore.com>
9519
9520 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
9521 <STRING_CST>: Likewise.
9522 <VECTOR_CST>: Likewise.
9523
c68a6e08
JW		 95242013-06-01 Janus Weil <janus@gcc.gnu.org>
9525 Mikael Morin <mikael@gcc.gnu.org>
9526
9527 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
9528 * config.in: Regenerated.
9529 * configure: Regenerated.
9530
38e55e5c
JH		 95312013-06-01 Jan Hubicka <jh@suse.cz>
9532
9533 PR middle-end/57366
9534 * cgraphunit.c (compile): When weakref is not supported,
9535 set up transparent aliases before final output pass.
9536 * varasm.c (assemble_alias): Do not try to do it here.
9537
eb51d2ff
JH		 95382013-06-01 Jan Hubicka <jh@suse.cz>
9539
9540 PR middle-end/57467
9541 * passes.c (for_per_function): Skip unanalyzed functions.
9542
40a7fe1e
JH		 95432013-06-01 Jan Hubicka <jh@suse.cz>
9544
9545 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
9546 (lto_symtab_merge_symbols_1): ... this one.
9547 (lto_symtab_merge_cgraph_nodes): Rename to ...
9548 (lto_symtab_merge_symbols): ... this one; simplify.
9549 * cgraph.c (same_body_aliases_done): Rename to ...
9550 (cpp_implicit_aliases_done): ... this one.
9551 (cgraph_create_function_alias): Update.
9552 (cgraph_same_body_alias): Update.
c1e183a9 9553 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 9554 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 9555 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH		 9556 (cgraph_node): Remove same_body_alias.
9557 (varpool_node): Remove alias_of and extra_name_alias.
9558 (same_body_aliases_done): Rename to ..
9559 (cpp_implicit_aliases_done): ... this one.
9560 (symtab_alias_ultimate_target): Add default parameter.
9561 (symtab_resolve_alias): New function.
9562 (fixup_same_cpp_alias_visibility): Declare.
9563 (cgraph_function_node): Add default parameter.
9564 (cgraph_node_asm_name): Likewise.
9565 (cgraph_function_or_thunk_node): Add default parameter; do
9566 not ICE when it is NULL.
9567 (varpool_variable_node): Likewise.
9568 * tree-emutls.c (create_emultls_var): Update.
9569 (ipa_lower_emutls): Update.
9570 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
9571 (cgraph_reset_node): Reset alias info.
9572 (cgraph_finalize_function): Update.
9573 (fixup_same_cpp_alias_visibility): Move to symtab.c.
9574 (analyze_function): Simplify.
9575 (cgraph_process_same_body_aliases): Simplify.
9576 (analyze_functions): Fixup same body aliases.
9577 (handle_alias_pairs): Simplify.
9578 (assemble_thunk): Update.
9579 (assemble_thunks_and_aliases): Update.
9580 (output_weakrefs): Rewrite.
9581 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
9582 (lto_output_varpool_node): Likewise.
9583 (compute_ltrans_boundary): Remve assert.
9584 (get_alias_symbol): New functoin.
9585 (input_node): Rewrite alias handling.
9586 (input_varpool_node): Likewise.
9587 * ipa-pure-const.c (propagate_pure_const): Fix formating.
9588 * ipa.c (process_references): Handle weakrefs correctly.
9589 (symtab_remove_unreachable_nodes): Likewise.
9590 * trans-mem.c (get_cg_data): Update.
9591 (ipa_tm_create_version_alias): Update.
9592 (ipa_tm_execute): Update.
9593 * symtab.c (dump_symtab_base): Dump aliases.
9594 (verify_symtab_base): Verify aliases.
9595 (symtab_node_availability): New function.
9596 (symtab_alias_ultimate_target): Simplify.
9597 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
9598 handle all the fixup cases.
9599 (symtab_resolve_alias): New function.
9600 * passes.c (ipa_write_summaries): Handle weakrefs.
9601 * varpool.c (varpool_analyze_node): Simplify.
9602 (assemble_aliases): Update.
9603 (varpool_create_variable_alias): Simplify.
9604 (varpool_extra_name_alias): Simplify.
9605 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
9606 (lto_symtab_merge_symbols): ... this one.
9607
21759881
DT		 96082013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
9609
9610 Revert
9611 PR rtl-optimization/57268
9612 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9613 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9614
daf4e940
TB		 96152013-06-01 Tobias Burnus <burnus@net-b.de>
9616
9617 Partially reverted:
9618 2013-05-31 Tobias Burnus <burnus@net-b.de>
9619
9620 PR middle-end/57073
9621 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
9622 further up.
9623
c359d8d0
DT		 96242013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9625
9626 PR rtl-optimization/57268
9627 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9628 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9629
4172215d
EB		 96302013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9631
9632 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
9633 unordered comparison operators when -fno-trapping-math is in effect
9634 on the e500.
9635 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
9636 and implement unordered comparison operators properly on the e500.
9637
a8c50132
EB		 96382013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9639
9640 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
9641 for constant scalar integers.
9642 (simplify_relational_operation_1): Likewise.
9643
44626634
SB		 96442013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
9645
9646 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
9647 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
9648 Fix comment.
9649
55805e54
YR		 96502013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9651 Igor Zamyatin <igor.zamyatin@intel.com>
9652
9653 Silvermont (SLM) architecture performance tuning.
9654 * config/i386/i386.h (enum ix86_tune_indices): Add
9655 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
9656 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
9657
9658 * config/i386/i386.c (initial_ix86_tune_features)
9659 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
9660 (ix86_lea_outperforms): Handle Silvermont tuning.
9661 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
9662 call.
9663 (ix86_use_lea_for_mov): Likewise.
9664 (ix86_avoid_lea_for_addr): Likewise.
9665 (ix86_lea_for_add_ok): Likewise.
9666 (exact_dependency_1): New function.
9667 (exact_store_load_dependency): Likewise.
9668 (ix86_adjust_cost): Handle Silvermont tuning.
9669 (do_reoder_for_imul): Likewise.
9670 (swap_top_of_ready_list): New function.
9671 (ix86_sched_reorder): Changed to handle Silvermont tuning.
9672
9673 * config/i386/i386.md (peepholes that split memory operand in fp
9674 converts): New.
9675
e19c9de2
MS		 96762013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9677
9678 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
9679 Remove un-necessary braces.
9680
38e6c9a6
MS		 96812013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9682
9683 * config/aarch64/aarch64.c (aarch64_classify_symbol):
9684 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
9685
e3530904
TB		 96862013-05-31 Tobias Burnus <burnus@net-b.de>
9687
9688 PR middle-end/57073
c1e183a9 9689 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 9690
b6af05a9
KT		 96912013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9692
9693 PR target/56315
9694 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
9695 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
9696 * config/arm/neon.md (iordi3_neon): Remove.
9697 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
9698 * config/arm/predicates.md (imm_for_neon_logic_operand):
9699 Move to earlier in the file.
9700 (neon_logic_op2): Likewise.
9701 (arm_iordi_operand_neon): New predicate.
9702
f800c166
RB		 97032013-05-31 Richard Biener <rguenther@suse.de>
9704
9705 PR tree-optimization/57478
9706 PR tree-optimization/57453
9707 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
9708 are life as well.
9709
e571fa59
KP		 97102013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9711
9712 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
9713 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
9714
0fa6e0ef
TB		 97152013-05-30 Tobias Burnus <burnus@net-b.de>
9716 Thomas Koenig <tkoenig@gcc.gnu.org>
9717
9718 PR middle-end/57073
9719 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
9720 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
9721
4688ddf5
SB		 97222013-05-30 Steven Bosscher <steven@gcc.gnu.org>
9723
0fa6e0ef 9724 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 9725
3b9ceb4b
VM		 97262013-05-30 Vladimir Makarov <vmakarov@redhat.com>
9727
9728 * target.def (register_usage_leveling_p): New hook.
9729 * targhooks.c (default_register_usage_leveling_p): New.
9730 * targhooks.h (default_register_usage_leveling_p): New prototype.
9731 * lra-assigns.c (register_usage_leveling_p): Use the hook.
9732 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
9733 * doc/tm.texi: Update.
9734 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
9735
65f9e789
IB		 97362013-05-30 Ian Bolton <ian.bolton@arm.com>
9737
9738 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
9739 (*insv_reg<mode>): New define_insn.
9740
12211b99 97412013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR		 9742
9743 PR rtl-optimization/57439
9744 * postreload.c (move2add_valid_value_p): Check that we have
9745 a zero subreg_regno_offset when accessing the register in
9746 the requested mode.
9747
0b871ccf
YR		 97482013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9749 Igor Zamyatin <igor.zamyatin@intel.com>
9750
9751 Silvermont (SLM) architecture pipeline model, tuning and
9752 insn selection.
9753 * config.gcc: Add slm config options and target.
9754
9755 * config/i386/slm.md: New.
9756
9757 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
9758
9759 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
9760 PROCESSOR_SLM.
9761 (ix86_target_macros_internal): Likewise.
9762
9763 * gcc/config/i386/i386.c (slm_cost): New cost.
9764 (m_SLM): New macro flag.
9765 (initial_ix86_tune_features): Set m_SLM.
9766 (x86_accumulate_outgoing_args): Likewise.
9767 (x86_arch_always_fancy_math_387): Likewise.
9768 (processor_target_table): Add slm cost.
9769 (cpu_names): Add slm cpu name.
9770 (x86_option_override_internal): Set SLM ISA.
9771 (ix86_issue_rate): New case PROCESSOR_SLM.
9772 (ia32_multipass_dfa_lookahead): Likewise.
9773 (fold_builtin_cpu): Add slm.
9774
9775 * config/i386/i386.h (TARGET_SLM): New target macro.
9776 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
9777 (processor_type): Add PROCESSOR_SLM.
9778
9779 * config/i386/i386.md (cpu): Add new value "slm".
9780 (slm.md): Include slm.md.
9781
24d5b097
XG		 97822013-05-30 Bernd Schmidt <bernds@codesourcery.com>
9783 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9784
9785 * config/arm/arm-protos.h: Add and update function protos.
9786 * config/arm/arm.c (use_simple_return_p): New added.
9787 (thumb2_expand_return): Check simple_return flag.
9788 * config/arm/arm.md: Add simple_return and conditional simple_return.
9789 * config/arm/iterators.md: Add iterator for return and simple_return.
9790
c1cccc15
ZC		 97912013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9792
9793 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
9794 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
9795 (arm_emit_vfp_multi_reg_pop): Likewise.
9796 (thumb2_emit_ldrd_pop): Likewise.
9797 (arm_expand_epilogue): Add misc REG_CFA notes.
9798 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
9799
3a4f280b
LC		 98002013-05-29 Lawrence Crowl <crowl@google.com>
9801
9802 * config/arm/t-arm: Update for below.
9803
9804 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
9805 Change type to hash_table. Update dependent calls and types.
9806
9807 * config/i386/t-cygming: Update for below.
9808
9809 * config/i386/t-interix: Update for below.
9810
9811 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
9812 Change type to hash_table. Update dependent calls and types.
9813 (i386_find_on_wrapper_list::wrappers): Likewise.
9814
9815 * config/ia64/t-ia64: Update for below.
9816
9817 * config/ia64/ia64.c (bundle_state_table):
9818 Change type to hash_table. Update dependent calls and types.
9819
9820 * config/mips/mips.c (mips_reorg_process_insns::htab):
9821 Change type to hash_table. Update dependent calls and types.
9822
9823 * config/sol2.c (solaris_comdat_htab):
9824 Change type to hash_table. Update dependent calls and types.
9825
9826 * config/t-sol2: Update for above.
9827
1388a0e3
TJ		 98282013-05-29 Teresa Johnson <tejohnson@google.com>
9829
9830 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
9831 functions are not yet marked as defined.
9832
a5965b52 98332013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 9834 Pat Haugen <pthaugen@us.ibm.com>
9835 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM		 9836
9837 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
9838 instructions.
9839 (VEC_A): Likewise.
9840 (VEC_C): Likewise.
9841 (vrotl<mode>3): Likewise.
9842 (vashl<mode>3): Likewise.
9843 (vlshr<mode>3): Likewise.
9844 (vashr<mode>3): Likewise.
9845
9846 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9847 support for power8 V2DI builtins.
9848
9849 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
9850 power8 V2DI builtins.
9851 (vupkhsw): Likewise.
9852 (vupklsw): Likewise.
9853 (vaddudm): Likewise.
9854 (vminsd): Likewise.
9855 (vmaxsd): Likewise.
9856 (vminud): Likewise.
9857 (vmaxud): Likewise.
9858 (vpkudum): Likewise.
9859 (vpksdss): Likewise.
9860 (vpkudus): Likewise.
9861 (vpksdus): Likewise.
9862 (vrld): Likewise.
9863 (vsld): Likewise.
9864 (vsrd): Likewise.
9865 (vsrad): Likewise.
9866 (vsubudm): Likewise.
9867 (vcmpequd): Likewise.
9868 (vcmpgtsd): Likewise.
9869 (vcmpgtud): Likewise.
9870 (vcmpequd_p): Likewise.
9871 (vcmpgtsd_p): Likewise.
9872 (vcmpgtud_p): Likewise.
9873 (vupkhsw): Likewise.
9874 (vupklsw): Likewise.
9875 (vaddudm): Likewise.
9876 (vmaxsd): Likewise.
9877 (vmaxud): Likewise.
9878 (vminsd): Likewise.
9879 (vminud): Likewise.
9880 (vpksdss): Likewise.
9881 (vpksdus): Likewise.
9882 (vpkudum): Likewise.
9883 (vpkudus): Likewise.
9884 (vrld): Likewise.
9885 (vsld): Likewise.
9886 (vsrad): Likewise.
9887 (vsrd): Likewise.
9888 (vsubudm): Likewise.
9889
9890 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
9891 support for power8 V2DI instructions.
9892
9893 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
9894 power8 V2DI instructions. Combine pack and unpack insns to use an
9895 iterator for each mode. Check whether a particular mode supports
9896 Altivec instructions instead of just checking TARGET_ALTIVEC.
9897 (UNSPEC_VPKUWUM): Likewise.
9898 (UNSPEC_VPKSHSS): Likewise.
9899 (UNSPEC_VPKSWSS): Likewise.
9900 (UNSPEC_VPKUHUS): Likewise.
9901 (UNSPEC_VPKSHUS): Likewise.
9902 (UNSPEC_VPKUWUS): Likewise.
9903 (UNSPEC_VPKSWUS): Likewise.
9904 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
9905 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
9906 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
9907 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
9908 (UNSPEC_VUPKHSB): Likewise.
9909 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
9910 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
9911 (UNSPEC_VUPKHSH): Likewise.
9912 (UNSPEC_VUPKLSB): Likewise.
9913 (UNSPEC_VUPKLSH): Likewise.
9914 (VI2): Likewise.
9915 (VI_char): Likewise.
9916 (VI_scalar): Likewise.
9917 (VI_unit): Likewise.
9918 (VP): Likewise.
9919 (VP_small): Likewise.
9920 (VP_small_lc): Likewise.
9921 (VU_char): Likewise.
9922 (add<mode>3): Likewise.
9923 (altivec_vaddcuw): Likewise.
9924 (altivec_vaddu<VI_char>s): Likewise.
9925 (altivec_vadds<VI_char>s): Likewise.
9926 (sub<mode>3): Likewise.
9927 (altivec_vsubcuw): Likewise.
9928 (altivec_vsubu<VI_char>s): Likewise.
9929 (altivec_vsubs<VI_char>s): Likewise.
9930 (altivec_vavgs<VI_char>): Likewise.
9931 (altivec_vcmpbfp): Likewise.
9932 (altivec_eq<mode>): Likewise.
9933 (altivec_gt<mode>): Likewise.
9934 (altivec_gtu<mode>): Likewise.
9935 (umax<mode>3): Likewise.
9936 (smax<mode>3): Likewise.
9937 (umin<mode>3): Likewise.
9938 (smin<mode>3): Likewise.
9939 (altivec_vpkuhum): Likewise.
9940 (altivec_vpkuwum): Likewise.
9941 (altivec_vpkshss): Likewise.
9942 (altivec_vpkswss): Likewise.
9943 (altivec_vpkuhus): Likewise.
9944 (altivec_vpkshus): Likewise.
9945 (altivec_vpkuwus): Likewise.
9946 (altivec_vpkswus): Likewise.
9947 (altivec_vpks<VI_char>ss): Likewise.
9948 (altivec_vpks<VI_char>us): Likewise.
9949 (altivec_vpku<VI_char>us): Likewise.
9950 (altivec_vpku<VI_char>um): Likewise.
9951 (altivec_vrl<VI_char>): Likewise.
9952 (altivec_vsl<VI_char>): Likewise.
9953 (altivec_vsr<VI_char>): Likewise.
9954 (altivec_vsra<VI_char>): Likewise.
9955 (altivec_vsldoi_<mode>): Likewise.
9956 (altivec_vupkhsb): Likewise.
9957 (altivec_vupkhs<VU_char>): Likewise.
9958 (altivec_vupkls<VU_char>): Likewise.
9959 (altivec_vupkhsh): Likewise.
9960 (altivec_vupklsb): Likewise.
9961 (altivec_vupklsh): Likewise.
9962 (altivec_vcmpequ<VI_char>_p): Likewise.
9963 (altivec_vcmpgts<VI_char>_p): Likewise.
9964 (altivec_vcmpgtu<VI_char>_p): Likewise.
9965 (abs<mode>2): Likewise.
9966 (vec_unpacks_hi_v16qi): Likewise.
9967 (vec_unpacks_hi_v8hi): Likewise.
9968 (vec_unpacks_lo_v16qi): Likewise.
9969 (vec_unpacks_hi_<VP_small_lc>): Likewise.
9970 (vec_unpacks_lo_v8hi): Likewise.
9971 (vec_unpacks_lo_<VP_small_lc>): Likewise.
9972 (vec_pack_trunc_v8h): Likewise.
9973 (vec_pack_trunc_v4si): Likewise.
9974 (vec_pack_trunc_<mode>): Likewise.
9975
9976 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
9977 V2DI builtins.
9978 (vec_vmaxsd): Likewise.
9979 (vec_vmaxud): Likewise.
9980 (vec_vminsd): Likewise.
9981 (vec_vminud): Likewise.
9982 (vec_vpksdss): Likewise.
9983 (vec_vpksdus): Likewise.
9984 (vec_vpkudum): Likewise.
9985 (vec_vpkudus): Likewise.
9986 (vec_vrld): Likewise.
9987 (vec_vsld): Likewise.
9988 (vec_vsrad): Likewise.
9989 (vec_vsrd): Likewise.
9990 (vec_vsubudm): Likewise.
9991 (vec_vupkhsw): Likewise.
9992 (vec_vupklsw): Likewise.
9993
e70670cf
JH		 99942013-05-29 Jan Hubicka <jh@suse.cz>
9995
9996 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
9997 flags; reorder rest of fields in more consistent way.
9998 (varpool_node): Remove analyzed, finalized and alias.
9999 (cgraph_ndoe): Likewise.
10000 (symtab_alias_ultimate_target): New function.
10001 (cgraph_function_node): Move offline.
10002 (cgraph_reset_node): Declare.
10003 (cgraph_comdat_can_be_unshared_p): Remove.
10004 (varpool_remove_initializer): Declare.
10005 (varpool_first_defined_variable, varpool_next_defined_variable
10006 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10007 (cgraph_function_with_gimple_body_p): Update.
10008 (varpool_all_refs_explicit_p): Update.
10009 (symtab_alias_target): New function.
10010 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10011 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB		 10012 (cgraph_function_or_thunk_node): Simplify using
10013 symtab_alias_ultimate_target.
e70670cf
JH		 10014 (varpool_variable_node): Likewise.
10015 * cgraph.c (cgraph_create_function_alias): Update.
10016 (cgraph_add_thunk): Update.
10017 (cgraph_remove_node): Update.
10018 (dump_cgraph_node): Do not dump removed flags.
10019 (cgraph_function_body_availability): Update.
10020 (cgraph_propagate_frequency): Update.
10021 (verify_cgraph_node): Check sanity of local flag.
10022 (cgraph_function_node): Move here from cgraph.h; revamp for
10023 cgraph_function_or_thunk_node.
10024 * lto-symtab.c (lto_varpool_replace_node): Update.
10025 (lto_symtab_resolve_can_prevail_p): Update.
10026 (lto_symtab_merge_cgraph_nodes): Update.
10027 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10028 propagate_constants_accross_call, devirtualization_time_bonus,
10029 ipcp_propagate_stage): Update.
10030 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB		 10031 * ipa-inline-transform.c (clone_inlined_nodes,
10032 preserve_function_body_p): Update.
e70670cf
JH		 10033 * ipa-reference.c (propagate): Update.
10034 (write_node_summary_p): Update.
10035 * toplev.c (wrapup_global_declaration_2): Update.
10036 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10037 (analyze_function) ... this one.
10038 (cgraph_process_new_functions): Update.
10039 (cgraph_reset_node): Export.
10040 (cgraph_finalize_function): Update.
10041 (cgraph_add_new_function): Update.
10042 (process_function_and_variable_attributes): Update.
10043 (varpool_finalize_decl): Update.
10044 (symbol_finalized): Remove.
10045 (symbol_finalized_and_needed): Rename to ...
10046 (symbol_defined_and_needed): ... update.
10047 (cgraph_analyze_functions): Update.
10048 (handle_alias_pairs): Update.
10049 (mark_functions_to_output): Update.
10050 (assemble_thunk): Update.
10051 (output_in_order): Update.
10052 (output_weakrefs): Update.
10053 (finalize_compilation_unit): Update.
10054 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10055 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10056 input_node, input_varpool_node): Update.
10057 * dbxout.c (dbxout_expand_expr): Update.
10058 * cgraphclones.c (cgraph_clone_node): Update.
10059 (cgraph_copy_node_for_versioning): Update.
10060 (cgraph_materialize_clone): Update.
10061 (cgraph_materialize_all_clones): Update.
10062 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10063 propagate_pure_const, propagate_nothrow): Update.
10064 * lto-streamer-out.c (lto_output, write_symbol): Update.
10065 * ipa-utils.c (ipa_reverse_postorder): Update.
10066 * ipa-inline.c (can_inline_edge_p): Update.
10067 (update_caller_keys, ipa_inline): Update.
10068 * dwarf2out.c (reference_to_unused,
10069 premark_types_used_by_global_vars_helper): Update.
10070 * tree-eh.c (tree_could_trap_p): Update.
10071 * ipa-split.c (consider_split, execute_split_functions): Update.
10072 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB		 10073 has_addr_references_p): Update; move ahead in file for better
10074 readability.
e70670cf
JH		 10075 (process_references): Simplify.
10076 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10077 bodies are removed.
10078 (cgraph_comdat_can_be_unshared_p): Make static.
10079 (cgraph_externally_visible_p): Update.
10080 (varpool_externally_visible_p): Update.
10081 (function_and_variable_visibility): Update.
10082 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10083 ipa_tm_mark_force_output_node): Update.
10084 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10085 estimate_edge_devirt_benefit, inline_generate_summary,
10086 inline_write_summary): Update.
10087 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10088 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB		 10089 (ipa_print_node_params, ipa_prop_read_section,
10090 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH		 10091 * varasm.c (mark_decl_referenced): Update.
10092 (assemble_alias, dump_tm_clone_pairs): Update.
10093 * tree-inline.c (copy_bb): Update.
10094 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
10095 Update.
10096 * symtab.c (dump_symtab_base): Print new flags.
10097 (verify_symtab_base): Verify new flags.
10098 (symtab_alias_ultimate_target): New function.
10099 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
10100 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
10101 Update.
c1e183a9
UB		 10102 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
10103 Update.
e70670cf
JH		 10104 * i386.c (ix86_get_function_versions_dispatcher,
10105 ix86_generate_version_dispatcher_body): Update.
10106 (fold_builtin_cpu): Use varpool_add_new_variable.
10107 * varpool.c (varpool_remove_initializer): Break out from ...
10108 (varpool_remove_node): ... this one.
10109 (dump_varpool_node, varpool_node_for_asm,
10110 cgraph_variable_initializer_availability, varpool_analyze_node,
10111 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB		 10112 varpool_finalize_named_section_flags, varpool_create_variable_alias):
10113 Update.
e70670cf 10114
182802ad
JH		 101152013-05-29 Jan Hubicka <jh@suse.cz>
10116
10117 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
10118
8190b609
ER		 101192013-05-29 Easwaran Raman <eraman@google.com>
10120
10121 PR tree-optimization/57442
10122 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
10123 when control exits the main loop.
10124
69f5aa9b
SKS		 101252013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
10126
10127 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
10128 and RX600.
c1e183a9 10129 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS		 10130 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
10131 * rx/t-rx: Add rx100 under multi library matches option for nofpu
10132 option.
10133
4b847da9
BS		 101342013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10135
10136 PR tree-optimization/57441
10137 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
10138 Don't limit size of incr_vec to number of candidates.
10139
4fc43c39
SE		 101402013-05-29 Steve Ellcey <sellcey@imgtec.com>
10141
10142 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
10143 and mips16 directories.
c1e183a9 10144 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE		 10145 (MULTILIB_DIRNAMES): Ditto.
10146 (MULTILIB_EXCEPTIONS): Add new exceptions.
10147 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
10148 (MULTILIB_DIRNAMES): Ditto.
10149 (MULTILIB_EXCEPTIONS): Add new exceptions.
10150
12211b99 101512012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS		 10152 Marcus Shawcroft <marcus.shawcroft@arm.com>
10153
10154 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
10155 SYMBOL_TINY_ABSOLUTE.
10156 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
10157 SYMBOL_TINY_ABSOLUTE.
10158 (aarch64_expand_mov_immediate): Likewise.
10159 (aarch64_classify_symbol): Likewise.
10160 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10161 Permit SYMBOL_TINY_ABSOLUTE.
10162 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10163
12211b99 101642013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS		 10165 Marcus Shawcroft <marcus.shawcroft@arm.com>
10166
10167 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10168 Refactor if/switch. Replace gcc_assert with if.
10169
c0186656
GG		 101702013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10171
10172 * config/i386/i386.c (initial_ix86_tune_features): Enable
10173 FP Reassociation for AMD bdver1 and bdver2.
10174
d20188f3
MJ		 101752013-05-29 Martin Jambor <mjambor@suse.cz>
10176
10177 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10178 and IMAGPART_EXPR do not occur within other handled_components.
10179
292cba13
RB		 101802013-05-29 Richard Biener <rguenther@suse.de>
10181
10182 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10183 access on whether the use is in the BB we currently try to
10184 vectorize.
10185 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10186 vectorize to vect_bb_slp_scalar_cost.
10187
6eddf228
RB		 101882013-05-29 Richard Biener <rguenther@suse.de>
10189
10190 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10191 computing scalar cost offsetted by stmts that are kept live
10192 by scalar uses.
10193 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10194 for computation of scalar cost.
10195
7df36117
SE		 101962013-05-28 Steve Ellcey <sellcey@mips.com>
10197
10198 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10199
36536d79
BI		 102002013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10201
10202 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10203 array notation built-in reduction functions.
10204 * doc/passes.texi (Passes): Added documentation about changes done
10205 for Cilk Plus.
10206 * doc/invoke.texi (C Dialect Options): Added documentation about
10207 the -fcilkplus flag.
10208 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10209 (BUILTINS_DEF): Depend on cilkplus.def.
10210 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10211 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10212 * cilkplus.def: New file.
10213
12211b99 102142013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR		 10215
10216 PR rtl-optimization/57439
10217 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10218
9daf6dbb
ER		 102192013-05-28 Easwaran Raman <eraman@google.com>
10220
10221 PR tree-optimization/57337
10222 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10223 (find_insert_point): Correctly identify the insertion point
10224 when two statements with the same UID is compared.
10225
fbd7e877
RB		 102262013-05-28 Richard Biener <rguenther@suse.de>
10227
10228 PR tree-optimization/56787
10229 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10230 from the list of data references.
10231 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10232 clobbers.
10233 (vect_analyze_loop_operations): Likewise.
10234 (vect_transform_loop): Remove clobbers.
10235
bbba1117
MJ		 102362013-05-28 Martin Jambor <mjambor@suse.cz>
10237
10238 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10239 and REALPART_EXPRs have scalar type.
10240
bd388c2a
RB		 102412013-05-28 Richard Biener <rguenther@suse.de>
10242
10243 PR tree-optimization/57411
10244 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10245 virtual operands.
10246 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10247 virtual operand propagation.
10248
2f56a311
EB		 102492013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10250
10251 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10252 destination register for bmasksi_vis.
10253 (vector_init_bshuffle): Likewise.
10254 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10255
5c3eacbb
EB		 102562013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10257
10258 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10259 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10260 mode if the instruction isn't available in the original mode.
10261 * config/sparc/sparc.opt (mfix-ut699): New option.
10262 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10263 (divdf3): Turn into expander.
10264 (divdf3_nofix): New insn.
10265 (divdf3_fix): Likewise.
10266 (divsf3): Disable if -mfix-ut699.
10267 (sqrtdf2): Turn into expander.
10268 (sqrtdf2_nofix): New insn.
10269 (sqrtdf2_fix): Likewise.
10270 (sqrtsf2): Disable if -mfix-ut699.
10271
a1756c0a
RB		 102722013-05-27 Richard Biener <rguenther@suse.de>
10273
10274 PR middle-end/57412
10275 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10276 block for the new loop.
10277
5a892248
RB		 102782013-05-27 Richard Biener <rguenther@suse.de>
10279
10280 PR tree-optimization/57343
10281 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10282 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10283 (number_of_iterations_cond): Do not build the folded tree.
10284
d1de852b
RB		 102852013-05-27 Richard Biener <rguenther@suse.de>
10286
10287 Revert
10288 PR middle-end/57381
10289 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10290 OEP_CONSTANT_ADDRESS_OF retained.
10291
10292 PR tree-optimization/57417
10293 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10294 for unchanged base.
10295 (set_ssa_val_to): Compare addresses using
10296 get_addr_base_and_unit_offset.
10297
12211b99 102982013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR		 10299
10300 PR rtl-optimization/56833
10301 * postreload.c (move2add_record_mode): New function.
10302 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10303 (move2add_use_add2_insn): Use move2add_record_sym_value.
10304 (move2add_use_add3_insn): Likewise.
10305 (reload_cse_move2add): Use move2add_valid_value_p and
10306 move2add_record_mode. Invalidate call-clobbered and REG_INC
10307 affected regs by setting reg_mode to VOIDmode.
10308 (move2add_note_store): Don't pretend the inside of a SUBREG is
10309 the actual destination. Invalidate single/leading registers by
10310 setting reg_mode to VOIDmode.
10311 Use move2add_record_sym_value, move2add_valid_value_p and
10312 move2add_record_mode.
10313
b03be25f
RB		 103142013-05-27 Richard Biener <rguenther@suse.de>
10315
10316 PR tree-optimization/57396
10317 * tree-affine.c (double_int_constant_multiple_p): Properly
10318 return false for val == 0 and div != 0.
10319
44e88db2
RS		 103202013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10321
10322 * config/mips/mips.h: Use #elif in preprocessor conditions.
10323
3b859704
RS		 103242013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10325
10326 PR target/53916
10327 * config/mips/constraints.md (kl): New constraint.
10328 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10329 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10330 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10331 from LO for MIPS16.
10332 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10333
c979d5f5
RS		 103342013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10335
10336 PR target/55777
10337 * config/mips/mips.c (mips_can_inline_p): New function.
10338 (TARGET_CAN_INLINE_P): Define.
10339
8e90de43
SB		 103402013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10341
10342 * sched-int.h (ds_t, dw_t): Make unsigned int.
10343 Fix documentation that describes how all the ds_t bits are used.
10344 Reserve the last bit for delayed-branch scheduling.
10345 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10346 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10347 (gen_dep_weak_1): Remove prototype.
10348 * sched-deps.c (get_dep_weak_1): Make static.
10349 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10350 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10351 * doc/tm.texi: Regenerate.
10352 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
10353
cb5cb194
SB		 103542013-05-24 Steven Bosscher <steven@gcc.gnu.org>
10355
10356 PR debug/56950
10357 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
10358
5bd93ff6
NS		 103592013-05-24 Nathan Sidwell <nathan@codesourcery.com>
10360 Sandra Loosemore <sandra@codesourcery.com>
10361
10362 * config.gcc (powerpc-*): Allow native for with-cpu.
10363
2343af65
JL		 103642013-05-24 Jeff Law <law@redhat.com>
10365
10366 PR tree-optimization/57124
10367 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
10368 conversion feeding a condition if the range has an overflow
10369 if -fstrict-overflow. Add warnings for when we do make the
10370 transformation.
10371
3ad695b9
DC		 103722013-05-24 Dehao Chen <dehao@google.com>
10373
04960246 10374 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC		 10375 (locus_discrim_hasher): Likewise.
10376 (locus_discrim_hasher::hash): Likewise.
10377 (locus_discrim_hasher::equal): Likewise.
10378
82338059
MJ		 103792013-05-24 Martin Jambor <mjambor@suse.cz>
10380
10381 PR tree-optimization/57294
10382 * cgraph.h (ipa_record_stmt_references): Declare.
10383 * cgraphbuild.c (ipa_record_stmt_references): New function.
10384 (build_cgraph_edges): Use ipa_record_stmt_references.
10385 (rebuild_cgraph_edges): Likewise.
10386 (cgraph_rebuild_references): Likewise.
10387 * ipa-prop.c (ipa_modify_call_arguments): Discard references
10388 associated with the old statement and build references from the
10389 newly built statements.
10390 * ipa-ref.c (ipa_remove_stmt_references): New function.
10391 * ipa-ref.h (ipa_remove_stmt_references): Declare.
10392
1ccd4874
VM		 103932013-05-24 Vladimir Makarov <vmakarov@redhat.com>
10394
55805e54 10395 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 10396 mem-mem moves.
55805e54 10397 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 10398 too.
55805e54 10399 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM		 10400 secondary memory moves.
10401 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
10402 reg set up in the current insn.
10403
25e25c73
DC		 104042013-05-24 Dehao Chen <dehao@google.com>
10405
04960246 10406 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC		 10407 hash function.
10408 (locus_descrim_hasher::equal): Likewise.
04960246 10409 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC		 10410 (make_edges): Likewise.
10411 (next_discriminator_for_locus): Likewise.
10412 (same_line_p): Likewise.
10413 (assign_discriminators): Likewise.
10414 (make_cond_expr_edges): Likewise.
10415 (make_gimple_switch_edges): Likewise.
10416 (make_goto_expr_edges): Likewise.
10417 (make_gimple_asm_edges): Likewise.
10418
50d38551
IB		 104192013-05-24 Ian Bolton <ian.bolton@arm.com>
10420
10421 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
10422 X format specifier to only display bottom 16 bits.
10423 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
10424 immediate to match for operand 2, since it will be masked.
10425
aea0101d
RB		 104262013-05-24 Richard Biener <rguenther@suse.de>
10427
10428 PR tree-optimization/57287
10429 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
10430 all SSA names that occur in abnormal PHIs.
10431
634e03d3
AI		 104322013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
10433
10434 PR tree-ssa/57385
10435 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
10436 that index is not negative.
10437
b17c024f
EB		 104382013-05-24 Eric Botcazou <ebotcazou@adacore.com>
10439
10440 PR rtl-optimization/55177
10441 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
10442 (simplify_byte_swapping_operation): New.
10443 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
10444 (simplify_relational_operation_1): Deal with BSWAP.
10445
46aeac1b
RH		 104462013-05-23 Richard Henderson <rth@redhat.com>
10447
10448 PR target/56742
10449 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
10450 (ix86_reorg): Call it.
10451
70cc1536
UB		 104522013-05-23 Uros Bizjak <ubizjak@gmail.com>
10453
10454 PR target/57379
10455 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
10456 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
10457 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
10458
b3851501
CB		 104592013-05-23 Christian Bruel <christian.bruel@st.com>
10460
10461 PR debug/57351
10462 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
10463
12211b99 104642013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS		 10465 Marcus Shawcroft <marcus.shawcroft@arm.com>
10466
10467 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
10468 * config/aarch64/constraints.md (Usa): Remove.
10469 * doc/md.texi (AArch64 Usa): Remove.
10470
12211b99 104712013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS		 10472 Marcus Shawcroft <marcus.shawcroft@arm.com>
10473
10474 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
10475 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
10476 * config/aarch64/predicates.md (aarch64_const_address): Remove.
10477 (aarch64_mov_operand): Use aarch64_mov_operand_p.
10478
12211b99 104792013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP		 10480
10481 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
10482 instruction (AdvSIMD).
10483 * config/aarch64/aarch64-builtins.c
10484 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
10485 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
10486
76c36cb1 104872013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ		 10488
10489 PR middle-end/57347
10490 * tree.h (contains_bitfld_component_ref_p): Declare.
10491 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB		 10492 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
10493 caller.
7d2fb524
MJ		 10494 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
10495 not access a bit-field. Assert all final offsets are byte-aligned.
10496
ce521ae6
RB		 104972013-05-23 Richard Biener <rguenther@suse.de>
10498
10499 PR tree-optimization/57380
10500 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
10501 least one invariant or re-used load.
10502 * passes.c (init_optimization_passes): Move pass_phiprop before
10503 pass_forwprop.
10504
75c7257f
JG		 105052013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
10506
10507 * config/aarch64/aarch64-simd.md
10508 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
10509
af6d44b5
RB		 105102013-05-23 Richard Biener <rguenther@suse.de>
10511
10512 PR middle-end/57381
10513 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10514 OEP_CONSTANT_ADDRESS_OF retained.
10515
bd3647bf
JJ		 105162013-05-23 Jakub Jelinek <jakub@redhat.com>
10517
10518 PR middle-end/57344
70cc1536
UB		 10519 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
10520 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 10521
a55757ea
RB		 105222013-05-23 Richard Biener <rguenther@suse.de>
10523
10524 PR rtl-optimization/57341
10525 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
10526 instead of true_dependence.
10527
cfb00b41
DM		 105282013-05-22 David Malcolm <dmalcolm@redhat.com>
10529
dd1c676f
DM		 10530 * bb-reorder.c (branch_threshold): Make const.
10531 (exec_threshold): Ditto.
cfb00b41 10532
f62511da 105332013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 10534 Pat Haugen <pthaugen@us.ibm.com>
10535 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 10536
10537 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
10538 documentation for the power8 crypto builtins.
10539
10540 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
10541
10542 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
10543 macros for defining power8 builtin functions.
10544 (BU_P8V_AV_2): Likewise.
10545 (BU_P8V_AV_P): Likewise.
10546 (BU_P8V_VSX_1): Likewise.
10547 (BU_P8V_OVERLOAD_1): Likewise.
10548 (BU_P8V_OVERLOAD_2): Likewise.
10549 (BU_CRYPTO_1): Likewise.
10550 (BU_CRYPTO_2): Likewise.
10551 (BU_CRYPTO_3): Likewise.
10552 (BU_CRYPTO_OVERLOAD_1): Likewise.
10553 (BU_CRYPTO_OVERLOAD_2): Likewise.
10554 (XSCVSPDP): Fix typo, point to the correct instruction.
10555 (VCIPHER): Add power8 crypto builtins.
10556 (VCIPHERLAST): Likewise.
10557 (VNCIPHER): Likewise.
10558 (VNCIPHERLAST): Likewise.
10559 (VPMSUMB): Likewise.
10560 (VPMSUMH): Likewise.
10561 (VPMSUMW): Likewise.
10562 (VPERMXOR_V2DI): Likewise.
10563 (VPERMXOR_V4SI: Likewise.
10564 (VPERMXOR_V8HI: Likewise.
10565 (VPERMXOR_V16QI: Likewise.
10566 (VSHASIGMAW): Likewise.
10567 (VSHASIGMAD): Likewise.
10568 (VPMSUM): Likewise.
10569 (VPERMXOR): Likewise.
10570 (VSHASIGMA): Likewise.
10571
10572 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
10573 __CRYPTO__ if the crypto instructions are available.
10574 (altivec_overloaded_builtins): Add support for overloaded power8
10575 builtins.
10576
10577 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
10578 support for power8 crypto builtins.
10579 (builtin_function_type): Likewise.
10580 (altivec_init_builtins): Add support for builtins that take vector
10581 long long (V2DI) arguments.
10582
10583 * config/rs6000/crypto.md: New file, define power8 crypto
10584 instructions.
10585
105862013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 10587 Pat Haugen <pthaugen@us.ibm.com>
10588 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 10589
10590 * doc/invoke.texi (Option Summary): Add power8 options.
10591 (RS/6000 and PowerPC Options): Likewise.
10592
10593 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
10594 constraints.md instead of rs6000.h. Reorder w* constraints. Add
10595 wm, wn, wr documentation.
10596
10597 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
10598 registers if direct move instructions are enabled.
10599 (wn): New constraint for no registers.
10600 (wq): New constraint for quad word even GPR registers.
10601 (wr): New constraint if 64-bit instructions are enabled.
10602 (wv): New constraint if power8 vector instructions are enabled.
10603 (wQ): New constraint for quad word memory locations.
10604
10605 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
10606 constraint for 0..15 for crypto instructions.
10607 (gpc_reg_operand): If VSX allow registers in VSX registers as well
10608 as GPR and floating point registers.
10609 (int_reg_operand): New predicate to match only GPR registers.
10610 (base_reg_operand): New predicate to match base registers.
10611 (quad_int_reg_operand): New predicate to match even GPR registers
10612 for quad memory operations.
10613 (vsx_reg_or_cint_operand): New predicate to allow vector logical
10614 operations in both GPR and VSX registers.
10615 (quad_memory_operand): New predicate for quad memory operations.
10616 (reg_or_indexed_operand): New predicate for direct move support.
10617
10618 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
10619 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
10620 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
10621 (POWERPC_MASKS): Add power8 options.
10622 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
10623 various options.
10624
10625 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10626 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
10627
10628 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
10629 (-mpower8-fusion): New power8 options.
10630 (-mpower8-fusion-sign): Likewise.
10631 (-mpower8-vector): Likewise.
10632 (-mcrypto): Likewise.
10633 (-mdirect-move): Likewise.
10634 (-mquad-memory): Likewise.
10635
10636 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
10637 power8.
10638 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
10639 registers.
70cc1536 10640 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM		 10641 (rs6000_debug_vector_unit): Add p8_vector.
10642 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
10643 definitions. Also print fusion state.
10644 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
10645 (rs6000_builtin_mask_calculate): Add power8 builtin support.
10646 (rs6000_option_override_internal): Add support for power8.
10647 (rs6000_common_init_builtins): Add debugging for skipped builtins
10648 if -mdebug=builtin.
10649 (rs6000_adjust_cost): Add power8 support.
10650 (rs6000_issue_rate): Likewise.
10651 (insn_must_be_first_in_group): Likewise.
10652 (insn_must_be_last_in_group): Likewise.
10653 (force_new_group): Likewise.
10654 (rs6000_register_move_cost): Likewise.
10655 (rs6000_opt_masks): Likewise.
10656
10657 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
10658 power8 capable assembler, default to power7 options.
10659 (TARGET_DIRECT_MOVE): Likewise.
10660 (TARGET_CRYPTO): Likewise.
10661 (TARGET_P8_VECTOR): Likewise.
10662 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
10663 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
10664 (VECTOR_MEM_P8_VECTOR_P): Likewise.
10665 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
10666 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
10667 (TARGET_XSCVDPSPN): Likewise.
10668 (TARGET_XSCVSPDPN): Likewsie.
10669 (TARGET_SYNC_HI_QI): Likewise.
10670 (TARGET_SYNC_TI): Likewise.
10671 (MASK_CRYPTO): Likewise.
10672 (MASK_DIRECT_MOVE): Likewise.
10673 (MASK_P8_FUSION): Likewise.
10674 (MASK_P8_VECTOR): Likewise.
70cc1536
UB		 10675 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
10676 temporary used by some of the direct move instructions to get two FP
10677 temporary registers does not force creation of a stack frame.
f62511da
MM		 10678 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
10679 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
10680 that any VSX registers are tieable, even if they are also an
10681 Altivec vector mode.
10682 (r6000_reg_class_enum): Add wm, wr, wv constraints.
10683 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
10684 (RS6000_BTM_CRYPTO): Likewise.
10685 (RS6000_BTM_COMMON): Likewise.
10686
10687 * config/rs6000/rs6000.md (cpu attribute): Add power8.
10688 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
10689 (enum rs6000_vector): Add power8 vector support.
10690
73a1a707
RR		 106912013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10692
10693 PR target/19599
10694 PR target/57340
10695 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
10696 (any_sibcall_could_use_r3): this and handle indirect calls.
10697 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
10698
d28073d4
BS		 106992013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10700
10701 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
10702
bee0b10c
RB		 107032013-05-22 Richard Biener <rguenther@suse.de>
10704
10705 PR middle-end/57349
10706 * profile.c (branch_prob): Do not split blocks that are
10707 abnormally receiving from ECF_RETURNS_TWICE functions.
10708
98409b51
RS		 107092013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10710
10711 * recog.c (offsettable_address_addr_space_p): Fix calculation of
10712 address mode. Move pointer mode initialization to the same place.
10713
c0602ab8
MZ		 107142013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10715
10716 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
10717 while it has any effect.
10718
4f7a634e
ER		 107192013-05-21 Easwaran Raman <eraman@google.com>
10720
10721 PR tree-optimization/57322
9daf6dbb
ER		 10722 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
10723 UID of the statement added to the BB to be 1.
4f7a634e 10724
a4ce1258
JJ		 107252013-05-21 Jakub Jelinek <jakub@redhat.com>
10726
10727 PR tree-optimization/57331
70cc1536
UB		 10728 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
10729 of conversion from pointer type to integral type with integer.
a4ce1258 10730
1b14621a
MJ		 107312013-05-21 Martin Jambor <mjambor@suse.cz>
10732
10733 PR lto/57289
10734 * ipa-prop.c (ipa_read_node_info): Process param_used and
10735 controlled_uses in the same order as when writing.
10736
e60661f0
MG		 107372013-05-21 Magnus Granberg <baldrick@free.fr>
10738
10739 PR plugins/56754
ee49aa34 10740 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 10741
4a61be9a
RB		 107422013-05-21 Richard Biener <rguenther@suse.de>
10743
10744 PR tree-optimization/57318
10745 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
10746 estimate stmts with side-effects as likely eliminated.
10747
c52da5f7
RB		 107482013-05-21 Richard Biener <rguenther@suse.de>
10749
10750 PR tree-optimization/57330
10751 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
10752 preserve the call stmts fntype.
10753
7ec67e2a
RB		 107542013-05-21 Richard Biener <rguenther@suse.de>
10755
10756 PR tree-optimization/57303
10757 * tree-ssa-sink.c (statement_sink_location): Improve killing
10758 stmt detection and properly handle self-assignments.
10759
b112d513
CB		 107602013-05-21 Christian Bruel <christian.bruel@st.com>
10761
55805e54
YR		 10762 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
10763 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB		 10764 registers. Set register size out of the PARALLEL loop.
10765
14c2ec26
OE		 107662013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
10767
10768 PR target/56547
10769 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
10770 (*fmasf4, *fmasf4_media): New insns.
10771
da734fa1
RS		 107722013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10773
10774 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
10775 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
10776 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
10777 (mips_idiv_insns): Update the comments to say that the returned
10778 instruction counts are in units of BASE_INSN_LENGTH.
10779 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
10780 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
10781 using 2 rather than 4 as the length of indirect MIPS16 and
10782 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
10783 length of a NOP. Don't divide MIPS16 lengths by 2.
10784 (mips16_split_long_branches): Assume a branch is long if the
10785 length is greater than 4 rather than 8.
10786 * config/mips/mips.md (length): Give MIPS16 lengths directly,
10787 rather than multiplying them by 2. Multiply instruction counts
10788 by BASE_INSN_LENGTH rather than 4.
10789 (*jump_mips16, tls_get_tp_mips16_<mode>)
10790 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
10791
13719e8b
RS		 107922013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10793
10794 * config/mips/mips.md (extended_mips16): Remove branch case.
10795 (length): Remove duplicated extended_mips16 test.
10796
c3850d14
RS		 107972013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10798
10799 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
10800
8da2e059
RS		 108012013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
10802
10803 * recog.h (Recog_data): Rename to...
10804 (recog_data_d): ...this.
10805 (recog_data): Update accordingly.
10806 * recog.c (recog_data): Likewise.
10807 * reload.c (save_recog_data): Likewise.
10808 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
10809 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
10810
2df013f3
JB		 108112013-05-17 Julian Brown <julian@codesourcery.com>
10812
10813 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
10814 found in a REG_EQUAL note, invalidate it.
10815
371e77e3 108162013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER		 10817
10818 * tree-ssa-reassoc.c (find_insert_point): New function.
10819 (insert_stmt_after): Likewise.
10820 (get_def_stmt): Likewise.
10821 (ensure_ops_are_available): Likewise.
10822 (not_dominated_by): Likewise.
10823 (rewrite_expr_tree): Do not move statements beyond what is
10824 necessary. Remove call to swap_ops_for_binary_stmt...
10825 (reassociate_bb): ... and move it here.
10826 (build_and_add_sum): Assign UIDs for new statements.
10827 (linearize_expr): Likewise.
10828 (do_reassoc): Renumber gimple statement UIDs.
10829
e01c7cca
JH		 108302013-05-17 Jan Hubicka <jh@suse.cz>
10831
10832 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
10833 weakrefs.
10834 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB		 10835 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
10836 weakrefs.
e01c7cca
JH		 10837 (output_weakrefs): Update.
10838
c3272a92
PCC		 108392013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
10840 Martin Jambor <mjambor@suse.cz>
10841
10842 PR middle-end/57276
10843 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
10844 value that corresponds to the given aggval is found in values vector.
10845
11c2aa39
UB		 108462013-05-17 Uros Bizjak <ubizjak@gmail.com>
10847
10848 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
10849 sse, sse2, sse3, ssse3 and sse4a flags to options.
10850
abecc8c6
DM		 108512013-05-17 David Malcolm <dmalcolm@redhat.com>
10852
10853 * gengtype-state.c: (s_expr_writer): New class, to handle
10854 prettifying of output layout of s-expressions.
10855 (state_writer): New class, to write out gtype.state.
10856 (state_written_type_count): Move this variable into member data of
10857 state_writer.
10858 (s_expr_writer::s_expr_writer): New code: constructor for new class
10859 (state_writer::state_writer(): ditto
10860 (s_expr_writer::write_new_line): New function
10861 (s_expr_writer::write_any_indent): ditto
10862 (s_expr_writer::begin_s_expr): ditto
10863 (s_expr_writer::end_s_expr): ditto
10864 (write_state_fileloc): convert to method of state_writer...
10865 (state_writer:: write_state_fileloc): ...and use methods of
10866 s_expr_writer to write indentation into the gtype.state output file
10867 to visually represent the hierarchical structure of the list
10868 structures
10869 (write_state_fields): ditto, renaming to...
10870 (state_writer::write_state_fields)
10871 (write_state_a_string): ditto, renaming to...
10872 (state_writer::write_state_a_string)
10873 (write_state_string_option): ditto, renaming to...
10874 (state_writer::write_state_string_option)
10875 (write_state_type_option): ditto, renaming to...
10876 (state_writer::write_state_type_option)
10877 (write_state_nested_option): ditto, renaming to...
10878 (state_writer::write_state_nested_option)
10879 (write_state_option): ditto, renaming to...
10880 (state_writer::write_state_option)
10881 (write_state_options): ditto, renaming to...
10882 (state_writer::write_state_options)
10883 (write_state_lang_bitmap): ditto, renaming to...
10884 (state_writer::write_state_lang_bitmap)
10885 (write_state_version): ditto, renaming to...
10886 (state_writer::write_state_version)
10887 (write_state_scalar_type): ditto, renaming to...
10888 (state_writer::write_state_scalar_type)
10889 (write_state_string_type): ditto, renaming to...
10890 (state_writer::write_state_string_type)
10891 (write_state_undefined_type): ditto, renaming to...
10892 (state_writer::write_state_undefined_type)
10893 (write_state_struct_union_type): ditto, renaming to...
10894 (state_writer::write_state_struct_union_type)
10895 (write_state_struct_type): ditto, renaming to...
10896 (state_writer::write_state_struct_type)
10897 (write_state_user_struct_type): ditto, renaming to...
10898 (state_writer::write_state_user_struct_type)
10899 (write_state_lang_struct_type): ditto, renaming to...
10900 (state_writer::write_state_lang_struct_type)
10901 (write_state_param_struct_type): ditto, renaming to...
10902 (state_writer::write_state_param_struct_type)
10903 (write_state_pointer_type): ditto, renaming to...
10904 (state_writer::write_state_pointer_type)
10905 (write_state_array_type): ditto, renaming to...
10906 (state_writer::write_state_array_type)
10907 (write_state_gc_used): ditto, renaming to...
10908 (state_writer::write_state_gc_used)
10909 (write_state_common_type_content): ditto, renaming to...
10910 (state_writer::write_state_common_type_content)
10911 (write_state_type): ditto, renaming to...
10912 (state_writer::write_state_type)
10913 (write_state_pair_list): ditto, renaming to...
10914 (state_writer::write_state_pair_list)
10915 (write_state_pair): ditto, renaming to...
10916 (state_writer::write_state_pair)
10917 (write_state_typedefs): ditto, renaming to...
10918 (state_writer::write_state_typedefs)
10919 (write_state_structures): ditto, renaming to...
10920 (state_writer::write_state_structures)
10921 (write_state_param_structs): ditto, renaming to...
10922 (state_writer::write_state_param_structs)
10923 (write_state_variables): ditto, renaming to...
10924 (state_writer::write_state_variables)
10925 (write_state_srcdir): ditto, renaming to...
10926 (state_writer::write_state_srcdir)
10927 (write_state_files_list): ditto, renaming to...
10928 (state_writer::write_state_files_list)
10929 (write_state_languages): ditto, renaming to...
10930 (state_writer::write_state_languages)
10931 (write_state): create a state_writer instance and use it when
10932 writing out the state file
10933
d6545f29
MS		 109342013-05-17 Mike Stump <mikestump@comcast.net>
10935
816a3f73 10936 PR rtl-optimization/57304
d6545f29
MS		 10937 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
10938 accessing DF_REF_REAL_LOC.
10939
38c821cf
JJ		 109402013-05-17 Jakub Jelinek <jakub@redhat.com>
10941
10942 PR rtl-optimization/57281
10943 PR rtl-optimization/57300
10944 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
10945 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
10946 what the other splitter did if the registers are dead.
10947
2a293391
RB		 109482013-05-17 Richard Biener <rguenther@suse.de>
10949
10950 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
10951 MEM_REF offsets.
10952
17042d2b
JJ		 109532013-05-17 Jakub Jelinek <jakub@redhat.com>
10954
10955 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
10956 linking.
10957
5b115c1f
MP		 109582013-05-17 Marek Polacek <polacek@redhat.com>
10959
11c2aa39
UB		 10960 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
10961 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 10962
52d84413
JJ		 109632013-05-17 Jakub Jelinek <jakub@redhat.com>
10964
68119618
JJ		 10965 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
10966 vect_external_def oprnd1 with loop_vinfo, try to emit
10967 optional cast, negation and and stmts on the loop preheader
10968 edge instead of into the pattern def seq.
10969
52d84413
JJ		 10970 PR tree-optimization/57051
10971 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
10972 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
10973
17b962bd
NC		 109742013-05-16 Nick Clifton <nickc@redhat.com>
10975
10976 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
10977 (rl78_is_naked_func): New function.
56aefbf7
UB		 10978 (rl78_expand_prologue): Skip prologue generation for naked functions.
10979 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC		 10980 * doc/extend.texi (naked): Add RL78 to the list of processors
10981 that supports this attribute.
10982
b1a0f84e
JL		 109832013-05-16 Jeff Law <law@redhat.com>
10984
10985 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
10986
af0e415b
UB		 109872013-05-16 Uros Bizjak <ubizjak@gmail.com>
10988
10989 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
10990 cache parameters using detect_caches_amd also for CYRIX,
10991 NSC and TM2 signatures.
10992
19db293a
UB		 109932013-05-16 Uros Bizjak <ubizjak@gmail.com>
10994 Dzianis Kahanovich <mahatma@eu.by>
10995
10996 PR target/45359
10997 PR target/46396
10998 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10999 VIA/Centaur processors and determine their cache parameters
11000 using detect_caches_amd.
11001
251a41b9
TJ		 110022013-05-16 Teresa Johnson <tejohnson@google.com>
11003
11004 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11005 (rtl_verify_edges): New function.
11006 (rtl_verify_bb_insns): Ditto.
11007 (rtl_verify_bb_pointers): Ditto.
11008 (rtl_verify_bb_insn_chain): Ditto.
11009 (rtl_verify_fallthru): Ditto.
11010 (rtl_verify_bb_layout): Ditto.
11011 (rtl_verify_flow_info_1): Outline checks into new functions.
11012 (rtl_verify_flow_info): Ditto.
11013
f14540b6
SE		 110142013-05-16 Steve Ellcey <sellcey@imgtec.com>
11015
11016 * cfghooks.c (copy_bbs): Add update_dominance argument.
11017 * cfghooks.h (copy_bbs): Update prototype.
11018 * tree-cfg.c (gimple_duplicate_sese_region):
11019 Add update_dominance argument.
11020 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11021 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11022 gimple_duplicate_sese_region call.
11023 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11024 Update copy_bbs call.
11025 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11026 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11027
7e9a3abb
JJ		 110282013-05-16 Jakub Jelinek <jakub@redhat.com>
11029
11030 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11031 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11032 vect_recog_rotate_pattern.
11033 (vect_recog_rotate_pattern): New function.
11034
427b248d
JM		 110352013-05-16 Jason Merrill <jason@redhat.com>
11036
11037 * Makefile.in (LLINKER): New variable.
11038 (mostlyclean): Remove link mutex.
11039 * configure.ac: Handle --enable-link-mutex.
11040 * lock-and-run.sh: New script.
11041
b871e3d2
RR		 110422013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11043
11044 PR target/19599
11045 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11046 for NULL decl.
11047
ce858126
RO		 110482013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11049
11050 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11051
798d3d04
GY		 110522013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11053
11054 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11055 * config/arm/arm.c (next_consecutive_mem): New function.
11056 (gen_movmem_ldrd_strd): Likewise.
11057 * config/arm/arm.md (movmemqi): Update condition and code.
11058 (unaligned_loaddi, unaligned_storedi): New patterns.
11059
0baddc45
RO		 110602013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11061
11062 * config.gcc: Obsolete *-*-solaris2.9*.
11063 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11064
aa9d5bcf
RB		 110652013-05-16 Richard Biener <rguenther@suse.de>
11066
11067 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11068 earlier, after GRAPHITE transforms and IV canonicalization.
11069
5a0f4dd3
JJ		 110702013-05-16 Jakub Jelinek <jakub@redhat.com>
11071
11072 * omp-low.c (extract_omp_for_data): For collapsed loops,
11073 if at least one of the loops is known at compile time to
11074 iterate zero times, set count to 0.
11075 (expand_omp_regimplify_p): New function.
11076 (expand_omp_for_generic): For collapsed loops, if at least
11077 one of the loops isn't known to iterate at least once,
11078 add runtime check with setting count to 0.
11079 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11080 For unsigned types if it isn't known at compile time that
11081 the loop will iterate at least once, add runtime check to bypass
11082 the whole loop if initial condition isn't true.
11083
e3753785
NS		 110842013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11085
11086 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11087
43bb4dd1
MG		 110882013-05-16 Marc Glisse <marc.glisse@inria.fr>
11089
11090 PR middle-end/57286
11091 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
11092 transformations to avoid an infinite loop.
11093
3571dde6
MP		 110942013-05-16 Marek Polacek <polacek@redhat.com>
11095
11096 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
11097
b8b3f0ca
LE		 110982013-05-15 Leif Ekblad <leif@rdos.net>
11099
11100 * config/i386/i386.c (ix86_decompose_address): Use
11101 DEFAULT_TLS_SEG_REG to access TLS segment register.
11102 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
11103 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
11104 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
11105
e299a383
RS		 111062013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
11107
11108 PR target/57260
11109 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
11110 sibling calls to functions that would normally be lazily bound,
11111 unless $gp is call-clobbered.
11112
e7208ea3
UB		 111132013-05-15 Uros Bizjak <ubizjak@gmail.com>
11114
19db293a 11115 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 11116 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB		 11117 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
11118 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB		 11119 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
11120 of TARGET_3DNOW.
11121 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
11122
e7413f3d
AS		 111232013-05-15 Andreas Schwab <schwab@suse.de>
11124
11125 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
11126 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
11127 third operand.
11128
ad4db775
TJ		 111292013-05-15 Teresa Johnson <tejohnson@google.com>
11130
11131 * loop-unroll.c (report_unroll_peel): Check decision before
11132 emitting unroll/peel message.
11133
af205f67
TJ		 111342013-05-15 Teresa Johnson <tejohnson@google.com>
11135
11136 * function.h (has_bb_partition): New rtl_data flag.
11137 (bb_reorder_complete): Ditto.
11138 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
11139 instead of flag_reorder_blocks_and_partition.
11140 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
11141 with some enhancements.
11142 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
11143 * bb-reorder.c (connect_traces): Check for has_bb_partition
11144 instead of flag_reorder_blocks_and_partition.
11145 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
11146 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
11147 verify_hot_cold_block_grouping.
11148 (partition_hot_cold_basic_blocks): Set has_bb_partition.
11149
9adcfa3c
RR		 111502013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11151
11152 PR target/19599
11153 * config/arm/predicates.md (call_insn_operand): New predicate.
11154 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
11155 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
11156 if insn is not a tail call.
11157 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11158 registers.
11159 * config/arm/arm.h (enum reg_class): New caller save register class.
11160 (REG_CLASS_NAMES): Likewise.
11161 (REG_CLASS_CONTENTS): Likewise.
11162 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11163 without decls.
11164
ccb3ad87
RB		 111652013-05-15 Richard Biener <rguenther@suse.de>
11166
11167 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11168 of MSG_OPTIMIZED_LOCATIONS.
11169 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11170 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11171 message.
11172 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11173 of MSG_OPTIMIZED_LOCATIONS.
11174 (execute_vect_slp): Likewise.
11175 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11176 (vect_create_cond_for_alias_checks): Likewise.
11177 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11178 (vect_recog_widen_mult_pattern): Likewise.
11179 (vect_recog_widen_sum_pattern): Likewise.
11180 (vect_recog_over_widening_pattern): Likewise.
11181 (vect_recog_widen_shift_pattern): Likewise.
11182 (vect_recog_vector_vector_shift_pattern): Likewise.
11183 (vect_recog_divmod_pattern): Likewise.
11184 (vect_recog_mixed_size_cond_pattern): Likewise.
11185 (vect_recog_bool_pattern): Likewise.
11186 (vect_pattern_recog_1): Likewise.
11187
48b1474e
MJ		 111882013-05-15 Martin Jambor <mjambor@suse.cz>
11189
11190 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11191 non-functions to builtin_unreachable.
11192 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11193 correct when new direct edges were discovered.
11194
9de04252
MJ		 111952013-05-15 Martin Jambor <mjambor@suse.cz>
11196
11197 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11198 header, print symbol order instead of node uid, print more information
11199 about indirect edge targets.
11200 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11201 uids.
11202 (ipa_make_edge_direct_to_target): Likewise.
11203 (remove_described_reference): Likewise.
11204 (propagate_controlled_uses): Likewise.
11205 (ipa_print_node_params): Also print symbol order.
11206 (ipcp_transform_function): Print symbol order instead of node uids.
11207 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11208 (cgraph_get_create_real_symbol_node): Likewise.
11209 * ipa-cp.c (print_lattice): Likewise.
11210 (print_all_lattices): Likewise.
11211 (determine_versionability): Likewise.
11212 (initialize_node_lattices): Likewise.
11213 (estimate_local_effects): Likewise.
11214 (update_profiling_info): Likewise.
11215 (create_specialized_node): Likewise.
11216 (perhaps_add_new_callers): Likewise.
11217 (decide_about_value): Likewise.
11218 (decide_whether_version_node): Likewise.
11219 (identify_dead_nodes): Likewise.
11220 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11221 (dump_inline_summary): Likewise.
11222 (estimate_node_size_and_time): Likewise.
11223 (inline_analyze_function): Likewise.
11224 * ipa-inline.c (report_inline_failed_reason): Likewise.
11225 (want_early_inline_function_p): Likewise.
11226 (edge_badness): Likewise.
11227 (update_edge_key): Likewise.
11228 (inline_small_functions): Likewise. Add dumping of order to two other
11229 dumps.
11230 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11231 instead of node uids.
11232 (propagate_pure_const): Likewise.
11233 (propagate_pure_const): Likewise.
11234 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11235 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11236 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11237 of node uids.
11238 * tree-pretty-print.c (dump_function_header): Likewise.
11239 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11240 Print symbol order instead of node uids.
11241
1dd03b91
AK		 112422013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11243
11244 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11245 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11246
fdf6a7b9
RB		 112472013-05-15 Richard Biener <rguenther@suse.de>
11248
11249 PR tree-optimization/57275
11250 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11251 return value for fail to do runtime alias checks for gather loads.
11252
2d6e4603
JH		 112532013-05-15 Jan Hubicka <jh@suse.cz>
11254
11255 PR lto/57038
11256 PR lto/47375
e7208ea3
UB		 11257 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11258 weakrefs are not external.
11259 (lto_symtab_merge_decls): Fix thinko when dealing with
11260 non-lto_symtab decls.
2d6e4603
JH		 11261 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11262 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11263 * varpool.c (dump_varpool_node): Dump more flags.
11264
83f44b39
GG		 112652013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11266
11267 * config/i386/i386.c (processor_alias_table): Add instruction
11268 FSGSBASE for AMD bdver3 architecture.
11269
2e55d062
JJ		 112702013-05-14 Jakub Jelinek <jakub@redhat.com>
11271
11272 * tree.c (warn_deprecated_use): Print file:line using locus color.
11273 * diagnostic.c (diagnostic_report_current_module): Print file:line
11274 and file:line:column using locus color.
11275
0bfdb81e
MS		 112762013-05-14 Mike Stump <mikestump@comcast.net>
11277
11278 * gdbinit.in: Add __null.
11279
a508ef22
MS		 112802013-05-14 Mike Stump <mikestump@comcast.net>
11281
11282 * recog.h: Rename struct recog_data to Recog_data.
11283 * recog.c: Likewise.
11284 * reload.c (can_reload_into): Likewise.
11285 * config/picochip/picochip.c: Likewise.
11286
e7180acb
MS		 112872013-05-14 Mike Stump <mikestump@comcast.net>
11288
11289 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11290
61216c26
SB		 112912013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11292
fd6beed4
SB		 11293 * resource.h (struct resources): Remove unch_memory member.
11294 (CLEAR_RESOURCE): Don't clear unch_memory.
11295 * resource.c (mark_referenced_resources): Don't set it.
11296 (mark_set_resources): Likewise.
11297 (mark_target_live_regs): Don't clear it.
11298 (init_resource_info): Likewise.
11299 * reorg.c (resource_conflicts_p): Don't compare it.
11300 (redundant_insn): Don't set it.
11301
61216c26
SB		 11302 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11303 Remove prototypes.
11304 * emit-rtl.c (next_label): Remove unused function.
11305 (skip_consecutive_labels, link_cc0_insns): Move to ...
11306 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11307 only place where these functions are used, and make them static.
11308
418d1b87
MG		 113092013-05-14 Marc Glisse <marc.glisse@inria.fr>
11310
11311 * fold-const.c (fold_negate_expr): Handle vectors.
11312 (fold_truth_not_expr): Make it static.
11313 (fold_invert_truthvalue): New static function.
11314 (invert_truthvalue_loc): Handle vectors. Do not call
11315 fold_truth_not_expr directly.
11316 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11317 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11318 (fold_comparison): Handle vector constants.
11319 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11320 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11321 * tree.h (fold_truth_not_expr): Remove declaration.
11322
fc21784d
JG		 113232013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11324
11325 * config/aarch64/aarch64-simd.md
11326 (aarch64_vcond_internal<mode>): Rename to...
11327 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11328 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11329 float modes. Clarify all iterator modes.
11330 (vcond<mode><mode>): Use new name for vcond expanders.
11331 (vcond<v_cmp_result><mode>): Likewise.
11332 (vcondu<mode><mode>: Likewise.
11333 * config/aarch64/iterators.md (VDQF_COND): New.
11334
d4c52634
MG		 113352013-05-14 Marc Glisse <marc.glisse@inria.fr>
11336
11337 PR bootstrap/57266
11338 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11339 variable for the shift amount. Check that we shift by non-negative
11340 amounts.
11341
2b261262
CLT		 113422013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11343
11344 PR target/42017
11345 * config/arm/arm.h (EPILOGUE_USES): Only return true
11346 for LR_REGNUM after epilogue_completed.
11347
12211b99 113482013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR		 11349
11350 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 11351 is error_mark_node.
6e022d7b 11352
56cf7859
RO		 113532013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11354
11355 PR target/57261
11356 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
11357 and Solaris 11+/x86 with gld.
11358 * configure: Regenerate.
11359
75776c6d
JJ		 113602013-05-14 Jakub Jelinek <jakub@redhat.com>
11361
11362 * expmed.c (expand_shift_1): Canonicalize rotates by
11363 constant bitsize / 2 to bitsize - 1.
7f998021 11364 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ		 11365 case ROTATERT>: Likewise.
11366
11367 Revert:
11368 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11369
11370 * config/i386/i386.md (rotateinv): New code attr.
11371 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11372 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11373 roll $31, %eax, etc.
11374
df35498a
RB		 113752013-05-14 Richard Biener <rguenther@suse.de>
11376
11377 PR middle-end/57235
11378 * tree-eh.c (sink_clobbers): Give up for successors with
11379 multiple predecessors and no virtual uses.
11380
cc6e7ece
EB		 113812013-05-14 Eric Botcazou <ebotcazou@adacore.com>
11382
11383 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
11384 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
11385
f778dd4d
JJ		 113862013-05-14 Jakub Jelinek <jakub@redhat.com>
11387
11388 PR middle-end/57251
11389 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
11390 the case when both op0 and op1 have VOIDmode.
11391
bad4df9b
KP		 113922013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11393
11394 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
11395 in multiply-accumulate mode.
11396
56f3e9ac
GW		 113972013-05-13 Guozhi Wei <carrot@google.com>
11398
11399 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
11400
a3d7ab92
KT		 114012013-05-13 Kai Tietz <ktietz@redhat.com>
11402
11403 PR target/56975
11404 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 11405 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 11406 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 11407 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT		 11408 (ix86_expand_prologue): Likewise.
11409 (ix86_expand_split_stack_prologue): Likewise.
11410 (legitimate_pic_address_disp_p): Likewise.
11411 (legitimize_pic_address): Likewise.
11412 (legitimize_tls_address): Likewise.
11413 (legitimize_pe_coff_symbol): Likewise.
11414 (output_pic_addr_const): Likewise.
11415 (construct_plt_address): Likewise.
11416 (ix86_expand_call): Likewise.
11417 (x86_output_mi_thunk): Likewise.
11418 (x86_function_profiler): Likewise.
11419
c59b7e28
SN		 114202013-05-13 Sofiane Naci <sofiane.naci@arm.com>
11421
11422 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
11423 similar switch cases.
11424 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
11425 (aarch64_simd_mov_to_<mode>low): Delete.
11426 (aarch64_simd_mov_to_<mode>high): Delete.
11427 (move_lo_quad_<mode>): Add w<-r alternative.
11428 (aarch64_simd_move_hi_quad_<mode>): Likewise.
11429 (aarch64_simd_mov_from_*): Update type attribute.
11430 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
11431 statement.
11432
5f28524a
JH		 114332013-05-13 Jan Hubicka <jh@suse.cz>
11434
11435 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
11436 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB		 11437 ix86_expand_epilogue, emit_i387_cw_initialization,
11438 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
11439 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 11440
ae6fa899
JJ		 114412013-05-13 Jakub Jelinek <jakub@redhat.com>
11442
11443 PR tree-optimization/45216
11444 PR tree-optimization/57157
11445 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
11446 the (-Y) & (B - 1) variant if OP is |.
11447 * expmed.c (expand_shift_1): For rotations by const0_rtx just
11448 return shifted. Use (-op1) & (prec - 1) as other_amount
11449 instead of prec - op1.
11450
4502fe8d
MJ		 114512013-05-13 Martin Jambor <mjambor@suse.cz>
11452
11453 PR middle-end/42371
11454 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
11455 (ipa_constant_data): New type.
11456 (ipa_jump_func): Use ipa_constant_data to hold information about
11457 constant jump functions.
11458 (ipa_get_jf_constant): Adjust to jump function type changes.
11459 (ipa_get_jf_constant_rdesc): New function.
11460 (ipa_param_descriptor): New field controlled_uses.
11461 (ipa_get_controlled_uses): New function.
11462 (ipa_set_controlled_uses): Likewise.
11463 * ipa-ref.h (ipa_find_reference): Declare.
11464 * ipa-prop.c (ipa_cst_ref_desc): New type.
11465 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
11466 changes.
11467 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
11468 New parameter cs. Adjust all callers.
11469 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
11470 (remove_described_reference): New function.
11471 (jfunc_rdesc_usable): Likewise.
11472 (try_make_edge_direct_simple_call): Decrement controlled use count,
11473 attempt to remove reference if it hits zero.
11474 (combine_controlled_uses_counters): New function.
11475 (propagate_controlled_uses): Likewise.
11476 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
11477 (ipa_edge_duplication_hook): Duplicate reference descriptions.
11478 (ipa_print_node_params): Print described use counter.
11479 (ipa_write_jump_function): Adjust to jump function type changes.
11480 (ipa_read_jump_function): New parameter CS, pass it to
11481 ipa_set_jf_constant. Adjust caller.
11482 (ipa_write_node_info): Stream controlled use count
11483 (ipa_read_node_info): Likewise.
11484 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
11485 asserting.
11486 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
11487 count. Remove cloning-added reference if it reaches zero.
11488 * ipa-ref.c (ipa_find_reference): New function.
11489
0864bfc2
GG		 114902013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11491
e7208ea3 11492 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG		 11493 alignment values for AMD BD and BT architectures.
11494
640bfeb2
MG		 114952013-05-13 Marc Glisse <marc.glisse@inria.fr>
11496
11497 * tree-vect-generic.c (uniform_vector_p): Move ...
11498 * tree.c (uniform_vector_p): ... here.
11499 * tree.h (uniform_vector_p): Declare it.
11500 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
11501 into a scalar.
11502
3a60f32b
JJ		 115032013-05-13 Jakub Jelinek <jakub@redhat.com>
11504
198fe1bf
JJ		 11505 PR tree-optimization/57230
11506 * tree-ssa-strlen.c (handle_char_store): Record length for
11507 array store from STRING_CST.
11508
3a60f32b
JJ		 11509 PR tree-optimization/57230
11510 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
11511 check.
11512
566be57c
JR		 115132013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
11514
11515 * config/epiphany/epiphany.c (epiphany_init): Check size of
11516 NUM_MODES_FOR_MODE_SWITCHING.
11517 (epiphany_expand_prologue):
11518 Remove CONFIG_REGNUM initial value handling code.
11519 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
11520 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 11521 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR		 11522 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
11523 Don't return 1 for FP_MODE_NONE.
11524 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
11525 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
11526 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
11527 * config/epiphany/epiphany.md (save_config): New pattern.
11528
0f2c2331
UB		 115292013-05-12 Uros Bizjak <ubizjak@gmail.com>
11530
11531 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
11532
5b3f93c7
UB		 115332013-05-10 Uros Bizjak <ubizjak@gmail.com>
11534
11535 * config/i386/i386.md (memory): Handle sseishft1.
11536 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
11537 (*vec_extractv2di_1): Ditto.
11538
1f873f0e
VM		 115392013-05-10 Vladimir Makarov <vmakarov@redhat.com>
11540
11541 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
11542 saved registers.
11543
341427fa 115442013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH		 11545
11546 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
11547 Add mthumb/march=armv7-a multilib.
11548 Add mthumb/march=armv7-r multilib.
11549 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
11550
9e69bdde
RC		 115512013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
11552
11553 * config/v850/t-rtems: Add more multilibs.
11554
9ff09a22
RB		 115552013-05-10 Richard Biener <rguenther@suse.de>
11556
11557 PR tree-optimization/57214
11558 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
11559 not propagate from SSA names that occur in abnormal PHI nodes.
11560
a5e0cd1d
MG		 115612013-05-10 Marc Glisse <marc.glisse@inria.fr>
11562
11563 * stor-layout.c (element_precision): New function.
11564 * machmode.h (element_precision): Declare it.
11565 * tree.c (build_minus_one_cst): New function.
11566 (element_precision): Likewise.
11567 * tree.h (build_minus_one_cst): Declare new function.
11568 (element_precision): Likewise.
11569 * fold-const.c (operand_equal_p): Use element_precision.
11570 (fold_binary_loc): Handle vector types.
11571 * convert.c (convert_to_integer): Use element_precision.
11572 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
11573 separately.
11574
cb2558bc
RS		 115752013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
11576
11577 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
11578 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
11579 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
11580 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
11581 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
11582 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
11583 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
11584 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
11585 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
11586 (Uuw8): New constraints.
11587 (Usb4): Move into alphabetical order.
11588 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
11589 (sd8_operand, ub8_operand, uw8_operand): New predicates.
11590 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
11591 previously unnamed patterns.
11592 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
11593 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
11594 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
11595 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
11596 of set_attr_alternative/if_then_else. Use extended_mips16 instead
11597 of specific lengths.
11598
cb3b8d33
JJ		 115992013-05-10 Jakub Jelinek <jakub@redhat.com>
11600
6f93c008
JJ		 11601 * config/i386/i386.md (rotateinv): New code attr.
11602 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11603 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11604 roll $31, %eax, etc.
11605
cb3b8d33
JJ		 11606 PR tree-optimization/45216
11607 PR tree-optimization/57157
11608 * tree-ssa-forwprop.c (simplify_rotate): New function.
11609 (ssa_forward_propagate_and_combine): Call it.
11610
afb119be
RB		 116112013-05-10 Richard Biener <rguenther@suse.de>
11612
11613 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
11614 disable peeling when we version for aliasing.
11615 (vector_alignment_reachable_p): Honor explicit user alignment.
11616 (vect_supportable_dr_alignment): Likewise.
11617 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
11618 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
11619 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
11620 then peeling to arrange for the cost-model check to come first.
11621
01ae4861
AM		 116222013-05-10 Alan Modra <amodra@gmail.com>
11623
11624 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
11625 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
11626 * configure: Regenerate.
11627
ebc9a431
AM		 116282013-05-10 Alan Modra <amodra@gmail.com>
11629
11630 PR target/55033
11631 * varasm.c (default_elf_select_section): Move !DECL_P check..
11632 (get_named_section): ..to here before calling get_section_name.
11633 Adjust assertion.
11634 (default_section_type_flags): Add DECL_P check.
11635 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
11636 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
11637
d4bca93c
JR		 116382013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
11639
11640 * config/epiphany/epiphany.c (epiphany_expand_prologue):
11641 When using gen_stack_adjust_str with a register offset, add a
11642 REG_FRAME_RELATED_EXPR note.
11643
60ca9a65
UB		 116442013-05-09 Uros Bizjak <ubizjak@gmail.com>
11645
11646 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
11647 (*vec_extractv4si_zext_mem): Ditto.
11648 (*vec_extractv2di): Add 0->x and x->x alternatives.
11649 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
11650 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
11651
a3409c02
JM		 116522013-05-09 Jason Merrill <jason@redhat.com>
11653
0138d6b2
JM		 11654 N3639 C++1y VLA support
11655 * gimplify.c (gimplify_vla_decl): Don't touch an existing
11656 DECL_VALUE_EXPR.
11657
a3409c02
JM		 11658 * tree.c (build_constructor_va): New.
11659 * tree.h: Declare it.
11660
66e6b990
MJ		 116612013-05-09 Martin Jambor <mjambor@suse.cz>
11662
11663 PR lto/57084
11664 * gimple-fold.c (canonicalize_constructor_val): Call
11665 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
11666
64cfa6c0
JH		 116672013-05-09 Jan Hubicka <jh@suse.cz>
11668 Richard Biener <rguenther@suse.de>
11669
11670 PR lto/54095
11671 * symtab.c (symtab_make_decl_local): Do not add private names.
11672
c3167b00
JH		 116732013-05-09 Jan Hubicka <jh@suse.cz>
11674
11675 PR lto/54095
11676 * symtab.c (insert_to_assembler_name_hash): Handle clones.
11677 (unlink_from_assembler_name_hash): Likewise.
11678 (symtab_prevail_in_asm_name_hash, symtab_register_node,
11679 symtab_unregister_node, symtab_initialize_asm_name_hash,
11680 change_decl_assembler_name): Update.
11681
12dc6974
SN		 116822013-05-09 Sofiane Naci <sofiane.naci@arm.com>
11683
11684 * config/aarch64/aarch64.md: New movtf split.
11685 (*movtf_aarch64): Update.
11686 (aarch64_movdi_tilow): Handle TF modes and rename to
11687 aarch64_movdi_<mode>low.
11688 (aarch64_movdi_tihigh): Handle TF modes and rename to
11689 aarch64_movdi_<mode>high
11690 (aarch64_movtihigh_di): Handle TF modes and rename to
11691 aarch64_mov<mode>high_di
11692 (aarch64_movtilow_di): Handle TF modes and rename to
11693 aarch64_mov<mode>low_di
11694 (aarch64_movtilow_tilow): Remove spurious whitespace.
11695 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
11696 splits.
11697 (aarch64_print_operand): Update.
11698
227eb343
AM		 116992013-05-09 Alan Modra <amodra@gmail.com>
11700
11701 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
11702 powerpc64le.
11703 * configure: Regenerate.
11704
0b013847
UB		 117052013-05-08 Uros Bizjak <ubizjak@gmail.com>
11706
11707 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
11708 splitter preparation statements.
11709 * config/i386/sse.md (*vec_extract* splitters): Ditto.
11710 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
11711 adjust_address_nv.
11712
1dc3d6e9
BS		 117132013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11714
11715 * gimple-ssa-strength-reduction.c (count_candidates): Change
11716 return value to int.
11717 (analyze_candidates_and_replace): Change type of length to int.
11718
e61e7d28
UB		 117192013-05-08 Uros Bizjak <ubizjak@gmail.com>
11720
11721 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
11722 (*vec_extract<mode>): Use VI12_128 mode iterator.
11723 (*vec_extract<mode>_mem): Ditto.
11724 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
11725 attribute.
11726
4fbfcf44
DN		 117272013-05-08 Diego Novillo <dnovillo@google.com>
11728
11729 PR bootstrap/54659
11730
11731 Revert:
4fbfcf44
DN		 11732 2012-08-17 Diego Novillo <dnovillo@google.com>
11733
e61e7d28
UB		 11734 PR bootstrap/54281
11735 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
11736 * config.in: Regenerate.
11737 * configure: Regenerate.
11738 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 11739
702d8703
JH		 117402013-05-08 Jan Hubicka <jh@suse.cz>
11741
11742 PR lto/54095
11743 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
11744 * cgraph.h (symtab_node_base): Add unique_name.
11745 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11746 input_overwrite_node, input_varpool_node): Stream unique_name.
11747 * cgraphclones.c (cgraph_create_virtual_clone,
11748 cgraph_function_versioning): Set unique_name.
11749 * ipa.c (function_and_variable_visibility): Set unique_name.
11750
8b28cf47
BS		 117512013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11752
11753 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
11754 (alloc_cand_and_find_basis): Restrict conditional candidate
11755 processing to CAND_MULTs.
11756
e86074fd
JH		 117572013-05-08 Jan Hubicka <jh@suse.cz>
11758
11759 PR lto/54095
11760 lto-symtab.c (lto_symtab_symbol_p): New function.
11761 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
11762 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
11763 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
11764 Skip static symbols.
11765
44398cbe
PC		 117662013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
11767
11768 PR tree-optimization/57200
11769 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
11770 Only call inform if the preceding warning_at returns true.
11771
f6bc1c4a
HS		 117722013-05-07 Han Shen <shenhan@google.com>
11773
11774 * cfgexpand.c (record_or_union_type_has_array_p): New function.
11775 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
11776 * common.opt (fstack-protector-strong): New option.
11777 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
11778 * doc/invoke.texi (Optimization Options): Document
11779 "-fstack-protector-strong".
11780 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
11781
4ffecb1f
SB		 117822013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11783
11784 * config/mips/mips.c (mips_machine_reorg2): Return 0.
11785
5a107a0f
VM		 117862013-05-07 Vladimir Makarov <vmakarov@redhat.com>
11787
11788 * ira.c (update_equiv_regs): Add insn having equiv memory even if
11789 it is not lhs of the insn.
11790 (setup_reg_equiv): Remove insn having equiv memory which it is not
11791 lhs of the insn.
11792 * lra-constraints.c (process_address): Try to improve generation
11793 code for address base + disp.
11794 (lra_constraints): Make correct the code for checking insn setting
11795 up backward equivalence. Remove insn only if it is in the init
11796 insn list.
11797 * lra-eliminations.c (update_reg_eliminate): Change return value.
11798 (lra_eliminate): Use the result.
11799
3f5783ea
UB		 118002013-05-07 Uros Bizjak <ubizjak@gmail.com>
11801
11802 * config/i386/sse.md (ssescalarnummask): New mode attribute.
11803 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
11804 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
11805 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
11806 register target operands.
11807 (*vec_extractv8hi_sse2): New pattern.
11808 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
11809 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
11810 (*vec_extract<mode>_mem): New insn and split pattern.
11811
8a5800b8
CL		 118122013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
11813
11814 * config/arm/arm.c (arm_asan_shadow_offset): New function.
11815 (TARGET_ASAN_SHADOW_OFFSET): Define.
11816 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
11817 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
11818
7bf55a70
BS		 118192013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11820
11821 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
11822 (incr_vec_index): Return -1 if increment not found.
11823 (create_add_on_incoming_edge): Assert if increment not found.
11824 (record_increment): Limit number of increments recorded.
11825 (all_phi_incrs_profitable): Return false if an increment not found.
11826 (replace_profitable_candidates): Don't process increments that were
11827 not recorded.
11828 (analyze_candidates_and_replace): Limit size of incr_vec.
11829
3f8825c0
RB		 118302013-05-07 Richard Biener <rguenther@suse.de>
11831
11832 * calls.c (special_function_p): setjmp-like functions are leaf.
11833 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
11834 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
11835
fd4842cd
SN		 118362013-05-07 Sofiane Naci <sofiane.naci@arm.com>
11837
11838 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
11839 (aarch64_simd_mov<mode>): New expander.
11840 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
11841 (aarch64_simd_mov_to_<mode>high): Likewise.
11842 (aarch64_simd_mov_from_<mode>low): Likewise.
11843 (aarch64_simd_mov_from_<mode>high): Likewise.
11844 (aarch64_dup_lane<mode>): Update.
11845 (aarch64_dup_lanedi): New instruction pattern.
11846 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
11847 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
11848
a7a7d10e
BS		 118492013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11850
11851 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
11852 (replace_mult_candidate): Remove unnecessary argument; remove
11853 unnecessary parameter from call to introduce_cast_before_cand.
11854 (replace_unconditional_candidate): Remove unnecessary parameter
11855 from call to replace_mult_candidate.
11856 (replace_conditional_candidate): Likewise.
11857 (insert_initializers): Use make_temp_ssa_name.
11858 (introduce_cast_before_cand): Remove unnecessary argument; use
11859 make_temp_ssa_name.
11860 (replace_one_candidate): Remove unnecessary argument; remove
11861 unnecessary parameter from calls to introduce_cast_before_cand.
11862 (replace_profitable_candidates): Remove unnecessary parameters
11863 from calls to replace_one_candidate.
11864
29105868
BS		 118652013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11866
11867 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
11868 phi def as possibly hiding a basis for a CAND_ADD whose operands
11869 have been commuted in the analysis.
11870 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
11871
4095f9fa
N		 118722013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11873
11874 * config/aarch64/aarch64.md
11875 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
11876 shift value between 0-4.
11877
39e843e8
RB		 118782013-05-07 Richard Biener <rguenther@suse.de>
11879
11880 * double-int.h (rshift): New overload.
11881 * double-int.c (rshift): New function.
11882 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
11883 (create_reference_ops_from_ref): Remove.
11884 (vn_reference_insert): Use shared ops for constructing the
11885 reference and copy it.
11886
0a1a83cb
RB		 118872013-05-07 Richard Biener <rguenther@suse.de>
11888
11889 PR middle-end/57190
11890 * tree-eh.c (sink_clobbers): Properly propagate
11891 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
11892
43303d6f
JJ		 118932013-05-07 Jakub Jelinek <jakub@redhat.com>
11894
ba7e83f8
JJ		 11895 PR tree-optimization/57149
11896 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
11897 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
11898 collect_phi_def_edges, execute_late_warn_uninitialized): Use
11899 uninit_undefined_value_p instead of ssa_undefined_value_p.
11900
43303d6f
JJ		 11901 PR debug/57184
11902 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
11903 for modifier == EXPAND_INITIALIZER.
11904
14523c25
AB		 119052013-05-07 Anton Blanchard <anton@samba.org>
11906
11907 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
11908 for powerpc64 little endian.
11909 * configure: Regenerate.
11910
cb7c8be9
GS		 119112013-05-06 Graham Stott <grahams@btinternet.com>
11912
11913 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
11914 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
11915 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
11916 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
11917
119182013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS		 11919
11920 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
11921 codes which allow non-lvalues.
11922
9a0ee7b0
MG		 119232013-05-06 Marc Glisse <marc.glisse@inria.fr>
11924
11925 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
11926 components are all 1s.
11927 (integer_minus_onep): New function.
11928 * tree.h (integer_minus_onep): Declare it.
11929 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
11930 integer_minus_onep instead of integer_all_onesp.
11931
f2c17ea9
OE		 119322013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11933
11934 PR target/52933
11935 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
11936 variations of these patterns.
11937
f75e6a51
UB		 119382013-05-06 Uros Bizjak <ubizjak@gmail.com>
11939
11940 * config/i386/i386.md (isa): Add x64_sse4 member.
11941 (enabled): Handle x64_sse4.
11942 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
11943 instruction for 64bit SSE4_1 targets. Update insn attributes.
11944 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
11945 instruction for SSE4_1 targets. Update insn attributes.
11946 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11947 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
11948 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
11949 const_1 selector.
11950 (*vec_extractv4si): Rename from *sse4_1_pextrd.
11951 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
11952 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
11953
a986d468
OE		 119542013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11955
11956 PR target/57108
11957 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
11958
382522cb
MK		 119592013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
11960
11961 * final.c (do_assembler_dialects): Don't handle curly braces and
11962 vertical bar escaped by % as dialect delimiters.
11963 (output_asm_insn): Print curly braces and vertical bar if escaped
11964 by % and ASSEMBLER_DIALECT defined.
11965 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
11966 * doc/tm.texi: Regenerated.
11967
fb0d5c60
SB		 119682013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11969
fb0d5c60
SB		 11970 * config/mips/mips.c: Include tree-pass.h.
11971 (mips_reorg): Split in pre- and post-dbr_schedule parts.
11972 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
11973 (pass_mips_machine_reorg2): New machine specific pass.
11974 (insert_pass_mips_machine_reorg2): New pass plugin definition.
11975 (mips_option_override): Register the new pass.
11976 * rtl.h (cleanup_barriers): Remove prototype.
11977 (dbr_schedule): Likewise.
11978 * jump.c (cleanup_barriers): Make static.
11979 * reorg.c (dbr_schedule): Likewise.
11980
aa06a978
RB		 119812013-05-06 Richard Biener <rguenther@suse.de>
11982
11983 PR tree-optimization/57185
11984 * tree-parloops.c (add_field_for_reduction): Handle anonymous
11985 SSA names properly.
11986
0b953bec
UB		 119872013-05-06 Uros Bizjak <ubizjak@gmail.com>
11988
11989 PR target/57106
11990 * config/i386/i386.c (add_parameter_dependencies): Add dependence
11991 between "first_arg" and "insn", not "last" and "insn".
11992
28708525
WS		 119932013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11994
11995 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
11996 (find_candidates_in_block): Re-enable slsr_process_phi.
11997 (create_phi_basis): Fix double counting of candidate adjustment.
11998
0107dca2
RB		 119992013-05-06 Richard Biener <rguenther@suse.de>
12000
12001 PR middle-end/57147
12002 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12003 the edge is also fallthru, preserve it and just clear the
12004 abnormal flag.
12005 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12006 also complex, preserve that and just clear the fallthru flag.
12007 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12008 update virtual operands.
12009
470d4d13
AM		 120102013-05-06 Alan Modra <amodra@gmail.com>
12011
12012 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12013 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12014 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12015 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12016 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12017 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12018
8f1dbf8d
AM		 120192013-05-06 Alan Modra <amodra@gmail.com>
12020
12021 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12022 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12023 (DEFAULT_ASM_ENDIAN): Define.
12024 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12025 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12026 Update -K PIC clause from sysv4.h.
12027 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12028 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12029
54890767
AM		 120302013-05-06 Alan Modra <amodra@gmail.com>
12031
12032 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12033 twice for little-endian.
12034 (ashrdi3_no_power, ashrdi3): Support little-endian.
12035
2353515d
OE		 120362013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12037
12038 PR target/55303
12039 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12040 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12041 related expanders.
12042 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12043 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12044 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12045 New predicates.
12046
33e67557
SB		 120472013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12048 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12049
12050 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12051 * config/pa/pa.opt: Make mbig-switch a no-op.
12052 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12053 (CASE_VECTOR_MODE): Always return SImode.
12054 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12055 for the !TARGET_BIG_SWITCH case.
12056 * config/pa/pa-linux.h: Likewise.
12057 * config/pa/pa-openbsd.h: Likewise.
12058 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12059 * config/pa/pa.md (short_jump): Remove define_insn.
12060 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12061 (casesi0): Remove define_insn.
12062 (type): Remove btable_branch.
12063 (pa_combine_type): Likewise.
12064 (in_nullified_branch_delay): Likewise.
12065 (in_call_delay): Likewise.
12066 (define_delay): Likewise.
12067 (define_insn_reservation "Z3"): Likewise.
12068 (define_insn_reservation "Z4"): Likewise.
12069 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12070 (pa_adjust_insn_length): Remove adjustment for btable branches.
12071 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12072 and mno-big-switch
12073
3095685e
UB		 120742013-05-05 Uros Bizjak <ubizjak@gmail.com>
12075
12076 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12077 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12078 Add m->r,x alternatives.
12079 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12080 splitters using SWI48x mode iterator.
12081 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12082 TARGET_64BIT. Add m->x alternative.
12083 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12084 Add o->x alternative. Enable for TARGET_SSE.
12085 (sse_storeq): Remove expander.
12086 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12087 with memory input operand.
12088 (*vec_extractv2di_1 splitter): New.
12089 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12090 * config/i386/i386.md (ssevecmodelower): New mode attribute.
12091
4b36ae28
SB		 120922013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12093
12094 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
12095 (INT_LOWPART): Delete.
12096 (extract_MB): Adjust.
12097 (extract_ME): Adjust.
12098 (print_operand): Adjust.
12099
da226db2
SB		 121002013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12101
12102 * config/rs6000/predicates.md (reg_or_add_cint_operand,
12103 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12104 (reg_or_logical_cint_operand, easy_fp_constant,
12105 logical_const_operand): Delete "CONST_DOUBLE" case.
12106 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
12107 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 12108 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB		 12109 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
12110 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
12111 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
12112 test.
12113 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
12114 CONST_DOUBLE DImode/VOIDmode case.
12115 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
12116 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12117 CONST_DOUBLE VOIDmode case.
12118 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12119 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
12120 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
12121 Delete CONST_DOUBLE case.
12122 (splitters for mov FMOVE64 const_double): Delete
12123 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12124 "HOST_BITS_PER_WIDE_INT >= 64" test.
12125 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
12126 case.
12127 (mov DI const_double): Delete.
12128
40de22d6
JJ		 121292013-05-04 Jakub Jelinek <jakub@redhat.com>
12130
12131 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
12132 on op shows all bits zero in mode of a lowpart subreg, return zero.
12133
5ec6aff2
MM		 121342013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
12135
12136 PR target/57150
12137 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
12138 to save TFmode registers and DImode to save TImode registers for
12139 caller save operations.
12140 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
12141 mark being partially clobbered since they only use the first
12142 double word.
12143
12144 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
12145 and TDmode only use the upper 64-bits of each VSX register.
12146
2cefad90
BS		 121472013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12148
12149 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
12150 (find_candidates_in_block): Disable slsr_process_phi.
12151
d6d7eee1
GW		 121522013-05-03 Guozhi Wei <carrot@google.com>
12153
12154 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB		 12155 constructor to ...
12156 (build_init_ctor): ... here.
d6d7eee1 12157
9b92d12b
BS		 121582013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12159
12160 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12161 (slsr_cand_d): Redefine def_phi.
12162 (stride_status, phi_adjust_status, count_phis_status): New enums.
12163 (find_phi_def): New.
12164 (find_basis_for_base_expr): New.
12165 (find_basis_for_candidate): Handle hidden bases.
12166 (alloc_cand_and_find_basis): Handle phi candidates.
12167 (slsr_process_phi): New.
12168 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12169 (create_mul_imm_cand): Likewise.
12170 (create_add_ssa_cand): Exclude phi base candidates.
12171 (create_add_imm_cand): Likewise.
12172 (slsr_process_cast): Likewise.
12173 (slsr_process_copy): Likewise.
12174 (find_candidates_in_block): Handle phi candidates.
12175 (dump_candidate): Likewise.
12176 (unconditional_cands): Delete.
12177 (unconditional_cands_with_known_stride_p): Delete.
12178 (phi_dependent_cand_p): New.
12179 (cand_increment): Handle phi-dependent candidates.
12180 (replace_dependent): Delete.
12181 (replace_mult_candidate): New.
12182 (replace_unconditional_candidate): New.
12183 (incr_vec_index): Move to avoid forward reference.
12184 (create_add_on_incoming_edge): New.
12185 (create_phi_basis): New.
12186 (replace_dependents): Delete.
12187 (replace_conditional_candidate): New.
12188 (phi_add_costs): New.
12189 (replace_uncond_cands_and_profitable_phis): New.
12190 (record_increment): Handle phi adjustments.
12191 (record_phi_increments): New.
12192 (record_increments): Handle phi adjustments.
12193 (phi_incr_cost): New.
12194 (lowest_cost_path): Handle phis.
12195 (total_savings): Likewise.
12196 (analyze_increments): Likewise.
12197 (ncd_with_phi): New.
12198 (ncd_of_cand_and_phis): New.
12199 (nearest_common_dominator_for_cands): Handle phi increments.
12200 (all_phi_incrs_profitable): New.
12201 (replace_profitable_candidates): Handle phi-dependent candidates.
12202 (analyze_candidates_and_replace): Likewise.
12203
68f073d4
TJ		 122042013-05-03 Teresa Johnson <tejohnson@google.com>
12205
12206 PR bootstrap/57154
12207 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12208 do not exceed REG_BR_PROB_BASE.
12209
a4ee7cb9
JL		 122102013-05-03 Jeff Law <law@redhat.com>
12211
ade67f70 12212 PR tree-optimization/57144
a4ee7cb9
JL		 12213 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12214 operand of the condition will bit into the new type when eliminating
12215 a cast feeding a condition.
12216
47954c4e
JJ		 122172013-05-03 Jakub Jelinek <jakub@redhat.com>
12218
12219 PR rtl-optimization/57130
3095685e
UB		 12220 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12221 of COMPARE as in_code to the recursive call if needed.
47954c4e 12222
3c21604f
UB		 122232013-05-03 Uros Bizjak <ubizjak@gmail.com>
12224
12225 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12226 (enabled): Handle new members.
12227 * config/i386/sse.md (*vec_concatv2si): Merge from
12228 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12229 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12230
12211b99 122312013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR		 12232
12233 PR tree-optimization/57027
12234 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12235 for fnms opportunity, check we got the prerequisite kind
12236 of tree / gimple before using accessor functions.
12237
07bfc9ec
RB		 122382013-05-03 Richard Biener <rguenther@suse.de>
12239
12240 * double-int.h (lshift): New overload without precision
12241 and arith argument.
12242 (operator *=, operator +=, operator -=): Move ...
12243 * double-int.c (operator *=, operator +=, operator -=): ... here
12244 and implement more efficiently.
12245 (mul_double_with_sign): Remove.
12246 (lshift_double): Adjust to take unsinged shift argument, push
12247 dispatching code to callers.
12248 (mul_double_wide_with_sign): Add early out for callers that
12249 are not interested in high parts or overflow.
12250 (lshift): New function.
12251 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12252 dispatch code here.
12253 (lrotate, rrotate): Use logical shifts.
12254 * expr.c (get_inner_reference): Use lshift.
12255 * fixed-value.c (do_fixed_divide): Likewise.
12256 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12257 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12258 (indirect_refs_may_alias_p): Likewise.
12259 (stmt_kills_ref_p_1): Likewise.
12260
7769bb64
VP		 122612013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12262
12263 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12264
1d0c8e5c
VP		 122652013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12266
12267 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12268 scalar form of FABD instruction.
12269
f15feaf9
VM		 122702013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12271
12272 * lra-constraints.c (process_alt_operands): Add checking alt
12273 number to choose the best alternative.
12274
d90e76d4
RB		 122752013-05-02 Richard Biener <rguenther@suse.de>
12276
12277 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12278 bitmap and its handling.
12279 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12280
9f8e7a96
RB		 122812013-05-02 Richard Biener <rguenther@suse.de>
12282
12283 PR middle-end/57140
12284 * tree-inline.c (copy_loops): Properly handle removed loops.
12285 (copy_cfg_body): Mark destination loops for fixup if source
12286 loops needed fixup.
12287
f3a81b39
GY		 122882013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12289
12290 PR target/56732
12291 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12292 generating simple_return for naked functions.
12293
7b920a9a
MJ		 122942013-05-02 Martin Jambor <mjambor@suse.cz>
12295
12296 PR middle-end/56988
12297 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12298 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12299 flags match.
12300 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12301 ipa_agg_replacement_value structures.
12302 (known_aggs_to_agg_replacement_list): Likewise.
12303 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12304 (read_agg_replacement_chain): Likewise.
12305 (ipcp_transform_function): Also check that by_ref flags match.
12306
2c41c19d
RB		 123072013-05-02 Richard Biener <rguenther@suse.de>
12308
12309 * graphds.h (struct graph): Add obstack member.
12310 * graphds.c (new_graph): Initialize obstack and allocate
12311 vertices from it.
12312 (add_edge): Allocate edge from the obstack.
3c21604f 12313 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 12314
8b47039c
TJ		 123152013-05-02 Teresa Johnson <tejohnson@google.com>
12316
12317 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12318 divides.
12319 * cfg.c (update_bb_profile_for_threading): Ditto.
12320 * tree-inline.c (copy_bb): Ditto.
12321 (copy_edges_for_bb): Ditto.
12322 (initialize_cfun): Ditto.
12323 (copy_cfg_body): Ditto.
12324 (expand_call_inline): Ditto.
12325 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12326 (estimate_node_size_and_time): Ditto.
12327 (inline_merge_summary): Ditto.
12328 * cgraphclones.c (cgraph_clone_edge): Ditto.
12329 (cgraph_clone_node): Ditto.
12330 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12331 (compute_trg_info): Ditto.
12332
da65928c
IB		 123332013-05-02 Ian Bolton <ian.bolton@arm.com>
12334
12335 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12336 S reg when fp attribute set.
12337 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12338
add3c965
IB		 123392013-05-02 Ian Bolton <ian.bolton@arm.com>
12340
12341 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12342 New pattern.
12343 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12344 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12345 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12346
fdd43ac4
RB		 123472013-05-02 Richard Biener <rguenther@suse.de>
12348
12349 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12350 (struct instantiate_cache_entry): New type.
12351 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12352 (struct instantiate_cache_type): New type.
12353 (set_instantiated_value, get_instantiated_value): Remove.
12354 (get_instantiated_value_entry): New function.
12355 (instantiate_scev_name): Use the new cache and adjust.
12356 (instantiate_scev_poly): Adjust.
12357 (instantiate_scev_binary): Likewise.
12358 (instantiate_array_ref): Likewise.
12359 (instantiate_scev_convert): Likewise.
12360 (instantiate_scev_not): Likewise.
12361 (instantiate_scev_3): Likewise.
12362 (instantiate_scev_2): Likewise.
12363 (instantiate_scev_r): Likewise.
12364 (instantiate_scev): Likewise.
12365 (resolve_mixers): Likewise.
12366
36ff9dfb
VM		 123672013-05-01 Vladimir Makarov <vmakarov@redhat.com>
12368
12369 PR target/57091
12370 * lra-constraints.c (best_small_class_operands_num): Remove.
12371 (process_alt_operands): Remove small_class_operands_num. Take
12372 small classes operands into losers and only if the operand is not
12373 matched. Modify debugging output.
12374 (curr_insn_transform): Remove best_small_class_operands_num.
12375 Print insn name.
12376
36054fab
JG		 123772013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12378
12379 * config/aarch64/aarch64-builtins.c
12380 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
12381 * config/aarch64/aarch64-simd-builtins.def
12382 (reduc_splus_): Add new modes.
12383 (reduc_uplus_): New.
12384 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
12385 (reduc_uplus_v4sf): Likewise.
12386 (reduc_splus_v4sf): Likewise.
12387 (aarch64_addv<mode>): Likewise.
12388 (reduc_uplus_<mode>): Likewise.
12389 (reduc_splus_<mode>): Likewise.
12390 (aarch64_addvv2di): Likewise.
12391 (reduc_uplus_v2di): Likewise.
12392 (reduc_splus_v2di): Likewise.
12393 (aarch64_addvv2si): Likewise.
12394 (reduc_uplus_v2si): Likewise.
12395 (reduc_splus_v2si): Likewise.
12396 (reduc_<sur>plus_<mode>): New.
12397 (reduc_<sur>plus_v2di): Likewise.
12398 (reduc_<sur>plus_v2si): Likewise.
12399 (reduc_<sur>plus_v4sf): Likewise.
12400 (aarch64_addpv4sf): Likewise.
12401 * config/aarch64/arm_neon.h
12402 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
12403 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
12404 add UNSPEC_SADDV, UNSPEC_UADDV.
12405 (SUADDV): New.
12406 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
12407
6dce23a8
JG		 124082013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12409
12410 * config/aarch64/arm_neon.h
12411 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
12412
1598945b
JG		 124132013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12414
12415 * config/aarch64/aarch64-builtins
12416 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
12417
998eaf97
JG		 124182013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12419
12420 * config/aarch64/aarch64-simd-builtins.def
12421 (reduc_smax_): New.
12422 (reduc_smin_): Likewise.
12423 (reduc_umax_): Likewise.
12424 (reduc_umin_): Likewise.
12425 (reduc_smax_nan_): Likewise.
12426 (reduc_smin_nan_): Likewise.
12427 (fmax): Remove.
12428 (fmin): Likewise.
12429 (smax): Update for V2SF, V4SF and V2DF modes.
12430 (smin): Likewise.
12431 (smax_nan): New.
12432 (smin_nan): Likewise.
12433 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
12434 (<su><maxmin><mode>3): ...This, refactor.
12435 (s<maxmin><mode>3): New.
12436 (<maxmin_uns><mode>3): Likewise.
12437 (reduc_<maxmin_uns>_<mode>): Refactor.
12438 (reduc_<maxmin_uns>_v4sf): Likewise.
12439 (reduc_<maxmin_uns>_v2si): Likewise.
12440 (aarch64_<fmaxmin><mode>: Remove.
12441 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
12442 new builtin names.
12443 (vmin<q>_f<32,64>): Likewise.
12444 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12445 (FMAXMIN): New.
12446 (su): Add mappings for smax, smin, umax, umin.
12447 (maxmin): New.
12448 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12449 (FMAXMIN): Rename as...
12450 (FMAXMIN_UNS): ...This.
12451 (maxminv): Remove.
12452 (fmaxminv): Likewise.
12453 (fmaxmin): Likewise.
12454 (maxmin_uns): New.
12455 (maxmin_uns_op): Likewise.
12456
bd11644e
JG		 124572013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12458
12459 * config/aarch64/arm_neon.h
12460 (vac<ge, gt><sd>_f<32, 64>): Rename to...
12461 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
12462 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
12463
75dd5ace
JG		 124642013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12465
12466 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
12467 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
12468
7c19979f
JG		 124692013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12470
12471 * config/aarch64/aarch64-simd.md
12472 (vcond<mode>_internal): Handle special cases for constant masks.
12473 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
12474 (vcondu<mode><mode>): Likewise.
12475 (vcond<v_cmp_result><mode>): New.
12476
bb60efd9
JG		 124772013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12478
12479 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
12480 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
12481 * config/aarch64/aarch64-simd-builtins.def
12482 (cmeq): Update to BUILTIN_VALLDI.
12483 (cmgt): Likewise.
12484 (cmge): Likewise.
12485 (cmle): Likewise.
12486 (cmlt): Likewise.
12487 * config/aarch64/arm_neon.h
12488 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
12489 to builtins or C as appropriate.
12490
889b9412
JG		 124912013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12492
12493 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
12494 (cmgeu): ...This.
12495 (cmhi): Rename to...
12496 (cmgtu): ...This.
12497 * config/aarch64/aarch64-simd.md
12498 (simd_mode): Add SF.
12499 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
12500 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
12501 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
12502 (cstore<mode>_neg): ...This.
12503 * config/aarch64/iterators.md
12504 (VALLF): new.
12505 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
12506 (COMPARISONS): New.
12507 (UCOMPARISONS): Likewise.
12508 (optab): Add missing comparisons.
12509 (n_optab): New.
12510 (cmp_1): Likewise.
12511 (cmp_2): Likewise.
12512 (CMP): Likewise.
12513 (cmp): Remove.
12514 (VCMP_S): Likewise.
12515 (VCMP_U): Likewise.
12516 (V_cmp_result): Add DF, SF modes.
12517 (v_cmp_result): Likewise.
12518 (v): Likewise.
12519 (vmtype): Likewise.
12520 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
12521
0a7dbb76
GY		 125222013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
12523
12524 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
12525 define_insn to define_insn_and_split.
12526 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
12527 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
12528 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
12529 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
12530 (thumb2_negscc): Likewise.
12531
fb614ca6
GY		 125322013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12533
12534 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
12535
9e64a0bf
GY		 125362013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12537
12538 * config/arm/thumb2.md: Remove trailing whitespaces.
12539
d6b28156
RS		 125402013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12541
12542 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
12543 Use gen_int_mode rather than GEN_INT.
12544
f91674c3
L		 125452013-04-30 H.J. Lu <hongjiu.lu@intel.com>
12546
b0dec607 12547 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L		 12548 debug_gimple_stmt.
12549
3551257c
RB		 125502013-04-30 Richard Biener <rguenther@suse.de>
12551
12552 PR middle-end/57122
3c21604f 12553 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 12554
4c1aff1c
RB		 125552013-04-30 Richard Biener <rguenther@suse.de>
12556
12557 PR middle-end/57107
12558 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
12559
54b8379a
AB		 125602013-04-30 Andrey Belevantsev <abel@ispras.ru>
12561
12562 PR rtl-optimization/56957
12563 PR rtl-optimization/57105
54b8379a
AB		 12564 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
12565 variable. Use just INSN_UID for determining whether an insn
12566 should be only disconnected from the insn stream.
12567 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
12568
abc27962
JJ		 125692013-04-30 Jakub Jelinek <jakub@redhat.com>
12570
12571 PR tree-optimization/57104
12572 * tsan.c (instrument_expr): Don't instrument accesses to
12573 DECL_HARD_REGISTER VAR_DECLs.
12574
0fc822d0
RB		 125752013-04-30 Richard Biener <rguenther@suse.de>
12576
12577 * function.h (loops_for_fn): New inline function.
12578 (set_loops_for_fn): Likewise.
12579 * cfgloop.h (place_new_loop): Add struct function parameter.
12580 (get_loop): Likewise.
12581 (get_loops): Likewise.
12582 (number_of_loops): Likewise.
12583 (fel_next): Adjust.
12584 (fel_init): Likewise.
12585 * cfg.c (get_loop_copy): Adjust.
12586 * cfgloop.c (flow_loops_dump): Likewise.
12587 (record_loop_exits): Likewise.
12588 (verify_loop_structure): Likewise.
12589 * cfgloopanal.c (mark_irreducible_loops): Likewise.
12590 (estimate_reg_pressure_cost): Likewise.
12591 (mark_loop_exit_edges): Likewise.
12592 * cfgloopmanip.c (place_new_loop): Likewise.
12593 (add_loop): Likewise.
12594 (duplicate_loop): Likewise.
12595 * graph.c (draw_cfg_nodes): Likewise.
12596 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
12597 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
12598 (extract_affine_chrec): Likewise.
12599 (build_scop_iteration_domain): Likewise.
12600 * graphite.c (graphite_initialize): Likewise.
12601 * ira-build.c (create_loop_tree_nodes): Likewise.
12602 (more_one_region_p): Likewise.
12603 (rebuild_regno_allocno_maps): Likewise.
12604 (mark_loops_for_removal): Likewise.
12605 (mark_all_loops_for_removal): Likewise.
12606 (remove_unnecessary_regions): Likewise.
12607 (ira_build): Likewise.
12608 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
12609 * loop-init.c (fix_loop_structure): Likewise.
12610 (gate_rtl_move_loop_invariants): Likewise.
12611 (gate_rtl_unswitch): Likewise.
12612 (gate_rtl_unroll_and_peel_loops): Likewise.
12613 (rtl_doloop): Likewise.
12614 * lto-streamer-in.c (input_cfg): Likewise.
12615 * lto-streamer-out.c (output_cfg): Likewise.
12616 * modulo-sched.c (sms_schedule): Likewise.
12617 * predict.c (tree_estimate_probability): Likewise.
12618 (tree_estimate_probability_driver): Likewise.
12619 (estimate_loops): Likewise.
12620 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
12621 (move_sese_region_to_fn): Likewise.
12622 (debug_loop_num): Likewise.
12623 * tree-chrec.c (chrec_evaluate): Likewise.
12624 (hide_evolution_in_other_loops_than_loop): Likewise.
12625 (chrec_component_in_loop_num): Likewise.
12626 (reset_evolution_in_loop): Likewise.
12627 (evolution_function_is_invariant_rec_p): Likewise.
12628 * tree-if-conv.c (main_tree_if_conversion): Likewise.
12629 * tree-inline.c (copy_loops): Likewise.
12630 (copy_cfg_body): Likewise.
12631 (tree_function_versioning): Likewise.
12632 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
12633 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
12634 Likewise.
12635 (add_to_evolution_1): Likewise.
12636 (scev_const_prop): Likewise.
12637 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
12638 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
12639 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
12640 (tree_ssa_lim_initialize): Likewise.
12641 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
12642 (verify_loop_closed_ssa): Likewise.
12643 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
12644 (tree_ssa_loop_im): Likewise.
12645 (tree_ssa_loop_unswitch): Likewise.
12646 (tree_vectorize): Likewise.
12647 (check_data_deps): Likewise.
12648 (tree_ssa_loop_ivcanon): Likewise.
12649 (tree_ssa_loop_bounds): Likewise.
12650 (tree_complete_unroll): Likewise.
12651 (tree_complete_unroll_inner): Likewise.
12652 (tree_parallelize_loops): Likewise.
12653 (tree_ssa_loop_prefetch): Likewise.
12654 (tree_ssa_loop_ivopts): Likewise.
12655 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
12656 * tree-vectorizer.c (vectorize_loops): Likewise.
12657
37953bd3
MF		 126582013-04-29 Mike Frysinger <vapier@gentoo.org>
12659
12660 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
12661 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
12662 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
12663 with EABI_LINK_SPEC.
12664
f9ed28db
UB		 126652013-04-29 Uros Bizjak <ubizjak@gmail.com>
12666
12667 PR target/44578
12668 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
12669 alternative.
12670
deca73f5
VM		 126712013-04-29 Vladimir Makarov <vmakarov@redhat.com>
12672
12673 PR target/57097
37953bd3 12674 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM		 12675 using memory for pseudos. Print cost dump for alternatives.
12676 Modify cost values for conflicts with early clobbers.
12677 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
12678
fdca7d03
UB		 126792013-04-29 Uros Bizjak <ubizjak@gmail.com>
12680
12681 PR target/57098
12682 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
12683
0b064172
IB		 126842013-04-29 Ian Bolton <ian.bolton@arm.com>
12685
12686 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
12687 from/to S register.
12688 (movdi_aarch64): Support LDR/STR from/to D register.
12689
473cec55
IB		 126902013-04-29 Ian Bolton <ian.bolton@arm.com>
12691
12692 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
12693 or higher by default.
12694
a6f30e66
RB		 126952013-04-29 Richard Biener <rguenther@suse.de>
12696
12697 PR middle-end/57075
12698 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
12699 even if not adding abnormal edges for calls that can make
12700 abnormal gotos.
12701
0c2b2040
RB		 127022013-04-29 Richard Biener <rguenther@suse.de>
12703
12704 PR middle-end/57103
12705 * tree-cfg.c (move_stmt_op): Fix condition under which to update
12706 TREE_BLOCK.
12707 (move_stmt_r): Remove redundant checking.
12708
f41f80f9
TJ		 127092013-04-29 Teresa Johnson <tejohnson@google.com>
12710
12711 PR bootstrap/57077
12712 * basic-block.h (apply_scale): New function.
12713 (apply_probability): Use apply_scale.
12714 * gimple-streamer-in.c (input_bb): Ditto.
12715 * lto-streamer-in.c (input_cfg): Ditto.
12716 * lto-cgraph.c (merge_profile_summaries): Ditto.
12717 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 12718 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ		 12719 (copy_edges_for_bb): Ditto.
12720 (copy_cfg_body): Ditto.
12721
315bbd2e
TV		 127222013-04-29 Tom de Vries <tom@codesourcery.com>
12723
12724 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
12725 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
12726 (tail_merge_optimize): Handle current_loops == NULL.
12727
ebbd90d8
JL		 127282013-04-26 Jeff Law <law@redhat.com>
12729
12730 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
12731 (simplify_cond_using_ranges): Generalize code to simplify
12732 COND_EXPRs where one argument is a constant and the other
12733 is an SSA_NAME created by an integral type conversion.
12734
8b9b57eb
KT		 127352013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12736
12737 * config/arm/arm.md (store_minmaxsi): Use only when
12738 optimize_insn_for_size_p.
12739
9498e5dc
CB		 127402013-04-29 Christian Bruel <christian.bruel@st.com>
12741
12742 PR target/57108
12743 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
12744
6093bc06
RB		 127452013-04-29 Richard Biener <rguenther@suse.de>
12746
12747 PR middle-end/57089
fdca7d03
UB		 12748 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
12749 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB		 12750 (expand_omp_for_generic): Properly add loops.
12751 (expand_omp_for_static_nochunk): Likewise.
12752 (expand_omp_for_static_chunk): Likewise.
12753 (expand_omp_for): For the degenerate case fixup loops.
12754 (expand_omp_sections): Fix default bb placement in loops.
12755 (expand_omp_atomic_pipeline): Properly add loops.
12756
84aacbfd
KT		 127572013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12758
12759 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
12760
8dee4479
TV		 127612013-04-29 Tom de Vries <tom@codesourcery.com>
12762
12763 * tree-ssa-tail-merge.c: Update header comment.
12764
47934dc4
JG		 127652013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12766
12767 * config/aarch64/arm_neon.h
12768 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
12769 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
12770 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
12771 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
12772 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
12773 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
12774 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
12775 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
12776
384be29f
JG		 127772013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12778
12779 * config/aarch64/aarch64-simd.md
12780 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
12781 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
12782 fix_trunc, fixuns_trunc.
12783 (ftrunc<VDQF:mode>2): New.
12784 * config/aarch64/iterators.md (optab): Add fix, fixuns.
12785 (fix_trunc_optab): New.
12786
0386b123
JG		 127872013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12788
12789 * config/aarch64/aarch64-builtins.c
12790 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
12791 iceilf, lround, iroundf.
12792
00fcb892
UB		 127932013-04-29 Uros Bizjak <ubizjak@gmail.com>
12794
12795 PR target/54349
12796 * config/i386/i386.h (enum ix86_tune_indices)
12797 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
12798 New, split from X86_TUNE_INTER_UNIT_MOVES.
12799 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
12800 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
12801 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
12802 (TARGET_INTER_UNIT_MOVES): Remove.
12803 * config/i386/i386.c (initial_ix86_tune_features): Update.
12804 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
12805 (ix86_expand_convert_uns_didf_sse): Use
12806 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12807 (ix86_expand_vector_init_one_nonzero): Ditto.
12808 (ix86_expand_vector_init_interleave): Ditto.
12809 (inline_secondary_memory_needed): Return true for moves from SSE class
12810 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
12811 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
12812 * config/i386/constraints.md (Yi, Ym): Depend on
12813 TARGET_INTER_UNIT_MOVES_TO_VEC.
12814 (Yj, Yn): New constraints.
12815 * config/i386/i386.md (*movdi_internal): Change constraints of
12816 operand 1 from Yi to Yj and from Ym to Yn.
12817 (*movsi_internal): Ditto.
12818 (*movdf_internal): Ditto.
12819 (*movsf_internal): Ditto.
12820 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
12821 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12822 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
12823 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
12824 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
12825 * config/i386/sse.md (movdi_to_sse): Ditto.
12826 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
12827 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
12828 TARGET_INTER_UNIT_MOVES.
12829 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
12830 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
12831 instead of TARGET_INTER_UNIT_MOVES.
12832 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
12833 operand 1 from Yi to Yj and from Ym to Yn.
12834
4c871069
JG		 128352013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12836
12837 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
12838 (float_truncate_hi_): Likewise.
12839 (float_extend_lo_): Likewise.
12840 (float_truncate_lo_): Likewise.
12841 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
12842 (aarch64_float_extend_lo_v2df): Likewise.
12843 (vec_unpacks_hi_v4sf): Likewise.
12844 (aarch64_float_truncate_lo_v2sf): Likewise.
12845 (aarch64_float_truncate_hi_v4sf): Likewise.
12846 (vec_pack_trunc_v2df): Likewise.
12847 (vec_pack_trunc_df): Likewise.
12848
1709ff9b
JG		 128492013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12850
12851 * config/aarch64/aarch64-builtins.c
12852 (aarch64_fold_builtin): Fold float conversions.
12853 * config/aarch64/aarch64-simd-builtins.def
12854 (floatv2si, floatv4si, floatv2di): New.
12855 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
12856 * config/aarch64/aarch64-simd.md
12857 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
12858 * config/aarch64/iterators.md (FLOATUORS): New.
12859 (optab): Add float, floatuns.
12860 (su_optab): Likewise.
12861
ce966824
JG		 128622013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12863
12864 * config/aarch64/aarch64-builtins.c
12865 (aarch64_builtin_vectorized_function): Use new names for
12866 fcvt builtins.
12867 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
12868 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
12869 (fcvtzu): Split as...
12870 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
12871 (fcvtas): Split as...
12872 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
12873 (fcvtau): Split as...
12874 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
12875 (fcvtps): Split as...
12876 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
12877 (fcvtpu): Split as...
12878 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
12879 (fcvtms): Split as...
12880 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
12881 (fcvtmu): Split as...
12882 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
12883 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
12884 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
12885 (lfrintnusf, lfrintnudf): Likewise.
12886 * config/aarch64/aarch64-simd.md
12887 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
12888 define_insn.
12889 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
12890 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
12891 (fcvt_pattern): Likewise.
12892
b9de24fe
JG		 128932013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12894
12895 * config/aarch64/aarch64-simd.md
12896 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
12897 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
12898
77a205be
JG		 128992013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12900
12901 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
12902 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
12903 (vrnd<a,m,n,p>_f32): Implement using builtins.
12904 (vrnd<i,x><q>_f<32, 64>): New.
12905
0659ce6f
JG		 129062013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12907
12908 * config/aarch64/aarch64-builtins.c
12909 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
12910 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
12911 (frintz): Rename to...
12912 (btrunc): ...this.
12913 (frintp): Rename to...
12914 (ceil): ...this.
12915 (frintm): Rename to...
12916 (floor): ...this.
12917 (frinti): Rename to...
12918 (nearbyint): ...this.
12919 (frintx): Rename to...
12920 (rint): ...this.
12921 (frinta): Rename to...
12922 (round): ...this.
12923 * config/aarch64/aarch64-simd.md
12924 (aarch64_frint<frint_suffix><mode>): Delete.
12925 (<frint_pattern><mode>2): Convert to insn.
12926 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
12927 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
12928 (frint_pattern): Likewise.
12929 (frint_suffix): Likewise.
12930
ea78906a
RB		 129312013-04-29 Richard Biener <rguenther@suse.de>
12932
12933 PR tree-optimization/57081
12934 * loop-init.c: Include tree-flow.h.
12935 (loop_optimizer_finalize): Free number of iteration estimates.
12936 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
12937
baee1763
JJ		 129382013-04-29 Jakub Jelinek <jakub@redhat.com>
12939
94dc5332
JJ		 12940 PR tree-optimization/57083
12941 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
12942 non-singleton shift count range, zero extend low_bound for uns case.
12943
baee1763
JJ		 12944 * config/i386/predicates.md (general_vector_operand): New predicate.
12945 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
12946 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
12947 if they aren't nonimmediate operands. If their original values
12948 satisfy const_vector_equal_evenodd_p, don't shift them.
12949 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
12950 predicates. For the SSE4.1 case force operands[{1,2}] into registers
12951 if not nonimmediate_operand.
12952 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
12953 instead of register_operand.
12954 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
12955
a9073727 129562013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB		 12957
12958 * stor-layout.c (finalize_size_functions): Allocate a structure and
12959 reset cfun before dumping the functions.
12960
ba8011e6
JJ		 129612013-04-27 Jakub Jelinek <jakub@redhat.com>
12962
d6fde69e
JJ		 12963 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
12964
ba8011e6
JJ		 12965 PR target/56866
12966 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
12967 use xop_pmacsdqh if uns_p.
12968 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
12969 the immediate rotate count.
12970
2c62cbaa
VM		 129712013-04-26 Vladimir Makarov <vmakarov@redhat.com>
12972
12973 * rtl.h (struct rtx_def): Add comment for field jump.
12974 (LRA_SUBREG_P): New macro.
12975 * recog.c (register_operand): Check LRA_SUBREG_P.
12976 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12977 stack frame.
12978 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12979 (lra_final_code_change): Skip subreg change for operators.
12980 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12981 if there are no operand changes.
12982 * lra-constraints.c (curr_insn_set): New.
12983 (match_reload): Set LRA_SUBREG_P.
12984 (emit_spill_move): Ditto.
12985 (check_and_process_move): Use curr_insn_set. Process only single
12986 set insns. Don't initialize sec_mem_p and change_p.
12987 (simplify_operand_subreg): Use LRA_SUBREG_P.
12988 (reg_in_class_p): New function.
12989 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12990 of #ifdef. Add code to remove cycling.
12991 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12992 non-null disp. Reload inner instead of disp when base and index
12993 are null. Try to put lo_sum into register.
12994 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12995 (check_and_process_move): Move code for move cost check to
12996 simple_move_p. Remove equiv_substitution.
12997 (simple_move_p): New function.
12998 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12999 curr_insn_set. Call check_and_process_move only for single set
13000 insns. Use the new function. Move call of check_and_process_move
13001 after operand equiv substitution and address process.
13002
e7d764f3
JJ		 130032013-04-26 Jakub Jelinek <jakub@redhat.com>
13004
13005 PR go/57045
13006 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13007 with nonlocal goto receivers or returns twice calls, ignore
13008 unininitialized values from abnormal edges to nl goto receiver
13009 or returns twice call.
13010
41e10689
JJ		 130112013-04-26 Jakub Jelinek <jakub@redhat.com>
13012
13013 PR tree-optimization/57051
13014 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13015 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13016 bitsize.
13017
d7ed20db
RB		 130182013-04-26 Richard Biener <rguenther@suse.de>
13019
13020 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13021 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13022 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13023 (fixup_loop_arrays_after_move): New function.
13024 (move_sese_region_to_fn): Properly outline the loop tree parts
13025 of the SESE region.
13026
df93505e
UB		 130272013-04-26 Uros Bizjak <ubizjak@gmail.com>
13028
13029 * config/i386/i386.md (type, unit): Fix long lines.
13030
dd366ec3
RB		 130312013-04-26 Richard Biener <rguenther@suse.de>
13032
13033 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13034 (lto-streamer-out.o): Likewise.
13035 * cfgloop.c (init_loops_structure): Export, add struct function
13036 argument and adjust.
13037 (flow_loops_find): Adjust.
13038 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13039 (init_loops_structure): Declare.
13040 * lto-streamer-in.c: Include cfgloop.h.
13041 (input_cfg): Input the loop tree.
13042 * lto-streamer-out.c: Include cfgloop.h.
13043 (output_cfg): Output the loop tree.
13044 (output_struct_function_base): Do not drop PROP_loops.
13045
a9e0d843
RB		 130462013-03-26 Richard Biener <rguenther@suse.de>
13047
13048 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13049 (pass_build_cfg): Provide PROP_loops.
13050 (move_sese_region_to_fn): Remove loops that are outlined into fn
13051 for now.
13052 * tree-inline.c: Include cfgloop.h.
13053 (initialize_cfun): Do not drop PROP_loops.
13054 (copy_loops): New function.
13055 (copy_cfg_body): Copy loop structure.
13056 (tree_function_versioning): Initialize destination loop tree.
13057 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13058 (pass_parallelize_loops): Do IL verification.
13059 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13060 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13061 the CFG make sure we fixup loops as well.
13062 * tree-ssa-tail-merge.c: Include cfgloop.h.
13063 (replace_block_by): When merging loop latches mark loops for fixup.
13064 * lto-streamer-out.c (output_struct_function_base): Drop
13065 PROP_loops for now.
13066 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13067 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13068 * ipa-split.c: Include cfgloop.h.
13069 (split_function): Add the new return block to the loop tree root.
13070 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13071 whether we have removed the forwarder block.
13072 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13073 * cfgloop.h (place_new_loop): Declare.
13074 * cfgloopmanip.c (place_new_loop): Export.
13075 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13076 (tree-switch-conversion.o): Likewise.
13077 (tree-complex.o): Likewise.
13078 (tree-inline.o): Likewise.
13079 (tree-ssa-tailmerge.o): Likewise.
13080 (ipa-split.o): Likewise.
13081 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13082 (tree-ssa-copy.o): Likewise.
13083 * tree-switch-conversion.c: Include cfgloop.h
13084 (process_switch): If we emit a bit-test cascade, schedule loops
13085 for fixup.
13086 * tree-complex.c: Include cfgloop.h.
13087 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13088 * asan.c: Include cfgloop.h.
13089 (create_cond_insert_point): Properly add new basic-blocks to
13090 loops, schedule loop fixup.
13091 * cfgloop.c (verify_loop_structure): Check that looks are not
13092 marked for fixup.
13093 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
13094 to loops.
13095 (expand_omp_for_generic): Likewise.
13096 (expand_omp_sections): Likewise.
13097 (expand_omp_atomic_pipeline): Schedule loops for fixup.
13098 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
13099 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
13100 is initialized, not when loops are present.
13101 * tree-parloops.c (parallelize_loops): Remove checking here.
13102 * passes.c (init_optimization_passes): Schedule a copy-propagation
13103 pass before complete unrolling of inner loops.
13104
e78e8a0b
JJ		 131052013-04-26 Jakub Jelinek <jakub@redhat.com>
13106
a2e836b2
JJ		 13107 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
13108 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
13109 (colorize_init): Add argument to _WIN32 version.
13110 * toplev.c: Include diagnostic-color.h.
13111 (process_options): Default to -fdiagnostics-color=auto if
13112 GCC_COLORS env var is in the environment.
13113 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB		 13114 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
13115 env var is in the environment, the default is auto rather than never.
a2e836b2 13116
e78e8a0b
JJ		 13117 * diagnostic.h (file_name_as_prefix): Add context argument.
13118 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
13119 the string as locus.
13120 * langhooks.c (lhd_print_error_function): Adjust caller.
13121
013e5ef9
LC		 131222013-04-25 Lawrence Crowl <crowl@google.com>
13123
13124 * var-tracking.c (shared_hash_def::htab):
13125 Change type to hash_table. Update dependent calls and types.
13126
4a8fb1a1
LC		 131272013-04-25 Lawrence Crowl <crowl@google.com>
13128
13129 * Makefile.in: Update as needed below.
13130
13131 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
13132 Move declaration to after the type's method definitons.
13133
13134 * attribs.c (htab_t scoped_attributes::attribute_hash):
13135 Change type to hash_table. Update dependent calls and types.
13136
13137 * bitmap.c (htab_t bitmap_desc_hash):
13138 Change type to hash_table. Update dependent calls and types.
13139
13140 * cselib.c (htab_t cselib_hash_table):
13141 Change type to hash_table. Update dependent calls and types.
13142
13143 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
13144 (hash_string_slot_node): Move implementation into lto-streamer.h
13145 struct string_slot_hasher.
13146 (eq_string_slot_node): Likewise.
13147
13148 * data-streamer-out.c: Update output_block::string_hash_table
13149 dependent calls and types.
13150
13151 * dwarf2cfi.c (htab_t trace_index):
13152 Change type to hash_table. Update dependent calls and types.
13153
13154 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
13155 Change type to hash_table. Update dependent calls and types.
13156 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
13157 (htab_t optimize_external_refs::map): Likewise.
13158 (htab_t output_comp_unit::extern_map): Likewise.
13159 (htab_t output_comdat_type_unit::extern_map): Likewise.
13160 (htab_t output_macinfo::macinfo_htab): Likewise.
13161 (htab_t optimize_location_lists::htab): Likewise.
13162 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13163
13164 * except.c (htab_t ehspec_hash_type):
13165 Change type to hash_table. Update dependent calls and types.
13166 (assign_filter_values::ttypes): Likewise.
13167 (assign_filter_values::ehspec): Likewise.
13168 (sjlj_assign_call_site_values::ar_hash): Likewise.
13169 (convert_to_eh_region_ranges::ar_hash): Likewise.
13170
13171 * gcse.c (htab_t pre_ldst_table):
13172 Change type to hash_table. Update dependent calls and types.
13173
13174 * ggc-common.c (htab_t saving_htab):
13175 Change type to hash_table. Update dependent calls and types.
13176 (htab_t loc_hash): Likewise.
13177 (htab_t ptr_hash): Likewise.
13178 (call_count): Rename ggc_call_count.
13179 (call_alloc): Rename ggc_call_alloc.
13180 (loc_descriptor): Rename make_loc_descriptor.
13181 (add_statistics): Rename ggc_add_statistics.
13182
13183 * ggc-common.c (saving_htab):
13184 Change type to hash_table. Update dependent calls and types.
13185
13186 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13187 (push_gimplify_context): Likewise.
13188 (pop_gimplify_context): Likewise.
13189 (struct gimple_temp_hash_elt): Added.
13190 (struct gimplify_hasher): Likewise.
13191 (struct gimplify_ctx.temp_htab):
13192 Change type to hash_table. Update dependent calls and types.
13193
13194 * gimple-fold.c: Include gimplify-ctx.h.
13195
13196 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13197 Change type to hash_table. Update dependent calls and types.
13198 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13199 avoid potential global name collision.
13200
13201 * gimplify.c: Include gimplify-ctx.h.
13202 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13203 (htab_t gimplify_ctx::temp_htab):
13204 Update dependent calls and types for new type hash_table.
13205 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13206 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13207
13208 * gimplify-ctx.h: New.
13209 (struct gimple_temp_hash_elt): Move from gimplify.c.
13210 (class gimplify_hasher): New.
13211 (struct gimplify_ctx): Move from gimple.h.
13212 (htab_t gimplify_ctx::temp_htab):
13213 Change type to hash_table. Update dependent calls and types.
13214
13215 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13216 (htab_t ivs_params::newivs_index):
13217 Change type to hash_table. Update dependent calls and types.
13218 (htab_t ivs_params::params_index): Likewise.
13219 (htab_t print_generated_program::params_index): Likewise.
13220 (htab_t gloog::newivs_index): Likewise.
13221 (htab_t gloog::params_index): Likewise.
13222
13223 * graphite.c: Include graphite-htab.h.
13224 4htab_t graphite_transform_loops::bb_pbb_mapping):
13225 Change type to hash_table. Update dependent calls and types.
13226
13227 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13228 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13229 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13230
13231 * graphite-dependences.c: Include graphite-htab.h.
13232 (loop_is_parallel_p): Change hash table type of parameter.
13233
13234 * graphite-htab.h: New.
13235 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13236 (extern find_pbb_via_hash): Move from graphite-poly.h.
13237 (extern loop_is_parallel_p): Move from graphite-poly.h.
13238 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13239
13240 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13241 (extern loop_is_parallel_p): Move to graphite-htab.h.
13242 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13243
13244 * haifa-sched.c (htab_t delay_htab):
13245 Change type to hash_table. Update dependent calls and types.
13246 (htab_t delay_htab_i2): Likewise.
13247
13248 * ira-color.c (htab_t allocno_hard_regs_htab):
13249 Change type to hash_table. Update dependent calls and types.
13250
13251 * ira-costs.c (htab_t cost_classes_htab):
13252 Change type to hash_table. Update dependent calls and types.
13253
13254 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13255 Change type to hash_table. Update dependent calls and types.
13256
13257 * loop-iv.c (htab_t bivs):
13258 Change type to hash_table. Update dependent calls and types.
13259
13260 * loop-unroll.c (htab_t opt_info::insns_to_split):
13261 Change type to hash_table. Update dependent calls and types.
13262 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13263
13264 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13265 (struct string_slot_hasher): New.
13266 (htab_t output_block::string_hash_table):
13267 Change type to hash_table. Update dependent calls and types.
13268
13269 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13270 (htab_t file_name_hash_table):
13271 Change type to hash_table. Update dependent calls and types.
13272
13273 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13274 calls and types.
13275
13276 * lto-streamer.c (htab_t tree_htab):
13277 Change type to hash_table. Update dependent calls and types.
13278
13279 * omp-low.c: Include gimplify-ctx.h.
13280
13281 * passes.c (htab_t name_to_pass_map):
13282 Change type to hash_table. Update dependent calls and types.
13283 (pass_traverse): Rename to passes_pass_traverse.
13284
13285 * plugin.c (htab_t event_tab):
13286 Change type to hash_table. Update dependent calls and types.
13287
13288 * postreload-gcse.c (htab_t expr_table):
13289 Change type to hash_table. Update dependent calls and types.
13290 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13291
13292 * sese.c (debug_rename_map_1): Make extern.
13293 (htab_t copy_bb_and_scalar_dependences::rename_map):
13294 Change type to hash_table. Update dependent calls and types.
13295
13296 * sese.h (extern debug_rename_map): Move to .c file.
13297
13298 * store-motion.c (htab_t store_motion_mems_table):
13299 Change type to hash_table. Update dependent calls and types.
13300
13301 * trans-mem.c (htab_t tm_new_mem_hash):
13302 Change type to hash_table. Update dependent calls and types.
13303
13304 * tree-browser.c (htab_t TB_up_ht):
13305 Change type to hash_table. Update dependent calls and types.
13306
13307 * tree-cfg.c (htab_t discriminator_per_locus):
13308 Change type to hash_table. Update dependent calls and types.
13309
13310 * tree-complex.c: Include tree-hasher.h
13311 (htab_t complex_variable_components):
13312 Change type to hash_table. Update dependent calls and types.
13313
13314 * tree-eh.c (htab_t finally_tree):
13315 Change type to hash_table. Update dependent calls and types.
13316
13317 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13318 struct int_tree_hasher.
13319 (extern int_tree_map_eq): Likewise.
13320 (uid_decl_map_hash): Removed.
13321 (extern decl_tree_map_eq): Likewise.
13322
13323 * tree-hasher.h: New.
13324 (struct int_tree_hasher): New.
13325 (typedef int_tree_htab_type): New.
13326
13327 * tree-inline.c: Include gimplify-ctx.h.
13328
13329 * tree-mudflap.c: Include gimplify-ctx.h.
13330
13331 * tree-parloops.c: Include tree-hasher.h.
13332 (htab_t eliminate_local_variables_stmt::decl_address):
13333 Change type to hash_table. Update dependent calls and types.
13334 (htab_t separate_decls_in_region::decl_copies): Likewise.
13335
13336 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13337 Change type to hash_table. Update dependent calls and types.
13338
13339 * tree-sra.c (candidates):
13340 Change type to hash_table. Update dependent calls and types.
13341
13342 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13343 in tree-flow.h.
13344 (int_tree_map_hash): Likewise.
13345
13346 * tree-ssa-dom.c (htab_t avail_exprs):
13347 Change type to hash_table. Update dependent calls and types.
13348
13349 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13350 Change type to hash_table. Update dependent calls and types.
13351
13352 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
13353 Change type to hash_table. Update dependent calls and types.
13354
13355 * tree-ssa-phiopt.c (seen_ssa_names):
13356 Change type to hash_table. Update dependent calls and types.
13357
13358 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
13359 Change type to hash_table. Update dependent calls and types.
13360
13361 * tree-ssa-uncprop.c (equiv):
13362 Change type to hash_table. Update dependent calls and types.
13363
c5a44004
JJ		 133642013-04-25 Jakub Jelinek <jakub@redhat.com>
13365
13366 PR rtl-optimization/57003
13367 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
13368 call note_stores with kill_clobbered_value callback again after
13369 killing regs_invalidated_by_call.
13370
09962a4a
JG		 133712013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13372
13373 * config/aarch64/aarch64-simd.md
13374 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
13375 (aarch64_simd_bsl<mode>): Likewise.
13376 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
13377
ea28bb0b
MP		 133782013-04-25 Marek Polacek <polacek@redhat.com>
13379
13380 PR tree-optimization/57066
3c21604f 13381 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 13382
96659611
JG		 133832013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
13384
13385 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
13386
9697e620
JG		 133872013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13388
13389 * config/aarch64/aarch64-builtins.c
13390 (aarch64_fold_builtin): New.
13391 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
13392 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
13393 * config/aarch64/aarch64-simd-builtins.def (abs): New.
13394 * config/aarch64/arm_neon.h
13395 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
13396
0ac198d3
JG		 133972013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13398 Tejas Belagod <tejas.belagod@arm.com>
13399
13400 * config/aarch64/aarch64-builtins.c
13401 (aarch64_gimple_fold_builtin): New.
13402 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
13403 * config/aarch64/aarch64-simd-builtins.def (addv): New.
13404 * config/aarch64/aarch64-simd.md (addpv4sf): New.
13405 (addvv4sf): Update.
13406 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
13407
58cff58c
N		 134082013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13409
df93505e 13410 * config/aarch64/aarch64.md
58cff58c
N		 13411 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
13412
7e0228bf
N		 134132013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13414
13415 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
13416 (*ngcsi_uxtw): New pattern.
13417
5819f96f 134182013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 13419 Julian Brown <julian@codesourcery.com>
5819f96f
KT		 13420
13421 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
13422 (TB_DREG): Add T_V4HF.
13423 (v4hf_UP): New macro.
13424 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 13425 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 13426 Handle initialisation of V4HF. Adjust initialisation of reinterpret
13427 built-ins.
df93505e 13428 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 13429 (arm_vector_mode_supported_p): Handle V4HF.
13430 (arm_mangle_map): Handle V4HFmode.
13431 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
13432 * config/arm/arm_neon_builtins.def: Add entries for
13433 vcvtv4hfv4sf, vcvtv4sfv4hf.
13434 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
13435 (neon_vcvtv4hfv4sf): Likewise.
13436 * config/arm/neon-gen.ml: Handle half-precision floating point
13437 features.
13438 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
13439 * config/arm/arm_neon.h: Regenerate.
13440 * config/arm/neon.ml (type elts): Add F16.
13441 (type vectype): Add T_float16x4, T_floatHF.
13442 (type vecmode): Add V4HF.
13443 (type features): Add Requires_FP_bit feature.
13444 (elt_width): Handle F16.
13445 (elt_class): Likewise.
13446 (elt_of_class_width): Likewise.
13447 (mode_of_elt): Refactor.
13448 (type_for_elt): Handle F16, fix error messages.
13449 (vectype_size): Handle T_float16x4.
13450 (vcvt_sh): New function.
13451 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
13452 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
13453 (string_of_mode): Handle V4HF.
13454 * doc/arm-neon-intrinsics.texi: Regenerate.
13455
1ef395e4
JG		 134562013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13457
13458 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
13459 format specifier in 'X' case.
13460
41c34e94
AM		 134612013-04-25 Alan Modra <amodra@gmail.com>
13462
13463 PR target/57052
13464 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
13465 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
13466 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
13467 Repeat for many other rotate/shift and mask patterns using subregs.
13468 Name lshiftrt insns.
13469 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
13470 on WORDS_BIG_ENDIAN.
13471
b9a7eb5d
AM		 134722013-04-25 Alan Modra <amodra@gmail.com>
13473
13474 * config.gcc: Support little-endian powerpc-linux targets.
13475 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
13476 (LINK_OS_LINUX_SPEC): Define.
13477 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
13478 Preserve MASK_LITTLE_ENDIAN.
13479 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
13480 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
13481 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
13482 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
13483 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
13484 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
13485 Correct fp word order for little-endian. Don't shift toc entries
13486 smaller than a word for little-endian.
13487 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
13488 (bswapdi2 splits): Correct low-part subreg for little-endian.
13489 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
13490 low/high where such is correct only for be.
13491 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
13492 little-endian for -mcall-aixdesc.
13493
87f73374
AM		 134942013-04-25 Alan Modra <amodra@gmail.com>
13495
13496 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
13497 replace_equiv_address_nv.
13498
cabf91cd
AM		 134992013-04-25 Alan Modra <amodra@gmail.com>
13500
13501 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
13502
0ae24cc8
VM		 135032013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13504
13505 Revert:
13506 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13507 * rtl.h (struct rtx_def): ...
cabf91cd 13508
77bce07c
VM		 135092013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13510
13511 PR rtl-optimizations/57046
13512 * lra-constraints (split_reg): Set up lra_risky_transformations_p
13513 for multi-reg splits.
13514
0db63e7f
L		 135152013-04-24 H.J. Lu <hongjiu.lu@intel.com>
13516
13517 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
13518
3362b6b6
SA		 135192013-04-24 Sterling Augustine <saugustine@google.com>
13520
13521 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
13522 (comp_dir_string, debug_str_dwo_section): New.
13523 (DEBUG_STR_DWO_SECTION): Rename to ...
13524 (DEBUG_DWO_STR_SECTION): ... this.
13525 (DEBUG_NORM_STR_SECTION): Delete.
13526 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
13527 (DEBUG_STR_DWO_SECTION_FLAGS): New.
13528 (find_AT_string): Move most logic to ...
13529 (find_AT_string_in_table): ... here. New.
13530 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
13531 add_skeleton_AT_string. Delete logic.
13532 (output_skeleton_debug_sections): Remove call to
13533 add_top_level_skeleton_die_attrs.
13534 (add_comp_dir_attribute): Move logic to comp_dir_string.
13535 (dwarf2out_init): Initialize debug_str_dwo_section.
13536 (output_indirect_string): Call find_string_form.
13537 (output_indirect_strings): Rewrite.
13538 (prune_unused_types): Empty skeleton_debug_str_hash.
13539 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
13540 (dwarf2out_finish): Call output_indirect_strings.
13541
e93e18e9
PC		 135422013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13543
13544 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
13545
f6ce35ac
VM		 135462013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13547
cabf91cd 13548 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM		 13549 (LRA_SUBREG_P): New macro.
13550 * recog.c (register_operand): Check LRA_SUBREG_P.
13551 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13552 stack frame.
13553 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13554 (lra_final_code_change): Skip subreg change for operators.
13555 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13556 if there are no operand changes.
13557 * lra-constraints.c (curr_insn_set): New.
13558 (match_reload): Set LRA_SUBREG_P.
13559 (emit_spill_move): Ditto.
13560 (check_and_process_move): Use curr_insn_set. Process only single
13561 set insns. Don't initialize sec_mem_p and change_p.
13562 (simplify_operand_subreg): Use LRA_SUBREG_P.
13563 (reg_in_class_p): New function.
13564 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13565 of #ifdef. Add code to remove cycling.
13566 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13567 non-null disp. Reload inner instead of disp when base and index
13568 are null. Try to put lo_sum into register.
13569 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 13570 (check_and_process_move): Move code for move cost check to
f6ce35ac 13571 simple_move_p. Remove equiv_substitution.
cabf91cd 13572 (simple_move_p): New function.
f6ce35ac
VM		 13573 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13574 curr_insn_set. Call check_and_process_move only for single set
13575 insns. Use the new function. Move call of check_and_process_move
13576 after operand equiv substitution and address process.
13577
38047d90
JG		 135782013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
13579
13580 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
13581 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
13582 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
13583
13f39b2e
PC		 135842013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13585
13586 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
13587
97c116dc
MP		 135882013-04-24 Marek Polacek <polacek@redhat.com>
13589
13590 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
13591 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
13592 (select_loops_exit_conditions): Likewise.
13593 (number_of_iterations_for_all_loops): Likewise.
13594 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
13595 (scev_analysis): Likewise.
13596
83082391 135972013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 13598 Chao-ying Fu <fu@mips.com>
83082391 13599
cabf91cd
AM		 13600 * config/mips/micromips.md (jraddiusp): New pattern.
13601 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
13602 instruction if possible.
83082391 13603
19e34aa2
AM		 136042013-04-24 Alan Modra <amodra@gmail.com>
13605
13606 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
13607
fdb6603c
JB		 136082013-04-24 Julian Brown <julian@codesourcery.com>
13609 Chung-Lin Tang <cltang@codesourcery.com>
13610
13611 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB		 13612 dependency behavior in enumeration type DIE generation. Add TODO note
13613 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 13614
bf190e8d
LC		 136152013-04-23 Lawrence Crowl <crowl@google.com>
13616
13617 * Makefile.in: Update as needed below.
13618
13619 * hash-table.h (class hash_table):
13620 Correct many methods with parameter types compare_type to the correct
13621 value_type. (Correct code was unlikely to notice the change.)
13622 (hash_table::elements_with_deleted) New.
13623 (class hashtable::iterator): New.
13624 (hashtable::begin()): New.
13625 (hashtable::end()): New.
13626 (FOR_EACH_HASH_TABLE_ELEMENT): New.
13627
13628 * statistics.c (statistics_hashes):
13629 Change type to hash_table. Update dependent calls and types.
13630
13631 * tree-into-ssa.c (var_infos):
13632 Change type to hash_table. Update dependent calls and types.
13633
13634 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
13635 Change type to hash_table. Update dependent calls and types.
13636
13637 * tree-ssa-loop-im.c (struct mem_ref.refs):
13638 Change type to hash_table. Update dependent calls and types.
13639
13640 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
13641 Change type to hash_table. Update dependent calls and types.
13642
13643 * tree-ssa-sccvn.c (vn_tables_s::nary):
13644 Change type to hash_table. Update dependent calls and types.
13645 (vn_tables_s::phis): Likewise.
13646 (vn_tables_s::references): Likewise.
13647
13648 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
13649 (vn_reference_eq): Update parameter and return types.
13650
13651 * tree-ssa-structalias.c (pointer_equiv_class_table):
13652 Change type to hash_table. Update dependent calls and types.
13653 (location_equiv_class_table): Likewise.
13654
13655 * tree-vect-data-refs.c: Consequential changes for making
13656 peeling a hash_table.
13657
13658 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
13659 (destroy_loop_vec_info): Dependent hash_table update.
13660
13661 * tree-vectorizer.h (peeling_htab):
13662 Change type to hash_table. Update dependent calls and types.
13663
d70a81dd
SC		 136642013-04-23 Shiva Chen <shiva0217@gmail.com>
13665
cabf91cd
AM		 13666 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
13667 to check the register content is equal or not.
13668 * lra-constraints.c (match_reload): Use lra_assign_reg_val
13669 to assign register content record.
13670 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 13671 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM		 13672 * lra-int.h (struct lra_reg): Add offset member.
13673 (lra_reg_val_equal_p): New static inline function.
13674 (lra_update_reg_val_offset): New static inline function.
13675 (lra_assign_reg_val): New static inline function.
13676 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
13677 to assign register content record.
13678 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 13679
b894a1f3
CM		 136802013-04-23 Catherine Moore <clm@codesourcery.com>
13681
13682 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
13683 operands. Record compression.
13684
ecd14de9
XDL		 136852013-04-23 Xinliang David Li <davidxl@google.com>
13686
13687 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
13688
92e776e9
RB		 136892013-04-23 Richard Biener <rguenther@suse.de>
13690
13691 PR middle-end/57036
13692 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
13693 parameter, only add abnormal goto edges from the copied body
13694 if the call could perform abnormal gotos.
13695 (copy_cfg_body): Adjust.
13696
a15ee567
SN		 136972013-04-23 Sofiane Naci <sofiane.naci@arm.com>
13698
13699 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
13700
08c52234
AS		 137012013-04-23 Andreas Schwab <schwab@linux-m68k.org>
13702
13703 * coretypes.h (gimple_stmt_iterator): Add struct to make
13704 compatible with C.
13705
999c1171
RB		 137062013-04-23 Richard Biener <rguenther@suse.de>
13707
13708 PR tree-optimization/57026
13709 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
13710 from SSA names occuring in abnormal PHI nodes.
13711
53669259
AK		 137122013-04-22 Andi Kleen <ak@linux.intel.com>
13713
13714 * lto/lto.c (print_lto_report_1): Fix LTO report names.
13715
1a0ad150
AK		 137162013-04-22 Andi Kleen <ak@linux.intel.com>
13717
13718 * lto/lto.c (print_lto_report_1): Declare early.
13719 (read_cgraph_and_symbols): Call print_lto_report_1 early.
13720
057f8f20
AK		 137212013-04-22 Andi Kleen <ak@linux.intel.com>
13722
13723 * common.opt (-flto-report-wpa): Add.
13724 * doc/invoke.texi (-flto-report-wpa): Add.
13725 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
13726 (lto_main): dito.
13727
473b1e05
XDL		 137282013-04-22 Xinliang David Li <davidxl@google.com>
13729
13730 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
13731 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
13732 * Makefile.in: New dependency
13733
cabf91cd 13734 David Daney <ddaney.cavm@gmail.com>
b1485a33 13735
cabf91cd
AM		 13736 * configure.ac (gcc_cv_as_micromips_support): Use the
13737 --fatal-warnings option.
13738 * configure: Regenerate.
b1485a33 13739
829d0168
MP		 137402013-04-22 Marek Polacek <polacek@redhat.com>
13741
13742 PR sanitizer/56990
13743 * tsan.c (instrument_expr): Don't instrument expression
13744 in case its size is zero.
13745
6d9b7208
UB		 137462013-04-22 Uros Bizjak <ubizjak@gmail.com>
13747
13748 PR target/57032
13749 Revert:
13750 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13751
13752 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13753
ea679d55
JG		 137542013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13755
13756 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
13757 (gimple_stmt_iterator): New typedef.
13758 * gimple.h (gimple_stmt_iterator): Rename to...
13759 (gimple_stmt_iterator_d): ... This.
13760 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
13761 trees be valid for GIMPLE and GENERIC.
13762 (TARGET_GIMPLE_FOLD_BUILTIN): New.
13763 * gimple-fold.c (gimple_fold_call): Call target hook
13764 gimple_fold_builtin.
13765 * hooks.c (hook_bool_gsiptr_false): New.
13766 * hooks.h (hook_bool_gsiptr_false): New.
13767 * target.def (fold_stmt): New.
13768 * doc/tm.texi: Regenerate.
13769
88a581da
VM		 137702013-04-22 Vladimir Makarov <vmakarov@redhat.com>
13771
13772 PR target/57018
13773 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
13774 a set sp if no stack realignment.
13775
92be22dc
NC		 137762013-04-22 Nick Clifton <nickc@redhat.com>
13777
13778 * config.gcc (tilegx-linux): Extend extra_objs rather than
13779 overwriting it.
13780 (tilepro-linux): Likewise.
13781
0ddec79f
JG		 137822013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13783
13784 * config/aarch64/aarch64-builtins.c
13785 (CF): Remove.
13786 (CF0, CF1, CF2, CF3, CF4, CF10): New.
13787 (VAR<1-12>): Add MAP parameter.
13788 (BUILTIN_*): Likewise.
13789 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
13790 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
13791 (aarch64_ushl_n<mode>): Likewise.
13792 (aarch64_sshr_n<mode>): Likewise.
13793 (aarch64_ushr_n<mode>): Likewise.
13794 (aarch64_<maxmin><mode>): Likewise.
13795 (aarch64_sqrt<mode>): Likewise.
13796 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
13797 (vshr<q>_n_*): Likewise.
13798
0050faf8
JG		 137992013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13800
13801 * config/aarch64/aarch64-builtins.c
13802 (aarch64_simd_builtin_type_mode): Handle SF types.
13803 (sf_UP): Define.
13804 (BUILTIN_GPF): Define.
13805 (aarch64_init_simd_builtins): Handle SF types.
13806 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
13807 (frecps): Likewise.
13808 (frecpx): Likewise.
13809 * config/aarch64/aarch64-simd.md
13810 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
13811 (aarch64_frecpe<mode>): New.
13812 (aarch64_frecps<mode>): Likewise.
13813 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
13814 (v8type): Add frecp<esx>.
13815 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
13816 (aarch64_frecps<mode>): Likewise.
13817 * config/aarch64/iterators.md (FRECP): New.
13818 (frecp_suffix): Likewise.
13819 * config/aarch64/arm_neon.h
13820 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
13821
0fad3dbc 138222013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB		 13823
13824 PR target/56995
13825 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
13826 (REG_CLASS_NAMES): Idem.
13827 (REG_CLASS_CONTENTS): Idem.
13828 (REGCLASS_HAS_FP_REG): Idem.
13829 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
13830 (sh_conditional_register_usage): Idem.
13831
3e8a33f9
JL		 138322013-04-21 Jeff Law <law@redhat.com>
13833
13834 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
13835 (ssa_forward_propagate_and_combine): Use it.
13836
f38e1b0a
VM		 138372013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13838
13839 * lra.c: Update the flow chart diagram.
13840
682303da
VM		 138412013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13842
13843 PR rtl-optimization/56847
13844 * lra-constraints.c (process_alt_operands): Discourage alternative
13845 with non-matche doffsettable memory constraint fro memory with
13846 known offset.
13847
f6b64c35
RB		 138482013-04-19 Richard Biener <rguenther@suse.de>
13849
13850 PR tree-optimization/56982
13851 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
13852 function.
13853 * gimplify.c (gimplify_call_expr): Notice special calls.
13854 (gimplify_modify_expr): Likewise.
13855 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
13856 abnormal control flow receivers.
13857 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
13858 in the same way as cfun->has_nonlocal_labels.
13859 (gimple_purge_dead_abnormal_call_edges): Likewise.
13860 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
13861 receivers start a basic-block.
13862
01d8bf07
RB		 138632013-04-19 Richard Biener <rguenther@suse.de>
13864
13865 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
13866 member ...
13867 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
13868 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
13869 (SLP_TREE_LOAD_PERMUTATION): Add.
13870 (vect_transform_slp_perm_load): Adjust prototype.
13871 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
13872 (vect_free_slp_instance): Likewise.
13873 (vect_create_new_slp_node): Likewise.
13874 (vect_supported_slp_permutation_p): Remove.
13875 (vect_slp_rearrange_stmts): Adjust.
13876 (vect_supported_load_permutation_p): Likewise. Inline
13877 vect_supported_slp_permutation_p here.
13878 (vect_analyze_slp_instance): Compute load permutations per
13879 slp node instead of per instance.
13880 (vect_get_slp_defs): Adjust.
13881 (vect_transform_slp_perm_load): Likewise.
13882 (vect_schedule_slp_instance): Remove redundant code.
13883 (vect_schedule_slp): Remove hack for PR56270, add it ...
13884 * tree-vect-stmts.c (vectorizable_load): ... here, do not
13885 CSE loads for SLP. Adjust.
13886
ede22fc3
GY		 138872013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13888
13889 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
13890 spelling in two comments.
13891
67bc84fb
GY		 138922013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13893
13894 PR target/56797
13895 * config/arm/arm.c (load_multiple_sequence): Require SP
13896 as base register for loads if SP is in the register list.
13897
e248d83f
MJ		 138982013-04-19 Martin Jambor <mjambor@suse.cz>
13899
13900 PR tree-optimization/56718
13901 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
13902 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
13903 and made public. Adjusted all callers.
13904 (ipa_intraprocedural_devirtualization): New function.
13905 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
13906 (ipa_intraprocedural_devirtualization): Likewise.
13907 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
13908
4891e8f8
RB		 139092013-04-19 Richard Biener <rguenther@suse.de>
13910
13911 PR tree-optimization/57000
13912 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
13913
dad89f7c
TG		 139142013-04-19 Terry Guo <terry.guo@arm.com>
13915
13916 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
13917 Replace with ...
13918 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
13919 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
13920 (cortex_m4_fmacs): Use new reservations.
13921 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
13922
72ea0d47
VM		 139232013-04-18 Vladimir Makarov <vmakarov@redhat.com>
13924
f1e6512c 13925 PR rtl-optimization/56999
72ea0d47
VM		 13926 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
13927 related code.
13928 (lra_coalesce): Remove split_origin_bitmap and related code.
13929 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
13930 ranges if necessary.
13931
780a5b71
UB		 139322013-04-18 Uros Bizjak <ubizjak@gmail.com>
13933
13934 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
13935 New array.
13936 (ix86_expand_call): Remove clobbered_registers array and use
13937 x86_64_ms_sysv_extra_clobbered_registers instead.
13938 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
13939 Declare here.
13940 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
13941 predicate.
13942 * config/i386/i386.md (*call_rex64_ms_sysv): Use
13943 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
13944 (*call_value_rex64_ms_sysv): Ditto.
13945
6f5a366a
CC		 139462013-04-18 Cary Coutant <ccoutant@google.com>
13947
13948 * dwarf2out.c (output_pubnames): Check die_perennial_p of
13949 parent instead of die_mark.
13950
475b8f37
DN		 139512013-04-18 Diego Novillo <dnovillo@google.com>
13952
13953 * gimple.c (create_gimple_tmp): New.
13954 (get_expr_type): New.
13955 (build_assign): New.
13956 (build_type_cast): New.
13957 * gimple.h (enum ssa_mode): Define.
13958 (gimple_seq_set_location): New.
13959 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 13960 to use build_assign and build_type_cast.
475b8f37 13961
08940f33
RB		 139622013-04-18 Richard Biener <rguenther@suse.de>
13963
13964 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
13965 handle negative step. Remove redundant checks.
13966 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
13967 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
13968 for negative step and grouped loads fail to vectorize.
13969
0e0f87d4
SB		 139702013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13971
13972 * emit-rtl.c (reset_insn_used_flags): New function.
13973 (reset_all_used_flags): Use it.
13974 (verify_insn_sharing): New function.
13975 (verify_rtl_sharing): Fix verification for SEQUENCEs.
13976
4c445590
JJ		 139772013-04-18 Jakub Jelinek <jakub@redhat.com>
13978
13979 PR tree-optimization/56984
13980 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
13981 and (x >> M) >= N don't register any assertion if N << M is the
13982 minimum value.
13983
6873ecab
SB		 139842013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13985
13986 * lower-subreg.c (resolve_simple_move): If called self-recursive,
13987 do not delete_insn insns that have not yet been emitted, only
13988 unlink them with remove_insn.
13989 * df-scan.c (df_insn_delete): Revert r197492.
13990
3ccb989e
SB		 139912013-04-17 Steven Bosscher <steven@gcc.gnu.org>
13992
13993 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
13994 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
13995
139962013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY		 13997
13998 * config/arm/arm.md (movsicc_insn): Convert define_insn into
13999 define_insn_and_split.
14000 (and_scc,ior_scc,negscc): Likewise.
14001 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14002
3ccb989e 140032013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY		 14004
14005 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14006 can benefit from using a sequence of LDRD instructions in epilogue
14007 instead of a single LDM instruction.
14008
6d10a203
MLI		 140092013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14010
14011 PR 45688
14012 * doc/extend.texi: Fix typo.
14013
6983e6b5
RB		 140142013-04-17 Richard Biener <rguenther@suse.de>
14015
14016 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14017 (vect_build_slp_tree): ... here.
14018 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14019 match. Remove special-casing of mismatched complex loads.
14020 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14021 re-try the match with swapped commutative operands.
14022 (vect_supported_load_permutation_p): Remove special-casing of
14023 mismatched complex loads.
14024 (vect_analyze_slp_instance): Adjust.
14025
ef23e6a2
RB		 140262013-04-17 Richard Biener <rguenther@suse.de>
14027
14028 PR rtl-optimization/56921
14029 * cfgloop.h (struct loop): Add simple_loop_desc member.
14030 (struct niter_desc): Mark with GTY(()).
14031 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14032 * loop-iv.c (get_simple_loop_desc): Likewise.
14033 (free_simple_loop_desc): Likewise.
14034
14035 Revert
14036 2013-04-16 Richard Biener <rguenther@suse.de>
14037
14038 PR rtl-optimization/56921
14039 * loop-init.c (pass_rtl_move_loop_invariants): Add
14040 TODO_do_not_ggc_collect to todo_flags_finish.
14041 (pass_rtl_unswitch): Same.
14042 (pass_rtl_unroll_and_peel_loops): Same.
14043 (pass_rtl_doloop): Same.
14044
fc6f94f5
EB		 140452013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14046
14047 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14048 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14049 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14050 references.
14051 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14052 * tree-streamer.c (record_common_node): Adjust reference in comment.
14053
10a88311
TG		 140542013-04-17 Terry Guo <terry.guo@arm.com>
14055
14056 * config/arm/cortex-m4.md: Add a new bypass.
14057
6d9b7208 140582013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N		 14059
14060 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14061 New pattern.
14062 (*subs_<optab><mode>_multp2): New pattern.
14063 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14064 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14065
6d9b7208 140662013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N		 14067
14068 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14069 (*subs_mul_imm_<mode>): New pattern.
14070
18a6701e
DE		 140712013-04-16 David Edelsohn <dje.gcc@gmail.com>
14072
14073 PR target/56948
14074 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14075 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14076 (vsx_movti_32bit): Change j->wa to O->wa.
14077
07c37b2f
RB		 140782013-04-16 Richard Biener <rguenther@suse.de>
14079
14080 PR rtl-optimization/56921
14081 * loop-init.c (pass_rtl_move_loop_invariants): Add
14082 TODO_do_not_ggc_collect to todo_flags_finish.
14083 (pass_rtl_unswitch): Same.
14084 (pass_rtl_unroll_and_peel_loops): Same.
14085 (pass_rtl_doloop): Same.
14086
0e0f87d4 140872013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY		 14088
14089 * config/arm/arm.c (emit_multi_reg_push): New declaration
14090 for an existing function.
14091 (arm_emit_strd_push): New function.
14092 (arm_expand_prologue): Used here.
14093 (arm_emit_ldrd_pop): New function.
14094 (arm_expand_epilogue): Used here.
14095 (arm_get_frame_offsets): Update condition.
14096 (arm_emit_multi_reg_pop): Add a special case for load of a single
14097 register with writeback.
14098
5e8e2af4
UB		 140992013-04-16 Uros Bizjak <ubizjak@gmail.com>
14100
14101 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
14102 description.
14103
9fd9ccf7
RB		 141042013-04-16 Richard Biener <rguenther@suse.de>
14105
14106 PR tree-optimization/56756
14107 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
14108 (first_mem_ref_loc): New.
14109 (execute_sm): Place the load temporarily before a previous
14110 access instead of in the latch edge to ensure its SSA dependencies
14111 are defined at points dominating the load.
14112
96fba521
SB		 141132013-04-16 Steven Bosscher <steven@gcc.gnu.org>
14114
4c8af858
SB		 14115 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
14116 correct fix by moving header and footer insn to the footer of
14117 the merged basic block. Clear BB_END of the merged-away block.
14118
96fba521
SB		 14119 PR middle-end/43631
14120 * emit-rtl.c (make_note_raw): New function.
14121 (link_insn_into_chain): New static inline function.
14122 (add_insn): Use it.
14123 (add_insn_before, add_insn_after): Factor insn chain linking code...
14124 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
14125 using link_insn_into_chain.
14126 (note_outside_basic_block_p): New helper function for emit_note_after
14127 and emit_note_before.
14128 (emit_note_after): Use nobb variant of add_insn_after if the note
14129 should not be contained in a basic block.
14130 (emit_note_before): Use nobb variant of add_insn_before if the note
14131 should not be contained in a basic block.
14132 (emit_note_copy): Use make_note_raw.
14133 (emit_note): Likewise.
14134 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
14135 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
14136 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
14137 the moved barrier the tail of the basic block it follows.
14138 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
14139
7b8265ba
JJ		 141402013-04-15 Jakub Jelinek <jakub@redhat.com>
14141
14142 PR tree-optimization/56962
14143 * gimple-ssa-strength-reduction.c (record_increment): Only set
14144 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
14145 either rhs1 or rhs2 is equal to c->base_expr.
14146
5185d248
RB		 141472013-04-15 Richard Biener <rguenther@suse.de>
14148
14149 PR tree-optimization/56933
14150 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
14151 member.
14152 (GROUP_READ_WRITE_DEPENDENCE): Remove.
14153 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
14154 * tree-vect-data-refs.c (vect_analyze_group_access): Move
14155 dependence check ...
14156 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14157 ... here.
14158 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14159 GROUP_READ_WRITE_DEPENDENCE.
14160
a24243a0
AK		 141612013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14162
14163 * emit-rtl.c (reset_all_used_flags): New function.
14164 (verify_rtl_sharing): Call reset_all_used_flags before and after
14165 performing the checks.
14166
1c50eada
KT		 141672013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14168
14169 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14170 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14171 * config/arm/constraints.md (De): New constraint.
14172 * config/arm/neon.md (anddi3_neon): Delete.
14173 (neon_vand<mode>): Expand to standard anddi3 pattern.
14174 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14175 Move earlier in the file.
14176 (neon_inv_logic_op2): Likewise.
14177 (arm_anddi_operand_neon): New predicate.
14178
e927b6ad
RO		 141792013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14180
14181 * configure.ac (gcc_cv_ld_as_needed): Set
14182 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14183 Use -z ignore, -z record on *-*-solaris2*.
14184 (HAVE_LD_AS_NEEDED): Update comment.
14185 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14186 * configure: Regenerate.
14187 * config.in: Regenerate.
14188 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14189 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14190 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14191 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14192 equivalents. Fix markup.
14193 * doc/tm.texi: Regenerate.
14194
e0ea8797
AH		 141952013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14196
14197 * config/i386/i386.opt: New option mstack-protector-guard=.
14198 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14199 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14200 TARGET_SSP_TLS_GUARD.
14201 * config/i386/i386.c (ix86_option_override_internal): Set
14202 ix86_stack_protector_guard.
14203 * config/i386/i386.md (stack_protect_set): Enable for
14204 TARGET_SSP_TLS_GUARD only.
14205 (stack_protect_set_<mode>): Ditto.
14206 (stack_protect_test): Ditto.
14207 (stack_protect_test_<mode>): Ditto.
14208 * doc/invoke.texi (i386 Option): Document.
14209
811b72f9
EB		 142102013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14211
14212 PR target/56890
14213 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14214 (S_MODES): Set H_MODE bit.
14215 (SF_MODES): Set only S_MODE and SF_MODE bits.
14216 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14217 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14218 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14219 <MODE_FLOAT>: Likewise.
14220
5529fdd6
JY		 142212013-04-15 Joey Ye <joey.ye@arm.com>
14222
14223 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14224
517b1da2
JY		 142252013-04-15 Joey Ye <joey.ye@arm.com>
14226
14227 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14228 for real far jump.
14229 (thumb_far_jump_used_p): Count instruction size and set
14230 far_jump_used.
14231
01007ae0
EB		 142322013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14233
14234 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14235 * resource.c (mark_target_live_regs): Reformat conditional block.
14236
c46f6580
SB		 142372013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14238
14239 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14240 notes, they are emitted only just before final.
14241 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14242
1f397f45
SB		 142432013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14244
14245 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14246 * cfgrtl.c (delete_insn): Call it here instead.
14247 * lra-spills.c (lra_final_code_change): Use delete_insn.
14248 * haifa-sched.c (sched_remove_insn): Likewise.
14249 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14250 returning to the nop pool.
14251 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14252 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14253
58a51369
SB		 142542013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14255
14256 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14257 * doc/tm.texi: Regenerated.
14258
33159866
UB		 142592013-04-12 Uros Bizjak <ubizjak@gmail.com>
14260
14261 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14262 QImode checks.
14263
226e378f
SB		 142642013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14265
14266 * df-core.c (df_find_def): Compare register numbers.
14267 (df_find_use): Likewise.
14268
fafb9b18
VM		 142692013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14270
14271 PR target/56903
14272 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14273 lra_in_progress for return.
14274
9a946fd6
GY		 142752013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14276
14277 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14278 define_insn into define_insn_and_split and emit movsicc patterns.
14279
41b83758
GY		 142802013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14281
14282 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14283
d3afd9aa
RB		 142842013-04-12 Richard Biener <rguenther@suse.de>
14285
14286 * tree-pass.h (TODO_do_not_ggc_collect): New.
14287 * passes.c (execute_one_ipa_transform_pass): Honor
14288 TODO_do_not_ggc_collect.
14289 (execute_one_pass): Likewise.
14290
14291 Revert
14292 2013-04-10 Richard Biener <rguenther@suse.de>
14293
14294 * passes.c (init_optimization_passes): Remove reload pass.
14295 * ira.c (do_reload): Merge into ...
14296 (ira): ... this.
14297 (rest_of_handle_reload): Remove.
14298 (pass_reload): Likewise.
14299 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14300 of reload for vzeroupper pass placement.
14301
06f9b387
JJ		 143022013-04-12 Jakub Jelinek <jakub@redhat.com>
14303
14304 PR tree-optimization/56918
14305 PR tree-optimization/56920
14306 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14307 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14308 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14309 use wide_mul_with_sign method.
14310
953094d2
RB		 143112013-04-12 Richard Biener <rguenther@suse.de>
14312
14313 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14314 not be considered a gimple constant.
14315
bb506982
MG		 143162013-04-12 Marc Glisse <marc.glisse@inria.fr>
14317
14318 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14319 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14320 operations.
14321
4b84d650
JJ		 143222013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14323 Jakub Jelinek <jakub@redhat.com>
14324
14325 * opts.c: Include diagnostic-color.h.
14326 (common_handle_option): Handle OPT_fdiagnostics_color_.
14327 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14328 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14329 (diagnostic-color.o): New.
14330 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14331 (diagnostic_color_rule): New enum.
14332 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14333 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14334 the location string.
14335 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14336 either NULL, or color kind.
14337 * diagnostic-color.c: New file.
14338 * diagnostic-color.h: New file.
14339 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14340 arguments.
14341 * doc/invoke.texi (-fdiagnostics-color): Document.
14342 * pretty-print.h (pp_show_color): Define.
14343 (struct pretty_print_info): Add show_color field.
14344 * diagnostic.c: Include diagnostic-color.h.
14345 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14346 macros. Colorize error:, warning: etc. strings and also the location
14347 string.
14348 (diagnostic_show_locus): Colorize the caret line.
14349 * pretty-print.c: Include diagnostic-color.h.
14350 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14351 inside of %< %> quotes or quoted through q format modifier.
14352
067a1e71
AK		 143532013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14354
33159866 14355 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 14356
33159866 143572013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N		 14358
14359 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
14360 code in CC_NZ mode.
14361 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
14362 pattern.
14363
7b55f98f
MP		 143642013-04-11 Marek Polacek <polacek@redhat.com>
14365
14366 PR tree-optimization/48184
33159866 14367 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 14368
966b587e
EB		 143692013-04-11 Eric Botcazou <ebotcazou@adacore.com>
14370
14371 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
14372 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
14373 (skip_simple_arithmetic): Tidy up.
14374 * tree.h (skip_simple_constant_arithmetic): Declare.
14375
33159866 143762013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N		 14377
14378 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
14379
1d42c1ec
RB		 143802013-04-11 Richard Biener <rguenther@suse.de>
14381
14382 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14383 generate vector constants.
14384
4ba5ea11
RB		 143852013-04-11 Richard Biener <rguenther@suse.de>
14386
14387 PR tree-optimization/56878
14388 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
14389 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
14390 New function.
14391 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
14392 Prefer to align the DR with the most invariant base address.
14393
f0defe58
SKS		 143942013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14395
14396 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
14397 comment.
14398
d07458be
JG		 143992013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
14400
14401 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
14402 floating-point vector comparisons against 0.
14403
146b8692
JJ		 144042013-04-11 Jakub Jelinek <jakub@redhat.com>
14405
14406 PR tree-optimization/56899
14407 * fold-const.c (extract_muldiv_1): Apply distributive law
14408 only if TYPE_OVERFLOW_WRAPS (ctype).
14409
b8578ff7
BC		 144102013-04-11 Bin Cheng <bin.cheng@arm.com>
14411
14412 PR target/56124
14413 * ira-costs.c (scan_one_insn): Check whether the source rtx of
14414 loading has side effect.
14415
0ea8a6f9
SB		 144162013-04-10 Steven Bosscher <steven@gcc.gnu.org>
14417
14418 * config/sparc/sparc.c: Include tree-pass.h.
14419 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
14420 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
14421 head of file. Change return type. Split off gate function.
14422 (sparc_gate_work_around_errata): New function.
14423 (pass_work_around_errata): New pass definition.
14424 (insert_pass_work_around_errata) New pass insert definition to
14425 insert pass_work_around_errata just after delayed-branch scheduling.
14426 (sparc_option_override): Insert the pass.
14427 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
14428
42e37616
DM		 144292013-04-10 David S. Miller <davem@davemloft.net>
14430
89deeb3b
DM		 14431 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
14432 or -mcpu=hypersparc.
14433
42e37616
DM		 14434 * target.def (cstore_mode): New hook.
14435 * target.h: Include insn-codes.h
14436 * targhooks.c: Likewise.
14437 (default_cstore_mode): New function.
14438 * targhooks.h: Declare it.
14439 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
14440 * doc/tm.texi: Rebuild.
14441 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
14442 target hook, rather than inspecting the insn_data.
14443 * config/sparc/sparc.c (sparc_cstore_mode): New function.
14444 (TARGET_CSTORE_MODE): Redefine.
14445 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
14446 result patterns.
14447 * config/sparc/predicates.md (cstore_result_operand): New special
14448 predicate.
14449 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
14450 Use it for operand 0.
14451 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
14452 (*snesi_special): Likewise.
14453 (*snesi_zero): Likewise.
14454 (*seqsi_zero): Likewise.
14455 (*sltu_insn): Likewise.
14456 (*sgeu_insn): Likewise.
14457 (*seqdi_special): Make operand 0 and comparison operation be of
14458 DImode.
14459 (*snedi_special): Likewise.
14460 (*snedi_special_vis3): Likewise.
14461 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
14462 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
14463 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
14464 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
14465 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
14466 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
14467 (*sltu_extend_sp64): Likewise.
14468 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
14469 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
14470 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
14471 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
14472 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
14473
95ca411e
YZ		 144742013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
14475
14476 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
14477 (aarch64_start_file): Use the new function.
14478
6782438d 144792013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 14480 Jason Merrill <jason@redhat.com>
6782438d
SKS		 14481
14482 * common.opt: Add -gdwarf.
14483 * opts.c (common_handle_option): Handle it.
14484 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
14485
bb313b93
RB		 144862013-04-10 Richard Biener <rguenther@suse.de>
14487
14488 * passes.c (execute_todo): Do not call ggc_collect conditional here.
14489 (execute_one_ipa_transform_pass): But unconditionally here.
14490 (execute_one_pass): And here.
14491 (init_optimization_passes): Remove reload pass.
14492 * tree-pass.h (TODO_ggc_collect): Remove.
14493 (pass_reload): Likewise.
14494 * ira.c (do_reload): Merge into ...
14495 (ira): ... this.
14496 (rest_of_handle_reload): Remove.
14497 (pass_reload): Likewise.
14498 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14499 of reload for vzeroupper pass placement.
14500 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
14501 and todo_flags_finish of all passes.
14502
793d9a16
RB		 145032013-04-10 Richard Biener <rguenther@suse.de>
14504
14505 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
14506 first_const_oprnd field, rename first_def_type to first_op_type.
14507 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
14508 (vect_get_and_check_slp_defs): Always use the type of the
14509 operand. Allow mixed vect_external_def, vect_constant_def types.
14510 (vect_get_constant_vectors): Handle mixed vect_external_def,
14511 vect_constant_def types.
14512
12211b99 145132013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR		 14514
14515 PR tree-optimization/55524
14516 * tree-ssa-math-opts.c
14517 (convert_mult_to_fma): Don't use an fms construct
14518 when we don't have an fms operation, but fnma, and it looks
14519 likely that we'll be able to use the latter.
14520
12211b99 145212013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ		 14522
14523 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
14524 function.
14525 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
14526 inline fail caused by overwritable functions.
14527
34ab4a5b
CJW		 145282013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
14529
14530 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
14531 unnecessary bits in the constant power of two case.
14532
abf9bfbc
RB		 145332013-04-10 Richard Biener <rguenther@suse.de>
14534
14535 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
14536 broken code swapping operands.
14537 (vect_build_slp_tree): Do not compute load permutations here.
14538 (vect_analyze_slp_instance): Compute load permutations here,
14539 after building the SLP tree.
14540
f408477e
CB		 145412013-04-09 Christian Bruel <christian.bruel@st.com>
14542
14543 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
14544 of next/prev_real_insn.
14545
5ac42672
JH		 145462013-04-09 Jan Hubicka <jh@suse.cz>
14547
abf9bfbc
RB		 14548 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14549 Drop aliased parameter.
5ac42672
JH		 14550 (function_and_variable_visibility): Do not handle alias pairs.
14551 * cgraph.c (varpool_externally_visible_p): Update prototype.
14552 * varpool.c (varpool_add_new_variable): Update.
14553
5017f1d2
KT		 145542013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14555
14556 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
14557
48eecbee
SB		 145582013-04-09 Steven Bosscher <steven@gcc.gnu.org>
14559
a949cf1c
SB		 14560 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
14561
48eecbee
SB		 14562 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
14563
75ef8e3d
MP		 145642013-04-09 Marek Polacek <polacek@redhat.com>
14565
14566 PR tree-optimization/48762
33159866 14567 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 14568
23847df4
RB		 145692013-04-09 Richard Biener <rguenther@suse.de>
14570
14571 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
14572 dealing with cost.
14573 (vect_build_slp_tree): Likewise.
14574 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
14575 calculating the cost of a SLP instance.
14576 (vect_analyze_slp_instance): Use it from here, after building
14577 the SLP tree.
14578
ea3a0fde
JJ		 145792013-04-09 Jakub Jelinek <jakub@redhat.com>
14580
14581 PR middle-end/56883
14582 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
14583 expand_omp_for_static_chunk): Use simple_p = true in
14584 force_gimple_operand_gsi calls when assigning to addressable decls.
14585
a32dfe9d
JL		 145862013-04-09 Jeff Law <law@redhat.com>
14587
14588 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
14589 when the boolean was created by converting a wider object which
14590 had a boolean range.
14591
d755c7ef
RB		 145922013-04-09 Richard Biener <rguenther@suse.de>
14593
14594 * tree-vectorizer.h (slp_void_p): Remove.
14595 (slp_tree): Typedef before _slp_tree declaration.
14596 (struct _slp_tree): Use a vector of slp_tree as children.
14597 (vect_get_place_in_interleaving_chain): Remove.
14598 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
14599 Move ...
14600 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
14601 and make static.
14602 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
14603 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
14604 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
14605 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
14606 Use slp_node instead of slp_void_p and adjust.
14607
3d741091
RB		 146082013-04-09 Richard Biener <rguenther@suse.de>
14609
14610 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
14611 work that is not necessary.
14612
39307ba7
JJ		 146132013-04-09 Jakub Jelinek <jakub@redhat.com>
14614
14615 PR tree-optimization/56854
14616 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
14617 forward into clobber stmts if it would change MEM_REF lhs into
14618 non-MEM_REF.
14619
343881fd
MK		 146202013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
14621
14622 * tree.c (type_hash_lookup, type_hash_add): Make static.
14623 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
14624
3922658a
RB		 146252013-04-09 Richard Biener <rguenther@suse.de>
14626
14627 * tree.h (unsave_expr_now): Remove.
14628 * tree-inline.c (mark_local_for_remap_r): Remove.
14629 (unsave_expr_1): Likewise.
14630 (unsave_r): Likewise.
14631 (unsave_expr_now): Likewise.
14632 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
14633 (propagate_tree_value): Likewise.
14634
9fb6b620
SB		 146352013-04-08 Steven Bosscher <steven@gcc.gnu.org>
14636
14637 * doc/rtl.texi (sequence): Rewrite documentation to match the
14638 current use of SEQUENCE rtl objects.
14639 * rtl.def (SEQUENCE): Likewise.
14640
14641 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
14642 Update documentation.
14643 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
14644 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
14645
14646 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
14647
8ddb5a29
TJ		 146482013-04-08 Teresa Johnson <tejohnson@google.com>
14649
14650 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
14651 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 14652 methods.
8ddb5a29
TJ		 14653 (estimate_edge_size_and_time): Add comment to suggest using rounding
14654 methods.
14655 (estimate_node_size_and_time): Ditto.
14656 (remap_edge_change_prob): Use helper rounding divide methods.
14657 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
14658 (gimple_mod_pow2_value_transform): Ditto.
14659 (gimple_mod_subtract_transform): Ditto.
14660 (gimple_ic_transform): Ditto.
14661 (gimple_stringops_transform): Ditto.
14662 * stmt.c (conditional_probability): Ditto.
14663 (emit_case_dispatch_table): Ditto.
14664 * lto-cgraph.c (merge_profile_summaries): Ditto.
14665 * tree-optimize.c (execute_fixup_cfg): Ditto.
14666 * cfgcleanup.c (try_forward_edges): Ditto.
14667 * cfgloopmanip.c (scale_loop_profile): Ditto.
14668 (loopify): Ditto.
14669 (duplicate_loop_to_header_edge): Ditto.
14670 (lv_adjust_loop_entry_edge): Ditto.
14671 * tree-vect-loop.c (vect_transform_loop): Ditto.
14672 * profile.c (compute_branch_probabilities): Ditto.
14673 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
14674 * lto-streamer-in.c (input_cfg): Ditto.
14675 * gimple-streamer-in.c (input_bb): Ditto.
14676 * ipa-cp.c (update_profiling_info): Ditto.
14677 (update_specialized_profile): Ditto.
14678 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
14679 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 14680 rounding methods.
8ddb5a29
TJ		 14681 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14682 (compute_trg_info): Ditto.
14683 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
14684 (purge_dead_edges): Ditto.
14685 * loop-unswitch.c (unswitch_loop): Ditto.
14686 * cgraphclones.c (cgraph_clone_edge): Ditto.
14687 (cgraph_clone_node): Ditto.
14688 * tree-inline.c (copy_bb): Ditto.
14689 (copy_edges_for_bb): Ditto.
14690 (initialize_cfun): Ditto.
14691 (copy_cfg_body): Ditto.
14692 (expand_call_inline): Ditto.
14693
661e6bd7
KT		 146942013-04-08 Kai Tietz <ktietz@redhat.com>
14695
14696 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
14697 TARGET_CYGWIN64 by TARGET_64BIT.
14698
105766f3
JR		 146992013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14700
14701 * config/epiphany/epiphany.md (GPR_1): New constant.
14702 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
14703 * config/epiphany/epiphany.c (gen_compare_reg):
14704 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
14705 is already in place.
14706 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
14707 Don't require being called during rtl expansion; If y operlaps r0,
14708 return 0.
14709 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
14710 (epiphany_expand_epilogue): Likewise.
14711
8afab237
JR		 14712 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
14713 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 14714 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 14715
fa7707d6
JR		 14716 * config/epiphany/constraints.md (CnL): New constraint.
14717 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
14718 * config/epiphany/predicates.md (add_operand): Allow 1024.
14719
5138e18d
JR		 14720 * config/epiphany/epiphany.md (logical_op): New code iterator.
14721 (op_mnc): New code attribute.
14722 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
14723 (mov_f+1, mov_f+2): New peephole2 patterns.
14724
2ccc703d
JR		 14725 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
14726 (cstoresi4): Also allow re-use of zero result when doing a NE
14727 comparison to a non-zero operand.
aefb0819 14728 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 14729
093ac0a5
JR		 14730 * config/epiphany/epiphany.md (<insn_opname>v2si3):
14731 Use gen_addsi3_i / gen_subsi3_i.
14732
f223bb13
JJ		 147332013-04-08 Jakub Jelinek <jakub@redhat.com>
14734
14735 PR c++/34949
14736 PR c++/50243
14737 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
14738 contain anything but clobbers, at most one __builtin_stack_restore,
14739 optionally debug stmts and final resx, and if it has at least one
14740 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
14741 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
14742 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
14743 which isn't defaut definition, remove them.
14744 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
14745 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
14746 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
14747 with MEM_REF LHS with SSA_NAME address.
14748
4481581f
JL		 147492013-04-08 Jeff Law <law@redhat.com>
14750
14751 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 14752
451dabda
RB		 147532013-04-08 Richard Biener <rguenther@suse.de>
14754
14755 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
14756 extra newline.
14757 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
14758 determined vector type.
14759 (vect_analyze_data_refs): Likewise.
14760 (vect_get_new_vect_var): Adjust.
14761 (vect_create_destination_var): Preserve SSA name versions.
14762 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
14763 not dump anything here.
14764
3b088b47
JR		 147652013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14766
14767 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
14768 Add member lr_slot_known.
14769 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
14770 if necessary.
14771 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14772 Remove code that sets lr_slot_offset according to what a previous
14773 version of epiphany_emit_save_restore used to do.
14774 (epiphany_emit_save_restore): When doing an lr save or restore,
14775 set/verify lr_slot_known and lr_slot_offset.
14776
d8484d41
XQ		 147772013-04-08 Xinyu Qi <xyqi@marvell.com>
14778
33159866
UB		 14779 PR target/54338
14780 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ		 14781 in ALL_REGS.
14782
481be1c4
RB		 147832013-04-08 Richard Biener <rguenther@suse.de>
14784
14785 * alias.c (find_base_term): Fix thinko in previous change.
14786
401f3a81
JJ		 147872013-04-08 Jakub Jelinek <jakub@redhat.com>
14788
14789 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
14790 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
14791 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
14792 if possible to compute val.
14793 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
14794 For QImode integers don't require anything about precision. Use
14795 const_with_all_bytes_same to find out if the constant doesn't have
14796 repeated bytes in it.
14797
03ed99a8
AK		 147982013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14799
14800 * config/s390/s390.c (s390_expand_insv): Only accept insertions
14801 within mode size.
14802
781b2e62
MP		 148032013-04-08 Marek Polacek <polacek@redhat.com>
14804
14805 PR rtl-optimization/48182
14806 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
14807 value to 1.
14808
27e430a2
JDA		 148092013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14810
14811 PR target/55487
14812 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
14813 nuses, make sure we have a label.
14814
4902aa64
BS		 148152013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14816
14817 PR target/56843
14818 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
14819 (rs6000_emit_swdiv_low_precision): Remove.
14820 (rs6000_emit_swdiv): Rewrite to handle between one and four
14821 iterations of Newton-Raphson generally; modify required number of
14822 iterations for some cases.
14823 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
14824
7bca81dc
SB		 148252013-04-05 Steven Bosscher <steven@gcc.gnu.org>
14826
14827 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
14828 set-but-unused variable.
14829
14830 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
14831 basic blocks of released function bodies garbage-collectable.
14832
14833 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
14834 (struct rtl_opt_pass): Add TODO_df_finish.
14835
14836 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
14837
4542a38a
GY		 148382013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14839
14840 * config/arm/constraints.md (q): New constraint.
14841 * config/arm/ldrdstrd.md: New file.
14842 * config/arm/arm.md (ldrdstrd.md) New include.
14843 (arm_movdi): Use "q" instead of "r" constraint
14844 for double-word memory access.
14845 (movdf_soft_insn): Likewise.
14846 * config/arm/vfp.md (movdi_vfp): Likewise.
14847 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 14848 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY		 14849 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
14850 (mem_ok_for_ldrd_strd): Likewise.
14851 (output_move_double): Update assertion.
14852
2385b218
GY		 148532013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14854
14855 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
14856
75fe1cb5
GY		 148572013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14858
14859 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
14860 define_insn_and_split.
14861 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
14862
dd660e8e
GY		 148632013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14864
14865 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
14866 define_insn_and_split.
33159866 14867 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY		 14868 (shiftsi3_compare): New pattern.
14869 (rrx): New pattern.
14870 * config/arm/unspecs.md (UNSPEC_RRX): New.
14871
045e472c
GY		 148722013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14873
14874 * config/arm/arm.md (negdi_extendsidi): New pattern.
14875 (negdi_zero_extendsidi): Likewise.
14876
3f3bf1a8
GY		 148772013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14878
14879 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
14880 define_insn_and_split.
14881 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
14882 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
14883
b0b49556
GY		 148842013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14885
14886 * config/arm/arm.md (arm_subdi3): Convert define_insn into
14887 define_insn_and_split.
14888 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
14889 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
14890
d633dd84
GY		 148912013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14892
14893 * config/arm/arm.md (subsi3_carryin): New pattern.
14894 (subsi3_carryin_const): Likewise.
14895 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
14896 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
14897
f4499066
GY		 148982013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14899
14900 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
14901
ceef6fd9
GY		 149022013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14903
14904 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 14905 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 14906
ddbdd8a7
KT		 149072013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14908
14909 * config/arm/arm.c (arm_expand_builtin): Change fcode
14910 type to unsigned int.
14911
8456d78a
RR		 149122013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14913
14914 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
14915
526c230f
ILT		 149162013-04-04 Ian Lance Taylor <iant@google.com>
14917
33159866
UB		 14918 * doc/standards.texi (Standards): The Go frontend supports the Go 1
14919 language standard.
526c230f 14920
3cfbe04d
SB		 149212013-04-04 Steven Bosscher <steven@gcc.gnu.org>
14922
14923 PR middle-end/56729
14924 * df-scan.c (df_insn_delete): Disable failing assert.
14925
dfa3f8d0
KT		 149262013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14927
14928 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
14929 New function prototype.
14930 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
14931 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
14932 (arm_builtin_vectorized_function): New function.
14933
f7837758
KT		 149342013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14935
14936 * config/arm/arm_neon_builtins.def: New file.
14937 * config/arm/arm.c (neon_builtin_data): Move contents to
14938 arm_neon_builtins.def.
14939 (enum arm_builtins): Include neon builtin definitions.
14940 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 14941 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 14942
39c1b6db
MP		 149432013-04-04 Marek Polacek <polacek@redhat.com>
14944
14945 PR tree-optimization/48186
14946 * predict.c (maybe_hot_frequency_p): Return false if
14947 HOT_BB_FREQUENCY_FRACTION is 0.
14948 (cgraph_maybe_hot_edge_p): Likewise.
14949
314f64eb
RB		 149502013-04-04 Richard Biener <rguenther@suse.de>
14951
14952 PR tree-optimization/56826
14953 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
14954 more accurately.
14955
51a905b2
RB		 149562013-04-04 Richard Biener <rguenther@suse.de>
14957
14958 PR tree-optimization/56213
14959 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 14960 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 14961
f1bf4f3a
RB		 149622013-04-04 Richard Biener <rguenther@suse.de>
14963
14964 PR tree-optimization/56837
14965 * tree-loop-distribution.c (classify_partition): For non-zero
14966 values require that the value has the same precision as its
14967 mode to be useful as memset value.
14968
0bca7ded
NC		 149692013-04-03 Nick Clifton <nickc@redhat.com>
14970
33159866 14971 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC		 14972 (fmssf4): Use fmsf.s on E3V5 architectures.
14973 (fnmasf4): Use fnmaf.s on E3V5 architectures.
14974 (fnmssf4): Use fnmsf.s on E3V5 architectures.
14975
b4019227
JL		 149762013-04-03 Jeff Law <law@redhat.com>
14977
14978 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
14979 (lra-eliminations.o): Likewise.
14980
f57ddb5b
TJ		 149812013-04-03 Teresa Johnson <tejohnson@google.com>
14982
14983 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 14984 compute_working_sets here from profile.c.
f57ddb5b 14985 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB		 14986 (gcov_working_set_t): Moved typedef here from basic-block.h
14987 (compute_working_set): Declare.
f57ddb5b
TJ		 14988 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
14989 (get_working_sets): Renamed from compute_working_set,
33159866 14990 replace most of body with call to new compute_working_sets.
f57ddb5b 14991 (get_exec_counts): Replace call to compute_working_sets
33159866
UB		 14992 to get_working_sets.
14993 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 14994 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 14995 to get_working_sets.
f57ddb5b
TJ		 14996 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
14997 * gcov-dump.c (dump_working_sets): New function.
14998
12211b99 149992013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ		 15000
15001 * hwint.c (sext_hwi, zext_hwi): New functions.
15002 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15003 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15004 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15005 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15006 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15007 (sext_hwi, zext_hwi): New functions.
0bca7ded 15008
be672e08
JL		 150092013-04-03 Jeff Law <law@redhat.com>
15010
15011 PR tree-optimization/56799
33159866
UB		 15012 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15013 back test for widening conversion erroneously dropped in prior change.
be672e08 15014
9d821fa5
KT		 150152013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15016
15017 PR target/56809
15018 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15019 instead of next_real_insn.
15020
71c581e7
MP		 150212013-04-03 Marek Polacek <polacek@redhat.com>
15022
15023 PR sanitizer/55702
33159866 15024 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 15025
4a32ef80
KT		 150262013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15027
15028 PR target/56809
15029 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15030 next_real_insn.
15031 (thumb1_output_casesi): Likewise.
15032 (thumb2_output_casesi): Likewise.
15033
1b2253d4
RB		 150342013-04-03 Richard Biener <rguenther@suse.de>
15035
15036 PR tree-optimization/56817
15037 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15038 Split out ...
15039 (tree_unroll_loops_completely_1): ... new function to manually
15040 walk the loop tree, properly defering outer loops of unrolled
15041 loops to later iterations.
15042
38000232
MG		 150432013-04-03 Marc Glisse <marc.glisse@inria.fr>
15044
15045 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15046 (vectorizable_load): Likewise.
15047 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15048 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15049
3fa35298
MG		 150502013-04-03 Marc Glisse <marc.glisse@inria.fr>
15051
15052 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15053 BIT_FIELD_REF.
15054
b3d45ff0
UW		 150552013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15056
15057 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15058
ec9202a8
BC		 150592013-04-03 Bin Cheng <bin.cheng@arm.com>
15060
15061 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15062
6805bd36
MG		 150632013-04-03 Marc Glisse <marc.glisse@inria.fr>
15064
15065 PR tree-optimization/56790
33159866
UB		 15066 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15067 folding.
6805bd36 15068
da694a77
MG		 150692013-04-03 Marc Glisse <marc.glisse@inria.fr>
15070
15071 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15072 Handle VEC_MERGE.
15073 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15074 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15075 equal arguments.
15076
4e7d7b3d
JJ		 150772013-04-03 Jakub Jelinek <jakub@redhat.com>
15078
15079 PR c/19449
15080 * tree.h (force_folding_builtin_constant_p): New decl.
15081 * builtins.c (force_folding_builtin_constant_p): New variable.
15082 (fold_builtin_constant_p): Fold immediately also if
15083 force_folding_builtin_constant_p.
15084
e6c9d234
RB		 150852013-04-03 Richard Biener <rguenther@suse.de>
15086
15087 PR tree-optimization/56812
15088 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15089 DRs of the same interleaving chain are independent.
15090
984d07dd
JM		 150912013-04-02 Jason Merrill <jason@redhat.com>
15092
15093 * gdbinit.in (pbb): Use debug fn.
15094
622849c9
LC		 150952013-04-02 Lawrence Crowl <crowl@google.com>
15096
15097 * sese.h (struct ivtype_map_elt_s): Remove unused.
15098 (extern debug_ivtype_map): Remove unused.
15099 (extern eq_ivtype_map_elts): Remove unused.
15100 * sese.c (debug_ivtype_map): Removed unused.
15101 (debug_ivtype_map_1): Removed unused.
15102 (debug_ivtype_elt): Remove unused.
15103 (eq_ivtype_map_elts): Remove unused.
15104
15105
82c0e1a0
KT		 151062013-04-02 Kai Tietz <ktietz@redhat.com>
15107
15108 PR target/52790
15109 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
15110 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB		 15111 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
15112 function.
82c0e1a0
KT		 15113 (legitimize_pe_coff_symbol): Likewise.
15114 (is_imported_p): New helper-function.
15115 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
15116 for Windows x64 targets.
15117 (ix86_expand_prologue): Optimize for pe-coff targets.
15118 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
15119 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
15120 medium/large code-model.
15121 (legitimize_pic_address): Likewise.
15122 (legitimize_tls_address): Likewise.
15123 (ix86_expand_call): Likewise.
15124 (x86_output_mi_thunk): Likewise.
15125 (get_dllimport_decl): Add new beimport argument.
15126 (construct_plt_address): Don't assert for x64 pe-coff targets.
15127 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
15128 targets.
15129 (SYMBOL_FLAG_STUBVAR): New macro.
15130 (SYMBOL_REF_STUBVAR_P): Likewise.
15131 * config/i386/winnt.c (stub_list): New structure.
15132 (stub_head): New local variable.
15133 (i386_pe_record_stub): New function.
15134 (i386_pe_file_end): Emit refptr-stubs.
15135
5d751b0c
JJ		 151362013-04-02 Jakub Jelinek <jakub@redhat.com>
15137
09bb4c99
JJ		 15138 PR rtl-optimization/56745
15139 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
15140 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
15141
a1d8947a
JJ		 15142 PR c++/34949
15143 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
15144 and both of them are MEM_REFs, just compare first argument for
15145 equality and attempt to deal even with differing offsets.
15146
5d751b0c
JJ		 15147 PR c++/34949
15148 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
15149 of gimple_clobber_p to be MEM_REF.
15150 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
15151 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
15152 after gimplification.
15153 * asan.c (get_mem_ref_of_assignment): Don't instrument
15154 gimple_clobber_p stmts.
15155 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
15156 gimple_clobber_p stmt if they have MEM_REF lhs and
15157 are dead because of another gimple_clobber_p stmt.
15158 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15159 gimple_clobber_p stmts like debug stmts.
15160 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15161 that refer to unused VAR_DECLs or uninitialized values.
15162 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15163 gimple_clobber_p stmts if they refer to removed parameters.
15164 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15165 formatting.
15166
e594716a
UB		 151672013-04-02 Uros Bizjak <ubizjak@gmail.com>
15168
15169 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15170 using SWI48 mode attribute.
15171
7a80735b
WM		 151722013-04-02 Wei Mi <wmi@google.com>
15173
15174 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15175 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15176 *<rotate_insn><mode>3_mask in i386.md.
15177
f423a9e4
AI		 151782013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15179
15180 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15181
90eb75f2
RB		 151822013-04-02 Richard Biener <rguenther@suse.de>
15183
15184 PR tree-optimization/56778
15185 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15186 Runtime alias tests are not supported for gather loads.
15187 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15188 stmts referenced from SSA operands before updating SSA form.
15189
d8c69a92
IC		 151902013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15191 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15192
15193 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB		 15194 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15195 * config/arm/cortex-a53.md: New file.
15196 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15197 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15198 * config/arm/arm.c (arm_issue_rate): Likewise.
15199 * config/arm/arm-tune.md: Regenerate
15200 * config/arm/arm-tables.opt: Regenerate.
15201 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 15202
239eb04c
ZC		 152032013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15204
15205 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15206 non-static link.
15207
c902d3c8
SN		 152082013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15209
d8c69a92
IC		 15210 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15211 scalar load/store operations using B/H registers.
c902d3c8
SN		 15212 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15213
051d0e2f
SN		 152142013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15215
15216 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15217 scalar move.
15218 * config/aarch64/aarch64.c
15219 (aarch64_simd_scalar_immediate_valid_for_move): New.
15220 * config/aarch64/aarch64-protos.h
15221 (aarch64_simd_scalar_immediate_valid_for_move): New.
15222 * config/aarch64/constraints.md (Dh, Dq): New.
15223 * config/aarch64/iterators.md (hq): New.
15224
0ee1e3d9
EB		 152252013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15226
15227 * reorg.c (get_branch_condition): Deal with conditional returns.
15228 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15229
136fb3f7
WM		 152302013-04-01 Wei Mi <wmi@google.com>
15231
15232 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15233 Truncate operand 2 using %b asm operand modifier.
15234 (*<shift_insn><mode>3_mask): Ditto.
15235 (*<rotate_insn><mode>3_mask): Ditto.
15236
6388c738
SB		 152372013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15238
15239 PR middle-end/56798
15240 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15241
f7a4d826
KK		 152422013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15243
15244 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15245 of next_real_insn.
15246 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15247
4b943a49
LC		 152482013-03-30 Lawrence Crowl <crowl@google.com>
15249
15250 * dse.c (clear_alias_sets): Remove never set.
15251 (disqualified_clear_alias_sets): Remove never set.
15252 (clear_alias_mode_pool): Remove never set.
15253 (dse_step0): Remove condition that is never true.
15254 (canon_address): Remove condition that is never true.
15255 (dse_step7): Remove condition that is never true.
15256 (rest_of_handle_dse): Remove condition that is never true.
15257 (rest_of_handle_dse::did_global): Remove never read from above.
15258 (dse_step2_spill): Remove never called from above.
15259 (dse_step5_spill): Remove never called from above.
15260
39718607
SB		 152612013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15262
da5c6bde
SB		 15263 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15264 JUMP_TABLE_DATA changes.
15265 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15266 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15267 (Insns) <jump_table_data>: New entry.
15268 * doc/tm.texi: Regenerate.
15269
39718607
SB		 15270 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15271
15272 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15273 for table jump at the end of a basic block using tablejump_p.
15274 * targhooks.c (default_invalid_within_doloop): Likewise.
15275 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15276 target hook implementation that is identical to the default hook.
15277 (rs6000_invalid_within_doloop): Remove.
15278
15279 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15280 unused variable from tablejump_p call.
15281
15282 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15283 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15284 (INSN_DELETED_P): Likewise.
15285 (emit_jump_table_data): New prototype.
15286 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15287 after 4th as unused.
15288 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15289 * sched-vis.c (print_insn): Likewise.
15290 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15291 insn for compatibility with back ends that use next_active_insn to
15292 identify jump table data.
15293 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15294 (remove_insn): Likewise.
15295 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15296 to be emitted.
15297 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15298 (emit_jump_table_data): New function.
15299
15300 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15301 basic block, a JUMP_TABLE_DATA never is.
15302 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15303 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15304 off from code handling real insns.
15305 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15306 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15307 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15308 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15309 is not a NONDEBUG_INSN_P.
15310 * ira-costs.c (scan_one_insn): Likewise.
15311 * jump.c (mark_all_labels): Likewise.
15312 (mark_jump_label_1): Likewise.
15313 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15314 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15315 a basic block.
15316 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15317 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15318 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15319 JUMP_TABLE_DATA_P insns.
15320 (calculate_elim_costs_all_insns): Likewise.
15321 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15322 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15323 (delete_output_reload): Code style fixups.
15324 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15325 insn flags on this non-insn.
15326 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15327 as scheduling barriers, for pre-change compatibility.
15328 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15329 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15330
15331 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15332 redundant JUMP_TABLE_DATA_P test.
15333 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15334 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15335 (frv_for_each_packet): Likewise.
15336 * config/i386/i386.c (min_insn_size): Likewise.
15337 (ix86_avoid_jump_mispredicts): Likewise.
15338 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15339 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15340 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15341 (mips16_insn_length): Robustify.
15342 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15343 (mips16_split_long_branches): Likewise.
15344 * config/pa/pa.c (pa_combine_instructions): Likewise.
15345 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15346 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15347 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15348 as contributing to pool range lengths.
15349 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15350 Remove redundant JUMP_TABLE_DATA_P test.
15351 (sh_loop_align): Likewise.
15352 (split_branches): Likewise.
15353 (sh_insn_length_adjustment): Likewise.
15354 * config/spu/spu.c (get_branch_target): Likewise.
15355
0208f7da
JH		 153562013-03-29 Jan Hubicka <jh@suse.cz>
15357
15358 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
15359 gcov streaming; stream hot bb threshold to ltrans.
15360 * predict.c (get_hot_bb_threshold): Break out from ....
15361 (maybe_hot_count_p): ... here.
15362 (set_hot_bb_threshold): New function.
15363 * lto-section-in.c (lto_section_name): Add profile.
15364 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
15365 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
15366 and data-streamer.h
15367 (histogram_entry): New structure.
15368 (histogram, histogram_pool): New global vars.
15369 (histogram_hash): New structure.
15370 (histogram_hash::hash): New method.
15371 (histogram_hash::equal): Likewise.
15372 (account_time_size): New function.
15373 (cmp_counts): New function.
15374 (dump_histogram): New function.
15375 (ipa_profile_generate_summary): New function.
15376 (ipa_profile_write_summary): New function.
15377 (ipa_profile_read_summary): New function.
15378 (ipa_profile): Decide on threshold.
e594716a
UB		 15379 (pass_ipa_profile): Add ipa_profile_write_summary and
15380 ipa_profile_read_summary.
0208f7da
JH		 15381 * Makefile.in (ipa.o): Update dependencies.
15382 * lto-streamer.h (LTO_section_ipa_profile): New section.
15383
5a6ccc94
GDR		 153842013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
15385
15386 * tree.h (VAR_P): New.
15387
39385fa6
PC		 153882013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
15389
15390 PR lto/56777
15391 * doc/invoke.texi ([-fwhole-program]): Fix typo.
15392
34f0d87a
SB		 153932013-03-29 Steven Bosscher <steven@gcc.gnu.org>
15394
15395 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
15396 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15397 (control_flow_insn_p): Likewise.
15398 * cfgrtl.c (duplicate_insn_chain): Likewise.
15399 * final.c (get_attr_length_1): Likewise.
15400 (shorten_branches): Likewise.
15401 (final_scan_insn): Likewise.
15402 * function.c (instantiate_virtual_regs): Likewise.
15403 * gcse.c (insert_insn_end_basic_block): Likewise.
15404 * ira-costs.c (scan_one_insn): Likewise.
15405 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15406 * lra.c (check_rtl): Likewise.
15407 * reload1.c (elimination_costs_in_insn): Likewise.
15408 * reorg.c (follow_jumps): Likewise.
15409
15410 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
15411 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15412 (thumb_far_jump_used_p): Likewise.
15413 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
15414 (workaround_speculation): Likewise.
15415 (add_sched_insns_for_speculation): Likewise.
15416 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
15417 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15418 (frv_for_each_packet): Likewise.
15419 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
15420 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
15421 (final_emit_insn_group_barriers): Likewise.
15422 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15423 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15424 (mips16_insn_length): Likewise.
15425 * config/pa/pa.c (pa_reorg): Likewise.
15426 (pa_combine_instructions): Likewise.
15427 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
15428 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
15429 (sh_reorg): Likewise.
15430 (split_branches): Likewise.
15431 * config/spu/spu.c (get_branch_target): Likewise.
15432
15433 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
15434 JUMP_TABLE_DATA_P.
15435
4ac761b0
KY		 154362013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
15437
39385fa6 15438 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY		 15439 Fix declaration name.
15440
58a49189
LC		 154412013-03-28 Lawrence Crowl <crowl@google.com>
15442
15443 * graphds.h (struct graph.indicies): Remove unused.
15444 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
15445 (SCOP_ORIGINAL_PDDRS): Remove unused.
15446 * sese.h (extern insert_loop_close_phis): Removed unused.
15447 (extern insert_guard_phis): Removed unused.
15448 (extern ivtype_map_elt_info): Removed unused.
15449 (new_ivtype_map_elt): Removed unused.
15450 * sese.c (ivtype_map_elt_info): Removed unused.
15451
7b3b6ae4
LC		 154522013-03-28 Lawrence Crowl <crowl@google.com>
15453
15454 * Makefile.in: Add several missing include dependences.
15455 (DUMPFILE_H): New.
15456 (test-dump.o): New. This object is not added to any executable,
15457 but is present for ad-hoc testing.
15458 * bitmap.c
15459 (debug (const bitmap_head_def &)): New.
15460 (debug (const bitmap_head_def *)): New.
15461 * bitmap.h
15462 (extern debug (const bitmap_head_def &)): New.
15463 (extern debug (const bitmap_head_def *)): New.
15464 * cfg.c
15465 (debug (edge_def &)): New.
15466 (debug (edge_def *)): New.
15467 * cfghooks.c
15468 (debug (basic_block_def &)): New.
15469 (debug (basic_block_def *)): New.
15470 * dumpfile.h
15471 (dump_node (const_tree, int, FILE *)): Correct source file.
15472 * dwarf2out.c
15473 (debug (die_struct &)): New.
15474 (debug (die_struct *)): New.
15475 * dwarf2out.h
15476 (extern debug (die_struct &)): New.
15477 (extern debug (die_struct *)): New.
15478 * gimple-pretty-print.c
15479 (debug (gimple_statement_d &)): New.
15480 (debug (gimple_statement_d *)): New.
15481 * gimple-pretty-print.h
15482 (extern debug (gimple_statement_d &)): New.
15483 (extern debug (gimple_statement_d *)): New.
15484 * ira-build.c
15485 (debug (ira_allocno_copy &)): New.
15486 (debug (ira_allocno_copy *)): New.
15487 (debug (ira_allocno &)): New.
15488 (debug (ira_allocno *)): New.
15489 * ira-int.h
15490 (extern debug (ira_allocno_copy &)): New.
15491 (extern debug (ira_allocno_copy *)): New.
15492 (extern debug (ira_allocno &)): New.
15493 (extern debug (ira_allocno *)): New.
15494 * ira-lives.c
15495 (debug (live_range &)): New.
15496 (debug (live_range *)): New.
15497 * lra-int.h
15498 (debug (lra_live_range &)): New.
15499 (debug (lra_live_range *)): New.
15500 * lra-lives.c
15501 (debug (lra_live_range &)): New.
15502 (debug (lra_live_range *)): New.
15503 * omega.c
15504 (debug (omega_pb_d &)): New.
15505 (debug (omega_pb_d *)): New.
15506 * omega.h
15507 (extern debug (omega_pb_d &)): New.
15508 (extern debug (omega_pb_d *)): New.
15509 * print-rtl.c
15510 (debug (const rtx_def &)): New.
15511 (debug (const rtx_def *)): New.
15512 * print-tree.c
15513 (debug_tree (tree): Move within file.
15514 (debug_raw (const tree_node &)): New.
15515 (debug_raw (const tree_node *)): New.
15516 (dump_tree_via_hooks (const tree_node *, int)): New.
15517 (debug (const tree_node &)): New.
15518 (debug (const tree_node *)): New.
15519 (debug_verbose (const tree_node &)): New.
15520 (debug_verbose (const tree_node *)): New.
15521 (debug_head (const tree_node &)): New.
15522 (debug_head (const tree_node *)): New.
15523 (debug_body (const tree_node &)): New.
15524 (debug_body (const tree_node *)): New.
15525 (debug_vec_tree (tree): Move and reimplement in terms of dump.
15526 (debug (vec<tree, va_gc> &)): New.
15527 (debug (vec<tree, va_gc> *)): New.
15528 * rtl.h
15529 (extern debug (const rtx_def &)): New.
15530 (extern debug (const rtx_def *)): New.
15531 * sbitmap.c
15532 (debug_raw (simple_bitmap_def &)): New.
15533 (debug_raw (simple_bitmap_def *)): New.
15534 (debug (simple_bitmap_def &)): New.
15535 (debug (simple_bitmap_def *)): New.
15536 * sbitmap.h
15537 (extern debug (simple_bitmap_def &)): New.
15538 (extern debug (simple_bitmap_def *)): New.
15539 (extern debug_raw (simple_bitmap_def &)): New.
15540 (extern debug_raw (simple_bitmap_def *)): New.
15541 * sel-sched-dump.c
15542 (debug (vinsn_def &)): New.
15543 (debug (vinsn_def *)): New.
15544 (debug_verbose (vinsn_def &)): New.
15545 (debug_verbose (vinsn_def *)): New.
15546 (debug (expr_def &)): New.
15547 (debug (expr_def *)): New.
15548 (debug_verbose (expr_def &)): New.
15549 (debug_verbose (expr_def *)): New.
15550 (debug (vec<rtx> &)): New.
15551 (debug (vec<rtx> *)): New.
15552 * sel-sched-dump.h
15553 (extern debug (vinsn_def &)): New.
15554 (extern debug (vinsn_def *)): New.
15555 (extern debug_verbose (vinsn_def &)): New.
15556 (extern debug_verbose (vinsn_def *)): New.
15557 (extern debug (expr_def &)): New.
15558 (extern debug (expr_def *)): New.
15559 (extern debug_verbose (expr_def &)): New.
15560 (extern debug_verbose (expr_def *)): New.
15561 (extern debug (vec<rtx> &)): New.
15562 (extern debug (vec<rtx> *)): New.
15563 * sel-sched-ir.h
15564 (_list_iter_cond_expr): Make inline instead of static.
15565 * sreal.c
15566 (debug (sreal &)): New.
15567 (debug (sreal *)): New.
15568 * sreal.h
15569 (extern debug (sreal &)): New.
15570 (extern debug (sreal *)): New.
15571 * tree.h
15572 (extern debug_raw (const tree_node &)): New.
15573 (extern debug_raw (const tree_node *)): New.
15574 (extern debug (const tree_node &)): New.
15575 (extern debug (const tree_node *)): New.
15576 (extern debug_verbose (const tree_node &)): New.
15577 (extern debug_verbose (const tree_node *)): New.
15578 (extern debug_head (const tree_node &)): New.
15579 (extern debug_head (const tree_node *)): New.
15580 (extern debug_body (const tree_node &)): New.
15581 (extern debug_body (const tree_node *)): New.
15582 (extern debug (vec<tree, va_gc> &)): New.
15583 (extern debug (vec<tree, va_gc> *)): New.
15584 * tree-cfg.c
15585 (debug (struct loop &)): New.
15586 (debug (struct loop *)): New.
15587 (debug_verbose (struct loop &)): New.
15588 (debug_verbose (struct loop *)): New.
15589 * tree-dump.c: Add header dependence.
15590 * tree-flow.h
15591 (extern debug (struct loop &)): New.
15592 (extern debug (struct loop *)): New.
15593 (extern debug_verbose (struct loop &)): New.
15594 (extern debug_verbose (struct loop *)): New.
15595 * tree-data-ref.c
15596 (debug (data_reference &)): New.
15597 (debug (data_reference *)): New.
15598 (debug (vec<data_reference_p> &)): New.
15599 (debug (vec<data_reference_p> *)): New.
15600 (debug (vec<ddr_p> &)): New.
15601 (debug (vec<ddr_p> *)): New.
15602 * tree-data-ref.h
15603 (extern debug (data_reference &)): New.
15604 (extern debug (data_reference *)): New.
15605 (extern debug (vec<data_reference_p> &)): New.
15606 (extern debug (vec<data_reference_p> *)): New.
15607 (extern debug (vec<ddr_p> &)): New.
15608 (extern debug (vec<ddr_p> *)): New.
15609 * tree-ssa-alias.c
15610 (debug (pt_solution &)): New.
15611 (debug (pt_solution *)): New.
15612 * tree-ssa-alias.h
15613 (extern debug (pt_solution &)): New.
15614 (extern debug (pt_solution *)): New.
15615 * tree-ssa-alias.c
15616 (debug (_var_map &)): New.
15617 (debug (_var_map *)): New.
15618 (debug (tree_live_info_d &)): New.
15619 (debug (tree_live_info_d *)): New.
15620 * tree-ssa-alias.h
15621 (extern debug (_var_map &)): New.
15622 (extern debug (_var_map *)): New.
15623 (extern debug (tree_live_info_d &)): New.
15624 (extern debug (tree_live_info_d *)): New.
15625
be77e1e5
JH		 156262013-03-28 Jan Hubicka <jh@suse.cz>
15627
15628 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
15629
777e6976
IB		 156302013-03-28 Ian Bolton <ian.bolton@arm.com>
15631
15632 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
15633 record only when desired or required.
15634
9cd347ae
UB		 156352013-03-28 Uros Bizjak <ubizjak@gmail.com>
15636
15637 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
15638 *vec_extractv2di_1_rex64. Use x64 isa attribute.
15639
33159866 156402013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N		 15641
15642 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
15643 (*andsi3_compare0_uxtw): New pattern.
15644 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
15645 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
15646
89ab31c1
JH		 156472013-03-28 Jan Hubicka <jh@suse.cz>
15648
15649 * data-streamer-in.c (streamer_read_gcov_count): New function.
15650 * gimple-streamer-out.c: Include value-prof.h.
15651 (output_gimple_stmt): Output histogram.
15652 (output_bb): Use streamer_write_gcov_count.
15653 * value-prof.c: Include data-streamer.h
15654 (dump_histogram_value): Add HIST_TYPE_MAX.
15655 (stream_out_histogram_value): New function.
15656 (stream_in_histogram_value): New function.
15657 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
15658 (stream_out_histogram_value, stream_in_histogram_value): Declare.
15659 * data-streamer-out.c (streamer_write_gcov_count): New function.
15660 (streamer_write_gcov_count_stream): New function.
15661 * lto-cgraph.c (lto_output_edge): Update counter streaming.
15662 (lto_output_node): Likewise.
15663 (input_node, input_edge): Likewise.
15664 * lto-streamer-out.c (output_cfg): Update streaming.
15665 * lto-streamer-in.c (input_cfg): Likewise.
15666 * data-streamer.h (streamer_write_gcov_count,
15667 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
15668 * gimple-streamer-in.c: Include value-prof.h
15669 (input_gimple_stmt): Input histograms.
15670 (input_bb): Update profile streaming.
15671
e594716a 156722013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 15673
33159866 15674 * genmodes.c (emit_max_int): New function.
8697be17 15675 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC		 15676 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
15677 Added doc.
8697be17 15678 * machmode.def: Fixed comment.
89ab31c1 15679
e594716a 156802013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ		 15681
15682 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ		 15683 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
15684
74946978
MP		 156852013-03-28 Marek Polacek <polacek@redhat.com>
15686 Richard Biener <rguenther@suse.de>
15687
15688 PR tree-optimization/56695
15689 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
15690 build signed result of a vector comparison.
15691 * tree-cfg.c (verify_gimple_comparison): Check that a result
15692 of a vector comparison has signed type.
15693
a64b9c26
RB		 156942013-03-28 Richard Biener <rguenther@suse.de>
15695
15696 PR tree-optimization/37021
15697 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
15698 do not restrict gaps between groups.
15699 * tree-vect-stmts.c (vectorizable_load): Properly account for
15700 a gap between groups.
15701
a9dc2a2f
EB		 157022013-03-28 Eric Botcazou <ebotcazou@adacore.com>
15703
15704 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
15705 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
15706 is not enabled.
15707
53cb97f9
GP		 157082013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
15709
15710 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
15711 * doc/extend.texi (Named Address Spaces): Ditto.
15712 (Variable Attributes): Ditto.
15713
b802ae5c
KT		 157142013-03-27 Kai Tietz <ktietz@redhat.com>
15715
eddae10a
KT		 15716 * config.build: Add support for cygwin x64 target.
15717 * config.gcc: Likewise.
15718 * config.host: Likewise.
15719 * configure.ac: Likewise
15720 * configure: Regenerated.
b802ae5c 15721
371e77e3 157222013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT		 15723
15724 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
15725 * config/i386/t-cygwin-w64: New file.
15726 * config/i386/cygwin-w64.h: New file.
15727 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
15728 and add support for x64-cygwin target.
15729 (CPP_SPEC): Likewise.
15730 (CXX_WRAP_SPEC_LIST): Undefine before define.
15731 (LIBGCJ_SONAME): Use 15 as version.
15732
f49b33cb
RB		 157332013-03-27 Richard Biener <rguenther@suse.de>
15734
15735 PR tree-optimization/56716
15736 * tree-ssa-structalias.c (perform_var_substitution): Adjust
15737 dumping for ref nodes.
15738
b37a6ce5
MJ		 157392013-03-27 Martin Jambor <mjambor@suse.cz>
15740
15741 PR tree-optimization/55334
15742 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
15743 restricted pointers to arrays.
15744
9469b9b2
GDR		 157452013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
15746
15747 * Makefile.in (.SUFFIXES): Add .cc.
15748 (.c.o): Apply same recipe for implicit rule .cc.o.
15749
7d24f650
RB		 157502013-03-27 Richard Biener <rguenther@suse.de>
15751
15752 PR tree-optimization/37021
15753 * tree-vect-data-refs.c (vect_check_strided_load): Allow
15754 REALPART/IMAGPART_EXPRs around the supported refs.
15755 * tree-ssa-structalias.c (find_func_aliases): Assume that
15756 floating-point values are not used to transfer pointers.
15757
2f251a05
AI		 157582013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
15759
d33d9e47
AI		 15760 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
15761 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
15762 * gcc/doc/tm.texi: Regenerate.
15763 * gcc/targhooks.h (default_has_ifunc_p): New.
15764 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
15765 * gcc/config/linux-protos.h: New file.
15766 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
2f251a05
AI		 15767 this hook for linux which disables support of indirect functions in
15768 android.
d33d9e47
AI		 15769 * gcc/config/linux-android.c: New file.
15770 * gcc/config/t-linux-android.c: Ditto.
15771 * gcc/config.gcc: Added new object file linux-android.o.
15772 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 15773 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
d33d9e47 15774 * gcc/varasm.c (do_assemble_alias): Likewise.
2f251a05
AI		 15775 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
15776 doesn't support indirect functions.
15777 * configure: Regenerate.
15778
78b4e425
BC		 157792013-03-27 Bin Cheng <bin.cheng@arm.com>
15780
15781 PR target/56102
15782 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
15783 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
15784 mult-word mode.
15785
a5ba7b92
AK		 157862013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15787
15788 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
15789
be624986
TG		 157902013-03-27 Terry Guo <terry.guo@arm.com>
15791
15792 * config/arm/arm-cores.def: Added core cortex-r7.
15793 * config/arm/arm-tune.md: Regenerated.
15794 * config/arm/arm-tables.opt: Regenerated.
15795 * doc/invoke.texi: Added entry for core cortex-r7.
15796
0a514f47
WL		 157972013-03-27 Walter Lee <walt@tilera.com>
15798
15799 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
15800 double-decrement of next_scratch_regno.
15801
9332b0d2
WL		 158022013-03-27 Walter Lee <walt@tilera.com>
15803
801d9b2a 15804 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL		 15805 input operands.
15806 (insn_v1mulus): Ditto.
15807 (insn_v2muls): Ditto.
15808
f54ea5dd
WL		 158092013-03-27 Walter Lee <walt@tilera.com>
15810
e594716a 15811 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL		 15812 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
15813
bf60f4ca
WL		 158142013-03-27 Walter Lee <walt@tilera.com>
15815
e594716a 15816 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL		 15817 (*sibcall_value): Ditto.
15818
e3b51eeb
WL		 158192013-03-27 Walter Lee <walt@tilera.com>
15820
15821 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
15822 (insn_mnz_v8qi): ... this ...
15823 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
15824 vector equivalent.
15825 (insn_v<n>mnz): Replaced by ...
15826 (insn_v1mnz): ... this ...
15827 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
15828 equivalent.
15829 (insn_mz_<mode>): Replaced by ...
15830 (insn_mz_v8qi): ... this ...
15831 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
15832 vector equivalent.
15833 (insn_v<n>mz): Replaced by ...
15834 (insn_v1mz): ... this ...
15835 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
15836 equivalent.
15837
4fc7b145
EB		 158382013-03-26 Eric Botcazou <ebotcazou@adacore.com>
15839
15840 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
15841
6f33817e
RM		 158422013-03-26 Roland McGrath <mcgrathr@google.com>
15843
15844 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
15845 than fprintf with a non-constant, non-format string.
15846
93a4145b
UB		 158472013-03-26 Uros Bizjak <ubizjak@gmail.com>
15848
15849 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
15850 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
15851 operand 0 predicate.
15852 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
15853 attribute. Use general_x64nomem_operand as operand 1 predicate.
15854 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
15855 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15856 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
15857 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15858 (mov<mode>_insv_1): Remove expander. Merge insn with
15859 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
15860 Use general_x64nomem_operand as operand 1 predicate.
15861 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
15862 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
15863 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15864 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
15865 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
15866 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
15867 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15868 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
15869 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
15870 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
15871 (general_x64nomem_operand): Ditto.
15872
c6a9ed5a 158732013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH		 15874
15875 * config/rtems.opt: Add -pthread option.
15876
31b0a960
RB		 158772013-03-26 Richard Biener <rguenther@suse.de>
15878
93a4145b
UB		 15879 * alias.c (find_base_term): Avoid redundant and not used recursion.
15880 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB		 15881 (true_dependence_1): Compute and pass base terms to base_alias_check.
15882 (write_dependence_p): Likewise.
15883 (may_alias_p): Likewise.
15884
79517551
SN		 158852013-03-26 Sofiane Naci <sofiane.naci@arm.com>
15886
15887 * config/aarch64/aarch64.c (aarch64_classify_address): Support
15888 PC-relative load in SI modes and above only.
15889
a76213b9
XQ		 158902013-03-26 Xinyu Qi <xyqi@marvell.com>
15891
15892 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB		 15893 * config/arm/iwmmxt.md (WCGR0): Update.
15894 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 15895
37ff9355
UB		 158962013-03-26 Uros Bizjak <ubizjak@gmail.com>
15897
15898 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
15899 Use x64 and nox64 isa attributes.
15900
30b0317c
RB		 159012013-03-26 Richard Biener <rguenther@suse.de>
15902
15903 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
15904 alignment computations and rely on get_object_alignment_1
15905 for the !TYPE_P case.
15906 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
15907
cbcdb140
WL		 159082013-03-26 Walter Lee <walt@tilera.com>
15909
15910 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
15911 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
15912
dd884525 159132013-03-25 Jeff Law <law@redhat.com>
be672e08 15914
33159866
UB		 15915 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
15916 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL		 15917 wrong version of prior patch.
15918
0051d3ec
WL		 159192013-03-25 Walter Lee <walt@tilera.com>
15920
15921 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
15922 TILEGX_INSN_SHUFFLEBYTES1.
15923 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
15924 shufflebytes1.
15925 (tilegx_builtins): Ditto.
15926 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
15927
065a3605
WL		 159282013-03-25 Walter Lee <walt@tilera.com>
15929
15930 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL		 15931 (floatunssisf2): New pattern.
15932 (floatsidf2): New pattern.
15933 (floatunssidf2): New pattern.
065a3605 15934
5b2a3c82
WL		 159352013-03-25 Walter Lee <walt@tilera.com>
15936
15937 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
15938 tests for constraint J, K, N, P.
15939
192ea533
WL		 159402013-03-25 Walter Lee <walt@tilera.com>
15941
15942 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
15943 Use indirect/pcrel encoding.
15944 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
15945 Ditto.
15946
855e0d0b
SE		 159472013-03-25 Steve Ellcey <sellcey@mips.com>
15948
15949 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
15950 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
15951 * config/mips/mips.c (mips_option_override): Set IMADD default.
15952 * config/mips/mips.h (PTF_AVOID_IMADD): New.
15953 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
15954 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
15955 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
15956 * doc/invoke.texi (-mimadd/-mno-imadd): New.
15957
39e45653
JL		 159582013-03-25 Jeff Law <law@redhat.com>
15959
15960 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
15961 slightly to avoid creating and folding useless trees. Simplify
15962 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
15963
7482c470
UB		 159642013-03-25 Uros Bizjak <ubizjak@gmail.com>
15965
15966 * config/i386/i386.md (*zero_extendsidi2): Merge with
15967 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
15968 * config/i386/predicates.md (x86_64_zext_operand): Rename from
15969 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
15970 targets. Clarify comment.
15971
4a53743e
MJ		 159722013-03-25 Martin Jambor <mjambor@suse.cz>
15973
15974 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
15975 pass-through jump functions differently.
15976 (ipa_read_jump_function): Likewise. Also use setter functions to set
15977 up jump functions.
15978
162712de
MJ		 159792013-03-25 Martin Jambor <mjambor@suse.cz>
15980
15981 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
15982 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
15983 process it.
15984 (ipa_get_indirect_edge_target): New function.
15985 (devirtualization_time_bonus): New parameter known_aggs, pass it to
15986 ipa_get_indirect_edge_target. Update all callers.
15987 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
15988 ipa_get_indirect_edge_target_1 instead of calling
15989 ipa_get_indirect_edge_target.
15990 (create_specialized_node): Pass aggvlas to
15991 ipcp_discover_new_direct_edges.
15992
a5a4c20a
KT		 159932013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15994
15995 * config/arm/arm.md (f_sels, f_seld): New types.
15996 (*cmov<mode>): New pattern.
15997 * config/arm/predicates.md (arm_vsel_comparison_operator): New
15998 predicate.
15999
f992b9fc
KT		 160002013-03-25 Kai Tietz <ktietz@redhat.com>
16001
16002 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16003 POSIX-printf for mingw-hosted builds.
16004
c16fd676
RB		 160052013-03-25 Richard Biener <rguenther@suse.de>
16006
16007 PR middle-end/56694
16008 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16009 must-not-throw stmt location.
16010
5ad29f12
KT		 160112013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16012
16013 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16014 Emit load-acquire versions when acq is true.
16015 (arm_emit_store_exclusive): Add rel parameter.
16016 Emit store-release versions when rel is true.
16017 (arm_split_compare_and_swap): Use acquire-release instructions
16018 instead.
16019 of barriers when appropriate.
16020 (arm_split_atomic_op): Likewise.
16021 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16022 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16023 (VUNSPEC_SLX): Likewise.
16024 (VUNSPEC_LDA): Likewise.
16025 (VUNSPEC_STL): Likewise.
16026 * config/arm/sync.md (atomic_load<mode>): New pattern.
16027 (atomic_store<mode>): Likewise.
16028 (arm_load_acquire_exclusive<mode>): Likewise.
16029 (arm_load_acquire_exclusivesi): Likewise.
16030 (arm_load_acquire_exclusivedi): Likewise.
16031 (arm_store_release_exclusive<mode>): Likewise.
16032
03a7dddb
CM		 160332013-03-25 Catherine Moore <clm@codesourcery.com>
16034
16035 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16036 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16037 * config/mip/predicates.md (lwsp_swsp_operand,
16038 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16039 sb16_operand, db4_operand, db7_operand, ib3_operand,
16040 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16041 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16042 andi16_operand): New predicates.
16043 * config/mips/mips.md (compression): New attribute.
16044 (enabled): New attribute.
16045 (length): Consider compression in computing length.
16046 (shift_compression): New code attribute.
16047 (*add<mode>3): New operands. Record compression.
16048 (sub<mode>3): Likewise.
16049 (one_cmpl<mode>2): Likewise.
16050 (*and<mode>3): Likewise.
16051 (*ior<mode>3): Likewise.
16052 (unnamed pattern for xor): Likewise.
16053 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16054 (*<optab><mode>3): Likewise.
16055 (*mov<mode>_internal: Likewise.
16056 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16057 (mips_unsigned_immediate_p): New.
16058 (umips_lwsp_swsp_address_p): New.
16059 (m16_based_address_p): New.
16060 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16061 (mips_unsigned_immediate_p): New prototype.
16062 (lwsp_swsp_address_p): New prototype.
16063 (m16_based_address_p): New prototype.
16064 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16065 (mips_signed_immediate_p): New function.
16066 (m16_based_address_p): New function.
16067 (lwsp_swsp_address_p): New function.
16068 (mips_print_operand_punctuation): Recognize short delay slot insns
16069 for microMIPS.add<mode>3"
16070
f35c297f 160712013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 16072
f35c297f
KT		 16073 PR target/56720
16074 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16075 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16076
051b9446
RB		 160772013-03-25 Richard Biener <rguenther@suse.de>
16078
16079 PR tree-optimization/56689
16080 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16081 any edge.
16082
374001cb
RB		 160832013-03-25 Richard Biener <rguenther@suse.de>
16084
16085 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16086 of bitmap.
16087 (memory_references): Likewise.
16088 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16089 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16090 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16091 (gather_mem_refs_in_loops): Fold into ...
16092 (analyze_memory_references): ... this. Move initialization
16093 to tree_ssa_lim_initialize.
16094 (fill_always_executed_in): Rename to ...
16095 (fill_always_executed_in_1): ... this.
16096 (fill_always_executed_in): Move contains_call computation to
16097 this new function from ...
16098 (tree_ssa_lim_initialize): ... here.
16099 (tree_ssa_lim): Call fill_always_executed_in.
16100
57895947
EB		 161012013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16102
16103 * postreload.c (reload_combine): Fix code detecting returns.
16104
1e1b18c1
EB		 161052013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16106
16107 * function.c (emit_use_return_register_into_block): On cc0 targets,
16108 do not emit the sequence between cc0 setter and user.
16109
9216baf1
KT		 161102013-03-25 Kai Tietz <ktietz@redhat.com>
16111
1e1b18c1
EB		 16112 * config/i386/predicates.md (local_symbolic_operand): Interpret
16113 dll-imported symbols as none-local.
9216baf1 16114
76421b44
RB		 161152013-03-25 Richard Biener <rguenther@suse.de>
16116
16117 * tree-ssa-loop-im.c (struct depend): Remove.
16118 (struct lim_aux_data): Make depends a vec of gimples.
16119 (free_lim_aux_data): Adjust.
16120 (add_dependency): Likewise.
16121 (set_level): Likewise.
16122
d154bfa2
RB		 161232013-03-25 Richard Biener <rguenther@suse.de>
16124
16125 PR middle-end/56434
16126 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
16127 the pointer returned by calls with ECF_MALLOC set.
16128
2cd9804e
UB		 161292013-03-24 Uros Bizjak <ubizjak@gmail.com>
16130
acef5fe0
UB		 16131 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
16132
161332013-03-24 Uros Bizjak <ubizjak@gmail.com>
16134
16135 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB		 16136 using MMXMODE mode iterator.
16137 (*move<mode>_internal): Merge with *movv2sf_internal and
16138 *movv2sf_internal_rex64 using MMXMODE mode iterator.
16139
225ccc68
SB		 161402013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16141
85c0f02d
SB		 16142 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
16143 (record_last_mem_set_info): Likewise.
16144
225ccc68
SB		 16145 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
16146 of XNEWVEC followed by memset.
16147 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
16148
b64925dc
SB		 161492013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16150
16151 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
16152 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
16153 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
16154 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
16155 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
16156 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
16157 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16158 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16159 BARRIER_P instead of GET_CODE.
16160
b9aaf52e
EB		 161612013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16162
16163 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16164 inaccuracy in the probing code.
16165
16166 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16167 (ctrapdi4): Likewise.
16168
66de4d7c
EB		 161692013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16170
16171 * calls.c (expand_call): Add missing guard to code handling return
16172 of non-BLKmode structures in MSB.
16173 * function.c (expand_function_end): Likewise.
16174
4f25c8fb
EB		 161752013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16176
16177 * combine.c (try_combine): Adjust comment. Do not add the set of
16178 insn #0 if the destination indirectly is set or dies in insn #2.
16179 Tidy up code to distribute a new note.
16180
23b7850d
UB		 161812013-03-22 Uros Bizjak <ubizjak@gmail.com>
16182
16183 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16184 also for alternatives 16 and 17.
16185
bed852cf
UB		 161862013-03-22 Uros Bizjak <ubizjak@gmail.com>
16187
16188 * config/i386/sse.md (*mov<mode>_internal): Merge with
16189 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16190 Emit insn template depending on type attribute. Use
16191 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16192 movd instead of movq mnemonic for interunit moves. Rewrite mode
16193 attribute calculation. Remove unit attribute calculation.
16194 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16195 Set prefix_data16 attribute for DImode ssemov types.
16196 Use Ym instead of y for SSE-MMX conversion alternatives.
16197 Reorder operand constraints.
16198
1e8a7937
SB		 161992013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16200
16201 * df.h (df_insn_delete): Adjust prototype.
16202 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16203 and let it decide whether mark the basic block dirty.
16204 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16205 * df-scan.c (df_insn_info_delete): New helper function, split
16206 off from df_insn_delete.
16207 (df_scan_free_bb_info): Use it.
16208 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16209 Likewise.
16210 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16211 that the insn is actually an insn and it has a non-NULL basic block.
16212 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16213
e14d094c
RB		 162142013-03-22 Richard Biener <rguenther@suse.de>
16215
16216 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16217 dep_ref members.
16218 (mem_ref_alloc): Do not allocate them.
16219 (refs_independent_p): Do not query or maintain a cache.
16220
e9cf7316
RB		 162212013-03-22 Richard Biener <rguenther@suse.de>
16222
16223 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16224 (gather_mem_refs_in_loops): Do not compute it.
16225 (analyze_memory_references): Do not allocate it.
16226 (tree_ssa_lim_finalize): Do not free it.
16227 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16228
f046e81b
RB		 162292013-03-22 Richard Biener <rguenther@suse.de>
16230
16231 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 16232
a05c0ddf
IB		 162332013-03-22 Ian Bolton <ian.bolton@arm.com>
16234
16235 * config/aarch64/aarch64.c (aarch64_print_operand): New
16236 format specifier for printing a constant in hex.
16237 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16238 format specifier for printing second operand.
16239
c00217fc
RB		 162402013-03-22 Richard Biener <rguenther@suse.de>
16241
16242 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16243 bitmaps.
16244 (gather_mem_refs_in_loops): Perform store accumulation here.
16245 (create_vop_ref_mapping_loop): Remove.
16246 (create_vop_ref_mapping): Likewise.
16247 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 16248 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB		 16249 (record_indep_loop): Remove.
16250 (record_dep_loop): New function.
16251 (ref_indep_loop_p_1): Adjust to only walk over references
16252 in the loop, not its subloops.
16253 (ref_indep_loop_p): Rename to ...
16254 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16255 maintaining a more fine-grained cache.
16256 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16257 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16258
15d19bf8
RB		 162592013-03-22 Richard Biener <rguenther@suse.de>
16260
16261 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16262 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16263 aggregate mem_ref_loc.
16264 (free_mem_ref_locs): Inline into ...
16265 (memref_free): ... this and adjust.
16266 (mem_ref_alloc): Adjust.
16267 (mem_ref_locs_alloc): Remove.
16268 (record_mem_ref_loc): Adjust.
16269 (get_all_locs_in_loop): Rewrite into ...
16270 (for_all_locs_in_loop): ... this iterator.
16271 (rewrite_mem_ref_loc): New functor.
16272 (rewrite_mem_refs): Use for_all_locs_in_loop.
16273 (sm_set_flag_if_changed): New functor.
16274 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16275 (ref_always_accessed): New functor.
16276 (ref_always_accessed_p): Use for_all_locs_in_loop.
16277
6f37411d
MG		 162782013-03-21 Marc Glisse <marc.glisse@inria.fr>
16279
16280 * tree-pass.h (PROP_gimple_lvec): New.
16281 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16282 (init_optimization_passes): Move pass_lower_vector.
16283 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16284 PROP_gimple_lvec.
16285 (pass_lower_vector): Provide PROP_gimple_lvec.
16286 (pass_lower_vector_ssa): Likewise.
16287 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16288
511ef689
MW		 162892013-03-21 Mark Wielaard <mjw@redhat.com>
16290
16291 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16292
cc1df30b
UB		 162932013-03-21 Uros Bizjak <ubizjak@gmail.com>
16294
16295 * config/i386/i386.md (*movdi_internal): Disparage slightly
16296 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16297 conversion alternatives.
16298
c845cfe1
JJ		 162992013-03-21 Jakub Jelinek <jakub@redhat.com>
16300
37e99116
JJ		 16301 PR middle-end/48087
16302 * diagnostic.def (DK_WERROR): New kind.
16303 * diagnostic.h (werrorcount): Define.
16304 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16305 promoted to DK_ERROR, increment DK_WERROR counter instead of
16306 DK_ERROR counter.
16307 * toplev.c (toplev_main): Call print_ignored_options even if
16308 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16309 even if just werrorcount is non-zero.
16310
c845cfe1
JJ		 16311 PR debug/55608
16312 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16313 on failure.
16314 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16315 (string_cst_pool_decl): New function.
16316 (optimize_one_addr_into_implicit_ptr): New function.
16317 (resolve_addr_in_expr): Optimize DWARF location expression
16318 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16319 which doesn't live in memory, but has DW_AT_location or
16320 DW_AT_const_value, or refers to a string literal, into
16321 DW_OP_GNU_implicit_pointer.
16322 (optimize_location_into_implicit_ptr): New function.
16323 (resolve_addr): If removing DW_AT_location of a variable because
16324 it was DW_OP_addr of address of the variable, but the variable doesn't
16325 live in memory, try to emit const value attribute for the initializer.
16326
08e0cda6
MG		 163272013-03-21 Marc Glisse <marc.glisse@inria.fr>
16328
16329 * tree.h (VECTOR_TYPE_P): New macro.
16330 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16331 TYPE_MODE): Use it.
16332 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16333 VEC_COND_EXPR cannot be lvalues.
16334 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16335
d08633b4
MG		 163362013-03-21 Marc Glisse <marc.glisse@inria.fr>
16337
16338 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16339 Restrict the transformation to equal modes.
16340
e6647190
RB		 163412013-03-21 Richard Biener <rguenther@suse.de>
16342
16343 PR tree-optimization/39326
16344 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16345 (MEM_ANALYZABLE): Adjust.
16346 (record_mem_ref_loc): Move bitmap ops ...
16347 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16348 unanalyzable refs, do not record locations for it.
16349 (analyze_memory_references): Allocate ref zero as shared
16350 unanalyzable ref.
16351 (refs_independent_p): Do not test for unanalyzed mems here.
16352 (ref_indep_loop_p_1): Special-case disambiguation against
16353 the unanalyzed ref.
cc1df30b 16354 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 16355
65074f54
CL		 163562013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
16357
16358 * config/arm/arm-protos.h (tune_params): Add
16359 prefer_neon_for_64bits field.
16360 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
16361 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
16362 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
16363 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
16364 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
16365 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
16366 (arm_option_override): Handle -mneon-for-64bits new option.
16367 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
16368 (prefer_neon_for_64bits): Declare new variable.
16369 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 16370 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL		 16371 (arch_enabled): Handle new arch types. Remove support for onlya8
16372 and nota8.
16373 (one_cmpldi2): Use new arch names.
bc5faa5b 16374 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL		 16375 * config/arm/arm.opt (mneon-for-64bits): Add option.
16376 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
16377 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
16378 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
16379 of onlya8.
16380 * doc/invoke.texi (-mneon-for-64bits): Document.
16381
5a2d2a79
RB		 163822013-03-21 Richard Biener <rguenther@suse.de>
16383
16384 PR tree-optimization/39326
16385 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
16386 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 16387 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 16388
5abe1e05
RB		 163892013-03-21 Richard Biener <rguenther@suse.de>
16390
16391 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
16392 (vect_insert_into_interleaving_chain): Likewise.
16393 (vect_drs_dependent_in_basic_block): Inline ...
16394 (vect_slp_analyze_data_ref_dependence): ... here. New function,
16395 split out from ...
16396 (vect_analyze_data_ref_dependence): ... here. Simplify.
16397 (vect_check_interleaving): Simplify.
16398 (vect_analyze_data_ref_dependences): Likewise. Split out ...
16399 (vect_slp_analyze_data_ref_dependences): ... this new function.
16400 (dr_group_sort_cmp): New function.
16401 (vect_analyze_data_ref_accesses): Compute data-reference groups
16402 here instead of in vect_analyze_data_ref_dependence. Use
16403 a more efficient algorithm.
16404 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
16405 vect_slp_analyze_data_ref_dependences. Call
16406 vect_analyze_data_ref_accesses earlier.
16407 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
16408 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
16409 (vect_slp_analyze_data_ref_dependences): New prototype.
16410
cad1735b
RB		 164112013-03-21 Richard Biener <rguenther@suse.de>
16412
16413 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
16414 ref is stored in the loop.
16415 (find_refs_for_sm): Walk only over all stores.
16416 (store_motion_loop): Allocate from lim_bitmap_obstack.
16417 (store_motion): Likewise.
16418
141310ef
RB		 164192013-03-21 Richard Biener <rguenther@suse.de>
16420
16421 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
16422 Update virtual SSA form.
16423
5022315a
RO		 164242013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16425
16426 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
16427 * configure: Regenerate.
16428 * config.in: Regenerate.
16429 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
16430 if !HAVE_LD_EH_FRAME_CIEV3.
16431
50d4421c
RB		 164322013-03-21 Richard Biener <rguenther@suse.de>
16433
16434 * tree-cfg.c (verify_expr_no_block): New function.
16435 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
16436 nor DECL_VALUE_EXPR have locations with associated blocks.
16437 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
16438 (clear_unused_block_pointer): Remove code dealing with
16439 blocks in DECL_DEBUG_EXPR locations.
16440
839b422f
RB		 164412013-03-21 Richard Biener <rguenther@suse.de>
16442
16443 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
16444 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
16445 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
16446 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
16447 instead of DECL_DEBUG_EXPR_IS_FROM.
16448 * gimplify.c (gimplify_modify_expr): Likewise.
16449 * tree-cfg.c (verify_expr_location_1): Likewise.
16450 * tree-complex.c (create_one_component_var): Likewise.
16451 * tree-sra.c (create_access_replacement): Likewise.
16452 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
16453 (clear_unused_block_pointer): Likewise.
16454 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
16455 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
16456 * var-tracking.c (var_debug_decl): Likewise.
16457 (track_expr_p): Likewise.
16458 * tree-inline.c (add_local_variables): Likewise. Set
16459 DECL_HAS_DEBUG_EXPR_P after copying it.
16460 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
16461 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
16462
fe04878d
UB		 164632013-03-21 Uros Bizjak <ubizjak@gmail.com>
16464
16465 PR bootstrap/56656
16466 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
16467 * configure: Regenerate.
16468 * config.in: Regenerate.
16469 * config/i386/i386.md (*movdf_internal): Use
16470 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16471 movd instead of movq mnemonic for interunit moves.
16472 (*movdi_internal): Ditto.
16473
bd059b26 164742013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N		 16475
16476 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
16477 (abd<mode>_3): New pattern.
16478 (aba<mode>_3): New pattern.
16479 (fabd<mode>_3): New pattern.
16480
bd059b26 164812013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N		 16482
16483 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
16484 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
16485 occurrence of REGISTER_PREFIX as its empty string.
16486
4f1f78b9
JL		 164872013-03-20 Jeff Law <law@redhat.com>
16488
16489 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
16490 addititional equivalences for equality comparisons between an SSA_NAME
16491 and a constant where the SSA_NAME was set from a widening conversion.
16492
327a1118
WL		 164932013-03-20 Walter Lee <walt@tilera.com>
16494
16495 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
16496
813e0036
UB		 164972013-03-20 Uros Bizjak <ubizjak@gmail.com>
16498
16499 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
16500 depending on type attribute.
16501 (*movti_internal): Ditto.
16502 (*movtf_internal): Ditto.
16503 (*movxf_internal): Ditto.
16504 (*movdf_internal): Ditto.
16505 (*movsf_internal): Ditto.
16506
7cf34aae
UB		 165072013-03-20 Uros Bizjak <ubizjak@gmail.com>
16508
16509 * config/i386/i386.md (*movti_internal): Set prefix attribute to
16510 maybe_vex for sselog1 and ssemov types.
16511 (*movdi_internal): Reorder operand constraints.
16512 (*movsi_internal): Ditto. Set prefix attribute to
16513 maybe_vex for sselog1 and ssemov types.
16514 (*movtf_internal): Set prefix attribute to maybe_vex
16515 for sselog1 and ssemov types.
16516 (*movdf_internal): Ditto. Set prefix_data16 attribute for
16517 DImode ssemov types. Reorder operand constraints.
16518 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
16519 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
16520 attribute for SImode ssemov types. Reorder operand constraints.
16521
19321415
MJ		 165222013-03-20 Martin Jambor <mjambor@suse.cz>
16523
16524 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
16525 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
16526
12211b99 165272013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM		 16528
16529 * config/rs6000/predicates.md (indexed_address, update_address_mem
16530 update_indexed_address_mem): New predicates.
16531 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
16532 attribute for load/store instructions.
16533 * config/rs6000/dfp.md (movsd_store): Likewise.
16534 (movsd_load): Likewise.
16535 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
16536 (unnamed HI->DI extend define_insn): Likewise.
16537 (unnamed SI->DI extend define_insn): Likewise.
16538 (unnamed QI->SI extend define_insn): Likewise.
16539 (unnamed QI->HI extend define_insn): Likewise.
16540 (unnamed HI->SI extend define_insn): Likewise.
16541 (unnamed HI->SI extend define_insn): Likewise.
16542 (extendsfdf2_fpr): Likewise.
16543 (movsi_internal1): Likewise.
16544 (movsi_internal1_single): Likewise.
16545 (movhi_internal): Likewise.
16546 (movqi_internal): Likewise.
16547 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
16548 attribute for load/store instructions.
16549 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
16550 instructions.
16551 (mov<mode>_softfloat): Likewise.
16552 (mov<mode>_hardfloat32): Likewise.
16553 (mov<mode>_hardfloat64): Likewise.
16554 (mov<mode>_softfloat64): Likewise.
16555 (movdi_internal32): Likewise.
16556 (movdi_internal64): Likewise.
16557 (probe_stack_<mode>): Likewise.
16558
165592013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
16560
16561 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
16562 floating point, and decimal floating point to reload iterator.
16563
16564 * config/rs6000/constraints.md (wl constraint): New constraints to
16565 return FLOAT_REGS if certain options are used to reduce the number
16566 of separate patterns that exist in the file.
16567 (wx constraint): Likewise.
16568 (wz constraint): Likewise.
16569
16570 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
16571 -mdebug=reg, print wg, wl, wx, and wz constraints.
16572 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
16573 Initialize the reload functions for 64-bit binary/decimal floating
16574 point types.
16575 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
16576 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
16577 create the buffer on the stack to overcome not having a 32-bit
16578 load and store.
16579 (rs6000_emit_move): Likewise.
16580 (rs6000_secondary_memory_needed_rtx): Likewise.
16581 (rs6000_alloc_sdmode_stack_slot): Likewise.
16582 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
16583 via xxlxor, just like DFmode 0.0.
16584
16585 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
16586 define as 1 if we are running on a power7 or newer.
16587 (enum r6000_reg_class_enum): Add new constraints.
16588
16589 * config/rs6000/dfp.md (movsd): Delete, combine with binary
16590 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
16591 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 16592 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM		 16593 (movsd splitter): Likewise.
16594 (movsd_hardfloat): Likewise.
16595 (movsd_softfloat): Likewise.
16596
16597 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
16598 binary and decimal floating point moves.
16599 (fmove_ok): New attributes to combine binary and decimal floating
16600 point moves, and to combine power6x (mfpgpr) moves along normal
16601 floating moves.
16602 (real_value_to_target): Likewise.
16603 (f32_lr): Likewise.
16604 (f32_lm): Likewise.
16605 (f32_li): Likewise.
16606 (f32_sr): Likewise.
16607 (f32_sm): Likewise.
16608 (f32_si): Likewise.
16609 (movsf): Combine binary and decimal floating point moves. Combine
16610 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 16611 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM		 16612 (mov<mode> for SFmode/SDmode); Likewise.
16613 (SFmode/SDmode splitters): Likewise.
16614 (movsf_hardfloat): Likewise.
16615 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
16616 (movsf_softfloat): Likewise.
16617 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
16618
16619 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
16620 wx and wz constraints.
16621
16622 * config/rs6000/constraints.md (wg constraint): New constraint to
16623 return FLOAT_REGS if -mmfpgpr (power6x) was used.
16624
16625 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
16626 constraint.
16627
16628 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
16629 -mdebug=reg, print wg, wl, wx, and wz constraints.
16630 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
16631 Initialize the reload functions for 64-bit binary/decimal floating
16632 point types.
16633 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
16634 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
16635 create the buffer on the stack to overcome not having a 32-bit
16636 load and store.
16637 (rs6000_emit_move): Likewise.
16638 (rs6000_secondary_memory_needed_rtx): Likewise.
16639 (rs6000_alloc_sdmode_stack_slot): Likewise.
16640 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
16641 via xxlxor, just like DFmode 0.0.
16642
c6d5ff83
MM		 16643 * config/rs6000/dfp.md (movdd): Delete, combine with binary
16644 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
16645 with other moves by using conditional constraits (wg). Use LFIWZX
16646 and STFIWX for loading SDmode on power7.
16647 (movdd splitters): Likewise.
16648 (movdd_hardfloat32): Likewise.
16649 (movdd_softfloat32): Likewise.
16650 (movdd_hardfloat64_mfpgpr): Likewise.
16651 (movdd_hardfloat64): Likewise.
16652 (movdd_softfloat64): Likewise.
16653
16654 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
16655 64-bit binary and decimal floating point moves.
16656 (FMOVE64X): Likewise.
16657 (movdf): Combine 64-bit binary and decimal floating point moves.
16658 Combine power6x (mfpgpr) moves with other moves by using
16659 conditional constraits (wg).
16660 (mov<mode> for DFmode/DDmode): Likewise.
16661 (DFmode/DDmode splitters): Likewise.
16662 (movdf_hardfloat32): Likewise.
16663 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
16664 (movdf_softfloat32): Likewise.
16665 (movdf_hardfloat64_mfpgpr): Likewise.
16666 (movdf_hardfloat64): Likewise.
16667 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
16668 (movdf_softfloat64): Likewise.
16669 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
16670 (reload_<mode>_load): Move to later in the file so they aren't in
16671 the middle of the floating point move insns.
16672 (reload_<mode>_store): Likewise.
16673
16674 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
16675 constraint.
16676
16677 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
16678 constraint if -mdebug=reg.
bd059b26
UB		 16679 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
16680 Enable using dd reload support if needed.
c6d5ff83
MM		 16681
16682 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
16683 binary and decimal floating point moves in rs6000.md.
16684 (movtd_internal): Likewise.
16685
16686 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
16687 decimal floating point moves.
16688 (movtf): Likewise.
16689 (movtf_internal): Likewise.
16690 (mov<mode>_internal, TDmode/TFmode): Likewise.
16691 (movtf_softfloat): Likewise.
16692 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
16693
16694 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
16695 movdi_internal64, using wg constraint for move direct operations.
16696 (movdi_internal64): Likewise.
16697
16698 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
16699 MODES_TIEABLE_P for selected modes. Print the numerical value of
16700 the various virtual registers. Use GPR/FPR first/last values,
16701 instead of hard coding the register numbers. Print which modes
16702 have reload functions registered.
bd059b26
UB		 16703 (rs6000_option_override_internal): If -mdebug=reg, trace the options
16704 settings before/after setting cpu, target and subtarget settings.
16705 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
16706 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM		 16707 (rs6000_secondary_reload_fail): Likewise.
16708 (rs6000_secondary_reload_inner): Likewise.
16709
16710 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
16711 macros for first/last GPR and FPR registers.
16712 (LAST_GPR_REGNO): Likewise.
16713 (FIRST_FPR_REGNO): Likewise.
16714 (LAST_FPR_REGNO): Likewise.
16715
16716 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
16717 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
16718 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
16719 (vcond<mode><mode>): Likewise.
16720 (vcondu<mode><mode>): Likewise.
16721 (vector_gtu<mode>): Likewise.
16722 (vector_gte<mode>): Likewise.
16723 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 16724 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM		 16725 (ior<mode>3): Likewise.
16726 (and<mode>3): Likewise.
16727 (one_cmpl<mode>2): Likewise.
16728 (nor<mode>3): Likewise.
16729 (andc<mode>3): Likewise.
16730
16731 * config/rs6000/constraints.md (wt constraint): New constraint
16732 that returns VSX_REGS if TImode is allowed in VSX registers.
16733
16734 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
16735 constant under VSX.
16736
16737 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
16738 similar to TImode, but it is restricted to being in the GPRs.
16739
16740 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
16741 TImode to occupy a single VSX register.
16742
16743 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
16744 -mvsx-timode for power7/power8.
16745 (power7 cpu): Likewise.
16746 (power8 cpu): Likewise.
16747
16748 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
16749 sure that TFmode/TDmode take up two registers if they are ever
16750 allowed in the upper VSX registers.
16751 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
16752 registers.
16753 (rs6000_init_hard_regno_mode_ok): Likewise.
16754 (rs6000_debug_reg_global): Add debugging for PTImode and wt
16755 constraint. Print if LRA is turned on.
16756 (rs6000_option_override_internal): Give an error if -mvsx-timode
16757 and VSX is not enabled.
16758 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
16759 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
16760 to reg+offset addressing. Use PTImode when checking offset
16761 addresses for validity.
16762 (reg_offset_addressing_ok_p): Likewise.
16763 (rs6000_legitimate_offset_address_p): Likewise.
16764 (rs6000_legitimize_address): Likewise.
16765 (rs6000_legitimize_reload_address): Likewise.
16766 (rs6000_legitimate_address_p): Likewise.
16767 (rs6000_eliminate_indexed_memrefs): Likewise.
16768 (rs6000_emit_move): Likewise.
16769 (rs6000_secondary_reload): Likewise.
16770 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
16771 reloads to fpr registers to continue to use reg+offset addressing,
16772 but 64-bit reloads to altivec registers need reg+reg addressing.
16773 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
16774 it. Treat LO_SUM like a PLUS operation.
16775 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 16776 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM		 16777 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
16778 registers to share a register with a smaller sized type, since VSX
16779 puts scalars in the upper 64-bits.
16780 (print_operand): Add support for PTImode.
16781 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
16782 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
16783 registers, but don't have arithmetic support.
16784 (rs6000_memory_move_cost): Add test for VSX.
16785 (rs6000_opt_masks): Add -mvsx-timode.
16786
16787 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
16788 for TImode.
16789 (VSs): Likewise.
16790 (VSr): Use wt constraint for TImode.
16791 (VSv): Drop TImode support.
16792 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
16793 (vsx_movti_64bit): Likewise.
16794 (vsx_movti_32bit): Likewise.
16795 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
16796 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
16797 one '?' on the appropriate output constraint. Do not allow TImode
16798 logical operations on 32-bit systems.
16799 (vsx_ior<mode>3): Likewise.
16800 (vsx_xor<mode>3): Likewise.
16801 (vsx_one_cmpl<mode>2): Likewise.
16802 (vsx_nor<mode>3): Likewise.
16803 (vsx_andc<mode>3): Likewise.
16804 (vsx_concat_<mode>): Likewise.
16805 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
16806
16807 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
16808 OPTION_MASK_VSX_TIMODE.
16809 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
16810 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
16811
16812 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
16813 (TI2 iterator): New iterator for TImode, PTImode.
16814 (wd mode attribute): Add values for vector types.
bd059b26
UB		 16815 (movti_string): Replace TI move operations with operations for TImode
16816 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM		 16817 (mov<mode>_string, TImode/PTImode): Likewise.
16818 (movti_ppc64): Likewise.
16819 (mov<mode>_ppc64, TImode/PTImode): Likewise.
16820 (TI mode splitters): Likewise.
16821
16822 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
16823 constraint.
16824
1fc5eced
MG		 168252013-03-20 Marc Glisse <marc.glisse@inria.fr>
16826
16827 PR tree-optimization/56355
16828 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
16829 Also handle integers with undefined overflow.
16830
22c4c869
CM		 168312013-03-20 Catherine Moore <clm@codesourcery.com>
16832 Maciej W. Rozycki <macro@codesourcery.com>
16833 Tom de Vries <tom@codesourcery.com>
12211b99 16834 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM		 16835 Iain Sandoe <iain@codesourcery.com>
16836 Nathan Froyd <froydnj@codesourcery.com>
12211b99 16837 Chao-ying Fu <fu@mips.com>
22c4c869
CM		 16838
16839 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 16840 Document new function attributes.
22c4c869
CM		 16841 * doc/invoke.texi (minterlink-compressed, mmicromips,
16842 m14k, m14ke, m14kec): Document new options.
16843 (minterlink-mips16): Update documentation.
16844 * doc/md.texi (ZC, ZD): Document new constraints.
16845 * configure.ac (gcc_cv_as_micromips): Check if linker
16846 supports the .set micromips directive.
16847 * configure: Regenerate.
16848 * config.in: Regenerate.
16849 * config/mips/mips-tables.opt: Regenerate.
16850 * config/mips/micromips.md: New file.
16851 * constraints.md (ZC, ZD): New constraints.
16852 * config/mips/predicates.md (movep_src_register): New predicate.
16853 (movep_src_operand): New predicate.
16854 (non_volatile_mem_operand): New predicate.
16855 * config/mips/mips.md (multimem): New type.
16856 (length): Differentiate between 17-bit and 18-bit branch offsets.
16857 (MOVEP1, MOVEP2): New mode iterator.
33159866 16858 (mov_<load>l): Use ZC constraint.
22c4c869
CM		 16859 (mov_<load>r): Likewise.
16860 (mov_<store>l): Likewise.
16861 (mov_<store>r): Likewise.
16862 (*branch_equality<mode>_inverted): Add microMIPS support.
16863 (*branch_equality<mode>): Likewise.
16864 (*jump_absolute): Likewise.
16865 (indirect_jump_<mode>): Likewise.
16866 (tablejump_<mode>): Likewise.
16867 (<optab>_internal): Likewise.
16868 (sibcall_internal): Likewise.
16869 (sibcall_value_internal): Likewise.
16870 (prefetch): Use constraint ZD.
16871 * config/mips/mips.opt (minterlink-compressed): New option.
16872 (minterlink-mips16): Now an alias for minterlink-compressed.
16873 (mmicromips): New option.
16874 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16875 (compare_and_swap_12): Likewise.
16876 (sync_add<mode>): Likewise.
16877 (sync_<optab>_12): Likewise.
16878 (sync_old_<optab>_12): Likewise.
16879 (sync_new_<optab>_12): Likewise.
16880 (sync_nand_12): Likewise.
16881 (sync_old_nand_12): Likewise.
16882 (sync_new_nand_12): Likewise.
16883 (sync_sub<mode>): Likewise.
16884 (sync_old_add<mode>): Likewise.
16885 (sync_old_sub<mode>): Likewise.
16886 (sync_new_add<mode>): Likewise.
16887 (sync_new_sub<mode>): Likewise.
16888 (sync_<optab><mode>): Likewise.
16889 (sync_old_<optab><mode>): Likewise.
16890 (sync_new_<optab><mode>): Likewise.
16891 (sync_nand<mode>): Likewise.
16892 (sync_old_nand<mode>): Likewise.
16893 (sync_new_nand<mode>): Likewise.
16894 (sync_lock_test_and_set<mode>): Likewise.
16895 (test_and_set_12): Likewise.
16896 (atomic_compare_and_swap<mode>): Likewise.
16897 (atomic_exchange<mode>_llsc): Likewise.
16898 (atomic_fetch_add<mode>_llsc): Likewise.
16899 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16900 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16901 (umips_save_restore_pattern_p): Likewise.
16902 (umips_load_store_pair_p): Likewise.
16903 (umips_output_load_store_pair): Likewise.
16904 (umips_movep_target_p): Likewise.
16905 (umips_12bit_offset_address_p): Likewise.
16906 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16907 (mips_base_mips16): Rename this...
16908 (mips_base_compression_flags): ...to this. Update all uses.
16909 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16910 (mips_mips16_decl_p): Delete.
16911 (mips_nomips16_decl_p): Delete.
33159866 16912 (mips_get_compress_on_flags): New function.
22c4c869
CM		 16913 (mips_get_compress_off_flags): New function.
16914 (mips_get_compress_mode): New function.
16915 (mips_get_compress_on_name): New function.
16916 (mips_get_compress_off_name): New function.
16917 (mips_insert_attributes): Support multiple compression types.
16918 (mips_merge_decl_attributes): Likewise.
16919 (umips_12bit_offset_address_p): New function.
16920 (mips_start_function_definition): Emit .set micromips directive.
16921 (mips_call_may_need_jalx_p): New function.
16922 (mips_function_ok_for_sibcall): Add microMIPS support.
16923 (mips_print_operand_punctuation): Support short delay slots and
16924 compact jumps.
16925 (umips_swm_mask, umips_swm_encoding): New.
16926 (umips_build_save_restore): New function.
16927 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16928 (was_mips16_p): Remove.
16929 (old_compression_mode): New.
16930 (mips_set_compression_mode): New function.
16931 (mips_set_current_function): Add microMIPS support.
16932 (mips_option_override): Likewise.
16933 (umips_save_restore_pattern_p): New function.
16934 (umips_output_save_restore): New function.
16935 (umips_load_store_pair_p_1): New function.
16936 (umips_load_store_pair_p): New function.
16937 (umips_output_load_store_pair_1): New function.
16938 (umips_output_load_store_pair): New function.
16939 (umips_movep_target_p) New function.
16940 (mips_prepare_pch_save): Add microMIPS support.
16941 * config/mips/mips.h (TARGET_COMPRESSION): New.
16942 (TARGET_CPU_CPP_BUILTINS): Update macro
16943 to use new compression flags and to support microMIPS.
16944 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16945 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16946 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16947 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16948 (ASM_SPEC): Support mmicromips and mno-micromips.
16949 (M16STORE_REG_P): New macro.
16950 (MIPS_CALL): Support TARGET_MICROMIPS.
16951 (MICROMIPS_J): New macro.
16952 (mips_base_mips16): Rename this...
16953 (mips_base_compression_flags): ...to this.
16954 (UMIPS_12BIT_OFFSET_P): New macro.
16955 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16956 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB		 169572013-03-20 Richard Biener <rguenther@suse.de>
16958
16959 PR tree-optimization/56661
16960 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
16961 the result does not have to be distinct.
16962
54714c68
RB		 169632013-03-20 Richard Biener <rguenther@suse.de>
16964
16965 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
16966 remap_gimple_op_r.
16967
cca1130d 169682013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 16969 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS		 16970
16971 PR rtl-optimization/56605
16972 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
16973
2a930377
UB		 169742013-03-20 Uros Bizjak <ubizjak@gmail.com>
16975
16976 PR bootstrap/56656
16977 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
16978 that require movd instead of movq.
16979
d6d305fe
RB		 169802013-03-20 Richard Biener <rguenther@suse.de>
16981
16982 * tree-ssa-structalias.c (struct variable_info): Add pointer
16983 to the first field of an aggregate with sub-vars. Make
16984 this and the pointer to the next subfield its ID.
16985 (vi_next): New function.
16986 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
16987 storedanything_id, integer_id): Increment by one.
16988 (new_var_info, get_call_vi, lookup_call_clobber_vi,
16989 get_call_clobber_vi): Adjust.
16990 (solution_set_expand): Simplify and speedup.
16991 (solution_set_add): Inline into ...
16992 (set_union_with_increment): ... this. Adjust accordingly.
16993 (do_sd_constraint): Likewise.
16994 (do_ds_constraint): Likewise.
16995 (do_complex_constraint): Simplify.
16996 (build_pred_graph): Adjust.
16997 (solve_graph): Likewise. Simplify and speedup.
16998 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
16999 get_constraint_for_component_ref, get_constraint_for_1,
17000 first_vi_for_offset, first_or_preceding_vi_for_offset,
17001 create_function_info_for, create_variable_info_for_1,
17002 create_variable_info_for, intra_create_variable_infos): Adjust.
17003 (init_base_vars): Push NULL for ID zero.
17004 (compute_points_to_sets): Adjust.
17005
a271b42d
RB		 170062013-03-20 Richard Biener <rguenther@suse.de>
17007
17008 * cfgloop.c (verify_loop_structure): Streamline and avoid
17009 ICEing on corrupt loop tree.
17010 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17011 loop tree.
17012
ebd65954
RB		 170132013-03-20 Richard Biener <rguenther@suse.de>
17014
17015 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17016 check whether an SSA update is needed.
17017
4547b7ee
RS		 170182013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17019
17020 * config/mips/constraints.md (T): Rename to...
17021 (Yf): ...this.
17022 (U): Rename to...
17023 (Yd): ...this.
17024 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17025 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17026
0b8cdc58
IB		 170272013-03-19 Ian Bolton <ian.bolton@arm.com>
17028
17029 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17030 (*subsi3_carryin_uxtw): Likewise.
17031
5977a10d
IB		 170322013-03-19 Ian Bolton <ian.bolton@arm.com>
17033
17034 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17035 (*rorsi3_insn_uxtw): Likewise.
17036
bd83ff2c
IB		 170372013-03-19 Ian Bolton <ian.bolton@arm.com>
17038
17039 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17040 (*extrsi5_insn_uxtw): Likewise.
17041
ba65123c
RB		 170422013-03-19 Richard Biener <rguenther@suse.de>
17043
17044 PR tree-optimization/56273
17045 * passes.c (init_optimization_passes): Move second VRP after DOM.
17046
2eac0476
UB		 170472013-03-19 Uros Bizjak <ubizjak@gmail.com>
17048
17049 * config/i386/i386.md (*movti_internal): Merge from
17050 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17051 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17052 nox64 isa attributes.
17053
6e55eda7
RB		 170542013-03-18 Richard Biener <rguenther@suse.de>
17055
17056 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17057 (unite): Likewise.
17058 (merge_node_constraints): Likewise.
17059 (build_succ_graph): Likewise.
17060 (valid_graph_edge): Inline into single caller.
17061 (unify_nodes): Likewise. Use bitmap_set_bit return value
17062 and cache varinfo.
17063 (scc_visit): Fix formatting and variable use.
17064 (do_sd_constraint): Use gcc_checking_assert.
17065 (do_ds_constraint): Likewise.
17066 (do_complex_constraint): Likewise.
17067 (condense_visit): Likewise. Cleanup.
17068 (dump_pred_graph): New function.
17069 (perform_var_substitution): Dump the pred-graph before
17070 variable substitution.
17071 (find_equivalent_node): Use gcc_checking_assert.
17072 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17073
4bdd44c4
RB		 170742013-03-18 Richard Biener <rguenther@suse.de>
17075
17076 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17077 Remove cond_expr_stmt_list argument and do not gimplify the
17078 built expression.
17079 (vect_loop_versioning): Adjust.
17080 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17081 Cleanup to use less temporaries.
17082 (vect_create_data_ref_ptr): Cleanup.
17083
38c56a5b
JJ		 170842013-03-18 Jakub Jelinek <jakub@redhat.com>
17085
17086 PR tree-optimization/56635
17087 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17088 require types_compatible_p types.
17089
20b2e6a0
NC		 170902013-03-18 Nick Clifton <nickc@redhat.com>
17091
a6178a25
NC		 17092 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
17093 spurious backslash.
17094
20b2e6a0
NC		 17095 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
17096 Add missing line to comment describing function.
17097
92608d0e
RB		 170982013-03-18 Richard Biener <rguenther@suse.de>
17099
17100 PR tree-optimization/56210
17101 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17102 Handle string / character search functions.
17103 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
17104
31348d52
RB		 171052013-03-18 Richard Biener <rguenther@suse.de>
17106
17107 PR middle-end/56483
17108 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
17109 and implement properly.
17110 * gimple.h (gimple_cond_single_var_p): Remove.
17111
fcac74a1
RB		 171122013-03-18 Richard Biener <rguenther@suse.de>
17113
17114 * tree-data-ref.h (find_data_references_in_loop): Declare.
17115 * tree-data-ref.c (get_references_in_stmt): Use a stack
17116 vector pre-allocated in the callers.
17117 (find_data_references_in_stmt): Adjust.
17118 (graphite_find_data_references_in_stmt): Likewise.
17119 (create_rdg_vertices): Likewise.
17120 (find_data_references_in_loop): Export.
17121 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
17122 Compute dependences here...
17123 (vect_analyze_data_refs): ...not here. When we encounter
17124 a non-vectorizable data reference in basic-block vectorization
17125 truncate the data reference vector. Do not bother to
17126 fixup data-dependence information for gather loads.
17127 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
17128 of data references, as reported.
17129
0d5a1b56
RB		 171302013-03-18 Richard Biener <rguenther@suse.de>
17131
17132 PR tree-optimization/3713
17133 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
17134 has_constants and expr.
17135 (stmt_has_constants): Properly valueize SSA names when deciding
17136 whether the stmt has constants.
17137
789c34e3
RB		 171382013-03-18 Richard Biener <rguenther@suse.de>
17139
17140 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
17141 whole function when there is nothing to do.
17142 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
17143 * tree-vectorizer.c (vectorize_loops): Update virtual and
17144 loop-closed SSA once.
17145 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
17146
076b4605
RB		 171472013-03-18 Richard Biener <rguenther@suse.de>
17148
17149 PR middle-end/56113
17150 * domwalk.c (bb_postorder): New global static.
17151 (cmp_bb_postorder): New function.
17152 (walk_dominator_tree): Replace scheme imposing an order for
17153 visiting dominator sons by one sorting them at the time they
17154 are pushed on the stack.
17155
bdb01696
RB		 171562013-03-18 Richard Biener <rguenther@suse.de>
17157
17158 PR tree-optimization/39326
17159 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17160 (struct mem_ref): Replace mem member with ao_ref typed member.
17161 (MEM_ANALYZABLE): Adjust.
17162 (memref_eq): Likewise.
17163 (mem_ref_alloc): Likewise.
17164 (gather_mem_refs_stmt): Likewise.
17165 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17166 (execute_sm_if_changed_flag_set): Adjust.
17167 (execute_sm): Likewise.
17168 (ref_always_accessed_p): Likewise.
17169 (refs_independent_p): Likewise.
17170 (can_sm_ref_p): Likewise.
17171
12d2dc5e
JJ		 171722013-03-18 Jakub Jelinek <jakub@redhat.com>
17173
17174 PR c/56566
17175 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17176 return 1 even for !unsignedp.
17177
286fb677
UB		 171782013-03-17 Uros Bizjak <ubizjak@gmail.com>
17179
17180 * config/i386/i386.md (isa): Add x64 and nox64.
17181 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17182 (*pushtf): Enable *roF alternative for x64 isa only.
17183 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17184 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17185 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17186 (*movtf_internal): Merge from *movtf_internal_rex64 and
17187 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17188 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17189 nox64 isa attributes.
17190 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17191 nox64 isa attributes.
17192 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17193
88b97037
UB		 171942013-03-17 Uros Bizjak <ubizjak@gmail.com>
17195
17196 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17197
9f4f1735
JJ		 171982013-03-17 Jakub Jelinek <jakub@redhat.com>
17199
17200 PR target/56640
17201 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17202 class names. Remove trailing comma after "ALL_REGS".
17203
48f4a6fa
JH		 172042013-03-16 Jan Hubicka <jh@suse.cz>
17205
17206 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17207 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17208 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17209 of cgraph_get_create_node.
17210 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17211
98e81123
JM		 172122013-03-16 Jason Merrill <jason@redhat.com>
17213
17214 PR debug/49090
17215 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17216 with DW_AT_default_value.
17217
8a298c04
JJ		 172182013-03-16 Jakub Jelinek <jakub@redhat.com>
17219
17220 * BASE-VER: Set to 4.9.0.
17221
4323afa0
AK		 172222013-03-14 Andi Kleen <ak@linux.intel.com>
17223
17224 PR target/56619
17225 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17226 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17227 Document _x* TSX intrinsics.
17228
b3c0d469
JJ		 172292013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17230 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI		 17231
17232 * configure.ac: Add MicroBlaze TLS support detection.
17233 * configure: Regenerate.
88b97037
UB		 17234 * config/microblaze/microblaze-protos.h
17235 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17236 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17237 * config/microblaze/microblaze.c (microblaze_address_type): Add
17238 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI		 17239 (microblaze_address_info): Add tls_reloc.
17240 (TARGET_HAVE_TLS): Define.
17241 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17242 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB		 17243 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17244 load_tls_operand, microblaze_call_tls_get_addr,
17245 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI		 17246 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17247 (get_base_reg): Use microblaze_tls_symbol_p.
17248 (microblaze_classify_address): Handle TLS.
88b97037
UB		 17249 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17250 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI		 17251 (microblaze_legitimize_address): Handle TLS.
17252 (microblaze_address_insns): Handle ADDRESS_TLS.
17253 (pic_address_needs_scratch): Handle TLS.
17254 (print_operand_address): Handle TLS.
17255 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17256 (microblaze_expand_move): Handle TLS.
88b97037
UB		 17257 (microblaze_legitimate_constant_p): Check
17258 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI		 17259 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17260 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17261 (PIC_OFFSET_TABLE_REGNUM): Set.
17262 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17263 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17264 (addsi3, movsi_internal2, movdf_internal): Update constraints
17265 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB		 17266 (move_operand): Redefine as move_src_operand,
17267 check microblaze_tls_referenced_p.
8cc9a5a5 17268
d803a491
IB		 172692013-03-14 Ian Bolton <ian.bolton@arm.com>
17270
17271 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17272 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17273
a8504f22
IB		 172742013-03-14 Ian Bolton <ian.bolton@arm.com>
17275
17276 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17277 CC mode for AND.
17278
df2dfaea
JJ		 172792013-03-14 Jakub Jelinek <jakub@redhat.com>
17280
fbd28bc3
JJ		 17281 PR tree-optimization/53265
17282 * common.opt (Waggressive-loop-optimizations): New option.
17283 * tree-ssa-loop-niter.c: Include tree-pass.h.
17284 (do_warn_aggressive_loop_optimizations): New function.
17285 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17286 if number_of_latch_executions returned constant.
17287 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17288 early. If number_of_latch_executions returned constant, set
17289 nb_iterations_upper_bound back to it.
17290 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17291 field.
17292 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17293 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17294
df2dfaea
JJ		 17295 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17296 (MULTILIB_OSDIRNAMES): Set.
17297 * genmultilib: If defaultosdirname doesn't start with :: , set
17298 defaultosdirname2 instead, clear it and emit two . multilib_raw
17299 entries instead of just one.
17300
ee0d2b68
KK		 173012013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17302
17303 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17304 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17305 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17306 (SUBTARGET_OVERRIDE_OPTIONS): New.
17307
decc676e
OE		 173082013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17309
17310 PR target/49880
17311 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17312 (musermode): Convert to Var(TARGET_USERMODE).
17313 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17314 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17315 * config/sh/sh.c (sh_option_override): Use
17316 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17317 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17318 condition.
17319 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17320 TARGET_SH4.
17321 (udivsi3_i4_single, divsi3_i4_single): Use
17322 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17323
f4b719c7
DK		 173242013-03-13 Dave Korn <dave.korn.cygwin@....>
17325
17326 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17327 default setting.
17328
c40eced0
RB		 173292013-03-13 Richard Biener <rguenther@suse.de>
17330
17331 PR tree-optimization/56608
17332 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17333 calls when vectorizing basic-blocks.
17334
1bfa5973
JJ		 173352013-03-13 Jakub Jelinek <jakub@redhat.com>
17336
17337 PR plugins/45078
17338 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17339 tm_file.
17340
6fcf5434
JJ		 173412013-03-12 Jakub Jelinek <jakub@redhat.com>
17342
17343 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17344
4f38fa8c
JH		 173452013-03-11 Jan Hubicka <jh@suse.cz>
17346
17347 PR lto/56557
17348 * lto-streamer-out.c (output_symbol_p): Skip references from
17349 constructors of external variables.
17350
c5c5ba89
JH		 173512013-03-11 Jan Hubicka <jh@suse.cz>
17352
17353 PR middle-end/56571
17354 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
17355 from pseudos.
17356 * emit-rtl.c (verify_rtx_sharing): Likewise.
17357 (copy_insn_1): Likewise.
17358 * rtl.c (copy_rtx): Likewise.
17359
c2a939b1
GJL		 173602013-03-11 Georg-Johann Lay <avr@gjlay.de>
17361
17362 PR target/56591
17363 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
17364 output_operand_lossage message.
88b97037 17365
3c3279fb
RE		 173662013-03-11 Richard Earnshaw <rearnsha@arm.com>
17367
17368 PR target/56470
17369 * arm.c (shift_op): Validate RTL pattern on the fly.
17370 (arm_print_operand, case 'S'): Don't use shift_operator to validate
17371 the RTL.
17372
aef5ef9d
JDA		 173732013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17374
17375 PR target/56347
17376 * config/pa/pa.md (call_value): Check for calls to powf and direct to
17377 new call patterns that clobber %fr12.
17378 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
17379 split and postreload patterns.
17380 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
17381 registers %fr12 and %fr12R as call used.
17382
456610d3
SB		 173832013-03-09 Steven Bosscher <steven@gcc.gnu.org>
17384
17385 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
17386 (canon_address, record_store, replace_read, check_mem_read_rtx,
17387 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
17388 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
17389 rest_of_handle_dse): Likewise.
17390
4b1baac8
RS		 173912013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
17392
17393 PR middle-end/56524
17394 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
17395 Add base_optabs.
17396 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
17397 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
17398 (save_optabs_if_changed): Replace with...
17399 (init_tree_optimization_optabs): ...this.
17400 * optabs.c (save_optabs_if_changed): Rename to...
17401 (init_tree_optimization_optabs): ...this. Take the optimization node
17402 as argument. Do nothing if the base optabs are already correct.
17403 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
17404 to recompute optabs.
17405 * function.h (function): Remove optabs field.
17406 * function.c (invoke_set_current_function_hook): Call
17407 init_tree_optimization_optabs. Use the result to initialize
17408 this_fn_optabs.
17409
b7a78683
AH		 174102013-02-27 Aldy Hernandez <aldyh@redhat.com>
17411
17412 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
17413 if GTMA_HAS_NO_INSTRUMENTATION.
17414 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
17415 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
17416 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
17417 * gimple-pretty-print.c (dump_gimple_transaction): Handle
17418 GTMA_HAS_NO_INSTRUMENTATION.
17419
6384c29b
JJ		 174202013-03-08 Jakub Jelinek <jakub@redhat.com>
17421
17422 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
17423 libasan_preinit.o.
17424
ecd4f20a
MP		 174252013-03-08 Marek Polacek <polacek@redhat.com>
17426 Jakub Jelinek <jakub@redhat.com>
17427
17428 PR tree-optimization/56478
17429 * predict.c (is_comparison_with_loop_invariant_p): Change the
17430 type of loop_step to tree.
17431 (predict_loops): Adjust.
17432 (predict_iv_comparison): Perform the computations on double_ints.
17433
64366d35
RB		 174342013-03-08 Richard Biener <rguenther@suse.de>
17435
17436 PR tree-optimization/56570
17437 * tree-cfg.c (verify_expr_location_1): Verify locations for
17438 DECL_DEBUG_EXPR.
17439 * tree-sra.c (create_access_replacement): Strip locations
17440 from DECL_DEBUG_EXPRs.
17441
a9d5a059
RB		 174422013-03-08 Richard Biener <rguenther@suse.de>
17443
17444 * tree-inline.c (expand_call_inline): Do not associate
17445 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
17446 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
17447
b7aa4e9a
RB		 174482013-03-08 Richard Biener <rguenther@suse.de>
17449
17450 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
17451 or block changes with -Og. Fix for location / block encoding
17452 changes and PHI arguments with locations.
17453
c4c2f9fa
SB		 174542013-03-07 Steven Bosscher <steven@gcc.gnu.org>
17455
17456 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
17457 for all counters.
17458 (struct output_info): Likewise.
17459 (register_overhead): Remove bad gcc_assert.
17460 (bitmap_find_bit): If there is only a single bitmap element, do not
17461 count a miss as a search.
17462 (print_statistics): Update for counter type changes.
17463 (dump_bitmap_statistics): Likewise. Print headers such that they
17464 are properly lined up with the printed counters.
17465
5bf6606a
JJ		 174662013-03-07 Jakub Jelinek <jakub@redhat.com>
17467
17468 PR tree-optimization/56559
17469 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
17470 check that it has only a single use.
17471
2c653d46
RB		 174722013-03-07 Richard Biener <rguenther@suse.de>
17473
17474 * doc/invoke.texi (fwhole-program): Discourage use in combination
17475 with -flto.
17476
a72d8780
JJ		 174772013-03-06 Jakub Jelinek <jakub@redhat.com>
17478
01a454df
JJ		 17479 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
17480
c1781047
JJ		 17481 PR tree-optimization/56539
17482 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
17483 instead of GSI_CONTINUE_LINKING as last argument to
17484 force_gimple_operand_gsi. Adjust function comment.
17485
9772c47a
JJ		 17486 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
17487 aarch64-cores.def.
17488
a72d8780
JJ		 17489 PR middle-end/56548
17490 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
17491 promoted mode, convert the result back to the original mode.
17492
fa5556de
RB		 174932013-03-06 Richard Biener <rguenther@suse.de>
17494
17495 PR middle-end/56294
17496 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
17497 (insert_updated_phi_nodes_compare_uids): New function.
17498 (update_ssa): Sort symbols_to_rename after UID before
17499 traversing it to insert PHI nodes.
17500
010403d1
RB		 175012013-03-06 Richard Biener <rguenther@suse.de>
17502
17503 PR middle-end/50494
17504 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
17505 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
17506
17507 Revert
17508 2013-02-13 Richard Biener <rguenther@suse.de>
17509
17510 PR lto/50494
17511 * varasm.c (output_constant_def_1): Get the decl representing
17512 the constant as argument.
17513 (output_constant_def): Wrap output_constant_def_1.
17514 (make_decl_rtl): Use output_constant_def_1 with the decl
17515 representing the constant.
17516 (build_constant_desc): Optionally re-use a decl already
17517 representing the constant.
17518 (tree_output_constant_def): Adjust.
17519
3c27ce4c
JY		 175202013-03-06 Joey Ye <joey.ye@arm.com>
17521
17522 PR lto/50293
17523 * gcc.c (convert_white_space): New function.
17524 (main): Handles white space in function name.
17525
8f6d1c86
OE		 175262013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
17527
17528 PR target/56529
17529 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
17530 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
17531 to SH_DIV_CALL_TABLE for TARGET_SH2.
17532 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
17533 list.
17534 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
17535 call-table options.
17536
798a209f
SA		 175372013-03-05 Sterling Augustine <saugustine@google.com>
17538 Cary Coutant <ccoutant@google.com>
17539
17540 PR debug/55364
17541 * dwarf2out.c (resolve_addr): Don't call
17542 remove_loc_list_addr_table_entries a second time for the same
17543 expression.
17544
6cfa417f 175452013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 17546
6cfa417f
JJ		 17547 PR debug/56510
17548 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
17549 (avoid_complex_debug_insns): New function.
17550 (expand_debug_locations): Call it.
17551
17552 PR rtl-optimization/56484
17553 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
17554 lifetimes of hard registers on small register class machines.
17555
175562013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
17557
17558 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 17559 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 17560 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH		 17561 fast_interrupt.
17562 (microblaze_fast_interrupt_function_p): New function.
17563 (microblaze_is_interrupt_handler): Rename to
17564 microblaze_is_interrupt_variant and add fast_interrupt check.
17565 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
17566 (save_restore_insns): Likewise.
17567 (compute_frame_size): Likewise.
17568 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
17569 (microblaze_globalize_label): Likewise.
6cfa417f
JJ		 17570 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
17571 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH		 17572 microblaze_is_interrupt_variant.
17573
731aefac
KT		 175742013-03-05 Kai Tietz <ktietz@redhat.com>
17575
17576 * sdbout.c (sdbout_one_type): Switch to current function's section
17577 supporting cold/hot.
17578
a72d8780 175792013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH		 17580
17581 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
17582 -mxl-reorder.
17583
334e71e8
JJ		 175842013-03-05 Jakub Jelinek <jakub@redhat.com>
17585
0b50e654
JJ		 17586 PR middle-end/56461
17587 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
17588 if VALGRIND_GET_VBITS is defined, temporarily make object
17589 memory all defined, and restore previous valgrind addressability
17590 and definability afterwards. Free this_object at the end.
17591
4ccf8f43
JJ		 17592 PR middle-end/56461
17593 * lra.c (lra): Call lra_clear_live_ranges if live_p,
17594 right before calling lra_create_live_ranges, also call it
17595 when clearing live_p. Only call lra_clear_live_ranges
17596 at the end if live_p.
17597
334e71e8
JJ		 17598 PR middle-end/56461
17599 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
17600
9ca966ca
RB		 176012013-03-05 Richard Biener <rguenther@suse.de>
17602
17603 PR tree-optimization/56521
17604 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
17605 value-id.
17606
d273b176
SB		 176072013-03-05 Steven Bosscher <steven@gcc.gnu.org>
17608
17609 PR c++/55135
17610 * except.h (remove_unreachable_eh_regions): New prototype.
17611 * except.c (remove_eh_handler_splicer): New function, split out
17612 of remove_eh_handler.
17613 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
17614 warning about running it on many EH regions one at a time.
17615 (remove_unreachable_eh_regions_worker): New function, walk the
17616 EH tree in depth-first order and remove non-marked regions.
17617 (remove_unreachable_eh_regions): New function.
17618 * tree-eh.c (mark_reachable_handlers): New function, split out
17619 from remove_unreachable_handlers.
17620 (remove_unreachable_handlers): Use mark_reachable_handlers and
17621 remove_unreachable_eh_regions.
17622 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
17623 and remove_unreachable_eh_regions.
17624
5e7f1aef
RB		 176252013-03-05 Richard Biener <rguenther@suse.de>
17626
17627 PR middle-end/56525
17628 * loop-init.c (fix_loop_structure): Remove loops in two stages,
17629 not freeing them until the end.
17630
f276b762
AK		 176312013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17632
17633 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
17634
12a43ab8
RB		 176352013-03-05 Richard Biener <rguenther@suse.de>
17636
17637 PR tree-optimization/56270
17638 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
17639 of loads after scheduling an SLP instance.
17640
93675444
JJ		 176412013-03-05 Jakub Jelinek <jakub@redhat.com>
17642
db4138e3
JJ		 17643 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
17644 tic6x.exp.
17645 (check_gcc_parallelize): Run guality.exp as a separate job from
17646 vect.exp with unsorted.exp and $(dg_target_exps) separately from
17647 struct-layout-1.exp with stackalign.exp.
17648
dd3d1ec0
JJ		 17649 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
17650
b8d381a3
JJ		 17651 PR middle-end/56461
17652 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
17653 load_index sbitmap even if some bit in it isn't set.
17654
b4f9786b
JJ		 17655 PR middle-end/56461
17656 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
17657 (discover_iteration_bound_by_body_walk): Change queues to
17658 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
17659 spelling in comment. Call safe_push on queues[bound_index] directly.
17660 Release queues[queue_index] in every iteration unconditionally.
17661 Release bounds vector.
17662
93675444
JJ		 17663 PR middle-end/56461
17664 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
17665 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
17666 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
17667 inner_phis vector.
17668
3e492e9c
RB		 176692013-03-05 Richard Biener <rguenther@suse.de>
17670
17671 PR lto/56515
17672 * tree-inline.c (remap_blocks_to_null): New function.
17673 (expand_call_inline): When expanding a call stmt without
17674 an associated block inline remap all callee blocks to NULL.
17675
a5d1569a
JJ		 176762013-03-05 Jakub Jelinek <jakub@redhat.com>
17677
86efb5cd
JJ		 17678 PR rtl-optimization/56494
17679 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
17680 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
17681 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
17682
85f5dbea
JJ		 17683 PR middle-end/56461
17684 * sel-sched-ir.c (free_sched_pools): Release
17685 succs_info_pool.stack[succs_info_pool.max_top] vectors too
17686 if succs_info_pool.max_top isn't -1.
17687
a5d1569a
JJ		 17688 PR bootstrap/56509
17689 * opts.c (opts_obstack, opts_concat): Moved to...
17690 * opts-common.c (opts_obstack, opts_concat): ... here.
17691
4432aa6c
JJ		 176922013-03-04 Jakub Jelinek <jakub@redhat.com>
17693
17694 PR middle-end/56461
17695 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
17696
f3a1fb91
MJ		 176972013-03-04 Martin Jambor <mjambor@suse.cz>
17698
17699 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
17700 all appropriate places.
17701
22110e6c
EB		 177022013-01-04 Eric Botcazou <ebotcazou@adacore.com>
17703
17704 PR tree-optimization/56424
17705 * ipa-split.c (split_function): Do not set the RSO flag if result is
17706 not by reference and its type is a register type.
17707
a72d8780 177082013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 17709
88b97037 17710 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 17711 (microblaze_legitimate_pic_operand): Likewise
88b97037 17712 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH		 17713 new function microblaze_legitimate_pic_operand
17714 * config/microblaze/microblaze-protos.h
17715 (microblaze_legitimate_pic_operand): Declare.
17716
a72d8780 177172013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 17718
a72d8780 17719 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 17720 New predicate for supported rtx code types.
a72d8780 17721 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI		 17722 call_insn_simple_operand predicate.
17723
541d9ac8
JJ		 177242013-03-04 Jakub Jelinek <jakub@redhat.com>
17725
5eb010bc
JJ		 17726 PR middle-end/56461
17727 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
17728 partitions.ordered_remove.
17729
30862efc
JJ		 17730 PR middle-end/56461
17731 * tree-vect-stmts.c (vectorizable_conversion): Don't call
17732 vec_oprnds0.create (1) for modifier == NONE.
17733
8930f723
JJ		 17734 PR middle-end/56461
17735 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
17736 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
17737 vec_oprnds1 right before pushing anything to it for
17738 scalar_shift_arg.
17739
541d9ac8
JJ		 17740 PR middle-end/56461
17741 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
17742 set nbbs to 0 instead of having separate code path.
17743 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
17744 instead of false as last argument if returning NULL.
17745
7aa7f2e3
SL		 177462013-03-03 Sandra Loosemore <sandra@codesourcery.com>
17747
17748 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
17749 the attribute is now called "target" instead of "option".
17750 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
17751 * doc/tm.texi.in (Target Attributes): Likewise document the correct
17752 attribute/pragma name for TARGET_OPTION_VALID_P and
17753 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
17754 * doc/tm.texi: Regenerated.
17755
8930f723 177562013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 17757
8930f723 17758 * config/microblaze/microblaze.c:
cb8a1637 17759 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ		 17760 * config/microblaze/microblaze.h: Add -mxl-reorder to
17761 DRIVER_SELF_SPECS.
17762 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
17763 instructions emitted if TARGET_REORDER.
88b97037
UB		 17764 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
17765 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 17766 separately.
cb8a1637 17767
4dc1d68c
XDL		 177682013-03-01 Xinliang David Li <davidxl@google.com>
17769
17770 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
17771 walk length.
17772
689eaba3
JJ		 177732013-03-01 Jakub Jelinek <jakub@redhat.com>
17774
9d676bf6
JJ		 17775 PR middle-end/56461
17776 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
17777 vector even when returning true. Fix up function comment formatting.
17778
caff7edf
JJ		 17779 PR middle-end/56461
17780 * ira-build.c (ira_loop_nodes_count): New variable.
17781 (create_loop_tree_nodes): Initialize it.
17782 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
17783
b6b9227d
JJ		 17784 PR middle-end/56461
17785 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
17786 method on dr_chain and result_chain.
17787 * tree-vect-stmts.c (vectorizable_store): Only call
17788 result_chain.create if j == 0.
17789
689eaba3
JJ		 17790 PR middle-end/56461
17791 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
17792 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
17793 before overwriting it.
17794
40bf31ed
TB		 177952013-03-01 Tobias Burnus <burnus@net-b.de>
17796
17797 * doc/extended.texi (C Extensions): Change order in @menu
17798 to match @node.
17799 (Other MIPS Built-in Functions): Move last MIPS entry before
17800 "picoChip Built-in Functions".
17801 (SH Built-in Functions): Move after RX Built-in Functions.
17802 * doc/gcc.texi (Introduction): Change order in @menu
17803 to match @node.
17804 * doc/md.texi (Constraints): Ditto.
17805 * gty.texi (Type Information): Ditto.
17806 (User-provided marking routines for template types): Make
17807 subsection.
17808 * doc/invoke.texi (AArch64 Options): Move before
17809 "Adapteva Epiphany Options".
17810
e664c61c
KS		 178112013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
17812 Jakub Jelinek <jakub@redhat.com>
17813
17814 PR sanitizer/56454
17815 * asan.c (gate_asan): Lookup no_sanitize_address instead of
17816 no_address_safety_analysis attribute.
17817 * doc/extend.texi (no_address_safety_attribute): Rename to
17818 no_sanitize_address attribute, mention no_address_safety_analysis
17819 attribute as deprecated alias.
17820
37b5ec8f
JJ		 178212013-02-28 Jakub Jelinek <jakub@redhat.com>
17822
17823 PR middle-end/56461
17824 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
17825 type to vec<vec<tree> > *.
17826 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
17827 to be vec<tree> instead of vec<tree> *, set vec_defs
17828 to vNULL and call vec_defs.create (number_of_vects), adjust other
17829 uses of vec_defs.
17830 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
17831 vectorizable_condition): Adjust vect_get_slp_defs callers.
17832
ba96cdfb
JG		 178332013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17834
17835 * config/aarch64/aarch64.c
17836 (aarch64_float_const_representable): Remove unused variable.
17837
6f549691
JG		 178382013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17839
17840 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
17841
af55e82d
JG		 178422013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17843
17844 * config/aarch64/aarch64-builtins.c
17845 (aarch64_init_simd_builtins): Make static.
17846
1df3f464
JG		 178472013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17848
17849 * config/aarch64/aarch64.c
17850 (aarch64_simd_make_constant): Make static.
17851
f8f42513
MJ		 178522013-02-28 Martin Jambor <mjambor@suse.cz>
17853
17854 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
17855 with no initialization to the RHS of debug statements.
17856
b48b3fc4
MJ		 178572013-02-28 Martin Jambor <mjambor@suse.cz>
17858
17859 PR tree-optimization/56294
17860 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
17861 Adjust dumping.
17862 (get_access_replacement): Do not call create_access_replacement.
17863 Assert a replacement exists.
17864 (get_repl_default_def_ssa_name): Create the replacement declaration
17865 itself.
17866
c3ae224c
RR		 178672013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17868
17869 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
17870 final_end_function.
17871
45fa0eef
MP		 178722013-02-28 Marek Polacek <polacek@redhat.com>
17873
17874 PR rtl-optimization/56466
17875 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
17876 if we're changing a loop.
17877 (peel_loops_completely): Likewise.
17878
502c067d
PC		 178792013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
17880
17881 PR c++/55813
17882 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
17883
ff24afc8
GJL		 178842013-02-28 Georg-Johann Lay <avr@gjlay.de>
17885
17886 PR target/56445
17887 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
17888 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
17889 INTX_FTYPE_FX, FX_FTYPE_INTX.
17890 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
17891
c7b1fc1b
GJL		 178922013-02-28 Georg-Johann Lay <avr@gjlay.de>
17893
17894 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
17895 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
17896 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
17897 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
17898 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
17899 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
17900 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
17901 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
17902 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
17903 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
17904 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
17905 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
17906 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
17907 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
17908 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
17909 (avrxmega6): Increase max flash segments from 5 to 6.
17910 * config/avr/t-multilib: Regenerate.
17911 * config/avr/avr-tables.opt: Regenerate.
17912 * doc/avr-mmcu.texi: Regenerate.
17913
4a0e3cfe
GJL		 179142013-02-28 Georg-Johann Lay <avr@gjlay.de>
17915
17916 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
17917 (avr_device_to_arch): Rename to avr_device_to_ld.
17918 (avr_device_to_as): New prototype.
17919 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
17920 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
17921 * config/avr/driver-avr.c (avr_device_to_as): New.
17922 (avr_device_to_arch): Rename to avr_device_to_ld.
17923
97785e52
JJ		 179242013-02-27 Jakub Jelinek <jakub@redhat.com>
17925
3f292312
JJ		 17926 PR middle-end/56461
17927 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
17928 method on dr_chain and result_chain.
17929
a344216b
JJ		 17930 PR middle-end/56461
17931 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
17932 pointer_set_destroy on not_executed_last_iteration.
17933
f121ad02 17934 PR middle-end/56461
88b97037 17935 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 17936
307f83a3
JJ		 17937 PR middle-end/56461
17938 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
17939 FOR_EACH_DEFINED_FUNCTION when freeing state.
17940
e19624ee
JJ		 17941 PR middle-end/56461
17942 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
17943 pool_free.
17944 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
17945 overwriting it.
17946
90e709fd
JJ		 17947 PR middle-end/56461
17948 * ipa-cp.c (decide_whether_version_node): Call vec_free on
17949 known_aggs[i].items and release known_aggs vector.
17950
97785e52
JJ		 17951 PR middle-end/56461
17952 * ipa-reference.c (propagate): Free node_info even for alias nodes.
17953
179542013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 17955
97785e52
JJ		 17956 * config/microblaze/microblaze.c (microblaze_emit_compare):
17957 Use xor for EQ/NE comparisions.
17958 * config/microblaze/microblaze.md (cstoresf4): Add constraints
17959 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 17960
6fa5e0ed
JJ		 179612013-02-27 Jakub Jelinek <jakub@redhat.com>
17962
17963 PR middle-end/56461
17964 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
17965 vector.
17966 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
17967 vec_safe_push, always update *slot.
17968 (redirect_edge_var_map_clear): Use vec_free.
17969 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
17970 (free_var_map_entry): Use vec_free.
17971 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
17972 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
17973
436a956a
AB		 179742013-02-27 Andrey Belevantsev <abel@ispras.ru>
17975
17976 PR middle-end/45472
436a956a
AB		 17977 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
17978 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB		 17979 Reorder tests for speculativeness in the logical and operator.
17980
f5c2caca
JJ		 179812013-02-27 Jakub Jelinek <jakub@redhat.com>
17982
0fcb564b
JJ		 17983 * incpath.c (add_standard_paths): Use reconcat instead of concat
17984 where appropriate and avoid leaking memory.
17985
dc357798
JJ		 17986 * opts.h: Include obstack.h.
17987 (opts_concat): New prototype.
17988 (opts_obstack): New declaration.
17989 * opts.c (opts_concat): New function.
17990 (opts_obstack): New variable.
17991 (init_options_struct): Call gcc_init_obstack on opts_obstack.
17992 (finish_options): Use opts_concat instead of concat
17993 and XOBNEWVEC instead of XNEWVEC.
17994 * opts-common.c (generate_canonical_option, decode_cmdline_option,
17995 generate_option): Likewise.
17996 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
17997 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
17998
f5c2caca
JJ		 17999 PR target/56455
18000 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18001 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18002
d0163673
JJ		 180032013-02-26 Jakub Jelinek <jakub@redhat.com>
18004
18005 PR middle-end/56461
18006 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18007
b5ad2b8e
JR		 180082013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18009
18010 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18011 (arm_block_move_unaligned_straight): Likewise.
18012 (arm_adjust_block_mem): Likewise.
18013
9b639e2c
JR		 180142013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18015
18016 PR target/48901
18017 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18018 temp, cond and label.
18019 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18020
e13a0ccb
JR		 18021 PR target/52500
18022 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18023 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18024
fbe4f171
JR		 18025 PR target/52501
18026 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18027 of prologue/epilogue functions.
18028
ae006543
JR		 18029 PR target/52550
18030 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18031 Remove unused variable cfa_offset.
18032 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18033
c81369fa
JR		 18034 PR target/54639
18035 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18036 type promotion to unsigned.
18037
f8a8fea7
JR		 18038 PR target/54640
18039 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18040 for HOST_WIDE_INT of 32 bit / same size as int.
18041 (arm_block_move_unaligned_straight): Likewise.
18042 (arm_adjust_block_mem): Likewise.
18043
f8be5169
JR		 18044 PR target/54662
18045 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18046 ALL_CFLAGS.
18047
f1ad3354
MP		 180482013-02-26 Marek Polacek <polacek@redhat.com>
18049
18050 PR tree-optimization/56426
88b97037 18051 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 18052
a6af0f42
RB		 180532013-02-26 Richard Biener <rguenther@suse.de>
18054
18055 PR target/56444
18056 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18057 unused variable loops.
18058
aca43c6c
JJ		 180592013-02-26 Jakub Jelinek <jakub@redhat.com>
18060
cecbe5d9
JJ		 18061 PR tree-optimization/56448
18062 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18063 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18064 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18065 later operands of the references, or even first operand for
18066 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18067
aca43c6c
JJ		 18068 PR tree-optimization/56443
18069 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18070 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18071 to type_for_mode langhook.
18072
53e2e141
MT		 180732013-02-25 Matt Turner <mattst88@gmail.com>
18074
18075 * doc/invoke.texi: Document r4700.
18076
259ee451
RB		 180772013-02-25 Richard Biener <rguenther@suse.de>
18078
18079 PR tree-optimization/56175
18080 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18081 split out from ...
18082 (simplify_bitwise_binary): ... here. Also guard the conversion
18083 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18084
6d65e8f1
CM		 180852013-02-25 Catherine Moore <clm@codesourcery.com>
18086
18087 Revert:
18088 2013-02-24 Catherine Moore <clm@codesourcery.com>
18089 Maciej W. Rozycki <macro@codesourcery.com>
18090 Tom de Vries <tom@codesourcery.com>
a72d8780 18091 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM		 18092 Iain Sandoe <iain@codesourcery.com>
18093 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 18094 Chao-ying Fu <fu@mips.com>
88b97037 18095
6d65e8f1 18096 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 18097 Document new function attributes.
6d65e8f1
CM		 18098 * doc/invoke.texi (minterlink-compressed, mmicromips,
18099 m14k, m14ke, m14kec): Document new options.
18100 (minterlink-mips16): Update documentation.
18101 * doc/md.texi (ZC, ZD): Document new constraints.
18102 * configure.ac (gcc_cv_as_micromips): Check if linker
18103 supports the .set micromips directive.
18104 * configure: Regenerate.
18105 * config.in: Regenerate.
18106 * config/mips/mips-tables.opt: Regenerate.
18107 * config/mips/micromips.md: New file.
18108 * constraints.md (ZC, AD): New constraints.
18109 * config/mips/predicates.md (movep_src_register): New predicate.
18110 (movep_src_operand): New predicate.
18111 (non_volatile_mem_operand): New predicate.
18112 * config/mips/mips.md (multimem): New type.
18113 (length): Differentiate between 17-bit and 18-bit branch offsets.
18114 (MOVEP1, MOVEP2): New mode iterator.
18115 (mov_<load>l): Use ZC constraint.
18116 (mov_<load>r): Likewise.
18117 (mov_<store>l): Likewise.
18118 (mov_<store>r): Likewise.
18119 (*branch_equality<mode>_inverted): Add microMIPS support.
18120 (*branch_equality<mode>): Likewise.
18121 (*jump_absolute): Likewise.
18122 (indirect_jump_<mode>): Likewise.
18123 (tablejump_<mode>): Likewise.
18124 (<optab>_internal): Likewise.
18125 (sibcall_internal): Likewise.
18126 (sibcall_value_internal): Likewise.
18127 (prefetch): Use constraint ZD.
18128 * config/mips/mips.opt (minterlink-compressed): New option.
18129 (minterlink-mips16): Now an alias for minterlink-compressed.
18130 (mmicromips): New option.
18131 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18132 (compare_and_swap_12): Likewise.
18133 (sync_add<mode>): Likewise.
18134 (sync_<optab>_12): Likewise.
18135 (sync_old_<optab>_12): Likewise.
18136 (sync_new_<optab>_12): Likewise.
18137 (sync_nand_12): Likewise.
18138 (sync_old_nand_12): Likewise.
18139 (sync_new_nand_12): Likewise.
18140 (sync_sub<mode>): Likewise.
18141 (sync_old_add<mode>): Likewise.
18142 (sync_old_sub<mode>): Likewise.
18143 (sync_new_add<mode>): Likewise.
18144 (sync_new_sub<mode>): Likewise.
18145 (sync_<optab><mode>): Likewise.
18146 (sync_old_<optab><mode>): Likewise.
18147 (sync_new_<optab><mode>): Likewise.
18148 (sync_nand<mode>): Likewise.
18149 (sync_old_nand<mode>): Likewise.
18150 (sync_new_nand<mode>): Likewise.
18151 (sync_lock_test_and_set<mode>): Likewise.
18152 (test_and_set_12): Likewise.
18153 (atomic_compare_and_swap<mode>): Likewise.
18154 (atomic_exchange<mode>_llsc): Likewise.
18155 (atomic_fetch_add<mode>_llsc): Likewise.
18156 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18157 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18158 (umips_save_restore_pattern_p): Likewise.
18159 (umips_load_store_pair_p): Likewise.
18160 (umips_output_load_store_pair): Likewise.
18161 (umips_movep_target_p): Likewise.
18162 (umips_12bit_offset_address_p): Likewise.
18163 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18164 (mips_base_mips16): Rename this...
18165 (mips_base_compression_flags): ...to this. Update all uses.
18166 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18167 (mips_mips16_decl_p): Delete.
18168 (mips_nomips16_decl_p): Delete.
18169 (mips_get_compress_on_flags): New function.
18170 (mips_get_compress_off_flags): New function.
18171 (mips_get_compress_mode): New function.
18172 (mips_get_compress_on_name): New function.
18173 (mips_get_compress_off_name): New function.
18174 (mips_insert_attributes): Support multiple compression types.
18175 (mips_merge_decl_attributes): Likewise.
18176 (umips_12bit_offset_address_p): New function.
18177 (mips_start_function_definition): Emit .set micromips directive.
18178 (mips_call_may_need_jalx_p): New function.
18179 (mips_function_ok_for_sibcall): Add microMIPS support.
18180 (mips_print_operand_punctuation): Support short delay slots and
18181 compact jumps.
18182 (umips_swm_mask, umips_swm_encoding): New.
18183 (umips_build_save_restore): New function.
18184 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18185 (was_mips16_p): Remove.
18186 (old_compression_mode): New.
18187 (mips_set_compression_mode): New function.
18188 (mips_set_current_function): Add microMIPS support.
18189 (mips_option_override): Likewise.
18190 (umips_save_restore_pattern_p): New function.
18191 (umips_output_save_restore): New function.
18192 (umips_load_store_pair_p_1): New function.
18193 (umips_load_store_pair_p): New function.
18194 (umips_output_load_store_pair_1): New function.
18195 (umips_output_load_store_pair): New function.
18196 (umips_movep_target_p) New function.
18197 (mips_prepare_pch_save): Add microMIPS support.
18198 * config/mips/mips.h (TARGET_COMPRESSION): New.
18199 (TARGET_CPU_CPP_BUILTINS): Update macro
18200 to use new compression flags and to support microMIPS.
18201 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18202 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18203 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18204 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18205 (ASM_SPEC): Support mmicromips and mno-micromips.
18206 (M16STORE_REG_P): New macro.
18207 (MIPS_CALL): Support TARGET_MICROMIPS.
18208 (MICROMIPS_J): New macro.
18209 (mips_base_mips16): Rename this...
18210 (mips_base_compression_flags): ...to this.
18211 (UMIPS_12BIT_OFFSET_P): New macro.
18212 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18213 (MULTILIB_DIRNAMES): Likewise.
18214
5e5df392
TV		 182152013-02-25 Tom de Vries <tom@codesourcery.com>
18216
18217 PR rtl-optimization/56131
18218 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18219 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18220 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18221
3fdb53c1
TB		 182222013-02-25 Tobias Burnus <burnus@net-b.de>
18223
18224 * doc/invoke.texi (-fsanitize=): Move from optimization
18225 to debugging options.
18226
ed358aea
AB		 182272013-02-25 Andrey Belevantsev <abel@ispras.ru>
18228
18229 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18230
f45e9053
AB		 182312013-02-25 Andrey Belevantsev <abel@ispras.ru>
18232 Alexander Monakov <amonakov@ispras.ru>
18233
18234 PR middle-end/56077
18235 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18236 flush pending lists also on non-jumps. Adjust comment.
18237
6941b508
CM		 182382013-02-24 Catherine Moore <clm@codesourcery.com>
18239 Maciej W. Rozycki <macro@codesourcery.com>
18240 Tom de Vries <tom@codesourcery.com>
a72d8780 18241 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM		 18242 Iain Sandoe <iain@codesourcery.com>
18243 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 18244 Chao-ying Fu <fu@mips.com>
6941b508
CM		 18245
18246 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 18247 Document new function attributes.
6941b508
CM		 18248 * doc/invoke.texi (minterlink-compressed, mmicromips,
18249 m14k, m14ke, m14kec): Document new options.
18250 (minterlink-mips16): Update documentation.
18251 * doc/md.texi (ZC, ZD): Document new constraints.
18252 * configure.ac (gcc_cv_as_micromips): Check if linker
18253 supports the .set micromips directive.
18254 * configure: Regenerate.
18255 * config.in: Regenerate.
18256 * config/mips/mips-tables.opt: Regenerate.
18257 * config/mips/micromips.md: New file.
18258 * constraints.md (ZC, AD): New constraints.
18259 * config/mips/predicates.md (movep_src_register): New predicate.
18260 (movep_src_operand): New predicate.
18261 (non_volatile_mem_operand): New predicate.
18262 * config/mips/mips.md (multimem): New type.
18263 (length): Differentiate between 17-bit and 18-bit branch offsets.
18264 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 18265 (mov_<load>l): Use ZC constraint.
6941b508
CM		 18266 (mov_<load>r): Likewise.
18267 (mov_<store>l): Likewise.
18268 (mov_<store>r): Likewise.
18269 (*branch_equality<mode>_inverted): Add microMIPS support.
18270 (*branch_equality<mode>): Likewise.
18271 (*jump_absolute): Likewise.
18272 (indirect_jump_<mode>): Likewise.
18273 (tablejump_<mode>): Likewise.
18274 (<optab>_internal): Likewise.
18275 (sibcall_internal): Likewise.
18276 (sibcall_value_internal): Likewise.
18277 (prefetch): Use constraint ZD.
18278 * config/mips/mips.opt (minterlink-compressed): New option.
18279 (minterlink-mips16): Now an alias for minterlink-compressed.
18280 (mmicromips): New option.
18281 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18282 (compare_and_swap_12): Likewise.
18283 (sync_add<mode>): Likewise.
18284 (sync_<optab>_12): Likewise.
18285 (sync_old_<optab>_12): Likewise.
18286 (sync_new_<optab>_12): Likewise.
18287 (sync_nand_12): Likewise.
18288 (sync_old_nand_12): Likewise.
18289 (sync_new_nand_12): Likewise.
18290 (sync_sub<mode>): Likewise.
18291 (sync_old_add<mode>): Likewise.
18292 (sync_old_sub<mode>): Likewise.
18293 (sync_new_add<mode>): Likewise.
18294 (sync_new_sub<mode>): Likewise.
18295 (sync_<optab><mode>): Likewise.
18296 (sync_old_<optab><mode>): Likewise.
18297 (sync_new_<optab><mode>): Likewise.
18298 (sync_nand<mode>): Likewise.
18299 (sync_old_nand<mode>): Likewise.
18300 (sync_new_nand<mode>): Likewise.
18301 (sync_lock_test_and_set<mode>): Likewise.
18302 (test_and_set_12): Likewise.
18303 (atomic_compare_and_swap<mode>): Likewise.
18304 (atomic_exchange<mode>_llsc): Likewise.
18305 (atomic_fetch_add<mode>_llsc): Likewise.
18306 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18307 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18308 (umips_save_restore_pattern_p): Likewise.
18309 (umips_load_store_pair_p): Likewise.
18310 (umips_output_load_store_pair): Likewise.
18311 (umips_movep_target_p): Likewise.
18312 (umips_12bit_offset_address_p): Likewise.
18313 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18314 (mips_base_mips16): Rename this...
18315 (mips_base_compression_flags): ...to this. Update all uses.
18316 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18317 (mips_mips16_decl_p): Delete.
18318 (mips_nomips16_decl_p): Delete.
3fdb53c1 18319 (mips_get_compress_on_flags): New function.
6941b508
CM		 18320 (mips_get_compress_off_flags): New function.
18321 (mips_get_compress_mode): New function.
18322 (mips_get_compress_on_name): New function.
18323 (mips_get_compress_off_name): New function.
18324 (mips_insert_attributes): Support multiple compression types.
18325 (mips_merge_decl_attributes): Likewise.
18326 (umips_12bit_offset_address_p): New function.
18327 (mips_start_function_definition): Emit .set micromips directive.
18328 (mips_call_may_need_jalx_p): New function.
18329 (mips_function_ok_for_sibcall): Add microMIPS support.
18330 (mips_print_operand_punctuation): Support short delay slots and
18331 compact jumps.
18332 (umips_swm_mask, umips_swm_encoding): New.
18333 (umips_build_save_restore): New function.
18334 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18335 (was_mips16_p): Remove.
18336 (old_compression_mode): New.
18337 (mips_set_compression_mode): New function.
18338 (mips_set_current_function): Add microMIPS support.
18339 (mips_option_override): Likewise.
18340 (umips_save_restore_pattern_p): New function.
18341 (umips_output_save_restore): New function.
18342 (umips_load_store_pair_p_1): New function.
18343 (umips_load_store_pair_p): New function.
18344 (umips_output_load_store_pair_1): New function.
18345 (umips_output_load_store_pair): New function.
18346 (umips_movep_target_p) New function.
18347 (mips_prepare_pch_save): Add microMIPS support.
18348 * config/mips/mips.h (TARGET_COMPRESSION): New.
18349 (TARGET_CPU_CPP_BUILTINS): Update macro
18350 to use new compression flags and to support microMIPS.
18351 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18352 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18353 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18354 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18355 (ASM_SPEC): Support mmicromips and mno-micromips.
18356 (M16STORE_REG_P): New macro.
18357 (MIPS_CALL): Support TARGET_MICROMIPS.
18358 (MICROMIPS_J): New macro.
18359 (mips_base_mips16): Rename this...
18360 (mips_base_compression_flags): ...to this.
18361 (UMIPS_12BIT_OFFSET_P): New macro.
18362 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18363 (MULTILIB_DIRNAMES): Likewise.
18364
cdebbc6d
JJ		 183652013-02-24 Jakub Jelinek <jakub@redhat.com>
18366
18367 PR target/52555
18368 * target-globals.c (save_target_globals): For init_reg_sets and
18369 target_reinit remporarily set this_fn_optabs to this_target_optabs.
18370
18c63565
JG		 183712013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
18372
18373 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
18374 * config/aarch64/t-aarch64
18375 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
18376
2194f7a2
VM		 183772013-02-22 Vladimir Makarov <vmakarov@redhat.com>
18378
18379 PR inline-asm/56148
18380 * lra-constraints.c (process_alt_operands): Reload operand
18381 conflicting with earlier clobber only if no more other conflicting
18382 operands.
18383
7d613735
JJ		 183842013-02-22 Jakub Jelinek <jakub@redhat.com>
18385
18386 PR sanitizer/56393
18387 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
18388 if not linking a shared library.
18389
ac8d93a7
SL		 183902013-02-22 Seth LaForge <sethml@google.com>
18391
18392 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
18393
e0237780
GY		 183942013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
18395
3fdb53c1
TB		 18396 * config/arm/arm.md (split for extendsidi): Update condition.
18397 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
18398 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
18399 (qhs_zextenddi_cstr): Likewise.
e0237780 18400
d7fde18c
JJ		 184012013-02-21 Jakub Jelinek <jakub@redhat.com>
18402
be63b77d
JJ		 18403 PR middle-end/56420
18404 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
18405 avoid signed wrapping.
18406 (expand_mult): Handle properly multiplication by
18407 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
18408 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
18409 in the compiler if coeff is HOST_WIDE_INT_MIN.
18410 (expand_divmod): Don't make ext_op1 static, change it's type to
18411 uhwi. Avoid undefined behavior in -INTVAL (op1).
18412
d7fde18c
JJ		 18413 PR rtl-optimization/50339
18414 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
18415 field.
18416 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
18417 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
18418 into splitting_ashiftrt field.
18419 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
18420 ASHIFTRT.
18421 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
18422 choices.
18423
6aad4455
AH		 184242013-02-20 Aldy Hernandez <aldyh@redhat.com>
18425
18426 PR middle-end/56108
18427 * trans-mem.c (execute_tm_mark): Do not expand transactions that
18428 are sure to go irrevocable.
18429
38fe784d
HPN		 184302013-02-21 Hans-Peter Nilsson <hp@axis.com>
18431
18432 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
18433 scalars are valid operands.
18434
0fd44da3
MJ		 184352013-02-21 Martin Jambor <mjambor@suse.cz>
18436
18437 PR tree-optimization/56310
18438 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
18439 only matching indices and non-negative final offsets.
18440 (intersect_aggregates_with_edge): Pass src_idx to
18441 agg_replacements_to_vector. Pass src_idx insstead of index to
18442 intersect_with_agg_replacements.
18443
7a92038b
MJ		 184442013-02-21 Martin Jambor <mjambor@suse.cz>
18445
18446 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
18447 instead of hard-wired defaults.
18448
c0da9c37
MR		 184492013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
18450
18451 * doc/invoke.texi (MIPS Options): Update documentation of the
18452 floating-point multiply-accumulate instruction restrictions.
18453
d247ea0c 184542013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS		 18455
18456 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
18457 asan_shadow_offset on x86_64 linux.
18458
22deefcb
RB		 184592013-02-21 Richard Biener <rguenther@suse.de>
18460
18461 PR tree-optimization/56415
18462 Revert
18463 2013-02-11 Richard Biener <rguenther@suse.de>
18464
18465 PR tree-optimization/56273
18466 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18467 first VRP run.
18468
7bcc6e75
JJ		 184692013-02-21 Jakub Jelinek <jakub@redhat.com>
18470
a0ad148f
JJ		 18471 PR bootstrap/56258
18472 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
18473 instead of @itemx.
18474
7bcc6e75
JJ		 18475 PR inline-asm/56405
18476 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
18477 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
18478
a0a7b611
JH		 184792013-02-20 Jan Hubicka <jh@suse.cz>
18480
18481 PR tree-optimization/56265
88b97037
UB		 18482 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
18483 when target is referenced for first time.
a0a7b611 18484
c0e50f72
RB		 184852013-02-20 Richard Biener <rguenther@suse.de>
18486
18487 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
18488 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
18489 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
18490 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
18491 not return anything.
18492 (rename_ssa_copies): Do not remove unused locals.
18493 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 18494 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB		 18495 * passes.c (execute_function_todo): Do not schedule unused locals
18496 removal if cleanup_tree_cfg did something.
18497 * tree-ssa-live.c (remove_unused_locals): Dump statistics
18498 about the number of removed locals.
18499
a52ca739
RB		 185002013-02-20 Richard Biener <rguenther@suse.de>
18501
18502 PR tree-optimization/56398
88b97037 18503 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 18504
ec9c9d1b
MJ		 185052013-02-20 Martin Jambor <mjambor@suse.cz>
18506
18507 PR tree-optimization/55334
18508 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
18509 restricted pointers to arrays.
18510
e91c8ed6 185112013-02-20 Richard Biener <rguenther@suse.de>
88b97037 18512 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB		 18513
18514 PR tree-optimization/56396
18515 * tree-ssa-ccp.c (n_const_val): New static variable.
18516 (get_value): Return NULL for SSA names we don't have a lattice
18517 entry for.
18518 (ccp_initialize): Initialize n_const_val.
18519 * tree-ssa-copy.c (n_copy_of): New static variable.
18520 (init_copy_prop): Initialize n_copy_of.
18521 (get_value): Return NULL_TREE for SSA names we don't have a
18522 lattice entry for.
18523
3d916479
MJ		 185242013-02-20 Martin Jambor <mjambor@suse.cz>
18525
18526 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
18527
71a86758
RB		 185282013-02-20 Richard Biener <rguenther@suse.de>
18529
18530 * genpreds.c (write_lookup_constraint): Do not compare first
18531 letter of the constraint again.
18532
79836a12
RB		 185332013-02-20 Richard Biener <rguenther@suse.de>
18534
18535 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
18536 and ceil_log2.
18537 (get_use_iv_cost): Terminate hashtable walk when coming across
18538 an empty entry.
18539
bbe4fb2c
IZ		 185402013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
18541
18542 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
18543 reassociation for avx2 targets.
18544
a72d8780 185452012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 18546
c77f83d5
EI		 18547 * config/microblaze/microblaze.c: microblaze_has_clz = 0
18548 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 18549 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI		 18550 version and TARGET_PATTERN_COMPARE check
18551 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 18552
a72d8780 185532012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 18554
a72d8780 18555 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI		 18556 function before branching.
18557
242387fa
AB		 185582012-02-19 Andrey Belevantsev <abel@ispras.ru>
18559
18560 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
18561 DUMP_INSN_RTX_UID.
18562 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
18563
b4979ab9
AB		 185642012-02-19 Andrey Belevantsev <abel@ispras.ru>
18565
18566 PR middle-end/55889
b4979ab9
AB		 18567 * sel-sched.c: Include ira.h.
18568 (implicit_clobber_conflict_p): New function.
18569 (moveup_expr): Use it.
88b97037 18570 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 18571
24d63016
RB		 185722013-02-19 Richard Biener <rguenther@suse.de>
18573
18574 PR tree-optimization/56384
18575 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
18576 (vn_hash_type): Split out from ...
18577 (vn_hash_constant_with_type): ... here.
18578 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
18579 (vn_phi_eq): Compare types from vn_phi_s structure.
18580 (vn_phi_lookup): Populate vn_phi_s type.
18581 (vn_phi_insert): Likewise.
18582
a475fd3d
JJ		 185832013-02-19 Jakub Jelinek <jakub@redhat.com>
18584
47cc28f5
JJ		 18585 PR tree-optimization/56350
18586 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
18587 if haven't found reduction or nested cycle operand, rather than
18588 asserting we must find it.
18589
a475fd3d
JJ		 18590 PR tree-optimization/56381
18591 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
18592 to fold_build3.
18593
185942013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH		 18595 Jakub Jelinek <jakub@redhat.com>
18596
18597 PR target/52555
18598 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
18599 (swap_optab_enable): Same.
18600 (init_all_optabs): Use argument instead of global.
88b97037 18601 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH		 18602 * expr.h (init_all_optabs): Add argument to prototype.
18603 (TREE_OPTIMIZATION_OPTABS): New.
18604 (save_optabs_if_changed): Protoize.
18605 * optabs.h: Declare this_fn_optabs.
18606 * optabs.c (save_optabs_if_changed): New.
18607 Declare this_fn_optabs.
18608 (init_optabs): Add argument to init_all_optabs() call.
18609 * function.c (invoke_set_current_function_hook): Handle per
18610 function optabs.
18611 * function.h (struct function): New field optabs.
18612 * config/mips/mips.c (mips_set_mips16_mode): Handle when
18613 optimization_current_node has changed.
18614 * target-globals.h (save_target_globals_default_opts): Protoize.
18615 * target-globals.c (save_target_globals_default_opts): New.
18616
3f587ca3
JDA		 186172013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18618
18619 PR target/56347
18620 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
18621 registers %fr12 and %fr12R as call used.
ceaca33e
JDA		 18622
18623 PR target/56214
18624 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB		 18625 and HImode, require all displacements to be an integer multiple of
18626 their mode size.
ceaca33e
JDA		 18627 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
18628 only allow QImode and HImode when reload is in progress and strict is
18629 true. Likewise for symbolic addresses. Use base14_operand to check
18630 displacements in REG+BASE addresses.
18631
fe0b4796
RB		 186322013-02-18 Richard Biener <rguenther@suse.de>
18633
18634 PR tree-optimization/56366
18635 * tree-vect-loop.c (get_initial_def_for_induction): Properly
18636 handle sign-conversion of outer-loop initial induction value.
18637
6aaf596b
RB		 186382013-02-18 Richard Biener <rguenther@suse.de>
18639
73db8ff1 18640 PR middle-end/56349
6aaf596b
RB		 18641 * cfghooks.c (merge_blocks): If we merge a latch into another
18642 block adjust references to it.
18643 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
18644 (verify_loop_structure): Verify that a recorded latch is in fact
18645 a latch.
18646
5e97dfb6
RB		 186472013-02-18 Richard Biener <rguenther@suse.de>
18648
18649 PR tree-optimization/56321
18650 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
18651 order SSA name release and virtual operand unlinking.
18652
825527e8
EI		 186532013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18654
18655 * config/microblaze/microblaze.md (save_stack_block): Define.
18656 (restore_stack_block): Likewise.
18657
debd11d9
EI		 186582013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18659
18660 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
18661 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
18662 * config/microblaze/microblaze.c (microblaze_option_override):
18663 Bail out early for PIC modes when target does not support PIC.
18664
8ec77be0
EI		 186652013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18666
88b97037 18667 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI		 18668 Replace with a microblaze version.
18669 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 18670 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI		 18671 microblaze.
18672
b41288b3
JJ		 186732013-02-16 Jakub Jelinek <jakub@redhat.com>
18674 Dodji Seketeli <dodji@redhat.com>
18675
18676 PR asan/56330
88b97037 18677 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ		 18678 (instrument_mem_region_access): Do not forget to always put
18679 instrumentation of the of 'base' and 'base + len' in a "if (len !=
18680 0) statement, even for cases where either 'base' or 'base + len'
18681 are not instrumented -- because they have been previously
18682 instrumented. Simplify the logic by putting all the statements
18683 instrument 'base + len' inside a sequence, and then insert that
18684 sequence right before the current insertion point. Then, to
18685 instrument 'base + len', just get an iterator on that statement.
18686 And do not forget to update the pointer to iterator the function
18687 received as argument.
18688
47918951
VM		 186892013-02-15 Vladimir Makarov <vmakarov@redhat.com>
18690
18691 PR rtl-optimization/56348
18692 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
18693
3fb7c699
SB		 186942013-02-15 Steven Bosscher <steven@gcc.gnu.org>
18695
18696 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
18697 (clean_graph_dump_file): Pass base to start_graph_dump.
18698
576fe41a
RH		 186992013-02-14 Richard Henderson <rth@redhat.com>
18700
18701 PR target/55941
18702 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
18703
bb3accfa
SB		 187042013-02-14 Steven Bosscher <steven@gcc.gnu.org>
18705
18706 * collect2-aix.h: Define F_LOADONLY.
18707
03143140
RB		 187082013-02-14 Richard Biener <rguenther@suse.de>
18709
18710 PR lto/50494
18711 * varasm.c (output_constant_def_1): Get the decl representing
18712 the constant as argument.
18713 (output_constant_def): Wrap output_constant_def_1.
18714 (make_decl_rtl): Use output_constant_def_1 with the decl
18715 representing the constant.
18716 (build_constant_desc): Optionally re-use a decl already
18717 representing the constant.
18718 (tree_output_constant_def): Adjust.
18719
8fb06726
DS		 187202013-02-14 Dodji Seketeli <dodji@redhat.com>
18721
18722 Fix an asan crash
18723 * asan.c (instrument_builtin_call): Really put the length of the
18724 second source argument into src1_len.
18725
c4bfe8bf
JJ		 187262013-02-13 Jakub Jelinek <jakub@redhat.com>
18727
18728 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
18729 argument. If it is false, don't create edge from then_bb to
18730 fallthru_bb.
18731 (insert_if_then_before_iter): Pass true to it.
18732 (build_check_stmt): Pass false to it.
18733 (transform_statements): Flush hash table only on extended basic
18734 block boundaries, rather than at the beginning of every bb.
18735 Don't flush hash table on nonfreeing_call_p calls.
18736 * tree-flow.h (nonfreeing_call_p): New prototype.
18737 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
18738
7afe2801
DM		 187392013-02-13 David S. Miller <davem@davemloft.net>
18740
18741 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
18742
70cc3288
VM		 187432013-02-13 Vladimir Makarov <vmakarov@redhat.com>
18744
18745 PR target/56184
18746 * ira.c (max_regno_before_ira): Move from ...
18747 (ira): ... here.
18748 (fix_reg_equiv_init): Use max_regno_before_ira instead of
18749 vec_safe_length.
18750
6422242b
JJ		 187512013-02-13 Jakub Jelinek <jakub@redhat.com>
18752
18753 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
18754
976a81ee
RB		 187552013-02-13 Richard Biener <rguenther@suse.de>
18756
18757 PR lto/56295
18758 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
18759 globals in MEM_REFs.
18760
c1874a87
RB		 187612013-02-13 Richard Biener <rguenther@suse.de>
18762
18763 * loop-init.c (loop_optimizer_init): Clear loop state when
18764 re-initializing preserved loops.
18765 * loop-unswitch.c (unswitch_single_loop): Return whether
18766 we unswitched the loop. Do not verify loop state here.
88b97037 18767 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 18768
b4ab7d34
KS		 187692013-02-13 Kostya Serebryany <kcc@google.com>
18770
88b97037
UB		 18771 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
18772 on x86_64 linux.
b4ab7d34
KS		 18773 * sanitizer.def: Rename __asan_init to __asan_init_v1.
18774
bdcbe80c
DS		 187752013-02-12 Dodji Seketeli <dodji@redhat.com>
18776
18777 Avoid instrumenting duplicated memory access in the same basic block
18778 * Makefile.in (asan.o): Add new dependency on hash-table.h
18779 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
18780 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
18781 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
18782 (free_mem_ref_resources, has_mem_ref_been_instrumented)
18783 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
18784 (get_mem_ref_of_assignment): New functions.
18785 (get_mem_refs_of_builtin_call): Extract from
18786 instrument_builtin_call and tweak a little bit to make it fit with
18787 the new signature.
18788 (instrument_builtin_call): Use the new
18789 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
18790 of is_gimple_builtin_call.
18791 (instrument_derefs, instrument_mem_region_access): Insert the
18792 instrumented memory reference into the hash table.
18793 (maybe_instrument_assignment): Renamed instrument_assignment into
18794 this, and change it to advance the iterator when instrumentation
18795 actually happened and return true in that case. This makes it
18796 homogeneous with maybe_instrument_assignment, and thus give a
18797 chance to callers to be more 'regular'.
18798 (transform_statements): Clear the memory reference hash table
18799 whenever we enter a new BB, when we cross a function call, or when
18800 we are done transforming statements. Use
18801 maybe_instrument_assignment instead of instrumentation. No more
18802 need to special case maybe_instrument_assignment and advance the
18803 iterator after calling it; it's now handled just like
18804 maybe_instrument_call. Update comment.
18805
4861a1f7
RB		 188062013-02-13 Richard Biener <rguenther@suse.de>
18807
18808 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
18809 Fix loop discovery code.
18810
1c86bd80
VM		 188112013-02-12 Vladimir Makarov <vmakarov@redhat.com>
18812
18813 PR inline-asm/56148
18814 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 18815 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM		 18816 the operand is not reloaded. Prefer to reload conflicting operand
18817 if earlyclobber and matching operands are the same.
18818
7cbda518
RB		 188192013-02-12 Richard Biener <rguenther@suse.de>
18820
18821 PR lto/56297
18822 * lto-streamer-out.c (write_symbol): Do not output symbols
18823 for hard register variables.
18824
e68a4ef6
GJL		 188252013-02-12 Georg-Johann Lay <avr@gjlay.de>
18826
18827 PR target/54222
18828 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
18829 (umulsidi3_insn, mulsidi3_insn): New insns.
18830
a72d8780 188312013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL		 18832
18833 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
18834 (struct tune_params): Add vec_costs field.
18835 * config/arm/arm.c (arm_builtin_vectorization_cost)
18836 (arm_add_stmt_cost): New functions.
18837 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
18838 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
18839 (arm_default_vec_cost): New struct of type cpu_vec_costs.
18840 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
18841 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
18842 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
18843 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
18844
43320568
RB		 188452013-02-12 Richard Biener <rguenther@suse.de>
18846
18847 PR lto/56295
18848 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
18849 decls again if possible.
18850
a011aa39
RB		 188512013-02-12 Richard Biener <rguenther@suse.de>
18852
18853 PR middle-end/56288
18854 * tree-ssa.c (verify_ssa_name): Fix check, move
18855 SSA_NAME_IN_FREE_LIST check up.
18856
6da26889
JJ		 188572013-02-12 Jakub Jelinek <jakub@redhat.com>
18858 Steven Bosscher <steven@gcc.gnu.org>
18859
18860 PR rtl-optimization/56151
18861 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
18862 equal to op0 or op1, and last_insn pattern is CODE operation
18863 with MEM dest and one of the operands matches that MEM.
18864
f80e0faf
ST		 188652013-02-11 Sriraman Tallam <tmsriramgoogle.com>
18866
18867 * doc/extend.texi: Document Function Multiversioning and "default"
18868 parameter string to target attribute.
18869 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
18870 target attribute parameter is "default".
18871 (ix86_compare_version_priority): Remove checks for target attribute.
18872 (ix86_mangle_function_version_assembler_name): Change error to sorry.
18873 Remove check for target attribute equal to NULL. Add assert.
18874 (ix86_generate_version_dispatcher_body): Change error to sorry.
18875
6c59ffd1
IS		 188762013-02-11 Iain Sandoe <iain@codesourcery.com>
18877 Jack Howarth <howarth@bromo.med.uc.edu>
18878 Patrick Marlier <patrick.marlier@gmail.com>
18879
18880 PR libitm/55693
18881 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
18882 define ENDFILE_SPEC as TM_DESTRUCTOR.
18883 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
18884
a72d8780 188852013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP		 18886 Jack Howarth <howarth@bromo.med.uc.edu>
18887 Jakub Jelinek <jakub@redhat.com>
18888
18889 PR sanitizer/55617
18890 * config/darwin.c (cdtor_record): Rename ctor_record.
18891 (sort_cdtor_records): Rename sort_ctor_records.
18892 (finalize_dtors): New routine to sort destructors by
18893 priority before use in assemble_integer.
18894 (machopic_asm_out_destructor): Use finalize_dtors if needed.
18895
b63fe007
UB		 188962013-02-11 Uros Bizjak <ubizjak@gmail.com>
18897
18898 PR rtl-optimization/56275
18899 * simplify-rtx.c (avoid_constant_pool_reference): Check that
18900 offset is non-negative and less than cmode size before
18901 calling simplify_subreg.
18902
8e89b5b5
RB		 189032013-02-11 Richard Biener <rguenther@suse.de>
18904
18905 PR tree-optimization/56264
18906 * cfgloop.h (fix_loop_structure): Adjust prototype.
18907 * loop-init.c (fix_loop_structure): Return the number of
18908 newly discovered loops.
18909 * tree-cfgcleanup.c (repair_loop_structures): When new loops
18910 are discovered, do a full loop-closed SSA rewrite.
18911
b4a4b56d
RB		 189122013-02-11 Richard Biener <rguenther@suse.de>
18913
18914 PR tree-optimization/56273
18915 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18916 first VRP run.
18917 (check_array_ref): Fix missing newline in dumps.
18918 (search_for_addr_array): Likewise.
18919
0c885229
DE		 189202013-02-09 David Edelsohn <dje.gcc@gmail.com>
18921
18922 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
18923
59ac9a55
JJ		 189242013-02-09 Jakub Jelinek <jakub@redhat.com>
18925
18926 PR target/56256
18927 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
18928
25bb0bb5
VM		 189292013-02-08 Vladimir Makarov <vmakarov@redhat.com>
18930
18931 PR rtl-optimization/56246
0c885229 18932 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM		 18933 reload pseudo.
18934 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
18935 constraints are satisfied.
18936
a698cc03
JL		 189372013-02-08 Jeff Law <law@redhat.com>
18938
18939 PR debug/53948
18940 * emit-rtl.c (reg_is_parm_p): New function.
18941 * regs.h (reg_is_parm_p): New prototype.
18942 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
18943 callee-clobbered registers.
18944
e1122ddd
MM		 189452013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
18946
18947 PR target/56043
18948 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
18949 If there is no implicit builtin declaration, just return NULL.
18950
19c5f6e6
UB		 189512013-02-08 Uros Bizjak <ubizjak@gmail.com>
18952
18953 * config/i386/sse.md (FMAMODEM): New mode iterator.
18954 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
18955 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
18956
2480f2ca 189572013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 18958
2480f2ca
UB		 18959 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
18960 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
18961 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
18962
189632013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18964
18965 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
18966 (microblaze*-*-elf): Likewise.
18967 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
18968 LINK_SPEC.
18969 * config/microblaze/microblaze-c.c: Add builtin defines for
18970 _LITTLE_ENDIAN and _BIG_ENDIAN.
18971 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
18972 add to TARGET_DEFAULT flags.
76ef61fb 18973 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB		 18974 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
18975 * config/microblaze/microblaze.md: Update extendsidi2 and
18976 movdi_internal instructions to use low-order / high-order reg
18977 print_operands.
18978 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
18979 options and inversemask / mask of LITTLE_ENDIAN.
18980 * config/microblaze/t-microblaze: Expand multilib options to
18981 include mlittle-endian (le) and update exceptions patterns.
18982
600a5961
JJ		 189832013-02-08 Jakub Jelinek <jakub@redhat.com>
18984
5df81313
JJ		 18985 PR rtl-optimization/56195
18986 * lra-constraints.c (get_reload_reg): Don't reuse regs
18987 if they have smaller mode than requested, if they have
18988 wider mode than requested, try to return a SUBREG.
18989
600a5961
JJ		 18990 PR tree-optimization/56250
18991 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
18992 if type is unsigned and code isn't MULT_EXPR.
18993
ff544649
GJL		 189942013-02-08 Georg-Johann Lay <avr@gjlay.de>
18995
18996 PR tree-optimization/56064
18997 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
18998 bits according to mode.
18999 * fixed-value.h (fixed_from_double_int)
19000 (const_fixed_from_double_int): Adjust comments.
19001
e45cde98
RB		 190022013-02-08 Richard Biener <rguenther@suse.de>
19003
19004 PR lto/56231
19005 * lto-streamer.h (struct data_in): Remove current_file, current_line
19006 and current_col members.
19007 * lto-streamer-out.c (lto_output_location): Stream changed bits
19008 en-block for efficiency.
19009 * lto-streamer-in.c (clear_line_info): Remove.
19010 (lto_input_location): Cache current file, line and column
19011 globally via local statics. Read changed bits en-block.
19012 (input_function): Do not call clear_line_info.
19013 (lto_read_body): Likewise.
19014 (lto_input_toplevel_asms): Likewise.
19015
c1ca73d8
MM		 190162013-02-08 Michael Matz <matz@suse.de>
19017
19018 PR tree-optimization/52448
19019 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19020 (nt_call_phase): New static.
19021 (add_or_mark_expr): Only mark accesses with newer phase than any
19022 call seen.
19023 (nonfreeing_call_p): New.
19024 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19025 (nt_fini_block): Keep blocks marked as visited.
19026 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19027
57548aa2
RB		 190282013-02-08 Richard Biener <rguenther@suse.de>
19029
19030 * ira.c (ira): Free broken dominator information.
19031
8e10366f
UB		 190322013-02-08 Uros Bizjak <ubizjak@gmail.com>
19033
19034 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19035
f64fb0fa
MP		 190362013-02-08 Marek Polacek <polacek@redhat.com>
19037
8e10366f 19038 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 19039
0375167b
RB		 190402013-02-08 Richard Biener <rguenther@suse.de>
19041
19042 PR middle-end/56181
19043 * cfgloop.h (flow_loops_find): Adjust.
19044 (bb_loop_header_p): Declare.
19045 * cfgloop.c (bb_loop_header_p): New function split out from ...
19046 (flow_loops_find): ... here. Adjust function signature,
19047 support incremental loop structure update.
19048 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19049 * cfgloopmanip.c (fix_loop_structure): Move ...
19050 * loop-init.c (fix_loop_structure): ... here.
19051 (apply_loop_flags): Split out from ...
19052 (loop_optimizer_init): ... here.
19053 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19054 in incremental mode, only remove dead loops here.
19055
85d768f3
GJL		 190562013-02-08 Georg-Johann Lay <avr@gjlay.de>
19057
19058 PR target/54222
19059 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19060 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19061 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19062 (*round<mode>3.libgcc): New insns for fixed-modes.
19063 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19064 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19065 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19066 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19067 implementations. Define to __builtin_avr_absFX,
19068 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19069 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19070 __builtin_avr_countlsFX, respectively.
19071 * config/avr/avr-c.c (target.h): Include it.
19072 (enum avr_builtin_id): New enum.
19073 (avr_resolve_overloaded_builtin): New static function.
19074 (avr_register_target_pragmas): Use it to set
19075 targetm.resolve_overloaded_builtin.
19076 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19077 tree nodes used by DEF_BUILTIN.
19078 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19079 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19080 <AVR_BUILTIN_xxBITS>: Same.
19081
661bc682
RB		 190822013-02-08 Richard Biener <rguenther@suse.de>
19083
19084 * cfgloop.c (verify_loop_structure): Properly handle
19085 a loop exiting to another loop header.
19086 * ira-int.h (ira_loops): Remove.
19087 * ira.c (ira_loops): Remove.
19088 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19089 (do_reload): Use loop_optimizer_finalize.
19090 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19091 number_of_loops to access the loop tree.
19092 (more_one_region_p): Likewise.
19093 (finish_loop_tree_nodes): Likewise.
19094 (rebuild_regno_allocno_maps): Likewise.
19095 (mark_loops_for_removal): Likewise.
19096 (mark_all_loops_for_removal): Likewise.
19097 (remove_unnecessary_regions): Likewise.
19098 (ira_build): Likewise.
19099 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19100
0d5049b2
RB		 191012013-02-08 Richard Biener <rguenther@suse.de>
19102
19103 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
19104 * ipa-pure-const.c (analyze_function): Avoid calling
19105 mark_irreducible_loops twice.
8e10366f 19106 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 19107
499f32e8
DM		 191082013-02-07 David S. Miller <davem@davemloft.net>
19109
19110 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
19111 on 'reg'.
19112 * var-tracking.c (vt_add_function_parameter): Test the presence of
19113 HAVE_window_save properly and do not remap argument registers when
19114 we have a leaf function.
19115
6edc3e32
UB		 191162013-02-07 Uros Bizjak <ubizjak@gmail.com>
19117
19118 PR bootstrap/56227
19119 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
19120 instead of "ll".
19121 * config/i386/i386.c (ix86_print_operand): Ditto.
19122
5306401f
VM		 191232013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19124
6edc3e32 19125 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 19126
027ece11
VM		 191272013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19128
19129 PR rtl-optimization/56225
19130 * lra-constraints.c (process_alt_operands): Check that reload hard
19131 reg can hold value for strict_low_part.
19132
f980dfdb
JJ		 191332013-02-07 Jakub Jelinek <jakub@redhat.com>
19134
19135 PR debug/56154
19136 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
19137 dwarf2out_end_function.
19138 (in_first_function_p, maybe_at_text_label_p,
19139 first_loclabel_num_not_at_text_label): New variables.
19140 (dwarf2out_var_location): In the first function find out
6edc3e32 19141 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ		 19142 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
19143 functions.
19144
73dd3123
EB		 191452013-02-07 Eric Botcazou <ebotcazou@adacore.com>
19146
19147 PR rtl-optimization/56178
19148 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
19149 SUBREG of a register. Tidy up related block of code.
19150 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
19151 note if the source is a register or a SUBREG of a register.
19152
e3936f47
JJ		 191532013-02-07 Jakub Jelinek <jakub@redhat.com>
19154
19155 PR target/56228
19156 * config/rs6000/rs6000.md (ptrm): New mode attr.
19157 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19158 call_value_indirect_aix<pttrsize>,
19159 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19160 m in constraints.
19161
d96d674b
MH		 191622013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19163
19164 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19165 if -bnortl. Convert to strcmp and strncmp.
19166
921f2dee
AM		 191672013-02-07 Alan Modra <amodra@gmail.com>
19168
19169 PR target/54009
19170 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19171 addresses won't wrap when offsetting.
19172 (rs6000_secondary_reload): Provide secondary reloads needed for
19173 wrapping LO_SUM addresses.
19174
d09c7dba
TS		 191752013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19176
19177 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19178 MACH, just __MACH__.
19179
a44bbd48
RB		 191802013-02-06 Richard Biener <rguenther@suse.de>
19181
19182 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19183 instead of calling fix_loop_structure.
19184
6d840d99
JJ		 191852013-02-06 Jakub Jelinek <jakub@redhat.com>
19186
19187 PR middle-end/56217
19188 * omp-low.c (use_pointer_for_field): Return false if
19189 lower_send_shared_vars doesn't generate any copy-out code.
19190
0f33baa9
TV		 191912013-02-06 Tom de Vries <tom@codesourcery.com>
19192
19193 PR rtl-optimization/56131
19194 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19195 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19196 of the label is NULL. Add comment.
19197
11f1e3ab
JJ		 191982013-02-05 Jakub Jelinek <jakub@redhat.com>
19199
31502f9f
JJ		 19200 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19201
a74db9bd
JJ		 19202 PR sanitizer/55374
19203 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19204 (STATIC_LIBTSAN_LIBS): Likewise.
19205 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19206 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19207 is defined, don't add anything else beyond that.
19208 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19209 (LINK_COMMAND_SPEC): Use them.
19210
11f1e3ab
JJ		 19211 PR tree-optimization/56205
19212 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19213 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19214 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19215
14ac6aa2
RB		 192162013-02-05 Richard Biener <rguenther@suse.de>
19217
19218 PR tree-optimization/53342
19219 PR tree-optimization/53185
19220 * tree-vectorizer.h (vect_check_strided_load): Remove.
19221 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19222 not disallow peeling for vectorized strided loads.
19223 (vect_check_strided_load): Make static and simplify.
19224 (vect_analyze_data_refs): Adjust.
19225 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19226 correctly when vectorizing strided loads.
19227
6f22445a
RB		 192282013-02-05 Richard Biener <rguenther@suse.de>
19229
19230 * doc/install.texi: Refer to ISL, not PPL.
19231
39f9719e
JH		 192322013-02-05 Jan Hubicka <jh@suse.cz>
19233
ec4224ac
JH		 19234 PR tree-optimization/55789
19235 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19236
192372013-02-05 Jan Hubicka <jh@suse.cz>
19238
19239 PR tree-optimization/55789
39f9719e
JH		 19240 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19241 the dead call anyway.
19242
956d3b33
EB		 192432013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19244
19245 PR sanitizer/55374
19246 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19247
192482013-02-04 Alexander Potapenko <glider@google.com>
19249 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP		 19250 Jakub Jelinek <jakub@redhat.com>
19251
19252 PR sanitizer/55617
19253 * config/darwin.c (sort_ctor_records): Stabilized qsort
19254 on constructor priority by using original position.
19255 (finalize_ctors): New routine to sort constructors by
19256 priority before use in assemble_integer.
19257 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19258
7ac3af38
JJ		 192592013-02-04 Jakub Jelinek <jakub@redhat.com>
19260
19261 PR libstdc++/54314
19262 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19263 about visibility on artificial decls.
19264 * config/sol2.c (solaris_assemble_visibility): Likewise.
19265
152689dc
KT		 192662013-02-04 Kai Tietz <ktietz@redhat.com>
19267
19268 PR target/56186
19269 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19270 argument and improve checking of return-argument types for 16-byte
19271 modes.
19272 (ix86_function_value_1): Add additional valtype argument on call
19273 of function_value_64.
19274 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19275 handling infunction_value_64 function.
19276
2b5987b5
MGD		 192772013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19278
7ac3af38 19279 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 19280
14597080
RB		 192812013-02-04 Richard Biener <rguenther@suse.de>
19282
19283 PR tree-optimization/56188
19284 * tree-ssa-structalias.c (label_visit): Consider case with
19285 initially non-empty points-to set.
19286 (perform_var_substitution): Dump node mapping and clean up.
19287
ed73881e
RG		 192882013-02-04 Richard Guenther <rguenther@suse.de>
19289
19290 PR lto/56168
19291 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19292 node prevail as last resort.
19293 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 19294 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 19295
139a0707
RB		 192962013-02-04 Richard Biener <rguenther@suse.de>
19297
19298 PR tree-optimization/56113
19299 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19300 Merge into ...
19301 (equiv_class_lookup_or_add): ... this.
19302 (label_visit): Adjust and fix error in previous patch.
19303 (perform_var_substitution): Adjust.
19304
50fe8924
OE		 193052013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19306
19307 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19308 * config/sh/sh4-300.md: Likewise.
19309 * config/sh/sh4a.md: Likewise.
19310 * config/sh/constraints.md: Likewise.
19311 * config/sh/sh.md: Likewise.
19312 * config/sh/netbsd-elf.h: Likewise.
19313 * config/sh/predicates.md: Likewise.
19314 * config/sh/sh-protos.h: Likewise.
19315 * config/sh/ushmedia.h: Likewise.
19316 * config/sh/linux.h: Likewise.
19317 * config/sh/sh.c: Likewise.
19318 * config/sh/superh.h: Likewise.
19319 * config/sh/elf.h: Likewise.
19320 * config/sh/sh4.md: Likewise.
19321 * config/sh/sh.h: Likewise.
19322
1a04ac2b
JDA		 193232013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19324
19325 * config/pa/constraints.md: Adjust unused letters. Change "T"
19326 constraint to match_test floating_point_store_memory_operand().
19327 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19328 (base14_operand): New.
19329 (floating_point_store_memory_operand): New.
19330 (integer_store_memory_operand): Revise to use base14_operand and
19331 reg_plus_base_memory_operand.
19332 (move_dest_operand): Allow symbolic_memory_operands.
19333 (symbolic_memory_operand): Check for LO_SOM.
19334 (symbolic_operand): Change default case to break.
19335 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19336 CONST_DOUBLE values to be reloaded by putting them into memory when
19337 the destination is a floating point register.
19338 (movdf): Remove code to handle CONST_DOUBLE.
19339 (movsf): Likewise.
19340 (reload_indf_r1): New.
19341 (reload_insf_r1): New.
19342 Consistently use "Q" and "T" constraints with integer and floating
19343 point move instructions, respectively.
19344 (movdi): Remove FAIL.
19345 Change predicate for source operand unamed DImode move from
19346 general_operand to move_src_operand.
19347 (umulsidi3): Change predicate for destination operand to
19348 register_operand.
19349 Likewise for similar unamed patterns.
19350 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19351 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19352 (hppa_legitimize_address): Simplify mask calculation.
19353 (pa_emit_move_sequence): Revised handling of secondary reloads from
19354 REG+D addresses for floating point loads and stores. Directly handle
19355 loading CONST0_RTX (mode) to a floating point register.
19356 (pa_secondary_reload): Handle reloading DF and SFmode constant values
19357 to floating point registers. Don't restrict secondary reloads to
19358 floating point registers to integer modes. Revise some comments and
19359 cleanup some code.
19360 (TARGET_LEGITIMATE_ADDRESS_P): Define.
19361 (pa_legitimate_address_p): New.
19362 (pa_legitimize_reload_address): New.
19363 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
19364 (STRICT_REG_OK_FOR_BASE_P): New.
19365 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
19366 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
19367
3bf65591
DE		 193682013-02-03 David Edelsohn <dje.gcc@gmail.com>
19369 Andrew Dixie <andrewd@gentrack.com>
19370
19371 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
19372 flag set.
19373
c0a8a3e6
RS		 193742013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
19375
19376 * expmed.c (extract_bit_field_1): Pass the full width of the
19377 structure to get_best_reg_extraction_insn.
19378
99113dff
DE		 193792013-02-01 David Edelsohn <dje.gcc@gmail.com>
19380
19381 PR target/54601
19382 * configure.ac (use_cxa_atexit): Add AIX.
19383 * configure: Regenerate.
19384
19385 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
19386
ddd84654
JJ		 193872013-02-01 Jakub Jelinek <jakub@redhat.com>
19388
19389 PR debug/54793
19390 * final.c (need_profile_function): New variable.
19391 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
19392 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
19393 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
19394 notes, targetm.asm_out.function_prologue doesn't emit anything,
19395 HAVE_prologue and profiler should be emitted before prologue,
19396 set need_profile_function instead of emitting it.
19397 (final_scan_insn): If need_profile_function, emit
19398 profile_function on the first NOTE_INSN_BASIC_BLOCK or
19399 NOTE_INSN_FUNCTION_BEG note.
19400
194012013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH		 19402
19403 * config/rs6000/rs6000.md (smulditi3): New.
19404 (umulditi3): New.
19405
ff2a9d88
RH		 19406 * config/alpha/alpha.md (umulditi3): New.
19407
14d52b90
DE		 194082013-02-01 David Edelsohn <dje.gcc@gmail.com>
19409
19410 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
19411 (ASM_OUTPUT_ALIGNED_LOCAL): New.
19412
8c7ca45c
RB		 194132013-02-01 Richard Biener <rguenther@suse.de>
19414
19415 PR tree-optimization/56113
19416 * tree-ssa-structalias.c (label_visit): Reduce work for
19417 single-predecessor nodes.
19418
9f419393
EB		 194192013-02-01 Eric Botcazou <ebotcazou@adacore.com>
19420
19421 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
19422 range isn't testing for zero.
19423
c5f4be84
SB		 194242013-01-31 Steven Bosscher <steven@gcc.gnu.org>
19425
19426 PR middle-end/56113
19427 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
19428
dbdbd982
NC		 194292013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
19430 Nick Clifton <nickc@redhat.com>
19431
19432 * config/v850/constraints.md (Q): Define as a memory constraint.
19433 * config/v850/predicates.md (label_ref_operand): New predicate.
19434 (e3v5_shift_operand): New predicate.
19435 (ior_operator): New predicate.
19436 * config/v850/t-v850: Add e3v5 multilib.
19437 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
19438 (v850_gen_movdi): Prototype.
19439 * config/v850/v850.c: Add support for e3v5 architecture.
19440 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
19441 TARGET_V850E_UP.
19442 (construct_save_jarl): Add e3v5 long JARL support.
19443 (v850_adjust_insn_length): New function. Adjust length of call
19444 insns when using e3v5 instructions.
19445 (v850_gen_movdi): New function: Generate instructions to move a
19446 DImode value.
19447 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
19448 (CPP_SPEC): Define __v850e3v5__ as appropriate.
19449 (TARGET_USE_FPU): Enable for e3v5.
19450 (CONST_OK_FOR_W): New macro.
19451 (ADJUST_INSN_LENGTH): Define.
19452 * config/v850/v850.md (UNSPEC_LOOP): Define.
19453 (attr cpu): Add v850e3v5.
19454 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
19455 (movdi): New pattern.
19456 (movdi_internal): New pattern.
19457 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
19458 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
19459 (cstoresf4): Likewise.
19460 (cstoredf4): Likewise.
19461 (insv): New pattern.
19462 (rotlso3_a): New pattern.
19463 (rotlsi3_b): New pattern
19464 (rotlsi3_v850e3v5): New pattern.
19465 (doloop_begin): New pattern.
19466 (fix_loop_counter): New pattern.
19467 (doloop_end): New pattern.
19468 (branch_normal): Add e3v5 long branch support.
19469 (branch_invert): Likewise.
19470 (branch_z_normal): Likewise.
19471 (branch_z_invert): Likewise.
19472 (branch_nz_normal): Likewise.
19473 (branch_nz_invert): Likewise.
19474 (call_internal_short): Add e3v5 register-indirect JARL support.
19475 (call_internal_long): Likewise.
19476 (call_value_internal_short): Likewise.
19477 (call_value_internal_long): Likewise.
19478 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
19479 (mloop): New option.
19480 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 19481 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 19482
73861a41
PK		 194832013-01-31 Paul Koning <ni1d@arrl.net>
19484
19485 PR debug/55059
19486 PR debug/54508
19487 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
19488 children if parent is a class.
19489 (prune_unused_types_prune): Don't add DW_AT_declaration.
19490
e44978dc
RB		 194912013-01-31 Richard Biener <rguenther@suse.de>
19492
19493 PR tree-optimization/56157
19494 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
19495 match up operand with SLP child.
19496
90b10dec
JM		 194972013-01-31 Jason Merrill <jason@redhat.com>
19498
04d2dadd 19499 PR debug/54410
90b10dec
JM		 19500 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
19501 parameters the first time.
19502 (gen_scheduled_generic_parms_dies): Check completeness here.
19503
6e616110
RB		 195042013-01-31 Richard Biener <rguenther@suse.de>
19505
19506 PR middle-end/53073
19507 * common.opt (faggressive-loop-optimizations): New flag,
19508 enabled by default.
19509 * doc/invoke.texi (faggressive-loop-optimizations): Document.
19510 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
19511 infer_loop_bounds_from_undefined by it.
19512
636f59cf
RB		 195132013-01-31 Richard Biener <rguenther@suse.de>
19514
19515 PR tree-optimization/56150
19516 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
19517 visit virtual operands.
19518 (find_uses_to_rename_bb): Likewise.
19519
195202013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB		 19521
19522 PR tree-optimization/56150
19523 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
19524 mixed store non-store stmts.
19525
32887460
JJ		 195262013-01-30 Jakub Jelinek <jakub@redhat.com>
19527
e60e09a0
JJ		 19528 PR sanitizer/55374
19529 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
19530 LIBASAN_EARLY_SPEC is defined.
19531 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
19532 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
19533 before %o.
19534 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
19535
32887460
JJ		 19536 PR c++/55742
19537 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
19538 invalid args instead of ICEing on it.
19539 (ix86_valid_target_attribute_tree): Return error_mark_node if
19540 ix86_valid_target_attribute_inner_p failed.
19541 (ix86_valid_target_attribute_p): Return false only if
19542 ix86_valid_target_attribute_tree returned error_mark_node. Allow
19543 target("default") attribute.
19544 (sorted_attr_string): Change argument from const char * to tree,
19545 merge in all target attribute arguments rather than just one.
19546 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
19547 instead of free. Avoid using strcat.
19548 (ix86_mangle_function_version_assembler_name): Mangle
19549 target("default") as if no target attribute is present. Adjust
19550 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
19551 instead of xmalloc and XDELETEVEC instead of free.
19552 (ix86_function_versions): Don't return true if one of the decls
19553 doesn't have target attribute. If they don't and one of the decls
19554 is DECL_FUNCTION_VERSIONED, report an error. Adjust
19555 sorted_attr_string caller. Use XDELETEVEC instead of free.
19556 (ix86_supports_function_versions): Remove.
19557 (make_name): Fix up formatting.
19558 (make_dispatcher_decl): Remove resolver_name and its initialization.
19559 Avoid leaking memory.
19560 (is_function_default_version): Return true if there is
19561 target("default") attribute rather than no target attribute at all.
19562 (make_resolver_func): Avoid leaking memory.
19563 (ix86_generate_version_dispatcher_body): Likewise.
19564 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
19565 * target.def (supports_function_versions): Remove.
19566 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
19567 * doc/tm.texi: Regenerated.
19568
73cca0cc
VM		 195692013-01-30 Vladimir Makarov <vmakarov@redhat.com>
19570
19571 PR rtl-optimization/56144
19572 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
19573 for values with side effects.
19574
11452e7b
RB		 195752013-01-30 Richard Biener <rguenther@suse.de>
19576
19577 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
19578 (sparseset_pop): Likewise.
19579 * cfganal.c (compute_idf): Likewise. Increase work-stack size
19580 to be able to use quick_push in the worker loop.
19581
01cb1ef5
MP		 195822013-01-30 Marek Polacek <polacek@redhat.com>
19583
19584 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
19585
e892936e
RB		 195862013-01-30 Richard Biener <rguenther@suse.de>
19587
19588 PR lto/56147
6edc3e32 19589 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 19590
cc06c01d
GJL		 195912013-01-30 Georg-Johann Lay <avr@gjlay.de>
19592
19593 PR tree-optimization/56064
19594 * fixed-value.c (fixed_from_double_int): New function.
19595 * fixed-value.h (fixed_from_double_int): New prototype.
19596 (const_fixed_from_double_int): New static inline function.
19597 * fold-const.c (native_interpret_fixed): New static function.
19598 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
19599 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
19600 (native_encode_fixed): New static function.
19601 (native_encode_expr) <FIXED_CST>: Use it.
19602 (native_interpret_int): Move double_int worker code to...
19603 * double-int.c (double_int::from_buffer): ...this new static method.
19604 * double-int.h (double_int::from_buffer): Prototype it.
19605
d394a308
RB		 196062013-01-30 Richard Biener <rguenther@suse.de>
19607
19608 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
19609 New pointer-map and obstack.
19610 (init_alias_vars): Allocate pointer-map and obstack.
19611 (delete_points_to_sets): Free them.
19612 (find_what_var_points_to): Cache result.
19613 (find_what_p_points_to): Adjust for changed interface of
19614 find_what_var_points_to.
19615 (compute_points_to_sets): Likewise.
19616 (ipa_pta_execute): Likewise.
19617
20804d96
RO		 196182013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19619
19620 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
19621 * configure: Regenerate.
19622 * config.in: Regenerate.
19623 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
19624 #nobits/#progbits if supported.
19625
a7ad88a2
OE		 196262013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
19627
19628 PR target/56121
19629 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
19630 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
19631 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
19632
7af79f92
GY		 196332013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19634
91bfca59
OE		 19635 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
19636 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 19637
753bcf7b
GY		 196382013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19639
91bfca59
OE		 19640 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
19641 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 19642
8cbc2ea8
GY		 196432013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19644
91bfca59
OE		 19645 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
19646 declaration.
19647 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
19648 * config/arm/cortex-a7.md: New bypasses using
19649 arm_mac_accumulator_is_result.
8cbc2ea8 19650
697a3325
GY		 196512013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19652
91bfca59 19653 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 19654 (cortex_a7_neon_mla): Likewise.
91bfca59
OE		 19655 (cortex_a7_fpfmad): New reservation.
19656 (cortex_a7_fpmacs): Use ffmas and update required units.
19657 (cortex_a7_fpmuld): Update required units and latency.
19658 (cortex_a7_fpmacd): Likewise.
19659 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
19660 (cortex_a7_neon). Likewise.
19661 (bypass) Update participating units.
697a3325 19662
29637783
GY		 196632013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19664
91bfca59
OE		 19665 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
19666 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
19667 from fmac to ffma.
19668 * config/arm/vfp11.md (vfp_farith): Use ffmas.
19669 (vfp_fmul): Use ffmad.
19670 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
19671 (cortex_r4_fmacd): Use ffmad.
19672 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
19673 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
19674 (cortex_a9_fmacd): Use ffmad.
19675 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
19676 (cortex_a8_vfp_macd): Use ffmad.
19677 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
19678 (cortex_a5_fpmacd): Use ffmad.
19679 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
19680 (cortex_a15_vfp_macd): Use ffmad.
19681 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 19682
2ee8a2d5
JM		 196832013-01-29 Jason Merrill <jason@redhat.com>
19684
19685 PR libstdc++/54314
19686 * varasm.c (default_assemble_visibility): Don't warn about
19687 visibility on artificial decls.
19688
78d087bc
RB		 196892013-01-29 Richard Biener <rguenther@suse.de>
19690
19691 PR tree-optimization/56113
19692 * tree-ssa-structalias.c (equiv_class_lookup): Also return
19693 the bitmap leader.
19694 (label_visit): Free duplicate bitmaps and record the leader instead.
19695 (perform_var_substitution): Adjust.
19696
83ba4d6f
RB		 196972013-01-29 Richard Biener <rguenther@suse.de>
19698
19699 PR tree-optimization/55270
19700 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
19701 the CFG, schedule loops for fixup.
19702
66dfe13f
NC		 197032013-01-29 Nick Clifton <nickc@redhat.com>
19704
19705 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
19706 SP_REG.
19707
5a579c3b
LE		 197082013-01-28 Leif Ekblad <leif@rdos.net>
19709
19710 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
19711 * config/i386/i386.h (TARGET_RDOS): New macro.
19712 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
19713 * config/i386/i386.c (ix86_option_override_internal): For 64bit
19714 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
19715 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
19716 DEFAULT_LARGE_SECTION_THRESHOLD.
19717 * config/i386/i386.md (R14_REG, R15_REG): New constants.
19718 * config/i386/rdos.h: New file.
19719 * config/i386/rdos64.h: New file.
19720
51e44392
BS		 197212013-01-28 Bernd Schmidt <bernds@codesourcery.com>
19722
19723 PR other/54814
19724 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
19725 TEST_HARD_REG_BIT.
19726
db1fb332
JJ		 197272013-01-28 Jakub Jelinek <jakub@redhat.com>
19728
19729 PR rtl-optimization/56117
19730 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
19731 call cselib_lookup_from_insn on the MEM before calling
19732 add_insn_mem_dependence.
19733
16917761
RB		 197342013-01-28 Richard Biener <rguenther@suse.de>
19735
19736 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
19737 to a stmt that didn't have one.
19738 (copy_phis_for_bb): Likewise for PHI arguments.
19739 (copy_debug_stmt): Likewise for debug stmts.
19740
b9fc0497
RB		 197412013-01-28 Richard Biener <rguenther@suse.de>
19742
19743 PR tree-optimization/56034
6edc3e32 19744 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB		 19745 (partition_builtin_p): Adjust.
19746 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
19747 it is the last partition.
19748 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
19749 up the vertex for the definition.
19750 (classify_partition): Classify whether a partition is a
19751 PKIND_REDUCTION, thus has uses outside of the loop.
19752 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
19753 Merge all PKIND_REDUCTION partitions into the last partition.
19754 (tree_loop_distribution): Seed partitions from reductions as well.
19755
aa710d25
JJ		 197562013-01-28 Jakub Jelinek <jakub@redhat.com>
19757
0bfbca58
JJ		 19758 PR tree-optimization/56125
19759 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
19760 pow(x,c) into sqrt(x) * powi(x, n/2) or
19761 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
19762 optimizing for size.
19763 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
19764 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
19765 integer.
19766
aa710d25
JJ		 19767 PR tree-optimization/56094
19768 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
19769 to UNKNOWN_LOCATION while gimplifying expr.
19770
77dc5297
UB		 197712013-01-27 Uros Bizjak <ubizjak@gmail.com>
19772
19773 PR target/56114
19774 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
19775 operand 0 in movabs insn template for -masm=intel asm alternative.
19776 (*movabs<mode>_2): Ditto for operand 1.
19777
0bfbca58 197782013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH		 19779
19780 PR target/54663
19781 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
19782 of microblaze-c.o
19783
0bfbca58 197842013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI		 19785
19786 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
19787 tm_file.
19788
0bfbca58 197892013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N		 19790
19791 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
19792 Undef to avoid warning.
19793
478f60f9
MH		 197942013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19795
19796 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
19797 * configure: Regenerate.
19798
d7fa6ee2
JJ		 197992013-01-25 Jakub Jelinek <jakub@redhat.com>
19800
19801 PR tree-optimization/56098
19802 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
19803 for stmts with volatile ops.
19804 (cond_store_replacement): Don't optimize if assign has volatile ops.
19805 (cond_if_else_store_replacement_1): Don't optimize if either
19806 then_assign or else_assign have volatile ops.
19807 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
19808 volatile ops.
19809
f8fe87bd
GJL		 198102013-01-25 Georg-Johann Lay <avr@gjlay.de>
19811
19812 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
19813
28398d0d
GJL		 198142013-01-25 Georg-Johann Lay <avr@gjlay.de>
19815
19816 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
19817 missing ':' in asm example.
19818
b7d7d917
TB		 198192013-01-25 Tejas Belagod <tejas.belagod@arm.com>
19820
19821 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
19822 entries into lane and laneq entries.
77dc5297
UB		 19823 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
19824 Remove AdvSIMD scalar modes.
b7d7d917
TB		 19825 (aarch64_sq<r>dmulh_laneq<mode>): New.
19826 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
19827 modes.
19828 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
19829 builtin implementations to relfect changes in RTL in aarch64-simd.md.
19830 * config/aarch64/iterators.md (VCOND): New.
19831 (VCONQ): New.
19832
556f9906
GJL		 198332013-01-25 Georg-Johann Lay <avr@gjlay.de>
19834
19835 PR target/54222
19836 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
19837 Add NULL LIBNAME argument to existing definitions.
19838 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
19839 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
19840 * config/avr/avr.c (DEF_BUILTIN): Same.
19841 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
19842 (avr_expand_builtin): Expand to a vanilla call if a libgcc
19843 implementation is available (DECL_ASSEMBLER_NAME is set).
19844 (avr_fold_absfx): New static function.
19845 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
19846 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
19847 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
19848 AVR_BUILTIN_ABSLLK.
19849 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
19850 (abshk, absk, abslk, absllk): Provide as static inline functions.
19851
1f546bbb
MP		 198522013-01-25 Marek Polacek <polacek@redhat.com>
19853
19854 PR tree-optimization/56035
19855 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
19856
7e184bd7
UB		 198572012-01-24 Uros Bizjak <ubizjak@gmail.com>
19858
19859 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
19860 (*movtf_internal_rex64): Add (!o,C) alternative
19861 (*movxf_internal_rex64): Ditto.
19862 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
19863
f5ba49ac
SM		 198642013-01-24 Shenghou Ma <minux.ma@gmail.com>
19865
19866 * doc/invoke.texi: fix typo.
19867 * doc/objc.texi: fix typo.
19868
a6343728
RS		 198692013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
19870
19871 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
19872 for the first two alternatives.
19873
cd030c07
DN		 198742013-01-24 Diego Novillo <dnovillo@google.com>
19875
77dc5297 19876 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN		 19877 (ggc-zone.o): Remove.
19878 * configure.ac: Remove option --with-gc.
19879 * configure: Re-generate.
19880 * doc/install.texi: Remove documentation for --with-gc.
19881 * gengtype.c (write_enum_defn): Remove. Update all users.
19882 (write_Types_process_field): Remove generation of gt_e_* argument.
19883 (output_type_enum): Remove. Update all users.
19884 (write_enum_defn): Remove. Update all users.
19885 (enum alloc_zone): Remove. Update all users.
77dc5297 19886 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN		 19887 * ggc-common.c (ggc_splay_alloc): Remove first argument.
19888 Update all callers.
19889 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 19890 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN		 19891 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
19892 Update all users.
aaf1e810 19893 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN		 19894 (struct alloc_zone): Remove.
19895 (ggc_internal_alloc_zone_stat): Remove.
19896 (ggc_internal_cleared_alloc_zone_stat): Remove.
19897 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB		 19898 (ggc_pch_count_object): Remove last argument. Update all users.
19899 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN		 19900 (struct alloc_zone): Remove.
19901 * ggc-zone.c: Remove.
77dc5297 19902 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN		 19903 (struct alloc_zone): Remove.
19904 (ggc_alloc_typed_stat): Remove.
19905 (ggc_alloc_typed): Remove.
19906 (ggc_splay_alloc): Remove first argument.
19907 (rtl_zone): Remove. Update all users.
19908 (tree_zone): Remove. Update all users.
19909 (tree_id_zone): Remove. Update all users.
19910 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 19911 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 19912 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 19913 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 19914
a861ffa4
GJL		 199152013-01-24 Georg-Johann Lay <avr@gjlay.de>
19916
19917 * config/avr/avr.c (avr_out_fract): Make register numbers that
19918 might be outside of source operand signed.
19919
593c0ddd
UB		 199202013-01-24 Uros Bizjak <ubizjak@gmail.com>
19921
19922 * config/i386/constraints.md (Yf): New constraint.
19923 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
19924 of f constraint to conditionaly disable x87 register preferences.
19925 (*movdf_internal): Ditto.
19926 (*movsf_internal): Ditto.
19927
e86c0101
SB		 199282013-01-24 Steven Bosscher <steven@gcc.gnu.org>
19929
19930 PR inline-asm/55934
19931 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
19932 that have operands with impossible constraints.
19933 Add a FIXME for a speed-up opportunity.
19934 * lra-constraints.c (process_alt_operands): Verify that a class
19935 selected from constraints on asms is valid for the operand mode.
19936 (curr_insn_transform): Remove incorrect comment.
19937
f6fee35f
DE		 199382013-01-23 David Edelsohn <dje.gcc@gmail.com>
19939
19940 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
19941 TOC operand is a valid symbol ref in the constant pool.
19942
199432013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 19944
aaf1e810 19945 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 19946
dc62d7d1
GJL		 199472013-01-23 Georg-Johann Lay <avr@gjlay.de>
19948
19949 PR target/54222
19950 * config/avr/stdfix.h: New file.
19951 * t-avr (stdfix-gcc.h): New rule to build it.
19952 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
19953
2660d12d
KS		 199542013-01-23 Kostya Serebryany <kcc@google.com>
19955
77dc5297
UB		 19956 * config/darwin.h: remove dependency on
19957 CoreFoundation (asan on Mac OS).
2660d12d 19958
a70418fc
JJ		 199592013-01-23 Jakub Jelinek <jakub@redhat.com>
19960
19961 PR target/49069
19962 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
19963 instead of cmpdi_operand for first comparison operand.
19964 Don't assert that comparison operands aren't both constants.
19965
47876a2a
JW		 199662013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19967
19968 * doc/install.texi (Downloading the Source): Update references to
19969 downloading separate components.
19970
199712013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW		 19972
19973 * doc/extend.texi (__int128): Improve grammar.
19974
47876a2a 199752013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB		 19976
19977 PR target/56028
19978 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
19979 alternative to (o,r).
19980 (*movdi_internal_rex64): Remove (!o,n) alternative.
19981 (DImode immediate->memory splitter): Remove.
19982 (DImode immediate->memory peephole2): Remove.
19983 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
19984 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
19985 alternative to (!o,*r).
19986 (*movtf_internal_sse): New pattern.
19987 (*movxf_internal_rex64): New pattern.
19988 (*movxf_internal): Disable for TARGET_64BIT.
19989 (*movdf_internal_rex64): Remove (!o,F) alternative.
19990
3a984f10
JJ		 199912013-01-22 Jakub Jelinek <jakub@redhat.com>
19992
502498d5
JJ		 19993 PR middle-end/56074
19994 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
19995 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
19996 * tree-vect-loop-manip.c (find_loop_location): Also ignore
19997 stmt locations where LOCATION_LOCUS of the stmt location is
19998 UNKNOWN_LOCATION or BUILTINS_LOCATION.
19999
3a984f10
JJ		 20000 PR target/55686
20001 * config/i386/i386.md (UNSPEC_STOS): New.
20002 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20003 *strsetqi_1): Add UNSPEC_STOS.
20004
fa817f7f
PC		 200052013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20006
20007 PR c++/56067
20008 * doc/invoke.texi: Remove left over -Wsynth example.
20009
8f498c1b
JJ		 200102013-01-21 Jakub Jelinek <jakub@redhat.com>
20011
20012 PR tree-optimization/56051
20013 * fold-const.c (fold_binary_loc): Don't fold
20014 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20015 a narrowing conversion, or widening conversion from signed
20016 to unsigned.
20017
47876a2a 200182013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB		 20019
20020 PR rtl-optimization/56023
20021 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20022 dependent on debug instruction.
20023
5b9db1bc
MJ		 200242013-01-21 Martin Jambor <mjambor@suse.cz>
20025
20026 PR middle-end/56022
20027 * function.c (allocate_struct_function): Call
20028 invoke_set_current_function_hook earlier.
20029
e8bb7d68
JJ		 200302013-01-21 Jakub Jelinek <jakub@redhat.com>
20031
20032 * reload1.c (init_reload): Only initialize reload_obstack
20033 during the first call.
20034
616a4e32
MP		 200352013-01-21 Marek Polacek <polacek@redhat.com>
20036
20037 * cfgloop.c (verify_loop_structure): Fix up grammar.
20038
4401981b
YHH		 200392013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20040
20041 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20042 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20043
8e87740b
RR		 200442013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20045
20046 PR target/56058
20047 * config/arm/marvell-pj4.md: Update copyright year.
20048 Fix up use of alu to alu_reg and simple_alu_imm.
20049
47876a2a 200502013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB		 20051
20052 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20053
89d56d79
VM		 200542013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20055
20056 PR target/55433
20057 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 20058 insn for secondary memory move when memory mode should be different.
89d56d79 20059
fe603553
JDA		 200602013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20061
20062 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20063 atomic_storedi_1): New patterns.
20064
01284895
VK		 200652013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20066
20067 btver2 pipeline descriptions.
20068 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20069 descriptions.
20070 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 20071 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK		 20072 type attributes.
20073 * config/i386/btver2.md: New file describing btver2 pipelines.
20074
5630e3e1
JL		 200752013-01-19 Andrew Pinski <apinski@cavium.com>
20076
20077 PR tree-optimization/52631
20078 * tree-ssa-sccvn (visit_use): Before looking up the original
20079 statement, try looking up the simplified expression.
20080
650ae806
AG		 200812013-01-19 Anthony Green <green@moxielogic.com>
20082
20083 * config/moxie/moxie.c (moxie_expand_prologue): Set
20084 current_function_static_stack_size.
20085
e300ec2d
JJ		 200862013-01-18 Jakub Jelinek <jakub@redhat.com>
20087
20088 PR tree-optimization/56029
20089 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20090 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20091
a3d7af04
SS		 200922013-01-18 Sharad Singhai <singhai@google.com>
20093
20094 PR tree-optimization/55995
20095 * dumpfile.c (dump_loc): Print location only if available.
20096 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
20097
66aa7879
VM		 200982013-01-18 Vladimir Makarov <vmakarov@redhat.com>
20099
20100 PR target/55433
20101 * lra-constraints.c (curr_insn_transform): Reuse original insn for
20102 secondary memory move.
20103 (inherit_reload_reg): Use rclass instead of cl for
20104 check_secondary_memory_needed_p.
20105
3f0fee7b
JJ		 201062013-01-18 Jakub Jelinek <jakub@redhat.com>
20107
20108 PR middle-end/56015
20109 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 20110 the case where writing real complex part of target modifies op1.
3f0fee7b 20111
70c67693
JG		 201122013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
20113
20114 * config/aarch64/aarch64-simd.md
20115 (aarch64_vcond_internal<mode>): Handle unordered cases.
20116 * config/aarch64/iterators.md (v_cmp_result): New.
20117
df8de9b3
YHH		 201182013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
20119 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20120
20121 * config/arm/marvell-pj4.md: New file.
20122 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
20123 * config/arm/arm.md (generic_sched): Add marvell_pj4.
20124 (generic_vfp): Likewise.
20125 * config/arm/arm-cores.def: Add marvell-pj4.
20126 * config/arm/arm-tune.md: Regenerate.
20127 * config/arm/arm-tables.opt: Regenerate.
20128 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
20129 * doc/invoke.texi: Document marvell-pj4.
20130
be30c356
TB		 201312013-01-18 Tejas Belagod <tejas.belagod@arm.com>
20132
20133 * config/aarch64/arm_neon.h: Map scalar types to standard types.
20134
0bfbca58 201352013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO		 20136
20137 PR debug/54114
20138 PR debug/54402
20139 PR debug/49888
20140 * var-tracking.c (negative_power_of_two_p): New.
20141 (global_get_addr_cache, local_get_addr_cache): New.
20142 (get_addr_from_global_cache, get_addr_from_local_cache): New.
20143 (vt_canonicalize_addr): Rewrite using the above. Adjust the
20144 heading comment.
20145 (vt_stack_offset_p): Remove.
20146 (vt_canon_true_dep): Always canonicalize loc's address.
20147 (clobber_overlapping_mems): Make sure we have a MEM.
20148 (local_get_addr_clear_given_value): New.
20149 (val_reset): Clear local cached entries.
20150 (compute_bb_dataflow): Create and release the local cache.
20151 Disable duplicate MEMs clobbering.
20152 (emit_notes_in_bb): Clobber MEMs likewise.
20153 (vt_emit_notes): Create and release the local cache.
20154 (vt_initialize, vt_finalize): Create and release the global
20155 cache, respectively.
1f6bc337 20156 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 20157
0bfbca58 201582013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO		 20159
20160 PR libmudflap/53359
20161 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20162 not found in the symtab.
20163
0bfbca58 201642013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 20165
c350ba53 20166 PR debug/56006
3aa03517
AO		 20167 PR rtl-optimization/55547
20168 PR rtl-optimization/53827
20169 PR debug/53671
20170 PR debug/49888
20171 * alias.c (offset_overlap_p): New, factored out of...
20172 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20173 the conservative special case for symbolic constants. Don't
20174 adjust zero sizes on alignment.
20175
c664546f
JL		 201762013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20177
20178 PR rtl-optimization/52573
20179 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20180 REG_UNUSED for the same register.
20181
1bd3f750
MP		 201822013-01-17 Richard Biener <rguenther@suse.de>
20183 Marek Polacek <polacek@redhat.com>
20184
20185 PR rtl-optimization/55833
20186 * loop-unswitch.c (unswitch_loops): Move loop verification...
20187 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20188 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20189 Set it to true when we're removing a loop from hierarchy tree in
20190 an irreducible region.
20191 (fix_bb_placements): Adjust caller.
20192 (fix_loop_placements): Likewise.
20193
e52a8b71
GJL		 201942013-01-17 Georg-Johann Lay <avr@gjlay.de>
20195
20196 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20197 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20198 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20199 Remove ID. Adjust comments.
20200 * config/avr/avr-c.c (avr_builtin_name): Remove.
20201 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20202 * config/avr/avr.c (avr_tolower): New static function.
20203 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20204 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20205 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20206 default expansion.
20207
8386a7ea
JH		 202082013-01-17 Jan Hubicka <jh@suse.cz>
20209
610fb637 20210 PR tree-optimization/55273
8386a7ea
JH		 20211 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20212
47876a2a 202132013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB		 20214
20215 PR target/55981
20216 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20217 store through atomic_store<mode>_1.
20218 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20219
8222c37e
MJ		 202202013-01-17 Martin Jambor <mjambor@suse.cz>
20221
20222 PR tree-optimizations/55264
20223 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20224 for virtual methods.
20225 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20226 virtual methods before inlining is over.
20227 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20228 virtual functions.
20229 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20230 non-virtual.
20231
79f01c76
VM		 202322013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20233
20234 PR rtl-optimization/56005
20235 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20236 pending reads for prefetch.
20237
d0b6bb1b
IB		 202382013-01-16 Ian Bolton <ian.bolton@arm.com>
20239
aaf1e810 20240 * config/aarch64/aarch64.md
d0b6bb1b
IB		 20241 (*cstoresi_neg_uxtw): New pattern.
20242 (*cmovsi_insn_uxtw): New pattern.
20243 (*<optab>si3_uxtw): New pattern.
20244 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20245 (*<optab>si3_insn_uxtw): New pattern.
20246 (*bswapsi2_uxtw): New pattern.
20247
cb9cf03b
RB		 202482013-01-16 Richard Biener <rguenther@suse.de>
20249
20250 * tree-inline.c (tree_function_versioning): Remove set but
20251 never used variable.
20252
2cfc56b9
RB		 202532013-01-16 Richard Biener <rguenther@suse.de>
20254
20255 PR tree-optimization/55964
20256 * tree-flow.h (rename_variables_in_loop): Remove.
20257 (rename_variables_in_bb): Likewise.
20258 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20259 (copy_loop_before): Adjust and delete update-ssa status.
20260 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 20261 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB		 20262 (rename_variables_in_loop): Remove.
20263 (slpeel_update_phis_for_duplicate_loop): Likewise.
20264 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20265 use available cfg machinery instead of duplicating it.
20266 Update PHI nodes and perform poor-mans SSA update here.
20267 (slpeel_tree_peel_loop_to_edge): Adjust.
20268
c25a0c60
RB		 202692013-01-16 Richard Biener <rguenther@suse.de>
20270
20271 PR tree-optimization/54767
20272 PR tree-optimization/53465
20273 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20274 (vrp_visit_phi_node): For PHI arguments coming via backedges
20275 drop all symbolical range information.
20276 (execute_vrp): Compute backedges.
20277
04b535af
RB		 202782013-01-16 Richard Biener <rguenther@suse.de>
20279
20280 * doc/install.texi: Update CLooG and ISL requirements to
20281 0.18.0 and 0.11.1.
20282
8b0a1e0b
CB		 202832013-01-16 Christian Bruel <christian.bruel@st.com>
20284
20285 PR target/55301
20286 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20287 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20288 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20289
202902013-01-16 DJ Delorie <dj@redhat.com>
20291
20292 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20293 (UNSPECV_SP_SWITCH_E): New.
20294 (sp_switch_1): Change to an unspec.
20295 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20296 replace $r15.
20297
202982013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB		 20299
20300 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20301 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20302 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20303 (expand_mem_thread_fence): Ditto.
20304 (expand_mem_signal_fence): Ditto.
20305 (expand_atomic_load): Ditto.
20306 (expand_atomic_store): Ditto.
20307
0bfbca58 203082013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO		 20309
20310 PR rtl-optimization/55547
20311 PR rtl-optimization/53827
20312 PR debug/53671
20313 PR debug/49888
20314 * alias.c (memrefs_conflict_p): Set sizes to negative after
20315 AND adjustments.
20316
305e3ac1
JJ		 203172013-01-15 Jakub Jelinek <jakub@redhat.com>
20318
20319 PR target/55940
20320 * function.c (thread_prologue_and_epilogue_insns): Always
20321 add crtl->drap_reg to set_up_by_prologue.set, even if
20322 stack_realign_drap is false.
20323
f78ac4f2
JBG		 203242013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20325
20326 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20327 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20328 *call): Fix indention.
20329
a78a8cc4
TV		 203302013-01-15 Tom de Vries <tom@codesourcery.com>
20331
20332 PR target/55876
20333 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20334 Update comment.
20335
0e80383f
VM		 203362013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20337
305e3ac1 20338 PR rtl-optimization/55153
0e80383f
VM		 20339 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20340
207b5956
MJ		 203412013-01-15 Martin Jambor <mjambor@suse.cz>
20342
20343 PR tree-optimization/55920
20344 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20345 accesses as grp_to_be_debug_replaced.
20346
a7818b54
JJ		 203472013-01-15 Jakub Jelinek <jakub@redhat.com>
20348
20349 PR tree-optimization/55920
20350 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20351 there is non-useless type conversion needed from debug rhs to lhs,
20352 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20353
b0fe107e
JM		 203542013-01-15 Joseph Myers <joseph@codesourcery.com>
20355 Mikael Pettersson <mikpe@it.uu.se>
20356
20357 PR target/43961
20358 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
20359 Thumb.
20360 (ASM_OUTPUT_CASE_LABEL): Remove.
20361 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
20362 * final.c (shorten_branches): Update alignment of labels before
20363 jump tables if CASE_VECTOR_SHORTEN_MODE.
20364
34ab62ee
RB		 203652013-01-15 Richard Biener <rguenther@suse.de>
20366
20367 PR bootstrap/55961
20368 * system.h: Do not include gmp.h for building host tools.
20369
783a3a05
RB		 203702013-01-15 Richard Biener <rguenther@suse.de>
20371
20372 PR middle-end/55882
20373 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
20374 account for bitpos when computing alignment.
20375
3a579e09
VY		 203762013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
20377
20378 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
20379 (ix86_target_macros_internal): Likewise.
20380
20381 * config/i386/i386.c (m_CORE2I7): Removed.
20382 (m_CORE_HASWELL): New macro.
20383 (m_CORE_ALL): Likewise.
20384 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
20385 (initial_ix86_arch_features): Likewise.
20386 (processor_target_table): Initializations for Core avx2.
20387 (cpu_names): New names "core-avx2".
20388 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
20389 PROCESSOR_CORE_HASWELL.
20390 (ix86_issue_rate): New case.
20391 (ia32_multipass_dfa_lookahead): Likewise.
20392 (ix86_sched_init_global): Likewise.
20393
20394 * config/i386/i386.h (TARGET_HASWELL): New macro.
20395 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
20396 (processor_type): New PROCESSOR_HASWELL.
20397
955f5a07
JJ		 203982013-01-15 Jakub Jelinek <jakub@redhat.com>
20399
ff784829
JJ		 20400 PR tree-optimization/55955
20401 * tree-vect-loop.c (vectorizable_reduction): Give up early on
20402 *SHIFT_EXPR and *ROTATE_EXPR codes.
20403
955f5a07
JJ		 20404 PR tree-optimization/48766
20405 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
20406 -ftrapv disable -fwrapv.
20407
aeb8b4e9
GJL		 204082013-01-14 Georg-Johann Lay <avr@gjlay.de>
20409
20410 PR target/55974
20411 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
20412 etc. to 1 and not to __flash.
20413 Use LL suffix for __INT24_MAX__ with -mint8.
20414 Use ULL suffix for __UINT24_MAX__ with -mint8.
20415
1c494c6a
GJL		 204162013-01-14 Georg-Johann Lay <avr@gjlay.de>
20417
20418 * config/avr/avr-arch.h
20419 (struct base_arch_s): Use typedef avr_arch_t instead.
20420 (struct arch_info_s): Use typedef avr_arch_info_t instead.
20421 (struct mcu_type_s): Use typedef avr_mcu_t instead.
20422 * config/avr/avr.c: Same.
20423 * config/avr/avr-devices.c: Same.
20424 * config/avr/driver-avr.c: Same.
20425 * config/avr/gen-avr-mmcu-texi.c: Same.
20426 * config/avr/avr-mcus.def: Adjust comment.
20427
a50344cb
TB		 204282013-01-14 Tejas Belagod <tejas.belagod@arm.com>
20429
88e784e6
UB		 20430 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
20431 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 20432
47876a2a 204332013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB		 20434 Andi Kleen <ak@linux.intel.com>
20435
20436 PR target/55948
20437 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
20438 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
20439 memmodel flag.
20440
00892272
GJL		 204412013-01-14 Georg-Johann Lay <avr@gjlay.de>
20442
20443 * config/avr/avr-stdint.h: Remove trailing blanks.
20444 * config/avr/avr-log.h: Same.
20445 * config/avr/avr-arch.h: Same.
20446 * config/avr/avr-devices.c: Same.
20447 * config/avr/avr-dimode.md: Same.
20448 * config/avr/predicates.md: Same.
20449 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 20450
00892272
GJL		 20451 * config/avr/avr-protos.h: Same. And:
20452 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
20453 (init_cumulative_args): Rename to avr_init_cumulative_args.
20454 (expand_prologue): Rename to avr_expand_prologue.
20455 (expand_epilogue): Rename to avr_expand_epilogue.
20456 (adjust_insn_length): Rename to avr_adjust_insn_length.
20457 (notice_update_cc): Rename to avr_notice_update_cc.
20458 (final_prescan_insn): Rename to avr_final_prescan_insn.
20459 * config/avr/avr.c: Same.
20460 * config/avr/avr.h: Same.
20461 * config/avr/avr.md: Remove trailing blanks.
20462 (prologue): Use avr_expand_prologue.
20463 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
20464
7bb01996
RB		 204652013-01-14 Richard Biener <rguenther@suse.de>
20466
20467 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
20468 verify_location, collect_subblocks): New functions.
20469 (verify_gimple_in_cfg): Verify that locations only reference
20470 BLOCKs in the functions BLOCK tree.
20471
2724573f
RB		 204722013-01-14 Richard Biener <rguenther@suse.de>
20473
20474 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
20475 PHI argument.
20476 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
20477 unshare reference.
20478 (insert_out_of_ssa_copy_on_edge): Likewise.
20479 (rewrite_close_phi_out_of_ssa): Likewise.
20480 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
20481 debug expressions.
20482 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
20483 propagated constants.
20484 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
20485 can not be shared.
20486
9a0bbab6
GJL		 204872013-01-14 Georg-Johann Lay <avr@gjlay.de>
20488
20489 * config/avr/avr-modes.def: Add GPL copyright notice.
20490
45805f17
UB		 204912013-01-13 Uros Bizjak <ubizjak@gmail.com>
20492
20493 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
20494 MEMMODEL_MASK to determine memory model.
20495 (atomic_store<mode>): Ditto from operands[2].
20496 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
20497
9d60be38
JJ		 204982013-01-13 Jakub Jelinek <jakub@redhat.com>
20499
20500 PR fortran/55935
45805f17 20501 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ		 20502 (fold_gimple_assign): Don't call unshare_expr here.
20503 (fold_ctor_reference): Call unshare_expr.
20504
e7f49d92
TG		 205052013-01-13 Terry Guo <terry.guo@arm.com>
20506
aaf1e810
EB		 20507 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
20508 * doc/fragments.texi: Document MULTILIB_REUSE.
20509 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 20510 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 20511 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG		 20512 (tmpmultilib4): Ditto.
20513 (multilib_reuse): New multilib argument.
20514
fbd03a27
RS		 205152013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
20516
20517 * Makefile.in: Update copyright.
20518
b3681f13
TV		 205192013-01-12 Tom de Vries <tom@codesourcery.com>
20520
20521 PR middle-end/55890
20522 * calls.c (expand_call): Check if arg_nr is valid.
20523
3f287e4b
MM		 205242013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
20525
20526 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
20527 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM		 20528 documentation. Add missing '__' in front of
20529 __builtin_ia32_packssdw256.
3f287e4b 20530
1abcd5eb
AK		 205312013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20532
20533 PR target/55719
20534 * config/s390/s390.c (s390_preferred_reload_class): Do not return
20535 NO_REGS for larl operands.
20536 (s390_reload_larl_operand): Use s390_load_address instead of
20537 emit_move_insn.
20538
980d0812
RB		 205392013-01-11 Richard Biener <rguenther@suse.de>
20540
20541 * tree-cfg.c (verify_node_sharing_1): Split out from ...
20542 (verify_node_sharing): ... here.
20543 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
20544
7cb9fd07
EB		 205452013-01-11 Eric Botcazou <ebotcazou@adacore.com>
20546
20547 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
20548 Substitute TREECHECKING.
20549 * configure: Regenerate.
20550 * Makefile.in (TREECHECKING): New.
20551
47876a2a 205522013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG		 20553
20554 PR tree-optimization/44061
20555 * tree-vrp.c (extract_range_basic): Compute zero as
20556 value-range for __builtin_constant_p of function parameters.
20557
d1e082c2
RS		 205582013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
20559
45805f17 20560 Update copyright years.
d1e082c2 20561
f85021b0
VM		 205622013-01-10 Vladimir Makarov <vmakarov@redhat.com>
20563
9d60be38 20564 PR rtl-optimization/55672
0160442c
VM		 20565 * lra-eliminations.c (mark_not_eliminable): Permit addition with
20566 const to be eliminable.
f85021b0 20567
7a8b1ec4
DE		 205682013-01-10 David Edelsohn <dje.gcc@gmail.com>
20569
20570 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
20571 * configure: Regenerate.
20572
ad2e5b71
RB		 205732013-01-10 Richard Biener <rguenther@suse.de>
20574
ddf9322d 20575 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 20576
ffc5b2bb
RB		 205772013-01-10 Richard Biener <rguenther@suse.de>
20578
20579 PR bootstrap/55792
20580 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
20581 locations for virtual PHI arguments.
20582 (rewrite_update_phi_arguments): Likewise.
20583
e1f674e4
JS		 205842013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
20585
20586 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
20587 on to assembler.
20588
a57fc743
JJ		 205892013-01-10 Jakub Jelinek <jakub@redhat.com>
20590
20591 PR tree-optimization/55921
20592 * tree-complex.c (expand_complex_asm): New function.
20593 (expand_complex_operations_1): Call it for GIMPLE_ASM.
20594
0ff4390d
AK		 205952013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20596
20597 PR target/55718
20598 * config/s390/s390.c (s390_symref_operand_p)
20599 (s390_loadrelative_operand_p): Merge the two functions.
20600 (s390_check_qrst_address, print_operand_address): Add parameters
20601 to s390_loadrelative_operand_p invokation.
20602 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
20603 (s390_reload_larl_operand, s390_secondary_reload): Use
20604 s390_loadrelative_operand_p instead of s390_symref_operand_p.
20605 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
20606
6f557e0e
MS		 206072013-01-09 Mike Stump <mikestump@comcast.net>
20608
20609 * dse.c (record_store): Remove unnecessary assert.
20610
7770c9e9
JH		 206112013-01-09 Jan Hubicka <jh@suse.cz>
20612
20613 PR tree-optimization/55569
20614 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
20615 * cfgloop.h (scale_loop_profile): Likewise.
20616
a19b1432
JH		 206172013-01-09 Jan Hubicka <jh@suse.cz>
20618
20619 PR lto/45375
ddf9322d
UB		 20620 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
20621 functions.
a19b1432
JH		 20622 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
20623
9a002da8
RS		 206242013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
20625
20626 PR middle-end/55114
20627 * expr.h (maybe_emit_group_store): Declare.
20628 * expr.c (maybe_emit_group_store): New function.
20629 * builtins.c (expand_builtin_int_roundingfn): Call it.
20630 (expand_builtin_int_roundingfn_2): Likewise.
20631
511dcace
VM		 206322013-01-09 Vladimir Makarov <vmakarov@redhat.com>
20633
e1f2b729 20634 PR rtl-optimization/55829
511dcace
VM		 20635 * lra-constraints.c (match_reload): Add code for absent output.
20636 (curr_insn_transform): Add code for reloads of matched inputs
20637 without output.
20638
7b0fe4f4
UB		 206392013-01-09 Uros Bizjak <ubizjak@gmail.com>
20640
20641 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
20642 attribute of movddup insn to DF.
20643 (*vec_interleave_lowv2df): Ditto.
20644 (vec_dupv2df): Ditto.
20645
870ca331
JH		 206462013-01-09 Jan Hubicka <jh@suse.cz>
20647
20648 PR tree-optimiation/55875
20649 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
20650 EVERY_ITERATION parameter.
7b0fe4f4 20651 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH		 20652 (idx_infer_loop_bounds): Similarly here.
20653 (n_of_executions_at_most): Simplify
20654 to only test for cases where statement is dominated by the
7b0fe4f4 20655 particular bound; handle correctly the "postdominance" test.
870ca331
JH		 20656 (scev_probably_wraps_p): Use max loop iterations info
20657 as a global bound first.
20658
6f575fe4 206592013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD		 20660 Nick Clifton <nickc@redhat.com>
20661
20662 * config/v850/v850.md (cbranchsf4): New pattern.
20663 (cstoresf4): New pattern.
20664 (cbranchdf4): New pattern.
20665 (cstoredf4): New pattern.
20666 (movsicc): Disallow floating point comparisons.
20667 (cmpsf_le_insn): Fix order of operators.
20668 (cmpsf_lt_insn): Likewise.
20669 (cmpsf_eq_insn): Likewise.
20670 (cmpdf_le_insn): Likewise.
20671 (cmpdf_lt_insn): Likewise.
20672 (cmpdf_eq_insn): Likewise.
20673 (cmpsf_ge_insn): Use LE comparison.
20674 (cmpdf_ge_insn): Likewise.
20675 (cmpsf_gt_insn): Use LT comparison.
20676 (cmpdf_gt_insn): Likewise.
20677 (cmpsf_ne_insn): Delete pattern.
20678 (cmpdf_ne_insn): Delete pattern.
20679 * config/v850/v850.c (v850_gen_float_compare): Use
20680 gen_cmpdf_eq_insn for NE comparison.
20681 (v850_float_z_comparison_operator)
20682 (v850_float_nz_comparison_operator): Move from here ...
20683 * config/v850/predicates.md: ... to here. Move GT and GE
20684 comparisons into v850_float_z_comparison_operator.
20685 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
20686 Delete prototype.
20687 (v850_float_nz_comparison_operator): Likewise.
20688
f0d54148
JDA		 206892013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20690
20691 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
20692 with calls to gen_insvsi/gen_insvdi.
20693
8f01beca
VK		 206942013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20695
20696 * config/i386/i386.c (initial_ix86_tune_features): Set up
20697 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
20698
2aa579ad
JJ		 206992013-01-09 Steven Bosscher <steven@gcc.gnu.org>
20700 Jakub Jelinek <jakub@redhat.com>
20701
20702 PR tree-optimization/48189
20703 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
20704 If nitercst is 0, don't predict the exit edge.
20705
6edc3e32 207062013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N		 20707
20708 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
20709 in asm_fprintf with reg_names.
20710 (aarch64_print_operand_address): Likewise.
20711 (aarch64_return_addr): Likewise.
20712 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
20713
f6f94d94
JDA		 207142013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20715
20716 * config/pa/pa.h (VAL_U6_BITS_P): Define.
20717 (INT_U6_BITS): Likewise.
20718 * config/pa/predicates.md (uint6_operand): New predicate.
20719 (shift5_operand, shift6_operand): Likewise.
20720 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
20721 arith32_operand.
20722 (lshrdi3): Use shift6_operand.
20723 (shrpsi4, shrpdi4): New insn patterns.
20724 (extzv): Delete expander.
20725 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
20726 predicates in unamed zero extract patterns. Tighten common constraint.
20727 (extv): Delete expander.
20728 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
20729 predicates in unamed sign extract patterns. Tighten common constraint.
20730 (insv): Delete expander.
20731 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
20732 predicates in unamed insert patterns. Tighten common constraint.
20733 Change uint32_operand predicate to uint6_operand predicate in unamed
20734 DImode pattern to insert constant values of type 1...1xxxx.
20735
36b72910
JH		 207362013-01-04 Jan Hubicka <jh@suse.cz>
20737
20738 PR tree-optimization/55823
7b0fe4f4
UB		 20739 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
20740 issue.
36b72910 20741
47876a2a 207422013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ		 20743 Uros Bizjak <ubizjak@gmail.com>
20744
20745 PR rtl-optimization/55845
20746 * df-problems.c (can_move_insns_across): Stop scanning at
20747 volatile_insn_p source instruction or give up if
20748 across_from .. across_to range contains any volatile_insn_p
20749 instructions.
20750
4369c11e
TB		 207512013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20752
7b0fe4f4
UB		 20753 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
20754 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
20755 Declare.
4369c11e 20756 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 20757 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 20758
aeb7e7c1
JJ		 207592013-01-08 Jakub Jelinek <jakub@redhat.com>
20760
20761 PR fortran/55341
20762 * asan.c (asan_clear_shadow): New function.
20763 (asan_emit_stack_protection): Use it.
20764
a02ad1aa
TB		 207652013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20766
20767 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
20768 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
20769 with tab instead of space.
20770
f90d87f5
NC		 207712013-01-08 Nick Clifton <nickc@redhat.com>
20772
20773 * config/rl78/rl78.c (rl78_expand_prologue): Always select
20774 register bank 0 at the start of an interrupt handler.
83ffd964
NC		 20775 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
20776 MDBH registers.
f90d87f5 20777
385eb93d
JG		 207782013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20779
20780 * config/aarch64/aarch64-simd.md
20781 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
20782 (aarch64_simd_bsl): Likewise.
20783 (aarch64_vcond_internal<mode>): Likewise.
20784 (vcond<mode><mode>): Likewise.
20785 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
20786 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
20787
4dcd1054
JG		 207882013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20789
20790 * config/aarch64/aarch64-builtins.c
20791 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
20792
4267a4a6
MJ		 207932013-01-08 Martin Jambor <mjambor@suse.cz>
20794
20795 PR debug/55579
20796 * tree-sra.c (analyze_access_subtree): Return true also after
20797 potentially creating a debug-only replacement.
20798
5f4e6de3
JJ		 207992013-01-08 Jakub Jelinek <jakub@redhat.com>
20800
3138f224
JJ		 20801 PR middle-end/55890
20802 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
20803
5f4e6de3
JJ		 20804 PR tree-optimization/54120
20805 * tree-vrp.c (range_fits_type_p): Don't allow
20806 src_precision < precision from signed vr to unsigned_p
20807 if vr->min or vr->max is negative.
20808 (simplify_float_conversion_using_ranges): Test can_float_p
20809 against CODE_FOR_nothing.
20810
f3ef18ff
JJ		 208112013-01-08 Jakub Jelinek <jakub@redhat.com>
20812 Richard Biener <rguenther@suse.de>
20813
20814 PR middle-end/55851
20815 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
20816 types instead of just INTEGER_TYPE types.
20817
25c210f9
MK		 208182013-01-07 Mark Kettenis <kettenis@openbsd.org>
20819
20820 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
20821 TF_SIZE): Define.
f90d87f5 20822
81826a7b
SE		 208232013-01-07 Steve Ellcey <sellcey@mips.com>
20824
20825 PR target/42661
20826 * config/mips/mips.opt: Change mad to mmad to match documentation.
20827
35678514
GJL		 208282013-01-07 Georg-Johann Lay <avr@gjlay.de>
20829
20830 PR target/55897
20831 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
20832 .progmemx.data now.
20833
e5669488
GJL		 208342013-01-07 Georg-Johann Lay <avr@gjlay.de>
20835
20836 PR target/55897
20837 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
20838 (avr_addrspace_t): Add .section_name field.
20839 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
20840 array size.
20841 (avr_addrspace): Same. Initialize .section_name. Remove last
20842 NULL entry. Put __memx into .progmemx.data.
20843 (progmem_section_prefix): Remove.
20844 (avr_asm_init_sections): No need to initialize progmem_section.
20845 (avr_asm_named_section): Use avr_addrspace[].section_name to get
20846 section name prefix.
20847 (avr_asm_select_section): Ditto. And use get_unnamed_section to
20848 retrieve the progmem section.
20849 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
20850 boundary to run over avr_addrspace[].
20851 (avr_register_target_pragmas): Ditto.
20852
656e6f37
JJ		 208532013-01-06 Jakub Jelinek <jakub@redhat.com>
20854
20855 * varasm.c (output_constant_def_contents): For asan_protect_global
20856 protected strings, adjust DECL_ALIGN if needed, before testing for
20857 anchored symbols.
20858 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
20859 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
20860 normal decls.
20861 (output_object_block): For asan protected decls, emit asan padding
20862 after their contents.
20863 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
20864 (asan_finish_file): Test it here instead.
20865
6edc3e32
UB		 208662013-01-07 Nick Clifton <nickc@redhat.com>
20867 Matthias Klose <doko@debian.org>
20868 Doug Kwan <dougkwan@google.com>
20869 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC		 20870
20871 PR driver/55470
20872 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
20873
20874 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
20875
20876 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
20877
7b0fe4f4 20878 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC		 20879
20880 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
20881
2a095093
GJL		 208822013-01-07 Georg-Johann Lay <avr@gjlay.de>
20883
f3b503f4 20884 PR target/54461
2a095093
GJL		 20885 * doc/install.texi (Cross-Compiler-Specific Options): Document
20886 --with-avrlibc.
20887
383f9b34
TB		 208882013-01-07 Tejas Belagod <tejas.belagod@arm.com>
20889
20890 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
20891 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
20892 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
20893 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
20894 vqmovun_high_s64): Fix source operand number and update copyright.
20895
3626621a
RB		 208962013-01-07 Richard Biener <rguenther@suse.de>
20897
20898 PR middle-end/55890
20899 * gimple.h (gimple_call_builtin_p): New overload.
20900 * gimple.c (validate_call): New function.
20901 (gimple_call_builtin_p): Likewise.
20902 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20903 Use gimple_call_builtin_p.
20904 (find_func_clobbers): Likewise.
20905 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
20906 (strlen_optimize_stmt): Likewise.
20907
8b2ea410
JG		 209082013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20909
20910 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
20911 (vld1q_dup_*): Likewise.
20912 (vld1_*): Likewise.
20913 (vld1q_*): Likewise.
20914 (vld1_lane_*): Likewise.
20915 (vld1q_lane_*): Likewise.
20916
9713d329
RB		 209172013-01-07 Richard Biener <rguenther@suse.de>
20918
20919 * lto-streamer.h (LTO_minor_version): Bump to 2.
20920
3520f7cc
JG		 209212013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20922
20923 * config/aarch64/aarch64-protos.h
20924 (aarch64_const_double_zero_rtx_p): Rename to...
20925 (aarch64_float_const_zero_rtx_p): ...this.
20926 (aarch64_float_const_representable_p): New.
20927 (aarch64_output_simd_mov_immediate): Likewise.
20928 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
20929 move immediate case.
20930 * config/aarch64/aarch64.c
20931 (aarch64_const_double_zero_rtx_p): Rename to...
20932 (aarch64_float_const_zero_rtx_p): ...this.
20933 (aarch64_print_operand): Allow printing of new constants.
20934 (aarch64_valid_floating_const): New.
20935 (aarch64_legitimate_constant_p): Check for valid floating-point
20936 constants.
20937 (aarch64_simd_valid_immediate): Likewise.
20938 (aarch64_vect_float_const_representable_p): New.
20939 (aarch64_float_const_representable_p): Likewise.
20940 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
20941 (aarch64_output_simd_mov_immediate): New.
20942 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
20943 (*movdf_aarch64): Likewise.
20944 * config/aarch64/constraints.md (Ufc): New.
20945 (Y): call aarch64_float_const_zero_rtx.
20946 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
20947
e167c04d
RB		 209482013-01-07 Richard Biener <rguenther@suse.de>
20949
20950 PR tree-optimization/55888
20951 PR tree-optimization/55862
20952 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
20953 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
20954 not if it is contained therein.
20955
0139aaab
GJL		 209562013-01-07 Georg-Johann Lay <avr@gjlay.de>
20957
20958 * config/avr/t-avr: Typo.
20959
4a176b23
GJL		 209602013-01-07 Georg-Johann Lay <avr@gjlay.de>
20961
20962 PR55243
20963 * config/avr/t-avr: Don't automatically rebuild
20964 $(srcdir)/config/avr/t-multilib
20965 $(srcdir)/config/avr/avr-tables.opt
20966 $(srcdir)/doc/avr-mmcu.texi
20967 (avr-mcus): New phony target to build them on request.
20968 (s-avr-mlib, s-avr-mmcu-texi): Remove.
20969 * avr/avr-mcus.def: Adjust comments.
20970
c7afdc98
UB		 209712013-01-07 Uros Bizjak <ubizjak@gmail.com>
20972
20973 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
20974
1ab05c31
RS		 209752013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20976
20977 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
20978
488e3acc
RS		 209792013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20980
20981 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
20982
a89599d2
DE		 209832013-01-05 David Edelsohn <dje.gcc@gmail.com>
20984
20985 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
20986 to generate profiling.
20987 * config/rs6000/aix64.h (LIB_SPEC): Same.
20988
70f09188
AP		 209892013-01-04 Andrew Pinski <apinski@cavium.com>
20990
20991 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
20992 New function.
20993 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
20994
918d445f
UB		 209952013-01-04 Uros Bizjak <ubizjak@gmail.com>
20996
20997 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
20998 unconditionally.
20999 (ix86_expand_move): Ditto.
21000 (ix86_zero_extend_to_Pmode): Ditto.
21001 (ix86_expand_call): Ditto.
21002 (ix86_expand_special_args_builtin): Ditto.
21003 (ix86_expand_builtin): Ditto.
21004
361618ec
RB		 210052013-01-04 Richard Biener <rguenther@suse.de>
21006
21007 PR tree-optimization/55862
21008 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21009 translating them through PHI nodes.
21010
87eab554
MJ		 210112013-01-04 Martin Jambor <mjambor@suse.cz>
21012
21013 PR tree-optimization/55755
21014 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21015 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21016
33879b9f
MP		 210172013-01-04 Marek Polacek <polacek@redhat.com>
21018
21019 PR middle-end/55859
21020 * opts.c (default_options_optimization): Clarify error message.
21021
3068819a
RB		 210222013-01-04 Richard Biener <rguenther@suse.de>
21023
21024 PR middle-end/55863
21025 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21026 reassociation.
21027
0e4ae794
JDA		 210282013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21029
21030 PR target/53789
21031 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA		 21032 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21033 references.
0e4ae794 21034
65c88cef
RH		 210352013-01-03 Richard Henderson <rth@redhat.com>
21036
21037 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21038 after eliminating TLS symbols.
21039
8ac16127
MG		 210402013-01-03 Marc Glisse <marc.glisse@inria.fr>
21041
21042 PR bootstrap/50167
21043 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21044 * graphite-poly.c (debug_gmp_value): Likewise.
21045
bb664f09
UB		 210462013-01-03 Uros Bizjak <ubizjak@gmail.com>
21047
21048 PR target/55712
21049 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21050 selected code model, define __code_mode_small__, __code_model_medium__,
21051 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21052 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21053 xchg temporary register with %k. Declare temporary register as
21054 early clobbered.
21055 [__x86_64__]: For medium and large code models, preserve %rbx register.
21056
bcf1ef00
RB		 210572013-01-03 Richard Biener <rguenther@suse.de>
21058
0506634a 21059 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB		 21060 (dump_subscript): Adjust.
21061 (finalize_ddr_dependent): Do not dump redundant info.
21062 (analyze_siv_subscript): Adjust.
21063 (subscript_dependence_tester): Likewise.
21064 (compute_affine_dependence): Likewise.
21065
59fd17e3
RB		 210662013-01-03 Richard Biener <rguenther@suse.de>
21067
21068 Revert
21069 2013-01-03 Richard Biener <rguenther@suse.de>
21070
21071 PR tree-optimization/55857
21072 * tree-vect-stmts.c (vectorizable_load): Do not setup
21073 re-alignment for invariant loads.
21074
21075 2013-01-02 Richard Biener <rguenther@suse.de>
21076
21077 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 21078 invariant load do not generate a vector load from the scalar location.
59fd17e3 21079
595c2679
RB		 210802013-01-03 Richard Biener <rguenther@suse.de>
21081
21082 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21083 for not vectorizing.
21084 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21085 not build INDIRECT_REFs, call get_name once only.
21086 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21087 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21088
90bb7d7a
RB		 210892013-01-03 Richard Biener <rguenther@suse.de>
21090
21091 PR tree-optimization/55857
21092 * tree-vect-stmts.c (vectorizable_load): Do not setup
21093 re-alignment for invariant loads.
21094
f09b77ca
RB		 210952013-01-03 Richard Biener <rguenther@suse.de>
21096
21097 PR lto/55848
21098 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
21099 prefer a built-in decl.
21100
3461a16e
JJ		 211012013-01-03 Jakub Jelinek <jakub@redhat.com>
21102
df8e2b4f
JJ		 21103 * gcc.c (process_command): Update copyright notice dates.
21104 * gcov.c (print_version): Likewise.
21105 * gcov-dump.c (print_version): Likewise.
21106
3461a16e
JJ		 21107 PR rtl-optimization/55838
21108 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
21109 iv0.step, iv1.step and step.
21110
8b5546d6
JJ		 211112013-01-03 Jakub Jelinek <jakub@redhat.com>
21112 Marc Glisse <marc.glisse@inria.fr>
21113
21114 PR tree-optimization/55832
21115 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
21116 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
21117 integer_{one,zero}_node.
21118
8ab1d2e9
JJ		 211192013-01-03 Jakub Jelinek <jakub@redhat.com>
21120
21121 PR debug/54402
21122 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
21123 * var-tracking.c (reverse_op): Don't add reverse ops to
21124 VALUEs that have already
0506634a 21125 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 21126
5b9ad1d4
GP		 211272013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
21128
21129 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
21130
e25a6711
TJ		 211312013-01-02 Teresa Johnson <tejohnson@google.com>
21132
21133 * dumpfile.c (dump_loc): Print filename with location.
21134 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
21135 new location_t parameter to emit complete unroll message with
21136 new dump framework.
21137 (canonicalize_loop_induction_variables): Compute loops location
21138 and pass to try_unroll_loop_completely.
21139 * loop-unroll.c (report_unroll_peel): New function.
21140 (peel_loops_completely): Use new dump format with location
21141 for main dumpfile message, and invoke report_unroll_peel on success.
21142 (decide_unrolling_and_peeling): Ditto.
21143 (decide_peel_once_rolling): Remove old dumpfile message subsumed
21144 by report_unroll_peel.
21145 (decide_peel_completely): Ditto.
21146 (decide_unroll_constant_iterations): Ditto.
21147 (decide_unroll_runtime_iterations): Ditto.
21148 (decide_peel_simple): Ditto.
21149 (decide_unroll_stupid): Ditto.
21150 * cfgloop.c (get_loop_location): New function.
21151 * cfgloop.h (get_loop_location): Declare.
21152
77878621
ST		 211532013-01-02 Sriraman Tallam <tmsriram@google.com>
21154
21155 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21156 NULL.
21157
9e65d03e
JDA		 211582013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21159
21160 PR middle-end/55198
21161 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21162 BLKmode objects when EXPAND_MEMORY is specified.
21163
6a7da30f
ST		 211642013-01-02 Sriraman Tallam <tmsriram@google.com>
21165
21166 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21167 in loop predicate.
21168 (fold_builtin_cpu): Do not share cpu model decls across statements.
21169
e78167f2
JM		 211702013-01-02 Jason Merrill <jason@redhat.com>
21171
21172 PR c++/55804
21173 * tree.c (build_array_type_1): Revert earlier change.
21174
8c075fb4
YZ		 211752013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21176
21177 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21178 "cortex-a57".
21179 * config/aarch64/aarch64-tune.md: Re-generate.
21180
0682ed3e
RB		 211812013-01-02 Richard Biener <rguenther@suse.de>
21182
21183 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 21184 invariant load do not generate a vector load from the scalar location.
0682ed3e 21185
b8f6a302
RB		 211862013-01-02 Richard Biener <rguenther@suse.de>
21187
21188 PR bootstrap/55784
21189 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21190 * configure: Regenerate.
21191
04b80dbb
RS		 211922013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21193
21194 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21195 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21196 (expand_builtin_int_roundingfn_2): Keep the original target around
21197 for the fallback case.
21198
635b0b0c
RS		 211992013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21200
21201 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21202 to be clear for sign changes.
21203
42d57399
JH		 212042013-01-01 Jan Hubicka <jh@suse.cz>
21205
21206 * ipa-inline-analysis.c: Fix formatting.
21207
5bb6669d
JJ		 212082013-01-01 Jakub Jelinek <jakub@redhat.com>
21209
21210 PR tree-optimization/55831
21211 * tree-vect-loop.c (get_initial_def_for_induction): Use
21212 gsi_after_labels instead of gsi_start_bb.
ad41bd84 21213\f
86a2db33 21214Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM		 21215
21216Copying and distribution of this file, with or without modification,
21217are permitted in any medium without royalty provided the copyright
21218notice and this notice are preserved.
Mirror of https://gcc.gnu.org/git/gcc.git