]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
re PR middle-end/58385 (likely wrong code bug)
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
427e6a14
JJ
12013-09-11 Jakub Jelinek <jakub@redhat.com>
2
3 PR tree-optimization/58385
4 * fold-const.c (build_range_check): If both low and high are NULL,
5 use omit_one_operand_loc to preserve exp side-effects.
6
df45c0cc
KT
72013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8
9 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
10
80ab0b19
RB
112013-09-11 Richard Biener <rguenther@suse.de>
12
13 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
14 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
15 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
16 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
17 stmts_from_loop, known_dependences_p, build_empty_rdg,
18 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
19 * tree-loop-distribution.c: ... here.
20 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
21 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
22 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
23 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
24 * tree-loop-distribution.c: ... here.
25 * tree-loop-distribution.c: Include gimple-pretty-print.h.
26 (struct partition_s): Add loops member.
27 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
28 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
29 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
30
85a77221
AI
312013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
32 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
33 Sergey Lega <sergey.s.lega@intel.com>
34 Anna Tikhonova <anna.tikhonova@intel.com>
35 Ilya Tocar <ilya.tocar@intel.com>
36 Andrey Turetskiy <andrey.turetskiy@intel.com>
37 Ilya Verbin <ilya.verbin@intel.com>
38 Kirill Yukhin <kirill.yukhin@intel.com>
39 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
40
41 * config/i386/constraints.md (k): New.
42 (Yk): Ditto.
43 * config/i386/i386.c (const regclass_map): Add new mask registers.
44 (dbx_register_map): Ditto.
45 (dbx64_register_map): Ditto.
46 (svr4_dbx_register_map): Ditto.
47 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
48 disabled.
49 (ix86_preferred_reload_class): Disable constants for mask registers.
50 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
51 (ix86_hard_regno_mode_ok): Support new mask registers.
52 (x86_order_regs_for_local_alloc): Ditto.
53 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
54 (FIXED_REGISTERS): Add new mask registers.
55 (CALL_USED_REGISTERS): Ditto.
56 (REG_ALLOC_ORDER): Ditto.
57 (VALID_MASK_REG_MODE): New.
58 (FIRST_MASK_REG): Ditto.
59 (LAST_MASK_REG): Ditto.
60 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
61 (MAYBE_MASK_CLASS_P): New.
62 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
63 (REG_CLASS_CONTENTS): Ditto.
64 (MASK_REGNO_P): New.
65 (ANY_MASK_REG_P): Ditto.
66 (HI_REGISTER_NAMES): Add new mask registers.
67 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG,
68 MASK3_REG, MASK4_REG, MASK5_REG, MASK6_REG,
69 MASK7_REG): Constants for new mask registers.
70 (attribute "type"): Add mskmov, msklog.
71 (attribute "length_immediate"): Support them.
72 (attribute "memory"): Ditto.
73 (attribute "prefix_0f"): Ditto.
74 (*movhi_internal): Support new mask registers.
75 (*movqi_internal): Ditto.
76 (define_split): Split out clobber pattern is a logic
77 insn on mask registers.
78 (*k<logic><mode>): New.
79 (*andhi_1): Extend to support mask regs.
80 (*andqi_1): Extend to support mask regs.
81 (kandn<mode>): New.
82 (define_split): Split and-not to and and not if operands
83 are not mask regs.
84 (*<code><mode>_1): Separate HI mode to new pattern...
85 (*<code>hi_1): This.
86 (*<code>qi_1): Extend to support mask regs.
87 (kxnor<mode>): New.
88 (kortestzhi): Ditto.
89 (kortestchi): Ditto.
90 (kunpckhi): Ditto.
91 (*one_cmpl<mode>2_1): Remove HImode and handle it...
92 (*one_cmplhi2_1): ...Here, now with mask registers support.
93 (*one_cmplqi2_1): Support new mask registers.
94 (HI/QImode arithmetics splitter): Don't split if mask registers are used.
95 (HI/QImode not splitter): Ditto.
96 * config/i386/predicated.md (mask_reg_operand): New.
97 (general_reg_operand): Ditto.
98
7e570821
AI
992013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
100
101 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
102 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
103
9e1376e9
JL
1042013-09-10 Jeff Law <law@redhat.com>
105
106 PR tree-optimization/58380
107 * tree-ssa-threadupdate.c (thread_block): Recognize another case
108 of threading through a buried loop header.
109
110 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
111 return value for single successor case.
112
5bbcb888
JH
1132013-09-10 Jan Hubicka <jh@suse.cz>
114
115 * ipa-devirt.c (ipa_devirt): Enable with LTO.
116
d742ff4b
RE
1172013-09-10 Richard Earnshaw <rearnsha@arm.com>
118
119 PR target/58361
120 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
121 support conditional execution.
122 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
123
b0681c9e
VM
1242013-09-10 Vladimir Makarov <vmakarov@redhat.com>
125
126 * lra.c (lra): Clear lra_optional_reload_pseudos before every
127 constraint pass.
128 * lra-constraints.c (curr_insn_transform): Switch on optional
129 reloads. Check destination too to check move insn.
130 (undo_optional_reloads): Add check that the original peudo did not
131 changed its allocation and the optional reload was inherited on
132 last inheritance pass. Break loop after deciding to keep optional
133 reload.
134 (lra_undo_inheritance): Add check that inherited pseudo still in
135 memory.
136
c9b1c957
JG
1372013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
138
139 * config/aarch64/aarch64.md (generic_sched): New.
140 * config/aarch64/aarch64-generic.md (load): Make conditional
141 on generic_sched attribute.
142 (nonload): Likewise.
143
82d618d3
JH
1442013-09-10 Jan Hubicka <jh@suse.cz>
145
146 * lto-cgraph.c: Include ipa-utils.h.
147 (compute_ltrans_boundary): Also add possible targets into the boundary.
148
668b6894
JH
1492013-09-10 Jan Hubicka <jh@suse.cz>
150
151 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
152 VAR_DECL of vtable rather than full expression.
153
4f90d3e0
JH
1542013-09-10 Jan Hubicka <jh@suse.cz>
155 Paolo Carlini <paolo.carlini@oracle.com>
156
157 * cgraphunit.c (analyze_functions): Save input_location, set it
158 to UNKNOWN_LOCATION and restore it at the end.
159
39e87baf
MJ
1602013-09-10 Martin Jambor <mjambor@suse.cz>
161
162 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
163 represented by a thunk.
164
581aedec
JL
1652013-09-10 Jeff Law <law@redhat.com>
166
167 PR tree-optimization/58343
168 * tree-ssa-threadupdate.c (thread_block): Identify and disable
169 jump threading requests through loop headers buried in the middle
170 of a jump threading path.
171
172 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
173 in return value/type.
174
84cf4ab6
JJ
1752013-09-10 Jakub Jelinek <jakub@redhat.com>
176
177 PR rtl-optimization/58365
178 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
179 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
180 it differs.
181
447dd906
RB
1822013-09-10 Richard Biener <rguenther@suse.de>
183
184 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
185 * tree-data-ref.c (create_rdg_vertices): Collect all data
186 references, signal failure to the caller, use data-ref API.
187 (build_rdg): Compute data references only once. Maintain
188 lifetime of data references and data dependences from within
189 RDG.
190 (free_rdg): Free dependence relations.
191 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
192 inventing extra dependences.
193 (distribute_loop): Update for RDG API changes.
194
f14a7324
KT
1952013-09-10 Kai Tietz <ktietz@redhat.com>
196
197 * doc/invoke.texi (fms-extensions): Document changed
198 behavior for ms-abi targets.
199 * config/i386/i386.c (ix86_option_override_internal):
200 Set default value of option -fms-extension for ms-abi targets.
201
ff131f26
MZ
2022013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
203
204 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
205
33115a2c
AM
2062013-09-10 Alan Modra <amodra@gmail.com>
207
208 PR target/58330
209 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
210
b6872a13
AM
2112013-09-10 Alan Modra <amodra@gmail.com>
212
213 * config/rs6000/predicates.md (add_cint_operand): New.
214 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
215 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
216 using add_cint_operand.
217 (largetoc_high_plus_aix): Likewise.
218
28fd0ba2
JJ
2192013-09-09 Jakub Jelinek <jakub@redhat.com>
220
221 PR tree-optimization/58364
222 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
223 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
224 the current range can't be an unconditional true or false.
225
552d99fe
JG
2262013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
227
228 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
229
f43245d1
UB
2302013-09-09 Uros Bizjak <ubizjak@gmail.com>
231
232 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
233
c9b0866a
PC
2342013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
235
236 PR c++/43452
237 * doc/invoke.texi (-Wdelete-incomplete): Document it.
238
78d8b9f0
IB
2392013-09-09 Ian Bolton <ian.bolton@arm.com>
240
241 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
242 NO_REGS for immediate that can't be moved directly into FP_REGS.
243
274b2532
KT
2442013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
245
246 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
247 comparison with negated operand.
c9b0866a
PC
248 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
249 RTL form.
274b2532 250
467a3558
RB
2512013-09-09 Richard Biener <rguenther@suse.de>
252
253 PR middle-end/58326
254 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
255 of a subloop record all its block as affecting loop-closed SSA form.
256
088c5368
RS
2572013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
258
259 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
260 of an rtx/bitpos pair.
261 (store_fixed_bit_field): Update accordingly.
262
69db2d57
RS
2632013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
264
265 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
266 GEN_INT.
267 * builtins.c (expand_errno_check): Likewise.
268 * dwarf2cfi.c (init_return_column_size): Likewise.
269 * except.c (sjlj_mark_call_sites): Likewise.
270 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
271 * lra-constraints.c (emit_inc): Likewise.
272 * ree.c (combine_set_extension): Likewise.
273 * regmove.c (fixup_match_2): Likewise.
274 * reload1.c (inc_for_reload): Likewise.
275
69a59f0f
RS
2762013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
277
278 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
279 (force_to_mode, simplify_shift_const_1, simplify_comparison):
280 Use gen_int_mode with the mode of the associated simplify_* call.
281 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
282 * expmed.c (expand_shift_1): Likewise.
283 * function.c (instantiate_virtual_regs_in_insn): Likewise.
284 * loop-iv.c (iv_number_of_iterations): Likewise.
285 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
286 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
287
2f1cd2eb
RS
2882013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
289
290 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
291 of the associated expand_* call.
292 (asan_emit_stack_protection): Likewise.
293 * builtins.c (round_trampoline_addr): Likewise.
294 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
295 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
296 (emit_store_flag): Likewise.
297 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
298 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
299 Likewise.
300 * function.c (instantiate_virtual_regs_in_insn): Likewise.
301 * ifcvt.c (noce_try_store_flag_constants): Likewise.
302 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
303 * modulo-sched.c (generate_prolog_epilog): Likewise.
304 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
305 (expand_ctz, expand_ffs, expand_unop): Likewise.
306
4789c0ce
RS
3072013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
308
309 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
310 of the associated gen_rtx_* call.
311 * caller-save.c (init_caller_save): Likewise.
312 * combine.c (find_split_point, make_extraction): Likewise.
313 (make_compound_operation): Likewise.
314 * dwarf2out.c (mem_loc_descriptor): Likewise.
315 * explow.c (plus_constant, probe_stack_range): Likewise.
316 * expmed.c (expand_mult_const): Likewise.
317 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
318 * reload1.c (init_reload): Likewise.
319 * valtrack.c (cleanup_auto_inc_dec): Likewise.
320 * var-tracking.c (adjust_mems): Likewise.
321 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
322 rather than gen_rtx_fmt_ee.
323
a811f7cb
JH
3242013-09-09 Jan Hubicka <jh@suse.cz>
325
326 PR middle-end/58294
327 * value-prof.c (gimple_ic): Copy also abnormal edges.
328
dcad1dd3
RS
3292013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
330
331 * asan.c (asan_shadow_cst): Use gen_int_mode.
332
daf5c770
JH
3332013-09-08 Jan Hubicka <jh@suse.cz>
334
335 * ipa-profile.c: Add toplevel comment.
336 (ipa_propagate_frequency_1): Be more conservative when profile is read.
337 (contains_hot_call_p): New function.
338 (ipa_propagate_frequency): Set frequencies based on counts when
339 profile is read.
340 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
341 profile; do not tamper with profile after inlining if it is read.
342
a12cd2db
JH
3432013-09-08 Jan Hubicka <jh@suse.cz>
344
345 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
346 speculative edges.
347
856a1588
JH
3482013-09-08 Jan Hubicka <jh@suse.cz>
349
350 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
351 summary generation.
352
4f4b0b73
JL
3532013-09-08 Jeff Law <law@redhat.com>
354
355 PR bootstrap/58340
356 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
357 of 'found'.
358
cc4029ee
AK
3592013-09-08 Andi Kleen <ak@linux.intel.com>
360
361 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
362
7b395ddd
JH
3632013-09-08 Jan Hubicka <jh@suse.cz>
364
365 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
366
c34c46dd
RS
3672013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
368
369 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
370 for non-debug insns.
371 * lra.c (new_insn_reg): Take the containing insn as a parameter.
372 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
373 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
374 accordingly.
375
3462aa02
JH
3762013-09-08 Jan Hubicka <jh@suse.cz>
377
378 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
379 targets and devirtualize to BUILT_IN_UNREACHABLE.
380 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
381 * ipa.c (walk_polymorphic_call_targets): New function.
382 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
383 functions; use the new timevar.
384 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
385 was removed from the program.
386 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
387 not consider it in the walk when its vtable is dead.
388 (possible_polymorphic_call_targets_1): Pass anonymous flag to
389 record_binfo.
390 (devirt_variable_node_removal_hook): New function.
391 (possible_polymorphic_call_targets): Also register
392 devirt_variable_node_removal_hook.
393 (ipa_devirt): Do not do non-speculative devirtualization.
394 (gate_ipa_devirt): One execute if devirtualizing speculatively.
395
26e5b0fd
JH
3962013-09-08 Jan Hubicka <jh@suse.cz>
397
398 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
399 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
400 varpool_remove_variable_insertion_hook): Declare.
401 * varpool.c (varpool_node_hook_list): New structure.
402 (first_varpool_node_removal_hook,
403 first_varpool_variable_insertion_hook): New variables.
404 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
405 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
406 varpool_remove_variable_insertion_hook,
407 varpool_call_variable_insertion_hooks): New functions.
408 (varpool_remove_node): Use it.
409
97f1f314 4102013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 411
97f1f314
PC
412 PR c++/54941
413 * diagnostic.c (diagnostic_build_prefix): When s.file is
414 "<built-in>" don't output line and column numbers.
6df2ca09 415
bcb650cb
JH
4162013-09-06 Jan Hubicka <jh@suse.cz>
417
418 * cgraphunit.c (expand_thunk): Get body before touching arguments.
419 * lto-streamer-out.c: Stream thunks, too.
420 * lto-streamer-in.c (input_function): Pop cfun here
421 (lto_read_body): Instead of here.
422
87f3fea8
CT
4232013-09-06 Caroline Tice <cmtice@google.com>
424
425 * doc/install.texi: Add documentation for the --enable-vtable-verify
426 and the --disable-libvtv configure options.
427
5e8d9b45
JL
4282013-09-06 Jeff Law <law@redhat.com>
429
430 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
431 edge implied equivalences into successor phis.
432
3a001aff
JR
4332013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
434
435 * resource.c (mark_referenced_resources): Handle COND_EXEC.
436
4372013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ
438
439 * resource.c (mark_target_live_regs): Compute resources taking
440 into account if a call is predicated or not.
441
2c641f8b
EB
4422013-09-06 Eric Botcazou <ebotcazou@adacore.com>
443
444 * toplev.c (output_stack_usage): Be prepared for suffixes created by
445 the compiler in the function names.
446
a81b0a3d
JH
4472013-09-06 Jan Hubicka <jh@suse.cz>
448
449 PR middle-end/58094
450 * ipa-inline.c (has_caller_p): New function.
451 (want_inline_function_to_all_callers_p): Use it.
452 (sum_callers, inline_to_all_callers): Break out from ...
453 (ipa_inline): ... here.
454
68e4ca7e
JH
4552013-09-06 Jan Hubicka <jh@suse.cz>
456
2c641f8b
EB
457 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
458 only when AVX is enabled.
68e4ca7e 459
3c73e2fe
JG
4602013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
461
462 * config/aarch64/aarch64.md
463 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
464 is fpsimd_<load/store>2.
465 (load_pair<mode>): Likewise.
466 (store_pair<mode>): Likewise.
467
8c48eecd
JG
4682013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
469
470 * config/arm/types.md (type): Add "mrs" type.
471 * config/aarch64/aarch64.md
472 (aarch64_load_tp_hard): Make type "mrs".
473 * config/arm/arm.md
474 (load_tp_hard): Make type "mrs".
475 * config/arm/cortex-a15.md: Update with new attributes.
476 * config/arm/cortex-a5.md: Update with new attributes.
477 * config/arm/cortex-a53.md: Update with new attributes.
478 * config/arm/cortex-a7.md: Update with new attributes.
479 * config/arm/cortex-a8.md: Update with new attributes.
480 * config/arm/cortex-a9.md: Update with new attributes.
481 * config/arm/cortex-m4.md: Update with new attributes.
482 * config/arm/cortex-r4.md: Update with new attributes.
483 * config/arm/fa526.md: Update with new attributes.
484 * config/arm/fa606te.md: Update with new attributes.
485 * config/arm/fa626te.md: Update with new attributes.
486 * config/arm/fa726te.md: Update with new attributes.
487
9ea187a5
JG
4882013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
489
490 * config/aarch64/aarch64.md
491 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
492 (*movtf_aarch64): Likewise.
493 * config/arm/arm.md
494 (thumb1_movdi_insn): Use "multiple" for type where more than one
495 instruction is used for a move.
496 (*arm32_movhf): Likewise.
497 (*thumb_movdf_insn): Likewise.
498
292b89b3
JG
4992013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
500
501 * config/arm/types.md (type): Rename fcpys to fmov.
502 * config/arm/vfp.md
503 (*arm_movsi_vfp): Rename type fcpys as fmov.
504 (*thumb2_movsi_vfp): Likewise
505 (*movhf_vfp_neon): Likewise
506 (*movhf_vfp): Likewise
507 (*movsf_vfp): Likewise
508 (*thumb2_movsf_vfp): Likewise
509 (*movsfcc_vfp): Likewise
510 (*thumb2_movsfcc_vfp): Likewise
511 * config/aarch64/aarch64-simd.md
512 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
513 * config/aarch64/aarch64.md
514 (*movsi_aarch64): Replace type mov_reg with fmovs.
515 (*movdi_aarch64): Likewise
516 (*movsf_aarch64): Likewise
517 (*movdf_aarch64): Likewise
518 * config/arm/arm.c
519 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
520 * config/arm/iwmmxt.md
521 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
522 * config/arm/arm1020e.md: Update with new attributes.
523 * config/arm/cortex-a15-neon.md: Update with new attributes.
524 * config/arm/cortex-a5.md: Update with new attributes.
525 * config/arm/cortex-a53.md: Update with new attributes.
526 * config/arm/cortex-a7.md: Update with new attributes.
527 * config/arm/cortex-a8-neon.md: Update with new attributes.
528 * config/arm/cortex-a9.md: Update with new attributes.
529 * config/arm/cortex-m4-fpu.md: Update with new attributes.
530 * config/arm/cortex-r4f.md: Update with new attributes.
531 * config/arm/marvell-pj4.md: Update with new attributes.
532 * config/arm/vfp11.md: Update with new attributes.
533
6a24a83e
JG
5342013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
535
536 * config/aarch64/aarch64.md
537 (*madd<mode>): Fix type attribute.
538 (*maddsi_uxtw): Likewise.
539 (*msub<mode>): Likewise.
540 (*msubsi_uxtw): Likewise.
541 (<su_optab>maddsidi4): Likewise.
542 (<su_optab>msubsidi4): Likewise.
543
b86923f0
JG
5442013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
545
546 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
547 * config/arm/arm.md (core_cycles): Remove fdiv.
548 * config/arm/vfp.md:
549 (*sqrtsf2_vfp): Update for attribute changes.
550 (*sqrtdf2_vfp): Likewise.
551 * config/aarch64/aarch64.md:
552 (sqrt<mode>2): Update for attribute changes.
553 * config/arm/arm1020e.md: Update with new attributes.
554 * config/arm/cortex-a15-neon.md: Update with new attributes.
555 * config/arm/cortex-a5.md: Update with new attributes.
556 * config/arm/cortex-a53.md: Update with new attributes.
557 * config/arm/cortex-a7.md: Update with new attributes.
558 * config/arm/cortex-a8-neon.md: Update with new attributes.
559 * config/arm/cortex-a9.md: Update with new attributes.
560 * config/arm/cortex-m4-fpu.md: Update with new attributes.
561 * config/arm/cortex-r4f.md: Update with new attributes.
562 * config/arm/marvell-pj4.md: Update with new attributes.
563 * config/arm/vfp11.md: Update with new attributes.
564
7b49c9e1
JG
5652013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
566
567 * config/arm/types.md
568 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
569 * config/aarch64/aarch64.md
570 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
571 new attributes.
572 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
573 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
574 (float<GPI:mode><GPF:mode>2): Likewise.
575 * config/arm/vfp.md
576 (*truncsisf2_vfp): Update with new attributes.
577 (*truncsidf2_vfp): Likewise.
578 (fixuns_truncsfsi2): Likewise.
579 (fixuns_truncdfsi2): Likewise.
580 (*floatsisf2_vfp): Likewise.
581 (*floatsidf2_vfp): Likewise.
582 (floatunssisf2): Likewise.
583 (floatunssidf2): Likewise.
584 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
585 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
586 * config/arm/arm1020e.md: Update with new attributes.
587 * config/arm/cortex-a15-neon.md: Update with new attributes.
588 * config/arm/cortex-a5.md: Update with new attributes.
589 * config/arm/cortex-a53.md: Update with new attributes.
590 * config/arm/cortex-a7.md: Update with new attributes.
591 * config/arm/cortex-a8-neon.md: Update with new attributes.
592 * config/arm/cortex-a9.md: Update with new attributes.
593 * config/arm/cortex-m4-fpu.md: Update with new attributes.
594 * config/arm/cortex-r4f.md: Update with new attributes.
595 * config/arm/marvell-pj4.md: Update with new attributes.
596 * config/arm/vfp11.md: Update with new attributes.
597
e7df8af8
JG
5982013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
599
600 * config/aarch64/arm_neon.h
601 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
602 (vqtbx<1,2,3,4><q>_s8): Likewise.
603
594726e4
JG
6042013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
605
f43245d1
UB
606 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
607 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG
608 (add<mode>3): Add type attribute.
609 (add<mode>3): Likewise.
610 (usadd<mode>3): Likewise.
611 (ssadd<mode>3): Likewise.
612 (sub<mode>3): Likewise.
613 (sub<mode>3): Likewise.
614 (ussub<mode>3): Likewise.
615 (sssub<mode>3): Likewise.
616 (ssmulsa3): Likewise.
617 (usmulusa3): Likewise.
618 (arm_usatsihi): Likewise.
619 * config/arm/vfp.md
620 (*movdi_vfp): Add types for all instructions.
621 (*movdi_vfp_cortexa8): Likewise.
622 (*movhf_vfp_neon): Likewise.
623 (*movhf_vfp): Likewise.
624 (*movdf_vfp): Likewise.
625 (*thumb2_movdf_vfp): Likewise.
626 (*thumb2_movdfcc_vfp): Likewise.
627 * config/arm/arm.md: Add type attribute to all insn patterns.
628 (*thumb1_adddi3): Add type attribute.
629 (*arm_adddi3): Likewise.
630 (*adddi_sesidi_di): Likewise.
631 (*adddi_zesidi_di): Likewise.
632 (*thumb1_addsi3): Likewise.
633 (addsi3_compare0): Likewise.
634 (*addsi3_compare0_scratch): Likewise.
635 (*compare_negsi_si): Likewise.
636 (cmpsi2_addneg): Likewise.
637 (*addsi3_carryin_<optab>): Likewise.
638 (*addsi3_carryin_alt2_<optab>): Likewise.
639 (*addsi3_carryin_clobercc_<optab>): Likewise.
640 (*subsi3_carryin): Likewise.
641 (*subsi3_carryin_const): Likewise.
642 (*subsi3_carryin_compare): Likewise.
643 (*subsi3_carryin_compare_const): Likewise.
644 (*arm_subdi3): Likewise.
645 (*thumb_subdi3): Likewise.
646 (*subdi_di_zesidi): Likewise.
647 (*subdi_di_sesidi): Likewise.
648 (*subdi_zesidi_di): Likewise.
649 (*subdi_sesidi_di): Likewise.
650 (*subdi_zesidi_ze): Likewise.
651 (thumb1_subsi3_insn): Likewise.
652 (*arm_subsi3_insn): Likewise.
653 (*anddi3_insn): Likewise.
654 (*anddi_zesidi_di): Likewise.
655 (*anddi_sesdi_di): Likewise.
656 (*ne_zeroextracts): Likewise.
657 (*ne_zeroextracts): Likewise.
658 (*ite_ne_zeroextr): Likewise.
659 (*ite_ne_zeroextr): Likewise.
660 (*anddi_notdi_di): Likewise.
661 (*anddi_notzesidi): Likewise.
662 (*anddi_notsesidi): Likewise.
663 (andsi_notsi_si): Likewise.
664 (thumb1_bicsi3): Likewise.
665 (*iordi3_insn): Likewise.
666 (*iordi_zesidi_di): Likewise.
667 (*iordi_sesidi_di): Likewise.
668 (*thumb1_iorsi3_insn): Likewise.
669 (*xordi3_insn): Likewise.
670 (*xordi_zesidi_di): Likewise.
671 (*xordi_sesidi_di): Likewise.
672 (*arm_xorsi3): Likewise.
673 (*andsi_iorsi3_no): Likewise.
674 (*smax_0): Likewise.
675 (*smax_m1): Likewise.
676 (*arm_smax_insn): Likewise.
677 (*smin_0): Likewise.
678 (*arm_smin_insn): Likewise.
679 (*arm_umaxsi3): Likewise.
680 (*arm_uminsi3): Likewise.
681 (*minmax_arithsi): Likewise.
682 (*minmax_arithsi_): Likewise.
683 (*satsi_<SAT:code>): Likewise.
684 (arm_ashldi3_1bit): Likewise.
685 (arm_ashrdi3_1bit): Likewise.
686 (arm_lshrdi3_1bit): Likewise.
687 (*arm_negdi2): Likewise.
688 (*thumb1_negdi2): Likewise.
689 (*arm_negsi2): Likewise.
690 (*thumb1_negsi2): Likewise.
691 (*negdi_extendsid): Likewise.
692 (*negdi_zero_extend): Likewise.
693 (*arm_abssi2): Likewise.
694 (*thumb1_abssi2): Likewise.
695 (*arm_neg_abssi2): Likewise.
696 (*thumb1_neg_abss): Likewise.
697 (one_cmpldi2): Likewise.
698 (extend<mode>di2): Likewise.
699 (*compareqi_eq0): Likewise.
700 (*arm_extendhisi2addsi): Likewise.
701 (*arm_movdi): Likewise.
702 (*thumb1_movdi_insn): Likewise.
703 (*arm_movt): Likewise.
704 (*thumb1_movsi_insn): Likewise.
705 (pic_add_dot_plus_four): Likewise.
706 (pic_add_dot_plus_eight): Likewise.
707 (tls_load_dot_plus_eight): Likewise.
708 (*thumb1_movhi_insn): Likewise.
709 (*thumb1_movsf_insn): Likewise.
710 (*movdf_soft_insn): Likewise.
711 (*thumb_movdf_insn): Likewise.
712 (cbranchsi4_insn): Likewise.
713 (cbranchsi4_scratch): Likewise.
714 (*negated_cbranchsi4): Likewise.
715 (*tbit_cbranch): Likewise.
716 (*tlobits_cbranch): Likewise.
717 (*tstsi3_cbranch): Likewise.
718 (*cbranchne_decr1): Likewise.
719 (*addsi3_cbranch): Likewise.
720 (*addsi3_cbranch_scratch): Likewise.
721 (*arm_cmpdi_insn): Likewise.
722 (*arm_cmpdi_unsig): Likewise.
723 (*arm_cmpdi_zero): Likewise.
724 (*thumb_cmpdi_zero): Likewise.
725 (*deleted_compare): Likewise.
726 (*mov_scc): Likewise.
727 (*mov_negscc): Likewise.
728 (*mov_notscc): Likewise.
729 (*cstoresi_eq0_thumb1_insn): Likewise.
730 (cstoresi_nltu_thumb1): Likewise.
731 (cstoresi_ltu_thu): Likewise.
732 (thumb1_addsi3_addgeu): Likewise.
733 (*arm_jump): Likewise.
734 (*thumb_jump): Likewise.
735 (*check_arch2): Likewise.
736 (arm_casesi_internal): Likewise.
737 (thumb1_casesi_dispatch): Likewise.
738 (*arm_indirect_jump): Likewise.
739 (*thumb1_indirect_jump): Likewise.
740 (nop): Likewise.
741 (*and_scc): Likewise.
742 (*ior_scc): Likewise.
743 (*compare_scc): Likewise.
744 (*cond_move): Likewise.
745 (*cond_arith): Likewise.
746 (*cond_sub): Likewise.
747 (*cmp_ite0): Likewise.
748 (*cmp_ite1): Likewise.
749 (*cmp_and): Likewise.
750 (*cmp_ior): Likewise.
751 (*ior_scc_scc): Likewise.
752 (*ior_scc_scc_cmp): Likewise.
753 (*and_scc_scc): Likewise.
754 (*and_scc_scc_cmp): Likewise.
755 (*and_scc_scc_nod): Likewise.
756 (*negscc): Likewise.
757 (movcond_addsi): Likewise.
758 (movcond): Likewise.
759 (*ifcompare_plus_move): Likewise.
760 (*if_plus_move): Likewise.
761 (*ifcompare_move_plus): Likewise.
762 (*if_move_plus): Likewise.
763 (*ifcompare_arith_arith): Likewise.
764 (*if_arith_arith): Likewise.
765 (*ifcompare_arith_move): Likewise.
766 (*if_arith_move): Likewise.
767 (*ifcompare_move_arith): Likewise.
768 (*if_move_arith): Likewise.
769 (*ifcompare_move_not): Likewise.
770 (*if_move_not): Likewise.
771 (*ifcompare_not_move): Likewise.
772 (*if_not_move): Likewise.
773 (*ifcompare_shift_move): Likewise.
774 (*if_shift_move): Likewise.
775 (*ifcompare_move_shift): Likewise.
776 (*if_move_shift): Likewise.
777 (*ifcompare_shift_shift): Likewise.
778 (*ifcompare_not_arith): Likewise.
779 (*ifcompare_arith_not): Likewise.
780 (*if_arith_not): Likewise.
781 (*ifcompare_neg_move): Likewise.
782 (*if_neg_move): Likewise.
783 (*ifcompare_move_neg): Likewise.
784 (*if_move_neg): Likewise.
785 (prologue_thumb1_interwork): Likewise.
786 (*cond_move_not): Likewise.
787 (*sign_extract_onebit): Likewise.
788 (*not_signextract_onebit): Likewise.
789 (stack_tie): Likewise.
790 (align_4): Likewise.
791 (align_8): Likewise.
792 (consttable_end): Likewise.
793 (consttable_1): Likewise.
794 (consttable_2): Likewise.
795 (consttable_4): Likewise.
796 (consttable_8): Likewise.
797 (consttable_16): Likewise.
798 (*thumb1_tablejump): Likewise.
799 (prefetch): Likewise.
800 (force_register_use): Likewise.
801 (thumb_eh_return): Likewise.
802 (load_tp_hard): Likewise.
803 (load_tp_soft): Likewise.
804 (tlscall): Likewise.
805 (*arm_movtas_ze): Likewise.
806 (*arm_rev): Likewise.
807 (*arm_revsh): Likewise.
808 (*arm_rev16): Likewise.
809 * config/arm/thumb2.md
810 (*thumb2_smaxsi3): Likewise.
811 (*thumb2_sminsi3): Likewise.
812 (*thumb32_umaxsi3): Likewise.
813 (*thumb2_uminsi3): Likewise.
814 (*thumb2_negdi2): Likewise.
815 (*thumb2_abssi2): Likewise.
816 (*thumb2_neg_abss): Likewise.
817 (*thumb2_movsi_insn): Likewise.
818 (tls_load_dot_plus_four): Likewise.
819 (*thumb2_movhi_insn): Likewise.
820 (*thumb2_mov_scc): Likewise.
821 (*thumb2_mov_negs): Likewise.
822 (*thumb2_mov_negs): Likewise.
823 (*thumb2_mov_nots): Likewise.
824 (*thumb2_mov_nots): Likewise.
825 (*thumb2_movsicc_): Likewise.
826 (*thumb2_movsfcc_soft_insn): Likewise.
827 (*thumb2_indirect_jump): Likewise.
828 (*thumb2_and_scc): Likewise.
829 (*thumb2_ior_scc): Likewise.
830 (*thumb2_ior_scc_strict_it): Likewise.
831 (*thumb2_cond_move): Likewise.
832 (*thumb2_cond_arith): Likewise.
833 (*thumb2_cond_ari): Likewise.
834 (*thumb2_cond_sub): Likewise.
835 (*thumb2_negscc): Likewise.
836 (*thumb2_movcond): Likewise.
837 (thumb2_casesi_internal): Likewise.
838 (thumb2_casesi_internal_pic): Likewise.
839 (*thumb2_alusi3_short): Likewise.
840 (*thumb2_mov<mode>_shortim): Likewise.
841 (*thumb2_addsi_short): Likewise.
842 (*thumb2_subsi_short): Likewise.
843 (thumb2_addsi3_compare0): Likewise.
844 (*thumb2_cbz): Likewise.
845 (*thumb2_cbnz): Likewise.
846 (*thumb2_one_cmplsi2_short): Likewise.
847 (*thumb2_negsi2_short): Likewise.
848 (*orsi_notsi_si): Likewise.
849 * config/arm/arm1020e.md: Update with new attributes.
850 * config/arm/arm1026ejs.md: Update with new attributes.
851 * config/arm/arm1136jfs.md: Update with new attributes.
852 * config/arm/arm926ejs.md: Update with new attributes.
853 * config/arm/cortex-a15.md: Update with new attributes.
854 * config/arm/cortex-a5.md: Update with new attributes.
855 * config/arm/cortex-a53.md: Update with new attributes.
856 * config/arm/cortex-a7.md: Update with new attributes.
857 * config/arm/cortex-a8.md: Update with new attributes.
858 * config/arm/cortex-a9.md: Update with new attributes.
859 * config/arm/cortex-m4.md: Update with new attributes.
860 * config/arm/cortex-r4.md: Update with new attributes.
861 * config/arm/fa526.md: Update with new attributes.
862 * config/arm/fa606te.md: Update with new attributes.
863 * config/arm/fa626te.md: Update with new attributes.
864 * config/arm/fa726te.md: Update with new attributes.
865
1c83b673
JG
8662013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
867
868 * config/aarch64/aarch64-simd.md
869 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
870 <vwx> iterator to ensure correct register choice.
871 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
872 (aarch64_sqdmull_n<mode>): Likewise.
873 (aarch64_sqdmull2_n<mode>_internal): Likewise.
874 * config/aarch64/arm_neon.h
875 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
876 (vml<as><q>_n_<su>16): Likewise.
877 (vml<as>l_high_lane<q>_<su>16): Likewise.
878 (vml<as>l_high_n_<su>16): Likewise.
879 (vml<as>l_lane<q>_<su>16): Likewise.
880 (vml<as>l_n_<su>16): Likewise.
881 (vmul<q>_lane<q>_<su>16): Likewise.
882 (vmul<q>_n_<su>16): Likewise.
883 (vmull_lane<q>_<su>16): Likewise.
884 (vmull_n_<su>16): Likewise.
885 (vmull_high_lane<q>_<su>16): Likewise.
886 (vmull_high_n_<su>16): Likewise.
887 (vqrdmulh<q>_n_s16): Likewise.
888
bb1ae543
TB
8892013-09-06 Tejas Belagod <tejas.belagod@arm.com>
890
891 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
892 have the correct lane parameter.
893
f23c0742
RB
8942013-09-06 Richard Biener <rguenther@suse.de>
895
896 * cfganal.c (control_dependences::~control_dependences):
897 Properly free all of the vector.
898
c1d49770
KY
8992013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
900
901 PR target/58269
902 * config/i386/i386.c (ix86_conditional_register_usage):
903 Proper initialize extended SSE registers.
904
6da47f52
JH
9052013-09-06 Jan Hubicka <jh@suse.cz>
906
907 PR tree-optimization/58311
908 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
909
9102013-09-06 Jan Hubicka <jh@suse.cz>
911
912 * Makefile.in (tree-sra.o): Update dependencies.
913 * tree-sra.c: Include ipa-utils.h
914 (scan_function): Use recursive_call_p.
915 (has_caller_p): New function.
916 (cgraph_for_node_and_aliases): Count also callers of aliases.
917
fc11f321
JH
9182013-09-06 Jan Hubicka <jh@suse.cz>
919
920 PR middle-end/58094
921 * cgraph.h (symtab_semantically_equivalent_p): Declare.
922 * tree-tailcall.c: Include ipa-utils.h.
923 (find_tail_calls): Use it.
924 * ipa-pure-const.c (check_call): Likewise.
925 * ipa-utils.c (recursive_call_p): New function.
926 * ipa-utils.h (recursive_call_p): Dclare.
927 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
928 (symtab_semantically_equivalent_p): New function.
929 * Makefile.in (tree-tailcall.o): Update dependencies.
930
9a6af450
EB
9312013-09-06 Eric Botcazou <ebotcazou@adacore.com>
932
933 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
934 non-inlinable part.
935
35f32ad4
RB
9362013-09-06 Richard Biener <rguenther@suse.de>
937
938 * lto-streamer.h (lto_global_var_decls): Remove.
939 * Makefile.in (OBJS): Remove lto-symtab.o.
940 (lto-symtab.o): Remove.
941 (GTFILES): Remove lto-symtab.c
942 * lto-symtab.c: Move to lto/
943
d12a76f3
AK
9442013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
945
946 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
947 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
948 (UNSPEC_FPINT_RINT): New constant definitions.
949 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
950 definition with 2 attributes.
951 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
952 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
953 definitions.
954
9552013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
956
f43245d1 957 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK
958 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
959 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
960 attribute to "z196_alone".
961 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
962 "zEC12_simple".
963
c8e9d8c3
RB
9642013-09-06 Richard Biener <rguenther@suse.de>
965
966 * basic-block.h (class control_dependences): New.
967 * tree-ssa-dce.c (control_dependence_map): Remove.
968 (cd): New global.
969 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
970 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
971 find_pdom, find_control_dependence, find_all_control_dependences):
972 Move to cfganal.c.
f43245d1
UB
973 (mark_control_dependent_edges_necessary,
974 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
975 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB
976 * cfganal.c (set_control_dependence_map_bit,
977 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
978 find_all_control_dependences): Move from tree-ssa-dce.c and
979 implement as methods of control_dependences class.
980 (control_dependences::control_dependences): New.
981 (control_dependences::~control_dependences): Likewise.
982 (control_dependences::get_edges_dependent_on): Likewise.
983 (control_dependences::get_edge): Likewise.
984
7d2ba471
JH
9852013-09-04 Jan Hubicka <jh@suse.cz>
986
987 * tree.c (types_same_for_odr): Drop overactive check.
988 * ipa-devirt.c (hash_type_name): Likewise.
989
e18412fc
JH
9902013-09-04 Jan Hubicka <jh@suse.cz>
991
992 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
993 (analyze_functions): ... here.
994
d352b245
JH
9952013-09-04 Jan Hubicka <jh@suse.cz>
996
997 PR middle-end/58201
998 * cgraphunit.c (analyze_functions): Clear AUX fields
999 after processing; initialize assembler name has.
1000
770da076
JL
10012013-09-05 Jeff Law <law@redhat.com>
1002
1003 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
1004 from thread_around_empty_block. Record threading path into PATH.
1005 Recurse if threading through the initial block is successful.
1006 (thread_across_edge): Corresponding changes to slightly simplify.
1007
fe6f68e2
JG
10082013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1009
1010 * config/aarch64/aarch64.md
1011 (type): Remove frecpe, frecps, frecpx.
1012 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
1013 fix to be a TARGET_SIMD instruction.
1014 (aarch64_frecps): Remove.
1015 * config/aarch64/aarch64-simd.md
1016 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
1017 (aarch64_frecps<mode>): Handle all float/vector of float modes.
1018
6e4150e1
JG
10192013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1020 Sofiane Naci <sofiane.naci@arm.com>
1021
f43245d1 1022 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 1023 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB
1024 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
1025 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
1026 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
1027 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
1028 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
1029 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
1030 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 1031 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 1032 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG
1033 shift* types.
1034 * config/arm/arm-fixed.md: Update for attribute changes
1035 all occurrences of arlo_* types.
1036 * config/arm/thumb2.md: Update for attribute changes all occurrences
1037 of arlo_* types.
1038 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
1039 (cortexa7_older_only): Likewise.
1040 (cortexa7_younger): Likewise.
1041 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
1042 (1020alu_shift_op): Likewise.
1043 (1020alu_shift_reg_op): Likewise.
1044 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
1045 (alu_shift_op): Likewise.
1046 (alu_shift_reg_op): Likewise.
f43245d1 1047 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG
1048 (11_alu_shift_op): Likewise.
1049 (11_alu_shift_reg_op): Likewise.
1050 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
1051 (9_alu_shift_reg_op): Likewise.
1052 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
1053 attribute changes.
1054 (cortex_a15_alu_shift): Likewise.
1055 (cortex_a15_alu_shift_reg): Likewise.
1056 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
1057 attribute changes.
1058 (cortex_a5_alu_shift): Likewise.
f43245d1
UB
1059 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
1060 attribute changes.
6e4150e1 1061 (cortex_a53_alu_shift): Likewise.
f43245d1
UB
1062 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
1063 attribute changes.
6e4150e1
JG
1064 (cortex_a7_alu_reg): Likewise.
1065 (cortex_a7_alu_shift): Likewise.
f43245d1
UB
1066 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
1067 attribute changes.
6e4150e1
JG
1068 (cortex_a8_alu_shift): Likewise.
1069 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 1070 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 1071 (cortex_a9_dp_shift): Likewise.
f43245d1
UB
1072 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
1073 attribute changes.
6e4150e1
JG
1074 * config/arm/cortex-r4.md
1075 (cortex_r4_alu): Update for attribute changes.
1076 (cortex_r4_mov): Likewise.
1077 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 1078 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 1079 (526_alu_shift_op): Likewise.
f43245d1
UB
1080 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
1081 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 1082 (626te_alu_shift_op): Likewise.
f43245d1 1083 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG
1084 (726te_alu_shift_op): Likewise.
1085 (726te_alu_shift_reg_op): Likewise.
1086 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
1087 (mp626_alu_shift_op): Likewise.
1088 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
1089 (pj4_alu_conds): Likewise.
1090 (pj4_shift): Likewise.
1091 (pj4_shift_conds): Likewise.
1092 (pj4_alu_shift): Likewise.
1093 (pj4_alu_shift_conds): Likewise.
1094 * config/aarch64/aarch64.md: Update for attribute change
1095 all occurrences of arlo_* and shift* types.
1096
e6523306
MS
10972013-09-05 Mike Stump <mikestump@comcast.net>
1098
1099 * tree.h: Move documentation for tree_function_decl to tree-core.h
1100 with the declaration.
1101
80a18161
PB
11022013-09-05 Peter Bergner <bergner@vnet.ibm.com>
1103
1104 PR target/58139
1105 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
1106 looking for widest mode.
1107
6994430a
EB
11082013-09-05 Eric Botcazou <ebotcazou@adacore.com>
1109
1110 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
1111
e67f39f7
RB
11122013-09-05 Richard Biener <rguenther@suse.de>
1113
1114 PR tree-optimization/58137
1115 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
1116 Do not create vectors of pointers.
1117 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
1118 types for the components of the vector initializer.
1119 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
1120 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
1121
568cda29
MJ
11222013-09-05 Martin Jambor <mjambor@suse.cz>
1123
1124 * ipa-prop.c (remove_described_reference): Accept missing references,
1125 return false if that hppens, otherwise return true.
1126 (cgraph_node_for_jfunc): New function.
1127 (try_decrement_rdesc_refcount): Likewise.
1128 (try_make_edge_direct_simple_call): Use them.
1129 (ipa_edge_removal_hook): Remove references from rdescs.
1130 (ipa_edge_duplication_hook): Clone rdescs and their references
1131 when the new edge has the same caller as the old one.
1132 * cgraph.c (cgraph_resolve_speculation): Remove speculative
1133 reference before removing any edges.
1134
5a200acb
RE
11352013-09-05 Richard Earnshaw <rearnsha@arm.com>
1136
1137 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
1138 initial store.
1139 * thumb2.md (thumb2_storewb_parisi): New pattern.
1140
5922847b
YZ
11412013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
1142
1143 * config/aarch64/aarch64-option-extensions.def: Add
1144 AARCH64_OPT_EXTENSION of 'crc'.
1145 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
1146 (AARCH64_ISA_CRC): Ditto.
1147 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
1148 description of the CRC extension.
1149
996746aa
AI
11502013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
1151
1152 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
1153 * config/rs6000/linux.h: Ditto.
1154 * alpha/linux.h: Ditto.
1155 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
1156 no_c99_libc_has_function.
1157 * config/c6x/uclinux-elf.h: Ditto.
1158 * config/lm32/uclinux-elf.h: Ditto.
1159 * config/m68k/uclinux.h: Ditto.
1160 * config/moxie/uclinux.h: Ditto.
1161 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
1162 (crisv32-*-linux*, cris-*-linux*): Ditto.
1163 * config/bfin/bfin.c: Include "tm_p.h".
1164
c370d99f
RB
11652013-09-05 Richard Biener <rguenther@suse.de>
1166
1167 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
1168 check for a definition without a basic-block.
1169
003bb7f3
JG
11702013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1171 Sofiane Naci <sofiane.naci@arm.com>
1172
1173 * config/aarch64/aarch64.md
1174 (*movti_aarch64): Rename r_2_f and f_2_r.
1175 (*movsf_aarch64): Likewise.
1176 (*movdf_aarch64): Likewise.
1177 (*movtf_aarch64): Likewise.
1178 (aarch64_movdi_<mode>low): Likewise.
1179 (aarch64_movdi_<mode>high): Likewise.
1180 (aarch64_mov<mode>high_di): Likewise.
1181 (aarch64_mov<mode>low_di): Likewise.
1182 (aarch64_movtilow_tilow): Likewise.
1183 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
1184 values to config/arm/types.md
1185 (attribute "conds"): Update for attribute change.
1186 (anddi3_insn): Likewise.
1187 (iordi3_insn): Likewise.
1188 (xordi3_insn): Likewise.
1189 (one_cmpldi2): Likewise.
1190 * config/arm/types.md (type): Add Neon types.
1191 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
1192 use "type" attribute.
1193 (movmisalign<mode>_neon_store): Likewise.
1194 (movmisalign<mode>_neon_load): Likewise.
1195 (vec_set<mode>_internal): Likewise.
1196 (vec_setv2di_internal): Likewise.
1197 (vec_extract<mode>): Likewise.
1198 (vec_extractv2di): Likewise.
1199 (add<mode>3_neon): Likewise.
1200 (adddi3_neon): Likewise.
1201 (sub<mode>3_neon): Likewise.
1202 (subdi3_neon): Likewise.
1203 (mul<mode>3_neon): Likewise.
1204 (mul<mode>3add<mode>_neon): Likewise.
1205 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
1206 (fma<VCVTF:mode>4)): Likewise.
1207 (fma<VCVTF:mode>4_intrinsic): Likewise.
1208 (fmsub<VCVTF:mode>4)): Likewise.
1209 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1210 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1211 (ior<mode>3): Likewise.
1212 (and<mode>3): Likewise.
1213 (anddi3_neon): Likewise.
1214 (orn<mode>3_neon): Likewise.
1215 (orndi3_neon): Likewise.
1216 (bic<mode>3_neon): Likewise.
1217 (bicdi3_neon): Likewise.
1218 (xor<mode>3): Likewise.
1219 (one_cmpl<mode>2): Likewise.
1220 (abs<mode>2): Likewise.
1221 (neg<mode>2): Likewise.
1222 (umin<mode>3_neon): Likewise.
1223 (umax<mode>3_neon): Likewise.
1224 (smin<mode>3_neon): Likewise.
1225 (smax<mode>3_neon): Likewise.
1226 (vashl<mode>3): Likewise.
1227 (vashr<mode>3_imm): Likewise.
1228 (vlshr<mode>3_imm): Likewise.
1229 (ashl<mode>3_signed): Likewise.
1230 (ashl<mode>3_unsigned): Likewise.
1231 (neon_load_count): Likewise.
1232 (ashldi3_neon_noclobber): Likewise.
1233 (signed_shift_di3_neon): Likewise.
1234 (unsigned_shift_di3_neon): Likewise.
1235 (ashrdi3_neon_imm_noclobber): Likewise.
1236 (lshrdi3_neon_imm_noclobber): Likewise.
1237 (widen_ssum<mode>3): Likewise.
1238 (widen_usum<mode>3): Likewise.
1239 (quad_halves_<code>v4si): Likewise.
1240 (quad_halves_<code>v4sf): Likewise.
1241 (quad_halves_<code>v8hi): Likewise.
1242 (quad_halves_<code>v16qi): Likewise.
1243 (reduc_splus_v2di): Likewise.
1244 (neon_vpadd_internal<mode>): Likewise.
1245 (neon_vpsmin<mode>): Likewise.
1246 (neon_vpsmax<mode>): Likewise.
1247 (neon_vpumin<mode>): Likewise.
1248 (neon_vpumax<mode>): Likewise.
1249 (ss_add<mode>_neon): Likewise.
1250 (us_add<mode>_neon): Likewise.
1251 (ss_sub<mode>_neon): Likewise.
1252 (us_sub<mode>_neon): Likewise.
1253 (neon_vadd<mode>_unspec): Likewise.
1254 (neon_vaddl<mode>): Likewise.
1255 (neon_vaddw<mode>): Likewise.
1256 (neon_vhadd<mode>): Likewise.
1257 (neon_vqadd<mode>): Likewise.
1258 (neon_vaddhn<mode>): Likewise.
1259 (neon_vmul<mode>): Likewise.
1260 (neon_vmla<mode>): Likewise.
1261 (neon_vmlal<mode>): Likewise.
1262 (neon_vmls<mode>): Likewise.
1263 (neon_vmlsl<mode>): Likewise.
1264 (neon_vqdmulh<mode>): Likewise.
1265 (neon_vqdmlal<mode>): Likewise.
1266 (neon_vqdmlsl<mode>): Likewise.
1267 (neon_vmull<mode>): Likewise.
1268 (neon_vqdmull<mode>): Likewise.
1269 (neon_vsub<mode>_unspec): Likewise.
1270 (neon_vsubl<mode>): Likewise.
1271 (neon_vsubw<mode>): Likewise.
1272 (neon_vqsub<mode>): Likewise.
1273 (neon_vhsub<mode>): Likewise.
1274 (neon_vsubhn<mode>): Likewise.
1275 (neon_vceq<mode>): Likewise.
1276 (neon_vcge<mode>): Likewise.
1277 (neon_vcgeu<mode>): Likewise.
1278 (neon_vcgt<mode>): Likewise.
1279 (neon_vcgtu<mode>): Likewise.
1280 (neon_vcle<mode>): Likewise.
1281 (neon_vclt<mode>): Likewise.
1282 (neon_vcage<mode>): Likewise.
1283 (neon_vcagt<mode>): Likewise.
1284 (neon_vtst<mode>): Likewise.
1285 (neon_vabd<mode>): Likewise.
1286 (neon_vabdl<mode>): Likewise.
1287 (neon_vaba<mode>): Likewise.
1288 (neon_vabal<mode>): Likewise.
1289 (neon_vmax<mode>): Likewise.
1290 (neon_vmin<mode>): Likewise.
1291 (neon_vpaddl<mode>): Likewise.
1292 (neon_vpadal<mode>): Likewise.
1293 (neon_vpmax<mode>): Likewise.
1294 (neon_vpmin<mode>): Likewise.
1295 (neon_vrecps<mode>): Likewise.
1296 (neon_vrsqrts<mode>): Likewise.
1297 (neon_vqabs<mode>): Likewise.
1298 (neon_vqneg<mode>): Likewise.
1299 (neon_vcls<mode>): Likewise.
1300 (clz<mode>2): Likewise.
1301 (popcount<mode>2): Likewise.
1302 (neon_vrecpe): Likewise.
1303 (neon_vrsqrte): Likewise.
1304 (neon_vget_lane<mode>_sext_internal): Likewise.
1305 (neon_vget_lane<mode>_zext_internal): Likewise.
1306 (neon_vdup_n<mode>): Likewise.
1307 (neon_vdup_nv2di): Likewise.
1308 (neon_vdpu_lane<mode>_internal): Likewise.
1309 (neon_vswp<mode>): Likewise.
1310 (float<mode><V_cvtto>2): Likewise.
1311 (floatuns<mode><V_cvtto>2): Likewise.
1312 (fix_trunc<mode><V_cvtto>)2): Likewise
1313 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1314 (neon_vcvt<mode>): Likewise.
1315 (neon_vcvtv4sfv4hf): Likewise.
1316 (neon_vcvtv4hfv4sf): Likewise.
1317 (neon_vcvt_n<mode>): Likewise.
1318 (neon_vmovn<mode>): Likewise.
1319 (neon_vqmovn<mode>): Likewise.
1320 (neon_vqmovun<mode>): Likewise.
1321 (neon_vmovl<mode>): Likewise.
1322 (neon_vmul_lane<mode>): Likewise.
1323 (neon_vmull_lane<mode>): Likewise.
1324 (neon_vqdmull_lane<mode>): Likewise.
1325 (neon_vqdmulh_lane<mode>): Likewise.
1326 (neon_vmla_lane<mode>): Likewise.
1327 (neon_vmlal_lane<mode>): Likewise.
1328 (neon_vqdmlal_lane<mode>): Likewise.
1329 (neon_vmls_lane<mode>): Likewise.
1330 (neon_vmlsl_lane<mode>): Likewise.
1331 (neon_vqdmlsl_lane<mode>): Likewise.
1332 (neon_vext<mode>): Likewise.
1333 (neon_vrev64<mode>): Likewise.
1334 (neon_vrev32<mode>): Likewise.
1335 (neon_vrev16<mode>): Likewise.
1336 (neon_vbsl<mode>_internal): Likewise.
1337 (neon_vshl<mode>): Likewise.
1338 (neon_vqshl<mode>): Likewise.
1339 (neon_vshr_n<mode>): Likewise.
1340 (neon_vshrn_n<mode>): Likewise.
1341 (neon_vqshrn_n<mode>): Likewise.
1342 (neon_vqshrun_n<mode>): Likewise.
1343 (neon_vshl_n<mode>): Likewise.
1344 (neon_vqshl_n<mode>): Likewise.
1345 (neon_vqshlu_n<mode>): Likewise.
1346 (neon_vshll_n<mode>): Likewise.
1347 (neon_vsra_n<mode>): Likewise.
1348 (neon_vsri_n<mode>): Likewise.
1349 (neon_vsli_n<mode>): Likewise.
1350 (neon_vtbl1v8qi): Likewise.
1351 (neon_vtbl2v8qi): Likewise.
1352 (neon_vtbl3v8qi): Likewise.
1353 (neon_vtbl4v8qi): Likewise.
1354 (neon_vtbx1v8qi): Likewise.
1355 (neon_vtbx2v8qi): Likewise.
1356 (neon_vtbx3v8qi): Likewise.
1357 (neon_vtbx4v8qi): Likewise.
1358 (neon_vtrn<mode>_internal): Likewise.
1359 (neon_vzip<mode>_internal): Likewise.
1360 (neon_vuzp<mode>_internal): Likewise.
1361 (neon_vld1<mode>): Likewise.
1362 (neon_vld1_lane<mode>): Likewise.
1363 (neon_vld1_dup<mode>): Likewise.
1364 (neon_vld1_dupv2di): Likewise.
1365 (neon_vst1<mode>): Likewise.
1366 (neon_vst1_lane<mode>): Likewise.
1367 (neon_vld2<mode>): Likewise.
1368 (neon_vld2_lane<mode>): Likewise.
1369 (neon_vld2_dup<mode>): Likewise.
1370 (neon_vst2<mode>): Likewise.
1371 (neon_vst2_lane<mode>): Likewise.
1372 (neon_vld3<mode>): Likewise.
1373 (neon_vld3qa<mode>): Likewise.
1374 (neon_vld3qb<mode>): Likewise.
1375 (neon_vld3_lane<mode>): Likewise.
1376 (neon_vld3_dup<mode>): Likewise.
1377 (neon_vst3<mode>): Likewise.
1378 (neon_vst3qa<mode>): Likewise.
1379 (neon_vst3qb<mode>): Likewise.
1380 (neon_vst3_lane<mode>): Likewise.
1381 (neon_vld4<mode>): Likewise.
1382 (neon_vld4qa<mode>): Likewise.
1383 (neon_vld4qb<mode>): Likewise.
1384 (neon_vld4_lane<mode>): Likewise.
1385 (neon_vld4_dup<mode>): Likewise.
1386 (neon_vst4<mode>): Likewise.
1387 (neon_vst4qa<mode>): Likewise.
1388 (neon_vst4qb<mode>): Likewise.
1389 (neon_vst4_lane<mode>): Likewise.
1390 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1391 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1392 (neon_vec_<US>mult_lo_<mode>): Likewise.
1393 (neon_vec_<US>mult_hi_<mode>): Likewise.
1394 (neon_vec_<US>shiftl_<mode>): Likewise.
1395 (neon_unpack<US>_<mode>): Likewise.
1396 (neon_vec_<US>mult_<mode>): Likewise.
1397 (vec_pack_trunc_<mode>): Likewise.
1398 (neon_vec_pack_trunk_<mode>): Likewise.
1399 (neon_vabd<mode>_2): Likewise.
1400 (neon_vabd<mode>_3): Likewise.
1401 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1402 (thumb2_movsi_vfp): Likewise.
1403 (movdi_vfp): Likewise.
1404 (movdi_vfp_cortexa8): Likewise.
1405 (movhf_vfp_neon): Likewise.
1406 (movhf_vfp): Likewiwse.
1407 (movsf_vfp): Likewiwse.
1408 (thumb2_movsf_vfp): Likewiwse.
1409 (movdf_vfp): Likewise.
1410 (thumb2_movdf_vfp): Likewise.
1411 (movsfcc_vfp): Likewise.
1412 (thumb2_movsfcc_vfp): Likewise.
1413 (movdfcc_vfp): Likewise.
1414 (thumb2_movdfcc_vfp): Likewise.
1415 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1416 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1417 (v10_v2c): Likewise.
1418 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1419 attribute change.
1420 (cortex_a15_neon_int_2): Likewise.
1421 (cortex_a15_neon_int_3): Likewise.
1422 (cortex_a15_neon_int_4): Likewise.
1423 (cortex_a15_neon_int_5): Likewise.
1424 (cortex_a15_neon_vqneg_vqabs): Likewise.
1425 (cortex_a15_neon_vmov): Likewise.
1426 (cortex_a15_neon_vaba): Likewise.
1427 (cortex_a15_neon_vaba_qqq): Likewise.
1428 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1429 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1430 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1431 scalar_64_32_long_scalar): Likewise.
1432 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1433 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1434 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1435 lotype_qdd_64_32_long): Likewise.
1436 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1437 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1438 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1439 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1440 (cortex_a15_neon_shift_1): Likewise.
1441 (cortex_a15_neon_shift_2): Likewise.
1442 (cortex_a15_neon_shift_3): Likewise.
1443 (cortex_a15_neon_vshl_ddd): Likewise.
1444 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1445 (cortex_a15_neon_vsra_vrsra): Likewise.
1446 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1447 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1448 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1449 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1450 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1451 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1452 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1453 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1454 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1455 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1456 (cortex_a15_neon_bp_simple): Likewise.
1457 (cortex_a15_neon_bp_2cycle): Likewise.
1458 (cortex_a15_neon_bp_3cycle): Likewise.
1459 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1460 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1461 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1462 (cortex_a15_neon_vld2_4_regs): Likewise.
1463 (cortex_a15_neon_vld3_vld4): Likewise.
1464 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1465 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1466 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1467 (cortex_a15_neon_vst3_vst4): Likewise.
1468 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1469 (cortex_a15_neon_vld3_vld4_lane" 10
1470 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1471 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1472 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1473 (cortex_a15_neon_ldm_2): Likewise.0
1474 (cortex_a15_neon_stm_2): Likewise.
1475 (cortex_a15_neon_mcr): Likewise.
1476 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1477 (cortex_a15_neon_mrc): Likewise.
1478 (cortex_a15_neon_mrrc): Likewise.
1479 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1480 change.
1481 (cortex_a15_alu_shift): Likewise.
1482 (cortex_a15_alu_shift_reg): Likewise.
1483 (cortex_a15_mult32): Likewise.
1484 (cortex_a15_mult64): Likewise.
1485 (cortex_a15_block): Likewise.
1486 (cortex_a15_branch): Likewise.
1487 (cortex_a15_load1): Likewise.
1488 (cortex_a15_load3): Likewise.
1489 (cortex_a15_store1): Likewise.
1490 (cortex_a15_store3): Likewise.
1491 (cortex_a15_call): Likewise.
f43245d1 1492 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG
1493 (cortex_a5_f2r): Likewise.
1494 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1495 change.
1496 (cortex_a53_f2r): Likewise.
1497 * config/arm/cortex-a7.md
1498 (cortex_a7_branch): Update for attribute change.
1499 (cortex_a7_call): Likewise.
1500 (cortex_a7_alu_imm): Likewise.
1501 (cortex_a7_alu_reg): Likewise.
1502 (cortex_a7_alu_shift): Likewise.
1503 (cortex_a7_mul): Likewise.
1504 (cortex_a7_load1): Likewise.
1505 (cortex_a7_store1): Likewise.
1506 (cortex_a7_load2): Likewise.
1507 (cortex_a7_store2): Likewise.
1508 (cortex_a7_load3): Likewise.
1509 (cortex_a7_store3): Likewise.
1510 (cortex_a7_load4): Likewise.
1511 (cortex_a7_store4): Likewise.
1512 (cortex_a7_fpalu): Likewise.
1513 (cortex_a7_fconst): Likewise.
1514 (cortex_a7_fpmuls): Likewise.
1515 (cortex_a7_neon_mul): Likewise.
1516 (cortex_a7_fpmacs): Likewise.
1517 (cortex_a7_neon_mla: Likewise.
1518 (cortex_a7_fpmuld: Likewise.
1519 (cortex_a7_fpmacd: Likewise.
1520 (cortex_a7_fpfmad: Likewise.
1521 (cortex_a7_fdivs: Likewise.
1522 (cortex_a7_fdivd: Likewise.
1523 (cortex_a7_r2f: Likewise.
1524 (cortex_a7_f2r: Likewise.
1525 (cortex_a7_f_flags: Likewise.
1526 (cortex_a7_f_loads: Likewise.
1527 (cortex_a7_f_loadd: Likewise.
1528 (cortex_a7_f_stores: Likewise.
1529 (cortex_a7_f_stored: Likewise.
1530 (cortex_a7_neon): Likewise.
1531 * config/arm/cortex-a8-neon.md
1532 (cortex_a8_neon_mrc): Update for attribute change.
1533 (cortex_a8_neon_mrrc): Likewise.
1534 (cortex_a8_neon_int_1): Likewise.
1535 (cortex_a8_neon_int_2): Likewise.
1536 (cortex_a8_neon_int_3): Likewise.
1537 (cortex_a8_neon_int_4): Likewise.
1538 (cortex_a8_neon_int_5): Likewise.
1539 (cortex_a8_neon_vqneg_vqabs): Likewise.
1540 (cortex_a8_neon_vmov): Likewise.
1541 (cortex_a8_neon_vaba): Likewise.
1542 (cortex_a8_neon_vaba_qqq): Likewise.
1543 (cortex_a8_neon_vsma): Likewise.
1544 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1545 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB
1546 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1547 long_scalar): Likewise.
003bb7f3
JG
1548 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1549 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1550 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1551 long_scalar_qdd_64_32_long): Likewise.
1552 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1553 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1554 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1555 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1556 (cortex_a8_neon_shift_1): Likewise.
1557 (cortex_a8_neon_shift_2): Likewise.
1558 (cortex_a8_neon_shift_3): Likewise.
1559 (cortex_a8_neon_vshl_ddd): Likewise.
1560 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1561 (cortex_a8_neon_vsra_vrsra): Likewise.
1562 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1563 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1564 (cortex_a8_neon_fp_vsum): Likewise.
1565 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1566 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1567 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1568 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1569 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1570 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1571 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1572 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1573 (cortex_a8_neon_bp_simple): Likewise.
1574 (cortex_a8_neon_bp_2cycle): Likewise.
1575 (cortex_a8_neon_bp_3cycle): Likewise.
1576 (cortex_a8_neon_ldr): Likewise.
1577 (cortex_a8_neon_str): Likewise.
1578 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1579 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1580 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1581 (cortex_a8_neon_vld2_4_regs): Likewise.
1582 (cortex_a8_neon_vld3_vld4): Likewise.
1583 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1584 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1585 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1586 (cortex_a8_neon_vst3_vst4): Likewise.
1587 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1588 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1589 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1590 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1591 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1592 (cortex_a8_neon_mcr): Likewise.
1593 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 1594 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG
1595 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
1596 change.
1597 (ca9_neon_mrrc): Likewise.
1598 (cortex_a9_neon_int_1): Likewise.
1599 (cortex_a9_neon_int_2): Likewise.
1600 (cortex_a9_neon_int_3): Likewise.
1601 (cortex_a9_neon_int_4): Likewise.
1602 (cortex_a9_neon_int_5): Likewise.
1603 (cortex_a9_neon_vqneg_vqabs): Likewise.
1604 (cortex_a9_neon_vmov): Likewise.
1605 (cortex_a9_neon_vaba): Likewise.
1606 (cortex_a9_neon_vaba_qqq): Likewise.
1607 (cortex_a9_neon_vsma): Likewise.
1608 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1609 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB
1610 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1611 long_scalar): Likewise.
003bb7f3
JG
1612 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1613 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1614 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1615 long_scalar_qdd_64_32_long): Likewise.
1616 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1617 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1618 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1619 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1620 (cortex_a9_neon_shift_1): Likewise.
1621 (cortex_a9_neon_shift_2): Likewise.
1622 (cortex_a9_neon_shift_3): Likewise.
1623 (cortex_a9_neon_vshl_ddd): Likewise.
1624 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1625 (cortex_a9_neon_vsra_vrsra): Likewise.
1626 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1627 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1628 (cortex_a9_neon_fp_vsum): Likewise.
1629 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1630 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1631 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1632 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1633 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1634 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1635 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1636 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1637 (cortex_a9_neon_bp_simple): Likewise.
1638 (cortex_a9_neon_bp_2cycle): Likewise.
1639 (cortex_a9_neon_bp_3cycle): Likewise.
1640 (cortex_a9_neon_ldr): Likewise.
1641 (cortex_a9_neon_str): Likewise.
1642 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1643 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1644 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1645 (cortex_a9_neon_vld2_4_regs): Likewise.
1646 (cortex_a9_neon_vld3_vld4): Likewise.
1647 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1648 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1649 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1650 (cortex_a9_neon_vst3_vst4): Likewise.
1651 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1652 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1653 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1654 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1655 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1656 (cortex_a9_neon_mcr): Likewise.
1657 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1658 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
1659 (cortex_a9_fps): Likewise.
1660 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
1661 change.
1662 (cortex_m4_fmuls): Likewise.
1663 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
1664 change.
1665 (cortex_r4_mrc): Likewise.
1666 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 1667 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 1668 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB
1669 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
1670 attribute change.
003bb7f3
JG
1671 (pj4_core_to_vfp): Likewise.
1672 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
1673 attribute change.
1674 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
1675 (vfp_fstore): Likewise.
1676 * doc/md.texi: Change references to neon_type to refer to type.
1677
fbe0dc78
DS
16782013-09-04 Dodji Seketeli <dodji@redhat.com>
1679
1680 * tree.h (DECL_BUILT_IN): Fix typo in comment.
1681
c167bc5b
DE
16822013-09-04 David Edelsohn <dje.gcc@gmail.com>
1683
1684 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
1685 lglobl if not weak.
1686
035cb59f
ER
16872013-09-04 Easwaran Raman <eraman@google.com>
1688
1689 PR middle-end/57370
1690 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
1691 (build_and_add_sum): Use it.
1692 (appears_later_in_bb): Simplify code.
1693
71191083
TJ
16942013-09-04 Teresa Johnson <tejohnson@google.com>
1695
1696 * dumpfile.c (dump_finish): Don't close stderr/stdout.
1697
0c488213
JG
16982013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
1699
1700 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
1701
61a74079
JH
17022013-09-04 Jan Hubicka <jh@suse.cz>
1703
1704 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
1705 * ipa-devirt.c: Include diganostic.h
1706 (odr_type_d): Add types and types_set.
1707 (hash_type_name): Work for types with vtables during LTO.
1708 (odr_hasher::remove): Fix comment; destroy types_set.
1709 (add_type_duplicate): New function,
1710 (get_odr_type): Use it.
1711 (dump_type_inheritance_graph): Dump type duplicates.
1712 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
1713 graph.
1714 * tree.c (types_same_for_odr): Give exact answers on types with
1715 virtual tables.
1716
24ba8195
DS
17172013-09-04 Dodji Seketeli <dodji@redhat.com>
1718
1719 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
1720 explaining their differences.
1721
0a04869e
SKS
17222013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
1723
1724 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
1725
3b18bc42
JL
17262013-09-03 Jeff Law <law@redhat.com>
1727
1728 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
1729 when not threading through a joiner block. Pass joiner/no joiner
1730 state to register_jump_thread.
1731 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
1732 state from argument rather than implying on path length.
1733 Dump the entire jump thread path into debugging dump.
1734 * tree-flow.h (register_jump_thread): Update prototype.
1735
ae84e151
XDL
17362013-08-29 Xinliang David Li <davidxl@google.com>
1737
1738 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1739 Remove a trivial gcc_assert.
1740
c716e67f
XDL
17412013-08-29 Xinliang David Li <davidxl@google.com>
1742
1743 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
1744 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
1745 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1746 Delay base decl alignment adjustment.
1747 * tree-vectorizer.c (vect_destroy_datarefs): New function.
1748 * tree-vectorizer.h: New data structure.
1749 (set_dr_misalignment): New function.
1750 (dr_misalignment): Ditto.
1751 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
1752 (vectorizable_load): Ditto.
1753 (ensure_base_align): New function.
1754 (vectorize_loops): Add dbg_cnt support.
1755 (execute_vect_slp): Ditto.
1756 * dbgcnt.def: New debug counter.
1757 * Makefile: New dependency.
1758
db8800bc
MI
17592013-09-03 Meador Inge <meadori@codesourcery.com>
1760
1761 Revert:
1762
1763 2013-08-30 Meador Inge <meadori@codesourcery.com>
1764
1765 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1766
0d9b4f5c
DE
17672013-09-03 David Edelsohn <dje.gcc@gmail.com>
1768
1769 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
1770 function descriptor.
1771
29e810b9
RB
17722013-09-03 Richard Biener <rguenther@suse.de>
1773
1774 * tree-affine.c (add_elt_to_tree): Fix association issue,
1775 avoid useless converts and make sure to always return a
1776 properly typed result.
1777
dbfc2894
RB
17782013-09-03 Richard Biener <rguenther@suse.de>
1779
1780 PR middle-end/57656
1781 * fold-const.c (negate_expr_p): Fix division case.
1782 (negate_expr): Likewise.
1783
fa3bf4ce
RB
17842013-09-03 Richard Biener <rguenther@suse.de>
1785
1786 PR lto/58285
1787 * tree-streamer-out.c: Include tm.h.
1788 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
1789
a64bbb3f
JH
17902013-09-03 Jan Hubicka <jh@suse.cz>
1791
1792 * tree-profile.c (tree_profiling): Cleanup CFG when done.
1793
93ed07e2
AM
17942013-09-03 Alan Modra <amodra@gmail.com>
1795
1796 * config.gcc (powerpc*-*-linux*): Add support for little-endian
1797 multilibs to big-endian target and vice versa.
1798 * config/rs6000/t-linux64: Use := assignment on all vars.
1799 (MULTILIB_EXTRA_OPTS): Remove fPIC.
1800 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
1801 * config/rs6000/t-linux64le: New file.
1802 * config/rs6000/t-linux64bele: New file.
1803 * config/rs6000/t-linux64lebe: New file.
1804
07d6cd64
JH
18052013-09-02 Jan Hubicka <jh@suse.cz>
1806
1807 * ipa-inline-transform.c (inline_transform): Do not
1808 optimize_inline_calls when not optimizing.
1809
8e4c9a10
JH
18102013-09-02 Jan Hubicka <jh@suse.cz>
1811
1812 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
1813 duplicated nodes for assembler names.
1814 * symtab.c (symtab_unregister_node): Do not attempt to unlink
1815 hard registers from assembler name hash.
1816
c91061e6
JH
18172013-09-02 Jan Hubicka <jh@suse.cz>
1818
1819 * ipa-split.c (execute_split_functions): Split externally visible
1820 functions called once.
1821
2fd0985c
MJ
18222013-09-02 Martin Jambor <mjambor@suse.cz>
1823
1824 PR ipa/58106
1825 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
1826 linked list. When finding the correct duplicate, also consider also
1827 the caller in additon to its inlined_to node.
1828
91bd4114
JG
18292013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
1830
1831 * config/aarch64/aarch64-simd-builtins.def
1832 (dup_lane_scalar): Remove.
1833 * config/aarch64/aarch64-simd.md
1834 (aarch64_simd_dup): Add 'w->w' alternative.
1835 (aarch64_dup_lane<mode>): Allow for VALL.
1836 (aarch64_dup_lane_scalar<mode>): Remove.
1837 (aarch64_dup_lane_<vswap_width_name><mode>): New.
1838 (aarch64_get_lane_signed<mode>): Add w->w altenative.
1839 (aarch64_get_lane_unsigned<mode>): Likewise.
1840 (aarch64_get_lane<mode>): Likewise.
1841 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
1842 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
1843 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
1844 (VCON): Change container of V2SF.
1845 (vswap_width_name): Likewise.
1846 * config/aarch64/arm_neon.h
1847 (__aarch64_vdup_lane_any): New.
1848 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1849 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
1850 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1851
d617d2d8
EB
18522013-09-02 Eric Botcazou <ebotcazou@adacore.com>
1853
1854 PR middle-end/56382
1855 * expr.c (emit_move_complex): Do not move complex FP values as parts if
1856 the source or the destination is a single hard register.
1857
63fdb7be
RB
18582013-09-02 Richard Biener <rguenther@suse.de>
1859
1860 PR middle-end/57511
1861 * tree-scalar-evolution.c (instantiate_scev_name): Allow
1862 non-linear SCEVs.
1863
78de2333
RB
18642013-09-02 Richard Biener <rguenther@suse.de>
1865
1866 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
1867 arithmetic to sizetype.
1868
85ff4ec6
BC
18692013-09-02 Bin Cheng <bin.cheng@arm.com>
1870
1871 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
1872 Find auto-increment use both before and after candidate.
1873
576016fe
MP
18742013-09-02 Marek Polacek <polacek@redhat.com>
1875
1876 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
1877
08f835dc
JH
18782013-09-01 Jan Hubicka <jh@suse.cz>
1879
1880 * Makefile.in: Add ipa-profile.o
1881 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
1882 * cgraph.c (struct cgraph_propagate_frequency_data,
1883 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
1884 ipa-profile.c; replace cgraph_ by ipa_ prefix.
1885 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB
1886 * ipa-inline-analysis.c: Include ipa-utils.h;
1887 drop duplicated cfgloop.h.
08f835dc
JH
1888 (inline_update_callee_summaries): Update.
1889 * ipa-profile.c: New file.
1890 * ipa-utils.h (ipa_propagate_frequency): Declare.
1891 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 1892 data-streamer.h, value-prof.h.
08f835dc
JH
1893 (symtab_remove_unreachable_nodes): Update profile.
1894 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
1895 account_time_size, cmp_counts, dump_histogram,
1896 ipa_profile_generate_summary, ipa_profile_write_summary,
1897 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
1898 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 1899 Move to ipa-profile.c.
08f835dc 1900
cec8583c
JDA
19012013-09-01 John David Anglin <danglin@gcc.gnu.org>
1902
1903 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
1904
bbc9396b
JH
19052013-09-01 Jan Hubicka <jh@suse.cz>
1906
1907 * common.opt (fdevirtualize-speculatively): New function.
1908 * invoke.texi (fdevirtualize-speculatively): Document.
1909 * ipa-devirt.c: Include ipa-inline.h
1910 (likely_target_p): New function.
1911 (ipa_devirt): New function.
1912 (gate_ipa_devirt): New function.
1913 (pass_data_ipa_devirt): New static var.
1914 (pass_ipa_devirt): Likewise.
1915 (make_pass_ipa_devirt): New function.
1916 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
1917 (common_handle_option): Disable devirtualization when
1918 value range profiling is available.
1919 * passes.def (pass_ipa_devirt): Add.
1920 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 1921 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 1922
0cea211e
IS
19232013-09-01 Iain Sandoe <iain@codesourcery.com>
1924
1925 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
1926 include sanitize(undefined).
1927
5e351e96
DN
19282013-08-31 Diego Novillo <dnovillo@google.com>
1929
1930 * Makefile.in (TREE_CORE_H): Define.
1931 (TREE_H): Use.
1932 (GTFILES): Add tree-core.h.
1933 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
1934 size the array.
1935 * tree-core.h: New file.
1936 Move all data structures, enum, typedefs, global
1937 declarations and constants from ...
1938 * tree.h: ... here.
1939
63bf9a90
JH
19402013-08-31 Jan Hubicka <jh@suse.cz>
1941
1942 * bulitins.c (expand_builtin): Do not early exit for gcov
1943 instrumented functions.
1944
5c0abd6a
MP
19452013-08-31 Marek Polacek <polacek@redhat.com>
1946
1947 * ubsan.c: Include tm_p.h.
1948
4843f032
JH
19492013-08-31 Jan Hubicka <jh@suse.cz>
1950
8356c89c
JH
1951 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
1952 warning.
1953
4843f032
JH
1954 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
1955 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 1956 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH
1957 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
1958 (ipa_merge_profiles): New function.
1959 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
1960 (lto_input_function_body): Likewise.
5e581212
JH
1961 * ipa-utils.h (ipa_merge_profiles): Declare.
1962 * lto-streamer.h (lto_input_function_body): Update prototype.
1963 (emit_label_in_global_context_p): Remove.
1964 * lto-symtab.c: Include ipa-utils.h
1965 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 1966
57292ce9
JH
19672013-08-31 Jan Hubicka <jh@suse.cz>
1968
1969 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
1970
b9cbfeeb
JH
19712013-08-31 Jan Hubicka <jh@suse.cz>
1972
1973 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
1974
41dedebd
UB
19752013-08-31 Uros Bizjak <ubizjak@gmail.com>
1976
1977 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
1978 "cmp" RTX before signed_comparison_operator check to account
1979 for "code" changes.
1980
06d65050
JH
19812013-08-30 Jan Hubicka <jh@suse.cz>
1982
41dedebd 1983 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH
1984 (detect_type_change_1): Rename to ...
1985 (detect_type_change): ... this one; early return on non-polymorphic
1986 types.
41dedebd 1987 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH
1988 use of detect_type_change.
1989 (compute_complex_assign_jump_func): Add param_type parameter;
1990 update use of detect_type_change_ssa.
1991 (compute_complex_ancestor_jump_func): Likewise.
1992 (ipa_get_callee_param_type): New function.
1993 (ipa_compute_jump_functions_for_edge): Compute parameter type;
1994 update calls to the jump function computation functions.
1995
600b5b1d
TJ
19962013-08-30 Teresa Johnson <tejohnson@google.com>
1997 Steven Bosscher <steven@gcc.gnu.org>
1998
1999 * cfgrtl.c (fixup_new_cold_bb): New routine.
2000 (commit_edge_insertions): Invoke fixup_partitions.
2001 (find_partition_fixes): New routine.
2002 (fixup_partitions): Ditto.
2003 (verify_hot_cold_block_grouping): Update comments.
2004 (rtl_verify_edges): Invoke find_partition_fixes.
2005 (rtl_verify_bb_pointers): Update comments.
2006 (rtl_verify_bb_layout): Ditto.
2007 * basic-block.h (probably_never_executed_edge_p): Declare.
2008 (fixup_partitions): Ditto.
2009 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
2010 * bb-reorder.c (sanitize_hot_paths): New function.
2011 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
2012 sanitize_hot_paths.
2013 * predict.c (probably_never_executed_edge_p): New routine.
2014 * cfg.c (check_bb_profile): Add partition insanity warnings.
2015
0a6ea5c9
MI
20162013-08-30 Meador Inge <meadori@codesourcery.com>
2017
2018 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2019
de5a5fa1
MP
20202013-08-30 Marek Polacek <polacek@redhat.com>
2021
2022 * Makefile.in (ubsan.o): Add.
2023 (c-family/c-ubsan.o): Add.
2024 (builtins.o): Add ubsan.h dependency.
2025 * ubsan.h: New file.
2026 * ubsan.c: New file.
2027 * common.opt: Add -fsanitize=undefined option.
2028 (flag_sanitize): Add variable.
2029 (fsanitize=): Add option. Add Driver.
2030 (fsanitize=thread): Remove option.
2031 (fsanitize=address): Likewise.
2032 (static-libubsan): New option.
2033 * doc/invoke.texi: Document the new flag and -static-libubsan.
2034 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
2035 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
2036 * builtin-attrs.def (ATTR_COLD): Define.
2037 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2038 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
2039 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
2040 * flag-types.h (sanitize_code): New enum.
2041 * opts.c (common_handle_option): Parse command line arguments
2042 of -fsanitize=. Add -fsanitize=unreachable option.
2043 * varasm.c (get_variable_section): Adjust.
2044 (assemble_noswitch_variable): Likewise.
2045 (assemble_variable): Likewise.
2046 (output_constant_def_contents): Likewise.
2047 (categorize_decl_for_section): Likewise.
2048 (place_block_symbol): Likewise.
2049 (output_object_block): Likewise.
2050 * builtins.def: Likewise.
2051 * toplev.c (compile_file): Likewise.
2052 (process_options): Likewise.
2053 * cppbuiltin.c: Likewise.
2054 * tsan.c (tsan_pass): Likewise.
2055 (tsan_gate): Likewise.
2056 (tsan_gate_O0): Likewise.
2057 * cfgexpand.c (partition_stack_vars): Likewise.
2058 (expand_stack_vars): Likewise.
2059 (defer_stack_allocation): Likewise.
2060 (expand_used_vars): Likewise.
2061 * cfgcleanup.c (old_insns_match_p): Likewise.
2062 * asan.c (asan_finish_file): Likewise.
2063 (asan_instrument): Likewise.
2064 (gate_asan): Likewise.
2065 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
2066 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2067 (asan_global_struct): Use pointer_sized_int_node instead
2068 calling build_nonstandard_integer_type.
2069 (initialize_sanitizer_builtins): Likewise.
2070 (asan_finish_file): Likewise.
2071 * gcc.c: Document %{%:function(args):X}.
2072 (static_spec_functions): Add sanitize.
2073 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
2074 store funcval != NULL there.
2075 (do_spec_1): Adjust handle_spec_function caller.
2076 (handle_braces): Allow %:function(args) as condition.
2077 (sanitize_spec_function): New function.
2078 (ADD_STATIC_LIBUBSAN_LIBS): Define.
2079 (LIBUBSAN_SPEC): Likewise.
2080 (LIBUBSAN_EARLY_SPEC): Likewise.
2081 (SANITIZER_SPEC): Handle libubsan.
2082 (SANITIZER_EARLY_SPEC): Likewise.
2083 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
2084 instead of fsanitize=address.
2085 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
2086 instead of fsanitize=address*.
2087 * builtins.c: Include ubsan.h.
2088 (fold_builtin_0): Instrument __builtin_unreachable.
2089 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
2090 instead of flag_asan.
2091 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
2092 (pointer_sized_int_node): Define.
41dedebd 2093 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 2094
f07f30cf
MS
20952013-08-30 Mike Stump <mikestump@comcast.net>
2096
2097 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
2098 with RE patterns.
2099
8a41354f
JH
21002013-08-29 Jan Hubicka <jh@suse.cz>
2101
2102 * cgraph.c (cgraph_function_body_availability): Handle weakref
2103 correctly.
2104 * passes.def: Remove pass_fixup_cfg.
2105 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
2106 track when we need to remove functions.
2107 (gate_ipa_inline): Execute inlining always; add comment why.
2108 (pass_data_ipa_inline): Remove TODO_remove_functions.
2109 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
2110 do not produce summaries.
2111 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
2112 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
2113 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
2114 constant pool and vtable.
2115
dc8d7a0f
TB
21162013-08-30 Tejas Belagod <tejas.belagod@arm.com>
2117
41dedebd
UB
2118 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
2119 New arm_neon.h's internal macros to specify 64-bit constants.
2120 Avoid using stdint.h's macros.
dc8d7a0f 2121
ac7eacd2
JR
21222013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
2123
2124 * recog.c (verify_changes): Verify that changes[i].old is non-zero
2125 before applying REG_P.
2126
8b29fd4e
JJ
21272013-08-30 Jakub Jelinek <jakub@redhat.com>
2128
2129 PR tree-optimization/58277
2130 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
2131 after seeing too many stmts with vdef in between dombb and current
2132 bb, invalidate everything.
2133
26d75703
RB
21342013-08-30 Richard Biener <rguenther@suse.de>
2135
2136 * fold-const.c (fold_single_bit_test): Fix overflow test.
2137
f7e088e7
EB
21382013-08-30 Eric Botcazou <ebotcazou@adacore.com>
2139
2140 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
2141 and which can live in a register, always retrieve the value on entry.
2142 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
2143 passed by invisible reference specially.
2144 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
2145 (vt_add_function_parameter): Correctly deal with a parameter passed by
2146 invisible reference.
2147
21482013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH
2149
2150 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
2151
6e8dad05
RB
21522013-08-30 Richard Biener <rguenther@suse.de>
2153
2154 PR tree-optimization/58228
2155 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
2156 allow invariant loads in nested loop vectorization.
2157
062ef2c8
RB
21582013-08-30 Richard Biener <rguenther@suse.de>
2159
2160 PR tree-optimization/58223
2161 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
2162 (has_anti_or_output_dependence): ... this and adjust to also
2163 look for output dependences.
2164 (mark_nodes_having_upstream_mem_writes): Adjust.
2165 (rdg_flag_uses): Likewise.
2166
7a764c60
RB
21672013-08-30 Richard Biener <rguenther@suse.de>
2168
2169 PR tree-optimization/58010
2170 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
2171 assert that we have a loop-closed PHI.
2172
82e9d642
JH
21732013-08-29 Jan Hubicka <jh@suse.cz>
2174
2175 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
2176 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
2177 * lto-section-in.c (lto_free_function_in_decl_state): New function.
2178 (lto_free_function_in_decl_state_for_node): New function.
2179
9cc1fb4b
XDL
21802013-08-29 Xinliang David Li <davidxl@google.com>
2181
41dedebd 2182 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL
2183 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
2184 Emit alignment peeling message with default -fopt-info.
2185 (vect_loop_versioning): Emit loop version info message.
41dedebd 2186 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL
2187 (execute_vect_slp): Ditto.
2188
440a5082
EB
21892013-08-29 Eric Botcazou <ebotcazou@adacore.com>
2190
2191 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
2192 of the clone from the DECL_NAME of the original function.
2193
3fa3690d
OE
21942013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
2195
2196 * passes.c (register_pass): Add overload.
2197 * tree-pass.h (register_pass): Forward declare it. Add comment.
2198
0170f33c
JH
21992013-08-29 Jan Hubicka <jh@suse.cz>
2200
41dedebd
UB
2201 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
2202 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
2203 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
2204 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
2205 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
2206 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 2207 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
2208 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
2209
103ff0d6
TJ
22102013-08-29 Teresa Johnson <tejohnson@google.com>
2211
2212 * dumpfile.c (dump_loc): Output column number.
2213 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
2214 * doc/invoke.texi: Document optall -fopt-info flag.
2215 * profile.c (read_profile_edge_counts): Use new dump framework.
2216 (compute_branch_probabilities): Ditto.
2217 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
2218 when pass not in any opt group.
2219 * pass_manager.h (pass_manager::get_pass_profile): New method.
2220 * value-prof.c (check_counter): Use new dump framework.
2221 (check_ic_target): Ditto.
2222 * coverage.c (get_coverage_counts): Ditto.
2223 (coverage_init): Setup new dump framework.
2224
301bbc16
RB
22252013-08-29 Richard Biener <rguenther@suse.de>
2226
2227 PR tree-optimization/58246
2228 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
2229 handle the dominance check inside a basic-block.
2230
1b275000
RB
22312013-08-29 Richard Biener <rguenther@suse.de>
2232
2233 PR middle-end/57287
2234 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
2235 of default defs that appear in abnormal PHI nodes.
2236
6b1184ba
RB
22372013-08-29 Richard Biener <rguenther@suse.de>
2238
2239 PR tree-optimization/57685
2240 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2241 single-use operands to avoid exponential complexity.
2242
1b08b734
DC
22432013-08-28 Dehao Chen <dehao@google.com>
2244
2245 * ipa-inline.c (edge_badness): Fix integer underflow.
2246
48a3fa69
UB
22472013-08-28 Uros Bizjak <ubizjak@gmail.com>
2248
2249 * gtm-builtins.def (_ITM_free): Declare leaf.
2250
215f73e6
JJ
22512013-08-28 Jakub Jelinek <jakub@redhat.com>
2252
2253 PR target/58067
2254 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2255 (*tls_local_dynamic_base_64_largepic): Likewise.
2256 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2257 Remove predicate from call operand.
2258 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2259 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2260
34554d1a
JL
22612013-08-28 Jeff Law <law@redhat.com>
2262
2263 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2264 checks for the number of predecessors and successors allowed.
2265 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2266 which require copying a joiner block if there is a request which
2267 is a subpath that requires no joiner block copying.
2268
c01c111b
JH
22692013-08-28 Jan Hubicka <jh@suse.cz>
2270
2271 * lto-streamer-out.c (DFS_write_tree_body): Drop
2272 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2273 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2274 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2275 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2276 Do not read DECL_ERROR_ISSUED.
2277 (unpack_ts_decl_with_vis_value_fields): Do not read
2278 DECL_DEFER_OUTPUT.
48a3fa69
UB
2279 (lto_input_ts_binfo_tree_pointers): Do not read
2280 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH
2281 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2282 write DECL_ERROR_ISSUED..
2283 (pack_ts_decl_with_vis_value_fields): Do not write
2284 DECL_DEFER_OUTPUT.
2285 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 2286 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH
2287 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2288 * tree.h (tree_decl_common): Update comment.
2289 (DECL_ERROR_ISSUED): Remove.
2290
4b128ece
JJ
22912013-08-28 Jakub Jelinek <jakub@redhat.com>
2292
2293 PR middle-end/58257
2294 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2295
12211b99 22962013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH
2297
2298 * builtins.def (free): Declare leaf.
2299
8dce4dbc
DM
23002013-08-27 David Malcolm <dmalcolm@redhat.com>
2301
2302 * gdbhooks.py: New.
2303 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2304 * configure: Regenerate.
2305
b8f6e610
MJ
23062013-08-27 Martin Jambor <mjambor@suse.cz>
2307
2308 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2309 (ipa_ancestor_jf_data): Likewise.
2310 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2311 (ipa_get_jf_pass_through_type_preserved): New function.
2312 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2313 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 2314 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ
2315 (ipa_get_jf_ancestor_result): Likewise.
2316 (propagate_vals_accross_pass_through): Use
2317 ipa_get_jf_pass_through_result to do all the value mappings.
2318 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2319 type_preserved flag.
2320 (ipa_set_jf_cst_copy): New function.
2321 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2322 (ipa_set_jf_arith_pass_through): Likewise.
2323 (ipa_set_ancestor_jf): Likewise.
2324 (compute_complex_assign_jump_func): Set type_preserved instead of
2325 punting.
2326 (ipa_compute_jump_functions_for_edge): Likewise.
2327 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2328 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 2329 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ
2330 (ipa_write_jump_function): Stream the type_preserved flags.
2331 (ipa_read_jump_function): Likewise.
2332
74bf76ed
JJ
23332013-08-27 Jakub Jelinek <jakub@redhat.com>
2334 Aldy Hernandez <aldyh@redhat.com>
2335
2336 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
2337 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
2338 * function.h (struct function): Add has_force_vect_loops and
2339 has_simduid_loops.
12211b99 2340 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ
2341 * gimple.c (gimple_build_omp_critical): Add KIND argument and
2342 handle it.
2343 * gimple.def: Update CLAUSES comments.
2344 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
2345 (gimple_build_omp_for): Add argument to prototype.
2346 (gimple_omp_for_kind): New.
2347 (gimple_omp_for_set_kind): New.
2348 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
2349 GOVD_DATA_SHARE_CLASS.
2350 (enum omp_region_type): Add ORT_SIMD.
2351 (gimple_add_tmp_var): Handle ORT_SIMD.
2352 (gimplify_var_or_parm_decl): Same.
2353 (is_gimple_stmt): Same.
2354 (omp_firstprivatize_variable): Same.
2355 (omp_add_variable): Only use splay_tree_insert if lookup failed.
2356 (omp_notice_variable): Handle ORT_SIMD.
12211b99 2357 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ
2358 (omp_check_private): Handle ORT_SIMD.
2359 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2360 OMP_CLAUSE_SAFELEN.
2361 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
2362 Handle OMP_CLAUSE_LASTPRIVATE.
2363 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2364 OMP_CLAUSE_SAFELEN.
2365 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
2366 (gimplify_expr): Handle OMP_SIMD.
2367 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
2368 (expand_GOMP_SIMD_VF): New.
2369 (expand_GOMP_SIMD_LAST_LANE): New.
2370 * internal-fn.def (GOMP_SIMD_LANE): New.
2371 (GOMP_SIMD_VF): New.
2372 (GOMP_SIMD_LAST_LANE): New.
2373 * omp-low.c: Include target.h.
2374 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
2375 OMP_CLAUSE_SAFELEN.
2376 (check_omp_nesting_restrictions): Same.
2377 (omp_max_vf): New.
2378 (lower_rec_simd_input_clauses): New.
2379 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
2380 OMP_CLAUSE_LINEAR.
2381 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
2382 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
2383 (expand_omp_build_assign): New.
2384 (expand_omp_for_init_counts): New.
2385 (expand_omp_for_init_vars): New.
2386 (extract_omp_for_update_vars): New.
2387 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
2388 and rewrite accordingly.
2389 (expand_omp_simd): New.
2390 (expand_omp_for): Use expand_omp_simd.
2391 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2392 (lower_omp_for): Do not lower the body.
2393 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2394 in their own loops.
2395 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB
2396 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2397 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ
2398 (gate_tree_if_conversion): Similarly.
2399 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2400 gimple_build_omp_for.
2401 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2402 * tree-parloops (create_parallel_loop): Pass kind argument to
2403 gimple_build_omp_for.
2404 * tree-pretty-print.c (dump_omp_clause): Add cases for
2405 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2406 OMP_CLAUSE__SIMDUID_.
2407 (dump_generic_node): Handle OMP_SIMD.
2408 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2409 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2410 unroll OMP_SIMD loops here.
12211b99 2411 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 2412 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 2413 loop->safelen.
74bf76ed
JJ
2414 (vect_analyze_data_refs): Handle simd loops.
2415 * tree-vect-loop.c (vectorizable_live_operation): Handle
2416 IFN_GOMP_SIMD*.
12211b99 2417 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ
2418 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2419 (vectorizable_load): Same.
12211b99 2420 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ
2421 (struct simduid_to_vf): New.
2422 (simduid_to_vf::hash): New.
2423 (simduid_to-vf::equal): New.
2424 (struct simd_array_to_simduid): New.
2425 (simd_array_to_simduid::hash): New.
2426 (simd_array_to_simduid::equal): New.
2427 (adjust_simduid_builtins): New.
2428 (struct note_simd_array_uses_struct): New.
2429 (note_simd_array_uses_cb): New.
2430 (note_simd_array_uses): New.
2431 (vectorize_loops): Handle simd hints and adjust simd builtins
2432 accordingly.
2433 * tree-vectorizer.h (struct _stmt_vec_info): Add
2434 simd_lane_access_p field.
2435 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2436 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2437 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2438 (omp_clause_code_name): Same.
2439 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2440 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2441 * tree.def (OMP_SIMD): New entry.
12211b99
UB
2442 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2443 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ
2444 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2445 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2446 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2447 (OMP_CLAUSE_LINEAR_STEP): New.
2448 (OMP_CLAUSE_SAFELEN_EXPR): New.
2449 (OMP_CLAUSE__SIMDUID__DECL): New.
2450 (find_omp_clause): New prototype.
2451
d0cf4e84
L
24522013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2453
2454 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2455 Haswell processor detection.
2456
992592ec
CW
24572013-08-27 Christian Widmer <shadow@umbrox.de>
2458
2459 PR target/57927
2460 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2461 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2462 AVX2 capable processors.
2463
9116eb22
TB
24642013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2465
2466 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2467 of vget_low_* with implementations in terms of other intrinsics.
2468
1c5abb53
MG
24692013-08-27 Marc Glisse <marc.glisse@inria.fr>
2470
2471 PR middle-end/57219
2472 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2473 values to -1, 0 and 1.
2474
2e100703
VP
24752013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2476
2477 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2478 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2479 (<optab><mode>3_insn): Remove.
2480 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2481 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2482 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2483 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2484 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2485 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2486 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2487 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2488 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2489 (ror<mode>3_insn): Likewise.
2490 * config/aarch64/predicates.md (aarch64_simd_register): New.
2491
4ded8276
RB
24922013-08-27 Richard Biener <rguenther@suse.de>
2493
2494 PR tree-optimization/57521
2495 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2496 one edge is non-critical.
2497 (find_phi_replacement_condition): Make sure to use a non-critical
2498 edge. Cleanup and remove old bug workarounds.
2499 (bb_postdominates_preds): Remove.
2500 (if_convertible_loop_p_1): Do not compute post-dominators.
2501 (combine_blocks): Do not free post-dominators.
2502 (main_tree_if_conversion): Likewise.
2503 (pass_data_if_conversion): Add TODO_verify_ssa.
2504
5aa11061
DD
25052013-08-27 DJ Delorie <dj@redhat.com>
2506
2507 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2508
fd91cfe3
YZ
25092013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2510
2511 * function.c (assign_parm_find_data_types): Set passed_mode and
2512 nominal_mode to the TYPE_MODE of nominal_type for the built
2513 pointer type in case of the struct-pass-by-reference.
2514
907555ce
JR
25152013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2516
2517 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2518 (UINT16_TYPE): Change default to "unsigned int".
2519
67518c93
JR
2520 * config/avr/avr.opt (mfract-convert-truncate): New option.
2521 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2522 is set, round negative fractional integers according to n1169
2523 when converting to integer types.
2524
d6d989fc
JH
25252013-08-26 Jan Hubicka <jh@suse.cz>
2526
2527 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2528 methods can not be called indirectly when their address is not taken.
2529
2aa3da06
JH
25302013-08-26 Jan Hubicka <jh@suse.cz>
2531
12211b99
UB
2532 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2533 ctor_for_folding.
2aa3da06 2534
0987ffe7
JH
25352013-08-26 Jan Hubicka <jh@suse.cz>
2536
2537 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2538 can be unshared.
2539
befe8647
JR
25402013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2541
2542 * reload.c (find_valid_class): Allow classes that do not include
2543 FIRST_PSEUDO_REGISTER - 1.
2544
a21e735e
JH
25452013-08-26 Jan Hubicka <jh@suse.cz>
2546
2547 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2548 fix edge count/frequency when speculation failed; fix type check
2549 for the direct call.
2550
e067bd43
JH
25512013-08-26 Jan Hubicka <jh@suse.cz>
2552
2553 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2554
e8aec975
JH
25552013-08-26 Jan Hubicka <jh@suse.cz>
2556
2557 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2558 to be changed by edge redirection.
2559
d0b66480
JH
25602013-08-26 Jan Hubicka <jh@suse.cz>
2561
12211b99
UB
2562 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2563 formating; add sanity check.
d0b66480
JH
2564 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2565 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2566 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2567 (resolve_noninline_speculation): Update callee keys, too.
2568
0f9aaac7
JH
25692013-08-26 Jan Hubicka <jh@suse.cz>
2570
2571 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2572 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2573
4b37444e
JR
25742013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2575
2576 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2577 into proper place.
2578
5339fc0c
UB
25792013-08-26 Uros Bizjak <ubizjak@gmail.com>
2580
2581 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2582 (x86_64_elf_select_section): Ditto.
2583 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2584 arguments.
2585 (ix86_pass_by_reference): Ditto.
5339fc0c
UB
2586 (output_set_got): Ditto.
2587 (ix86_unary_operator_ok): Ditto.
2588 (ix86_expand_builtin): Ditto.
2589
d5c3d3ef
JH
25902013-08-23 Jan Hubicka <jh@suse.cz>
2591
5339fc0c 2592 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 2593
5b1e7435
JH
25942013-08-23 Jan Hubicka <jh@suse.cz>
2595
2596 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
2597 (tree_decl_with_vis): Add FINAL field.
2598
ead69dac
JL
25992013-08-23 Jeff Law <law@redhat.com>
2600
2601 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
2602 the debugging dump when the expression is fully redundant.
2603
025311c4
GDR
26042013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2605
2606 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
2607 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
2608 * pretty-print.c (pp_formatted_text_data): Likewise.
2609 (pp_write_text_to_stream): Likewise.
2610 (pp_write_text_as_dot_label_to_stream): Likewise.
2611 (pp_append_r): Likewise.
2612 (pp_format): Likewise.
2613 (pp_flush): Likewise.
2614 (pp_clear_output_area): Likewise.
2615 (pp_append_text): Likewise.
2616 (pp_formatted_text): Likewise.
2617 (pp_remaining_character_count_for_line): Likewise.
2618 (pp_newline): Likewise.
2619 (pp_character): Likewise.
2620 (output_buffer::~output_buffer): Define.
2621 (pretty_printer::~pretty_printer): Destruct output buffer.
2622 * pretty-print.h (output_buffer::~output_buffer): Declare.
2623 (pretty_printer::~pretty_printer): Declare virtual.
2624
0cadbfaa
MG
26252013-08-24 Marc Glisse <marc.glisse@inria.fr>
2626
2627 PR other/57324
2628 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
2629 HOST_WIDE_INT_M1U): New macros.
2630 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
2631 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
2632 unsigned -1 for lshift.
2633 * cse.c (cse_insn): Likewise.
2634 * double-int.c (rshift_double, lshift_double): Likewise.
2635 * builtins.c (fold_builtin_bitop): Likewise.
2636 * combine.c (force_to_mode): Likewise.
2637 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
2638 * simplify-rtx.c (simplify_const_unary_operation,
2639 simplify_const_binary_operation): Likewise.
2640 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
2641 check_va_list_escapes): Likewise.
2642 * rtlanal.c (nonzero_bits1): Likewise.
2643 * expmed.c (expand_smod_pow2): Likewise.
2644 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
2645
009e5353
JH
26462013-08-23 Jan Hubicka <jh@suse.cz>
2647
2648 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
2649 as having address taken.
2650
c4be6568
JH
26512013-08-23 Jan Hubicka <jh@suse.cz>
2652
64cbf23d
JH
2653 * ipa-utils.h (method_class_type): Declare.
2654 * ipa-devirt.c (method_class_type): Export.
2655
c4be6568
JH
2656 * cgraphunit.c (analyze_functions): Do basic devirtualization;
2657 do not walk base classes of anonymous types.
2658
97aba8e9
KK
26592013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
2660
2661 PR rtl-optimization/58220
2662 PR regression/58221
2663 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
2664 handle SEQUENCE insns properly.
2665
0fc80001
GDR
26662013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2667
2668 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
2669 definition.
2670 (pp_newline_and_indent): Likewise.
2671 (pp_separate_with): Likewise.
2672 * pretty-print.c (pp_newline_and_flush): Define.
2673 (pp_newline_and_indent): Likewise.
2674 (pp_separate_with): Likewise.
2675
520a5868
JJ
26762013-08-23 Jakub Jelinek <jakub@redhat.com>
2677
2678 PR target/58218
2679 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
2680 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
2681
e500c62a
KY
26822013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
2683
f43245d1
UB
2684 * config/i386/predicates.md (ext_sse_reg_operand): New.
2685 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY
2686 predicate to determine if EVEX is needed.
2687 (*movsi_internal): Ditto.
2688 (*movdf_internal): Ditto.
2689 (*movsf_internal): Ditto.
f43245d1 2690 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 2691
6b00d7dd
JJ
26922013-08-23 Jakub Jelinek <jakub@redhat.com>
2693
2694 PR tree-optimization/58209
2695 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
2696 (find_tail_calls): Give up for pointer result types if m is non-NULL.
2697 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
2698 emit POINTER_PLUS_EXPR.
2699 (create_tailcall_accumulator): For pointer result type accumulate in
2700 sizetype type.
2701
4f2a9f90
PC
27022013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
2703
2704 * configure.ac: Add backslashes missing from the last change.
2705 * configure: Regenerate.
2706
7bf4274e
JH
27072013-08-22 Jan Hubicka <jh@susue.cz>
2708
4f2a9f90
PC
2709 * ipa.c (function_and_variable_visibility): First remember function
2710 was global and then make it local.
7bf4274e 2711
1bea243d
JB
27122013-08-22 Julian Brown <julian@codesourcery.com>
2713
2714 * configure.ac: Add aarch64 to list of arches which use "nop" in
2715 debug_line test.
2716 * configure: Regenerate.
2717
bcba7ecc
AK
27182013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2719
2720 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
2721 gnu_libc_has_function.
2722 * config/s390/tpf.h: Likewise.
2723
4c495b0d
JH
27242013-08-22 Jan Hubicka <jh@susue.cz>
2725
2726 * timevar.c (validate_phases): Add cast.
2727
5e302bca
JH
27282013-08-22 Jan Hubicka <jh@susue.cz>
2729
2730 * timevar.c (validate_phases): Use size_t for memory.
2731 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
2732
da6ca2b5
GDR
27332013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
2734
2735 * pretty-print.h (output_buffer::output_buffer): Declare.
2736 (pretty_printer::pretty_printer): Likewise.
2737 (pp_construct): Remove.
2738 * pretty-print.c (output_buffer::output_buffer): Define.
2739 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
2740 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
2741 (print_gimple_expr): Likewise.
2742 (print_gimple_seq): Likewise.
2743 (gimple_dump_bb): Likewise.
2744 * sched-vis.c (dump_value_slim): Likewise.
2745 (dump_insn_slim): Likewise.
2746 (dump_rtl_slim): Likewise.
2747 (str_pattern_slim): Likewise.
2748 * tree-mudflap.c (mf_varname_tree): Likewise.
2749 * graph.c (print_graph_cfg): Likewise.
2750 (start_graph_dump): Likewise.
2751 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
2752 placement-new.
2753 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
2754 pretty printer initialization.
2755 * coretypes.h (diagnostic_context): Remove superflous type alias
2756 declaration.
2757 (pretty_printer): Likewise. Declare directly as a class.
2758 (pretty_print_info): Remove declaration as class.
2759 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
2760 and pp_clear_output_area.
2761 (asan_add_global): Likewise.
2762
0e1474e5
JH
27632013-08-22 Jan Hubicka <jh@suse.cz>
2764
2765 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
2766 * ipa-utils.h (update_type_inheritance_graph): Declare.
2767 (possible_polymorphic_call_target_p): Declare.
2768 (possible_polymorphic_call_target_p): New.
2769 * ipa-devirt.c: Update toplevel comments.
2770 (cached_polymorphic_call_targets): Move up.
2771 (odr_type_d): Move ID down.
2772 (polymorphic_type_binfo_p): Update comment.
2773 (odr_hasher::remove): Likewise;
2774 (get_odr_type): Set anonymous_namespace.
2775 (dump_odr_type): Dump it.
2776 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
2777 (maybe_record_node): Record node in cached_polymorphic_call_targets.
2778 (record_binfo): Add comment.
5339fc0c
UB
2779 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
2780 built.
0e1474e5
JH
2781 (devirt_node_removal_hook): Do not iCE when cache is freed.
2782 (possible_polymorphic_call_target_p): New predicate.
2783 (update_type_inheritance_graph): New function.
2784
3f97cb0b
AI
27852013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
2786 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2787 Sergey Lega <sergey.s.lega@intel.com>
2788 Anna Tikhonova <anna.tikhonova@intel.com>
2789 Ilya Tocar <ilya.tocar@intel.com>
2790 Andrey Turetskiy <andrey.turetskiy@intel.com>
2791 Ilya Verbin <ilya.verbin@intel.com>
2792 Kirill Yukhin <kirill.yukhin@intel.com>
2793 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2794
2795 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
2796 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
2797 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
2798 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
2799 (OPTION_MASK_ISA_AVX2_UNSET): Update.
2800 (OPTION_MASK_ISA_AVX512F_UNSET): New.
2801 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
2802 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
2803 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
2804 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
2805 OPT_mavx512pf, OPT_mavx512er cases.
2806 * config/i386/constraints.md (v): New constraint.
2807 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
2808 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
2809 (bit_AVX512CD): New.
2810 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
2811 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
2812 * config/i386/i386-c.c (ix86_target_macros_internal):
2813 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
2814 __AVX512PF__.
2815 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
2816 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
2817 * config/i386/i386.c (regclass_map, dbx_register_map)
2818 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
2819 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
2820 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
2821 -mavx512pf options.
2822 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
2823 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
2824 -mavx512pf options. Fix formatting.
2825 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
2826 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
2827 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
2828 -mavx512cd, -mavx512pf options.
2829 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
2830 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
2831 (ix86_preferred_output_reload_class): Replace SSE_REGS with
2832 ALL_SSE_REGS.
2833 (ix86_hard_regno_mode_ok): Support 512-bit registers.
2834 (ix86_set_reg_reg_cost): Ditto.
2835 (x86_order_regs_for_local_alloc): Ditto.
2836 (MAX_VECT_LEN): Extend to 64-byte.
2837 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
2838 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
2839 (TARGET_AVX512ER, TARGET_AVX512CD): New.
2840 (BIGGEST_ALIGNMENT): Extend to 512-bits.
2841 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
2842 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
2843 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
2844 (SSE_REG_MODE_P): Support new modes.
2845 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
2846 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
2847 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
2848 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
2849 (REG_CLASS_CONTENTS): Add new registers.
2850 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
2851 (EXT_REX_SSE_REGNO_P): New.
2852 (HI_REGISTER_NAMES): Add new registers.
2853 * config/i386/i386.md: Define constants for new registers.
2854 (mode): Add new 512-bit modes.
2855 (prefix): Support evex prefix.
2856 (isa): Support avx512f, noavx512f, fma_avx512f.
2857 (ssemodesuffix): Add new 512-bit modes.
2858 (movxi): New.
2859 (*movxi_internal_avx512f): Ditto.
2860 (*movdi_internal): Replace constraint "x" with the new constraint "v".
2861 Support MODE_XI.
2862 (*movsi_internal): Likewise.
2863 (*movdf_internal): Likewise.
2864 (*movsf_internal): Likewise.
2865 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
2866 (<code><mode>3): Likewise.
5339fc0c
UB
2867 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
2868 New.
3f97cb0b
AI
2869 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
2870 with the new constraint "v".
2871 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
2872 modes.
2873 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
2874 with the new constraint "v".
2875 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
2876 (<sse2>_storedqu<avxsizesuffix>): Likewise.
2877 (*<plusminus_insn><mode>3): Likewise.
2878 (<sse>_vm<plusminus_insn><mode>3): Likewise.
2879 (*mul<mode>3): Likewise.
2880 (<sse>_vmmul<mode>3): Likewise.
2881 (<sse>_div<mode>3): Likewise.
2882 (<sse>_vmdiv<mode>3): Likewise.
2883 (<sse>_sqrt<mode>2): Likewise.
2884 (<sse>_vmsqrt<mode>2): Likewise.
2885 (*<code><mode>3_finite): Likewise.
2886 (*<code><mode>3) <smaxmin>: Likewise.
2887 (<sse>_vm<code><mode>3): Likewise.
2888 (*<code><mode>3) <any_logic>: Likewise.
2889 (*fma_fmadd_<mode>): Likewise.
2890 (*fma_fmsub_<mode>): Likewise.
2891 (*fma_fnmadd_<mode>): Likewise.
2892 (*fma_fnmsub_<mode>): Likewise.
2893 (*fma_fmaddsub_<mode>): Likewise.
2894 (*fma_fmsubadd_<mode>): Likewise.
2895 (*fmai_fmadd_<mode>): Likewise.
2896 (*fmai_fmsub_<mode>): Likewise.
2897 (*fmai_fnmadd_<mode>): Likewise.
2898 (*fmai_fnmsub_<mode>): Likewise.
2899 (sse_cvtsi2ss): Likewise.
2900 (sse_cvtsi2ssq): Likewise.
2901 (sse_cvtss2si): Likewise.
2902 (sse_cvtss2si_2): Likewise.
2903 (sse_cvtss2siq): Likewise.
2904 (sse_cvtss2siq_2): Likewise.
2905 (sse_cvttss2si): Likewise.
2906 (sse_cvtss2siq_2): Likewise.
2907 (float<sseintvecmodelower><mode>2): Likewise.
2908 (sse2_cvtsd2si_2): Likewise.
2909 (sse2_cvtsd2siq_2): Likewise.
2910 (*<plusminus_insn><mode>3): Likewise.
2911 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
2912 (*<sse4_1_avx2>_mul<mode>3): Likewise.
2913 (ashr<mode>3): Likewise.
2914 (<shift_insn><mode>3): Likewise.
2915 (avx2_<code><mode>3): Likewise.
2916 (*avx2_<code><mode>3): Likewise.
2917 (*andnot<mode>3): Likewise.
2918 (*<code><mode>3) <any_logic>: Likewise.
2919 (abs<mode>2): Likewise.
2920 (avx2_permvar<mode>): Likewise.
2921 (avx2_perm<mode>_1): Likewise.
2922 (*avx_vpermilp<mode>): Likewise.
2923 (avx_vpermilvar<mode>3): Likewise.
2924 (avx2_ashrv<mode>): Likewise.
2925 (avx2_<shift_insn>v<mode>): Likewise.
2926 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
2927 -mavx512cd.
2928 * doc/rtl.texi: Document XImode.
2929
dd5e8423
JL
29302013-08-21 Jeff Law <law@redhat.com>
2931
b9ebee5d
JL
2932 * tree-flow.h (register_jump_thread): Pass vector of edges
2933 instead of each important edge.
2934 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
2935 thread path into a vector and pass that to register_jump_thread.
2936 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
2937 passed in edge vector to the current 3-edge form.
2938
dd5e8423
JL
2939 Revert:
2940 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2941
9bb6628e 2942 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL
2943 delete_unmarked_insns removed anything.
2944
445dc8df
JR
29452013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
2946
2947 * reload.h (struct reg_equivs): Rename to ..
2948 (struct reg_equivs_s): .. this.
2949
e0df53dd
ML
29502013-08-20 Martin Liska <marxin.liska@gmail.com>
2951
2952 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
2953
e7d1d3eb
RO
29542013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2955
5339fc0c 2956 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 2957
5562e26e
JL
29582013-08-21 Jeff Law <law@redhat.com>
2959
2960 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
2961 simplify assignments too. If the RHS collapses to a singleton
2962 range, then return the value for the range.
2963
c7ecdec6
KY
29642013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
2965
2966 * config/i386/sse.md (V16): Rename to...
2967 (VMOVE): this.
2968 (mov<mode>): Update iterator name.
2969 (*mov<mode>_internal): Ditto.
2970 (push<mode>1): Ditto.
2971 (movmisalign<mode>): Ditto.
2972
bfa3b50a
JH
29732013-08-20 Jan Hubicka <jh@suse.cz>
2974
2975 PR bootstrap/58186
2976 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
2977 entry for direct edges.
2978 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
2979
6cd4d135
DM
29802013-08-20 David Malcolm <dmalcolm@redhat.com>
2981
2982 Revert my last two changes, r201865 and r201864:
2983
2984 Revert r201865:
2985 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2986
2987 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2988 instances can own GC refs.
2989
2990 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2991 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2992 (gcc::context::gt_pch_nx): Likewise.
2993 (gcc::context::gt_pch_nx): Likewise.
2994 * ggc.h (gt_ggc_mx <T>): New.
2995 (gt_pch_nx_with_op <T>): New.
2996 (gt_pch_nx <T>): New.
2997 * passes.c (opt_pass::gt_ggc_mx): New.
2998 (opt_pass::gt_pch_nx): New.
2999 (opt_pass::gt_pch_nx_with_op): New.
3000 (pass_manager::gt_ggc_mx): New.
3001 (pass_manager::gt_pch_nx): New.
3002 (pass_manager::gt_pch_nx_with_op): New.
3003 (pass_manager::operator new): Use
3004 ggc_internal_cleared_alloc_stat rather than xcalloc.
3005 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3006 (pass_manager::gt_ggc_mx): New.
3007 (pass_manager::gt_pch_nx): New.
3008 (pass_manager::gt_pch_nx_with_op): New.
3009 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3010 (opt_pass::operator new): New.
3011 (opt_pass::gt_ggc_mx): New.
3012 (opt_pass::gt_pch_nx): New.
3013 (opt_pass::gt_pch_nx_with_op): New.
3014
3015 Revert r201864:
3016 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3017
3018 * Makefile.in (GTFILES): Add context.h.
3019 * context.c (gcc::context::operator new): New.
3020 (gcc::context::gt_ggc_mx): New.
3021 (gcc::context::gt_pch_nx): New.
3022 (gcc::context::gt_pch_nx): New.
3023 * context.h (gcc::context): Add GTY((user)) marking.
3024 (gcc::context::operator new): New.
3025 (gcc::context::gt_ggc_mx): New.
3026 (gcc::context::gt_pch_nx): New.
3027 (gcc::context::gt_pch_nx): New.
3028 (g): Add GTY marking.
3029 (gt_ggc_mx (gcc::context *)): New.
3030 (gt_pch_nx (gcc::context *)): New.
3031 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3032 void *cookie)): New.
3033 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3034
a046136a
AM
30352013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3036
3037 * dce.c (fini_dce): Call df_analyze again just in case
3038 delete_unmarked_insns removed anything.
3039
67598720
TJ
30402013-08-20 Teresa Johnson <tejohnson@google.com>
3041
3042 PR rtl-optimizations/57451
3043 * final.c (reemit_insn_block_notes): Prevent lexical blocks
3044 from crossing split section boundaries.
3045
f8693faf
MGD
30462013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
3047
3048 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
3049 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
3050 with MULTLIB_DEFAULTS.
3051
7d0b9a9c
NC
30522013-08-20 Nick Clifton <nickc@redhat.com>
3053
3054 * target.def (narrow_volatile_bitfield): Note that the default
3055 value is false, not !TARGET_STRICT_ALIGN.
3056 * doc/tm.texi: Regenerate.
3057
8edb8dc8
PC
30582013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
3059
3060 Fix LIB_SPEC for systems without libpthread.
3061
3062 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
3063 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
3064 for Android.
3065 * config/i386/linux-common.h: Likewise.
3066 * config/mips/linux-common.h: Likewise.
3067
12211b99 30682013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ
3069
3070 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
3071 checks.
3072
e42c64cb
DM
30732013-08-20 David Malcolm <dmalcolm@redhat.com>
3074
3075 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3076 instances can own GC refs.
3077
3078 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3079 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3080 (gcc::context::gt_pch_nx): Likewise.
3081 (gcc::context::gt_pch_nx): Likewise.
3082 * ggc.h (gt_ggc_mx <T>): New.
3083 (gt_pch_nx_with_op <T>): New.
3084 (gt_pch_nx <T>): New.
3085 * passes.c (opt_pass::gt_ggc_mx): New.
3086 (opt_pass::gt_pch_nx): New.
3087 (opt_pass::gt_pch_nx_with_op): New.
3088 (pass_manager::gt_ggc_mx): New.
3089 (pass_manager::gt_pch_nx): New.
3090 (pass_manager::gt_pch_nx_with_op): New.
3091 (pass_manager::operator new): Use
3092 ggc_internal_cleared_alloc_stat rather than xcalloc.
3093 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3094 (pass_manager::gt_ggc_mx): New.
3095 (pass_manager::gt_pch_nx): New.
3096 (pass_manager::gt_pch_nx_with_op): New.
3097 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3098 (opt_pass::operator new): New.
3099 (opt_pass::gt_ggc_mx): New.
3100 (opt_pass::gt_pch_nx): New.
3101 (opt_pass::gt_pch_nx_with_op): New.
3102
910c02a0
DM
31032013-08-20 David Malcolm <dmalcolm@redhat.com>
3104
3105 * Makefile.in (GTFILES): Add context.h.
3106 * context.c (gcc::context::operator new): New.
3107 (gcc::context::gt_ggc_mx): New.
3108 (gcc::context::gt_pch_nx): New.
3109 (gcc::context::gt_pch_nx): New.
3110 * context.h (gcc::context): Add GTY((user)) marking.
3111 (gcc::context::operator new): New.
3112 (gcc::context::gt_ggc_mx): New.
3113 (gcc::context::gt_pch_nx): New.
3114 (gcc::context::gt_pch_nx): New.
3115 (g): Add GTY marking.
3116 (gt_ggc_mx (gcc::context *)): New.
3117 (gt_pch_nx (gcc::context *)): New.
3118 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3119 void *cookie)): New.
3120 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3121
af4b4236
AM
31222013-08-20 Alan Modra <amodra@gmail.com>
3123
3124 PR target/57865
3125 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
3126 (rs6000_emit_epilogue): Likewise.
3127
12211b99 31282013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC
3129
3130 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
3131
e2323f5b
PB
31322013-08-19 Peter Bergner <bergner@vnet.ibm.com>
3133 Jakub Jelinek <jakub@redhat.com>
3134
3135 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
3136 (BUILT_IN_FABSD64): Likewise.
3137 (BUILT_IN_FABSD128): Likewise.
3138 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
3139 (fold_builtin_1): Likewise.
5339fc0c
UB
3140 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
3141 destination and source operands.
e2323f5b
PB
3142 (*abstd2_fpr): Likewise.
3143 (*nabstd2_fpr): Likewise.
3144
cfbf3ee8
RS
31452013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
3146
3147 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
3148 JUMP_P and INSN_P.
3149
d3136aeb
AH
31502013-08-19 Aldy Hernandez <aldyh@redhat.com>
3151
3152 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
3153 incomplete.
3154
d33d9e47
AI
31552013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
3156
3157 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
3158 * builtins.c (default_libc_has_function): New.
3159 (gnu_libc_has_function): Ditto.
3160 (no_c99_libc_has_function): Ditto.
3161 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
3162 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
3163 (fold_builtin_sincos): Likewise.
3164 (fold_builtin_cexp): Likewise.
3165 * builtins.def (DEF_C94_BUILTIN): Likewise.
3166 (DEF_C99_BUILTIN): Likewise.
3167 (DEF_C99_C90RES_BUILTIN): Likewise.
3168 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
3169 definitions to using this define.
3170 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 3171 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI
3172 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
3173 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
3174 * config/darwin.h: Ditto.
3175 * config/elfos.h: Ditto.
3176 * config/freebsd.h: Ditto.
3177 * config/i386/cygming.h: Ditto.
3178 * config/i386/djgpp.h: Ditto.
3179 * config/i386/i386-interix.h: Ditto.
3180 * config/microblaze/microblaze.h: Ditto.
3181 * config/mmix/mmix.h: Ditto.
3182 * config/gnu-user.h: Ditto.
3183 * config/ia64/hpux.h: Ditto.
3184 * config/pa/pa-hpux.h: Ditto.
3185 * config/pdp11/pdp11.h: Ditto.
3186 * config/picochip/picochip.h: Ditto.
3187 * config/linux.h: Ditto.
3188 * config/netbsd.h: Ditto.
3189 * config/openbsd.h: Ditto.
3190 * config/rs6000/aix43.h: Ditto.
3191 * config/rs6000/aix51.h: Ditto.
3192 * config/rs6000/aix52.h: Ditto.
3193 * config/rs6000/aix53.h: Ditto.
3194 * config/rs6000/aix61.h: Ditto.
3195 * config/rs6000/darwin.h: Ditto.
3196 * config/rs6000/linux.h: Ditto.
3197 * config/rs6000/linux64.h: Ditto.
3198 * config/s390/tpf.h: Ditto.
3199 * config/sol2-10.h: Ditto.
3200 * config/sol2.h: Ditto.
3201 * config/vms/vms.h: Ditto.
3202 * config/vxworks.h: Ditto.
3203 * config/linux-android.c (linux_android_libc_has_function):
3204 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
3205 * config/linux-protos.h (linux_android_libc_has_function):
3206 New declaration.
3207 * config/i386/i386.c (ix86_libc_has_function): New.
3208 * config/i386/i386-protos.h
3209 (ix86_libc_has_function): New declaration.
3210 * config/i386/i386.md
3211 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
3212 ("isinf<mode>2): Likewise.
3213 * convert.c (convert_to_integer): Using new target hook
3214 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
3215 TARGET_C99_FUNCTIONS.
3216 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
3217 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
3218 * coretypes.h (function_class): New enum for different
3219 classes of functions.
3220 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
3221 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
3222 (TARGET_HAS_SINCOS): Likewise.
3223 (TARGET_LIBC_HAS_FUNCTION): New.
3224 * doc/tm.texi: Regenerated.
3225 * targhooks.h (default_libc_has_function): New declaration.
3226 (no_c99_libc_has_function): Ditto.
3227 (gnu_libc_has_function): Ditto.
3228 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
3229 and TARGET_HAS_SINCOS.
3230
eefe9a99
JH
32312013-08-18 Jan Hubicka <jh@suse.cz>
3232
3233 * Makeifle-in (ipa-devirt.o): New.
3234 (GTFILES): Add ipa-utils.h and ipa-devirt.c
3235 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
3236 (analyze_functions): Look into possible targets of polymorphic call.
3237 * dumpfile.c (dump_files): Add type-inheritance dump.
3238 * dumpfile.h (TDI_inheritance): New.
3239 * ipa-devirt.c: New file.
3240 * ipa-utils.h (odr_type_d): Forward declare.
3241 (odr_type): New type.
3242 (build_type_inheritance_graph): Declare.
3243 (possible_polymorphic_call_targets): Declare and introduce inline
3244 variant when only edge is pased.
3245 (dump_possible_polymorphic_call_targets): Likewise.
3246 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3247 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3248 (types_same_for_odr): ... here.
3249 * tree.h (type_in_anonymous_namespace_p): Declare.
3250
29b89442
JJ
32512013-08-18 Jakub Jelinek <jakub@redhat.com>
3252
3253 PR tree-optimization/58006
3254 * tree-parloops.c (take_address_of): Don't ICE if get_name
3255 returns NULL.
3256 (eliminate_local_variables_stmt): Remove clobber stmts.
3257
4f219961
EB
32582013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3259
3260 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3261 error message is issued for an alias to undefined symbol.
3262
1d5755ef
JH
32632013-08-18 Jan Hubicka <jh@suse.cz>
3264
3265 * cgraph.c (cgraph_create_indirect_edge): Discover
3266 polymorphic calls and record basic info into indirect_info.
3267 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3268 devirtualization, ignore objc function calls.
3269 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3270 call with no parm index info.
3271 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3272 * tree.c (virtual_method_call_p): New function.
3273 * tree.h (virtual_method_call_p): Declare.
3274
5945bebf
JH
32752013-08-16 Jan Hubicka <jh@suse.cz>
3276
3277 PR middle-end/58179
3278 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3279
47fa5d23
DE
32802013-08-16 David Edelsohn <dje.gcc@gmail.com>
3281
3282 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3283 attribute.
3284
c74f54a0
DM
32852013-08-16 David Malcolm <dmalcolm@redhat.com>
3286
3287 * gengtype.c (type_for_name): Add special-case support for
3288 locating types within the "gcc::" namespace.
3289 (open_base_files): Emit a "using namespace gcc" directive.
3290
f758f299
MM
32912013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3292
3293 PR target/58160
3294 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3295 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3296
3297 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3298 array instead of each individual operand as a separate argument.
3299 (emit_fusion_gpr_load): Likewise.
3300 (expand_fusion_gpr_load): Add new function declaration.
3301
3302 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3303 signature to have the operands passed as an array, instead of as
3304 separate arguments. Allow ZERO_EXTEND to be in the memory
3305 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3306 depend on the register live/dead flags when peepholes are run.
3307 (expand_fusion_gpr_load): New function to be called from the
3308 peephole2 pass, to change the register that addis sets to be the
3309 target register.
3310 (emit_fusion_gpr_load): Change the calling signature to have the
3311 operands passed as an array, instead of as separate arguments.
3312 Allow ZERO_EXTEND to be in the memory address, and also
3313 SIGN_EXTEND if -mpower8-fusion-sign.
3314
3315 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3316 unspec enumeration.
3317 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3318 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 3319 depend on the register live/dead state when the peephole pass is done.
f758f299 3320
158f4e4f
DM
33212013-08-16 David Malcolm <dmalcolm@redhat.com>
3322
3323 * gengtype.c (create_user_defined_type): Ensure that the kind
3324 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3325 declaration is seen before the GTY((user)) marking.
3326
22f8cea5
BE
33272013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
3328
3329 PR target/58105
3330 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
3331
c49bdb2e
JH
33322013-08-16 Jan Hubicka <jh@suse.cz>
3333
3334 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
3335 arugment expected_type.
3336 (gimple_fold_call): Use it.
3337 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
3338 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 3339 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH
3340 (try_make_edge_direct_virtual_call): Likewise.
3341 * tree.c (obj_type_ref_class): New.
3342 * tree.h (obj_type_ref_class): Use it.
3343
4042dca9
GDR
33442013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
3345
3346 * sched-vis.c (rtl_slim_pp_initialized): Remove.
3347 (rtl_slim_pp): Likewise.
3348 (init_rtl_slim_pretty_print): Likewise.
3349 (dump_value_slim): Don't call it. Use local pretty printer.
3350 (dump_insn_slim): Likewise.
3351 (dump_rtl_slim): Likewise.
3352 (str_pattern_slim): Likewise.
3353 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
3354 Simplify.
3355
7a460594
JJ
33562013-08-16 Jakub Jelinek <jakub@redhat.com>
3357
639dc669
JJ
3358 PR tree-optimization/58164
3359 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
3360 walk gimple_goto_dest of GIMPLE_GOTO.
3361
7a460594
JJ
3362 PR tree-optimization/58165
3363 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
3364 bi_call must be the last stmt in a bb, don't split_block, instead
3365 use fallthru edge from it and give up if there is none.
3366 Release conds vector when returning early.
3367
e3c02cc7
XDL
33682013-08-14 Xinliang David Li <davidxl@google.com>
3369
3370 * config/i386/i386.c (ix86_option_override_internal):
3371 Remove unused variable and field.
3372
6469da2c
BS
33732013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3374
3375 PR target/57949
22f8cea5 3376 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS
3377 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
3378 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
3379 and Linux, correct BLKmode alignment when 128-bit alignment is
3380 required and compatibility flag is not set.
22f8cea5
BE
3381 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
3382 for zero-size arguments when compatibility flag is not set.
6469da2c 3383
4ca890e2
JJ
33842013-08-14 Jakub Jelinek <jakub@redhat.com>
3385
3386 PR tree-optimization/58145
3387 * tree-sra.c (build_ref_for_offset): If prev_base has
3388 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
3389
daec1759 33902013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 3391
daec1759 3392 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 3393 Fix uninitialized variable error.
daec1759 3394
3ad20bd4
XDL
33952013-08-14 Xinliang David Li <davidxl@google.com>
3396
3397 * config/i386/i386.opt: Define two new options.
3398 * config/i386/x86-tune.def: Add arch selector field in macros.
3399 * config/i386/i386.h: Adjust macro definition.
3400 * config/i386/i386.c (ix86_option_override_internal):
3401 Refactor the code.
3402 (parse_mtune_ctrl_str): New function.
3403 (set_ix86_tune_features): New function.
3404 (ix86_function_specific_restore): Call the new helper function.
3405
e839e2a9
AB
34062013-08-14 Andrey Belevantsev <abel@ispras.ru>
3407
3408 PR rtl-optimization/57662
3409 * sel-sched.c (code_motion_process_successors): When the current insn
3410 is removed after the recursive traversal, break from the loop.
3411 Add comments and debug printouts.
3412
27ed665f
JJ
34132013-08-14 Jakub Jelinek <jakub@redhat.com>
3414 Alexandre Oliva <aoliva@redhat.com>
3415
3416 PR target/58067
3417 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3418 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3419 there also UNSPEC_PLTOFF.
3420
61c1a609
MP
34212013-08-14 Marek Polacek <polacek@redhat.com>
3422
3423 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3424 NUM_CONDITIONS bit positions.
3425
0e901c67
CC
34262013-08-13 Cary Coutant <ccoutant@google.com>
3427
3428 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3429 (attr_checksum): Hash vector contents instead of pointer.
3430 (attr_checksum_ordered): Likewise.
3431
061eff6d
UB
34322013-08-13 Uros Bizjak <ubizjak@gmail.com>
3433
3434 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3435 when Pmode != word_mode. Add length_address attribute.
3436 (sse3_monitor_<mode>): Merge from sse3_monitor and
3437 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3438 Pmode != word_mode. Update insn length attribute.
3439 * config/i386/i386.c (ix86_option_override_internal): Update
3440 ix86_gen_monitor selection for merged sse3_monitor insn.
3441
8ec9249b
JB
34422013-08-13 Julian Brown <julian@codesourcery.com>
3443
3444 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3445 perform invalid legitimization on greater-than-word-size modes for
3446 TARGET_E500_DOUBLE.
3447
761a8eb7
VM
34482013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3449
3450 * ira.c (setup_class_translate_array): Use aclass instead of cl
3451 for classes not fully covered by allocno classes.
3452
99c2bd54
JJ
34532013-08-13 Jakub Jelinek <jakub@redhat.com>
3454
4029a5e0
JJ
3455 PR tree-optimization/57661
3456 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3457 * tree-inline.c (tree_function_versioning): Initialize it.
3458 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3459 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3460 that is not being copied.
3461
99c2bd54
JJ
3462 PR sanitizer/56417
3463 * asan.c (instrument_strlen_call): Fix typo in comment.
3464 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3465
b5f8f063
SE
34662013-08-13 Steve Ellcey <sellcey@mips.com>
3467
3468 * config/mips/mips.md (prefetch): Use lw instead of ld on
3469 loongson in 32bit mode.
3470
4de80584
NC
34712013-08-13 Nick Clifton <nickc@redhat.com>
3472
3473 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3474
09ce3660
JH
34752013-08-13 Jan Hubicka <jh@suse.cz>
3476
3477 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3478 introduced edge; fix typo in sanity check.
3479 (cgraph_resolve_speculation): Export; improve diagnostic.
3480 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3481 speculation at type mismatch.
3482 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3483 (cgraph_resolve_speculation): Declare.
3484 (symtab_can_be_discarded): New function.
3485 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3486 * ipa-inline-transform.c (speculation_removed): New global var.
3487 (clone_inlined_nodes): See if speculation can be removed.
3488 (inline_call): If speculations was removed, we growths may not match.
3489 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3490 (speculation_useful_p): New function.
3491 (resolve_noninline_speculation): New function.
3492 (inline_small_functions): Resolve useless speculations.
3493 * ipa-inline.h (speculation_useful_p): Declare
3494 * ipa.c (can_replace_by_local_alias): Simplify.
3495 (ipa_profile): Produce speculative calls in non-lto, too;
3496 add simple cost model; produce local aliases.
3497
f971dc24
DM
34982013-08-13 David Malcolm <dmalcolm@redhat.com>
3499
3500 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 3501 PASS_MANAGER_H.
f971dc24 3502
2e507b9a
PC
35032013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3504
3505 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3506
c97fd4b8
UB
35072013-08-12 Uros Bizjak <ubizjak@gmail.com>
3508
3509 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3510 iterator instead of X87MODEF.
3511
35122013-08-12 Perez Read <netfirewall@gmail.com>
3513
3514 PR target/58132
3515 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3516 operand 0 for intel asm alternative.
3517 (*movabs<mode>_2): Ditto for operand 1.
3518
f4f4204c
JG
35192013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3520
3521 * config/aarch64/arm_none.h
3522 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3523
133b1a8e
NC
35242013-08-12 Nick Clifton <nickc@redhat.com>
3525
3526 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3527
15dd8b3a
YR
35282013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3529
3530 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3531 expand for QI/HImode operand to produce more effictive code for
3532 unsigned char(short) --> float(double) conversion.
3533
94109a6a
AM
35342013-08-12 Alexander Monakov <amonakov@ispras.ru>
3535
3536 * doc/invoke.texi: Mention that -ftls-model does not force the final
3537 model.
3538
4e115102
MP
35392013-08-12 Marek Polacek <polacek@redhat.com>
3540 Marc Glisse <marc.glisse@inria.fr>
3541
3542 PR tree-optimization/57980
3543 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3544 when creating -1 constant.
3545
748d71f3
JH
35462013-08-10 Jan Hubicka <jh@suse.cz>
3547
1f97e193 3548 Workaround binutils PR14342.
748d71f3
JH
3549 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3550 (gimple_init_edge_profiler): Likewise.
3551 (gimple_gen_ic_func_profiler): Likewise.
3552
5979aa54
JH
35532013-08-09 Jan Hubicka <jh@suse.cz>
3554
3555 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3556
ad83025e
XDL
35572013-08-09 Xinliang David Li <davidxl@google.com>
3558
3559 * config/i386/stringop.def: New file.
3560 * config/i386/stringop.opt: New file.
3561 * config/i386/i386-opts.h: Include stringopt.def.
3562 * config/i386/i386.opt: Include stringopt.opt.
3563 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 3564 Override default size based stringop inline strategies with options.
ad83025e
XDL
3565 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3566 New function.
3567
e59df5fd
JH
35682013-08-09 Jan Hubicka <jh@suse.cz>
3569
c97fd4b8 3570 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 3571
634ab819
JH
35722013-08-09 Jan Hubicka <jh@suse.cz>
3573
3574 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3575 CGRAPH_FREQ_MAX.
3576 (dump_cgraph_node): Dump profile-id.
3577 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3578 and common_target_probability.
3579 * lto-cgraph.c (lto_output_edge): Stream common targets.
3580 (lto_output_node): Stream profile ids.
3581 (input_node): Stream profile ids.
3582 (input_edge): Stream common targets.
3583 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3584 * ipa.c: Include value-prof.h
3585 (ipa_profile_generate_summary): Turn indirect call statement histograms
3586 into common targets.
3587 (ipa_profile): Turn common targets into speculative edges.
3588
2fa3d31b
JH
35892013-08-09 Jan Hubicka <jh@suse.cz>
3590
3591 * cgraph.h (cgraph_node): Add profile_id.
3592 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 3593 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH
3594 (del_node_map): Update.
3595 (find_func_by_funcdef_no): Replace by ...
3596 (find_func_by_profile_id): ... this one.
3597 (gimple_ic_transform): Do not remove useful histograms when
3598 speculation is not done; dump info when indirect call removal
3599 can happen at LTO.
3600 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
3601 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
3602 (__gcov_indirect_call_profiler_v2): .. this one.
3603 * profile.h (init_node_map): Update.
3604 * coverage.c (coverage_compute_profile_id): New function.
3605 * coverage.h (coverage_compute_profile_id): Declare.
3606 * tree-profile.c (init_ic_make_global_vars): Make
3607 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
3608 (gimple_init_edge_profiler): Update prototype of
3609 __gcov_indirect_call_profiler.
3610 (gimple_gen_ic_func_profiler): Simplify.
3611 (tree_profiling): Use init_node_map
3612
042ae7d2
JH
36132013-08-09 Jan Hubicka <jh@suse.cz>
3614
c97fd4b8
UB
3615 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
3616 non-speculative refs.
042ae7d2
JH
3617 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
3618 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
3619 (cgraph_set_call_stmt): Likewise.
3620 (cgraph_create_edge_1): Fix release checking compilatoin;
3621 clear lto_stmt_uid.
3622 (cgraph_free_edge): Free indirect info.
3623 (cgraph_turn_edge_to_speculative): New function.
3624 (cgraph_speculative_call_info): New function.
3625 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 3626 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH
3627 (dump_cgraph_node): Dump speculation.
3628 (verify_edge_count_and_frequency): Accept speculative edges.
3629 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
3630 (verify_cgraph_node): Handle speculation.
3631 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
3632 (cgraph_set_call_stmt): Update prototype.
3633 (cgraph_make_edge_direct): Update prototype.
3634 (cgraph_speculative_call_info): Declare.
3635 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
3636 to change; update call of ipa_find_references.
3637 * ipa-ref.c (ipa_record_reference): Fix return value; clear
3638 lto_stmt_uid and speculative flags.
3639 (ipa_dump_references): Dump speculation.
3640 (ipa_clone_references): Clone speculative flag.
3641 (ipa_clone_referring): Likewise.
3642 (ipa_clone_ref): New function.
3643 (ipa_find_reference): Look into lto_stmt_uids
3644 (ipa_clear_stmts_in_references): Do not clear speculative calls.
3645 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
3646 (ipa_find_reference): Update declaration.
3647 (ipa_clone_ref): Declare.
3648 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
3649 stream speculative flag.
3650 (lto_output_ref): Stream statements uids and speculation.
3651 (input_ref): Likewise.
3652 (input_edge): Stream speuclation.
3653 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
3654 (cgraph_set_call_stmt_including_clones): Handle speculation.
3655 * ipa-inline.c (heap_edge_removal_hook): New function.
3656 (inline_small_functions): Register it.
3657 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
3658 also initialize refs.
3659 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
3660 edge to change.
3661 (try_make_edge_direct_simple_call): Likewise.
3662 (try_make_edge_direct_simple_call): Likewise.
3663 (update_indirect_edges_after_inlining): Likewise.
3664 (remove_described_reference): Look proper lto_stmt_uid.
3665 (propagate_controlled_uses): Likewise.
3666 (propagate_controlled_uses): Liekwise.
3667 * tree-inline.c (copy_bb): Copy speculative edges.
3668 (redirect_all_calls): New function.
c97fd4b8
UB
3669 (copy_cfg_body): Do redirection after loop info is updated.
3670 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 3671
aa1e10cc
JH
36722013-08-09 Jan Hubicka <jh@suse.cz>
3673
3674 * lto-streamer-out.c (output_function): Renumber PHIs.
3675 * lto-streamer-in.c (input_function): Likewise.
3676
66adb8eb
JG
36772013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
3678
3679 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
3680 (get_lane_unsigned): Likewise.
3681 (dup_lane_scalar): Likewise.
3682 (get_lane): enable for VALL.
3683 * config/aarch64/aarch64-simd.md
3684 (aarch64_dup_lane_scalar<mode>): Remove.
3685 (aarch64_get_lane_signed<mode>): Likewise.
3686 (aarch64_get_lane_unsigned<mode>): Likewise.
3687 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
3688 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
3689 (aarch64_get_lane<mode>): Enable for all vector modes.
3690 (aarch64_get_lanedi): Remove misleading constraints.
3691 * config/aarch64/arm_neon.h
3692 (__aarch64_vget_lane_any): Define.
3693 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
3694 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
3695 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
3696 * config/aarch64/iterators.md (VDQQH): New.
3697 (VDQQHS): Likewise.
3698 (vwcore): Likewise.
3699
1f65ae7a
EB
37002013-08-09 Eric Botcazou <ebotcazou@adacore.com>
3701
3702 * configure.ac: Add GAS check for LEON instructions on SPARC.
3703 * configure: Regenerate.
3704 * config.in: Likewise.
3705 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
3706 sparc*-*-* block.
3707 * config/sparc/sparc.opt (LEON, LEON3): New masks.
3708 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
3709 for LEON or LEON3.
3710 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
3711 (AS_LEON_FLAG): New macro.
3712 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
3713 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
3714 Deal with LEON and LEON3 for the memory model.
23a6cb78 3715 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB
3716 (atomic_compare_and_swap<mode>_1): Likewise.
3717 (*atomic_compare_and_swap<mode>_1): Likewise.
3718
ff522f7f
ZC
37192013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3720
3721 * config/arm/neon.md (vcond): Fix floating-point vector
3722 comparisons against 0.
3723
cb1cca12
VM
37242013-08-08 Vladimir Makarov <vmakarov@redhat.com>
3725
3726 * lra-constraints.c (emit_spill_move): Remove assert.
3727 (process_alt_operands): Add more debugging
3728 output. Increase reject for spilling into memory. Decrease
3729 reject for reloading scratch.
3730 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
3731
a3719e31
SE
37322013-08-08 Steve Ellcey <sellcey@mips.com>
3733
3734 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
3735 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
3736 micromips incompatible. Add nan2008.
3737 (MULTILIB_DIRNAMES): Add nan2008.
3738 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3739 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
3740 and micromips incompatible. Add nan2008.
3741 (MULTILIB_DIRNAMES): Add nan2008.
3742 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3743
f40423e2
RS
37442013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
3745
3746 PR rtl-optimization/58079
3747 * combine.c (combine_simplify_rtx): Avoid using SUBST if
3748 simplify_comparison has widened a comparison with an integer.
3749
869b9125
KT
37502013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3751
3752 * config/arm/neon.md (movmisalign<mode>): Disable when we
3753 don't allow unaligned accesses.
3754 (*movmisalign<mode>_neon_store): Likewise.
3755 (*movmisalign<mode>_neon_load): Likewise.
3756 (*movmisalign<mode>_neon_store): Likewise.
3757 (*movmisalign<mode>_neon_load): Likewise.
3758
71cafea9
JH
37592013-08-08 Jan Hubicka <jh@suse.cz>
3760
3761 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
3762 (make_pass_rebuild_cgraph_edges): Also clear references.
3763 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
3764 * ipa-inline-transform.c (inline_transform): Remove all references
3765 after inlining.
c97fd4b8
UB
3766 * cgraphunit.c (expand_function): Remove all references after
3767 expansion.
71cafea9
JH
3768 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
3769 (ipa_find_reference): Rewrite to iterator.
3770 (remove_stmt_references): Likewise.
3771 (ipa_clear_stmts_in_references): New function.
3772 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB
3773 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
3774 clear references.
71cafea9
JH
3775 * ipa-split.c (split_function): Remove references in split function.
3776
27d2e612
RE
37772013-08-08 Richard Earnshaw <rearnsha@arm.com>
3778
3779 PR target/57431
1f65ae7a 3780 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE
3781 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
3782
5efc661c
RE
37832013-08-08 Richard Earnshaw <rearnsha@arm.com>
3784
3785 PR target/56979
1f65ae7a 3786 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE
3787 suggested mode for the assignment isn't compatible with the
3788 registers required.
3789
f276d31d
BE
37902013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
3791
3792 PR target/58065
3793 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
3794
4b8bc035
XDL
37952013-08-07 Xinliang David Li <davidxl@google.com>
3796
3797 * config/i386/i386.opt: New option -mtune-ctrl=.
3798 * config/i386/x86-tune.def: New file.
3799 * config/i386/i386.h: include x86-tune.def.
3800 * config/i386/i386.c (ix86_option_override_internal):
3801 Parsing -mtune-ctrl= option and set tune features.
3802
12211b99 38032013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE
3804
3805 PR other/12081
03b0ee0a 3806 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE
3807 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
3808 to gen_2arg_fn_t.
3809
5cf6635b
EB
38102013-08-07 Eric Botcazou <ebotcazou@adacore.com>
3811
3812 * rtl.h (update_alignments): Declare.
3813 * final.c (grow_label_align): New function extracted from...
3814 (shorten_branches): ...here. Call it.
3815 (update_alignments): New function.
3816 * reorg.c (sibling_labels): New variable.
3817 (get_label_before): Add SIBLING parameter. If it is non-zero, push
3818 the new label along with it onto the sibling_labels vector.
3819 (fill_simple_delay_slots): Adjust call to get_label_before.
3820 (fill_slots_from_thread): Likewise.
3821 (relax_delay_slots): Likewise.
3822 (make_return_insns): Likewise.
3823 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
3824
3c8ca1ab
EB
38252013-08-07 Eric Botcazou <ebotcazou@adacore.com>
3826
3827 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
3828 document its semantics.
3829 (diagnostic_report_diagnostic): Adjust accordingly.
3830
5ee5b32c
DM
38312013-08-07 David Malcolm <dmalcolm@redhat.com>
3832
03b0ee0a 3833 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM
3834 (sparc_option_override): ...and port to new C++ pass API.
3835 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
3836
abd566fa
PB
38372013-08-07 Peter Bergner <bergner@vnet.ibm.com>
3838
3839 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
3840
2077db1b
CT
38412013-08-06 Caroline Tice <cmtice@google.com>
3842
3843 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
3844 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
3845 * tree-pass.h: Add pass_vtable_verify.
3846 * varasm.c (assemble_variable): Add code to properly set the comdat
3847 section and name for the .vtable_map_vars section.
3848 (assemble_vtyv_preinit_initializer): New function.
3849 (default_sectin_type_flags): Make sure .vtable_map_vars section has
3850 LINK_ONCE flag.
3851 * output.h: Add function decl for assemble_vtv_preinit_initializer.
3852 * vtable-verify.c: New file.
3853 * vtable-verify.h: New file.
3854 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
3855 initialiation levels.
3856 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
3857 * passes.def: Insert pass_vtable_verify.
3858 * aclocal.m4: Reorder includes.
3859 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
3860 -fvtv-counts options.
3861 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
3862 as appropriate, if -fvtable-verify=... is used.
3863 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
3864 -fvtable-verify=... is used.
3865 * Makefile.in (OBJS): Add vtable-verify.o to list.
3866 (vtable-verify.o): Add new build rule.
3867 (GTFILES): Add vtable-verify.c to list.
3868 * common.opt (fvtable-verify=): New flag.
3869 (vtv_priority): Values for fvtable-verify= flag.
3870 (fvtv-counts): New flag.
3871 (fvtv-debug): New flag.
3872 * tree.h (save_vtable_map_decl): New extern function decl.
3873
03085d1c
DM
38742013-08-07 David Malcolm <dmalcolm@redhat.com>
3875
3876 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
3877 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
3878 (pass_data_rl78_devirt): ...new pass_data instance and...
3879 (make_pass_rl78_devirt): ...new function.
3880 (rl78_asm_file_start): Port pass registration to new C++ API.
3881
05555c4a
DM
38822013-08-07 David Malcolm <dmalcolm@redhat.com>
3883
3884 * coretypes.h (rtl_opt_pass): Add.
3885 (gcc::context): Add.
3886 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
3887 (epiphany_init): Port to new C++ pass API.
3888 (epiphany_optimize_mode_switching): Likewise.
3889 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
3890 (pass_manager::get_pass_mode_switching): New.
3891 (pass_manager::get_pass_peephole2): New.
3892 * mode-switching.c (pass_mode_switching): Add clone method.
3893 * recog.c (pass_peephole2): Add clone method.
3894 (pass_split_all_insns): Add clone method.
3895
c5a12e26
DM
38962013-08-06 David Malcolm <dmalcolm@redhat.com>
3897
03b0ee0a
UB
3898 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
3899 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 3900
a2e2a668
JH
39012013-08-06 Jan Hubicka <jh@suse.cz>
3902
3903 * cgraph.c (cgraph_get_body): New function based on lto.c
3904 implementation.
3905 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB
3906 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
3907 LTO paths.
a2e2a668
JH
3908 * cgraphunit.c (expand_function): Get body prior expanding.
3909 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB
3910 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
3911 really need.
a2e2a668
JH
3912 * passes.c (do_per_function_toporder): Get body.
3913 * tree-inline.c (expand_call_inline): Get body prior inlining it.
3914 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
3915
15682f24
MJ
39162013-08-06 Martin Jambor <mjambor@suse.cz>
3917
3918 PR fortran/57987
3919 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
3920 is not re-finalized. Rename second parameter to no_collect.
3921
78f6dd68
MJ
39222013-08-06 Martin Jambor <mjambor@suse.cz>
3923
3924 PR middle-end/58041
3925 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
3926 MEM_REF has proper alignment information.
3927
12211b99 39282013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE
3929
3930 PR other/12081
3931 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
3932 class insn_gen_fn.
3933 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
3934 rtx (*) (rtx, ...) with insn_gen_fn.
3935 * genoutput.c (output_insn_data): Cast gen_? function pointers to
3936 insn_gen_fn::stored_funcptr. Add initializer braces.
3937
8ac69a6c
DM
39382013-08-05 David Malcolm <dmalcolm@redhat.com>
3939
3940 Rewrite how instances of passes are cloned to remove assumptions
3941 about their sizes (thus allowing pass subclasses to have
3942 additional data fields, albeit non-GC-managed ones at this point).
3943
3944 * passes.c (make_pass_instance): Now that passes have clone
3945 methods, rewrite this function to eliminate XNEW and memcpy
3946 calls that used hardcoded sizes. Since this function no longer
3947 creates pass instances, rename it to...
03b0ee0a
UB
3948 (add_pass_instance): ...this. Document the old way that passes were
3949 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM
3950 (next_pass_1): Add an initial_pass argument for use by
3951 add_pass_instance.
3952 (position_pass): When adding multiple instances of a pass, use
3953 the pass's clone method, rather than relying on the XNEW/memcpy
3954 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB
3955 (pass_manager::pass_manager): When invoking next_pass_1, also supply
3956 the initial instance of the current pass within the pass manager.
8ac69a6c 3957
27a4cd48
DM
39582013-08-05 David Malcolm <dmalcolm@redhat.com>
3959
3960 This is the automated part of the conversion of passes from C
3961 structs to C++ classes.
3962
3963 Patch autogenerated by refactor_passes.py from
3964 https://github.com/davidmalcolm/gcc-refactoring-scripts
3965 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
3966
3967 * asan.c (pass_asan): Convert from a global struct to a subclass of
3968 gimple_opt_pass along with...
3969 (pass_data_asan): ...new pass_data instance and...
3970 (make_pass_asan): ...new function.
3971 (pass_asan_O0): Convert from a global struct to a subclass of
3972 gimple_opt_pass along with...
3973 (pass_data_asan_O0): ...new pass_data instance and...
3974 (make_pass_asan_O0): ...new function.
3975 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
3976 subclass of rtl_opt_pass along with...
3977 (pass_data_inc_dec): ...new pass_data instance and...
3978 (make_pass_inc_dec): ...new function.
3979 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
3980 a subclass of rtl_opt_pass along with...
3981 (pass_data_reorder_blocks): ...new pass_data instance and...
3982 (make_pass_reorder_blocks): ...new function.
3983 (pass_duplicate_computed_gotos): Convert from a global struct to a
3984 subclass of rtl_opt_pass along with...
3985 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
3986 (make_pass_duplicate_computed_gotos): ...new function.
3987 (pass_partition_blocks): Convert from a global struct to a subclass of
3988 rtl_opt_pass along with...
3989 (pass_data_partition_blocks): ...new pass_data instance and...
3990 (make_pass_partition_blocks): ...new function.
3991 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
3992 struct to a subclass of rtl_opt_pass along with...
3993 (pass_data_branch_target_load_optimize1): ...new pass_data instance
3994 and...
3995 (make_pass_branch_target_load_optimize1): ...new function.
3996 (pass_branch_target_load_optimize2): Convert from a global struct to a
3997 subclass of rtl_opt_pass along with...
3998 (pass_data_branch_target_load_optimize2): ...new pass_data instance
3999 and...
4000 (make_pass_branch_target_load_optimize2): ...new function.
4001 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
4002 of rtl_opt_pass along with...
4003 (pass_data_jump): ...new pass_data instance and...
4004 (make_pass_jump): ...new function.
4005 (pass_jump2): Convert from a global struct to a subclass of
4006 rtl_opt_pass along with...
4007 (pass_data_jump2): ...new pass_data instance and...
4008 (make_pass_jump2): ...new function.
4009 * cfgexpand.c (pass_expand): Convert from a global struct to a
4010 subclass of rtl_opt_pass along with...
4011 (pass_data_expand): ...new pass_data instance and...
4012 (make_pass_expand): ...new function.
4013 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
4014 of rtl_opt_pass along with...
4015 (pass_data_free_cfg): ...new pass_data instance and...
4016 (make_pass_free_cfg): ...new function.
4017 (pass_into_cfg_layout_mode): Convert from a global struct to a
4018 subclass of rtl_opt_pass along with...
4019 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
4020 (make_pass_into_cfg_layout_mode): ...new function.
4021 (pass_outof_cfg_layout_mode): Convert from a global struct to a
4022 subclass of rtl_opt_pass along with...
4023 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
4024 (make_pass_outof_cfg_layout_mode): ...new function.
4025 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
4026 struct to a subclass of gimple_opt_pass along with...
4027 (pass_data_build_cgraph_edges): ...new pass_data instance and...
4028 (make_pass_build_cgraph_edges): ...new function.
4029 (pass_rebuild_cgraph_edges): Convert from a global struct to a
4030 subclass of gimple_opt_pass along with...
4031 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
4032 (make_pass_rebuild_cgraph_edges): ...new function.
4033 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
4034 subclass of gimple_opt_pass along with...
4035 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
4036 and...
4037 (make_pass_remove_cgraph_callee_edges): ...new function.
4038 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
4039 struct to a subclass of rtl_opt_pass along with...
4040 (pass_data_stack_adjustments): ...new pass_data instance and...
4041 (make_pass_stack_adjustments): ...new function.
4042 * combine.c (pass_combine): Convert from a global struct to a subclass
4043 of rtl_opt_pass along with...
4044 (pass_data_combine): ...new pass_data instance and...
4045 (make_pass_combine): ...new function.
4046 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
4047 global struct to a subclass of rtl_opt_pass along with...
4048 (pass_data_compare_elim_after_reload): ...new pass_data instance
4049 and...
4050 (make_pass_compare_elim_after_reload): ...new function.
4051 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
4052 of rtl_opt_pass along with...
4053 (pass_data_rtl_cprop): ...new pass_data instance and...
4054 (make_pass_rtl_cprop): ...new function.
4055 * cse.c (pass_cse): Convert from a global struct to a subclass of
4056 rtl_opt_pass along with...
4057 (pass_data_cse): ...new pass_data instance and...
4058 (make_pass_cse): ...new function.
4059 (pass_cse2): Convert from a global struct to a subclass of
4060 rtl_opt_pass along with...
4061 (pass_data_cse2): ...new pass_data instance and...
4062 (make_pass_cse2): ...new function.
4063 (pass_cse_after_global_opts): Convert from a global struct to a
4064 subclass of rtl_opt_pass along with...
4065 (pass_data_cse_after_global_opts): ...new pass_data instance and...
4066 (make_pass_cse_after_global_opts): ...new function.
4067 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
4068 of rtl_opt_pass along with...
4069 (pass_data_ud_rtl_dce): ...new pass_data instance and...
4070 (make_pass_ud_rtl_dce): ...new function.
4071 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
4072 rtl_opt_pass along with...
4073 (pass_data_fast_rtl_dce): ...new pass_data instance and...
4074 (make_pass_fast_rtl_dce): ...new function.
4075 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
4076 a subclass of rtl_opt_pass along with...
4077 (pass_data_df_initialize_opt): ...new pass_data instance and...
4078 (make_pass_df_initialize_opt): ...new function.
4079 (pass_df_initialize_no_opt): Convert from a global struct to a
4080 subclass of rtl_opt_pass along with...
4081 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
4082 (make_pass_df_initialize_no_opt): ...new function.
4083 (pass_df_finish): Convert from a global struct to a subclass of
4084 rtl_opt_pass along with...
4085 (pass_data_df_finish): ...new pass_data instance and...
4086 (make_pass_df_finish): ...new function.
4087 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
4088 rtl_opt_pass along with...
4089 (pass_data_rtl_dse1): ...new pass_data instance and...
4090 (make_pass_rtl_dse1): ...new function.
4091 (pass_rtl_dse2): Convert from a global struct to a subclass of
4092 rtl_opt_pass along with...
4093 (pass_data_rtl_dse2): ...new pass_data instance and...
4094 (make_pass_rtl_dse2): ...new function.
4095 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
4096 subclass of rtl_opt_pass along with...
4097 (pass_data_dwarf2_frame): ...new pass_data instance and...
4098 (make_pass_dwarf2_frame): ...new function.
4099 * except.c (pass_set_nothrow_function_flags): Convert from a global
4100 struct to a subclass of rtl_opt_pass along with...
4101 (pass_data_set_nothrow_function_flags): ...new pass_data instance
4102 and...
4103 (make_pass_set_nothrow_function_flags): ...new function.
4104 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
4105 subclass of rtl_opt_pass along with...
4106 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
4107 and...
4108 (make_pass_convert_to_eh_region_ranges): ...new function.
4109 * final.c (pass_compute_alignments): Convert from a global struct to a
4110 subclass of rtl_opt_pass along with...
4111 (pass_data_compute_alignments): ...new pass_data instance and...
4112 (make_pass_compute_alignments): ...new function.
4113 (pass_final): Convert from a global struct to a subclass of
4114 rtl_opt_pass along with...
4115 (pass_data_final): ...new pass_data instance and...
4116 (make_pass_final): ...new function.
4117 (pass_shorten_branches): Convert from a global struct to a subclass of
4118 rtl_opt_pass along with...
4119 (pass_data_shorten_branches): ...new pass_data instance and...
4120 (make_pass_shorten_branches): ...new function.
4121 (pass_clean_state): Convert from a global struct to a subclass of
4122 rtl_opt_pass along with...
4123 (pass_data_clean_state): ...new pass_data instance and...
4124 (make_pass_clean_state): ...new function.
4125 * function.c (pass_instantiate_virtual_regs): Convert from a global
4126 struct to a subclass of rtl_opt_pass along with...
4127 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
4128 (make_pass_instantiate_virtual_regs): ...new function.
4129 (pass_leaf_regs): Convert from a global struct to a subclass of
4130 rtl_opt_pass along with...
4131 (pass_data_leaf_regs): ...new pass_data instance and...
4132 (make_pass_leaf_regs): ...new function.
4133 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
4134 subclass of rtl_opt_pass along with...
4135 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
4136 and...
4137 (make_pass_thread_prologue_and_epilogue): ...new function.
4138 (pass_match_asm_constraints): Convert from a global struct to a
4139 subclass of rtl_opt_pass along with...
4140 (pass_data_match_asm_constraints): ...new pass_data instance and...
4141 (make_pass_match_asm_constraints): ...new function.
4142 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
4143 subclass of rtl_opt_pass along with...
4144 (pass_data_rtl_fwprop): ...new pass_data instance and...
4145 (make_pass_rtl_fwprop): ...new function.
4146 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
4147 rtl_opt_pass along with...
4148 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
4149 (make_pass_rtl_fwprop_addr): ...new function.
4150 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
4151 rtl_opt_pass along with...
4152 (pass_data_rtl_pre): ...new pass_data instance and...
4153 (make_pass_rtl_pre): ...new function.
4154 (pass_rtl_hoist): Convert from a global struct to a subclass of
4155 rtl_opt_pass along with...
4156 (pass_data_rtl_hoist): ...new pass_data instance and...
4157 (make_pass_rtl_hoist): ...new function.
4158 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
4159 subclass of gimple_opt_pass along with...
4160 (pass_data_lower_cf): ...new pass_data instance and...
4161 (make_pass_lower_cf): ...new function.
4162 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
4163 from a global struct to a subclass of gimple_opt_pass along with...
4164 (pass_data_strength_reduction): ...new pass_data instance and...
4165 (make_pass_strength_reduction): ...new function.
4166 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
4167 of rtl_opt_pass along with...
4168 (pass_data_rtl_ifcvt): ...new pass_data instance and...
4169 (make_pass_rtl_ifcvt): ...new function.
4170 (pass_if_after_combine): Convert from a global struct to a subclass of
4171 rtl_opt_pass along with...
4172 (pass_data_if_after_combine): ...new pass_data instance and...
4173 (make_pass_if_after_combine): ...new function.
4174 (pass_if_after_reload): Convert from a global struct to a subclass of
4175 rtl_opt_pass along with...
4176 (pass_data_if_after_reload): ...new pass_data instance and...
4177 (make_pass_if_after_reload): ...new function.
4178 * init-regs.c (pass_initialize_regs): Convert from a global struct to
4179 a subclass of rtl_opt_pass along with...
4180 (pass_data_initialize_regs): ...new pass_data instance and...
4181 (make_pass_initialize_regs): ...new function.
4182 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
4183 of ipa_opt_pass_d along with...
4184 (pass_data_ipa_cp): ...new pass_data instance and...
4185 (make_pass_ipa_cp): ...new function.
4186 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
4187 global struct to a subclass of gimple_opt_pass along with...
4188 (pass_data_inline_parameters): ...new pass_data instance and...
4189 (make_pass_inline_parameters): ...new function.
4190 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
4191 subclass of gimple_opt_pass along with...
4192 (pass_data_early_inline): ...new pass_data instance and...
4193 (make_pass_early_inline): ...new function.
4194 (pass_ipa_inline): Convert from a global struct to a subclass of
4195 ipa_opt_pass_d along with...
4196 (pass_data_ipa_inline): ...new pass_data instance and...
4197 (make_pass_ipa_inline): ...new function.
4198 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
4199 struct to a subclass of gimple_opt_pass along with...
4200 (pass_data_local_pure_const): ...new pass_data instance and...
4201 (make_pass_local_pure_const): ...new function.
4202 (pass_ipa_pure_const): Convert from a global struct to a subclass of
4203 ipa_opt_pass_d along with...
4204 (pass_data_ipa_pure_const): ...new pass_data instance and...
4205 (make_pass_ipa_pure_const): ...new function.
4206 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
4207 to a subclass of ipa_opt_pass_d along with...
4208 (pass_data_ipa_reference): ...new pass_data instance and...
4209 (make_pass_ipa_reference): ...new function.
4210 * ipa-split.c (pass_split_functions): Convert from a global struct to
4211 a subclass of gimple_opt_pass along with...
4212 (pass_data_split_functions): ...new pass_data instance and...
4213 (make_pass_split_functions): ...new function.
4214 (pass_feedback_split_functions): Convert from a global struct to a
4215 subclass of gimple_opt_pass along with...
4216 (pass_data_feedback_split_functions): ...new pass_data instance and...
4217 (make_pass_feedback_split_functions): ...new function.
4218 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
4219 global struct to a subclass of simple_ipa_opt_pass along with...
4220 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
4221 instance and...
4222 (make_pass_ipa_function_and_variable_visibility): ...new function.
4223 (pass_ipa_free_inline_summary): Convert from a global struct to a
4224 subclass of simple_ipa_opt_pass along with...
4225 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
4226 (make_pass_ipa_free_inline_summary): ...new function.
4227 (pass_ipa_whole_program_visibility): Convert from a global struct to a
4228 subclass of ipa_opt_pass_d along with...
4229 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
4230 and...
4231 (make_pass_ipa_whole_program_visibility): ...new function.
4232 (pass_ipa_profile): Convert from a global struct to a subclass of
4233 ipa_opt_pass_d along with...
4234 (pass_data_ipa_profile): ...new pass_data instance and...
4235 (make_pass_ipa_profile): ...new function.
4236 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
4237 ipa_opt_pass_d along with...
4238 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4239 (make_pass_ipa_cdtor_merge): ...new function.
4240 * ira.c (pass_ira): Convert from a global struct to a subclass of
4241 rtl_opt_pass along with...
4242 (pass_data_ira): ...new pass_data instance and...
4243 (make_pass_ira): ...new function.
4244 (pass_reload): Convert from a global struct to a subclass of
4245 rtl_opt_pass along with...
4246 (pass_data_reload): ...new pass_data instance and...
4247 (make_pass_reload): ...new function.
4248 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4249 subclass of rtl_opt_pass along with...
4250 (pass_data_cleanup_barriers): ...new pass_data instance and...
4251 (make_pass_cleanup_barriers): ...new function.
4252 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4253 of rtl_opt_pass along with...
4254 (pass_data_loop2): ...new pass_data instance and...
4255 (make_pass_loop2): ...new function.
4256 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4257 rtl_opt_pass along with...
4258 (pass_data_rtl_loop_init): ...new pass_data instance and...
4259 (make_pass_rtl_loop_init): ...new function.
4260 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4261 rtl_opt_pass along with...
4262 (pass_data_rtl_loop_done): ...new pass_data instance and...
4263 (make_pass_rtl_loop_done): ...new function.
4264 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4265 subclass of rtl_opt_pass along with...
4266 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4267 (make_pass_rtl_move_loop_invariants): ...new function.
4268 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4269 rtl_opt_pass along with...
4270 (pass_data_rtl_unswitch): ...new pass_data instance and...
4271 (make_pass_rtl_unswitch): ...new function.
4272 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4273 subclass of rtl_opt_pass along with...
4274 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4275 and...
4276 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4277 (pass_rtl_doloop): Convert from a global struct to a subclass of
4278 rtl_opt_pass along with...
4279 (pass_data_rtl_doloop): ...new pass_data instance and...
4280 (make_pass_rtl_doloop): ...new function.
4281 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4282 a subclass of rtl_opt_pass along with...
4283 (pass_data_lower_subreg): ...new pass_data instance and...
4284 (make_pass_lower_subreg): ...new function.
4285 (pass_lower_subreg2): Convert from a global struct to a subclass of
4286 rtl_opt_pass along with...
4287 (pass_data_lower_subreg2): ...new pass_data instance and...
4288 (make_pass_lower_subreg2): ...new function.
4289 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4290 struct to a subclass of ipa_opt_pass_d along with...
4291 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4292 (make_pass_ipa_lto_gimple_out): ...new function.
4293 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4294 of ipa_opt_pass_d along with...
4295 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4296 (make_pass_ipa_lto_finish_out): ...new function.
4297 * mode-switching.c (pass_mode_switching): Convert from a global struct
4298 to a subclass of rtl_opt_pass along with...
4299 (pass_data_mode_switching): ...new pass_data instance and...
4300 (make_pass_mode_switching): ...new function.
4301 * modulo-sched.c (pass_sms): Convert from a global struct to a
4302 subclass of rtl_opt_pass along with...
4303 (pass_data_sms): ...new pass_data instance and...
4304 (make_pass_sms): ...new function.
4305 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4306 subclass of gimple_opt_pass along with...
4307 (pass_data_expand_omp): ...new pass_data instance and...
4308 (make_pass_expand_omp): ...new function.
4309 (pass_lower_omp): Convert from a global struct to a subclass of
4310 gimple_opt_pass along with...
4311 (pass_data_lower_omp): ...new pass_data instance and...
4312 (make_pass_lower_omp): ...new function.
4313 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4314 of gimple_opt_pass along with...
4315 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4316 (make_pass_diagnose_omp_blocks): ...new function.
4317 * passes.c (pass_early_local_passes): Convert from a global struct to
4318 a subclass of simple_ipa_opt_pass along with...
4319 (pass_data_early_local_passes): ...new pass_data instance and...
4320 (make_pass_early_local_passes): ...new function.
4321 (pass_all_early_optimizations): Convert from a global struct to a
4322 subclass of gimple_opt_pass along with...
4323 (pass_data_all_early_optimizations): ...new pass_data instance and...
4324 (make_pass_all_early_optimizations): ...new function.
4325 (pass_all_optimizations): Convert from a global struct to a subclass
4326 of gimple_opt_pass along with...
4327 (pass_data_all_optimizations): ...new pass_data instance and...
4328 (make_pass_all_optimizations): ...new function.
4329 (pass_all_optimizations_g): Convert from a global struct to a subclass
4330 of gimple_opt_pass along with...
4331 (pass_data_all_optimizations_g): ...new pass_data instance and...
4332 (make_pass_all_optimizations_g): ...new function.
4333 (pass_rest_of_compilation): Convert from a global struct to a subclass
4334 of rtl_opt_pass along with...
4335 (pass_data_rest_of_compilation): ...new pass_data instance and...
4336 (make_pass_rest_of_compilation): ...new function.
4337 (pass_postreload): Convert from a global struct to a subclass of
4338 rtl_opt_pass along with...
4339 (pass_data_postreload): ...new pass_data instance and...
4340 (make_pass_postreload): ...new function.
4341 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
4342 subclass of rtl_opt_pass along with...
4343 (pass_data_gcse2): ...new pass_data instance and...
4344 (make_pass_gcse2): ...new function.
4345 * postreload.c (pass_postreload_cse): Convert from a global struct to
4346 a subclass of rtl_opt_pass along with...
4347 (pass_data_postreload_cse): ...new pass_data instance and...
4348 (make_pass_postreload_cse): ...new function.
4349 * predict.c (pass_profile): Convert from a global struct to a subclass
4350 of gimple_opt_pass along with...
4351 (pass_data_profile): ...new pass_data instance and...
4352 (make_pass_profile): ...new function.
4353 (pass_strip_predict_hints): Convert from a global struct to a subclass
4354 of gimple_opt_pass along with...
4355 (pass_data_strip_predict_hints): ...new pass_data instance and...
4356 (make_pass_strip_predict_hints): ...new function.
4357 * recog.c (pass_peephole2): Convert from a global struct to a subclass
4358 of rtl_opt_pass along with...
4359 (pass_data_peephole2): ...new pass_data instance and...
4360 (make_pass_peephole2): ...new function.
4361 (pass_split_all_insns): Convert from a global struct to a subclass of
4362 rtl_opt_pass along with...
4363 (pass_data_split_all_insns): ...new pass_data instance and...
4364 (make_pass_split_all_insns): ...new function.
4365 (pass_split_after_reload): Convert from a global struct to a subclass
4366 of rtl_opt_pass along with...
4367 (pass_data_split_after_reload): ...new pass_data instance and...
4368 (make_pass_split_after_reload): ...new function.
4369 (pass_split_before_regstack): Convert from a global struct to a
4370 subclass of rtl_opt_pass along with...
4371 (pass_data_split_before_regstack): ...new pass_data instance and...
4372 (make_pass_split_before_regstack): ...new function.
4373 (pass_split_before_sched2): Convert from a global struct to a subclass
4374 of rtl_opt_pass along with...
4375 (pass_data_split_before_sched2): ...new pass_data instance and...
4376 (make_pass_split_before_sched2): ...new function.
4377 (pass_split_for_shorten_branches): Convert from a global struct to a
4378 subclass of rtl_opt_pass along with...
4379 (pass_data_split_for_shorten_branches): ...new pass_data instance
4380 and...
4381 (make_pass_split_for_shorten_branches): ...new function.
4382 * ree.c (pass_ree): Convert from a global struct to a subclass of
4383 rtl_opt_pass along with...
4384 (pass_data_ree): ...new pass_data instance and...
4385 (make_pass_ree): ...new function.
4386 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
4387 subclass of rtl_opt_pass along with...
4388 (pass_data_stack_regs): ...new pass_data instance and...
4389 (make_pass_stack_regs): ...new function.
4390 (pass_stack_regs_run): Convert from a global struct to a subclass of
4391 rtl_opt_pass along with...
4392 (pass_data_stack_regs_run): ...new pass_data instance and...
4393 (make_pass_stack_regs_run): ...new function.
4394 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4395 subclass of rtl_opt_pass along with...
4396 (pass_data_cprop_hardreg): ...new pass_data instance and...
4397 (make_pass_cprop_hardreg): ...new function.
4398 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4399 subclass of rtl_opt_pass along with...
4400 (pass_data_reginfo_init): ...new pass_data instance and...
4401 (make_pass_reginfo_init): ...new function.
4402 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4403 of rtl_opt_pass along with...
4404 (pass_data_regmove): ...new pass_data instance and...
4405 (make_pass_regmove): ...new function.
4406 * regrename.c (pass_regrename): Convert from a global struct to a
4407 subclass of rtl_opt_pass along with...
4408 (pass_data_regrename): ...new pass_data instance and...
4409 (make_pass_regrename): ...new function.
4410 * reorg.c (pass_delay_slots): Convert from a global struct to a
4411 subclass of rtl_opt_pass along with...
4412 (pass_data_delay_slots): ...new pass_data instance and...
4413 (make_pass_delay_slots): ...new function.
4414 (pass_machine_reorg): Convert from a global struct to a subclass of
4415 rtl_opt_pass along with...
4416 (pass_data_machine_reorg): ...new pass_data instance and...
4417 (make_pass_machine_reorg): ...new function.
4418 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4419 of rtl_opt_pass along with...
4420 (pass_data_sched): ...new pass_data instance and...
4421 (make_pass_sched): ...new function.
4422 (pass_sched2): Convert from a global struct to a subclass of
4423 rtl_opt_pass along with...
4424 (pass_data_sched2): ...new pass_data instance and...
4425 (make_pass_sched2): ...new function.
4426 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4427 to a subclass of rtl_opt_pass along with...
4428 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4429 (make_pass_stack_ptr_mod): ...new function.
4430 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4431 to a subclass of rtl_opt_pass along with...
4432 (pass_data_rtl_store_motion): ...new pass_data instance and...
4433 (make_pass_rtl_store_motion): ...new function.
4434 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4435 of gimple_opt_pass along with...
4436 (pass_data_tracer): ...new pass_data instance and...
4437 (make_pass_tracer): ...new function.
4438 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4439 to a subclass of gimple_opt_pass along with...
4440 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4441 (make_pass_diagnose_tm_blocks): ...new function.
4442 (pass_lower_tm): Convert from a global struct to a subclass of
4443 gimple_opt_pass along with...
4444 (pass_data_lower_tm): ...new pass_data instance and...
4445 (make_pass_lower_tm): ...new function.
4446 (pass_tm_init): Convert from a global struct to a subclass of
4447 gimple_opt_pass along with...
4448 (pass_data_tm_init): ...new pass_data instance and...
4449 (make_pass_tm_init): ...new function.
4450 (pass_tm_mark): Convert from a global struct to a subclass of
4451 gimple_opt_pass along with...
4452 (pass_data_tm_mark): ...new pass_data instance and...
4453 (make_pass_tm_mark): ...new function.
4454 (pass_tm_edges): Convert from a global struct to a subclass of
4455 gimple_opt_pass along with...
4456 (pass_data_tm_edges): ...new pass_data instance and...
4457 (make_pass_tm_edges): ...new function.
4458 (pass_tm_memopt): Convert from a global struct to a subclass of
4459 gimple_opt_pass along with...
4460 (pass_data_tm_memopt): ...new pass_data instance and...
4461 (make_pass_tm_memopt): ...new function.
4462 (pass_ipa_tm): Convert from a global struct to a subclass of
4463 simple_ipa_opt_pass along with...
4464 (pass_data_ipa_tm): ...new pass_data instance and...
4465 (make_pass_ipa_tm): ...new function.
4466 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4467 subclass of gimple_opt_pass along with...
4468 (pass_data_call_cdce): ...new pass_data instance and...
4469 (make_pass_call_cdce): ...new function.
4470 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4471 subclass of gimple_opt_pass along with...
4472 (pass_data_build_cfg): ...new pass_data instance and...
4473 (make_pass_build_cfg): ...new function.
4474 (pass_split_crit_edges): Convert from a global struct to a subclass of
4475 gimple_opt_pass along with...
4476 (pass_data_split_crit_edges): ...new pass_data instance and...
4477 (make_pass_split_crit_edges): ...new function.
4478 (pass_warn_function_return): Convert from a global struct to a
4479 subclass of gimple_opt_pass along with...
4480 (pass_data_warn_function_return): ...new pass_data instance and...
4481 (make_pass_warn_function_return): ...new function.
4482 (pass_warn_function_noreturn): Convert from a global struct to a
4483 subclass of gimple_opt_pass along with...
4484 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4485 (make_pass_warn_function_noreturn): ...new function.
4486 (pass_warn_unused_result): Convert from a global struct to a subclass
4487 of gimple_opt_pass along with...
4488 (pass_data_warn_unused_result): ...new pass_data instance and...
4489 (make_pass_warn_unused_result): ...new function.
4490 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4491 a subclass of gimple_opt_pass along with...
4492 (pass_data_merge_phi): ...new pass_data instance and...
4493 (make_pass_merge_phi): ...new function.
4494 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4495 a subclass of gimple_opt_pass along with...
4496 (pass_data_lower_complex): ...new pass_data instance and...
4497 (make_pass_lower_complex): ...new function.
4498 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4499 gimple_opt_pass along with...
4500 (pass_data_lower_complex_O0): ...new pass_data instance and...
4501 (make_pass_lower_complex_O0): ...new function.
4502 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4503 subclass of gimple_opt_pass along with...
4504 (pass_data_lower_eh): ...new pass_data instance and...
4505 (make_pass_lower_eh): ...new function.
4506 (pass_refactor_eh): Convert from a global struct to a subclass of
4507 gimple_opt_pass along with...
4508 (pass_data_refactor_eh): ...new pass_data instance and...
4509 (make_pass_refactor_eh): ...new function.
4510 (pass_lower_resx): Convert from a global struct to a subclass of
4511 gimple_opt_pass along with...
4512 (pass_data_lower_resx): ...new pass_data instance and...
4513 (make_pass_lower_resx): ...new function.
4514 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4515 of gimple_opt_pass along with...
4516 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4517 (make_pass_lower_eh_dispatch): ...new function.
4518 (pass_cleanup_eh): Convert from a global struct to a subclass of
4519 gimple_opt_pass along with...
4520 (pass_data_cleanup_eh): ...new pass_data instance and...
4521 (make_pass_cleanup_eh): ...new function.
4522 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4523 to a subclass of simple_ipa_opt_pass along with...
4524 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4525 (make_pass_ipa_lower_emutls): ...new function.
4526 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4527 a subclass of gimple_opt_pass along with...
4528 (pass_data_if_conversion): ...new pass_data instance and...
4529 (make_pass_if_conversion): ...new function.
4530 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4531 subclass of gimple_opt_pass along with...
4532 (pass_data_build_ssa): ...new pass_data instance and...
4533 (make_pass_build_ssa): ...new function.
4534 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4535 global struct to a subclass of gimple_opt_pass along with...
4536 (pass_data_loop_distribution): ...new pass_data instance and...
4537 (make_pass_loop_distribution): ...new function.
4538 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4539 subclass of gimple_opt_pass along with...
4540 (pass_data_mudflap_1): ...new pass_data instance and...
4541 (make_pass_mudflap_1): ...new function.
4542 (pass_mudflap_2): Convert from a global struct to a subclass of
4543 gimple_opt_pass along with...
4544 (pass_data_mudflap_2): ...new pass_data instance and...
4545 (make_pass_mudflap_2): ...new function.
4546 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4547 subclass of gimple_opt_pass along with...
4548 (pass_data_mudflap_1): ...new pass_data instance and...
4549 (make_pass_mudflap_1): ...new function.
4550 (pass_mudflap_2): Convert from a global struct to a subclass of
4551 gimple_opt_pass along with...
4552 (pass_data_mudflap_2): ...new pass_data instance and...
4553 (make_pass_mudflap_2): ...new function.
4554 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4555 gimple_opt_pass along with...
4556 (pass_data_nrv): ...new pass_data instance and...
4557 (make_pass_nrv): ...new function.
4558 (pass_return_slot): Convert from a global struct to a subclass of
4559 gimple_opt_pass along with...
4560 (pass_data_return_slot): ...new pass_data instance and...
4561 (make_pass_return_slot): ...new function.
4562 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4563 to a subclass of gimple_opt_pass along with...
4564 (pass_data_object_sizes): ...new pass_data instance and...
4565 (make_pass_object_sizes): ...new function.
4566 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4567 global struct to a subclass of gimple_opt_pass along with...
4568 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4569 and...
4570 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4571 (pass_fixup_cfg): Convert from a global struct to a subclass of
4572 gimple_opt_pass along with...
4573 (pass_data_fixup_cfg): ...new pass_data instance and...
4574 (make_pass_fixup_cfg): ...new function.
4575 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4576 (make_pass_mudflap_1): ...new function.
4577 (pass_mudflap_2): Replace declaration with that of...
4578 (make_pass_mudflap_2): ...new function.
4579 (pass_asan): Replace declaration with that of...
4580 (make_pass_asan): ...new function.
4581 (pass_asan_O0): Replace declaration with that of...
4582 (make_pass_asan_O0): ...new function.
4583 (pass_tsan): Replace declaration with that of...
4584 (make_pass_tsan): ...new function.
4585 (pass_tsan_O0): Replace declaration with that of...
4586 (make_pass_tsan_O0): ...new function.
4587 (pass_lower_cf): Replace declaration with that of...
4588 (make_pass_lower_cf): ...new function.
4589 (pass_refactor_eh): Replace declaration with that of...
4590 (make_pass_refactor_eh): ...new function.
4591 (pass_lower_eh): Replace declaration with that of...
4592 (make_pass_lower_eh): ...new function.
4593 (pass_lower_eh_dispatch): Replace declaration with that of...
4594 (make_pass_lower_eh_dispatch): ...new function.
4595 (pass_lower_resx): Replace declaration with that of...
4596 (make_pass_lower_resx): ...new function.
4597 (pass_build_cfg): Replace declaration with that of...
4598 (make_pass_build_cfg): ...new function.
4599 (pass_early_tree_profile): Replace declaration with that of...
4600 (make_pass_early_tree_profile): ...new function.
4601 (pass_cleanup_eh): Replace declaration with that of...
4602 (make_pass_cleanup_eh): ...new function.
4603 (pass_sra): Replace declaration with that of...
4604 (make_pass_sra): ...new function.
4605 (pass_sra_early): Replace declaration with that of...
4606 (make_pass_sra_early): ...new function.
4607 (pass_early_ipa_sra): Replace declaration with that of...
4608 (make_pass_early_ipa_sra): ...new function.
4609 (pass_tail_recursion): Replace declaration with that of...
4610 (make_pass_tail_recursion): ...new function.
4611 (pass_tail_calls): Replace declaration with that of...
4612 (make_pass_tail_calls): ...new function.
4613 (pass_tree_loop): Replace declaration with that of...
4614 (make_pass_tree_loop): ...new function.
4615 (pass_tree_loop_init): Replace declaration with that of...
4616 (make_pass_tree_loop_init): ...new function.
4617 (pass_lim): Replace declaration with that of...
4618 (make_pass_lim): ...new function.
4619 (pass_tree_unswitch): Replace declaration with that of...
4620 (make_pass_tree_unswitch): ...new function.
4621 (pass_predcom): Replace declaration with that of...
4622 (make_pass_predcom): ...new function.
4623 (pass_iv_canon): Replace declaration with that of...
4624 (make_pass_iv_canon): ...new function.
4625 (pass_scev_cprop): Replace declaration with that of...
4626 (make_pass_scev_cprop): ...new function.
4627 (pass_empty_loop): Replace declaration with that of...
4628 (make_pass_empty_loop): ...new function.
4629 (pass_record_bounds): Replace declaration with that of...
4630 (make_pass_record_bounds): ...new function.
4631 (pass_graphite): Replace declaration with that of...
4632 (make_pass_graphite): ...new function.
4633 (pass_graphite_transforms): Replace declaration with that of...
4634 (make_pass_graphite_transforms): ...new function.
4635 (pass_if_conversion): Replace declaration with that of...
4636 (make_pass_if_conversion): ...new function.
4637 (pass_loop_distribution): Replace declaration with that of...
4638 (make_pass_loop_distribution): ...new function.
4639 (pass_vectorize): Replace declaration with that of...
4640 (make_pass_vectorize): ...new function.
4641 (pass_slp_vectorize): Replace declaration with that of...
4642 (make_pass_slp_vectorize): ...new function.
4643 (pass_complete_unroll): Replace declaration with that of...
4644 (make_pass_complete_unroll): ...new function.
4645 (pass_complete_unrolli): Replace declaration with that of...
4646 (make_pass_complete_unrolli): ...new function.
4647 (pass_parallelize_loops): Replace declaration with that of...
4648 (make_pass_parallelize_loops): ...new function.
4649 (pass_loop_prefetch): Replace declaration with that of...
4650 (make_pass_loop_prefetch): ...new function.
4651 (pass_iv_optimize): Replace declaration with that of...
4652 (make_pass_iv_optimize): ...new function.
4653 (pass_tree_loop_done): Replace declaration with that of...
4654 (make_pass_tree_loop_done): ...new function.
4655 (pass_ch): Replace declaration with that of...
4656 (make_pass_ch): ...new function.
4657 (pass_ccp): Replace declaration with that of...
4658 (make_pass_ccp): ...new function.
4659 (pass_phi_only_cprop): Replace declaration with that of...
4660 (make_pass_phi_only_cprop): ...new function.
4661 (pass_build_ssa): Replace declaration with that of...
4662 (make_pass_build_ssa): ...new function.
4663 (pass_build_alias): Replace declaration with that of...
4664 (make_pass_build_alias): ...new function.
4665 (pass_build_ealias): Replace declaration with that of...
4666 (make_pass_build_ealias): ...new function.
4667 (pass_dominator): Replace declaration with that of...
4668 (make_pass_dominator): ...new function.
4669 (pass_dce): Replace declaration with that of...
4670 (make_pass_dce): ...new function.
4671 (pass_dce_loop): Replace declaration with that of...
4672 (make_pass_dce_loop): ...new function.
4673 (pass_cd_dce): Replace declaration with that of...
4674 (make_pass_cd_dce): ...new function.
4675 (pass_call_cdce): Replace declaration with that of...
4676 (make_pass_call_cdce): ...new function.
4677 (pass_merge_phi): Replace declaration with that of...
4678 (make_pass_merge_phi): ...new function.
4679 (pass_split_crit_edges): Replace declaration with that of...
4680 (make_pass_split_crit_edges): ...new function.
4681 (pass_pre): Replace declaration with that of...
4682 (make_pass_pre): ...new function.
4683 (pass_profile): Replace declaration with that of...
4684 (make_pass_profile): ...new function.
4685 (pass_strip_predict_hints): Replace declaration with that of...
4686 (make_pass_strip_predict_hints): ...new function.
4687 (pass_lower_complex_O0): Replace declaration with that of...
4688 (make_pass_lower_complex_O0): ...new function.
4689 (pass_lower_complex): Replace declaration with that of...
4690 (make_pass_lower_complex): ...new function.
4691 (pass_lower_vector): Replace declaration with that of...
4692 (make_pass_lower_vector): ...new function.
4693 (pass_lower_vector_ssa): Replace declaration with that of...
4694 (make_pass_lower_vector_ssa): ...new function.
4695 (pass_lower_omp): Replace declaration with that of...
4696 (make_pass_lower_omp): ...new function.
4697 (pass_diagnose_omp_blocks): Replace declaration with that of...
4698 (make_pass_diagnose_omp_blocks): ...new function.
4699 (pass_expand_omp): Replace declaration with that of...
4700 (make_pass_expand_omp): ...new function.
4701 (pass_expand_omp_ssa): Replace declaration with that of...
4702 (make_pass_expand_omp_ssa): ...new function.
4703 (pass_object_sizes): Replace declaration with that of...
4704 (make_pass_object_sizes): ...new function.
4705 (pass_strlen): Replace declaration with that of...
4706 (make_pass_strlen): ...new function.
4707 (pass_fold_builtins): Replace declaration with that of...
4708 (make_pass_fold_builtins): ...new function.
4709 (pass_stdarg): Replace declaration with that of...
4710 (make_pass_stdarg): ...new function.
4711 (pass_early_warn_uninitialized): Replace declaration with that of...
4712 (make_pass_early_warn_uninitialized): ...new function.
4713 (pass_late_warn_uninitialized): Replace declaration with that of...
4714 (make_pass_late_warn_uninitialized): ...new function.
4715 (pass_cse_reciprocals): Replace declaration with that of...
4716 (make_pass_cse_reciprocals): ...new function.
4717 (pass_cse_sincos): Replace declaration with that of...
4718 (make_pass_cse_sincos): ...new function.
4719 (pass_optimize_bswap): Replace declaration with that of...
4720 (make_pass_optimize_bswap): ...new function.
4721 (pass_optimize_widening_mul): Replace declaration with that of...
4722 (make_pass_optimize_widening_mul): ...new function.
4723 (pass_warn_function_return): Replace declaration with that of...
4724 (make_pass_warn_function_return): ...new function.
4725 (pass_warn_function_noreturn): Replace declaration with that of...
4726 (make_pass_warn_function_noreturn): ...new function.
4727 (pass_cselim): Replace declaration with that of...
4728 (make_pass_cselim): ...new function.
4729 (pass_phiopt): Replace declaration with that of...
4730 (make_pass_phiopt): ...new function.
4731 (pass_forwprop): Replace declaration with that of...
4732 (make_pass_forwprop): ...new function.
4733 (pass_phiprop): Replace declaration with that of...
4734 (make_pass_phiprop): ...new function.
4735 (pass_tree_ifcombine): Replace declaration with that of...
4736 (make_pass_tree_ifcombine): ...new function.
4737 (pass_dse): Replace declaration with that of...
4738 (make_pass_dse): ...new function.
4739 (pass_nrv): Replace declaration with that of...
4740 (make_pass_nrv): ...new function.
4741 (pass_rename_ssa_copies): Replace declaration with that of...
4742 (make_pass_rename_ssa_copies): ...new function.
4743 (pass_sink_code): Replace declaration with that of...
4744 (make_pass_sink_code): ...new function.
4745 (pass_fre): Replace declaration with that of...
4746 (make_pass_fre): ...new function.
4747 (pass_check_data_deps): Replace declaration with that of...
4748 (make_pass_check_data_deps): ...new function.
4749 (pass_copy_prop): Replace declaration with that of...
4750 (make_pass_copy_prop): ...new function.
4751 (pass_vrp): Replace declaration with that of...
4752 (make_pass_vrp): ...new function.
4753 (pass_uncprop): Replace declaration with that of...
4754 (make_pass_uncprop): ...new function.
4755 (pass_return_slot): Replace declaration with that of...
4756 (make_pass_return_slot): ...new function.
4757 (pass_reassoc): Replace declaration with that of...
4758 (make_pass_reassoc): ...new function.
4759 (pass_rebuild_cgraph_edges): Replace declaration with that of...
4760 (make_pass_rebuild_cgraph_edges): ...new function.
4761 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
4762 (make_pass_remove_cgraph_callee_edges): ...new function.
4763 (pass_build_cgraph_edges): Replace declaration with that of...
4764 (make_pass_build_cgraph_edges): ...new function.
4765 (pass_local_pure_const): Replace declaration with that of...
4766 (make_pass_local_pure_const): ...new function.
4767 (pass_tracer): Replace declaration with that of...
4768 (make_pass_tracer): ...new function.
4769 (pass_warn_unused_result): Replace declaration with that of...
4770 (make_pass_warn_unused_result): ...new function.
4771 (pass_diagnose_tm_blocks): Replace declaration with that of...
4772 (make_pass_diagnose_tm_blocks): ...new function.
4773 (pass_lower_tm): Replace declaration with that of...
4774 (make_pass_lower_tm): ...new function.
4775 (pass_tm_init): Replace declaration with that of...
4776 (make_pass_tm_init): ...new function.
4777 (pass_tm_mark): Replace declaration with that of...
4778 (make_pass_tm_mark): ...new function.
4779 (pass_tm_memopt): Replace declaration with that of...
4780 (make_pass_tm_memopt): ...new function.
4781 (pass_tm_edges): Replace declaration with that of...
4782 (make_pass_tm_edges): ...new function.
4783 (pass_split_functions): Replace declaration with that of...
4784 (make_pass_split_functions): ...new function.
4785 (pass_feedback_split_functions): Replace declaration with that of...
4786 (make_pass_feedback_split_functions): ...new function.
4787 (pass_strength_reduction): Replace declaration with that of...
4788 (make_pass_strength_reduction): ...new function.
4789 (pass_ipa_lower_emutls): Replace declaration with that of...
4790 (make_pass_ipa_lower_emutls): ...new function.
4791 (pass_ipa_function_and_variable_visibility): Replace declaration with
4792 that of...
4793 (make_pass_ipa_function_and_variable_visibility): ...new function.
4794 (pass_ipa_tree_profile): Replace declaration with that of...
4795 (make_pass_ipa_tree_profile): ...new function.
4796 (pass_early_local_passes): Replace declaration with that of...
4797 (make_pass_early_local_passes): ...new function.
4798 (pass_ipa_whole_program_visibility): Replace declaration with that
4799 of...
4800 (make_pass_ipa_whole_program_visibility): ...new function.
4801 (pass_ipa_lto_gimple_out): Replace declaration with that of...
4802 (make_pass_ipa_lto_gimple_out): ...new function.
4803 (pass_ipa_increase_alignment): Replace declaration with that of...
4804 (make_pass_ipa_increase_alignment): ...new function.
4805 (pass_ipa_inline): Replace declaration with that of...
4806 (make_pass_ipa_inline): ...new function.
4807 (pass_ipa_free_lang_data): Replace declaration with that of...
4808 (make_pass_ipa_free_lang_data): ...new function.
4809 (pass_ipa_free_inline_summary): Replace declaration with that of...
4810 (make_pass_ipa_free_inline_summary): ...new function.
4811 (pass_ipa_cp): Replace declaration with that of...
4812 (make_pass_ipa_cp): ...new function.
4813 (pass_ipa_reference): Replace declaration with that of...
4814 (make_pass_ipa_reference): ...new function.
4815 (pass_ipa_pure_const): Replace declaration with that of...
4816 (make_pass_ipa_pure_const): ...new function.
4817 (pass_ipa_pta): Replace declaration with that of...
4818 (make_pass_ipa_pta): ...new function.
4819 (pass_ipa_lto_finish_out): Replace declaration with that of...
4820 (make_pass_ipa_lto_finish_out): ...new function.
4821 (pass_ipa_tm): Replace declaration with that of...
4822 (make_pass_ipa_tm): ...new function.
4823 (pass_ipa_profile): Replace declaration with that of...
4824 (make_pass_ipa_profile): ...new function.
4825 (pass_ipa_cdtor_merge): Replace declaration with that of...
4826 (make_pass_ipa_cdtor_merge): ...new function.
4827 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
4828 of...
4829 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4830 (pass_init_datastructures): Replace declaration with that of...
4831 (make_pass_init_datastructures): ...new function.
4832 (pass_fixup_cfg): Replace declaration with that of...
4833 (make_pass_fixup_cfg): ...new function.
4834 (pass_expand): Replace declaration with that of...
4835 (make_pass_expand): ...new function.
4836 (pass_instantiate_virtual_regs): Replace declaration with that of...
4837 (make_pass_instantiate_virtual_regs): ...new function.
4838 (pass_rtl_fwprop): Replace declaration with that of...
4839 (make_pass_rtl_fwprop): ...new function.
4840 (pass_rtl_fwprop_addr): Replace declaration with that of...
4841 (make_pass_rtl_fwprop_addr): ...new function.
4842 (pass_jump): Replace declaration with that of...
4843 (make_pass_jump): ...new function.
4844 (pass_jump2): Replace declaration with that of...
4845 (make_pass_jump2): ...new function.
4846 (pass_lower_subreg): Replace declaration with that of...
4847 (make_pass_lower_subreg): ...new function.
4848 (pass_cse): Replace declaration with that of...
4849 (make_pass_cse): ...new function.
4850 (pass_fast_rtl_dce): Replace declaration with that of...
4851 (make_pass_fast_rtl_dce): ...new function.
4852 (pass_ud_rtl_dce): Replace declaration with that of...
4853 (make_pass_ud_rtl_dce): ...new function.
4854 (pass_rtl_dce): Replace declaration with that of...
4855 (make_pass_rtl_dce): ...new function.
4856 (pass_rtl_dse1): Replace declaration with that of...
4857 (make_pass_rtl_dse1): ...new function.
4858 (pass_rtl_dse2): Replace declaration with that of...
4859 (make_pass_rtl_dse2): ...new function.
4860 (pass_rtl_dse3): Replace declaration with that of...
4861 (make_pass_rtl_dse3): ...new function.
4862 (pass_rtl_cprop): Replace declaration with that of...
4863 (make_pass_rtl_cprop): ...new function.
4864 (pass_rtl_pre): Replace declaration with that of...
4865 (make_pass_rtl_pre): ...new function.
4866 (pass_rtl_hoist): Replace declaration with that of...
4867 (make_pass_rtl_hoist): ...new function.
4868 (pass_rtl_store_motion): Replace declaration with that of...
4869 (make_pass_rtl_store_motion): ...new function.
4870 (pass_cse_after_global_opts): Replace declaration with that of...
4871 (make_pass_cse_after_global_opts): ...new function.
4872 (pass_rtl_ifcvt): Replace declaration with that of...
4873 (make_pass_rtl_ifcvt): ...new function.
4874 (pass_into_cfg_layout_mode): Replace declaration with that of...
4875 (make_pass_into_cfg_layout_mode): ...new function.
4876 (pass_outof_cfg_layout_mode): Replace declaration with that of...
4877 (make_pass_outof_cfg_layout_mode): ...new function.
4878 (pass_loop2): Replace declaration with that of...
4879 (make_pass_loop2): ...new function.
4880 (pass_rtl_loop_init): Replace declaration with that of...
4881 (make_pass_rtl_loop_init): ...new function.
4882 (pass_rtl_move_loop_invariants): Replace declaration with that of...
4883 (make_pass_rtl_move_loop_invariants): ...new function.
4884 (pass_rtl_unswitch): Replace declaration with that of...
4885 (make_pass_rtl_unswitch): ...new function.
4886 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
4887 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4888 (pass_rtl_doloop): Replace declaration with that of...
4889 (make_pass_rtl_doloop): ...new function.
4890 (pass_rtl_loop_done): Replace declaration with that of...
4891 (make_pass_rtl_loop_done): ...new function.
4892 (pass_web): Replace declaration with that of...
4893 (make_pass_web): ...new function.
4894 (pass_cse2): Replace declaration with that of...
4895 (make_pass_cse2): ...new function.
4896 (pass_df_initialize_opt): Replace declaration with that of...
4897 (make_pass_df_initialize_opt): ...new function.
4898 (pass_df_initialize_no_opt): Replace declaration with that of...
4899 (make_pass_df_initialize_no_opt): ...new function.
4900 (pass_reginfo_init): Replace declaration with that of...
4901 (make_pass_reginfo_init): ...new function.
4902 (pass_inc_dec): Replace declaration with that of...
4903 (make_pass_inc_dec): ...new function.
4904 (pass_stack_ptr_mod): Replace declaration with that of...
4905 (make_pass_stack_ptr_mod): ...new function.
4906 (pass_initialize_regs): Replace declaration with that of...
4907 (make_pass_initialize_regs): ...new function.
4908 (pass_combine): Replace declaration with that of...
4909 (make_pass_combine): ...new function.
4910 (pass_if_after_combine): Replace declaration with that of...
4911 (make_pass_if_after_combine): ...new function.
4912 (pass_ree): Replace declaration with that of...
4913 (make_pass_ree): ...new function.
4914 (pass_partition_blocks): Replace declaration with that of...
4915 (make_pass_partition_blocks): ...new function.
4916 (pass_match_asm_constraints): Replace declaration with that of...
4917 (make_pass_match_asm_constraints): ...new function.
4918 (pass_regmove): Replace declaration with that of...
4919 (make_pass_regmove): ...new function.
4920 (pass_split_all_insns): Replace declaration with that of...
4921 (make_pass_split_all_insns): ...new function.
4922 (pass_fast_rtl_byte_dce): Replace declaration with that of...
4923 (make_pass_fast_rtl_byte_dce): ...new function.
4924 (pass_lower_subreg2): Replace declaration with that of...
4925 (make_pass_lower_subreg2): ...new function.
4926 (pass_mode_switching): Replace declaration with that of...
4927 (make_pass_mode_switching): ...new function.
4928 (pass_sms): Replace declaration with that of...
4929 (make_pass_sms): ...new function.
4930 (pass_sched): Replace declaration with that of...
4931 (make_pass_sched): ...new function.
4932 (pass_ira): Replace declaration with that of...
4933 (make_pass_ira): ...new function.
4934 (pass_reload): Replace declaration with that of...
4935 (make_pass_reload): ...new function.
4936 (pass_clean_state): Replace declaration with that of...
4937 (make_pass_clean_state): ...new function.
4938 (pass_branch_prob): Replace declaration with that of...
4939 (make_pass_branch_prob): ...new function.
4940 (pass_value_profile_transformations): Replace declaration with that
4941 of...
4942 (make_pass_value_profile_transformations): ...new function.
4943 (pass_postreload_cse): Replace declaration with that of...
4944 (make_pass_postreload_cse): ...new function.
4945 (pass_gcse2): Replace declaration with that of...
4946 (make_pass_gcse2): ...new function.
4947 (pass_split_after_reload): Replace declaration with that of...
4948 (make_pass_split_after_reload): ...new function.
4949 (pass_branch_target_load_optimize1): Replace declaration with that
4950 of...
4951 (make_pass_branch_target_load_optimize1): ...new function.
4952 (pass_thread_prologue_and_epilogue): Replace declaration with that
4953 of...
4954 (make_pass_thread_prologue_and_epilogue): ...new function.
4955 (pass_stack_adjustments): Replace declaration with that of...
4956 (make_pass_stack_adjustments): ...new function.
4957 (pass_peephole2): Replace declaration with that of...
4958 (make_pass_peephole2): ...new function.
4959 (pass_if_after_reload): Replace declaration with that of...
4960 (make_pass_if_after_reload): ...new function.
4961 (pass_regrename): Replace declaration with that of...
4962 (make_pass_regrename): ...new function.
4963 (pass_cprop_hardreg): Replace declaration with that of...
4964 (make_pass_cprop_hardreg): ...new function.
4965 (pass_reorder_blocks): Replace declaration with that of...
4966 (make_pass_reorder_blocks): ...new function.
4967 (pass_branch_target_load_optimize2): Replace declaration with that
4968 of...
4969 (make_pass_branch_target_load_optimize2): ...new function.
4970 (pass_leaf_regs): Replace declaration with that of...
4971 (make_pass_leaf_regs): ...new function.
4972 (pass_split_before_sched2): Replace declaration with that of...
4973 (make_pass_split_before_sched2): ...new function.
4974 (pass_compare_elim_after_reload): Replace declaration with that of...
4975 (make_pass_compare_elim_after_reload): ...new function.
4976 (pass_sched2): Replace declaration with that of...
4977 (make_pass_sched2): ...new function.
4978 (pass_stack_regs): Replace declaration with that of...
4979 (make_pass_stack_regs): ...new function.
4980 (pass_stack_regs_run): Replace declaration with that of...
4981 (make_pass_stack_regs_run): ...new function.
4982 (pass_df_finish): Replace declaration with that of...
4983 (make_pass_df_finish): ...new function.
4984 (pass_compute_alignments): Replace declaration with that of...
4985 (make_pass_compute_alignments): ...new function.
4986 (pass_duplicate_computed_gotos): Replace declaration with that of...
4987 (make_pass_duplicate_computed_gotos): ...new function.
4988 (pass_variable_tracking): Replace declaration with that of...
4989 (make_pass_variable_tracking): ...new function.
4990 (pass_free_cfg): Replace declaration with that of...
4991 (make_pass_free_cfg): ...new function.
4992 (pass_machine_reorg): Replace declaration with that of...
4993 (make_pass_machine_reorg): ...new function.
4994 (pass_cleanup_barriers): Replace declaration with that of...
4995 (make_pass_cleanup_barriers): ...new function.
4996 (pass_delay_slots): Replace declaration with that of...
4997 (make_pass_delay_slots): ...new function.
4998 (pass_split_for_shorten_branches): Replace declaration with that of...
4999 (make_pass_split_for_shorten_branches): ...new function.
5000 (pass_split_before_regstack): Replace declaration with that of...
5001 (make_pass_split_before_regstack): ...new function.
5002 (pass_convert_to_eh_region_ranges): Replace declaration with that
5003 of...
5004 (make_pass_convert_to_eh_region_ranges): ...new function.
5005 (pass_shorten_branches): Replace declaration with that of...
5006 (make_pass_shorten_branches): ...new function.
5007 (pass_set_nothrow_function_flags): Replace declaration with that of...
5008 (make_pass_set_nothrow_function_flags): ...new function.
5009 (pass_dwarf2_frame): Replace declaration with that of...
5010 (make_pass_dwarf2_frame): ...new function.
5011 (pass_final): Replace declaration with that of...
5012 (make_pass_final): ...new function.
5013 (pass_rtl_seqabstr): Replace declaration with that of...
5014 (make_pass_rtl_seqabstr): ...new function.
5015 (pass_release_ssa_names): Replace declaration with that of...
5016 (make_pass_release_ssa_names): ...new function.
5017 (pass_early_inline): Replace declaration with that of...
5018 (make_pass_early_inline): ...new function.
5019 (pass_inline_parameters): Replace declaration with that of...
5020 (make_pass_inline_parameters): ...new function.
5021 (pass_update_address_taken): Replace declaration with that of...
5022 (make_pass_update_address_taken): ...new function.
5023 (pass_convert_switch): Replace declaration with that of...
5024 (make_pass_convert_switch): ...new function.
5025 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
5026 to a subclass of simple_ipa_opt_pass along with...
5027 (pass_data_ipa_tree_profile): ...new pass_data instance and...
5028 (make_pass_ipa_tree_profile): ...new function.
5029 * tree-sra.c (pass_sra_early): Convert from a global struct to a
5030 subclass of gimple_opt_pass along with...
5031 (pass_data_sra_early): ...new pass_data instance and...
5032 (make_pass_sra_early): ...new function.
5033 (pass_sra): Convert from a global struct to a subclass of
5034 gimple_opt_pass along with...
5035 (pass_data_sra): ...new pass_data instance and...
5036 (make_pass_sra): ...new function.
5037 (pass_early_ipa_sra): Convert from a global struct to a subclass of
5038 gimple_opt_pass along with...
5039 (pass_data_early_ipa_sra): ...new pass_data instance and...
5040 (make_pass_early_ipa_sra): ...new function.
5041 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
5042 subclass of gimple_opt_pass along with...
5043 (pass_data_ccp): ...new pass_data instance and...
5044 (make_pass_ccp): ...new function.
5045 (pass_fold_builtins): Convert from a global struct to a subclass of
5046 gimple_opt_pass along with...
5047 (pass_data_fold_builtins): ...new pass_data instance and...
5048 (make_pass_fold_builtins): ...new function.
5049 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
5050 subclass of gimple_opt_pass along with...
5051 (pass_data_copy_prop): ...new pass_data instance and...
5052 (make_pass_copy_prop): ...new function.
5053 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
5054 global struct to a subclass of gimple_opt_pass along with...
5055 (pass_data_rename_ssa_copies): ...new pass_data instance and...
5056 (make_pass_rename_ssa_copies): ...new function.
5057 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
5058 subclass of gimple_opt_pass along with...
5059 (pass_data_dce): ...new pass_data instance and...
5060 (make_pass_dce): ...new function.
5061 (pass_dce_loop): Convert from a global struct to a subclass of
5062 gimple_opt_pass along with...
5063 (pass_data_dce_loop): ...new pass_data instance and...
5064 (make_pass_dce_loop): ...new function.
5065 (pass_cd_dce): Convert from a global struct to a subclass of
5066 gimple_opt_pass along with...
5067 (pass_data_cd_dce): ...new pass_data instance and...
5068 (make_pass_cd_dce): ...new function.
5069 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
5070 subclass of gimple_opt_pass along with...
5071 (pass_data_dominator): ...new pass_data instance and...
5072 (make_pass_dominator): ...new function.
5073 (pass_phi_only_cprop): Convert from a global struct to a subclass of
5074 gimple_opt_pass along with...
5075 (pass_data_phi_only_cprop): ...new pass_data instance and...
5076 (make_pass_phi_only_cprop): ...new function.
5077 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
5078 subclass of gimple_opt_pass along with...
5079 (pass_data_dse): ...new pass_data instance and...
5080 (make_pass_dse): ...new function.
5081 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
5082 a subclass of gimple_opt_pass along with...
5083 (pass_data_forwprop): ...new pass_data instance and...
5084 (make_pass_forwprop): ...new function.
5085 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
5086 struct to a subclass of gimple_opt_pass along with...
5087 (pass_data_tree_ifcombine): ...new pass_data instance and...
5088 (make_pass_tree_ifcombine): ...new function.
5089 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
5090 subclass of gimple_opt_pass along with...
5091 (pass_data_ch): ...new pass_data instance and...
5092 (make_pass_ch): ...new function.
5093 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
5094 subclass of gimple_opt_pass along with...
5095 (pass_data_tree_loop): ...new pass_data instance and...
5096 (make_pass_tree_loop): ...new function.
5097 (pass_tree_loop_init): Convert from a global struct to a subclass of
5098 gimple_opt_pass along with...
5099 (pass_data_tree_loop_init): ...new pass_data instance and...
5100 (make_pass_tree_loop_init): ...new function.
5101 (pass_lim): Convert from a global struct to a subclass of
5102 gimple_opt_pass along with...
5103 (pass_data_lim): ...new pass_data instance and...
5104 (make_pass_lim): ...new function.
5105 (pass_tree_unswitch): Convert from a global struct to a subclass of
5106 gimple_opt_pass along with...
5107 (pass_data_tree_unswitch): ...new pass_data instance and...
5108 (make_pass_tree_unswitch): ...new function.
5109 (pass_predcom): Convert from a global struct to a subclass of
5110 gimple_opt_pass along with...
5111 (pass_data_predcom): ...new pass_data instance and...
5112 (make_pass_predcom): ...new function.
5113 (pass_vectorize): Convert from a global struct to a subclass of
5114 gimple_opt_pass along with...
5115 (pass_data_vectorize): ...new pass_data instance and...
5116 (make_pass_vectorize): ...new function.
5117 (pass_graphite): Convert from a global struct to a subclass of
5118 gimple_opt_pass along with...
5119 (pass_data_graphite): ...new pass_data instance and...
5120 (make_pass_graphite): ...new function.
5121 (pass_graphite_transforms): Convert from a global struct to a subclass
5122 of gimple_opt_pass along with...
5123 (pass_data_graphite_transforms): ...new pass_data instance and...
5124 (make_pass_graphite_transforms): ...new function.
5125 (pass_check_data_deps): Convert from a global struct to a subclass of
5126 gimple_opt_pass along with...
5127 (pass_data_check_data_deps): ...new pass_data instance and...
5128 (make_pass_check_data_deps): ...new function.
5129 (pass_iv_canon): Convert from a global struct to a subclass of
5130 gimple_opt_pass along with...
5131 (pass_data_iv_canon): ...new pass_data instance and...
5132 (make_pass_iv_canon): ...new function.
5133 (pass_scev_cprop): Convert from a global struct to a subclass of
5134 gimple_opt_pass along with...
5135 (pass_data_scev_cprop): ...new pass_data instance and...
5136 (make_pass_scev_cprop): ...new function.
5137 (pass_record_bounds): Convert from a global struct to a subclass of
5138 gimple_opt_pass along with...
5139 (pass_data_record_bounds): ...new pass_data instance and...
5140 (make_pass_record_bounds): ...new function.
5141 (pass_complete_unroll): Convert from a global struct to a subclass of
5142 gimple_opt_pass along with...
5143 (pass_data_complete_unroll): ...new pass_data instance and...
5144 (make_pass_complete_unroll): ...new function.
5145 (pass_complete_unrolli): Convert from a global struct to a subclass of
5146 gimple_opt_pass along with...
5147 (pass_data_complete_unrolli): ...new pass_data instance and...
5148 (make_pass_complete_unrolli): ...new function.
5149 (pass_parallelize_loops): Convert from a global struct to a subclass
5150 of gimple_opt_pass along with...
5151 (pass_data_parallelize_loops): ...new pass_data instance and...
5152 (make_pass_parallelize_loops): ...new function.
5153 (pass_loop_prefetch): Convert from a global struct to a subclass of
5154 gimple_opt_pass along with...
5155 (pass_data_loop_prefetch): ...new pass_data instance and...
5156 (make_pass_loop_prefetch): ...new function.
5157 (pass_iv_optimize): Convert from a global struct to a subclass of
5158 gimple_opt_pass along with...
5159 (pass_data_iv_optimize): ...new pass_data instance and...
5160 (make_pass_iv_optimize): ...new function.
5161 (pass_tree_loop_done): Convert from a global struct to a subclass of
5162 gimple_opt_pass along with...
5163 (pass_data_tree_loop_done): ...new pass_data instance and...
5164 (make_pass_tree_loop_done): ...new function.
5165 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
5166 struct to a subclass of gimple_opt_pass along with...
5167 (pass_data_cse_reciprocals): ...new pass_data instance and...
5168 (make_pass_cse_reciprocals): ...new function.
5169 (pass_cse_sincos): Convert from a global struct to a subclass of
5170 gimple_opt_pass along with...
5171 (pass_data_cse_sincos): ...new pass_data instance and...
5172 (make_pass_cse_sincos): ...new function.
5173 (pass_optimize_bswap): Convert from a global struct to a subclass of
5174 gimple_opt_pass along with...
5175 (pass_data_optimize_bswap): ...new pass_data instance and...
5176 (make_pass_optimize_bswap): ...new function.
5177 (pass_optimize_widening_mul): Convert from a global struct to a
5178 subclass of gimple_opt_pass along with...
5179 (pass_data_optimize_widening_mul): ...new pass_data instance and...
5180 (make_pass_optimize_widening_mul): ...new function.
5181 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
5182 subclass of gimple_opt_pass along with...
5183 (pass_data_phiopt): ...new pass_data instance and...
5184 (make_pass_phiopt): ...new function.
5185 (pass_cselim): Convert from a global struct to a subclass of
5186 gimple_opt_pass along with...
5187 (pass_data_cselim): ...new pass_data instance and...
5188 (make_pass_cselim): ...new function.
5189 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
5190 subclass of gimple_opt_pass along with...
5191 (pass_data_phiprop): ...new pass_data instance and...
5192 (make_pass_phiprop): ...new function.
5193 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
5194 subclass of gimple_opt_pass along with...
5195 (pass_data_pre): ...new pass_data instance and...
5196 (make_pass_pre): ...new function.
5197 (pass_fre): Convert from a global struct to a subclass of
5198 gimple_opt_pass along with...
5199 (pass_data_fre): ...new pass_data instance and...
5200 (make_pass_fre): ...new function.
5201 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
5202 subclass of gimple_opt_pass along with...
5203 (pass_data_reassoc): ...new pass_data instance and...
5204 (make_pass_reassoc): ...new function.
5205 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
5206 subclass of gimple_opt_pass along with...
5207 (pass_data_sink_code): ...new pass_data instance and...
5208 (make_pass_sink_code): ...new function.
5209 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
5210 subclass of gimple_opt_pass along with...
5211 (pass_data_strlen): ...new pass_data instance and...
5212 (make_pass_strlen): ...new function.
5213 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
5214 struct to a subclass of gimple_opt_pass along with...
5215 (pass_data_build_alias): ...new pass_data instance and...
5216 (make_pass_build_alias): ...new function.
5217 (pass_build_ealias): Convert from a global struct to a subclass of
5218 gimple_opt_pass along with...
5219 (pass_data_build_ealias): ...new pass_data instance and...
5220 (make_pass_build_ealias): ...new function.
5221 (pass_ipa_pta): Convert from a global struct to a subclass of
5222 simple_ipa_opt_pass along with...
5223 (pass_data_ipa_pta): ...new pass_data instance and...
5224 (make_pass_ipa_pta): ...new function.
5225 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
5226 subclass of gimple_opt_pass along with...
5227 (pass_data_uncprop): ...new pass_data instance and...
5228 (make_pass_uncprop): ...new function.
5229 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
5230 global struct to a subclass of gimple_opt_pass along with...
5231 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
5232 (make_pass_late_warn_uninitialized): ...new function.
5233 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
5234 to a subclass of gimple_opt_pass along with...
5235 (pass_data_init_datastructures): ...new pass_data instance and...
5236 (make_pass_init_datastructures): ...new function.
5237 (pass_early_warn_uninitialized): Convert from a global struct to a
5238 subclass of gimple_opt_pass along with...
5239 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5240 (make_pass_early_warn_uninitialized): ...new function.
5241 (pass_update_address_taken): Convert from a global struct to a
5242 subclass of gimple_opt_pass along with...
5243 (pass_data_update_address_taken): ...new pass_data instance and...
5244 (make_pass_update_address_taken): ...new function.
5245 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5246 struct to a subclass of gimple_opt_pass along with...
5247 (pass_data_release_ssa_names): ...new pass_data instance and...
5248 (make_pass_release_ssa_names): ...new function.
5249 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5250 subclass of gimple_opt_pass along with...
5251 (pass_data_stdarg): ...new pass_data instance and...
5252 (make_pass_stdarg): ...new function.
5253 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5254 global struct to a subclass of gimple_opt_pass along with...
5255 (pass_data_convert_switch): ...new pass_data instance and...
5256 (make_pass_convert_switch): ...new function.
5257 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5258 to a subclass of gimple_opt_pass along with...
5259 (pass_data_tail_recursion): ...new pass_data instance and...
5260 (make_pass_tail_recursion): ...new function.
5261 (pass_tail_calls): Convert from a global struct to a subclass of
5262 gimple_opt_pass along with...
5263 (pass_data_tail_calls): ...new pass_data instance and...
5264 (make_pass_tail_calls): ...new function.
5265 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5266 struct to a subclass of gimple_opt_pass along with...
5267 (pass_data_lower_vector): ...new pass_data instance and...
5268 (make_pass_lower_vector): ...new function.
5269 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5270 gimple_opt_pass along with...
5271 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5272 (make_pass_lower_vector_ssa): ...new function.
5273 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5274 to a subclass of gimple_opt_pass along with...
5275 (pass_data_slp_vectorize): ...new pass_data instance and...
5276 (make_pass_slp_vectorize): ...new function.
5277 (pass_ipa_increase_alignment): Convert from a global struct to a
5278 subclass of simple_ipa_opt_pass along with...
5279 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5280 (make_pass_ipa_increase_alignment): ...new function.
5281 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5282 gimple_opt_pass along with...
5283 (pass_data_vrp): ...new pass_data instance and...
5284 (make_pass_vrp): ...new function.
5285 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5286 subclass of simple_ipa_opt_pass along with...
5287 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5288 (make_pass_ipa_free_lang_data): ...new function.
5289 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5290 gimple_opt_pass along with...
5291 (pass_data_tsan): ...new pass_data instance and...
5292 (make_pass_tsan): ...new function.
5293 (pass_tsan_O0): Convert from a global struct to a subclass of
5294 gimple_opt_pass along with...
5295 (pass_data_tsan_O0): ...new pass_data instance and...
5296 (make_pass_tsan_O0): ...new function.
5297 * var-tracking.c (pass_variable_tracking): Convert from a global
5298 struct to a subclass of rtl_opt_pass along with...
5299 (pass_data_variable_tracking): ...new pass_data instance and...
5300 (make_pass_variable_tracking): ...new function.
5301 * web.c (pass_web): Convert from a global struct to a subclass of
5302 rtl_opt_pass along with...
5303 (pass_data_web): ...new pass_data instance and...
5304 (make_pass_web): ...new function.
5305 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5306 declaration with that of...
5307 (make_pass_mode_switch_use): ...new function.
5308 (pass_resolve_sw_modes): Replace declaration with that of...
5309 (make_pass_resolve_sw_modes): ...new function.
5310 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5311 from a global struct to a subclass of rtl_opt_pass along with...
5312 (pass_data_mode_switch_use): ...new pass_data instance and...
5313 (make_pass_mode_switch_use): ...new function.
5314 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5315 from a global struct to a subclass of rtl_opt_pass along with...
5316 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5317 (make_pass_resolve_sw_modes): ...new function.
5318 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5319 struct to a subclass of rtl_opt_pass along with...
5320 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5321 (make_pass_insert_vzeroupper): ...new function.
5322 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5323 global struct to a subclass of rtl_opt_pass along with...
5324 (pass_data_work_around_errata): ...new pass_data instance and...
5325 (make_pass_work_around_errata): ...new function.
5326 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5327 struct to a subclass of rtl_opt_pass along with...
5328 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
5329 (make_pass_mips_machine_reorg2): ...new function.
5330
6a389ed5
DM
53312013-08-05 David Malcolm <dmalcolm@redhat.com>
5332
5333 * passes.c (pass_manager::operator new): New.
5334
f7695dbf
DM
53352013-08-05 David Malcolm <dmalcolm@redhat.com>
5336
5337 Handwritten part of conversion of passes to C++ classes.
5338
5339 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
5340 (toplev.o): Add dep on PASS_MANAGER_H.
5341 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
5342 of early local pases to reflect this moving from a global to a
5343 member of gcc::pass_manager.
5344 (cgraph_add_new_function): Likewise.
5345 * lto-cgraph.c (lto_output_node): Update for conversion of
5346 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
5347 * passes.c (opt_pass::clone): New.
5348 (opt_pass::gate): New.
5349 (opt_pass::execute): New.
5350 (opt_pass::opt_pass): New.
5351 (pass_manager::execute_early_local_passes): New.
5352 (pass_manager::execute_pass_mode_switching): new.
5353 (finish_optimization_passes): Convert to...
5354 (pass_manager::finish_optimization_passes): ...this.
5355 (finish_optimization_passes): Update for conversion of passes to
5356 C++ classes.
5357 (register_dump_files_1): Use has_gate since we cannot portably
5358 check a vtable entry against NULL.
5359 (dump_one_pass): Likewise.
5360 (ipa_write_summaries_2): Likewise.
5361 (ipa_write_optimization_summaries_1): Likewise.
5362 (ipa_read_summaries_1): Likewise.
5363 (ipa_read_optimization_summaries_1): Likewise.
5364 (execute_ipa_stmt_fixups): Likewise.
5365 (pass_manager::pass_manager): Rewrite pass-creation, invoking
5366 pass-creation functions rather than wiring up globals, and
5367 storing the results in fields of pass_manager generated using
5368 pass-instances.def.
5369 (pass_manager::dump_profile_report): Update for conversion of
5370 passes to C++ classes.
5371 (pass_manager::execute_ipa_summary_passes): Likewise.
5372 (execute_one_ipa_transform_pass): Likewise.
5373 (execute_one_pass): Use has_gate and has_execute since we cannot
5374 portably check a vtable entry against NULL.
5375 * pass_manager.h (pass_manager::finish_optimization_passes): New.
5376 (pass_manager): Use pass-instances.def to add fields for the
5377 various pass instances.
5378 * toplev.c (finalize): Update for move of
5379 finish_optimization_passes to a method of gcc::pass_manager.
5380 * toplev.h (finish_optimization_passes): Move to method of class
5381 pass_manager.
5382 * tree-pass.h (struct pass_data): New.
03b0ee0a 5383 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM
5384 (opt_pass::gate): Convert to virtual function.
5385 (opt_pass::~opt_pass): New.
5386 (opt_pass::clone): New.
5387 (opt_pass::execute): Convert to virtual function.
5388 (opt_pass::opt_pass): New.
5389 (opt_pass::ctxt_): new.
5390 (gimple_opt_pass): Convert to subclass of opt_pass.
5391 (gimple_opt_pass::gimple_opt_pass): New.
5392 (rtl_opt_pass): Convert to subclass of opt_pass.
5393 (rtl_opt_pass::rtl_opt_pass): New.
5394 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5395 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5396 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5397 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5398 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5399 invocation of pass_mode_switching to reflect this moving from a
5400 global to a member of gcc::pass_manager.
5401 (ix86_option_override): Rework how pass_insert_vzeroupper is
5402 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 5403 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 5404
d0b2f831
RE
54052013-08-05 Richard Earnshaw <rearnsha@arm.com>
5406
5407 PR rtl-optimization/57708
5408 * recog.c (peep2_find_free_register): Validate all regs in a
5409 multi-reg mode.
5410
51a5c0c2
JH
54112013-08-05 Jan Hubicka <jh@suse.cz>
5412
5413 PR lto/57602
03b0ee0a
UB
5414 * cgraph.c (verify_cgraph_node): Accept local flags from other
5415 partitions.
51a5c0c2
JH
5416 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5417 (function_and_variable_visibility): Likewise.
5418 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5419
65f0a120
GDR
54202013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5421
5422 * graph.c (init_graph_slim_pretty_print): Remove.
5423 (print_graph_cfg): Do not call it. Use local pretty printer.
5424 (start_graph_dump): Likewise.
5425
b3f80694
GDR
54262013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5427
5428 * gimple-pretty-print.c (buffer): Remove.
5429 (initialized): Likewise.
5430 (maybe_init_pretty_print): Likewise.
5431 (print_gimple_stmt): Do not call it. Use non-static local
5432 pretty_printer variable.
5433 (print_gimple_expr): Likewise.
5434 (print_gimple_seq): Likewise.
5435 (gimple_dump_bb): Likewise.
5436
11a877b3
GDR
54372013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5438
5439 * asan.c (asan_pp): Remove.
5440 (asan_pp_initialized): Likewise.
5441 (asan_pp_initialize): Likewise.
5442 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5443 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5444 (asan_add_global): Likewise.
5445
b066401f
GDR
54462013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5447
5448 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5449 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5450 * pretty-print.c (pp_formatted_text_data): Likewise.
5451 (pp_indent): Rename from pp_base_indent.
5452 (pp_format): Rename from pp_base_format.
5453 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5454 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5455 (pp_flush): Rename from pp_base_flush.
5456 (pp_set_line_maximum_length): Rename from
5457 pp_base_set_line_maximum_length.
5458 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5459 (pp_set_prefix): Rename from pp_base_set_prefix.
5460 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5461 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5462 (pp_append_text): Rename from pp_base_append_text.
5463 (pp_formatted_text): Rename from pp_base_formatted_text.
5464 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5465 (pp_remaining_character_count_for_line): Rename from
5466 pp_base_remaining_character_count_for_line.
5467 (pp_newline): Rename from pp_base_newline.
5468 (pp_character): Rename from pp_base_character.
5469 (pp_string): Rename from pp_base_string.
5470 (pp_maybe_space): Rename from pp_base_maybe_space.
5471 * asan.c (asan_pp_string): Adjust.
5472 (asan_emit_stack_protection): Likewise.
5473 (asan_add_global): Likewise.
5474 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5475 * tree-mudflap.c (mf_varname_tree): Likewise.
5476 * tree-pretty-print.c (pp_tree_identifier): Rename from
5477 pp_base_tree_identifier.
5478 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5479 Declare as function.
5480
137a1a27
GDR
54812013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5482
5483 * pretty-print.h (pp_bar_bar): New.
5484 (pp_ampersand_ampersand): Likewise.
5485 (pp_less_equal): Likewise.
5486 (pp_greater_equal): Likewise.
5487 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5488 printer functions instead of pp_string or operators and punctuators.
5489 (dump_gimple_call): Likewise.
5490 (dump_gimple_omp_for): Likewise.
5491 (dump_gimple_transaction): Likewise.
5492 (dump_gimple_phi): Likewise.
5493 (pp_gimple_stmt_1): Likewise.
5494 * sched-vis.c (print_insn): Likewise.
5495 * tree-mudflap.c (mf_varname_tree): Likewise.
5496 * tree-pretty-print.c (dump_block_node): Likewise.
5497 (dump_generic_node): Likewise.
5498
815effe1
JH
54992013-08-02 Jan Hubicka <jh@suse.cz>
5500
5501 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5502 boundaries.
5503 * lto-streamer-out.c (tree_is_indexable): Results decls and
5504 parm decls are not indexable.
5505 (DFS_write_tree_body): Do not follow args and results.
5506 (hash_tree): Likewise.
5507 (output_functions): Rearrange so struct function is needed
5508 only when real body is output; be able to also ouptut abstract
5509 functions; output DECL_ARGUMENTS and DECL_RESULT.
5510 (lto_output): When not in WPA, ale store abstract functions.
5511 (write_symbol): Do not care about RESULT_DECL.
5512 (output_symbol_p): Handle correctly sbtract decls.
5513 * lto-streamer-in.c (input_function): Rearrange so struct
5514 function can be NULL at entry; allow streaming of
5515 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5516 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5517 sanity check during LTO.
5518 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5519 RESULT_DECl and DECL_ARGUMENTS.
5520 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5521 Likewise.
5522
07838b13
GDR
55232013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5524
5525 * pretty-print.h (pp_underscore): New.
5526 (pp_comma): Tidy.
5527 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5528 printer functions instead of pp_character.
5529 (dump_binary_rhs): Likewise.
5530 (dump_ternary_rhs): Likewise.
5531 (dump_gimple_call_args): Likewise.
5532 (pp_points_to_solution): Likewise.
5533 (dump_gimple_call): Likewise.
5534 (dump_gimple_switch): Likewise.
5535 (dump_gimple_cond): Likewise.
5536 (dump_gimple_bind): Likewise.
5537 (dump_gimple_try): Likewise.
5538 (dump_gimple_omp_for): Likewise.
5539 (dump_gimple_omp_continue): Likewise.
5540 (dump_gimple_omp_single): Likewise.
5541 (dump_gimple_omp_sections): Likewise.
5542 (dump_gimple_omp_block): Likewise.
5543 (dump_gimple_omp_critical): Likewise.
5544 (dump_gimple_transaction): Likewise.
5545 (dump_gimple_asm): Likewise.
5546 (dump_gimple_phi): Likewise.
5547 (dump_gimple_omp_parallel): Likewise.
5548 (dump_gimple_omp_task): Likewise.
5549 (dump_gimple_omp_atomic_load): Likewise.
5550 (dump_gimple_omp_atomic_store): Likewise.
5551 (dump_gimple_mem_ops): Likewise.
5552 (pp_gimple_stmt_1): Likewise.
5553 (pp_cfg_jump): Likewise.
5554 (dump_implicit_edges): Likewise.
5555 (gimple_dump_bb_for_graph): Likewise.
5556 * graph.c (draw_cfg_node): Likewise.
5557 * langhooks.c (lhd_print_error_function): Likewise.
5558 * sched-vis.c (print_exp): Likewise.
5559 (print_value): Likewise.
5560 (print_pattern): Likewise.
5561 (print_insn): Likewise.
5562 (rtl_dump_bb_for_graph): Likewise.
5563 * tree-pretty-print.c (dump_function_declaration): Likewise.
5564 (dump_array_domain): Likewise.
5565 (dump_omp_clause): Likewise.
5566 (dump_location): Likewise.
5567 (dump_generic_node): Likewise.
5568 (print_struct_decl): Likewise.
5569 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5570
bb0d2039
BS
55712013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5572
5573 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5574 candidate table when replacing a candidate statement.
5575 (replace_rhs_if_not_dup): Likewise.
5576 (replace_one_candidate): Likewise.
5577
bc0ec027 55782013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 5579 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH
5580
5581 * cgraphunit.c (add_new_function): Fix logic when adding from
5582 late IPA pass.
5583 (assemble_thunk): Rename to ...
03b0ee0a
UB
5584 (expand_thunk); .. this one; export; get it working with
5585 general functions; make produced gimple valid.
bc0ec027
JH
5586 * cgraph.h (expand_thunk): Declare.
5587
0e8853ee
JH
55882013-08-02 Jan Hubicka <jh@suse.cz>
5589
03b0ee0a
UB
5590 * ipa-cp.c (gather_context_independent_values): Use
5591 ipa_get_param_move_cost.
5592 (get_replacement_map): Remove PARAM; move parameter folding
5593 into tree-inline.c
0e8853ee
JH
5594 (create_specialized_node): Update.
5595 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
5596 assert that we have gimple body; update move_cost.
5597 (count_formal_params): Assert that we have gimple body.
5598 (ipa_dump_param): New function.
5599 (ipa_alloc_node_params): Break out from ...
5600 (ipa_initialize_node_params): ... here.
5601 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
5602 (ipa_write_node_info): Stream move costs.
5603 (ipa_read_node_info): Read move costs.
5604 (ipa_update_after_lto_read): Do not recompute node params.
5605 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
5606 (ipa_get_param): Check we are not in WPA.
5607 (ipa_get_param_move_cost): New.
5608 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
5609 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
5610 parm numbers to be present.
5611
ecee672b
VM
56122013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5613
5614 PR rtl-optimization/58048
5615 * lra-constraints.c (process_alt_operands): Don't check asm
5616 operand on register.
5617
13ec0527
EB
56182013-08-02 Eric Botcazou <ebotcazou@adacore.com>
5619
5620 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
5621 the implied StoreLoad barrier for atomic operations if before.
5622
af15184a
JH
56232013-08-02 Jan Hubicka <jh@suse.cz>
5624 Martin Liska <marxin.liska@gmail.com>
5625
03b0ee0a
UB
5626 * cgraph.c (cgraph_function_body_availability): Do not check
5627 cgraph flags.
af15184a
JH
5628 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
5629 symtab_node_availability): Declare.
5630 * ipa.c (can_replace_by_local_alias): New.
5631 (function_and_variable_visibility): Use it.
03b0ee0a
UB
5632 * symtab.c (symtab_for_node_and_aliases,
5633 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 5634
01e54ef8
VM
56352013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5636
5637 PR rtl-optimization/57963
03b0ee0a 5638 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM
5639 (lra_constraints): Use them.
5640
e2fd7ca7
SN
56412013-08-02 Sofiane Naci <sofiane.naci@arm.com>
5642
03b0ee0a
UB
5643 * config/arm/types.md (define_attr "type"): Add "load_acq"
5644 and "store_rel".
e2fd7ca7
SN
5645 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
5646 changes.
5647 (cortex_a53_store1): Likewise.
5648
0dcc26c3
JH
56492013-08-01 Jan Hubicka <jh@suse.cz>
5650
03b0ee0a
UB
5651 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
5652 partitions are not needed.
0dcc26c3 5653
0bd72901
UB
56542013-08-01 Uros Bizjak <ubizjak@gmail.com>
5655
5656 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
5657 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
5658 MAYBE_NON_Q_CLASS_P where appropriate.
5659
12123452
JH
56602013-08-01 Jan Hubicka <jh@suse.cz>
5661
5662 * cgraph.h (release_function_body): Declare.
5663 * tree.c (free_lang_data_in_decl): Free, parameters and return values
5664 of unused delcarations.
5665
ce852f9c
KT
56662013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5667
03b0ee0a
UB
5668 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
5669 RTL form when subtracting a constant.
ce852f9c 5670
c743b246
KT
56712013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5672
5673 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
5674 Generate canonical plus rtx with negated immediate instead of minus
5675 where appropriate.
5676 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
5677
c0c123ef
JH
56782013-08-01 Jan Hubicka <jh@suse.cz>
5679
5680 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
5681 (cgraph_release_function_body): Likewise.
5682 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5683 * cgraph.h (cgrpah_node): Rename abstract_and_needed
5684 to used_as_abstract_origin.
5685 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
5686 symbols used as abstract origins.
5687 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB
5688 * ipa.c (symtab_remove_unreachable_nodes): Recompute
5689 used_as_abstract_origin.
c0c123ef 5690 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB
5691 used_as_abstract_origin; be ready for DECL_RESULT and
5692 DECL_ARGUMENTS to be NULL.
c0c123ef 5693
0bd72901
UB
5694 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
5695 for abstract functions.
5696 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
5697 real symbols.
c0c123ef 5698
1f029433
JH
56992013-08-01 Jan Hubicka <jh@suse.cz>
5700
5701 * profile.c (compute_value_histograms): Fix thinko.
5702
c451f4d6
SN
57032013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5704
5705 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
5706 aarch-common-protos.h to extra_headers.
5707 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5708 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
5709 * config/aarch64/t-aarch64 (aarch-common.o): Define.
5710
34b512f5
SN
57112013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5712
5713 * config/aarch64/aarch64.md (define_attr "type"): Delete.
5714 Include "../arm/types.md". Define "type" attribute for all patterns.
5715 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
5716 attribute changes.
5717
d86e633a
MM
57182013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
5719
5720 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
5721 to support power8 load fusion.
5722 (fusion_gpr_mem_load): Likewise.
5723
5724 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
5725
5726 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
5727 declarations for power8 load fusion.
5728 (emit_fusion_gpr_load): Likewise.
5729
5730 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
5731 tuning for power8, turn on fusion mode by default. Turn on sign
5732 extending fusion mode if normal fusion mode is on, and we are at
5733 -O2 or -O3.
5734 (fusion_gpr_load_p): New function, return true if we can fuse an
5735 addis instruction with a dependent load to a GPR.
5736 (emit_fusion_gpr_load): Emit the instructions for power8 load
5737 fusion to GPRs.
5738
0bd72901 5739 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM
5740 (VSX load fusion peepholes): New peepholes to fuse together an
5741 addi instruction with a VSX load instruction.
5742
5743 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
5744 peepholes to fuse an addis instruction with a load to a GPR base
5745 register. If we are supporting sign extending fusions, convert
5746 sign extending loads to zero extending loads and add an explicit
5747 sign extension.
5748
c3f35647
SN
57492013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5750
5751 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
5752 aarch-common-protos.h to extra_headers.
5753 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5754 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
5755 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
5756 (arm_no_early_alu_shift_dep): Likewise.
5757 (arm_no_early_alu_shift_value_dep): Likewise.
5758 (arm_no_early_mul_dep): Likewise.
5759 (arm_no_early_store_addr_dep): Likewise.
5760 (arm_mac_accumulator_is_mul_result): Likewise.
5761 (arm_mac_accumulator_is_result): Likewise.
c3f35647 5762 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB
5763 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
5764 here to ...
c3f35647 5765 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
5766 (arm_no_early_alu_shift_dep): Likewise.
5767 (arm_no_early_alu_shift_value_dep): Likewise.
5768 (arm_no_early_mul_dep): Likewise.
5769 (arm_no_early_store_addr_dep): Likewise.
5770 (arm_mac_accumulator_is_mul_result): Likewise.
5771 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN
5772 * config/arm/aarch-common-protos.h: ... here. New file.
5773 * config/arm/t-arm (aarch-common.o): Define.
5774
57752013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5776
5777 * config/arm/arm.md: Include new file "types.md".
5778 (define_attr "type"): Move from here to ...
5779 (define_attr "mul32"): Likewise.
5780 (define_attr "mul64"): Likewise.
5781 * config/arm/types.md: ... here. New file.
5782
228c1313
SH
57832013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
5784
5785 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
5786 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
5787
d6e9046d 57882013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG
5789
5790 * gen-pass-instances.awk: Fix offset of substr().
5791
a167b052
DM
57922013-07-31 David Malcolm <dmalcolm@redhat.com>
5793
5794 * Makefile.in (pass-instances.def): New.
5795 (passes.o): Replace dependency on passes.def with one on
5796 pass-instances.def
5797
5798 * gen-pass-instances.awk: New.
5799
5800 * passes.c (pass_manager::pass_manager): Use pass-instances.def
5801 rather than passes.def, updating local definition of NEXT_PASS
5802 macro to add an extra NUM parameter (currently unused).
5803
315f8c0e
DM
58042013-07-30 David Malcolm <dmalcolm@redhat.com>
5805
5806 * Makefile.in (PASS_MANAGER_H): New.
5807 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
5808 (passes.o): Likewise.
5809 (statistics.o): Likewise.
5810 (cgraphunit.o): Likewise.
5811 (context.o): Depend on PASS_MANAGER_H.
5812
5813 * pass_manager.h: New.
5814
5815 * cgraphunit.c (cgraph_add_new_function): Update for moves
5816 of globals to fields of pass_manager.
5817 (analyze_function): Likewise.
5818 (expand_function): Likewise.
5819 (ipa_passes): Likewise.
5820 (compile): Likewise.
5821
5822 * context.c (context::context): New.
5823 * context.h (context::context): New.
5824 (context::get_passes): New.
5825 (context::passes_): New.
5826
5827 * lto-cgraph.c (input_node): Update for moves of globals to
5828 fields of pass_manager.
5829
5830 * passes.c (all_passes): Remove, in favor of a field of the
5831 same name within the new class pass_manager.
5832 (all_small_ipa_passes): Likewise.
5833 (all_lowering_passes): Likewise.
5834 (all_regular_ipa_passes): Likewise.
5835 (all_late_ipa_passes): Likewise.
5836 (all_lto_gen_passes): Likewise.
5837 (passes_by_id): Likewise.
5838 (passes_by_id_size): Likewise.
5839 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
5840 the new class pass_manager.
5841 (set_pass_for_id): Convert to...
5842 (pass_manager::set_pass_for_id): ...method.
5843 (get_pass_for_id): Convert to...
5844 (pass_manager::get_pass_for_id): ...method.
5845 (register_one_dump_file): Move body of implementation into...
5846 (pass_manager::register_one_dump_file): ...here.
5847 (register_dump_files_1): Convert to...
5848 (pass_manager::register_dump_files_1): ...method.
5849 (register_dump_files): Convert to...
5850 (pass_manager::register_dump_files): ...method.
5851 (create_pass_tab): Update for moves of globals to fields of
5852 pass_manager.
5853 (dump_passes): Move body of implementation into...
5854 (pass_manager::dump_passes): ...here.
5855 (register_pass): Move body of implementation into...
5856 (pass_manager::register_pass): ...here.
5857 (init_optimization_passes): Convert into...
5858 (pass_manager::pass_manager): ...constructor for new
5859 pass_manager class, and initialize the pass_lists array.
5860 (check_profile_consistency): Update for moves of globals to
5861 fields of pass_manager.
5862 (dump_profile_report): Move body of implementation into...
5863 (pass_manager::dump_profile_report): ...here.
5864 (ipa_write_summaries_1): Update for moves of pass lists from
5865 being globals to fields of pass_manager.
5866 (ipa_write_optimization_summaries): Likewise.
5867 (ipa_read_summaries): Likewise.
5868 (ipa_read_optimization_summaries): Likewise.
5869 (execute_all_ipa_stmt_fixups): Likewise.
5870
5871 * statistics.c (statistics_fini): Update for moves of globals to
5872 fields of pass_manager.
5873
5874 * toplev.c (general_init): Replace call to
5875 init_optimization_passes with construction of the pass_manager
5876 instance.
5877
5878 * tree-pass.h (all_passes): Remove, in favor of a field of the
5879 same name within the new class pass_manager.
5880 (all_small_ipa_passes): Likewise.
5881 (all_lowering_passes): Likewise.
5882 (all_regular_ipa_passes): Likewise.
5883 (all_lto_gen_passes): Likewise.
5884 (all_late_ipa_passes): Likewise.
5885 (passes_by_id): Likewise.
5886 (passes_by_id_size): Likewise.
5887 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
5888 the new class pass_manager.
5889 (get_pass_for_id): Remove.
5890
02cab1c1
RE
58912013-07-30 Richard Earnshaw <rearnsha@arm.com>
5892
5893 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
5894 configs.
5895
b2d3f886
RE
58962013-07-30 Richard Earnshaw <rearnsha@arm.com>
5897
5898 * arm.md (mulhi3): New expand pattern.
5899
605e86fa
JH
59002013-07-30 Jan Hubicka <jh@suse.cz>
5901 Martin Liska <marxin.liska@gmail.com>
5902
5903 * profile.c (compute_value_histograms): Do not ICE when
5904 there is mismatch only on some counters.
5905
14a87636
ZC
59062013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5907
5908 PR rtl-optimization/57637
5909 * function.c (move_insn_for_shrink_wrap): Also check the
5910 GEN set of the LIVE problem for the liveness analysis
5911 if it exists, otherwise give up.
5912
0100cd3f
BS
59132013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5914
5915 PR tree-optimization/57993
5916 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
5917 replaced statement in the candidate table.
5918 (phi_add_costs): Return infinite cost when the hidden basis does
5919 not dominate all phis on which the candidate is dependent.
5920 (replace_one_candidate): Record replaced statement in the
5921 candidate table.
5922
f0d811f7
JR
59232013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
5924
5925 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
5926 (ashlv2si3): New expander.
5927 (*ashlv2si3_i): New define_insn_and_split.
5928 * predicates.md (float_operation): Allow patterns with three
5929 basic sub-patterns.
5930
bba33211
JR
5931 PR rtl-optimization/58021
5932 * mode-switching.c (create_pre_exit): Always split off preceding
5933 insns if we are not at the basic block head.
5934
ff3f3951
MR
59352013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
5936
5937 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
5938 (UCLIBC_DYNAMIC_LINKER): New macro.
5939 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
5940 `-mnan=2008'.
5941 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
5942 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
5943 `-mnan=2008'.
5944 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
5945 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
5946 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
5947 for SF and DF modes. Use ieee_quad_format for TF mode.
5948 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
5949 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
5950 (mips_option_override): Handle `-mnan=legacy'.
5951 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
5952 `-mabs=2008' and `-mnan=2008'.
5953 (OPTION_DEFAULT_SPECS): Add "nan" default.
5954 (ASM_SPEC): Handle `-mnan='.
5955 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
5956 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
5957 comment accordingly.
5958 (neg<mode>2): Likewise.
5959 * config/mips/mips.opt (mabs, mnan): New options.
5960 * doc/install.texi (Configuration): Document `--with-nan=' option.
5961 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
5962 `-mnan=' options.
5963 (MIPS Options): Document them.
5964 * config.gcc <mips*-*-*>: Handle `--with-nan='.
5965 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
5966 * configure: Regenerate.
5967 * config.in: Regenerate.
5968
576e2f37
UB
59692013-07-29 Uros Bizjak <ubizjak@gmail.com>
5970
5971 * config/i386/i386.md (float post-reload splitters): Do not check
5972 for subregs of SSE registers.
5973
59742013-07-29 Uros Bizjak <ubizjak@gmail.com>
5975 H.J. Lu <hongjiu.lu@intel.com>
5976
5977 PR target/57954
5978 PR target/57988
5979 * config/i386/i386.md (post-reload splitter
5980 to avoid partial SSE reg dependency stalls): New pattern.
5981
a71f0749
DV
59822013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
5983
5984 * config/s390/s390.md ("movcc"): Swap load and store instructions.
5985
4c97f1cc
JR
59862013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
5987
5988 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
5989 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
5990
b0c733d8
CC
59912013-07-26 Cary Coutant <ccoutant@google.com>
5992
5993 * dwarf2out.c (die_checksum_ordered): Don't include template
5994 instantiations in signature.
5995 (is_template_parameter): New function.
5996 (is_template_instantiation): New function.
5997 (generate_skeleton_bottom_up): Don't include template instantiations
5998 in type unit DIE.
5999 (generate_skeleton): Likewise.
6000 (break_out_comdat_types): Move recursive call to break out nested
6001 types earlier.
6002 (prune_unused_types_mark_generic_parms_dies): Call
6003 is_template_parameter.
6004
040d8a1c
IB
60052013-07-26 Ian Bolton <ian.bolton@arm.com>
6006
6007 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
6008 uses vector registers.
89fdc743 6009 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 6010
85bd4ac6 60112013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 6012 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT
6013
6014 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 6015 where appropriate.
85bd4ac6
KT
6016 * config/arm/arm.md (movcond_addsi): New splitter.
6017
b6051207
SA
60182013-07-25 Sterling Augustine <saugustine@google.com>
6019
6020 * dwarf2out.c (size_of_pubnames): Move code to...
6021 (include_pubname_in_output): ...here. New.
6022 (want_pubnames): Rearrange.
6023 (output_pubnames): Call include_pubname_in_output. Move assertion.
6024
41a7c215
CM
60252013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6026
6027 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
6028
60292013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6030
6031 PR target/38836
6032 * doc/extend.texi: Remove obsolete builtins. Fix
6033 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
6034
49bde175
JH
60352013-07-25 Jan Hubicka <jh@suse.cz>
6036
6037 * cgraph.c (release_function_body): Break out from ...
6038 (cgraph_release_function_body): ... this one; also release DECL_RESULT
6039 and DECL_ARGUMENTS.
6040 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
6041 old_tree in the map.
6042 (create_specialized_node): Update.
6043 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
6044 into index.
41a7c215
CM
6045 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
6046 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH
6047 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
6048 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
6049 DECL_RESULT.
6050
4870352d
KT
60512013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6052
6053 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
6054 addsi3_carryin_alt2_<optab>): Correct output template.
6055
1d6aee1c
KT
60562013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6057
6058 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
6059 Adjust for arm_restrict_it.
6060 Remove trailing whitespace.
6061
c7cafd75
MK
60622013-07-25  Mark Kettenis  <kettenis@openbsd.org>
6063
017d38f5
MK
6064 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
6065 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
6066
c7cafd75
MK
6067 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
6068
00b7527b
VM
60692013-07-25 Vladimir Makarov <vmakarov@redhat.com>
6070
6071 PR rtl-optimization/57960
6072 * lra-constraints.c (process_alt_operands): Use the right mode
6073 when checking strict_low.
6074
4f63dfc6
JH
60752013-07-25 Jan Hubicka <jh@suse.cz>
6076
6077 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
6078 * cgraph.c (cgraph_remove_node): Do not release function body
6079 when in cgraph streaming.
6080 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM
6081 in other partitions are not considered reachable; fix handling of
6082 clones.
4f63dfc6 6083
e6dcfa18
RR
60842013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6085
6086 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
6087
0b93d3b6
RR
60882013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6089
6090 PR target/19599
6091 PR target/57731
bb80c2eb 6092 PR target/57837
e6dcfa18 6093 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 6094 Ss with US. Adjust output for v5 and v4t.
41a7c215 6095 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 6096
0b93d3b6
RR
6097 * config/arm/constraints.md ("Ss"): Rename to US.
6098
690688b3
TG
60992013-07-25 Terry Guo <terry.guo@arm.com>
6100
6101 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
6102 shift_add/shift_sub0/shift_sub1 RTXs.
6103
a4ad093b 61042013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 6105 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS
6106
6107 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
6108 (altivec_vpks<VI_char>ss): Likewise.
6109 (altivec_vpks<VI_char>us): Likewise.
6110 (altivec_vpku<VI_char>us): Likewise.
6111 (altivec_vpku<VI_char>um): Likewise.
6112
8c5005ce
DM
61132013-07-24 David Malcolm <dmalcolm@redhat.com>
6114
6115 Introduce context class.
6116
6117 * Makefile.in (CONTEXT_H): New.
6118 (OBJS): Add context.o.
6119 (toplev.o): Add CONTEXT_H to dependencies.
6120 (context.o): New.
6121
41a7c215 6122 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM
6123
6124 * context.c: New.
6125
6126 * context.h: New.
6127
ce4a9422
JR
61282013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
6129
6130 PR rtl-optimization/57968
6131 * mode-switching.c (create_pre_exit): Allow instructions that
6132 don't set a return register to need a non-exit mode.
6133
493f4c9e 61342013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 6135 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 6136
b5860fd3
WS
6137 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
6138 operands to vperm for little endian.
6139 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
6140 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 6141
73792b92 61422013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 6143 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS
6144
6145 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6146 two operands for little-endian.
6147
14c21302
SE
61482013-07-23 Steve Ellcey <sellcey@mips.com>
6149
6150 * config/mips/mips.c (mips_case_values_threshold): New.
6151 (TARGET_CASE_VALUES_THRESHOLD): Define.
6152
a6056198 61532013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 6154 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS
6155
6156 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
6157 selection of field for vector splat in little endian mode.
6158
dd7a40e1
MM
61592013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6160
6161 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
6162 expanders to rs6000.md.
6163 (ior<mode>3): Likewise.
6164 (and<mode>3): Likewise.
6165 (one_cmpl<mode>2): Likewise.
6166 (nor<mode>3): Likewise.
6167 (andc<mode>3): Likewise.
6168 (eqv<mode>3): Likewise.
6169 (nand<mode>3): Likewise.
6170 (orc<mode>3): Likewise.
6171
6172 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
6173 declaration.
6174
6175 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
6176 to split multi-word logical operations.
6177 (rs6000_split_logical_di): Likewise.
6178 (rs6000_split_logical): Likewise.
6179
6180 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
6181 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
6182 and allow TImode operations in 32-bit.
6183 (vsx_and<mode>3_64bit): Likewise.
6184 (vsx_ior<mode>3_32bit): Likewise.
6185 (vsx_ior<mode>3_64bit): Likewise.
6186 (vsx_xor<mode>3_32bit): Likewise.
6187 (vsx_xor<mode>3_64bit): Likewise.
6188 (vsx_one_cmpl<mode>2_32bit): Likewise.
6189 (vsx_one_cmpl<mode>2_64bit): Likewise.
6190 (vsx_nor<mode>3_32bit): Likewise.
6191 (vsx_nor<mode>3_64bit): Likewise.
6192 (vsx_andc<mode>3_32bit): Likewise.
6193 (vsx_andc<mode>3_64bit): Likewise.
6194 (vsx_eqv<mode>3_32bit): Likewise.
6195 (vsx_eqv<mode>3_64bit): Likewise.
6196 (vsx_nand<mode>3_32bit): Likewise.
6197 (vsx_nand<mode>3_64bit): Likewise.
6198 (vsx_orc<mode>3_32bit): Likewise.
6199 (vsx_orc<mode>3_64bit): Likewise.
6200
6201 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
6202 logical types in GPRs.
6203
6204 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
6205 logical insns to rs6000.md, and allow TImode operations in
6206 32-bit.
6207 (altivec_ior<mode>3): Likewise.
6208 (altivec_xor<mode>3): Likewise.
6209 (altivec_one_cmpl<mode>2): Likewise.
6210 (altivec_nor<mode>3): Likewise.
6211 (altivec_andc<mode>3): Likewise.
6212
6213 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
6214 attributes for moving the 128-bit logical operations into
6215 rs6000.md.
6216 (BOOL_REGS_OUTPUT): Likewise.
6217 (BOOL_REGS_OP1): Likewise.
6218 (BOOL_REGS_OP2): Likewise.
6219 (BOOL_REGS_UNARY): Likewise.
6220 (BOOL_REGS_AND_CR0): Likewise.
6221 (one_cmpl<mode>2): Add support for DI logical operations on
6222 32-bit, splitting the operations to 32-bit.
6223 (anddi3): Likewise.
6224 (iordi3): Likewise.
6225 (xordi3): Likewise.
6226 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
6227 changes to combine the 32/64-bit code, allow logical operations on
6228 TI mode in 32-bit, and to use similar match_operator patterns like
6229 scalar mode uses. Combine the Altivec and VSX code for logical
6230 operations, and move it here.
6231 (ior<mode>3, 128-bit types): Likewise.
6232 (xor<mode>3, 128-bit types): Likewise.
6233 (one_cmpl<mode>3, 128-bit types): Likewise.
6234 (nor<mode>3, 128-bit types): Likewise.
6235 (andc<mode>3, 128-bit types): Likewise.
6236 (eqv<mode>3, 128-bit types): Likewise.
6237 (nand<mode>3, 128-bit types): Likewise.
6238 (orc<mode>3, 128-bit types): Likewise.
6239 (and<mode>3_internal): Likewise.
6240 (bool<mode>3_internal): Likewise.
6241 (boolc<mode>3_internal1): Likewise.
6242 (boolc<mode>3_internal2): Likewise.
6243 (boolcc<mode>3_internal1): Likewise.
6244 (boolcc<mode>3_internal2): Likewise.
6245 (eqv<mode>3_internal1): Likewise.
6246 (eqv<mode>3_internal2): Likewise.
6247 (one_cmpl1<mode>3_internal): Likewise.
6248
12211b99 62492013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH
6250
6251 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6252 Rename flag_stack_usage to flag_stack_usage_info.
6253
12211b99 62542013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 6255
a5f257fd
ME
6256 * config/microblaze/sync.md: New file.
6257 * config/microblaze/microblaze.md: Include sync.md
6258 * config/microblaze/microblaze.c: Add print_operand 'y'.
6259 * config/microblaze/constraints.md: Add memory_contraint
6260 'Q' which is a single register.
450b0ebb 6261
49fba14f
EB
62622013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6263
6264 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6265
c7b3b99f
PCC
62662013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6267
6268 * reload.c (find_reloads): Exit loop once we find this operand
6269 cannot be reloaded somehow for this alternative.
6270
6271 * reload.c (find_reloads): Exit loop once we find a hard register.
6272
6273 * rtlanal.c (computed_jump_p): Exit loop once we find label
6274 reference is used.
6275
6276 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6277
6278 * cfgloopmanip.c (remove_path): Exit loop after setting
6279 irred_invalidated.
6280
6281 * gensupport.c (subst_dup): Avoid loop if code is not
6282 MATCH_DUP nor MATCH_OP_DUP.
6283
c8fbf1fa
NBJ
62842013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6285
6286 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6287
9259db42
YZ
62882013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6289
6290 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6291 true for SP_REGNUM if mode == ptr_mode.
6292 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6293 with value R0_REGNUM + 31.
6294
00edcfbe
YZ
62952013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6296
6297 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6298 pad pointer-typed argument downward.
6299
43be9a95
YZ
63002013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6301
6302 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6303 and __ILP32__ when the ILP32 model is in use.
6304
28514dda
YZ
63052013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6306
6307 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6308 (aarch64_load_symref_appropriately): In the case of
6309 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6310 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6311 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6312 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6313 if 'mode' doesn't equal to 'ptr_mode'.
6314 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6315 'vcall_offset'; change to call aarch64_emit_move differently depending
6316 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6317 to calculate the upper bound of 'vcall_offset'.
6318 (aarch64_cannot_force_const_mem): Change to also return true if
6319 mode != ptr_mode.
6320 (aarch64_legitimize_reload_address): In the case of large
6321 displacements, add new local variable 'xmode' and an assertion
6322 based on it; change to use 'xmode' to generate the new rtx and
6323 reload.
6324 (aarch64_asm_trampoline_template): Change to generate the template
6325 differently depending on TARGET_ILP32 or not; change to use
6326 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6327 (aarch64_trampoline_size): Removed.
6328 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
6329 and replace immediate literals with it. Change to use 'ptr_mode'
6330 instead of 'DImode' and call convert_memory_address if the mode
6331 of 'fnaddr' doesn't equal to 'ptr_mode'.
6332 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
6333 to output symbol.
6334 (aarch64_elf_asm_destructor): Likewise.
6335 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
6336 on TARGET_ILP32 instead of aarch64_trampoline_size.
6337 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
6338 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
6339 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
6340 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
6341 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
6342 (storewb_pair<GPI:mode>_<P:mode>): ... this.
6343 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
6344 depending on the value of 'mode'.
6345 (add_losym_<mode>): New.
6346 (ldr_got_small_<mode>): New, based on ldr_got_small.
6347 (ldr_got_small): Remove.
6348 (ldr_got_small_sidi): New.
6349 * config/aarch64/iterators.md (P): New.
6350 (PTR): Change to 'ptr_mode' in the condition.
6351
17a819cb
YZ
63522013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6353
6354 * config.gcc (aarch64*-*-*): Support --with-abi.
6355 (aarch64*-*-elf): Support --with-multilib-list.
6356 (aarch64*-*-linux*): Likewise.
6357 (supported_defaults): Add abi to aarch64*-*-*.
6358 * configure.ac: Mention AArch64 for --with-multilib-list.
6359 * configure: Re-generated.
6360 * config/aarch64/biarchilp32.h: New file.
6361 * config/aarch64/biarchlp64.h: New file.
6362 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
6363 (ABI_SPEC): Ditto.
6364 (MULTILIB_DEFAULTS): Ditto.
6365 (DRIVER_SELF_SPECS): Ditto.
6366 (ASM_SPEC): Update to also substitute -mabi.
6367 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
6368 file whose name depends on -mabi= and -mbig-endian.
6369 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
6370 TARGET_ILP32.
6371 (POINTER_SIZE): New define.
6372 (POINTERS_EXTEND_UNSIGNED): Ditto.
6373 (enum aarch64_abi_type): New enumeration tag.
6374 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
6375 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
6376 (TARGET_ILP32): New define.
6377 * config/aarch64/aarch64.opt (mabi): New.
6378 (aarch64_abi): New.
6379 (ilp32, lp64): New values for -mabi.
6380 * config/aarch64/t-aarch64 (comma): New define.
6381 (MULTILIB_OPTIONS): Ditto.
6382 (MULTILIB_DIRNAMES): Ditto.
6383 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
6384 * doc/invoke.texi: Document -mabi for AArch64.
6385
9f9cbdce
GJL
63862013-07-23 Georg-Johann Lay <avr@gjlay.de>
6387
6388 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
6389
7061977b 63902013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 6391 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS
6392
6393 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6394 endianness when selecting field to splat.
6395
75a41878
EC
63962013-07-22 Eric Christopher <echristo@gmail.com>
6397
6398 * dwarf2out.c (die_odr_checksum): New function to use
6399 CHECKSUM_ macros and ULEB128 for DIE tag.
6400 (generate_type_signature): Use.
41a7c215 6401
38ae58ca
EB
64022013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6403
6404 * config.gcc (sparc*-*-*): Accept leon3 processor.
6405 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6406 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6407 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6408 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6409 (mfix-ut699): Adjust comment.
6410 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6411 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6412 (CPP_CPU_SPEC): Likewise.
6413 (ASM_CPU_SPEC): Likewise.
6414 * config/sparc/sparc.c (leon3_cost): New constant.
6415 (sparc_option_override): Add leon3 support.
6416 (mem_ref): New function.
6417 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6418 (sparc_do_work_around_errata): Look into the instruction in the delay
6419 slot and adjust accordingly. Add fix for the data cache nullify issues
6420 of the UT699. Change insertion position for the NOP.
6421 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6422 (leon3_load): New reservation.
6423 (leon_store): Bump latency to 2.
6424 (grfpu): New automaton.
6425 (grfpu_alu): New unit.
6426 (grfpu_ds): Likewise.
6427 (leon_fp_alu): Adjust.
6428 (leon_fp_mult): Delete.
6429 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6430 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6431 * config/sparc/sparc.md (cpu): Add leon3.
6432 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6433 (swapsi): Likewise.
6434 (atomic_test_and_set): Likewise.
6435 (ldstub): Likewise.
6436
04dfc6df
JU
64372013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6438
6439 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6440 default for R5900 targets.
6441 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6442 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6443 * config/mips/mips.c (mips_option_override): Report an error for
6444 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6445 for -march=r5900 -mhard-float.
6446
c7b3b99f 64472013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC
6448
6449 * df-problems.c (can_move_insns_across): Exit loop once we
6450 find a non-fixed, non-global register.
6451
6452 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6453 setting can_throw.
6454
6455 * omega.c (omega_eliminate_red): Break after setting red_found.
6456 (omega_problem_has_red_equations): Similarly after setting found.
6457 (omega_query_variable): Similarly after setting coupled.
6458
9b6e6981
MP
64592013-07-22 Marek Polacek <polacek@redhat.com>
6460
6461 * gimplify.c: Don't include gimple.h twice.
6462
5e5f7673
KT
64632013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6464
6465 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6466 instead of TARGET_THUMB1.
6467 (Pz): New constraint.
6468 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6469 encodings.
6470 (compare_negsi_si): Likewise.
6471 (compare_addsi2_op0): Likewise.
6472 (compare_addsi2_op1): Likewise.
6473 (addsi3_carryin_<optab>): Likewise.
6474 (addsi3_carryin_alt2_<optab>): Likewise.
6475 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6476 for arm_restrict_it.
6477 (subsi3_carryin): Likewise.
6478 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6479 (minmax_arithsi): Disable for arm_restrict_it.
6480 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6481 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6482 (satsi_<SAT:code>_shift): Likewise.
6483 (arm_shiftsi3): Add alternative for 16-bit encoding.
6484 (arm32_movhf): Disable for arm_restrict_it.
6485 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6486 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6487
859abddd
SN
64882013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6489
6490 * config/arm/arm.md (attribute "insn"): Delete.
6491 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6492 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6493 (not_shiftsi): Update for attribute change.
6494 (not_shiftsi_compare0): Likewise.
6495 (not_shiftsi_compare0_scratch): Likewise.
6496 (arm_one_cmplsi2): Likewise.
6497 (thumb1_one_cmplsi2): Likewise.
6498 (notsi_compare0): Likewise.
6499 (notsi_compare0_scratch): Likewise.
6500 (thumb1_movdi_insn): Likewise.
6501 (arm_movsi_insn): Likewise.
6502 (movhi_insn_arch4): Likewise.
6503 (movhi_bytes): Likewise.
6504 (arm_movqi_insn): Likewise.
6505 (thumb1_movqi_insn): Likewise.
6506 (arm32_movhf): Likewise.
6507 (thumb1_movhf): Likewise.
6508 (arm_movsf_soft_insn): Likewise.
6509 (thumb1_movsf_insn): Likewise.
6510 (thumb_movdf_insn): Likewise.
6511 (movsicc_insn): Likewise.
6512 (movsfcc_soft_insn): Likewise.
6513 (and_scc): Likewise.
6514 (cond_move): Likewise.
6515 (if_move_not): Likewise.
6516 (if_not_move): Likewise.
6517 (if_shift_move): Likewise.
6518 (if_move_shift): Likewise.
6519 (if_shift_shift): Likewise.
6520 (if_not_arith): Likewise.
6521 (if_arith_not): Likewise.
6522 (cond_move_not): Likewise.
6523 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6524 (neon_mov<mode>): Likewise.
6525 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6526 (thumb2_movsi_vfp): Likewise.
6527 (movsf_vfp): Likewise.
6528 (thumb2_movsf_vfp): Likewise.
371e77e3
UB
6529 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6530 change.
859abddd
SN
6531 (cortexa7_older_only): Likewise.
6532 (cortexa7_younger): Likewise.
6533 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6534 (1020alu_shift_op): Likewise.
6535 (1020alu_shift_reg_op): Likewise.
6536 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6537 (alu_shift_op): Likewise.
6538 (alu_shift_reg_op): Likewise.
6539 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6540 (11_alu_shift_op): Likewise.
6541 (11_alu_shift_reg_op): Likewise.
6542 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6543 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
6544 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6545 change.
859abddd
SN
6546 (cortex_a15_alu_shift): Likewise.
6547 (cortex_a15_alu_shift_reg): Likewise.
6548 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6549 (cortex_a5_alu_shift): Likewise.
371e77e3
UB
6550 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6551 change.
859abddd 6552 (cortex_a53_alu_shift): Likewise.
371e77e3
UB
6553 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6554 change.
859abddd
SN
6555 (cortex_a7_alu_reg): Likewise.
6556 (cortex_a7_alu_shift): Likewise.
6557 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6558 (cortex_a8_alu_shift): Likewise.
6559 (cortex_a8_alu_shift_reg): Likewise.
6560 (cortex_a8_mov): Likewise.
6561 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6562 (cortex_a9_dp_shift): Likewise.
6563 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6564 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6565 (cortex_r4_mov): Likewise.
6566 (cortex_r4_alu_shift): Likewise.
6567 (cortex_r4_alu_shift_reg): Likewise.
6568 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6569 (526_alu_shift_op): Likewise.
6570 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6571 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6572 (626te_alu_shift_op): Likewise.
6573 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6574 (726te_alu_op): Likewise.
6575 (726te_alu_shift_op): Likewise.
6576 (726te_alu_shift_reg_op): Likewise.
6577 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6578 (mp626_alu_shift_op): Likewise.
6579 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6580 (pj4_alu_e1_conds): Likewise.
6581 (pj4_alu): Likewise.
6582 (pj4_alu_conds): Likewise.
6583 (pj4_shift): Likewise.
6584 (pj4_shift_conds): Likewise.
6585 (pj4_alu_shift): Likewise.
6586 (pj4_alu_shift_conds): Likewise.
6587
95b97fac
KT
65882013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6589
6590 * config/arm/predicates.md (shiftable_operator_strict_it):
6591 New predicate.
6592 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
6593 Disable cond_exec version for arm_restrict_it.
6594 (thumb2_smaxsi3): Convert to generate cond_exec.
6595 (thumb2_sminsi3): Likewise.
6596 (thumb32_umaxsi3): Likewise.
6597 (thumb2_uminsi3): Likewise.
6598 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
6599 (thumb2_neg_abssi2): Likewise.
6600 (thumb2_mov_scc): Add alternative for 16-bit encoding.
6601 (thumb2_movsicc_insn): Adjust alternatives.
6602 (thumb2_mov_negscc): Disable for arm_restrict_it.
6603 (thumb2_mov_negscc_strict_it): New pattern.
6604 (thumb2_mov_notscc_strict_it): New pattern.
6605 (thumb2_mov_notscc): Disable for arm_restrict_it.
6606 (thumb2_ior_scc): Likewise.
6607 (thumb2_ior_scc_strict_it): New pattern.
6608 (thumb2_cond_move): Adjust for arm_restrict_it.
6609 (thumb2_cond_arith): Disable for arm_restrict_it.
6610 (thumb2_cond_arith_strict_it): New pattern.
6611 (thumb2_cond_sub): Adjust for arm_restrict_it.
6612 (thumb2_movcond): Likewise.
6613 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
6614 (thumb2_zero_extendhisi2_v6): Likewise.
6615 (thumb2_zero_extendqisi2_v6): Likewise.
6616 (orsi_notsi_si): Likewise.
6617 (orsi_not_shiftsi_si): Likewise.
6618
d5a6ef82
GJL
66192013-07-22 Georg-Johann Lay <avr@gjlay.de>
6620
6621 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
6622 instruction sequence is 1 byte shorter.
6623
d9e71154
UB
66242013-07-22 Uros Bizjak <ubizjak@gmail.com>
6625
6626 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
6627 it is not needed after split.
6628
8c6381ec
IS
66292013-07-20 Iain Sandoe <iain@codesourcery.com>
6630
6631 PR target/51784
6632 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
6633 second label for nonlocal goto receivers. Don't output pic base labels
6634 unless we're producing PIC; mark that action unreachable().
6635 (ix86_save_reg): If the function contains a nonlocal label, save the
6636 PIC base reg.
6637 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 6638 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS
6639 (update_pic_label_number_if_needed): New.
6640 (machopic_output_function_base_name): Adjust for nonlocal receiver
6641 case.
6642 (machopic_should_output_picbase_label): New.
6643 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
6644 (nonlocal_goto_receiver): New insn and split.
6645
d05d0709
JG
66462013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
6647
6648 * config/aarch64/aarch64-builtins.c
6649 (aarch64_fold_builtin): Fold abs in all modes.
6650 * config/aarch64/aarch64-simd-builtins.def
6651 (abs): Enable for all modes.
6652 * config/aarch64/arm_neon.h
6653 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
6654 (vabs_f64): Add missing intrinsic.
6655
ba590e63
IB
66562013-07-19 Ian Bolton <ian.bolton@arm.com>
6657
6658 * config/aarch64/arm_neon.h (vabs_s64): New function
6659
02371798
GJL
66602013-07-19 Georg-Johann Lay <avr@gjlay.de>
6661
6662 PR target/57516
6663 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
6664 * config/avr/avr.md (adjust_len): Add `round'.
6665 * config/avr/avr-protos.h (avr_out_round): New prototype.
6666 (avr_out_plus): Add `out_label' argument.
6667 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
6668 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
6669 Handle the case where `insn' is just a pattern.
6670 (avr_out_bitop): Handle the case where `insn' is just a pattern.
6671 (avr_out_round): New function.
6672 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
6673
12211b99 66742013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH
6675
6676 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6677 Add check for flag_stack_usage to handle -fstack-usage support
6678
12211b99 66792013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH
6680
6681 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
6682 interaction for new Power8 flags and VSX.
6683
df9dda2d
ST
66842013-07-18 Sriraman Tallam <tmsriram@google.com>
6685
6686 PR middle-end/57698
6687 * tree-inline.c (expand_call_inline): Emit errors during
6688 early_inlining only if optimization is not turned on.
6689
0f9cdef4
DM
66902013-07-18 David Malcolm <dmalcolm@redhat.com>
6691
6692 * passes.def: New.
6693
6694 * passes.c (init_optimization_passes): Move the construction of
6695 the pass hierarchy into a new passes.def file.
6696
6697 * Makefile.in (passes.o): Add dependency on passes.def.
6698
2efa4087
DM
66992013-07-18 David Malcolm <dmalcolm@redhat.com>
6700
6701 * passes.c (init_optimization_passes): Introduce macros for
6702 constructing the tree of passes (INSERT_PASSES_AFTER,
6703 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
6704 TERMINATE_PASS_LIST).
6705
bc404e1b
VM
67062013-07-18 Vladimir Makarov <vmakarov@redhat.com>
6707 Wei Mi <wmi@google.com>
6708
6709 PR rtl-optimization/57878
6710 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
6711 top.
6712 (reload_pseudo_compare_func): Check nregs first for reload
6713 pseudos.
6714
d675843f
DM
67152013-07-18 David Malcolm <dmalcolm@redhat.com>
6716
6717 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
6718
c7b3b99f 67192013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC
6720
6721 * read-rtl.c (validate_const_int): Once an invalid character is
6722 seen, quit the loop.
6723
6724 * gengtype.c (write_roots): Similarly once we find the "deletable"
6725 or "if_marked" option.
6726
52fceb44
SN
67272013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6728
6729 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
6730 "xtab" and "sat". Move value "clz" from here to ...
6731 (attriubte "type"): ... here.
6732 (satsi_<SAT:code>): Delete "insn" attribute.
6733 (satsi_<SAT:code>_shift): Likewise.
6734 (arm_zero_extendqisi2addsi): Likewise.
6735 (arm_extendqisi2addsi): Likewise.
6736 (clzsi2): Update for attribute changes.
6737 (rbitsi2): Likewise.
371e77e3
UB
6738 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
6739 attribute.
52fceb44
SN
6740 (arm_usatsihi): Likewise.
6741 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6742
006bd006
SN
67432013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6744
6745 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB
6746 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
6747 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN
6748 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
6749 in alphabetical order.
6750 (attribute "core_cycles"): Update for attribute changes.
6751 (arm_addsi3): Likewise.
6752 (addsi3_compare0): Likewise.
6753 (addsi3_compare0_scratch): Likewise.
6754 (addsi3_compare_op1): Likewise.
6755 (addsi3_compare_op2): Likewise.
6756 (compare_addsi2_op0): Likewise.
6757 (compare_addsi2_op1): Likewise.
6758 (addsi3_carryin_shift_<optab>): Likewise.
6759 (subsi3_carryin_shift): Likewise.
6760 (rsbsi3_carryin_shift): Likewise.
6761 (arm_subsi3_insn): Likewise.
6762 (subsi3_compare0): Likewise.
6763 (subsi3_compare): Likewise.
6764 (arm_andsi3_insn): Likewise.
6765 (thumb1_andsi3_insn): Likewise.
6766 (andsi3_compare0): Likewise.
6767 (andsi3_compare0_scratch): Likewise.
6768 (zeroextractsi_compare0_scratch
6769 (andsi_not_shiftsi_si): Likewise.
6770 (iorsi3_insn): Likewise.
6771 (iorsi3_compare0): Likewise.
6772 (iorsi3_compare0_scratch): Likewise.
6773 (arm_xorsi3): Likewise.
6774 (thumb1_xorsi3_insn): Likewise.
6775 (xorsi3_compare0): Likewise.
6776 (xorsi3_compare0_scratch): Likewise.
6777 (satsi_<SAT:code>_shift): Likewise.
6778 (rrx): Likewise.
6779 (arm_shiftsi3): Likewise.
6780 (shiftsi3_compare0): Likewise.
6781 (not_shiftsi): Likewise.
6782 (not_shiftsi_compare0): Likewise.
6783 (not_shiftsi_compare0_scratch): Likewise.
6784 (arm_one_cmplsi2): Likewise.
6785 (thumb_one_complsi2): Likewise.
6786 (notsi_compare0): Likewise.
6787 (notsi_compare0_scratch): Likewise.
6788 (thumb1_zero_extendhisi2): Likewise.
6789 (arm_zero_extendhisi2): Likewise.
6790 (arm_zero_extendhisi2_v6): Likewise.
6791 (arm_zero_extendhisi2addsi): Likewise.
6792 (thumb1_zero_extendqisi2): Likewise.
6793 (thumb1_zero_extendqisi2_v6): Likewise.
6794 (arm_zero_extendqisi2): Likewise.
6795 (arm_zero_extendqisi2_v6): Likewise.
6796 (arm_zero_extendqisi2addsi): Likewise.
6797 (thumb1_extendhisi2): Likewise.
6798 (arm_extendhisi2): Likewise.
6799 (arm_extendhisi2_v6): Likewise.
6800 (arm_extendqisi): Likewise.
6801 (arm_extendqisi_v6): Likewise.
6802 (arm_extendqisi2addsi): Likewise.
6803 (thumb1_extendqisi2): Likewise.
6804 (thumb1_movdi_insn): Likewise.
6805 (arm_movsi_insn): Likewise.
6806 (movsi_compare0): Likewise.
6807 (movhi_insn_arch4): Likewise.
6808 (movhi_bytes): Likewise.
6809 (arm_movqi_insn): Likewise.
6810 (thumb1_movqi_insn): Likewise.
6811 (arm32_movhf): Likewise.
6812 (thumb1_movhf): Likewise.
6813 (arm_movsf_soft_insn): Likewise.
6814 (thumb1_movsf_insn): Likewise.
6815 (movdf_soft_insn): Likewise.
6816 (thumb_movdf_insn): Likewise.
6817 (arm_cmpsi_insn): Likewise.
6818 (cmpsi_shiftsi): Likewise.
6819 (cmpsi_shiftsi_swp): Likewise.
6820 (arm_cmpsi_negshiftsi_si): Likewise.
6821 (movsicc_insn): Likewise.
6822 (movsfcc_soft_insn): Likewise.
6823 (arith_shiftsi): Likewise.
6824 (arith_shiftsi_compare0
6825 (arith_shiftsi_compare0_scratch
6826 (sub_shiftsi): Likewise.
6827 (sub_shiftsi_compare0
6828 (sub_shiftsi_compare0_scratch
6829 (and_scc): Likewise.
6830 (cond_move): Likewise.
6831 (if_plus_move): Likewise.
6832 (if_move_plus): Likewise.
6833 (if_move_not): Likewise.
6834 (if_not_move): Likewise.
6835 (if_shift_move): Likewise.
6836 (if_move_shift): Likewise.
6837 (if_shift_shift): Likewise.
6838 (if_not_arith): Likewise.
6839 (if_arith_not): Likewise.
6840 (cond_move_not): Likewise.
6841 (thumb1_ashlsi3): Set type attribute.
6842 (thumb1_ashrsi3): Likewise.
6843 (thumb1_lshrsi3): Likewise.
6844 (thumb1_rotrsi3): Likewise.
6845 (shiftsi3_compare0_scratch): Likewise.
6846 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
6847 (neon_mov<mode>): Likewise.
371e77e3
UB
6848 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
6849 attribute changes.
006bd006
SN
6850 (thumb2_movsi_insn): Likewise.
6851 (thumb2_cmpsi_neg_shiftsi): Likewise.
6852 (thumb2_extendqisi_v6): Likewise.
6853 (thumb2_zero_extendhisi2_v6): Likewise.
6854 (thumb2_zero_extendqisi2_v6): Likewise.
6855 (thumb2_shiftsi3_short): Likewise.
6856 (thumb2_addsi3_compare0_scratch): Likewise.
6857 (orsi_not_shiftsi_si): Likewise.
6858 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6859 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
6860 changes.
6861 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6862 (1020alu_shift_op): Likewise.
6863 (1020alu_shift_reg_op): Likewise.
6864 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6865 (alu_shift_op): Likewise.
6866 (alu_shift_reg_op): Likewise.
6867 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6868 (11_alu_shift_op): Likewise.
6869 (11_alu_shift_reg_op): Likewise.
6870 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6871 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
6872 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6873 changes.
006bd006
SN
6874 (cortex_a15_alu_shift): Likewise.
6875 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB
6876 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
6877 changes.
006bd006
SN
6878 (cortex_a5_alu_shift): Likewise.
6879 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
6880 changes.
6881 (cortex_a53_alu_shift): Likewise.
6882 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6883 changes.
6884 (cortex_a7_alu_reg): Likewise.
6885 (cortex_a7_alu_shift): Likewise.
371e77e3
UB
6886 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
6887 changes.
006bd006
SN
6888 (cortex_a8_alu_shift): Likewise.
6889 (cortex_a8_alu_shift_reg): Likewise.
6890 (cortex_a8_mov): Likewise.
6891 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6892 (cortex_a9_dp_shift): Likewise.
371e77e3
UB
6893 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
6894 changes.
6895 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
6896 changes.
006bd006
SN
6897 (cortex_r4_mov): Likewise.
6898 (cortex_r4_alu_shift): Likewise.
6899 (cortex_r4_alu_shift_reg): Likewise.
6900 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6901 (526_alu_shift_op): Likewise.
6902 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6903 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6904 (626te_alu_shift_op): Likewise.
6905 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
6906 (726te_alu_op): Likewise.
6907 (726te_alu_shift_op): Likewise.
6908 (726te_alu_shift_reg_op): Likewise.
6909 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6910 (mp626_alu_shift_op): Likewise.
6911 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
6912 (pj4_alu_e1_conds): Likewise.
6913 (pj4_alu): Likewise.
6914 (pj4_alu_conds): Likewise.
6915 (pj4_shift): Likewise.
6916 (pj4_shift_conds): Likewise.
6917 (pj4_alu_shift): Likewise.
6918 (pj4_alu_shift_conds): Likewise.
371e77e3
UB
6919 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6920 changes.
006bd006
SN
6921 (cortexa7_older_only): Likewise.
6922 (cortexa7_younger): Likewise.
6923
651df1b2
DM
69242013-07-18 David Malcolm <dmalcolm@redhat.com>
6925
6926 * ipa-pure-const.c (generate_summary): Rename to...
6927 (pure_const_generate_summary): ... this.
6928
ea717bd7
IS
69292013-07-17 Iain Sandoe <iain@codesourcery.com>
6930
6931 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
6932
a1259a13
YR
69332013-07-17 Yvan Roux <yvan.roux@linaro.org>
6934
6935 PR target/57909
6936 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
6937 usage in HI mode.
6938
167f68ed
AK
69392013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6940
6941 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
6942 enabled without -march=zEC12.
6943 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
6944 flags to be set.
6945
f900a982
MR
69462013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6947
6948 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
6949 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
6950 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
6951 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
6952 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
6953 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
6954 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
6955 ISA_HAS_FP4.
6956 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
6957 and ISA_HAS_NMADD3_NMSUB3.
6958 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
6959 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
6960 (nmsub4<mode>, nmsub3<mode>): Likewise.
6961 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
6962
d4ed27eb
MR
69632013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6964
6965 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
6966 TARGET_MIPS5400 checking.
6967
7a5add18
PB
69682013-07-16 Jakub Jelinek <jakub@redhat.com>
6969 Peter Bergner <bergner@vnet.ibm.com>
6970
6971 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
6972 registers in the comment.
6973 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
6974 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
6975 rather than FIRST_PSEUDO_REGISTERS.
6976
69772013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB
6978
6979 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
6980 enable extra ISA flags with TARGET_HTM.
6981
d2ab0929
MR
69822013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6983
6984 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
6985 Fix comment typos.
6986
839c74bc
CH
69872013-07-15 Cong Hou <congh@google.com>
6988
6989 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
6990 in compare function for sorting.
6991
7a5add18 69922013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB
6993
6994 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
6995 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
6996 * config/rs6000/rs6000.opt: Add -mhtm option.
6997 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
6998 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
6999 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7000 __HTM__ if the HTM instructions are available.
7001 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
7002 htm_spr_reg_operand): New define_predicates.
7003 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
7004 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
7005 Include htm.md.
7006 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
7007 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
7008 HTM builtin functions.
7009 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
7010 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
7011 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
7012 (rs6000_builtin_mask_calculate): Likewise.
7013 (rs6000_option_override_internal): Likewise.
7014 (bdesc_htm): Add new HTM builtin support.
7015 (htm_spr_num): New function.
7016 (htm_spr_regno): Likewise.
7017 (rs6000_htm_spr_icode): Likewise.
7018 (htm_expand_builtin): Likewise.
7019 (htm_init_builtins): Likewise.
7020 (rs6000_expand_builtin): Add support for HTM builtin functions.
7021 (rs6000_init_builtins): Likewise.
371e77e3
UB
7022 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
7023 option.
0258b6e4
PB
7024 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
7025 (TARGET_HTM, MASK_HTM): Define macros.
7026 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
7027 (FIXED_REGISTERS): Likewise.
7028 (CALL_USED_REGISTERS): Likewise.
7029 (CALL_REALLY_USED_REGISTERS): Likewise.
7030 (REG_ALLOC_ORDER): Likewise.
7031 (enum reg_class): Likewise.
7032 (REG_CLASS_NAMES): Likewise.
7033 (REG_CLASS_CONTENTS): Likewise.
7034 (REGISTER_NAMES): Likewise.
7035 (ADDITIONAL_REGISTER_NAMES): Likewise.
7036 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
7037 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
7038 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
7039 * config/rs6000/htm.md: New file.
7040 * config/rs6000/htmintrin.h: New file.
7041 * config/rs6000/htmxlintrin.h: New file.
7042
87dd8ab0
MS
70432013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
7044
7045 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7046 Define SYMBOL_TINY_GOT, update comment.
7047 * config/aarch64/aarch64.c
7048 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
7049 (aarch64_expand_mov_immediate): Likewise.
7050 (aarch64_print_operand): Likewise.
7051 (aarch64_classify_symbol): Likewise.
7052 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
7053 (ldr_got_tiny): Define.
7054
52d676b6
TG
70552013-07-13 Tobias Grosser <tobias@grosser.es>
7056
7057 PR tree-optimization/54094
7058 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
7059 scheduling dimension for the parallelism check from the polyhedral
7060 information in the AST.
7061 * graphite-dependences.c (carries_deps): Do not assume the schedule is
7062 in 2D + 1 form.
7063
286e8fc1
JM
70642013-07-13 Jason Merrill <jason@redhat.com>
7065
7066 * print-tree.c (debug_vec_tree): Use debug_raw.
7067 (debug_raw (vec<tree, va_gc> &)): New.
7068 (debug_raw (vec<tree, va_gc> *)): New.
7069 * tree.h: Declare them.
7070
e43257e8
BC
70712013-07-13 Bin Cheng <bin.cheng@arm.com>
7072
7073 * ifcvt.c (ifcvt_after_combine): New static variable.
7074 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
7075 for size.
7076 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
7077 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
7078 rest_of_handle_if_after_reload): Pass new argument for if_convert.
7079
da582d46
MR
70802013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
7081
7082 * config/mips/mips.c (mips_expand_call): Remove empty statement.
7083
18b0ea8f
MM
70842013-07-12 Michael Matz <matz@suse.de>
7085
7086 PR middle-end/55771
7087 * convert.c (convert_to_real): Reject non-float inner types.
7088
e4f0f84d
TB
70892013-07-12 Tejas Belagod <tejas.belagod@arm.com>
7090
7091 * config/aarch64/aarch64-protos.h
7092 (aarch64_simd_immediate_valid_for_move): Remove.
7093 * config/aarch64/aarch64.c (simd_immediate_info): New member.
7094 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
7095 cases.
7096 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
7097
76500022
SE
70982013-07-11 Steve Ellcey <sellcey@mips.com>
7099
7100 * config/mips/mips.c (mips_conditional_register_usage): Do not
7101 use t[0-7] registers in MIPS16 mode when optimizing for size.
7102
78039734
ST
71032013-07-11 Sriraman Tallam <tmsriram@google.com>
7104
7105 * config/i386/i386.c (dispatch_function_versions): Fix array
7106 indexing of function_version_info to match actual_versions.
7107
7b24b675
TJ
71082013-07-11 Teresa Johnson <tejohnson@google.com>
7109
7110 * vec.h (struct va_gc): Move release out-of-line.
7111 (va_gc::release): Call ggc_free on released vec.
7112
85d4cbb8
UW
71132013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7114
7115 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
7116 Require GOT register as additional operand in UNSPEC.
7117 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
7118 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7119 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7120 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
7121 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
7122 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
7123 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
7124
52b3f9e2
GJL
71252013-07-11 Georg-Johann Lay <avr@gjlay.de>
7126
7127 PR target/57631
7128 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
7129 name seen by assembler/linker rather if available.
7130
3f3e266e
AS
71312013-07-11 Andreas Schwab <schwab@suse.de>
7132
7133 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
7134
21231ca6
VM
71352013-07-10 Vladimir Makarov <vmakarov@redhat.com>
7136
371e77e3 7137 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 7138
d331d374
JM
71392013-07-10 Joseph Myers <joseph@codesourcery.com>
7140
673c2f63
JM
7141 * doc/tm.texi.in: Move hook documentation to ....
7142 * target.def: ... here.
7143
d331d374
JM
7144 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
7145 text on @hook line.
7146 * doc/tm.texi: Regenerate.
7147
dfeadaa0
PC
71482013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
7149
7150 PR c++/57869
7151 * doc/invoke.texi: Document -Wconditionally-supported.
7152
bb6f2bac
GJL
71532013-07-10 Georg-Johann Lay <avr@gjlay.de>
7154
7155 PR target/57844
7156 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
7157 of my_fp.
7158
5219b232
GJL
71592013-07-10 Georg-Johann Lay <avr@gjlay.de>
7160
7161 PR target/57506
7162 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
7163 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
7164 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
7165 Remove duplicate devices.
7166 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
7167 * config/avr/t-multilib: Regenerate.
7168 * config/avr/avr-tables.opt: Regenerate.
7169 * doc/avr-mmcu.texi: Regenerate.
7170
283045c0
GJL
71712013-07-10 Georg-Johann Lay <avr@gjlay.de>
7172
7173 PR target/56987
7174 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
7175
ceb2d59f 71762013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 7177
ceb2d59f
GS
7178 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
7179 the cost of MULT when optimizing for size.
7180
055e0a99
JBG
71812013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7182
7183 * config/cr16/cr16-protos.h: Don't include target.h.
7184
ef43e868
JM
71852013-07-09 Joseph Myers <joseph@codesourcery.com>
7186
7187 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
7188 adjust register size for TDmode and TFmode for VSX registers.
7189
4a283fdf
KT
71902013-07-08 Kai Tietz <ktietz@redhat.com>
7191
7192 PR target/56892
7193 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
7194 hook_bool_const_tree_true.
7195
2cf4c39e
AK
71962013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7197
7198 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
7199 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
7200 * config/s390/s390.md: Define FPR*_REGNUM constants.
7201 Fix FPR2_REGNUM constant (18 -> 17).
7202 ("*trunc<BFP:mode><DFP_ALL:mode>2")
7203 ("*trunc<DFP_ALL:mode><BFP:mode>2")
7204 ("trunc<BFP:mode><DFP_ALL:mode>2")
7205 ("trunc<DFP_ALL:mode><BFP:mode>2")
7206 ("*extend<BFP:mode><DFP_ALL:mode>2")
7207 ("*extend<DFP_ALL:mode><BFP:mode>2")
7208 ("extend<BFP:mode><DFP_ALL:mode>2")
7209 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
7210 FPR4_REGNUM.
7211
e6ac0270 72122013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 7213
6a2fa4b2 7214 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 7215
b89b22fc
AK
72162013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7217
7218 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
7219 and cfun_fpr_bit_p to cfun_fpr_save_p.
7220 (s390_frame_area, s390_register_info, s390_frame_info)
7221 (s390_emit_prologue, s390_emit_epilogue)
7222 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
7223 register numbers.
7224 * config/s390/s390.h: Define *_REGNUM macros for floating point
7225 register numbers.
7226
07f398aa
EB
72272013-07-08 Eric Botcazou <ebotcazou@adacore.com>
7228
7229 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
7230
75fefa91
PCC
72312013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
7232
7233 PR rtl-optimization/57786
7234 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
7235 and break out of the loop when it is set to false.
7236
046f1eee
JJ
72372013-07-08 Jakub Jelinek <jakub@redhat.com>
7238
8140c065
JJ
7239 PR target/57819
7240 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7241 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7242 (const_int 63)) 0)).
7243 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7244 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7245 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7246
046f1eee
JJ
7247 PR rtl-optimization/57829
7248 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7249 mask bits outside of mode are just sign-extension from mode to HWI.
7250
930b700b
MZ
72512013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7252
7253 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7254 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7255 adjust_address instead of change_address to keep info about alignment.
7256 (emit_strmov): Remove.
7257 (emit_memmov): New function.
7258 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7259 (expand_movmem_epilogue): Likewise and return updated rtx for
7260 destination.
7261 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7262 destination and source.
7263 (decide_alignment): Refactor, handle vector_loop.
7264 (ix86_expand_movmem): Likewise.
7265 (ix86_expand_setmem): Likewise.
7266 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7267
7b9d1bd8
UB
72682013-07-07 Uros Bizjak <ubizjak@gmail.com>
7269
7270 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7271 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7272
eabb5f48
UB
72732013-07-06 Uros Bizjak <ubizjak@gmail.com>
7274
7275 * config/i386/sse.md (sse_movlhps): Change alternative 3
7276 of operand 2 to "m".
7277
72782013-07-06 Uros Bizjak <ubizjak@gmail.com>
7279
7280 PR target/57807
7281 * config/i386/sse.md (iptr): New mode attribute.
7282 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7283 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7284 (<sse>_vmmul<mode>3): Ditto.
7285 (<sse>_vmdiv<mode>3): Ditto.
7286 (sse_vmrcpv4sf2): Ditto.
7287 (<sse>_vmsqrt<mode>2): Ditto.
7288 (sse_vmrsqrtv4sf2): Ditto.
7289 (<sse>_vm<code><mode>3): Ditto.
7290 (avx_vmcmp<mode>3): Ditto.
7291 (<sse>_vmmaskcmp<mode>3): Ditto.
7292 (<sse>_comi): Ditto.
7293 (<sse>_ucomi): Ditto.
7294 (*xop_vmfrcz_<mode>): Ditto.
7295 (*fmai_fmadd_<mode>): Ditto.
7296 (*fmai_fmsub_<mode>): Ditto.
7297 (*fmai_fnmadd_<mode>): Ditto.
7298 (*fmai_fnmsub_<mode>): Ditto.
7299 (*fma4i_vmfmadd_<mode>): Ditto.
7300 (*fma4i_vmfmsub_<mode>): Ditto.
7301 (*fma4i_vmfnmadd_<mode>): Ditto.
7302 (*fma4i_vmfnmsub_<mode>): Ditto.
7303 (*xop_vmfrcz_<mode>): Ditto.
7304 (sse_cvtps2pi): Ditto.
7305 (sse_cvttps2pi): Ditto.
7306 (sse_cvtss2si): Ditto.
7307 (sse_cvtss2si_2): Ditto.
7308 (sse_cvtss2siq_2): Ditto.
7309 (sse_cvttss2si): Ditto.
7310 (sse_cvttss2siq): Ditto.
7311 (sse_cvtsd2si): Ditto.
7312 (sse_cvtsd2si_2): Ditto.
7313 (sse_cvtsd2siq_2): Ditto.
7314 (sse_cvttsd2si): Ditto.
7315 (sse_cvttsd2siq): Ditto.
7316 (sse_cvtsd2ss): Ditto.
7317 (sse_cvtss2sd): Ditto.
7318 (avx2_pbroadcast<mode>): Ditto.
7319 (avx2_pbroadcast<mode>_1): Ditto.
7320 (*avx_vperm_broadcast_v4sf): Ditto.
7321
7322 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7323 (sse_movlhps): Ditto.
7324 (sse_storehps): Ditto.
7325 (sse_loadhps): Ditto.
7326 (sse_storelps): Ditto.
7327 (sse_loadlps): Ditto.
7328 (*vec_concatv4sf): Ditto.
7329 (*vec_interleave_highv2df): Ditto.
7330 (*vec_interleave_lowv2df): Ditto.
7331 (*vec_extractv2df_1_sse): Ditto.
7332 (*vec_extractv2df_0_sse): Ditto.
7333 (sse2_storelpd): Ditto.
7334 (sse2_loadlpd): Ditto.
7335 (sse2_movsd): Ditto.
7336 (*vec_concatv4si): Ditto.
7337 (vec_concatv2di): Ditto.
7338
7339 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
7340 for Intel asm dialect.
7341 (mmx_punpcklwd): Ditto.
7342 (mmx_punpckldq): Ditto.
7343
7344 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
7345 for intel assembler dialect.
7346
1f6eac90
JJ
73472013-07-06 Jakub Jelinek <jakub@redhat.com>
7348
7349 PR target/29776
7350 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
7351 for BUILT_IN_C{LZ,LRSB}*.
7352 * tree.h (CASE_INT_FN): Add FN##IMAX case.
7353 * tree-vrp.c (extract_range_basic): Handle
7354 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
7355 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
7356 fall thru to code calling set_value*.
7357 * builtins.c (expand_builtin): Remove *IMAX cases.
7358 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
7359 if width is bigger than 2*HWI.
7360
2b778c9d
VM
73612013-07-05 Vladimir Makarov <vmakarov@redhat.com>
7362
7363 PR rtl-optimization/55342
7364 * lra-int.h (lra_subreg_reload_pseudos): New.
7365 * lra.c: Add undoing optional reloads to the block diagram.
7366 (lra_subreg_reload_pseudos): New.
7367 (lra_optional_reload_pseudos): Change comments.
7368 (lra): Init and clear lra_subreg_reload_pseudos. Clear
7369 lra_optional_reload_pseudos after undo transformations.
7370 * lra-assigns.c (pseudo_prefix_title): New.
7371 (lra_setup_reg_renumber): Use it.
7372 (spill_for): Ditto. Check subreg reload pseudos too.
7373 (assign_by_spills): Consider subreg reload pseudos too.
7374 * lra-constraints.c (simplify_operand_subreg): Use
7375 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
7376 (curr_insn_transform): Recognize and do optional reloads.
7377 (undo_optional_reloads): New.
7378 (lra_undo_inheritance): Call undo_optional_reloads.
7379
207156e4
TQ
73802013-07-05 Thomas Quinot <quinot@adacore.com>
7381
7382 * tree-complex.c (expand_complex_operations_1): Fix typo.
7383
8990e73a
TB
73842013-07-04 Tejas Belagod <tejas.belagod@arm.com>
7385
7386 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
7387 (tune_params): New member 'const vec_costs'.
7388 * config/aarch64/aarch64.c (generic_vector_cost): New.
7389 (generic_tunings): New member 'generic_vector_cost'.
7390 (aarch64_builtin_vectorization_cost): New.
7391 (aarch64_add_stmt_cost): New.
7392 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7393 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7394
f7bc421d
JJ
73952013-07-03 Jakub Jelinek <jakub@redhat.com>
7396
7397 PR target/57777
7398 * config/i386/predicates.md (vsib_address_operand): Disallow
7399 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7400
e90d1568
HPN
74012013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7402
7403 PR middle-end/55030
7404 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7405 expand_builtin_setjmp_receiver.
7406 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7407 with NULL for the label parameter.
7408 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7409 the frame-pointer. Adjust comments.
7410 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7411 only if LABEL is non-NULL.
7412
cbd7413b
YZ
74132013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7414
7415 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7416 (ARM_ABI_AAPCS64): Ditto.
7417 (arm_abi): Ditto.
7418 (ARM_DEFAULT_ABI): Ditto.
7419
dec11868
JG
74202013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7421
7422 * config/aarch64/aarch64-builtins.c
7423 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7424 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7425 (st1): Likewise.
7426 * config/aarch64/aarch64-simd.md
7427 (aarch64_ld1<VALL:mode>): New.
7428 (aarch64_st1<VALL:mode>): Likewise.
7429 * config/aarch64/arm_neon.h
7430 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7431
786fb9b6
ST
74322013-07-02 Sriraman Tallam <tmsriram@google.com>
7433
0ba6aff3
UB
7434 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7435 target ISA is AVX.
786fb9b6
ST
7436 (ix86_option_override_internal):Turn on all -mavx target flags by
7437 default as they are dependent on AVX anyway.
7438
4b043553
CC
74392013-07-02 Cary Coutant <ccoutant@google.com>
7440
7441 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7442 deterministic hash.
7443 (loc_checksum_ordered): Likewise.
7444 (hash_loc_operands): Remove inline keyword.
7445
c4e87a13
JJ
74462013-07-02 Jakub Jelinek <jakub@redhat.com>
7447
7448 PR tree-optimization/57741
7449 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7450 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7451 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7452 Allow REAL_CST step_exprs if flag_associative_math.
7453 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7454
26366d28
IB
74552013-07-02 Ian Bolton <ian.bolton@arm.com>
7456
0ba6aff3 7457 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 7458
2879bb2b
IB
74592013-07-02 Ian Bolton <ian.bolton@arm.com>
7460
7461 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7462
afc5e8a6
KT
74632013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7464
7465 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7466 encoding.
7467 (iorsi3_insn): Likewise.
7468 (arm_xorsi3): Likewise.
7469
31f8442b
SN
74702013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7471
7472 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7473 to ...
7474 (attribute "type"): ... here, and prefix with "wmmx_".
7475 (attribute "core_cycles"): Update for attribute changes.
7476 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7477 (tbcstv4hi): Likewise.
7478 (tbcstv2si): Likewise.
7479 (iwmmxt_iordi3): Likewise.
7480 (iwmmxt_xordi3): Likewise.
7481 (iwmmxt_anddi3): Likewise.
7482 (iwmmxt_nanddi3): Likewise.
7483 (iwmmxt_arm_movdi): Likewise.
7484 (iwmmxt_movsi_insn): Likewise.
7485 (mov<mode>_internal): Likewise.
7486 (and<mode>3_iwmmxt): Likewise.
7487 (ior<mode>3_iwmmxt): Likewise.
7488 (xor<mode>3_iwmmxt): Likewise.
7489 (add<mode>3_iwmmxt): Likewise.
7490 (ssaddv8qi3): Likewise.
7491 (ssaddv4hi3): Likewise.
7492 (ssaddv2si3): Likewise.
7493 (usaddv8qi3): Likewise.
7494 (usaddv4hi3): Likewise.
7495 (usaddv2si3): Likewise.
7496 (sub<mode>3_iwmmxt): Likewise.
7497 (sssubv8qi3): Likewise.
7498 (sssubv4hi3): Likewise.
7499 (sssubv2si3): Likewise.
7500 (ussubv8qi3): Likewise.
7501 (ussubv4hi3): Likewise.
7502 (ussubv2si3): Likewise.
7503 (mulv4hi3_iwmmxt): Likewise.
7504 (smulv4hi3_highpart): Likewise.
7505 (umulv4hi3_highpart): Likewise.
7506 (iwmmxt_wmacs): Likewise.
7507 (iwmmxt_wmacsz): Likewise.
7508 (iwmmxt_wmacu): Likewise.
7509 (iwmmxt_wmacuz): Likewise.
7510 (iwmmxt_clrdi): Likewise.
7511 (iwmmxt_clrv8qi): Likewise.
7512 (iwmmxt_clr4hi): Likewise.
7513 (iwmmxt_clr2si): Likewise.
7514 (iwmmxt_uavgrndv8qi3): Likewise.
7515 (iwmmxt_uavgrndv4hi3): Likewise.
7516 (iwmmxt_uavgv8qi3): Likewise.
7517 (iwmmxt_uavgv4hi3): Likewise.
7518 (iwmmxt_tinsrb): Likewise.
7519 (iwmmxt_tinsrh): Likewise.
7520 (iwmmxt_tinsrw): Likewise.
7521 (iwmmxt_textrmub): Likewise.
7522 (iwmmxt_textrmsb): Likewise.
7523 (iwmmxt_textrmuh): Likewise.
7524 (iwmmxt_textrmsh): Likewise.
7525 (iwmmxt_textrmw): Likewise.
7526 (iwmxxt_wshufh): Likewise.
7527 (eqv8qi3): Likewise.
7528 (eqv4hi3): Likewise.
7529 (eqv2si3): Likewise.
7530 (gtuv8qi3): Likewise.
7531 (gtuv4hi3): Likewise.
7532 (gtuv2si3): Likewise.
7533 (gtv8qi3): Likewise.
7534 (gtv4hi3): Likewise.
7535 (gtv2si3): Likewise.
7536 (smax<mode>3_iwmmxt): Likewise.
7537 (umax<mode>3_iwmmxt): Likewise.
7538 (smin<mode>3_iwmmxt): Likewise.
7539 (umin<mode>3_iwmmxt): Likewise.
7540 (iwmmxt_wpackhss): Likewise.
7541 (iwmmxt_wpackwss): Likewise.
7542 (iwmmxt_wpackdss): Likewise.
7543 (iwmmxt_wpackhus): Likewise.
7544 (iwmmxt_wpackwus): Likewise.
7545 (iwmmxt_wpackdus): Likewise.
7546 (iwmmxt_wunpckihb): Likewise.
7547 (iwmmxt_wunpckihh): Likewise.
7548 (iwmmxt_wunpckihw): Likewise.
7549 (iwmmxt_wunpckilb): Likewise.
7550 (iwmmxt_wunpckilh): Likewise.
7551 (iwmmxt_wunpckilw): Likewise.
7552 (iwmmxt_wunpckehub): Likewise.
7553 (iwmmxt_wunpckehuh): Likewise.
7554 (iwmmxt_wunpckehuw): Likewise.
7555 (iwmmxt_wunpckehsb): Likewise.
7556 (iwmmxt_wunpckehsh): Likewise.
7557 (iwmmxt_wunpckehsw): Likewise.
7558 (iwmmxt_wunpckelub): Likewise.
7559 (iwmmxt_wunpckeluh): Likewise.
7560 (iwmmxt_wunpckeluw): Likewise.
7561 (iwmmxt_wunpckelsb): Likewise.
7562 (iwmmxt_wunpckelsh): Likewise.
7563 (iwmmxt_wunpckelsw): Likewise.
7564 (ror<mode>3): Likewise.
7565 (ashr<mode>3_iwmmxt): Likewise.
7566 (lshr<mode>3_iwmmxt): Likewise.
7567 (ashl<mode>3_iwmmxt): Likewise.
7568 (ror<mode>3_di): Likewise.
7569 (ashr<mode>3_di): Likewise.
7570 (lshr<mode>3_di): Likewise.
7571 (ashl<mode>3_di): Likewise.
7572 (iwmmxt_wmadds): Likewise.
7573 (iwmmxt_wmaddu): Likewise.
7574 (iwmmxt_tmia): Likewise.
7575 (iwmmxt_tmiaph): Likewise.
7576 (iwmmxt_tmiabb): Likewise.
7577 (iwmmxt_tmiatb): Likewise.
7578 (iwmmxt_tmiabt): Likewise.
7579 (iwmmxt_tmiatt): Likewise.
7580 (iwmmxt_tmovmskb): Likewise.
7581 (iwmmxt_tmovmskh): Likewise.
7582 (iwmmxt_tmovmskw): Likewise.
7583 (iwmmxt_waccb): Likewise.
7584 (iwmmxt_wacch): Likewise.
7585 (iwmmxt_waccw): Likewise.
7586 (iwmmxt_waligni): Likewise.
7587 (iwmmxt_walignr): Likewise.
7588 (iwmmxt_walignr0): Likewise.
7589 (iwmmxt_walignr1): Likewise.
7590 (iwmmxt_walignr2): Likewise.
7591 (iwmmxt_walignr3): Likewise.
7592 (iwmmxt_wsadb): Likewise.
7593 (iwmmxt_wsadh): Likewise.
7594 (iwmmxt_wsadbz): Likewise.
7595 (iwmmxt_wsadhz): Likewise.
7596 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
7597 (iwmmxt_wabsdiffb): Likewise.
7598 (iwmmxt_wabsdiffh): Likewise.
7599 (iwmmxt_wabsdiffw): Likewise.
7600 (iwmmxt_waddsubhx): Likewise
7601 (iwmmxt_wsubaddhx): Likewise.
7602 (addc<mode>3): Likewise.
7603 (iwmmxt_avg4): Likewise.
7604 (iwmmxt_avg4r): Likewise.
7605 (iwmmxt_wmaddsx): Likewise.
7606 (iwmmxt_wmaddux): Likewise.
7607 (iwmmxt_wmaddsn): Likewise.
7608 (iwmmxt_wmaddun): Likewise.
7609 (iwmmxt_wmulwsm): Likewise.
7610 (iwmmxt_wmulwum): Likewise.
7611 (iwmmxt_wmulsmr): Likewise.
7612 (iwmmxt_wmulumr): Likewise.
7613 (iwmmxt_wmulwsmr): Likewise.
7614 (iwmmxt_wmulwumr): Likewise.
7615 (iwmmxt_wmulwl): Likewise.
7616 (iwmmxt_wqmulm): Likewise.
7617 (iwmmxt_wqmulwm): Likewise.
7618 (iwmmxt_wqmulmr): Likewise.
7619 (iwmmxt_wqmulwmr): Likewise.
7620 (iwmmxt_waddbhusm): Likewise.
7621 (iwmmxt_waddbhusl): Likewise.
7622 (iwmmxt_wqmiabb): Likewise.
7623 (iwmmxt_wqmiabt): Likewise.
7624 (iwmmxt_wqmiatb): Likewise.
7625 (iwmmxt_wqmiatt): Likewise.
7626 (iwmmxt_wqmiabbn): Likewise.
7627 (iwmmxt_wqmiabtn): Likewise.
7628 (iwmmxt_wqmiatbn): Likewise.
7629 (iwmmxt_wqmiattn): Likewise.
7630 (iwmmxt_wmiabb): Likewise.
7631 (iwmmxt_wmiabt): Likewise.
7632 (iwmmxt_wmiatb): Likewise.
7633 (iwmmxt_wmiatt): Likewise.
7634 (iwmmxt_wmiabbn): Likewise.
7635 (iwmmxt_wmiabtn): Likewise.
7636 (iwmmxt_wmiatbn): Likewise.
7637 (iwmmxt_wmiattn): Likewise.
7638 (iwmmxt_wmiawbb): Likewise.
7639 (iwmmxt_wmiawbt): Likewise.
7640 (iwmmxt_wmiawtb): Likewise.
7641 (iwmmxt_wmiawtt): Likewise.
7642 (iwmmxt_wmiawbbn): Likewise.
7643 (iwmmxt_wmiawbtn): Likewise.
7644 (iwmmxt_wmiawtbn): Likewise.
7645 (iwmmxt_wmiawttn): Likewise.
7646 (iwmmxt_wmerge): Likewise.
7647 (iwmmxt_tandc<mode>3): Likewise.
7648 (iwmmxt_torc<mode>3): Likewise.
7649 (iwmmxt_torvsc<mode>3): Likewise.
7650 (iwmmxt_textrc<mode>3): Likewise.
7651 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
7652 (wmmxt_pack): Likewise.
7653 (wmmxt_mult_c1): Likewise.
7654 (wmmxt_mult_c2): Likewise.
7655 (wmmxt_alu_c1): Likewise.
7656 (wmmxt_alu_c2): Likewise.
7657 (wmmxt_alu_c3): Likewise.
7658 (wmmxt_transfer_c1): Likewise.
7659 (wmmxt_transfer_c2): Likewise.
7660 (wmmxt_transfer_c3): Likewise.
7661 (marvell_f_iwmmxt_wstr): Likewise.
7662 (marvell_f_iwmmxt_wldr): Likewise.
7663
bdb7bf8a
YZ
76642013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
7665
0ba6aff3 7666 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 7667
e6631d38
VM
76682013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7669
7670 Revert:
7671 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7672 * lra-constraints.c (need_for_split_p): Check call used hard regs
7673 living through calls.
7674
7675 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
7676 call used regs for call insn.
7677
53c1275b
JJ
76782013-06-28 Jakub Jelinek <jakub@redhat.com>
7679
7680 PR target/57736
0ba6aff3
UB
7681 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
7682 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 7683
1141ed3f
BI
76842013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7685
7686 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 7687
e32e4c4a
VM
76882013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7689
7690 * lra-constraints.c (need_for_split_p): Check call used hard regs
7691 living through calls.
7692
f161bfd3
MM
76932013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
7694
7695 PR target/57744
7696 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
7697 to tie with any other modes. Eliminate Altivec vector mode tests,
7698 since these are a subset of ALTIVEC or VSX vector modes. Simplify
7699 code, to return 0 if testing MODE2 for a condition, if we've
7700 already tested MODE1 for the same condition.
7701
7eda14e1
MS
77022013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7703
7704 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
7705 layout.
7706
c822f852
MS
77072013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7708
7709 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7710 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
7711
da4f13a4
MS
77122013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7713
0ba6aff3
UB
7714 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
7715 Define.
da4f13a4
MS
7716 (aarch64_symbolic_constant_p): Remove.
7717 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
7718 static. Fix line length and white space.
7719 (aarch64_symbolic_constant_p): Remove.
7720 * config/aarch64/predicates.md (aarch64_valid_symref):
7721 Use aarch64_classify_symbol_expression.
7722
956a95a5
KT
77232013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7724
7725 * config/arm/constraints.md (Ts): New constraint.
7726 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
7727 16-bit encodings.
7728 (compare_scc): Use "Ts" constraint for operand 0.
7729 (ior_scc_scc): Likewise.
7730 (and_scc_scc): Likewise.
7731 (and_scc_scc_nodom): Likewise.
7732 (ior_scc_scc_cmp): Likewise for operand 7.
7733 (and_scc_scc_cmp): Likewise.
7734 * config/arm/thumb2.md (thumb2_movsi_insn):
7735 Add alternatives for 16-bit encodings.
7736 (thumb2_movhi_insn): Likewise.
7737 (thumb2_movsicc_insn): Likewise.
7738 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
7739 (thumb2_negscc): Use "Ts" constraint.
7740 Move mvn instruction outside cond_exec block.
7741 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
7742 for 16-bit encodings.
7743
31dfce10
KT
77442013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7745
7746 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
7747 encoding.
7748 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
7749 (mulsi3subsi): Likewise.
7750 (mulsidi3adddi): Likewise.
7751 (mulsidi3_v6): Likewise.
7752 (umulsidi3_v6): Likewise.
7753 (umulsidi3adddi_v6): Likewise.
7754 (smulsi3_highpart_v6): Likewise.
7755 (umulsi3_highpart_v6): Likewise.
7756 (mulhisi3tb): Likewise.
7757 (mulhisi3bt): Likewise.
7758 (mulhisi3tt): Likewise.
7759 (maddhisi4): Likewise.
7760 (maddhisi4tb): Likewise.
7761 (maddhisi4tt): Likewise.
7762 (maddhidi4): Likewise.
7763 (maddhidi4tb): Likewise.
7764 (maddhidi4tt): Likewise.
7765 (zeroextractsi_compare0_scratch): Likewise.
7766 (insv_zero): Likewise.
7767 (insv_t2): Likewise.
7768 (anddi_notzesidi_di): Likewise.
7769 (anddi_notsesidi_di): Likewise.
7770 (andsi_notsi_si): Likewise.
7771 (iordi_zesidi_di): Likewise.
7772 (xordi_zesidi_di): Likewise.
7773 (andsi_iorsi3_notsi): Likewise.
7774 (smax_0): Likewise.
7775 (smax_m1): Likewise.
7776 (smin_0): Likewise.
7777 (not_shiftsi): Likewise.
7778 (unaligned_loadsi): Likewise.
7779 (unaligned_loadhis): Likewise.
7780 (unaligned_loadhiu): Likewise.
7781 (unaligned_storesi): Likewise.
7782 (unaligned_storehi): Likewise.
7783 (extv_reg): Likewise.
7784 (extzv_t2): Likewise.
7785 (divsi3): Likewise.
7786 (udivsi3): Likewise.
7787 (arm_zero_extendhisi2addsi): Likewise.
7788 (arm_zero_extendqisi2addsi): Likewise.
7789 (compareqi_eq0): Likewise.
7790 (arm_extendhisi2_v6): Likewise.
7791 (arm_extendqisi2addsi): Likewise.
7792 (arm_movt): Likewise.
7793 (thumb2_ldrd): Likewise.
7794 (thumb2_ldrd_base): Likewise.
7795 (thumb2_ldrd_base_neg): Likewise.
7796 (thumb2_strd): Likewise.
7797 (thumb2_strd_base): Likewise.
7798 (thumb2_strd_base_neg): Likewise.
7799 (arm_negsi2): Add alternative for 16-bit encoding.
7800 (arm_one_cmplsi2): Likewise.
7801
1572e697
KT
78022013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7803
7804 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
7805 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
7806 (movdfcc): Likewise.
7807 * config/arm/vfp.md (*thumb2_movsf_vfp):
7808 Disable predication for arm_restrict_it.
7809 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
7810 (*thumb2_movdfcc_vfp): Likewise.
7811 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
7812 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
7813 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
7814 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
7815 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
7816 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
7817 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
7818 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
7819 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
7820 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
7821 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
7822 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
7823 Disable predication for arm_restrict_it.
7824
ebf8f0ea
KY
78252013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
7826
7827 * config/i386/bmiintrin.h (_bextr_u32): New.
7828 (_bextr_u64): Ditto.
7829
a290fcda
RS
78302013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
7831
7832 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
7833 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
7834 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
7835 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
7836 * config/mips/n32-elf.h: ...this new file.
7837
ea169996
MG
78382013-06-27 Marc Glisse <marc.glisse@inria.fr>
7839
7840 PR target/57224
7841 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
7842 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
7843
12211b99 78442013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM
7845
7846 * config/mips/mips-tables.opt: Regenerate.
7847 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 7848 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 7849 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 7850
2444244e
JJ
78512013-06-27 Jakub Jelinek <jakub@redhat.com>
7852
cd06d2a2
JJ
7853 PR target/57623
7854 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
7855 constraints of operand 1 and 2.
7856
2444244e
JJ
7857 PR target/57623
7858 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
7859 to match RTL canonicalization. Swap predicates and
7860 constraints of operand 1 and 2.
7861
2f259720
VM
78622013-06-27 Vladimir Makarov <vmakarov@redhat.com>
7863
0ba6aff3
UB
7864 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
7865 Process OP_INOUT regs for splitting too.
2f259720 7866
fef4d2b3
JJ
78672013-06-27 Jakub Jelinek <jakub@redhat.com>
7868
7869 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
7870 decl before the loop, initialize to NULL.
7871 (vectorizable_load): Initialize ptr_incr to NULL.
7872
79ee9826
MJ
78732013-06-27 Martin Jambor <mjambor@suse.cz>
7874
7875 PR lto/57208
7876 * ipa-ref.h (ipa_maybe_record_reference): Declare.
7877 * ipa-ref.c (ipa_maybe_record_reference): New function.
7878 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
7879 * ipa-cp.c (create_specialized_node): Record potential references from
7880 aggvals.
7881 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
7882
e18b4a81
YZ
78832013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7884
7885 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
7886 parameter 'mode' of type 'enum machine_mode mode'; change to pass
7887 'mode' to force_reg.
7888 (aarch64_add_offset): Update calls to aarch64_force_temporary.
7889 (aarch64_expand_mov_immediate): Likewise.
7890
9c023bf0
YZ
78912013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7892
7893 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
7894 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
7895
5a3fe9b6
AK
78962013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7897
7898 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
7899 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7900 (struct machine_function): Add tbegin_p.
7901 (s390_canonicalize_comparison): Fold CC mode compares to
7902 conditional jump if possible.
7903 (s390_emit_jump): Return the emitted jump.
7904 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
7905 Handle CCRAWmode compares.
7906 (s390_option_override): Default to -mhtm if available.
7907 (s390_reg_clobbered_rtx): Handle floating point regs as well.
7908 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
7909 FPRs instead of df_regs_ever_live_p.
7910 (s390_optimize_nonescaping_tx): New function.
7911 (s390_init_frame_layout): Extend clobbered_regs array to cover
7912 FPRs as well.
7913 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
7914 (s390_expand_tbegin): New function.
7915 (enum s390_builtin): New enum definition.
7916 (code_for_builtin): New array definition.
7917 (s390_init_builtins): New function.
7918 (s390_expand_builtin): New function.
7919 (TARGET_INIT_BUILTINS): Define.
7920 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 7921 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK
7922 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
7923 (s390_alc_comparison): Likewise.
7924 * config/s390/s390-modes.def: Add CCRAWmode.
7925 * config/s390/s390.h (processor_flags): Add PF_TX.
7926 (TARGET_CPU_HTM): Define macro.
7927 (TARGET_HTM): Define macro.
7928 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
7929 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
7930 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7931 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 7932 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK
7933 (TBEGIN_MASK, TBEGINC_MASK): New constants.
7934 ("*cc_to_int"): Move up.
7935 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
7936 constants other than 0.
7937 ("*ccraw_to_int"): New insn and splitter definition.
7938 ("tbegin", "tbegin_nofloat", "tbegin_retry")
7939 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
7940 ("tx_assist"): New expander.
7941 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
7942 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
7943 * config/s390/s390.opt: Add -mhtm option.
7944 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
7945 * config/s390/htmxlintrin.h: New file.
7946 * config/s390/htmintrin.h: New file.
7947 * config/s390/s390intrin.h: New file.
7948 * doc/extend.texi: Document htm builtins.
7949 * config.gcc: Add the new header files to extra_headers.
7950
9e216629
TS
79512013-06-26 Thomas Schwinge <thomas@codesourcery.com>
7952
7953 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 7954 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 7955
fd8c65e7 79562013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
7957 Pat Haugen <pthaugen@us.ibm.com>
7958 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM
7959
7960 * config/rs6000/power8.md: New.
7961 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
7962 setting for power8 entry.
7963 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
7964 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
7965 test for Power4/Power5 only.
7966 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
7967 support.
7968 (force_new_group): Adjust comment.
7969 * config/rs6000/rs6000.md: Include power8.md.
7970
b24a2ce5
GY
79712013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
7972
7973 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
7974 * config/arm/arm-protos.h (arm_max_conditional_execute): New
7975 declaration.
7976 (tune_params): Update comment.
7977 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
7978 (arm_max_conditional_execute): New function.
7979 (thumb2_final_prescan_insn): Use max_insn_skipped and
7980 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
7981
f6fdeb62
JJ
79822013-06-25 Jakub Jelinek <jakub@redhat.com>
7983
7984 PR tree-optimization/57705
7985 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
7986 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 7987 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ
7988 (get_initial_def_for_induction): Handle SSA_NAME IV step.
7989
c13bc3d9
MJ
79902013-06-25 Martin Jambor <mjambor@suse.cz>
7991
7992 PR middle-end/57670
7993 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
7994 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
7995 calls in the dump.
7996 (ipa_note_param_call): Initialize member_ptr flag.
7997 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
7998 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
7999 (ipa_write_indirect_edge_info): Stream member_ptr flag.
8000 (ipa_read_indirect_edge_info): Likewise.
8001
8a845901
RB
80022013-06-25 Richard Biener <rguenther@suse.de>
8003
8004 PR middle-end/56977
8005 * passes.c (init_optimization_passes): Move pass_fold_builtins
8006 and pass_dce earlier with -Og.
8007
33be0bec
EB
80082013-06-25 Eric Botcazou <ebotcazou@adacore.com>
8009
8010 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
8011 <BIT_FIELD_REF>: Remove trailing TAB.
8012 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
8013 remove blank line.
8014
5fe8e757
MJ
80152013-06-24 Martin Jambor <mjambor@suse.cz>
8016
8017 PR tree-optimization/57358
8018 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
8019 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
8020 (ipa_analyze_params_uses): Generate pessimistic info when true.
8021
44a60244
MJ
80222013-06-24 Martin Jambor <mjambor@suse.cz>
8023
8024 PR tree-optimization/57539
8025 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
8026 global.inlined_to of the new node to it. All callers changed.
8027 * ipa-inline-transform.c (clone_inlined_nodes): New variable
8028 inlining_into, pass it to cgraph_clone_node.
8029 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
8030 ipa_free_edge_args_substructures.
8031 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
8032 rdesc linked list. Do not assert rdesc edges have inlined caller.
8033 Assert we have found an rdesc in the rdesc list.
8034
7c5848b8
RB
80352013-06-24 Richard Biener <rguenther@suse.de>
8036
0ba6aff3 8037 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB
8038 (pointer_set_lookup): Declare.
8039 (class pointer_map): New template class implementing a
8040 generic pointer to T map.
8041 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
8042 pointer_map<T>::contains, pointer_map<T>::insert,
8043 pointer_map<T>::traverse): New functions.
8044 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
8045 (pointer_set_lookup): New function.
8046 (pointer_set_contains): Use pointer_set_lookup.
8047 (pointer_set_insert): Likewise.
8048 (insert_aux): Remove.
8049 (struct pointer_map_t): Embed a pointer_set_t.
8050 (pointer_map_create): Adjust.
8051 (pointer_map_destroy): Likewise.
8052 (pointer_map_contains): Likewise.
8053 (pointer_map_insert): Likewise.
8054 (pointer_map_traverse): Likewise.
8055 * tree-streamer.h (struct streamer_tree_cache_d): Use a
8056 pointer_map<unsigned> instead of a pointer_map_t.
8057 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
8058 (streamer_tree_cache_lookup): Likewise.
8059 (streamer_tree_cache_create): Likewise.
8060 (streamer_tree_cache_delete): Likewise.
8061 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
8062 pointer_map<unsigned> instead of a pointer_map_t.
8063 (lto_init_tree_ref_encoder): Adjust.
8064 (lto_destroy_tree_ref_encoder): Likewise.
8065 * lto-section-out.c (lto_output_decl_index): Likewise.
8066 (lto_record_function_out_decl_state): Likewise.
8067 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
8068
3dbc97a9
RB
80692013-06-24 Richard Biener <rguenther@suse.de>
8070
8071 PR tree-optimization/57488
8072 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
8073
8345965a
AM
80742013-06-24 Alan Modra <amodra@gmail.com>
8075
8076 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
8077 (gen_easy_altivec_constant): Likewise.
8078 * config/rs6000/predicates.md (easy_vector_constant_add_self,
8079 easy_vector_constant_msb): Likewise.
8080
b7df379f
JJ
80812013-06-23 Jakub Jelinek <jakub@redhat.com>
8082
8083 PR target/57688
8084 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
8085 add missing return true.
8086
12211b99 80872013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE
8088
8089 PR target/52483
8090 * config/sh/predicates.md (general_extend_operand): Invoke
8091 general_movsrc_operand for memory operands.
8092 (general_movsrc_operand): Allow reg+reg addressing, do not use
8093 general_operand for memory operands.
8094
97db2bf7
ST
80952013-06-23 Sriraman Tallam <tmsriram@google.com>
8096
8097 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
8098 when current target options does not apply.
8099 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
8100 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB
8101 * config/i386/bmiintrin.h: Pass appropriate target
8102 attributes to header.
97db2bf7
ST
8103 * config/i386/mmintrin.h: Ditto.
8104 * config/i386/nmmintrin.h: Ditto.
8105 * config/i386/avx2intrin.h: Ditto.
8106 * config/i386/fxsrintrin.h: Ditto.
8107 * config/i386/tbmintrin.h: Ditto.
8108 * config/i386/xsaveintrin.h: Ditto.
8109 * config/i386/f16cintrin.h: Ditto.
8110 * config/i386/xtestintrin.h: Ditto.
8111 * config/i386/xsaveoptintrin.h: Ditto.
8112 * config/i386/bmi2intrin.h: Ditto.
8113 * config/i386/lzcntintrin.h: Ditto.
8114 * config/i386/smmintrin.h: Ditto.
8115 * config/i386/wmmintrin.h: Ditto.
8116 * config/i386/x86intrin.h: Remove all header include guards.
8117 * config/i386/prfchwintrin.h: Ditto.
8118 * config/i386/pmmintrin.h: Ditto.
8119 * config/i386/tmmintrin.h: Ditto.
8120 * config/i386/xmmintrin.h: Ditto.
8121 * config/i386/popcntintrin.h: Ditto.
8122 * config/i386/rdseedintrin.h: Ditto.
8123 * config/i386/ammintrin.h: Ditto.
8124 * config/i386/emmintrin.h: Ditto.
8125 * config/i386/immintrin.h: Remove all header include guards.
8126 * config/i386/fma4intrin.h: Ditto.
8127 * config/i386/lwpintrin.h: Ditto.
8128 * config/i386/xopintrin.h: Ditto.
8129 * config/i386/ia32intrin.h: Ditto.
8130 * config/i386/avxintrin.h: Ditto.
8131 * config/i386/rtmintrin.h: Ditto.
8132 * config/i386/fmaintrin.h: Ditto.
8133 * config/i386/mm3dnow.h: Ditto.
8134
495e6879
ST
81352013-06-22 Sriraman Tallam <tmsriram@google.com>
8136
8137 * common/config/i386/i386-common.c: Handle LZCNT.
8138
11e8f687
AK
81392013-06-22 Andi Kleen <ak@linux.intel.com>
8140
8141 * doc/extend.texi: Use __atomic_store_n instead of
8142 __atomic_store in HLE example.
8143
12211b99 81442013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE
8145
8146 * config/sh/sh.c: Remove <cstdlib> workaround.
8147
927f908b
AK
81482013-06-21 Andi Kleen <ak@linux.intel.com>
8149
0ba6aff3 8150 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 8151
cca410b6
AK
81522013-06-21 Andi Kleen <ak@linux.intel.com>
8153
8154 * doc/extend.texi: Document that __atomic_clear and
8155 __atomic_test_and_set should only be used with bool.
8156
d84db1ed
JH
81572013-06-20 Jan Hubicka <jh@suse.cz>
8158
8159 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
8160 types_same_for_odr.
8161 * tree.c (decls_same_for_odr): New function.
8162 (same_for_edr): New function.
8163 (types_same_for_odr): New function.
8164 (get_binfo_at_offset): Use it.
8165 * tree.h (types_same_for_odr): Declare.
8166
12211b99 81672013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE
8168 Jason Merrill <jason@redhat.com>
8169
8170 * system.h: Include <cstdlib> as well as <stdlib.h>.
8171
49b4ff63
UB
81722013-06-20 Uros Bizjak <ubizjak@gmail.com>
8173
8174 PR target/57655
8175 * config/i386/i386.c (construct_container): Report error if
8176 long double is used with disabled x87 float returns.
8177
7a89ad00
JH
81782013-06-20 Jan Hubicka <jh@suse.cz>
8179
8180 * lto-cgraph.c (input_symtab): Do not set cgraph state.
8181
12211b99 81822013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR
8183
8184 PR rtl-optimization/57425
8185 PR rtl-optimization/57569
8186 * alias.c (write_dependence_p): Remove parameters mem_mode and
8187 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
8188 Changed all callers.
8189 (canon_anti_dependence): Get comments and semantics in sync.
8190 Add parameter mem_canonicalized. Changed all callers.
8191 * rtl.h (canon_anti_dependence): Update prototype.
8192
d16e9a99
RB
81932013-06-20 Richard Biener <rguenther@suse.de>
8194
8195 * data-streamer-in.c (streamer_read_uhwi): Optimize single
8196 byte case, inline streamer_read_uchar and defer section
8197 overrun check.
8198
c3a9b91b
RB
81992013-06-20 Richard Biener <rguenther@suse.de>
8200
8201 PR tree-optimization/57584
8202 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
8203 SSA names into the expanded expression that take part in
8204 abnormal coalescing.
8205
371e77e3 82062013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS
8207
8208 * gcov.c (print_usage): Handle new option.
8209 (process_args): Ditto.
8210 (get_gcov_intermediate_filename): New function.
8211 (output_intermediate_file): New function.
8212 (output_gcov_file): New function
8213 (generate_results): Handle new option.
8214 (release_function): Relase demangled name.
8215 (read_graph_file): Handle demangled name.
8216 (output_lines): Ditto.
8217 * doc/gcov.texi: Document gcov intermediate format.
8218
11f5b71a
VM
82192013-06-19 Vladimir Makarov <vmakarov@redhat.com>
8220
8221 PR bootstrap/57604
8222 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
8223 (lra_emit_add): Use the functions. Add comment about Y as an
8224 address segment.
8225
e1f9340b
DE
82262013-06-19 David Edelsohn <dje.gcc@gmail.com>
8227
8228 PR driver/57652
8229 * collect2.c (collect_atexit): New.
8230 (collect_exit): Delete.
8231 (main): Register collect_atexit with atexit.
8232 (collect_wait): Change collect_exit to exit.
8233 (do_wait): Same.
8234 * collect2.h (collect_exit): Delete.
8235 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
8236
e3f9e0ac
WM
82372013-06-19 Wei Mi <wmi@google.com>
8238
8239 PR rtl-optimization/57518
8240 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8241 if regno is used in paradoxical subreg.
8242 (update_equiv_regs): Check pdx_subregs[regno] before
8243 set a reg to be equivalent with a mem.
8244
a09f1a76
MK
82452013-06-19 Matthias Klose <doko@ubuntu.com>
8246
8247 PR driver/57651
8248 * file-find.h (find_a_file): Add a mode parameter.
8249 * file-find.c (find_a_file): Likewise.
8250 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8251 with X_OK for the executables.
8252 * collect2.c (main): Call find_a_file with X_OK.
8253
7d18b0ad
SE
82542013-06-19 Steve Ellcey <sellcey@mips.com>
8255
8256 PR target/56942
8257 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8258 Use NEXT_INSN instead of next_real_insn.
8259
6a6dac52
JH
82602013-06-19 Jan Hubicka <jh@suse.cz>
8261
8262 * cgraph.h (const_value_known_p): Replace by ...
8263 (ctor_for_folding): .. this one.
8264 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8265 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8266 * expr.c (expand_expr_real_1): Likewise.
8267 (string_constant): Likewise.
8268 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8269 * ipa.c (process_references): Likewise.
8270 (symtab_remove_unreachable_nodes): Likewise.
8271 * ipa-inline-analysis.c (param_change_prob): Likewise.
8272 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8273 (get_base_constructor): Likwise.
8274 * varpool.c (varpool_remove_node): Likewise.
8275 (varpool_remove_initializer): LIkewise.
8276 (dump_varpool_node): LIkwise.
8277 (const_value_known_p): Rewrite to ...
8278 (ctor_for_folding): ... this one.
8279
216c12ab
JJ
82802013-06-19 Jakub Jelinek <jakub@redhat.com>
8281
8282 PR driver/57651
8283 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8284 PERSONALITY in $PATH derived prefixes.
8285
d53e2f99
JL
82862013-06-19 Jeff Law <law@redhat.com>
8287
ecdbf306
JL
8288 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8289 in comment.
8290
d53e2f99
JL
8291 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8292 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8293 booleans.
8294
89e7fc04
SN
82952013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8296
8297 * config/arm/vfp.md: Move VFP instruction classification documentation
8298 to ...
8299 * config/arm/arm.md: ... here. Update instruction classification
8300 documentation.
8301
5e7f6aaa
RE
83022013-06-19 Richard Earnshaw <rearnsha@arm.com>
8303
8304 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8305 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8306 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8307
5fdc1e5d
SB
83082013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8309
8310 PR target/57609
8311 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8312 with NEXT_INSN. Use tablejump_p to check for jump table data
8313 insns.
8314
e952afba
PC
83152013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8316
8317 PR c++/56544
8318 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8319 that now in C++ the value is correct per the C++ standards.
8320
3fd0303a
RB
83212013-06-19 Richard Biener <rguenther@suse.de>
8322
8323 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8324 for global context.
8325
b04e4b08
AK
83262013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8327
8328 Revert:
8329 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8330
8331 PR target/57609
8332 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8333 with next_active_insn.
8334
bef8491a
ST
83352013-06-18 Sriraman Tallam <tmsriram@google.com>
8336
8337 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
8338 functions are inlined during failures to flag an error.
8339 * tree-inline.c (expand_call_inline): Allow the error to be flagged
8340 in early inline pass.
22fc7d3d 8341
c21b257d
L
83422013-06-18 H.J. Lu <hongjiu.lu@intel.com>
8343
8344 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
8345 in comments.
8346
33255ae3
JB
83472013-06-18 Julian Brown <julian@codesourcery.com>
8348
8349 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
8350 Permit virtual register pre-reload if !strict.
8351 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
8352 change.
8353 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
8354 prototype.
8355 * config/arm/neon.md (movmisalign<mode>): Use
8356 neon_perm_struct_or_reg_operand instead of
8357 neon_struct_or_register_operand.
8358 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
8359 neon_permissive_struct_operand instead of neon_struct_operand.
8360 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
8361 neon_vector_mem_operand.
8362 * config/arm/predicates.md (neon_struct_operand): Adjust call to
8363 neon_vector_mem_operand.
8364 (neon_permissive_struct_operand): New.
8365 (neon_struct_or_register_operand): Rename to...
8366 (neon_perm_struct_or_reg_operand): This. Adjust call to
8367 neon_vector_mem_operand.
8368
d579fcda
RB
83692013-06-18 Richard Biener <rguenther@suse.de>
8370
8371 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
8372 * lto-streamer.h: Include pointer-set.h.
8373 (struct lto_decl_slot): Remove.
8374 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
8375 Remove next_index entry.
8376 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8377 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8378 (lto_init_tree_ref_encoder): Adjust.
8379 (lto_destroy_tree_ref_encoder): Likewise.
8380 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8381 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8382 (lto_output_decl_index): Adjust.
8383 (lto_new_out_decl_state): Likewise.
8384 (lto_record_function_out_decl_state): Likewise.
8385 * lto-streamer-out.c (copy_function): Likewise.
8386
31ee20ba
RB
83872013-06-18 Richard Biener <rguenther@suse.de>
8388
8389 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
8390 * cgraphunit.c: Include cfgloop.h.
8391 (init_lowered_empty_function): Initialize the loop tree.
8392 (assemble_thunk): Insert new BBs into loops.
8393
ac0511f2
RB
83942013-06-18 Richard Biener <rguenther@suse.de>
8395
8396 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8397 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8398 the map from cache entry to cache index optional.
8399 (streamer_tree_cache_replace_tree): Adjust accordingly.
8400 (streamer_tree_cache_append): Likewise.
8401 (streamer_tree_cache_delete): Likewise.
8402 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8403 streamer cache map from cache entry to cache index.
8404 * lto-streamer-out.c (create_output_block): Adjust.
8405
09485a08
SN
84062013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8407
22fc7d3d
UB
8408 * config/arm/arm.md (attribute "insn"): Move multiplication and
8409 division attributes to...
09485a08
SN
8410 (attribute "type"): ... here. Remove mult.
8411 (attribute "mul32"): New attribute.
8412 (attribute "mul64"): Add umaal.
8413 (*arm_mulsi3): Update attributes.
8414 (*arm_mulsi3_v6): Likewise.
8415 (*thumb_mulsi3): Likewise.
8416 (*thumb_mulsi3_v6): Likewise.
8417 (*mulsi3_compare0): Likewise.
8418 (*mulsi3_compare0_v6): Likewise.
8419 (*mulsi_compare0_scratch): Likewise.
8420 (*mulsi_compare0_scratch_v6): Likewise.
8421 (*mulsi3addsi): Likewise.
8422 (*mulsi3addsi_v6): Likewise.
8423 (*mulsi3addsi_compare0): Likewise.
8424 (*mulsi3addsi_compare0_v6): Likewise.
8425 (*mulsi3addsi_compare0_scratch): Likewise.
8426 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8427 (*mulsi3subsi): Likewise.
8428 (*mulsidi3adddi): Likewise.
8429 (*mulsi3addsi_v6): Likewise.
8430 (*mulsidi3adddi_v6): Likewise.
8431 (*mulsidi3_nov6): Likewise.
8432 (*mulsidi3_v6): Likewise.
8433 (*umulsidi3_nov6): Likewise.
8434 (*umulsidi3_v6): Likewise.
8435 (*umulsidi3adddi): Likewise.
8436 (*umulsidi3adddi_v6): Likewise.
8437 (*smulsi3_highpart_nov6): Likewise.
8438 (*smulsi3_highpart_v6): Likewise.
8439 (*umulsi3_highpart_nov6): Likewise.
8440 (*umulsi3_highpart_v6): Likewise.
8441 (mulhisi3): Likewise.
8442 (*mulhisi3tb): Likewise.
8443 (*mulhisi3bt): Likewise.
8444 (*mulhisi3tt): Likewise.
8445 (maddhisi4): Likewise.
8446 (*maddhisi4tb): Likewise.
8447 (*maddhisi4tt): Likewise.
8448 (maddhidi4): Likewise.
8449 (*maddhidi4tb): Likewise.
8450 (*maddhidi4tt): Likewise.
8451 (divsi3): Likewise.
8452 (udivsi3): Likewise.
8453 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8454 (thumb2_mulsi_short_compare0): Likewise.
8455 (thumb2_mulsi_short_compare0_scratch): Likewise.
8456 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8457 (1020mult2): Likewise.
8458 (1020mult3): Likewise.
8459 (1020mult4): Likewise.
8460 (1020mult5): Likewise.
8461 (1020mult6): Likewise.
22fc7d3d
UB
8462 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8463 change.
09485a08
SN
8464 (cortex_a15_mult64): Likewise.
8465 (cortex_a15_sdiv): Likewise.
8466 (cortex_a15_udiv): Likewise.
8467 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8468 (mult2): Likewise.
8469 (mult3): Likewise.
8470 (mult4): Likewise.
8471 (mult5): Likewise.
8472 (mult6): Likewise.
8473 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8474 (pj4_ir_div): Likewise.
8475 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8476 (11_mult2): Likewise.
8477 (11_mult3): Likewise.
8478 (11_mult4): Likewise.
8479 (11_mult5): Likewise.
8480 (11_mult6): Likewise.
8481 (11_mult7): Likewise.
8482 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8483 (cortex_a8_mla): Likewise.
8484 (cortex_a8_mull): Likewise.
8485 (cortex_a8_smulwy): Likewise.
8486 (cortex_a8_smlald): Likewise.
8487 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8488 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8489 (cortex_r4_mul_3): Likewise.
8490 (cortex_r4_mla_4): Likewise.
8491 (cortex_r4_mla_3): Likewise.
8492 (cortex_r4_smlald): Likewise.
8493 (cortex_r4_mull): Likewise.
8494 (cortex_r4_sdiv): Likewise.
8495 (cortex_r4_udiv): Likewise.
8496 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8497 (cortex_a7_idiv): Likewise.
8498 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8499 (9_mult2): Likewise.
8500 (9_mult3): Likewise.
8501 (9_mult4): Likewise.
8502 (9_mult5): Likewise.
8503 (9_mult6): Likewise.
8504 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8505 (cortex_a53_sdiv): Likewise.
8506 (cortex_a53_udiv): Likewise.
8507 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8508 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8509 (mp626_mult2): Likewise.
8510 (mp626_mult3): Likewise.
8511 (mp626_mult4): Likewise.
8512 * config/arm/fa526.md (526_mult1): Update attribute change.
8513 (526_mult2): Likewise.
8514 * config/arm/arm-generic.md (mult): Update attribute change.
8515 (mult_ldsched_strongarm): Likewise.
8516 (mult_ldsched): Likewise.
8517 (multi_cycle): Likewise.
8518 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8519 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8520 (606te_mult2): Likewise.
8521 (606te_mult3): Likewise.
8522 (606te_mult4): Likewise.
8523 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8524 (cortex_a9_mac16): Likewise.
8525 (cortex_a9_multiply): Likewise.
8526 (cortex_a9_mac): Likewise.
8527 (cortex_a9_multiply_long): Likewise.
8528 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8529 (626te_mult2): Likewise.
8530 (626te_mult3): Likewise.
8531 (626te_mult4): Likewise.
8532
8951f345
RB
85332013-06-18 Richard Biener <rguenther@suse.de>
8534
8535 PR lto/57334
8536 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8537
d8101d05
AK
85382013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8539
8540 PR target/57609
8541 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8542 with next_active_insn.
8543
69eff9da
AM
85442013-06-18 Alan Modra <amodra@gmail.com>
8545
8546 * config/rs6000/rs6000.h (enum data_align): New.
8547 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8548 (DATA_ABI_ALIGNMENT): Define.
8549 (CONSTANT_ALIGNMENT): Correct comment.
8550 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8551 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8552
8ed00d76
DM
85532013-06-17 David Malcolm <dmalcolm@redhat.com>
8554
8555 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8556 ATTRIBUTE_UNUSED marking.
8557
3d33d06b
SN
85582013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8559
8560 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8561 alternative and update.
8562 (aarch64_dup_lanedi): Delete.
8563 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8564 * config/aarch64/aarch64-simd-builtins.def: Update.
8565
ee03e71d
RB
85662013-06-17 Richard Biener <rguenther@suse.de>
8567
8568 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8569 (lto_input_scc): Declare.
8570 (lto_input_tree_1): Likewise.
8571 (struct lto_stats_d): Add num_tree_bodies_output and
8572 num_pickle_refs_output.
8573 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8574 (lto_read_tree_1): Split out from ...
8575 (lto_read_tree): ... this.
8576 (lto_input_scc): New function.
8577 (lto_input_tree_1): Split out from ...
8578 (lto_input_tree): ... this. Handle LTO_tree_scc.
8579 (lto_data_in_create): Create the streamer cache without hashes.
8580 * lto-streamer-out.c (create_output_block): Create the streamer
8581 cache with hashes when not doing WPA.
8582 (lto_write_tree_1): Split out from ...
8583 (lto_write_tree): ... this.
8584 (get_symbol_initial_value): New function.
8585 (lto_output_tree_1): Split out from ...
8586 (lto_output_tree): ... this. Write trees as series of SCCs
8587 using a DFS walk via DFS_write_tree.
8588 (struct sccs, struct scc_entry): New types.
8589 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
8590 (DFS_write_tree_body): New function.
8591 (DFS_write_tree): Likewise.
8592 (hash_tree): Likewise.
8593 (scc_entry_compare): Likewise.
8594 (hash_scc): Likewise.
8595 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
8596 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
8597 TREE_CHAIN as regular reference.
8598 (streamer_read_integer_cst): Remove.
8599 (streamer_get_pickled_tree): Adjust.
8600 * tree-streamer-out.c (streamer_write_chain): Disable streaming
8601 of DECL_EXTERNALs in BLOCK_VARS for now.
8602 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
8603 reference.
8604 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
8605 Add hash value argument and record that if hashes are recorded
8606 in the cache.
8607 (streamer_tree_cache_insert_1): Adjust.
8608 (streamer_tree_cache_insert): Likewise.
8609 (streamer_tree_cache_insert_at): Rename to ...
8610 (streamer_tree_cache_replace_tree): ... this and adjust.
8611 (streamer_tree_cache_append): Adjust.
8612 (record_common_node): Likewise.
8613 (streamer_tree_cache_create): Add argument whether to
8614 record hash values together with trees.
8615 (streamer_tree_cache_delete): Adjust.
8616 * tree-streamer.h (struct streamer_tree_cache_d): Add
8617 vector of hashes.
8618 (streamer_read_integer_cst): Remove.
8619 (streamer_tree_cache_insert): Adjust.
8620 (streamer_tree_cache_append): Likewise.
8621 (streamer_tree_cache_insert_at): Rename to ...
8622 (streamer_tree_cache_replace_tree): ... this and adjust.
8623 (streamer_tree_cache_create): Add argument whether to record hashes.
8624 (streamer_tree_cache_get): Rename to ...
8625 (streamer_tree_cache_get_tree): ... this.
8626 (streamer_tree_cache_get_hash): New function.
8627 * tree.c (cache_integer_cst): New function.
8628 * tree.h (cache_integer_cst): Declare.
8629 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
8630 * lto-symtab.c (lto_varpool_replace_node): Only release
8631 DECL_INITIAL of non-prevailing decls.
8632 * varpool.c (varpool_remove_initializer): Do not release
8633 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
8634
0de86a92
JU
86352013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8636
8637 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
8638 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
8639 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
8640 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
8641 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
8642 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
8643 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
8644 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
8645 instead of TARGET_64BIT.
8646 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
8647 Require ISA_HAS_<D>DIV.
8648
74240413
RS
86492013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
8650
8651 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
8652 (mips*-*-linux*): Move default with_llsc setting to where other
8653 defaults are set.
8654 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
8655 with_arch block.
8656 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
8657 Likewise. Remove default with_tune setting. Move default float
8658 setting to its own block. Handle with_llsc in the same block as above.
8659
12211b99 86602013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR
8661
8662 PR rtl-optimization/57425
8663 PR rtl-optimization/57569
299a5f6a 8664 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR
8665 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
8666 Changed all callers.
8667 (canon_anti_dependence): New function.
8668 * cse.c (check_dependence): Use canon_anti_dependence.
8669 * cselib.c (cselib_invalidate_mem): Likewise.
8670 * rtl.h (canon_anti_dependence): Declare.
8671
b259d352
JU
86722013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8673
8674 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
8675 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
8676 ".set mips3" for 64-bit targets.
8677
4de09b85
DC
86782013-06-15 Dehao Chen <dehao@google.com>
8679
8680 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
8681 * gimple-low.c (gimple_check_call_matching_types): Likewise.
8682 (gimple_check_call_args): Likewise.
8683 * value-prof.c (check_ic_target): Likewise.
8684 * ipa-inline.c (early_inliner): Likewise.
8685 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
8686 * cgraph.c (cgraph_create_edge_1): Likewise.
8687 (cgraph_make_edge_direct): Likewise.
8688
d1aa4a76
MM
86892013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
8690
8691 PR target/57615
8692 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
8693 rs6000_output_move_128bit to handle emitting quad memory
8694 operations. Set attribute length to 8 bytes.
8695
12211b99 86962013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP
8697
8698 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
8699 New pattern.
8700 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
8701 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
8702 (aarch64_<su>mlsl<mode>): Likewise.
8703
f5273574
MS
87042013-06-14 Mike Stump <mikestump@comcast.net>
8705
8706 * Makefile.in (TARGET_H): Add insn-codes.h.
8707
ef7e7ab2
AM
87082013-06-14 Alan Modra <amodra@gmail.com>
8709
8710 PR middle-end/57134
8711 PR middle-end/57586
8712 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
8713 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
8714 bitfield expansion when EXPAND_MEMORY.
8715 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
8716
135faab6
MM
87172013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
8718
8719 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
8720 test for clearing quad memory on 32-bit later.
8721
948a1fd9
MG
87222013-06-13 Marc Glisse <marc.glisse@inria.fr>
8723
8724 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
8725 (fold_negate_expr): Likewise.
8726 (fold_real_zero_addition_p): Handle vectors.
8727 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
8728
0597fb9c
AM
87292013-06-14 Alan Modra <amodra@gmail.com>
8730
8731 * varasm.c (force_const_mem): Revert 2013-06-07 change.
8732
e5b962d0
JH
87332013-06-13 Jan Hubicka <jh@suse.cz>
8734
8735 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
8736 Local comdats are not externally visible.
8737 * symtab.c (dump_symtab_base): Dump externally visible.
8738 (verify_symtab_base): Verify back links in the symtab hash.
8739
a8355e51
BC
87402013-06-13 Bin Cheng <bin.cheng@arm.com>
8741
8742 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
8743 CONVERT_EXPR as equal nodes.
8744
8f7e6e33
BC
87452013-06-13 Bin Cheng <bin.cheng@arm.com>
8746
22fc7d3d 8747 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 8748
5d418483
MG
87492013-06-13 Marc Glisse <marc.glisse@inria.fr>
8750
8751 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
8752 Generalize to complex and vector.
8753 * tree.c (build_all_ones_cst): New function.
8754 * tree.h (build_all_ones_cst): Declare it.
8755
50751417
AM
87562013-06-13 Alan Modra <amodra@gmail.com>
8757
8758 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
8759 * config/rs6000/rs6000.md (signbittf2): New insn.
8760 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
8761 (abstf2_internal, cmptf_internal2): Likewise.
8762 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
8763
4b02c962 87642013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
8765 Pat Haugen <pthaugen@us.ibm.com>
8766 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM
8767
8768 * config/rs6000/rs6000.c (emit_load_locked): Add support for
8769 power8 byte, half-word, and quad-word atomic instructions.
8770 (emit_store_conditional): Likewise.
8771 (rs6000_expand_atomic_compare_and_swap): Likewise.
8772 (rs6000_expand_atomic_op): Likewise.
8773
8774 * config/rs6000/sync.md (larx): Add new modes for power8.
8775 (stcx): Likewise.
8776 (AINT): New mode iterator to include TImode as well as normal
8777 integer modes on power8.
8778 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
8779 that VSX registers are not considered. Use AINT mode iterator
8780 instead of INT1 to allow inclusion of quad word atomic operations
8781 on power8.
8782 (load_locked<mode>): Likewise.
8783 (store_conditional<mode>): Likewise.
8784 (atomic_compare_and_swap<mode>): Likewise.
8785 (atomic_exchange<mode>): Likewise.
8786 (atomic_nand<mode>): Likewise.
8787 (atomic_fetch_<fetchop_name><mode>): Likewise.
8788 (atomic_nand_fetch<mode>): Likewise.
8789 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
8790 each type.
8791 (ATOMIC): On power8, add QImode, HImode modes.
8792 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
8793 modes that promote to SImode.
8794 (load_lockedti): Convert TImode arguments to PTImode, so that we
8795 get a guaranteed even/odd register pair.
8796 (load_lockedpti): Likewise.
8797 (store_conditionalti): Likewise.
8798 (store_conditionalpti): Likewise.
8799
8800 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
8801 atomic load/store instructions.
8802 (HSI): Likewise.
8803
710b6cc1
RS
88042013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
8805
8806 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
8807 loads.
8808 (insn_count): New attribute, with most cases extracted from...
8809 (length): ...here. Redefine most cases in terms of insn_count.
8810 (single_insn): Delete.
8811 (can_delay): Use insn_count to check for single instructions.
8812 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
8813 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
8814 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
8815 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
8816 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
8817 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
8818 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
8819 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
8820 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
8821 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
8822 rather than "length".
8823 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
8824 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
8825 Use "insn_count" rather than "length".
8826 * config/mips/mips-dsp.md
8827 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
8828 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
8829 length attributes.
8830
1951f101
MG
88312013-06-12 Marc Glisse <marc.glisse@inria.fr>
8832
8833 PR tree-optimization/57361
8834 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
8835
8b033a8a
SN
88362013-06-12 Sofiane Naci <sofiane.naci@arm.com>
8837
22fc7d3d
UB
8838 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
8839 to split.
8b033a8a
SN
8840 (aarch64_simd_combine<mode>): New instruction expansion.
8841 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
8842 function prototype.
8843 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
8844 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
8845
ca0f62a8
JH
88462013-06-12 Jan Hubicka <jh@suse.cz>
8847
8848 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
8849 decl has when in streaming stage.
8850 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
8851 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
8852
12211b99 88532013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS
8854
8855 PR target/57578
8856 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
8857
bfdeda2c
JJ
88582013-06-12 Jakub Jelinek <jakub@redhat.com>
8859
8860 PR tree-optimization/57537
8861 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
8862 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
8863
a4fa02d1
RB
88642013-06-12 Richard Biener <rguenther@suse.de>
8865
8866 * data-streamer.h (streamer_write_char_stream): CSE
8867 obs->current_pointer.
8868 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
8869 streamer_write_char_stream manually and optimize the resulting loop.
8870 (streamer_write_hwi_stream): Likewise.
8871
bbf9ad07
JH
88722013-06-12 Jan Hubicka <jh@suse.cz>
8873
8874 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
8875 * cgraph.h (varpool_create_empty_node): Declare.
8876 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
8877 duplicated nodes.
8878 * symtab.c (symtab_unregister_node): Be lax about missin entries
8879 in node hash.
8880 (symtab_get_node): Update comment.
8881 * varpool.c (varpool_create_empty_node): Break out from ...
8882 (varpool_node_for_decl): ... here.
8883 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
8884
a53f90ad
EB
88852013-06-12 Eric Botcazou <ebotcazou@adacore.com>
8886
8887 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
8888 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
8889 part. Use straight-line flow at the end.
8890 <COMPONENT_REF>: Remove superfluous else.
8891 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
8892
946f9306
JJ
88932013-06-12 Jakub Jelinek <jakub@redhat.com>
8894
8895 PR target/56564
8896 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
8897 target hook even for !TREE_PUBLIC decls. If no resolution info
8898 is available, return false for common and external decls.
8899
a7fca6f0
KP
89002013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8901
8902 * config/rl78/constraints.md (U): New constraint.
8903 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
8904 valloc attribute.
8905
c7d8f446
MM
89062013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8907
8908 PR target/57589
8909 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
8910 to allow returning address to AT_PLATFORM name.
8911
08346abd
JH
89122013-06-11 Jan Hubicka <jh@suse.cz>
8913
8914 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
8915 * cgraph.h (symtab_node_base): Add weakref flag.
8916 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
8917 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
8918 (output_weakrefs): Use weakref flag.
8919 * fold-const.c (simple_operand_p): Handle WEAK.
8920 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
8921 * ipa.c (varpool_externally_visible_p): Drop weakref.
8922 (function_and_variable_visibility): Update comment; fix weakref
8923 sanity checks; do not clear DECL_WEAK on them.
8924 * lto-cgraph.c (lto_output_node): update.
8925 (lto_output_varpool_node): Update.
8926 (input_overwrite_node): Update.
8927 (input_node): Update.
8928 (input_varpool_node): Update.
8929 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
8930 (lto_symtab_merge_symbols): Add sanity check.
8931 (lto_symtab_prevailing_decl): Do not special case weakrefs.
8932 * passes.c (rest_of_decl_compilation): Set static flag, too.
8933 * symtab.c (dump_symtab_base): Dump weakref.
8934 (verify_symtab_base): Sanity check weakrefs.
8935 (symtab_make_decl_local): Remove duplicated code.
8936 (symtab_alias_ultimate_target): Simplify.
8937 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 8938
2e6a2f88
TV
89392013-06-11 Tom de Vries <tom@codesourcery.com>
8940
8941 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
8942 sequence_vect == NULL.
8943
d9af3652
DD
89442013-06-11 DJ Delorie <dj@redhat.com>
8945
8946 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
8947 (rl78_unwind_word_mode): New.
8948
ffa66012
DM
89492013-06-11 David Malcolm <dmalcolm@redhat.com>
8950
8951 * final.c (debug_prefix_maps): Make static.
8952
783f2d35
DM
89532013-06-11 David Malcolm <dmalcolm@redhat.com>
8954
8955 * function.c (initial_trampoline): Remove stray copy.
8956
7af26def
SN
89572013-06-11 Sofiane Naci <sofiane.naci@arm.com>
8958
8959 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
8960
97756c0e
MJ
89612013-06-11 Martin Jambor <mjambor@suse.cz>
8962
8963 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
8964 within bounds at the beginning of the function.
8965
2c7b8bf6
AM
89662013-06-11 Alan Modra <amodra@gmail.com>
8967
8968 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
8969 reporting.
8970 (get_named_section): Don't NULL !DECL_P decl.
8971
77cea46e
IZ
89722013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
8973
8974 * doc/invoke.texi (core-avx2): Document.
8975 (slm): Likewise.
8976 (atom): Updated with MOVBE.
8977
888f0920
RB
89782013-06-11 Richard Biener <rguenther@suse.de>
8979
22fc7d3d 8980 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 8981
6b438d58
AB
89822013-06-11 Anton Blanchard <anton@samba.org>
8983
8984 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
8985 correct shift value in little-endian mode.
8986
3bc66938
JJ
89872013-06-11 Jakub Jelinek <jakub@redhat.com>
8988
8989 PR target/56564
8990 * varasm.c (get_variable_align): Move #endif to the right place.
8991
a5dc7f88
CC
89922013-06-10 Cary Coutant <ccoutant@google.com>
8993
8994 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
8995 for hash so that hash table traversal order is deterministic.
8996
19be72ab 89972013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
8998 Pat Haugen <pthaugen@us.ibm.com>
8999 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM
9000
9001 * config/rs6000/vector.md (GPR move splitter): Do not split moves
9002 of vectors in GPRS if they are direct moves or quad word load or
9003 store moves.
9004
9005 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
9006 declaration.
9007 (direct_move_p): Likewise.
9008 (quad_load_store_p): Likewise.
9009
9010 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
9011 classes into bins based on the physical register type.
9012 (reg_class_to_reg_type): Likewise.
9013 (IS_STD_REG_TYPE): Likewise.
9014 (IS_FP_VECT_REG_TYPE): Likewise.
9015 (reload_fpr_gpr): Arrays to determine what insn to use if we can
9016 use direct move instructions.
9017 (reload_gpr_vsx): Likewise.
9018 (reload_vsx_gpr): Likewise.
9019 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
9020 information that is a simplification of register classes. Also
9021 precalculate direct move reload helpers.
9022 (direct_move_p): New function to return true if the operation can
9023 be done as a direct move instruciton.
9024 (quad_load_store_p): New function to return true if the operation
9025 is a quad memory operation.
9026 (rs6000_legitimize_address): If quad memory, only allow register
9027 indirect for TImode addresses.
9028 (rs6000_legitimate_address_p): Likewise.
9029 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
9030 (rs6000_reload_register_type): Likewise.
9031 (register_to_reg_type): Return register type.
9032 (rs6000_secondary_reload_simple_move): New helper function for
9033 secondary reload and secondary memory needed to identify anything
9034 that is a simple move, and does not need reloading.
9035 (rs6000_secondary_reload_direct_move): New helper function for
9036 secondary reload to identify cases that can be done with several
9037 instructions via the direct move instructions.
9038 (rs6000_secondary_reload_move): New helper function for secondary
9039 reload to identify moves between register types that can be done.
9040 (rs6000_secondary_reload): Add support for quad memory operations
9041 and for direct move.
9042 (rs6000_secondary_memory_needed): Likewise.
9043 (rs6000_debug_secondary_memory_needed): Change argument names.
9044 (rs6000_output_move_128bit): New function to return the move to
9045 use for 128-bit moves, including knowing about the various
9046 limitations of quad memory operations.
9047
9048 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
9049 memory operations. call rs6000_output_move_128bit for the actual
9050 instruciton(s) to generate.
9051 (vsx_movti_64bit): Likewise.
9052
9053 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
9054 (UNSPEC_P8V_MTVSRWZ): Likewise.
9055 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
9056 (UNSPEC_P8V_MTVSRD): Likewise.
9057 (UNSPEC_P8V_XXPERMDI): Likewise.
9058 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
9059 (UNSPEC_FUSION_GPR): Likewise.
9060 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 9061 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM
9062 (f32_sv): Likewise.
9063 (f32_dm): Likewise.
9064 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
9065 loads and direct move instructions.
9066 (zero_extendsidi2_lfiwzx): Likewise.
9067 (extendsidi2_lfiwax): Likewise.
9068 (extendsidi2_nocell): Likewise.
9069 (floatsi<mode>2_lfiwax): Likewise.
9070 (lfiwax): Likewise.
9071 (floatunssi<mode>2_lfiwzx): Likewise.
9072 (lfiwzx): Likewise.
9073 (fix_trunc<mode>_stfiwx): Likewise.
9074 (fixuns_trunc<mode>_stfiwx): Likewise.
9075 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
9076 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
9077 (parity<mode>2_cmpb): Set length/type attr.
9078 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
9079 for 'mr.' to fast_compare.
9080 (bpermd_<mode>): Change type attr to popcnt.
9081 (p8_fmrgow_<mode>): New insns for power8 direct move support.
9082 (p8_mtvsrwz_1): Likewise.
9083 (p8_mtvsrwz_2): Likewise.
9084 (reload_fpr_from_gpr<mode>): Likewise.
9085 (p8_mtvsrd_1): Likewise.
9086 (p8_mtvsrd_2): Likewise.
9087 (p8_xxpermdi_<mode>): Likewise.
9088 (reload_vsx_from_gpr<mode>): Likewise.
9089 (reload_vsx_from_gprsf): Likewise.
9090 (p8_mfvsrd_3_<mode>): LIkewise.
9091 (reload_gpr_from_vsx<mode>): Likewise.
9092 (reload_gpr_from_vsxsf): Likewise.
9093 (p8_mfvsrd_4_disf): Likewise.
9094 (multi-word GPR splits): Do not split direct moves or quad memory
9095 operations.
9096
16876bdc
DM
90972013-06-10 David Malcolm <dmalcolm@redhat.com>
9098
9099 * tree-into-ssa.c (interesting_blocks): Make static.
9100
df8a1d28
JJ
91012013-06-10 Jakub Jelinek <jakub@redhat.com>
9102
9103 PR target/56564
9104 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
9105 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
9106 Use DATA_ABI_ALIGNMENT for that case instead if defined.
9107 (get_variable_align): New function.
9108 (get_variable_section, emit_bss, emit_common,
9109 assemble_variable_contents, place_block_symbol): Use
9110 get_variable_align instead of DECL_ALIGN.
9111 (assemble_noswitch_variable): Add align argument, use it
9112 instead of DECL_ALIGN.
9113 (assemble_variable): Adjust caller. Use get_variable_align
9114 instead of DECL_ALIGN.
9115 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
9116 caller.
9117 (DATA_ABI_ALIGNMENT): Define.
9118 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
9119 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
9120 opt is false, only return the psABI mandated alignment increase.
9121 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
9122 (DATA_ABI_ALIGNMENT): ... this.
9123 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
9124 (DATA_ABI_ALIGNMENT): ... this.
9125 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
9126 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
9127 (DATA_ABI_ALIGNMENT): ... this.
9128 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
9129 * doc/tm.texi: Regenerated.
9130
57b29ca6
UB
91312013-06-10 Uros Bizjak <ubizjak@gmail.com>
9132
9133 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
9134 cmp_code to construct REG_EQUAL note.
9135
3e56ed50
JJ
91362013-06-09 Jakub Jelinek <jakub@redhat.com>
9137
9138 PR target/57568
9139 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
9140 that operands[2] doesn't overlap with operands[0].
9141
74fe2a1d
DE
91422013-06-09 David Edelsohn <dje.gcc@gmail.com>
9143 Jan Hubicka <jh@suse.cz>
9144
9145 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
9146 hack to mark symbols as used.
9147
ed52a84e
VM
91482013-06-08 Vladimir Makarov <vmakarov@redhat.com>
9149
9150 PR rtl-optimization/57559
9151 * lra-constraints.c (process_alt_operands): Don't discourage
9152 memory with known offset for offsetable memory constraint.
9153 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
9154
299316ed
EB
91552013-06-08 Eric Botcazou <ebotcazou@adacore.com>
9156
9157 * varasm.c (struct oc_local_state): Reorder fields.
9158 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
9159 and adjust accordingly.
9160 (output_constructor): Reorder initialization code and adjust call to
9161 output_constructor_bitfield.
9162
88ac60d3
JH
91632013-06-07 Jan Hubicka <jh@suse.cz>
9164
9165 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
9166
4cdce1a8
DM
91672013-06-07 David Malcolm <dmalcolm@redhat.com>
9168
9169 * tree-object-size.c (unknown): Make const.
9170
b5e0425c
AK
91712013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9172
9173 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
9174 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
9175 for last alternative in the cpu_facility attribute.
9176
79678d04
KT
91772013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9178
9179 PR target/56315
9180 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
9181 (xordi3): Change operand 2 constraint to arm_xordi_operand.
9182 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
9183 * config/arm/constraints.md (Dg): New constraint.
9184 * config/arm/neon.md (xordi3_neon): Remove.
9185 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
9186 * config/arm/predicates.md (arm_xordi_operand): New predicate.
9187
b31ddbdb
KT
91882013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9189
9190 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
9191 Clean up alternatives.
9192
8f90b7d4
AM
91932013-06-07 Alan Modra <amodra@gmail.com>
9194
9195 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
9196 va_list_gpr_size.
9197
961ce119
AM
91982013-06-07 Alan Modra <amodra@gmail.com>
9199
9200 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
9201
33e49835
KT
92022013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9203
9204 * config/arm/constraints.md (Df): New constraint.
9205 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
9206 Correct length attribute for last two alternatives.
9207
acd17ae6
AM
92082013-06-07 Alan Modra <amodra@gmail.com>
9209
9210 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9211 override user -mfp-in-toc.
9212 (offsettable_ok_by_alignment): Consider just the current access
9213 rather than the whole object, unless BLKmode. Handle
9214 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
9215 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
9216 for -mcmodel=medium.
9217 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
9218 override user -mfp-in-toc or -msum-in-toc. Default to
9219 -mno-fp-in-toc for -mcmodel=medium.
9220
73310b0e
DD
92212013-06-06 DJ Delorie <dj@redhat.com>
9222
9223 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
9224 TARGET_VALID_POINTER_MODE.
9225
0bd62dca 92262013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
9227 Pat Haugen <pthaugen@us.ibm.com>
9228 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM
9229
9230 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9231 Document new power8 builtins.
9232
9233 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
9234 condition code register, to allow 128-bit logical operations to be
9235 done in the VSX or GPR registers.
9236 (nor<mode>3): Use the canonical form for nor.
9237 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9238 vclz*, and vpopcnt* vector instructions.
9239 (nand<mode>3): Likewise.
9240 (orc<mode>3): Likewise.
9241 (clz<mode>2): LIkewise.
9242 (popcount<mode>2): Likewise.
9243
9244 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9245 that only the GPRs are recognized.
9246
9247 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9248 support for new power8 builtins.
9249
9250 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9251 builtin functions.
9252 (xscvdpspn): Likewise.
9253 (vclz): Likewise.
9254 (vclzb): Likewise.
9255 (vclzh): Likewise.
9256 (vclzw): Likewise.
9257 (vclzd): Likewise.
9258 (vpopcnt): Likewise.
9259 (vpopcntb): Likewise.
9260 (vpopcnth): Likewise.
9261 (vpopcntw): Likewise.
9262 (vpopcntd): Likewise.
9263 (vgbbd): Likewise.
9264 (vmrgew): Likewise.
9265 (vmrgow): Likewise.
9266 (eqv): Likewise.
9267 (eqv_v16qi3): Likewise.
9268 (eqv_v8hi3): Likewise.
9269 (eqv_v4si3): Likewise.
9270 (eqv_v2di3): Likewise.
9271 (eqv_v4sf3): Likewise.
9272 (eqv_v2df3): Likewise.
9273 (nand): Likewise.
9274 (nand_v16qi3): Likewise.
9275 (nand_v8hi3): Likewise.
9276 (nand_v4si3): Likewise.
9277 (nand_v2di3): Likewise.
9278 (nand_v4sf3): Likewise.
9279 (nand_v2df3): Likewise.
9280 (orc): Likewise.
9281 (orc_v16qi3): Likewise.
9282 (orc_v8hi3): Likewise.
9283 (orc_v4si3): Likewise.
9284 (orc_v2di3): Likewise.
9285 (orc_v4sf3): Likewise.
9286 (orc_v2df3): Likewise.
9287
9288 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9289 allow power8 quad mode in 64-bit.
9290 (rs6000_builtin_vectorized_function): Add support to vectorize
9291 ISA 2.07 count leading zeros, population count builtins.
9292 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9293 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9294 (builtin_function_type): Add vgbbd builtin function which takes an
9295 unsigned argument.
9296 (altivec_expand_vec_perm_const): Add support for new power8 merge
9297 instructions.
9298
9299 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9300 that does not include TImdoe for use with 32-bit.
9301 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9302 instructions.
9303 (UNSPEC_VSX_CVDPSPN): Likewise.
9304 (vsx_xscvdpspn): Likewise.
9305 (vsx_xscvspdpn): Likewise.
9306 (vsx_xscvdpspn_scalar): Likewise.
9307 (vsx_xscvspdpn_directmove): Likewise.
9308 (vsx_and<mode>3): Split logical operations into 32-bit and
9309 64-bit. Add support to do logical operations on TImode as well as
9310 VSX vector types. Allow logical operations to be done in either
9311 VSX registers or in general purpose registers in 64-bit mode. Add
9312 splitters if GPRs were used. For AND, add clobber of CCmode to
9313 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9314 encoding.
9315 (vsx_and<mode>3_32bit): Likewise.
9316 (vsx_and<mode>3_64bit): Likewise.
9317 (vsx_ior<mode>3): Likewise.
9318 (vsx_ior<mode>3_32bit): Likewise.
9319 (vsx_ior<mode>3_64bit): Likewise.
9320 (vsx_xor<mode>3): Likewise.
9321 (vsx_xor<mode>3_32bit): Likewise.
9322 (vsx_xor<mode>3_64bit): Likewise.
9323 (vsx_one_cmpl<mode>2): Likewise.
9324 (vsx_one_cmpl<mode>2_32bit): Likewise.
9325 (vsx_one_cmpl<mode>2_64bit): Likewise.
9326 (vsx_nor<mode>3): Likewise.
9327 (vsx_nor<mode>3_32bit): Likewise.
9328 (vsx_nor<mode>3_64bit): Likewise.
9329 (vsx_andc<mode>3): Likewise.
9330 (vsx_andc<mode>3_32bit): Likewise.
9331 (vsx_andc<mode>3_64bit): Likewise.
9332 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
9333 and xxlorc instructions.
9334 (vsx_eqv<mode>3_64bit): Likewise.
9335 (vsx_nand<mode>3_32bit): Likewise.
9336 (vsx_nand<mode>3_64bit): Likewise.
9337 (vsx_orc<mode>3_32bit): Likewise.
9338 (vsx_orc<mode>3_64bit): Likewise.
9339
9340 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
9341
9342 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
9343 instruction.
9344 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
9345 (p8_vmrgow): Likewise.
9346 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
9347 GPRs to be split under VSX.
9348 (p8v_clz<mode>2): Add power8 count leading zero support.
9349 (p8v_popcount<mode>2): Add power8 population count support.
9350 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
9351 support.
9352
9353 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
9354 instruction.
9355
9356 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
9357 builtin functions.
9358 (vec_nand): Likewise.
9359 (vec_vclz): Likewise.
9360 (vec_vclzb): Likewise.
9361 (vec_vclzd): Likewise.
9362 (vec_vclzh): Likewise.
9363 (vec_vclzw): Likewise.
9364 (vec_vgbbd): Likewise.
9365 (vec_vmrgew): Likewise.
9366 (vec_vmrgow): Likewise.
9367 (vec_vpopcnt): Likewise.
9368 (vec_vpopcntb): Likewise.
9369 (vec_vpopcntd): Likewise.
9370 (vec_vpopcnth): Likewise.
9371 (vec_vpopcntw): Likewise.
9372
37684c46
VM
93732013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9374
9375 PR rtl-optimization/57468
9376 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
9377 spilled pseudos.
9378
d89ae27c
VM
93792013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9380
9381 PR rtl-optimization/57459
9382 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
9383 type when setting live regs.
9384
3597e113
VM
93852013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9386
9387 * config/s390/s390.opt (mlra): New option.
9388 * config/s390/s390.c (s390_decompose_address): Check displacement
9389 for all registers for LRA.
9390 (s390_secondary_reload): Don't used secondary reloads for LRA.
9391 (s390_lra_p): New function.
9392 (TARGET_LRA_P): Define.
9393 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
9394 of attribute cpu_facility to zarch for the last alternative.
9395 (*cmpmem_short): Ditto.
9396
01037aeb
EB
93972013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9398
9399 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9400 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9401 (arm_expand_prologue): Likewise.
9402
3371a64f
TJ
94032013-06-06 Teresa Johnson <tejohnson@google.com>
9404
9405 PR c++/53743
9406 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9407 as this is now done by redirect_edge_and_branch_force.
9408 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9409 barriers, and fix interaction with splitting.
9410 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9411 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9412 reflect changes made in the routine.
9413 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9414 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9415 since this is called in cfglayout mode, and replace partition fixup
9416 with assert as that is now done by force_nonfallthru_and_redirect.
9417 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9418 already be marked with region crossing note.
9419 (insert_section_boundary_note): Make non-static, gate on flag
9420 has_bb_partition, rewrite to also check for multiple partitions.
9421 (rest_of_handle_reorder_blocks): Remove call to
9422 insert_section_boundary_note, now done later during free_cfg.
9423 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9424 * bb-reorder.h (insert_section_boundary_note): Declare.
9425 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9426 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9427 invoke insert_section_boundary_note.
9428 (try_redirect_by_replacing_jump): Remove unnecessary
9429 check for region crossing note.
9430 (fixup_partition_crossing): New function.
9431 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9432 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9433 in non-cfglayout mode.
9434 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9435 remove old code that tried to do this. Emit barrier correctly
9436 when we are in cfglayout mode.
9437 (last_bb_in_partition): New function.
9438 (rtl_split_edge): Correctly fixup partition boundaries.
9439 (commit_one_edge_insertion): Remove old code that tried to
9440 fixup region crossing edge since this is now handled in
9441 split_block, and set up insertion point correctly since
9442 block may now end in a jump.
9443 (verify_hot_cold_block_grouping): Guard against checking when not in
9444 linearized RTL mode.
9445 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9446 notes.
9447 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9448 rtl_verify_flow_info, so not called in cfglayout mode.
9449 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9450 (fixup_reorder_chain): Remove old code that attempted to fixup region
9451 crossing note as this is now handled in force_nonfallthru_and_redirect.
9452 (duplicate_insn_chain): Don't duplicate switch section notes.
9453 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9454 note.
9455 * basic-block.h (emit_barrier_after_bb): Declare.
9456
66071e10
KT
94572013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9458
9459 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9460 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9461 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9462
0e26bf3d
KT
94632013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9464
9465 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9466 where appropriate.
9467 * config/arm/ldmstm.md: Regenerate.
9468
12b4e7ef
KT
94692013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9470
9471 * config/arm/sync.md (atomic_loaddi_1):
9472 Disable predication for arm_restrict_it.
9473 (arm_load_exclusive<mode>): Likewise.
9474 (arm_load_exclusivesi): Likewise.
9475 (arm_load_exclusivedi): Likewise.
9476 (arm_load_acquire_exclusive<mode>): Likewise.
9477 (arm_load_acquire_exclusivesi): Likewise.
9478 (arm_load_acquire_exclusivedi): Likewise.
9479 (arm_store_exclusive<mode>): Likewise.
9480 (arm_store_exclusive<mode>): Likewise.
9481 (arm_store_release_exclusivedi): Likewise.
9482 (arm_store_release_exclusive<mode>): Likewise.
9483
15d16c8a
RB
94842013-06-06 Richard Biener <rguenther@suse.de>
9485
9486 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9487 after LTO_null.
9488 (lto_tag_is_tree_code_p): Adjust.
9489 (lto_tag_is_gimple_code_p): Likewise.
9490 (lto_gimple_code_to_tag): Likewise.
9491 (lto_tag_to_gimple_code): Likewise.
9492 (lto_tree_code_to_tag): Likewise.
9493 (lto_tag_to_tree_code): Likewise.
9494 * data-streamer.h (streamer_write_hwi_in_range): Use
9495 uhwi streaming to stream the normalized range.
9496 (streamer_read_hwi_in_range): Likewise.
9497
17a7fc37
KT
94982013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9499
9500 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9501 (predicable_short_it): Likewise.
9502 (predicated): Likewise.
9503 (enabled): Handle above.
9504 (define_cond_exec): Set predicated attribute to yes.
9505
b57ca59b
MS
95062013-06-05 Mike Stump <mikestump@comcast.net>
9507
9508 * gdbinit.in (__FUNCTION__): Add.
9509
c1e183a9
UB
95102013-06-05 Uros Bizjak <ubizjak@gmail.com>
9511
9512 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9513 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9514
0378bddb
JH
95152013-06-05 Jan Hubicka <jh@suse.cz>
9516
9517 * varasm.c (mark_decl_referenced): Revert the removal until targets
9518 are fixed.
9519
8f8a2057
DE
95202013-06-05 David Edelsohn <dje.gcc@gmail.com>
9521
9522 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9523 instead of mark_decl_referenced.
9524
edb983b2
JH
95252013-06-05 Jan Hubicka <jh@suse.cz>
9526
9527 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9528 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9529 and symtab_used_from_object_file_p.
9530 (cgraph_make_node_local_1): Clear forced_by_abi.
9531 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9532 * cgraph.h (symtab_node_base): Add forced_by_abi.
9533 (decide_is_variable_needed): Remove.
9534 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9535 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9536 (decide_is_symbol_needed): ... this one; handle symbols in general;
9537 always analyze virtuals; honnor forced_by_abi.
9538 (cgraph_finalize_function): Update.
9539 (varpool_finalize_decl): Update.
9540 (symbol_defined_and_needed): Remove.
9541 (analyze_functions): Update.
9542 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9543 output_refs, input_overwrite_node): Handle forced_by_abi.
9544 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9545 (address_taken_from_non_vtable_p): ... this one.
9546 (comdat_can_be_unshared_p_1): New function.
9547 (cgraph_comdat_can_be_unshared_p): Rename to ...
9548 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9549 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9550 (function_and_variable_visibility): Clear forced_by_abi as needed.
9551 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9552 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH
9553 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9554 * varpool.c (decide_is_variable_needed): Remove.
9555
9912dbe5
KT
95562013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9557
9558 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9559 (arm_option_override): Override arm_restrict_it where appropriate.
9560 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9561 * config/arm/arm.opt (mrestrict-it): New command-line option.
9562 * doc/invoke.texi: Document -mrestrict-it.
9563
34c136b6
DM
95642013-06-05 David Malcolm <dmalcolm@redhat.com>
9565
9566 * tsan.c (tsan_atomic_table): Make const.
9567
a367df53
RB
95682013-06-05 Richard Biener <rguenther@suse.de>
9569
9570 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9571 index associated with the tree we are supposed to replace.
9572 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9573 TREE_ASM_WRITTEN as zero for everything but SSA names.
9574
70d51a19
DM
95752013-06-05 David Malcolm <dmalcolm@redhat.com>
9576
9577 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9578
b9bd2075
JH
95792013-06-04 Jan Hubicka <jh@suse.cz>
9580
9581 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9582 (input_node, input_varpool_node): Handle correctly external same
9583 body aliases.
9584 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9585 nodes at ltrans stage.
9586
aaae719d
JH
95872013-06-04 Jan Hubicka <jh@suse.cz>
9588
9589 * ipa-inline.c (update_caller_keys): Fix availability test.
9590 (update_callee_keys): Likewise.
9591 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
9592 to follow ELF standard.
9593
107eea2c
JU
95942013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
9595
9596 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
9597 (mips64r5900el-*-elf*): New configurations.
9598 * config/mips/mips-cpus.def (r5900): New processor.
9599 * config/mips/mips-tables.opt: Regenerate.
9600 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
9601 (mips_issue_rate): Handle PROCESSOR_R5900.
9602 (mips_reorg_process_insns): Force reorder mode for the R5900.
9603 * config/mips/mips.h (TARGET_MIPS5900): Define.
9604 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
9605 TARGET_MIPS5900.
9606 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
9607 TARGET_MIPS5900.
9608 * config/mips/mips.md (processor): Add r5900.
9609 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
9610
b7342d25
IB
96112013-06-04 Ian Bolton <ian.bolton@arm.com>
9612
9613 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
9614 into function to generate MOVI instruction.
c1e183a9 9615 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB
9616 (aarch64_preferred_simd_mode): Turn into wrapper.
9617 (aarch64_output_scalar_simd_mov_immediate): New function.
9618 * config/aarch64/aarch64-protos.h: Add prototype for above.
9619
81c2dfb9
IB
96202013-06-04 Ian Bolton <ian.bolton@arm.com>
9621
9622 * config/aarch64/aarch64.c (simd_immediate_info): Remove
9623 element_char member.
9624 (sizetochar): Return signed char.
9625 (aarch64_simd_valid_immediate): Remove elchar and other
9626 unnecessary variables.
9627 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
9628 Calculate element_char as required.
9629 * config/aarch64/aarch64-protos.h: Update and move prototype
9630 for aarch64_output_simd_mov_immediate.
9631 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
9632 Update arguments.
9633
48063b9d
IB
96342013-06-04 Ian Bolton <ian.bolton@arm.com>
9635
9636 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
9637 information completed by aarch64_simd_valid_immediate.
9638 (aarch64_legitimate_constant_p): Update arguments.
9639 (aarch64_simd_valid_immediate): Work with struct rather than many
9640 pointers.
9641 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
9642 (aarch64_simd_make_constant): Update arguments.
9643 (aarch64_output_simd_mov_immediate): Work with struct rather than
9644 many pointers. Output immediate directly rather than as operand.
9645 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9646 Update prototype.
9647 * config/aarch64/constraints.md (Dn): Update arguments.
9648
3ea63f60
IB
96492013-06-04 Ian Bolton <ian.bolton@arm.com>
9650
9651 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
9652 longer static.
9653 (aarch64_simd_immediate_valid_for_move): Remove.
9654 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
9655 (aarch64_simd_make_constant): Update call.
9656 (aarch64_output_simd_mov_immediate): Update call.
9657 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9658 Add prototype.
9659 * config/aarch64/constraints.md (Dn): Update call.
9660
d8edd899
IB
96612013-06-04 Ian Bolton <ian.bolton@arm.com>
9662
9663 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
9664 return type to bool for prototype.
9665 (aarch64_legitimate_constant_p): Check for true instead of not -1.
9666 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
9667 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
9668
44b20bb8
CM
96692013-06-04 Catherine Moore <clm@codesourcery.com>
9670
9671 * config/mips/mips.opt (meva): New.
c1e183a9 9672 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM
9673 (ASM_SPEC): Handle -meva.
9674 * doc/invoke.texi (meva): Document.
9675
52befbd8
AM
96762013-06-04 Alan Modra <amodra@gmail.com>
9677
9678 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
9679 constant output.
9680
aadaf24e
KT
96812013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9682
9683 * rtl.def: Add extra fourth optional field to define_cond_exec.
9684 * gensupport.c (process_one_cond_exec): Process attributes from
9685 define_cond_exec.
9686 * doc/md.texi: Document fourth field in define_cond_exec.
9687
0cd9e9ee
EB
96882013-06-04 Eric Botcazou <ebotcazou@adacore.com>
9689
9690 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
9691 out the processing order as in store_bit_field_1.
9692
cb0f665d
JH
96932013-06-04 Jan Hubicka <jh@suse.cz>
9694
9695 PR middle-end/57500
9696 * cgraphunit.c (cgraph_process_same_body_aliases): Create
9697 non-VAR_DECL node if it does not exist yet.
9698
53984b9b
RS
96992013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
9700
9701 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
9702 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
9703 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
9704 target_cpu_default setting.
9705
5d318fd4
TJ
97062013-06-03 Teresa Johnson <tejohnson@google.com>
9707
9708 * dumpfile.c (opt_info_switch_p): Change -fopt-info
9709 default to -fopt-info=optimized instead of all.
9710 * doc/invoke.texi: Ditto.
9711 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
9712 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
9713 (execute_vect_slp): Emit BB vectorization success under
9714 MSG_OPTIMIZED_LOCATIONS.
9715 * tree-vect-slp.c (vect_slp_transform_bb): Change
9716 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
9717 * tree-vect-loop.c (vect_transform_loop): Ditto.
9718
ec3af349
JM
97192013-06-03 Jason Merrill <jason@redhat.com>
9720
9721 PR c++/57415
9722 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9723 Use TARGET_EXPR for C++.
9724
1baf9159
JJ
97252013-06-03 Jakub Jelinek <jakub@redhat.com>
9726
9727 PR rtl-optimization/57268
0cd9e9ee 9728 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ
9729 if DEBUG_INSN_P (insn).
9730
9731 Reapply
9732 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9733
9734 PR rtl-optimization/57268
9735 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9736 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9737
0f1d3965
YR
97382013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9739
9740 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
9741 (ix86_avoid_lea_for_addr): Likewise.
9742 (exact_dependency_1): Likewise.
9743 (ix86_adjust_cost): Likewise.
9744 (swap_top_of_ready_list): Fix formatting and !reload_completed check
9745 removed.
9746 (do_reorder_for_imul): Fix typo, formatting and
9747 !reload_completed check removed.
9748 (ix86_sched_reorder): Fix typo and formatting.
9749 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
9750 list.
9751
3bc30db4
SN
97522013-06-03 Sofiane Naci <sofiane.naci@arm.com>
9753
9754 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
9755
19cc6fac
EB
97562013-06-03 Eric Botcazou <ebotcazou@adacore.com>
9757
9758 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
9759 <STRING_CST>: Likewise.
9760 <VECTOR_CST>: Likewise.
9761
c68a6e08
JW
97622013-06-01 Janus Weil <janus@gcc.gnu.org>
9763 Mikael Morin <mikael@gcc.gnu.org>
9764
9765 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
9766 * config.in: Regenerated.
9767 * configure: Regenerated.
9768
38e55e5c
JH
97692013-06-01 Jan Hubicka <jh@suse.cz>
9770
9771 PR middle-end/57366
9772 * cgraphunit.c (compile): When weakref is not supported,
9773 set up transparent aliases before final output pass.
9774 * varasm.c (assemble_alias): Do not try to do it here.
9775
eb51d2ff
JH
97762013-06-01 Jan Hubicka <jh@suse.cz>
9777
9778 PR middle-end/57467
9779 * passes.c (for_per_function): Skip unanalyzed functions.
9780
40a7fe1e
JH
97812013-06-01 Jan Hubicka <jh@suse.cz>
9782
9783 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
9784 (lto_symtab_merge_symbols_1): ... this one.
9785 (lto_symtab_merge_cgraph_nodes): Rename to ...
9786 (lto_symtab_merge_symbols): ... this one; simplify.
9787 * cgraph.c (same_body_aliases_done): Rename to ...
9788 (cpp_implicit_aliases_done): ... this one.
9789 (cgraph_create_function_alias): Update.
9790 (cgraph_same_body_alias): Update.
c1e183a9 9791 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 9792 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 9793 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH
9794 (cgraph_node): Remove same_body_alias.
9795 (varpool_node): Remove alias_of and extra_name_alias.
9796 (same_body_aliases_done): Rename to ..
9797 (cpp_implicit_aliases_done): ... this one.
9798 (symtab_alias_ultimate_target): Add default parameter.
9799 (symtab_resolve_alias): New function.
9800 (fixup_same_cpp_alias_visibility): Declare.
9801 (cgraph_function_node): Add default parameter.
9802 (cgraph_node_asm_name): Likewise.
9803 (cgraph_function_or_thunk_node): Add default parameter; do
9804 not ICE when it is NULL.
9805 (varpool_variable_node): Likewise.
9806 * tree-emutls.c (create_emultls_var): Update.
9807 (ipa_lower_emutls): Update.
9808 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
9809 (cgraph_reset_node): Reset alias info.
9810 (cgraph_finalize_function): Update.
9811 (fixup_same_cpp_alias_visibility): Move to symtab.c.
9812 (analyze_function): Simplify.
9813 (cgraph_process_same_body_aliases): Simplify.
9814 (analyze_functions): Fixup same body aliases.
9815 (handle_alias_pairs): Simplify.
9816 (assemble_thunk): Update.
9817 (assemble_thunks_and_aliases): Update.
9818 (output_weakrefs): Rewrite.
9819 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
9820 (lto_output_varpool_node): Likewise.
9821 (compute_ltrans_boundary): Remve assert.
9822 (get_alias_symbol): New functoin.
9823 (input_node): Rewrite alias handling.
9824 (input_varpool_node): Likewise.
9825 * ipa-pure-const.c (propagate_pure_const): Fix formating.
9826 * ipa.c (process_references): Handle weakrefs correctly.
9827 (symtab_remove_unreachable_nodes): Likewise.
9828 * trans-mem.c (get_cg_data): Update.
9829 (ipa_tm_create_version_alias): Update.
9830 (ipa_tm_execute): Update.
9831 * symtab.c (dump_symtab_base): Dump aliases.
9832 (verify_symtab_base): Verify aliases.
9833 (symtab_node_availability): New function.
9834 (symtab_alias_ultimate_target): Simplify.
9835 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
9836 handle all the fixup cases.
9837 (symtab_resolve_alias): New function.
9838 * passes.c (ipa_write_summaries): Handle weakrefs.
9839 * varpool.c (varpool_analyze_node): Simplify.
9840 (assemble_aliases): Update.
9841 (varpool_create_variable_alias): Simplify.
9842 (varpool_extra_name_alias): Simplify.
9843 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
9844 (lto_symtab_merge_symbols): ... this one.
9845
21759881
DT
98462013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
9847
9848 Revert
9849 PR rtl-optimization/57268
9850 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9851 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9852
daf4e940
TB
98532013-06-01 Tobias Burnus <burnus@net-b.de>
9854
9855 Partially reverted:
9856 2013-05-31 Tobias Burnus <burnus@net-b.de>
9857
9858 PR middle-end/57073
9859 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
9860 further up.
9861
c359d8d0
DT
98622013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9863
9864 PR rtl-optimization/57268
9865 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9866 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9867
4172215d
EB
98682013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9869
9870 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
9871 unordered comparison operators when -fno-trapping-math is in effect
9872 on the e500.
9873 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
9874 and implement unordered comparison operators properly on the e500.
9875
a8c50132
EB
98762013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9877
9878 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
9879 for constant scalar integers.
9880 (simplify_relational_operation_1): Likewise.
9881
44626634
SB
98822013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
9883
9884 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
9885 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
9886 Fix comment.
9887
55805e54
YR
98882013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9889 Igor Zamyatin <igor.zamyatin@intel.com>
9890
9891 Silvermont (SLM) architecture performance tuning.
9892 * config/i386/i386.h (enum ix86_tune_indices): Add
9893 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
9894 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
9895
9896 * config/i386/i386.c (initial_ix86_tune_features)
9897 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
9898 (ix86_lea_outperforms): Handle Silvermont tuning.
9899 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
9900 call.
9901 (ix86_use_lea_for_mov): Likewise.
9902 (ix86_avoid_lea_for_addr): Likewise.
9903 (ix86_lea_for_add_ok): Likewise.
9904 (exact_dependency_1): New function.
9905 (exact_store_load_dependency): Likewise.
9906 (ix86_adjust_cost): Handle Silvermont tuning.
9907 (do_reoder_for_imul): Likewise.
9908 (swap_top_of_ready_list): New function.
9909 (ix86_sched_reorder): Changed to handle Silvermont tuning.
9910
9911 * config/i386/i386.md (peepholes that split memory operand in fp
9912 converts): New.
9913
e19c9de2
MS
99142013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9915
9916 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
9917 Remove un-necessary braces.
9918
38e6c9a6
MS
99192013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9920
9921 * config/aarch64/aarch64.c (aarch64_classify_symbol):
9922 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
9923
e3530904
TB
99242013-05-31 Tobias Burnus <burnus@net-b.de>
9925
9926 PR middle-end/57073
c1e183a9 9927 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 9928
b6af05a9
KT
99292013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9930
9931 PR target/56315
9932 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
9933 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
9934 * config/arm/neon.md (iordi3_neon): Remove.
9935 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
9936 * config/arm/predicates.md (imm_for_neon_logic_operand):
9937 Move to earlier in the file.
9938 (neon_logic_op2): Likewise.
9939 (arm_iordi_operand_neon): New predicate.
9940
f800c166
RB
99412013-05-31 Richard Biener <rguenther@suse.de>
9942
9943 PR tree-optimization/57478
9944 PR tree-optimization/57453
9945 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
9946 are life as well.
9947
e571fa59
KP
99482013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9949
9950 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
9951 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
9952
0fa6e0ef
TB
99532013-05-30 Tobias Burnus <burnus@net-b.de>
9954 Thomas Koenig <tkoenig@gcc.gnu.org>
9955
9956 PR middle-end/57073
9957 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
9958 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
9959
4688ddf5
SB
99602013-05-30 Steven Bosscher <steven@gcc.gnu.org>
9961
0fa6e0ef 9962 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 9963
3b9ceb4b
VM
99642013-05-30 Vladimir Makarov <vmakarov@redhat.com>
9965
9966 * target.def (register_usage_leveling_p): New hook.
9967 * targhooks.c (default_register_usage_leveling_p): New.
9968 * targhooks.h (default_register_usage_leveling_p): New prototype.
9969 * lra-assigns.c (register_usage_leveling_p): Use the hook.
9970 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
9971 * doc/tm.texi: Update.
9972 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
9973
65f9e789
IB
99742013-05-30 Ian Bolton <ian.bolton@arm.com>
9975
9976 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
9977 (*insv_reg<mode>): New define_insn.
9978
12211b99 99792013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR
9980
9981 PR rtl-optimization/57439
9982 * postreload.c (move2add_valid_value_p): Check that we have
9983 a zero subreg_regno_offset when accessing the register in
9984 the requested mode.
9985
0b871ccf
YR
99862013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9987 Igor Zamyatin <igor.zamyatin@intel.com>
9988
9989 Silvermont (SLM) architecture pipeline model, tuning and
9990 insn selection.
9991 * config.gcc: Add slm config options and target.
9992
9993 * config/i386/slm.md: New.
9994
9995 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
9996
f43245d1 9997 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR
9998 PROCESSOR_SLM.
9999 (ix86_target_macros_internal): Likewise.
10000
f43245d1 10001 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR
10002 (m_SLM): New macro flag.
10003 (initial_ix86_tune_features): Set m_SLM.
10004 (x86_accumulate_outgoing_args): Likewise.
10005 (x86_arch_always_fancy_math_387): Likewise.
10006 (processor_target_table): Add slm cost.
10007 (cpu_names): Add slm cpu name.
10008 (x86_option_override_internal): Set SLM ISA.
10009 (ix86_issue_rate): New case PROCESSOR_SLM.
10010 (ia32_multipass_dfa_lookahead): Likewise.
10011 (fold_builtin_cpu): Add slm.
10012
10013 * config/i386/i386.h (TARGET_SLM): New target macro.
10014 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
10015 (processor_type): Add PROCESSOR_SLM.
10016
10017 * config/i386/i386.md (cpu): Add new value "slm".
10018 (slm.md): Include slm.md.
10019
24d5b097
XG
100202013-05-30 Bernd Schmidt <bernds@codesourcery.com>
10021 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10022
10023 * config/arm/arm-protos.h: Add and update function protos.
10024 * config/arm/arm.c (use_simple_return_p): New added.
10025 (thumb2_expand_return): Check simple_return flag.
10026 * config/arm/arm.md: Add simple_return and conditional simple_return.
10027 * config/arm/iterators.md: Add iterator for return and simple_return.
10028
c1cccc15
ZC
100292013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10030
10031 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
10032 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
10033 (arm_emit_vfp_multi_reg_pop): Likewise.
10034 (thumb2_emit_ldrd_pop): Likewise.
10035 (arm_expand_epilogue): Add misc REG_CFA notes.
10036 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
10037
3a4f280b
LC
100382013-05-29 Lawrence Crowl <crowl@google.com>
10039
10040 * config/arm/t-arm: Update for below.
10041
10042 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
10043 Change type to hash_table. Update dependent calls and types.
10044
10045 * config/i386/t-cygming: Update for below.
10046
10047 * config/i386/t-interix: Update for below.
10048
10049 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
10050 Change type to hash_table. Update dependent calls and types.
10051 (i386_find_on_wrapper_list::wrappers): Likewise.
10052
10053 * config/ia64/t-ia64: Update for below.
10054
10055 * config/ia64/ia64.c (bundle_state_table):
10056 Change type to hash_table. Update dependent calls and types.
10057
10058 * config/mips/mips.c (mips_reorg_process_insns::htab):
10059 Change type to hash_table. Update dependent calls and types.
10060
10061 * config/sol2.c (solaris_comdat_htab):
10062 Change type to hash_table. Update dependent calls and types.
10063
10064 * config/t-sol2: Update for above.
10065
1388a0e3
TJ
100662013-05-29 Teresa Johnson <tejohnson@google.com>
10067
10068 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
10069 functions are not yet marked as defined.
10070
a5965b52 100712013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
10072 Pat Haugen <pthaugen@us.ibm.com>
10073 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM
10074
10075 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
10076 instructions.
10077 (VEC_A): Likewise.
10078 (VEC_C): Likewise.
10079 (vrotl<mode>3): Likewise.
10080 (vashl<mode>3): Likewise.
10081 (vlshr<mode>3): Likewise.
10082 (vashr<mode>3): Likewise.
10083
10084 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10085 support for power8 V2DI builtins.
10086
10087 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
10088 power8 V2DI builtins.
10089 (vupkhsw): Likewise.
10090 (vupklsw): Likewise.
10091 (vaddudm): Likewise.
10092 (vminsd): Likewise.
10093 (vmaxsd): Likewise.
10094 (vminud): Likewise.
10095 (vmaxud): Likewise.
10096 (vpkudum): Likewise.
10097 (vpksdss): Likewise.
10098 (vpkudus): Likewise.
10099 (vpksdus): Likewise.
10100 (vrld): Likewise.
10101 (vsld): Likewise.
10102 (vsrd): Likewise.
10103 (vsrad): Likewise.
10104 (vsubudm): Likewise.
10105 (vcmpequd): Likewise.
10106 (vcmpgtsd): Likewise.
10107 (vcmpgtud): Likewise.
10108 (vcmpequd_p): Likewise.
10109 (vcmpgtsd_p): Likewise.
10110 (vcmpgtud_p): Likewise.
10111 (vupkhsw): Likewise.
10112 (vupklsw): Likewise.
10113 (vaddudm): Likewise.
10114 (vmaxsd): Likewise.
10115 (vmaxud): Likewise.
10116 (vminsd): Likewise.
10117 (vminud): Likewise.
10118 (vpksdss): Likewise.
10119 (vpksdus): Likewise.
10120 (vpkudum): Likewise.
10121 (vpkudus): Likewise.
10122 (vrld): Likewise.
10123 (vsld): Likewise.
10124 (vsrad): Likewise.
10125 (vsrd): Likewise.
10126 (vsubudm): Likewise.
10127
10128 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
10129 support for power8 V2DI instructions.
10130
10131 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
10132 power8 V2DI instructions. Combine pack and unpack insns to use an
10133 iterator for each mode. Check whether a particular mode supports
10134 Altivec instructions instead of just checking TARGET_ALTIVEC.
10135 (UNSPEC_VPKUWUM): Likewise.
10136 (UNSPEC_VPKSHSS): Likewise.
10137 (UNSPEC_VPKSWSS): Likewise.
10138 (UNSPEC_VPKUHUS): Likewise.
10139 (UNSPEC_VPKSHUS): Likewise.
10140 (UNSPEC_VPKUWUS): Likewise.
10141 (UNSPEC_VPKSWUS): Likewise.
10142 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
10143 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
10144 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
10145 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
10146 (UNSPEC_VUPKHSB): Likewise.
10147 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
10148 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
10149 (UNSPEC_VUPKHSH): Likewise.
10150 (UNSPEC_VUPKLSB): Likewise.
10151 (UNSPEC_VUPKLSH): Likewise.
10152 (VI2): Likewise.
10153 (VI_char): Likewise.
10154 (VI_scalar): Likewise.
10155 (VI_unit): Likewise.
10156 (VP): Likewise.
10157 (VP_small): Likewise.
10158 (VP_small_lc): Likewise.
10159 (VU_char): Likewise.
10160 (add<mode>3): Likewise.
10161 (altivec_vaddcuw): Likewise.
10162 (altivec_vaddu<VI_char>s): Likewise.
10163 (altivec_vadds<VI_char>s): Likewise.
10164 (sub<mode>3): Likewise.
10165 (altivec_vsubcuw): Likewise.
10166 (altivec_vsubu<VI_char>s): Likewise.
10167 (altivec_vsubs<VI_char>s): Likewise.
10168 (altivec_vavgs<VI_char>): Likewise.
10169 (altivec_vcmpbfp): Likewise.
10170 (altivec_eq<mode>): Likewise.
10171 (altivec_gt<mode>): Likewise.
10172 (altivec_gtu<mode>): Likewise.
10173 (umax<mode>3): Likewise.
10174 (smax<mode>3): Likewise.
10175 (umin<mode>3): Likewise.
10176 (smin<mode>3): Likewise.
10177 (altivec_vpkuhum): Likewise.
10178 (altivec_vpkuwum): Likewise.
10179 (altivec_vpkshss): Likewise.
10180 (altivec_vpkswss): Likewise.
10181 (altivec_vpkuhus): Likewise.
10182 (altivec_vpkshus): Likewise.
10183 (altivec_vpkuwus): Likewise.
10184 (altivec_vpkswus): Likewise.
10185 (altivec_vpks<VI_char>ss): Likewise.
10186 (altivec_vpks<VI_char>us): Likewise.
10187 (altivec_vpku<VI_char>us): Likewise.
10188 (altivec_vpku<VI_char>um): Likewise.
10189 (altivec_vrl<VI_char>): Likewise.
10190 (altivec_vsl<VI_char>): Likewise.
10191 (altivec_vsr<VI_char>): Likewise.
10192 (altivec_vsra<VI_char>): Likewise.
10193 (altivec_vsldoi_<mode>): Likewise.
10194 (altivec_vupkhsb): Likewise.
10195 (altivec_vupkhs<VU_char>): Likewise.
10196 (altivec_vupkls<VU_char>): Likewise.
10197 (altivec_vupkhsh): Likewise.
10198 (altivec_vupklsb): Likewise.
10199 (altivec_vupklsh): Likewise.
10200 (altivec_vcmpequ<VI_char>_p): Likewise.
10201 (altivec_vcmpgts<VI_char>_p): Likewise.
10202 (altivec_vcmpgtu<VI_char>_p): Likewise.
10203 (abs<mode>2): Likewise.
10204 (vec_unpacks_hi_v16qi): Likewise.
10205 (vec_unpacks_hi_v8hi): Likewise.
10206 (vec_unpacks_lo_v16qi): Likewise.
10207 (vec_unpacks_hi_<VP_small_lc>): Likewise.
10208 (vec_unpacks_lo_v8hi): Likewise.
10209 (vec_unpacks_lo_<VP_small_lc>): Likewise.
10210 (vec_pack_trunc_v8h): Likewise.
10211 (vec_pack_trunc_v4si): Likewise.
10212 (vec_pack_trunc_<mode>): Likewise.
10213
10214 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
10215 V2DI builtins.
10216 (vec_vmaxsd): Likewise.
10217 (vec_vmaxud): Likewise.
10218 (vec_vminsd): Likewise.
10219 (vec_vminud): Likewise.
10220 (vec_vpksdss): Likewise.
10221 (vec_vpksdus): Likewise.
10222 (vec_vpkudum): Likewise.
10223 (vec_vpkudus): Likewise.
10224 (vec_vrld): Likewise.
10225 (vec_vsld): Likewise.
10226 (vec_vsrad): Likewise.
10227 (vec_vsrd): Likewise.
10228 (vec_vsubudm): Likewise.
10229 (vec_vupkhsw): Likewise.
10230 (vec_vupklsw): Likewise.
10231
e70670cf
JH
102322013-05-29 Jan Hubicka <jh@suse.cz>
10233
10234 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
10235 flags; reorder rest of fields in more consistent way.
10236 (varpool_node): Remove analyzed, finalized and alias.
10237 (cgraph_ndoe): Likewise.
10238 (symtab_alias_ultimate_target): New function.
10239 (cgraph_function_node): Move offline.
10240 (cgraph_reset_node): Declare.
10241 (cgraph_comdat_can_be_unshared_p): Remove.
10242 (varpool_remove_initializer): Declare.
10243 (varpool_first_defined_variable, varpool_next_defined_variable
10244 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10245 (cgraph_function_with_gimple_body_p): Update.
10246 (varpool_all_refs_explicit_p): Update.
10247 (symtab_alias_target): New function.
10248 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10249 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB
10250 (cgraph_function_or_thunk_node): Simplify using
10251 symtab_alias_ultimate_target.
e70670cf
JH
10252 (varpool_variable_node): Likewise.
10253 * cgraph.c (cgraph_create_function_alias): Update.
10254 (cgraph_add_thunk): Update.
10255 (cgraph_remove_node): Update.
10256 (dump_cgraph_node): Do not dump removed flags.
10257 (cgraph_function_body_availability): Update.
10258 (cgraph_propagate_frequency): Update.
10259 (verify_cgraph_node): Check sanity of local flag.
10260 (cgraph_function_node): Move here from cgraph.h; revamp for
10261 cgraph_function_or_thunk_node.
10262 * lto-symtab.c (lto_varpool_replace_node): Update.
10263 (lto_symtab_resolve_can_prevail_p): Update.
10264 (lto_symtab_merge_cgraph_nodes): Update.
10265 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10266 propagate_constants_accross_call, devirtualization_time_bonus,
10267 ipcp_propagate_stage): Update.
10268 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB
10269 * ipa-inline-transform.c (clone_inlined_nodes,
10270 preserve_function_body_p): Update.
e70670cf
JH
10271 * ipa-reference.c (propagate): Update.
10272 (write_node_summary_p): Update.
10273 * toplev.c (wrapup_global_declaration_2): Update.
10274 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10275 (analyze_function) ... this one.
10276 (cgraph_process_new_functions): Update.
10277 (cgraph_reset_node): Export.
10278 (cgraph_finalize_function): Update.
10279 (cgraph_add_new_function): Update.
10280 (process_function_and_variable_attributes): Update.
10281 (varpool_finalize_decl): Update.
10282 (symbol_finalized): Remove.
10283 (symbol_finalized_and_needed): Rename to ...
10284 (symbol_defined_and_needed): ... update.
10285 (cgraph_analyze_functions): Update.
10286 (handle_alias_pairs): Update.
10287 (mark_functions_to_output): Update.
10288 (assemble_thunk): Update.
10289 (output_in_order): Update.
10290 (output_weakrefs): Update.
10291 (finalize_compilation_unit): Update.
10292 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10293 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10294 input_node, input_varpool_node): Update.
10295 * dbxout.c (dbxout_expand_expr): Update.
10296 * cgraphclones.c (cgraph_clone_node): Update.
10297 (cgraph_copy_node_for_versioning): Update.
10298 (cgraph_materialize_clone): Update.
10299 (cgraph_materialize_all_clones): Update.
10300 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10301 propagate_pure_const, propagate_nothrow): Update.
10302 * lto-streamer-out.c (lto_output, write_symbol): Update.
10303 * ipa-utils.c (ipa_reverse_postorder): Update.
10304 * ipa-inline.c (can_inline_edge_p): Update.
10305 (update_caller_keys, ipa_inline): Update.
10306 * dwarf2out.c (reference_to_unused,
10307 premark_types_used_by_global_vars_helper): Update.
10308 * tree-eh.c (tree_could_trap_p): Update.
10309 * ipa-split.c (consider_split, execute_split_functions): Update.
10310 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB
10311 has_addr_references_p): Update; move ahead in file for better
10312 readability.
e70670cf
JH
10313 (process_references): Simplify.
10314 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10315 bodies are removed.
10316 (cgraph_comdat_can_be_unshared_p): Make static.
10317 (cgraph_externally_visible_p): Update.
10318 (varpool_externally_visible_p): Update.
10319 (function_and_variable_visibility): Update.
10320 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10321 ipa_tm_mark_force_output_node): Update.
10322 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10323 estimate_edge_devirt_benefit, inline_generate_summary,
10324 inline_write_summary): Update.
10325 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10326 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB
10327 (ipa_print_node_params, ipa_prop_read_section,
10328 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH
10329 * varasm.c (mark_decl_referenced): Update.
10330 (assemble_alias, dump_tm_clone_pairs): Update.
10331 * tree-inline.c (copy_bb): Update.
10332 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
10333 Update.
10334 * symtab.c (dump_symtab_base): Print new flags.
10335 (verify_symtab_base): Verify new flags.
10336 (symtab_alias_ultimate_target): New function.
10337 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
10338 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
10339 Update.
c1e183a9
UB
10340 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
10341 Update.
e70670cf
JH
10342 * i386.c (ix86_get_function_versions_dispatcher,
10343 ix86_generate_version_dispatcher_body): Update.
10344 (fold_builtin_cpu): Use varpool_add_new_variable.
10345 * varpool.c (varpool_remove_initializer): Break out from ...
10346 (varpool_remove_node): ... this one.
10347 (dump_varpool_node, varpool_node_for_asm,
10348 cgraph_variable_initializer_availability, varpool_analyze_node,
10349 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB
10350 varpool_finalize_named_section_flags, varpool_create_variable_alias):
10351 Update.
e70670cf 10352
182802ad
JH
103532013-05-29 Jan Hubicka <jh@suse.cz>
10354
10355 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
10356
8190b609
ER
103572013-05-29 Easwaran Raman <eraman@google.com>
10358
10359 PR tree-optimization/57442
10360 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
10361 when control exits the main loop.
10362
69f5aa9b
SKS
103632013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
10364
10365 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
10366 and RX600.
c1e183a9 10367 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS
10368 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
10369 * rx/t-rx: Add rx100 under multi library matches option for nofpu
10370 option.
10371
4b847da9
BS
103722013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10373
10374 PR tree-optimization/57441
10375 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
10376 Don't limit size of incr_vec to number of candidates.
10377
4fc43c39
SE
103782013-05-29 Steve Ellcey <sellcey@imgtec.com>
10379
10380 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
10381 and mips16 directories.
c1e183a9 10382 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE
10383 (MULTILIB_DIRNAMES): Ditto.
10384 (MULTILIB_EXCEPTIONS): Add new exceptions.
10385 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
10386 (MULTILIB_DIRNAMES): Ditto.
10387 (MULTILIB_EXCEPTIONS): Add new exceptions.
10388
12211b99 103892012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS
10390 Marcus Shawcroft <marcus.shawcroft@arm.com>
10391
10392 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
10393 SYMBOL_TINY_ABSOLUTE.
10394 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
10395 SYMBOL_TINY_ABSOLUTE.
10396 (aarch64_expand_mov_immediate): Likewise.
10397 (aarch64_classify_symbol): Likewise.
10398 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10399 Permit SYMBOL_TINY_ABSOLUTE.
10400 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10401
12211b99 104022013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS
10403 Marcus Shawcroft <marcus.shawcroft@arm.com>
10404
10405 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10406 Refactor if/switch. Replace gcc_assert with if.
10407
c0186656
GG
104082013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10409
10410 * config/i386/i386.c (initial_ix86_tune_features): Enable
10411 FP Reassociation for AMD bdver1 and bdver2.
10412
d20188f3
MJ
104132013-05-29 Martin Jambor <mjambor@suse.cz>
10414
10415 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10416 and IMAGPART_EXPR do not occur within other handled_components.
10417
292cba13
RB
104182013-05-29 Richard Biener <rguenther@suse.de>
10419
10420 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10421 access on whether the use is in the BB we currently try to
10422 vectorize.
10423 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10424 vectorize to vect_bb_slp_scalar_cost.
10425
6eddf228
RB
104262013-05-29 Richard Biener <rguenther@suse.de>
10427
10428 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10429 computing scalar cost offsetted by stmts that are kept live
10430 by scalar uses.
10431 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10432 for computation of scalar cost.
10433
7df36117
SE
104342013-05-28 Steve Ellcey <sellcey@mips.com>
10435
10436 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10437
36536d79
BI
104382013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10439
10440 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10441 array notation built-in reduction functions.
10442 * doc/passes.texi (Passes): Added documentation about changes done
10443 for Cilk Plus.
10444 * doc/invoke.texi (C Dialect Options): Added documentation about
10445 the -fcilkplus flag.
10446 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10447 (BUILTINS_DEF): Depend on cilkplus.def.
10448 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10449 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10450 * cilkplus.def: New file.
10451
12211b99 104522013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR
10453
10454 PR rtl-optimization/57439
10455 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10456
9daf6dbb
ER
104572013-05-28 Easwaran Raman <eraman@google.com>
10458
10459 PR tree-optimization/57337
10460 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10461 (find_insert_point): Correctly identify the insertion point
10462 when two statements with the same UID is compared.
10463
fbd7e877
RB
104642013-05-28 Richard Biener <rguenther@suse.de>
10465
10466 PR tree-optimization/56787
10467 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10468 from the list of data references.
10469 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10470 clobbers.
10471 (vect_analyze_loop_operations): Likewise.
10472 (vect_transform_loop): Remove clobbers.
10473
bbba1117
MJ
104742013-05-28 Martin Jambor <mjambor@suse.cz>
10475
10476 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10477 and REALPART_EXPRs have scalar type.
10478
bd388c2a
RB
104792013-05-28 Richard Biener <rguenther@suse.de>
10480
10481 PR tree-optimization/57411
10482 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10483 virtual operands.
10484 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10485 virtual operand propagation.
10486
2f56a311
EB
104872013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10488
10489 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10490 destination register for bmasksi_vis.
10491 (vector_init_bshuffle): Likewise.
10492 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10493
5c3eacbb
EB
104942013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10495
10496 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10497 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10498 mode if the instruction isn't available in the original mode.
10499 * config/sparc/sparc.opt (mfix-ut699): New option.
10500 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10501 (divdf3): Turn into expander.
10502 (divdf3_nofix): New insn.
10503 (divdf3_fix): Likewise.
10504 (divsf3): Disable if -mfix-ut699.
10505 (sqrtdf2): Turn into expander.
10506 (sqrtdf2_nofix): New insn.
10507 (sqrtdf2_fix): Likewise.
10508 (sqrtsf2): Disable if -mfix-ut699.
10509
a1756c0a
RB
105102013-05-27 Richard Biener <rguenther@suse.de>
10511
10512 PR middle-end/57412
10513 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10514 block for the new loop.
10515
5a892248
RB
105162013-05-27 Richard Biener <rguenther@suse.de>
10517
10518 PR tree-optimization/57343
10519 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10520 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10521 (number_of_iterations_cond): Do not build the folded tree.
10522
d1de852b
RB
105232013-05-27 Richard Biener <rguenther@suse.de>
10524
10525 Revert
10526 PR middle-end/57381
10527 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10528 OEP_CONSTANT_ADDRESS_OF retained.
10529
10530 PR tree-optimization/57417
10531 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10532 for unchanged base.
10533 (set_ssa_val_to): Compare addresses using
10534 get_addr_base_and_unit_offset.
10535
12211b99 105362013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR
10537
10538 PR rtl-optimization/56833
10539 * postreload.c (move2add_record_mode): New function.
10540 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10541 (move2add_use_add2_insn): Use move2add_record_sym_value.
10542 (move2add_use_add3_insn): Likewise.
10543 (reload_cse_move2add): Use move2add_valid_value_p and
10544 move2add_record_mode. Invalidate call-clobbered and REG_INC
10545 affected regs by setting reg_mode to VOIDmode.
10546 (move2add_note_store): Don't pretend the inside of a SUBREG is
10547 the actual destination. Invalidate single/leading registers by
10548 setting reg_mode to VOIDmode.
10549 Use move2add_record_sym_value, move2add_valid_value_p and
10550 move2add_record_mode.
10551
b03be25f
RB
105522013-05-27 Richard Biener <rguenther@suse.de>
10553
10554 PR tree-optimization/57396
10555 * tree-affine.c (double_int_constant_multiple_p): Properly
10556 return false for val == 0 and div != 0.
10557
44e88db2
RS
105582013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10559
10560 * config/mips/mips.h: Use #elif in preprocessor conditions.
10561
3b859704
RS
105622013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10563
10564 PR target/53916
10565 * config/mips/constraints.md (kl): New constraint.
10566 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10567 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10568 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10569 from LO for MIPS16.
10570 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10571
c979d5f5
RS
105722013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10573
10574 PR target/55777
10575 * config/mips/mips.c (mips_can_inline_p): New function.
10576 (TARGET_CAN_INLINE_P): Define.
10577
8e90de43
SB
105782013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10579
10580 * sched-int.h (ds_t, dw_t): Make unsigned int.
10581 Fix documentation that describes how all the ds_t bits are used.
10582 Reserve the last bit for delayed-branch scheduling.
10583 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10584 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10585 (gen_dep_weak_1): Remove prototype.
10586 * sched-deps.c (get_dep_weak_1): Make static.
10587 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10588 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10589 * doc/tm.texi: Regenerate.
10590 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
10591
cb5cb194
SB
105922013-05-24 Steven Bosscher <steven@gcc.gnu.org>
10593
10594 PR debug/56950
10595 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
10596
5bd93ff6
NS
105972013-05-24 Nathan Sidwell <nathan@codesourcery.com>
10598 Sandra Loosemore <sandra@codesourcery.com>
10599
10600 * config.gcc (powerpc-*): Allow native for with-cpu.
10601
2343af65
JL
106022013-05-24 Jeff Law <law@redhat.com>
10603
10604 PR tree-optimization/57124
10605 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
10606 conversion feeding a condition if the range has an overflow
10607 if -fstrict-overflow. Add warnings for when we do make the
10608 transformation.
10609
3ad695b9
DC
106102013-05-24 Dehao Chen <dehao@google.com>
10611
04960246 10612 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC
10613 (locus_discrim_hasher): Likewise.
10614 (locus_discrim_hasher::hash): Likewise.
10615 (locus_discrim_hasher::equal): Likewise.
10616
82338059
MJ
106172013-05-24 Martin Jambor <mjambor@suse.cz>
10618
10619 PR tree-optimization/57294
10620 * cgraph.h (ipa_record_stmt_references): Declare.
10621 * cgraphbuild.c (ipa_record_stmt_references): New function.
10622 (build_cgraph_edges): Use ipa_record_stmt_references.
10623 (rebuild_cgraph_edges): Likewise.
10624 (cgraph_rebuild_references): Likewise.
10625 * ipa-prop.c (ipa_modify_call_arguments): Discard references
10626 associated with the old statement and build references from the
10627 newly built statements.
10628 * ipa-ref.c (ipa_remove_stmt_references): New function.
10629 * ipa-ref.h (ipa_remove_stmt_references): Declare.
10630
1ccd4874
VM
106312013-05-24 Vladimir Makarov <vmakarov@redhat.com>
10632
55805e54 10633 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 10634 mem-mem moves.
55805e54 10635 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 10636 too.
55805e54 10637 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM
10638 secondary memory moves.
10639 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
10640 reg set up in the current insn.
10641
25e25c73
DC
106422013-05-24 Dehao Chen <dehao@google.com>
10643
04960246 10644 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC
10645 hash function.
10646 (locus_descrim_hasher::equal): Likewise.
04960246 10647 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC
10648 (make_edges): Likewise.
10649 (next_discriminator_for_locus): Likewise.
10650 (same_line_p): Likewise.
10651 (assign_discriminators): Likewise.
10652 (make_cond_expr_edges): Likewise.
10653 (make_gimple_switch_edges): Likewise.
10654 (make_goto_expr_edges): Likewise.
10655 (make_gimple_asm_edges): Likewise.
10656
50d38551
IB
106572013-05-24 Ian Bolton <ian.bolton@arm.com>
10658
10659 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
10660 X format specifier to only display bottom 16 bits.
10661 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
10662 immediate to match for operand 2, since it will be masked.
10663
aea0101d
RB
106642013-05-24 Richard Biener <rguenther@suse.de>
10665
10666 PR tree-optimization/57287
10667 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
10668 all SSA names that occur in abnormal PHIs.
10669
634e03d3
AI
106702013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
10671
10672 PR tree-ssa/57385
10673 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
10674 that index is not negative.
10675
b17c024f
EB
106762013-05-24 Eric Botcazou <ebotcazou@adacore.com>
10677
10678 PR rtl-optimization/55177
10679 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
10680 (simplify_byte_swapping_operation): New.
10681 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
10682 (simplify_relational_operation_1): Deal with BSWAP.
10683
46aeac1b
RH
106842013-05-23 Richard Henderson <rth@redhat.com>
10685
10686 PR target/56742
10687 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
10688 (ix86_reorg): Call it.
10689
70cc1536
UB
106902013-05-23 Uros Bizjak <ubizjak@gmail.com>
10691
10692 PR target/57379
10693 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
10694 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
10695 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
10696
b3851501
CB
106972013-05-23 Christian Bruel <christian.bruel@st.com>
10698
10699 PR debug/57351
10700 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
10701
12211b99 107022013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS
10703 Marcus Shawcroft <marcus.shawcroft@arm.com>
10704
10705 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
10706 * config/aarch64/constraints.md (Usa): Remove.
10707 * doc/md.texi (AArch64 Usa): Remove.
10708
12211b99 107092013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS
10710 Marcus Shawcroft <marcus.shawcroft@arm.com>
10711
10712 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
10713 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
10714 * config/aarch64/predicates.md (aarch64_const_address): Remove.
10715 (aarch64_mov_operand): Use aarch64_mov_operand_p.
10716
12211b99 107172013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP
10718
10719 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
10720 instruction (AdvSIMD).
10721 * config/aarch64/aarch64-builtins.c
10722 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
10723 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
10724
76c36cb1 107252013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ
10726
10727 PR middle-end/57347
10728 * tree.h (contains_bitfld_component_ref_p): Declare.
10729 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB
10730 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
10731 caller.
7d2fb524
MJ
10732 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
10733 not access a bit-field. Assert all final offsets are byte-aligned.
10734
ce521ae6
RB
107352013-05-23 Richard Biener <rguenther@suse.de>
10736
10737 PR tree-optimization/57380
10738 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
10739 least one invariant or re-used load.
10740 * passes.c (init_optimization_passes): Move pass_phiprop before
10741 pass_forwprop.
10742
75c7257f
JG
107432013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
10744
10745 * config/aarch64/aarch64-simd.md
10746 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
10747
af6d44b5
RB
107482013-05-23 Richard Biener <rguenther@suse.de>
10749
10750 PR middle-end/57381
10751 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10752 OEP_CONSTANT_ADDRESS_OF retained.
10753
bd3647bf
JJ
107542013-05-23 Jakub Jelinek <jakub@redhat.com>
10755
10756 PR middle-end/57344
70cc1536
UB
10757 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
10758 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 10759
a55757ea
RB
107602013-05-23 Richard Biener <rguenther@suse.de>
10761
10762 PR rtl-optimization/57341
10763 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
10764 instead of true_dependence.
10765
cfb00b41
DM
107662013-05-22 David Malcolm <dmalcolm@redhat.com>
10767
dd1c676f
DM
10768 * bb-reorder.c (branch_threshold): Make const.
10769 (exec_threshold): Ditto.
cfb00b41 10770
f62511da 107712013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
10772 Pat Haugen <pthaugen@us.ibm.com>
10773 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
10774
10775 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
10776 documentation for the power8 crypto builtins.
10777
10778 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
10779
10780 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
10781 macros for defining power8 builtin functions.
10782 (BU_P8V_AV_2): Likewise.
10783 (BU_P8V_AV_P): Likewise.
10784 (BU_P8V_VSX_1): Likewise.
10785 (BU_P8V_OVERLOAD_1): Likewise.
10786 (BU_P8V_OVERLOAD_2): Likewise.
10787 (BU_CRYPTO_1): Likewise.
10788 (BU_CRYPTO_2): Likewise.
10789 (BU_CRYPTO_3): Likewise.
10790 (BU_CRYPTO_OVERLOAD_1): Likewise.
10791 (BU_CRYPTO_OVERLOAD_2): Likewise.
10792 (XSCVSPDP): Fix typo, point to the correct instruction.
10793 (VCIPHER): Add power8 crypto builtins.
10794 (VCIPHERLAST): Likewise.
10795 (VNCIPHER): Likewise.
10796 (VNCIPHERLAST): Likewise.
10797 (VPMSUMB): Likewise.
10798 (VPMSUMH): Likewise.
10799 (VPMSUMW): Likewise.
10800 (VPERMXOR_V2DI): Likewise.
10801 (VPERMXOR_V4SI: Likewise.
10802 (VPERMXOR_V8HI: Likewise.
10803 (VPERMXOR_V16QI: Likewise.
10804 (VSHASIGMAW): Likewise.
10805 (VSHASIGMAD): Likewise.
10806 (VPMSUM): Likewise.
10807 (VPERMXOR): Likewise.
10808 (VSHASIGMA): Likewise.
10809
10810 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
10811 __CRYPTO__ if the crypto instructions are available.
10812 (altivec_overloaded_builtins): Add support for overloaded power8
10813 builtins.
10814
10815 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
10816 support for power8 crypto builtins.
10817 (builtin_function_type): Likewise.
10818 (altivec_init_builtins): Add support for builtins that take vector
10819 long long (V2DI) arguments.
10820
10821 * config/rs6000/crypto.md: New file, define power8 crypto
10822 instructions.
10823
108242013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
10825 Pat Haugen <pthaugen@us.ibm.com>
10826 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
10827
10828 * doc/invoke.texi (Option Summary): Add power8 options.
10829 (RS/6000 and PowerPC Options): Likewise.
10830
10831 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
10832 constraints.md instead of rs6000.h. Reorder w* constraints. Add
10833 wm, wn, wr documentation.
10834
f43245d1 10835 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM
10836 registers if direct move instructions are enabled.
10837 (wn): New constraint for no registers.
10838 (wq): New constraint for quad word even GPR registers.
10839 (wr): New constraint if 64-bit instructions are enabled.
10840 (wv): New constraint if power8 vector instructions are enabled.
10841 (wQ): New constraint for quad word memory locations.
10842
f43245d1 10843 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM
10844 constraint for 0..15 for crypto instructions.
10845 (gpc_reg_operand): If VSX allow registers in VSX registers as well
10846 as GPR and floating point registers.
10847 (int_reg_operand): New predicate to match only GPR registers.
10848 (base_reg_operand): New predicate to match base registers.
10849 (quad_int_reg_operand): New predicate to match even GPR registers
10850 for quad memory operations.
10851 (vsx_reg_or_cint_operand): New predicate to allow vector logical
10852 operations in both GPR and VSX registers.
10853 (quad_memory_operand): New predicate for quad memory operations.
10854 (reg_or_indexed_operand): New predicate for direct move support.
10855
f43245d1 10856 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM
10857 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
10858 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
10859 (POWERPC_MASKS): Add power8 options.
10860 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
10861 various options.
10862
f43245d1 10863 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM
10864 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
10865
f43245d1 10866 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM
10867 (-mpower8-fusion): New power8 options.
10868 (-mpower8-fusion-sign): Likewise.
10869 (-mpower8-vector): Likewise.
10870 (-mcrypto): Likewise.
10871 (-mdirect-move): Likewise.
10872 (-mquad-memory): Likewise.
10873
f43245d1 10874 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM
10875 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
10876 registers.
70cc1536 10877 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM
10878 (rs6000_debug_vector_unit): Add p8_vector.
10879 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
10880 definitions. Also print fusion state.
10881 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
10882 (rs6000_builtin_mask_calculate): Add power8 builtin support.
10883 (rs6000_option_override_internal): Add support for power8.
10884 (rs6000_common_init_builtins): Add debugging for skipped builtins
10885 if -mdebug=builtin.
10886 (rs6000_adjust_cost): Add power8 support.
10887 (rs6000_issue_rate): Likewise.
10888 (insn_must_be_first_in_group): Likewise.
10889 (insn_must_be_last_in_group): Likewise.
10890 (force_new_group): Likewise.
10891 (rs6000_register_move_cost): Likewise.
10892 (rs6000_opt_masks): Likewise.
10893
10894 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
10895 power8 capable assembler, default to power7 options.
10896 (TARGET_DIRECT_MOVE): Likewise.
10897 (TARGET_CRYPTO): Likewise.
10898 (TARGET_P8_VECTOR): Likewise.
10899 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
10900 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
10901 (VECTOR_MEM_P8_VECTOR_P): Likewise.
10902 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
10903 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
10904 (TARGET_XSCVDPSPN): Likewise.
10905 (TARGET_XSCVSPDPN): Likewsie.
10906 (TARGET_SYNC_HI_QI): Likewise.
10907 (TARGET_SYNC_TI): Likewise.
10908 (MASK_CRYPTO): Likewise.
10909 (MASK_DIRECT_MOVE): Likewise.
10910 (MASK_P8_FUSION): Likewise.
10911 (MASK_P8_VECTOR): Likewise.
70cc1536
UB
10912 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
10913 temporary used by some of the direct move instructions to get two FP
10914 temporary registers does not force creation of a stack frame.
f62511da
MM
10915 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
10916 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
10917 that any VSX registers are tieable, even if they are also an
10918 Altivec vector mode.
10919 (r6000_reg_class_enum): Add wm, wr, wv constraints.
10920 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
10921 (RS6000_BTM_CRYPTO): Likewise.
10922 (RS6000_BTM_COMMON): Likewise.
10923
10924 * config/rs6000/rs6000.md (cpu attribute): Add power8.
10925 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
10926 (enum rs6000_vector): Add power8 vector support.
10927
73a1a707
RR
109282013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10929
10930 PR target/19599
10931 PR target/57340
10932 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
10933 (any_sibcall_could_use_r3): this and handle indirect calls.
10934 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
10935
d28073d4
BS
109362013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10937
10938 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
10939
bee0b10c
RB
109402013-05-22 Richard Biener <rguenther@suse.de>
10941
10942 PR middle-end/57349
10943 * profile.c (branch_prob): Do not split blocks that are
10944 abnormally receiving from ECF_RETURNS_TWICE functions.
10945
98409b51
RS
109462013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10947
10948 * recog.c (offsettable_address_addr_space_p): Fix calculation of
10949 address mode. Move pointer mode initialization to the same place.
10950
c0602ab8
MZ
109512013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10952
10953 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
10954 while it has any effect.
10955
4f7a634e
ER
109562013-05-21 Easwaran Raman <eraman@google.com>
10957
10958 PR tree-optimization/57322
9daf6dbb
ER
10959 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
10960 UID of the statement added to the BB to be 1.
4f7a634e 10961
a4ce1258
JJ
109622013-05-21 Jakub Jelinek <jakub@redhat.com>
10963
10964 PR tree-optimization/57331
70cc1536
UB
10965 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
10966 of conversion from pointer type to integral type with integer.
a4ce1258 10967
1b14621a
MJ
109682013-05-21 Martin Jambor <mjambor@suse.cz>
10969
10970 PR lto/57289
10971 * ipa-prop.c (ipa_read_node_info): Process param_used and
10972 controlled_uses in the same order as when writing.
10973
e60661f0
MG
109742013-05-21 Magnus Granberg <baldrick@free.fr>
10975
10976 PR plugins/56754
ee49aa34 10977 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 10978
4a61be9a
RB
109792013-05-21 Richard Biener <rguenther@suse.de>
10980
10981 PR tree-optimization/57318
10982 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
10983 estimate stmts with side-effects as likely eliminated.
10984
c52da5f7
RB
109852013-05-21 Richard Biener <rguenther@suse.de>
10986
10987 PR tree-optimization/57330
10988 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
10989 preserve the call stmts fntype.
10990
7ec67e2a
RB
109912013-05-21 Richard Biener <rguenther@suse.de>
10992
10993 PR tree-optimization/57303
10994 * tree-ssa-sink.c (statement_sink_location): Improve killing
10995 stmt detection and properly handle self-assignments.
10996
b112d513
CB
109972013-05-21 Christian Bruel <christian.bruel@st.com>
10998
55805e54
YR
10999 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
11000 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB
11001 registers. Set register size out of the PARALLEL loop.
11002
14c2ec26
OE
110032013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
11004
11005 PR target/56547
11006 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
11007 (*fmasf4, *fmasf4_media): New insns.
11008
da734fa1
RS
110092013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11010
11011 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
11012 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
11013 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
11014 (mips_idiv_insns): Update the comments to say that the returned
11015 instruction counts are in units of BASE_INSN_LENGTH.
11016 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
11017 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
11018 using 2 rather than 4 as the length of indirect MIPS16 and
11019 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
11020 length of a NOP. Don't divide MIPS16 lengths by 2.
11021 (mips16_split_long_branches): Assume a branch is long if the
11022 length is greater than 4 rather than 8.
11023 * config/mips/mips.md (length): Give MIPS16 lengths directly,
11024 rather than multiplying them by 2. Multiply instruction counts
11025 by BASE_INSN_LENGTH rather than 4.
11026 (*jump_mips16, tls_get_tp_mips16_<mode>)
11027 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
11028
13719e8b
RS
110292013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11030
11031 * config/mips/mips.md (extended_mips16): Remove branch case.
11032 (length): Remove duplicated extended_mips16 test.
11033
c3850d14
RS
110342013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11035
11036 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
11037
8da2e059
RS
110382013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
11039
11040 * recog.h (Recog_data): Rename to...
11041 (recog_data_d): ...this.
11042 (recog_data): Update accordingly.
11043 * recog.c (recog_data): Likewise.
11044 * reload.c (save_recog_data): Likewise.
11045 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
11046 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
11047
2df013f3
JB
110482013-05-17 Julian Brown <julian@codesourcery.com>
11049
11050 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
11051 found in a REG_EQUAL note, invalidate it.
11052
371e77e3 110532013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER
11054
11055 * tree-ssa-reassoc.c (find_insert_point): New function.
11056 (insert_stmt_after): Likewise.
11057 (get_def_stmt): Likewise.
11058 (ensure_ops_are_available): Likewise.
11059 (not_dominated_by): Likewise.
11060 (rewrite_expr_tree): Do not move statements beyond what is
11061 necessary. Remove call to swap_ops_for_binary_stmt...
11062 (reassociate_bb): ... and move it here.
11063 (build_and_add_sum): Assign UIDs for new statements.
11064 (linearize_expr): Likewise.
11065 (do_reassoc): Renumber gimple statement UIDs.
11066
e01c7cca
JH
110672013-05-17 Jan Hubicka <jh@suse.cz>
11068
11069 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
11070 weakrefs.
11071 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB
11072 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
11073 weakrefs.
e01c7cca
JH
11074 (output_weakrefs): Update.
11075
c3272a92
PCC
110762013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
11077 Martin Jambor <mjambor@suse.cz>
11078
11079 PR middle-end/57276
11080 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
11081 value that corresponds to the given aggval is found in values vector.
11082
11c2aa39
UB
110832013-05-17 Uros Bizjak <ubizjak@gmail.com>
11084
11085 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
11086 sse, sse2, sse3, ssse3 and sse4a flags to options.
11087
abecc8c6
DM
110882013-05-17 David Malcolm <dmalcolm@redhat.com>
11089
11090 * gengtype-state.c: (s_expr_writer): New class, to handle
11091 prettifying of output layout of s-expressions.
11092 (state_writer): New class, to write out gtype.state.
11093 (state_written_type_count): Move this variable into member data of
11094 state_writer.
11095 (s_expr_writer::s_expr_writer): New code: constructor for new class
11096 (state_writer::state_writer(): ditto
11097 (s_expr_writer::write_new_line): New function
11098 (s_expr_writer::write_any_indent): ditto
11099 (s_expr_writer::begin_s_expr): ditto
11100 (s_expr_writer::end_s_expr): ditto
11101 (write_state_fileloc): convert to method of state_writer...
11102 (state_writer:: write_state_fileloc): ...and use methods of
11103 s_expr_writer to write indentation into the gtype.state output file
11104 to visually represent the hierarchical structure of the list
11105 structures
11106 (write_state_fields): ditto, renaming to...
11107 (state_writer::write_state_fields)
11108 (write_state_a_string): ditto, renaming to...
11109 (state_writer::write_state_a_string)
11110 (write_state_string_option): ditto, renaming to...
11111 (state_writer::write_state_string_option)
11112 (write_state_type_option): ditto, renaming to...
11113 (state_writer::write_state_type_option)
11114 (write_state_nested_option): ditto, renaming to...
11115 (state_writer::write_state_nested_option)
11116 (write_state_option): ditto, renaming to...
11117 (state_writer::write_state_option)
11118 (write_state_options): ditto, renaming to...
11119 (state_writer::write_state_options)
11120 (write_state_lang_bitmap): ditto, renaming to...
11121 (state_writer::write_state_lang_bitmap)
11122 (write_state_version): ditto, renaming to...
11123 (state_writer::write_state_version)
11124 (write_state_scalar_type): ditto, renaming to...
11125 (state_writer::write_state_scalar_type)
11126 (write_state_string_type): ditto, renaming to...
11127 (state_writer::write_state_string_type)
11128 (write_state_undefined_type): ditto, renaming to...
11129 (state_writer::write_state_undefined_type)
11130 (write_state_struct_union_type): ditto, renaming to...
11131 (state_writer::write_state_struct_union_type)
11132 (write_state_struct_type): ditto, renaming to...
11133 (state_writer::write_state_struct_type)
11134 (write_state_user_struct_type): ditto, renaming to...
11135 (state_writer::write_state_user_struct_type)
11136 (write_state_lang_struct_type): ditto, renaming to...
11137 (state_writer::write_state_lang_struct_type)
11138 (write_state_param_struct_type): ditto, renaming to...
11139 (state_writer::write_state_param_struct_type)
11140 (write_state_pointer_type): ditto, renaming to...
11141 (state_writer::write_state_pointer_type)
11142 (write_state_array_type): ditto, renaming to...
11143 (state_writer::write_state_array_type)
11144 (write_state_gc_used): ditto, renaming to...
11145 (state_writer::write_state_gc_used)
11146 (write_state_common_type_content): ditto, renaming to...
11147 (state_writer::write_state_common_type_content)
11148 (write_state_type): ditto, renaming to...
11149 (state_writer::write_state_type)
11150 (write_state_pair_list): ditto, renaming to...
11151 (state_writer::write_state_pair_list)
11152 (write_state_pair): ditto, renaming to...
11153 (state_writer::write_state_pair)
11154 (write_state_typedefs): ditto, renaming to...
11155 (state_writer::write_state_typedefs)
11156 (write_state_structures): ditto, renaming to...
11157 (state_writer::write_state_structures)
11158 (write_state_param_structs): ditto, renaming to...
11159 (state_writer::write_state_param_structs)
11160 (write_state_variables): ditto, renaming to...
11161 (state_writer::write_state_variables)
11162 (write_state_srcdir): ditto, renaming to...
11163 (state_writer::write_state_srcdir)
11164 (write_state_files_list): ditto, renaming to...
11165 (state_writer::write_state_files_list)
11166 (write_state_languages): ditto, renaming to...
11167 (state_writer::write_state_languages)
11168 (write_state): create a state_writer instance and use it when
11169 writing out the state file
11170
d6545f29
MS
111712013-05-17 Mike Stump <mikestump@comcast.net>
11172
816a3f73 11173 PR rtl-optimization/57304
d6545f29
MS
11174 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
11175 accessing DF_REF_REAL_LOC.
11176
38c821cf
JJ
111772013-05-17 Jakub Jelinek <jakub@redhat.com>
11178
11179 PR rtl-optimization/57281
11180 PR rtl-optimization/57300
11181 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
11182 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
11183 what the other splitter did if the registers are dead.
11184
2a293391
RB
111852013-05-17 Richard Biener <rguenther@suse.de>
11186
11187 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
11188 MEM_REF offsets.
11189
17042d2b
JJ
111902013-05-17 Jakub Jelinek <jakub@redhat.com>
11191
11192 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
11193 linking.
11194
5b115c1f
MP
111952013-05-17 Marek Polacek <polacek@redhat.com>
11196
11c2aa39
UB
11197 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
11198 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 11199
52d84413
JJ
112002013-05-17 Jakub Jelinek <jakub@redhat.com>
11201
68119618
JJ
11202 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
11203 vect_external_def oprnd1 with loop_vinfo, try to emit
11204 optional cast, negation and and stmts on the loop preheader
11205 edge instead of into the pattern def seq.
11206
52d84413
JJ
11207 PR tree-optimization/57051
11208 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
11209 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
11210
17b962bd
NC
112112013-05-16 Nick Clifton <nickc@redhat.com>
11212
11213 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
11214 (rl78_is_naked_func): New function.
56aefbf7
UB
11215 (rl78_expand_prologue): Skip prologue generation for naked functions.
11216 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC
11217 * doc/extend.texi (naked): Add RL78 to the list of processors
11218 that supports this attribute.
11219
b1a0f84e
JL
112202013-05-16 Jeff Law <law@redhat.com>
11221
11222 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
11223
af0e415b
UB
112242013-05-16 Uros Bizjak <ubizjak@gmail.com>
11225
11226 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
11227 cache parameters using detect_caches_amd also for CYRIX,
11228 NSC and TM2 signatures.
11229
19db293a
UB
112302013-05-16 Uros Bizjak <ubizjak@gmail.com>
11231 Dzianis Kahanovich <mahatma@eu.by>
11232
11233 PR target/45359
11234 PR target/46396
11235 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11236 VIA/Centaur processors and determine their cache parameters
11237 using detect_caches_amd.
11238
251a41b9
TJ
112392013-05-16 Teresa Johnson <tejohnson@google.com>
11240
11241 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11242 (rtl_verify_edges): New function.
11243 (rtl_verify_bb_insns): Ditto.
11244 (rtl_verify_bb_pointers): Ditto.
11245 (rtl_verify_bb_insn_chain): Ditto.
11246 (rtl_verify_fallthru): Ditto.
11247 (rtl_verify_bb_layout): Ditto.
11248 (rtl_verify_flow_info_1): Outline checks into new functions.
11249 (rtl_verify_flow_info): Ditto.
11250
f14540b6
SE
112512013-05-16 Steve Ellcey <sellcey@imgtec.com>
11252
11253 * cfghooks.c (copy_bbs): Add update_dominance argument.
11254 * cfghooks.h (copy_bbs): Update prototype.
11255 * tree-cfg.c (gimple_duplicate_sese_region):
11256 Add update_dominance argument.
11257 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11258 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11259 gimple_duplicate_sese_region call.
11260 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11261 Update copy_bbs call.
11262 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11263 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11264
7e9a3abb
JJ
112652013-05-16 Jakub Jelinek <jakub@redhat.com>
11266
11267 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11268 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11269 vect_recog_rotate_pattern.
11270 (vect_recog_rotate_pattern): New function.
11271
427b248d
JM
112722013-05-16 Jason Merrill <jason@redhat.com>
11273
11274 * Makefile.in (LLINKER): New variable.
11275 (mostlyclean): Remove link mutex.
11276 * configure.ac: Handle --enable-link-mutex.
11277 * lock-and-run.sh: New script.
11278
b871e3d2
RR
112792013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11280
11281 PR target/19599
11282 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11283 for NULL decl.
11284
ce858126
RO
112852013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11286
11287 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11288
798d3d04
GY
112892013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11290
11291 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11292 * config/arm/arm.c (next_consecutive_mem): New function.
11293 (gen_movmem_ldrd_strd): Likewise.
11294 * config/arm/arm.md (movmemqi): Update condition and code.
11295 (unaligned_loaddi, unaligned_storedi): New patterns.
11296
0baddc45
RO
112972013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11298
11299 * config.gcc: Obsolete *-*-solaris2.9*.
11300 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11301
aa9d5bcf
RB
113022013-05-16 Richard Biener <rguenther@suse.de>
11303
11304 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11305 earlier, after GRAPHITE transforms and IV canonicalization.
11306
5a0f4dd3
JJ
113072013-05-16 Jakub Jelinek <jakub@redhat.com>
11308
11309 * omp-low.c (extract_omp_for_data): For collapsed loops,
11310 if at least one of the loops is known at compile time to
11311 iterate zero times, set count to 0.
11312 (expand_omp_regimplify_p): New function.
11313 (expand_omp_for_generic): For collapsed loops, if at least
11314 one of the loops isn't known to iterate at least once,
11315 add runtime check with setting count to 0.
11316 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11317 For unsigned types if it isn't known at compile time that
11318 the loop will iterate at least once, add runtime check to bypass
11319 the whole loop if initial condition isn't true.
11320
e3753785
NS
113212013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11322
11323 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11324
43bb4dd1
MG
113252013-05-16 Marc Glisse <marc.glisse@inria.fr>
11326
11327 PR middle-end/57286
11328 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
11329 transformations to avoid an infinite loop.
11330
3571dde6
MP
113312013-05-16 Marek Polacek <polacek@redhat.com>
11332
11333 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
11334
b8b3f0ca
LE
113352013-05-15 Leif Ekblad <leif@rdos.net>
11336
11337 * config/i386/i386.c (ix86_decompose_address): Use
11338 DEFAULT_TLS_SEG_REG to access TLS segment register.
11339 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
11340 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
11341 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
11342
e299a383
RS
113432013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
11344
11345 PR target/57260
11346 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
11347 sibling calls to functions that would normally be lazily bound,
11348 unless $gp is call-clobbered.
11349
e7208ea3
UB
113502013-05-15 Uros Bizjak <ubizjak@gmail.com>
11351
19db293a 11352 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 11353 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB
11354 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
11355 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB
11356 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
11357 of TARGET_3DNOW.
11358 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
11359
e7413f3d
AS
113602013-05-15 Andreas Schwab <schwab@suse.de>
11361
11362 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
11363 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
11364 third operand.
11365
ad4db775
TJ
113662013-05-15 Teresa Johnson <tejohnson@google.com>
11367
11368 * loop-unroll.c (report_unroll_peel): Check decision before
11369 emitting unroll/peel message.
11370
af205f67
TJ
113712013-05-15 Teresa Johnson <tejohnson@google.com>
11372
11373 * function.h (has_bb_partition): New rtl_data flag.
11374 (bb_reorder_complete): Ditto.
11375 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
11376 instead of flag_reorder_blocks_and_partition.
11377 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
11378 with some enhancements.
11379 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
11380 * bb-reorder.c (connect_traces): Check for has_bb_partition
11381 instead of flag_reorder_blocks_and_partition.
11382 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
11383 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
11384 verify_hot_cold_block_grouping.
11385 (partition_hot_cold_basic_blocks): Set has_bb_partition.
11386
9adcfa3c
RR
113872013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11388
11389 PR target/19599
11390 * config/arm/predicates.md (call_insn_operand): New predicate.
11391 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
11392 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
11393 if insn is not a tail call.
11394 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11395 registers.
11396 * config/arm/arm.h (enum reg_class): New caller save register class.
11397 (REG_CLASS_NAMES): Likewise.
11398 (REG_CLASS_CONTENTS): Likewise.
11399 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11400 without decls.
11401
ccb3ad87
RB
114022013-05-15 Richard Biener <rguenther@suse.de>
11403
11404 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11405 of MSG_OPTIMIZED_LOCATIONS.
11406 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11407 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11408 message.
11409 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11410 of MSG_OPTIMIZED_LOCATIONS.
11411 (execute_vect_slp): Likewise.
11412 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11413 (vect_create_cond_for_alias_checks): Likewise.
11414 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11415 (vect_recog_widen_mult_pattern): Likewise.
11416 (vect_recog_widen_sum_pattern): Likewise.
11417 (vect_recog_over_widening_pattern): Likewise.
11418 (vect_recog_widen_shift_pattern): Likewise.
11419 (vect_recog_vector_vector_shift_pattern): Likewise.
11420 (vect_recog_divmod_pattern): Likewise.
11421 (vect_recog_mixed_size_cond_pattern): Likewise.
11422 (vect_recog_bool_pattern): Likewise.
11423 (vect_pattern_recog_1): Likewise.
11424
48b1474e
MJ
114252013-05-15 Martin Jambor <mjambor@suse.cz>
11426
11427 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11428 non-functions to builtin_unreachable.
11429 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11430 correct when new direct edges were discovered.
11431
9de04252
MJ
114322013-05-15 Martin Jambor <mjambor@suse.cz>
11433
11434 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11435 header, print symbol order instead of node uid, print more information
11436 about indirect edge targets.
11437 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11438 uids.
11439 (ipa_make_edge_direct_to_target): Likewise.
11440 (remove_described_reference): Likewise.
11441 (propagate_controlled_uses): Likewise.
11442 (ipa_print_node_params): Also print symbol order.
11443 (ipcp_transform_function): Print symbol order instead of node uids.
11444 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11445 (cgraph_get_create_real_symbol_node): Likewise.
11446 * ipa-cp.c (print_lattice): Likewise.
11447 (print_all_lattices): Likewise.
11448 (determine_versionability): Likewise.
11449 (initialize_node_lattices): Likewise.
11450 (estimate_local_effects): Likewise.
11451 (update_profiling_info): Likewise.
11452 (create_specialized_node): Likewise.
11453 (perhaps_add_new_callers): Likewise.
11454 (decide_about_value): Likewise.
11455 (decide_whether_version_node): Likewise.
11456 (identify_dead_nodes): Likewise.
11457 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11458 (dump_inline_summary): Likewise.
11459 (estimate_node_size_and_time): Likewise.
11460 (inline_analyze_function): Likewise.
11461 * ipa-inline.c (report_inline_failed_reason): Likewise.
11462 (want_early_inline_function_p): Likewise.
11463 (edge_badness): Likewise.
11464 (update_edge_key): Likewise.
11465 (inline_small_functions): Likewise. Add dumping of order to two other
11466 dumps.
11467 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11468 instead of node uids.
11469 (propagate_pure_const): Likewise.
11470 (propagate_pure_const): Likewise.
11471 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11472 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11473 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11474 of node uids.
11475 * tree-pretty-print.c (dump_function_header): Likewise.
11476 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11477 Print symbol order instead of node uids.
11478
1dd03b91
AK
114792013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11480
11481 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11482 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11483
fdf6a7b9
RB
114842013-05-15 Richard Biener <rguenther@suse.de>
11485
11486 PR tree-optimization/57275
11487 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11488 return value for fail to do runtime alias checks for gather loads.
11489
2d6e4603
JH
114902013-05-15 Jan Hubicka <jh@suse.cz>
11491
11492 PR lto/57038
11493 PR lto/47375
e7208ea3
UB
11494 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11495 weakrefs are not external.
11496 (lto_symtab_merge_decls): Fix thinko when dealing with
11497 non-lto_symtab decls.
2d6e4603
JH
11498 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11499 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11500 * varpool.c (dump_varpool_node): Dump more flags.
11501
83f44b39
GG
115022013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11503
11504 * config/i386/i386.c (processor_alias_table): Add instruction
11505 FSGSBASE for AMD bdver3 architecture.
11506
2e55d062
JJ
115072013-05-14 Jakub Jelinek <jakub@redhat.com>
11508
11509 * tree.c (warn_deprecated_use): Print file:line using locus color.
11510 * diagnostic.c (diagnostic_report_current_module): Print file:line
11511 and file:line:column using locus color.
11512
0bfdb81e
MS
115132013-05-14 Mike Stump <mikestump@comcast.net>
11514
11515 * gdbinit.in: Add __null.
11516
a508ef22
MS
115172013-05-14 Mike Stump <mikestump@comcast.net>
11518
11519 * recog.h: Rename struct recog_data to Recog_data.
11520 * recog.c: Likewise.
11521 * reload.c (can_reload_into): Likewise.
11522 * config/picochip/picochip.c: Likewise.
11523
e7180acb
MS
115242013-05-14 Mike Stump <mikestump@comcast.net>
11525
11526 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11527
61216c26
SB
115282013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11529
fd6beed4
SB
11530 * resource.h (struct resources): Remove unch_memory member.
11531 (CLEAR_RESOURCE): Don't clear unch_memory.
11532 * resource.c (mark_referenced_resources): Don't set it.
11533 (mark_set_resources): Likewise.
11534 (mark_target_live_regs): Don't clear it.
11535 (init_resource_info): Likewise.
11536 * reorg.c (resource_conflicts_p): Don't compare it.
11537 (redundant_insn): Don't set it.
11538
61216c26
SB
11539 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11540 Remove prototypes.
11541 * emit-rtl.c (next_label): Remove unused function.
11542 (skip_consecutive_labels, link_cc0_insns): Move to ...
11543 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11544 only place where these functions are used, and make them static.
11545
418d1b87
MG
115462013-05-14 Marc Glisse <marc.glisse@inria.fr>
11547
11548 * fold-const.c (fold_negate_expr): Handle vectors.
11549 (fold_truth_not_expr): Make it static.
11550 (fold_invert_truthvalue): New static function.
11551 (invert_truthvalue_loc): Handle vectors. Do not call
11552 fold_truth_not_expr directly.
11553 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11554 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11555 (fold_comparison): Handle vector constants.
11556 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11557 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11558 * tree.h (fold_truth_not_expr): Remove declaration.
11559
fc21784d
JG
115602013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11561
11562 * config/aarch64/aarch64-simd.md
11563 (aarch64_vcond_internal<mode>): Rename to...
11564 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11565 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11566 float modes. Clarify all iterator modes.
11567 (vcond<mode><mode>): Use new name for vcond expanders.
11568 (vcond<v_cmp_result><mode>): Likewise.
11569 (vcondu<mode><mode>: Likewise.
11570 * config/aarch64/iterators.md (VDQF_COND): New.
11571
d4c52634
MG
115722013-05-14 Marc Glisse <marc.glisse@inria.fr>
11573
11574 PR bootstrap/57266
11575 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11576 variable for the shift amount. Check that we shift by non-negative
11577 amounts.
11578
2b261262
CLT
115792013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11580
11581 PR target/42017
11582 * config/arm/arm.h (EPILOGUE_USES): Only return true
11583 for LR_REGNUM after epilogue_completed.
11584
12211b99 115852013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR
11586
11587 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 11588 is error_mark_node.
6e022d7b 11589
56cf7859
RO
115902013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11591
11592 PR target/57261
11593 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
11594 and Solaris 11+/x86 with gld.
11595 * configure: Regenerate.
11596
75776c6d
JJ
115972013-05-14 Jakub Jelinek <jakub@redhat.com>
11598
11599 * expmed.c (expand_shift_1): Canonicalize rotates by
11600 constant bitsize / 2 to bitsize - 1.
7f998021 11601 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ
11602 case ROTATERT>: Likewise.
11603
11604 Revert:
11605 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11606
11607 * config/i386/i386.md (rotateinv): New code attr.
11608 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11609 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11610 roll $31, %eax, etc.
11611
df35498a
RB
116122013-05-14 Richard Biener <rguenther@suse.de>
11613
11614 PR middle-end/57235
11615 * tree-eh.c (sink_clobbers): Give up for successors with
11616 multiple predecessors and no virtual uses.
11617
cc6e7ece
EB
116182013-05-14 Eric Botcazou <ebotcazou@adacore.com>
11619
11620 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
11621 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
11622
f778dd4d
JJ
116232013-05-14 Jakub Jelinek <jakub@redhat.com>
11624
11625 PR middle-end/57251
11626 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
11627 the case when both op0 and op1 have VOIDmode.
11628
bad4df9b
KP
116292013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11630
11631 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
11632 in multiply-accumulate mode.
11633
56f3e9ac
GW
116342013-05-13 Guozhi Wei <carrot@google.com>
11635
11636 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
11637
a3d7ab92
KT
116382013-05-13 Kai Tietz <ktietz@redhat.com>
11639
11640 PR target/56975
11641 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 11642 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 11643 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 11644 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT
11645 (ix86_expand_prologue): Likewise.
11646 (ix86_expand_split_stack_prologue): Likewise.
11647 (legitimate_pic_address_disp_p): Likewise.
11648 (legitimize_pic_address): Likewise.
11649 (legitimize_tls_address): Likewise.
11650 (legitimize_pe_coff_symbol): Likewise.
11651 (output_pic_addr_const): Likewise.
11652 (construct_plt_address): Likewise.
11653 (ix86_expand_call): Likewise.
11654 (x86_output_mi_thunk): Likewise.
11655 (x86_function_profiler): Likewise.
11656
c59b7e28
SN
116572013-05-13 Sofiane Naci <sofiane.naci@arm.com>
11658
11659 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
11660 similar switch cases.
11661 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
11662 (aarch64_simd_mov_to_<mode>low): Delete.
11663 (aarch64_simd_mov_to_<mode>high): Delete.
11664 (move_lo_quad_<mode>): Add w<-r alternative.
11665 (aarch64_simd_move_hi_quad_<mode>): Likewise.
11666 (aarch64_simd_mov_from_*): Update type attribute.
11667 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
11668 statement.
11669
5f28524a
JH
116702013-05-13 Jan Hubicka <jh@suse.cz>
11671
11672 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
11673 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB
11674 ix86_expand_epilogue, emit_i387_cw_initialization,
11675 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
11676 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 11677
ae6fa899
JJ
116782013-05-13 Jakub Jelinek <jakub@redhat.com>
11679
11680 PR tree-optimization/45216
11681 PR tree-optimization/57157
11682 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
11683 the (-Y) & (B - 1) variant if OP is |.
11684 * expmed.c (expand_shift_1): For rotations by const0_rtx just
11685 return shifted. Use (-op1) & (prec - 1) as other_amount
11686 instead of prec - op1.
11687
4502fe8d
MJ
116882013-05-13 Martin Jambor <mjambor@suse.cz>
11689
11690 PR middle-end/42371
11691 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
11692 (ipa_constant_data): New type.
11693 (ipa_jump_func): Use ipa_constant_data to hold information about
11694 constant jump functions.
11695 (ipa_get_jf_constant): Adjust to jump function type changes.
11696 (ipa_get_jf_constant_rdesc): New function.
11697 (ipa_param_descriptor): New field controlled_uses.
11698 (ipa_get_controlled_uses): New function.
11699 (ipa_set_controlled_uses): Likewise.
11700 * ipa-ref.h (ipa_find_reference): Declare.
11701 * ipa-prop.c (ipa_cst_ref_desc): New type.
11702 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
11703 changes.
11704 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
11705 New parameter cs. Adjust all callers.
11706 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
11707 (remove_described_reference): New function.
11708 (jfunc_rdesc_usable): Likewise.
11709 (try_make_edge_direct_simple_call): Decrement controlled use count,
11710 attempt to remove reference if it hits zero.
11711 (combine_controlled_uses_counters): New function.
11712 (propagate_controlled_uses): Likewise.
11713 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
11714 (ipa_edge_duplication_hook): Duplicate reference descriptions.
11715 (ipa_print_node_params): Print described use counter.
11716 (ipa_write_jump_function): Adjust to jump function type changes.
11717 (ipa_read_jump_function): New parameter CS, pass it to
11718 ipa_set_jf_constant. Adjust caller.
11719 (ipa_write_node_info): Stream controlled use count
11720 (ipa_read_node_info): Likewise.
11721 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
11722 asserting.
11723 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
11724 count. Remove cloning-added reference if it reaches zero.
11725 * ipa-ref.c (ipa_find_reference): New function.
11726
0864bfc2
GG
117272013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11728
e7208ea3 11729 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG
11730 alignment values for AMD BD and BT architectures.
11731
640bfeb2
MG
117322013-05-13 Marc Glisse <marc.glisse@inria.fr>
11733
11734 * tree-vect-generic.c (uniform_vector_p): Move ...
11735 * tree.c (uniform_vector_p): ... here.
11736 * tree.h (uniform_vector_p): Declare it.
11737 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
11738 into a scalar.
11739
3a60f32b
JJ
117402013-05-13 Jakub Jelinek <jakub@redhat.com>
11741
198fe1bf
JJ
11742 PR tree-optimization/57230
11743 * tree-ssa-strlen.c (handle_char_store): Record length for
11744 array store from STRING_CST.
11745
3a60f32b
JJ
11746 PR tree-optimization/57230
11747 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
11748 check.
11749
566be57c
JR
117502013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
11751
11752 * config/epiphany/epiphany.c (epiphany_init): Check size of
11753 NUM_MODES_FOR_MODE_SWITCHING.
11754 (epiphany_expand_prologue):
11755 Remove CONFIG_REGNUM initial value handling code.
11756 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
11757 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 11758 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR
11759 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
11760 Don't return 1 for FP_MODE_NONE.
11761 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
11762 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
11763 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
11764 * config/epiphany/epiphany.md (save_config): New pattern.
11765
0f2c2331
UB
117662013-05-12 Uros Bizjak <ubizjak@gmail.com>
11767
11768 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
11769
5b3f93c7
UB
117702013-05-10 Uros Bizjak <ubizjak@gmail.com>
11771
11772 * config/i386/i386.md (memory): Handle sseishft1.
11773 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
11774 (*vec_extractv2di_1): Ditto.
11775
1f873f0e
VM
117762013-05-10 Vladimir Makarov <vmakarov@redhat.com>
11777
11778 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
11779 saved registers.
11780
341427fa 117812013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH
11782
11783 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
11784 Add mthumb/march=armv7-a multilib.
11785 Add mthumb/march=armv7-r multilib.
11786 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
11787
9e69bdde
RC
117882013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
11789
11790 * config/v850/t-rtems: Add more multilibs.
11791
9ff09a22
RB
117922013-05-10 Richard Biener <rguenther@suse.de>
11793
11794 PR tree-optimization/57214
11795 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
11796 not propagate from SSA names that occur in abnormal PHI nodes.
11797
a5e0cd1d
MG
117982013-05-10 Marc Glisse <marc.glisse@inria.fr>
11799
11800 * stor-layout.c (element_precision): New function.
11801 * machmode.h (element_precision): Declare it.
11802 * tree.c (build_minus_one_cst): New function.
11803 (element_precision): Likewise.
11804 * tree.h (build_minus_one_cst): Declare new function.
11805 (element_precision): Likewise.
11806 * fold-const.c (operand_equal_p): Use element_precision.
11807 (fold_binary_loc): Handle vector types.
11808 * convert.c (convert_to_integer): Use element_precision.
11809 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
11810 separately.
11811
cb2558bc
RS
118122013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
11813
11814 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
11815 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
11816 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
11817 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
11818 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
11819 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
11820 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
11821 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
11822 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
11823 (Uuw8): New constraints.
11824 (Usb4): Move into alphabetical order.
11825 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
11826 (sd8_operand, ub8_operand, uw8_operand): New predicates.
11827 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
11828 previously unnamed patterns.
11829 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
11830 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
11831 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
11832 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
11833 of set_attr_alternative/if_then_else. Use extended_mips16 instead
11834 of specific lengths.
11835
cb3b8d33
JJ
118362013-05-10 Jakub Jelinek <jakub@redhat.com>
11837
6f93c008
JJ
11838 * config/i386/i386.md (rotateinv): New code attr.
11839 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11840 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11841 roll $31, %eax, etc.
11842
cb3b8d33
JJ
11843 PR tree-optimization/45216
11844 PR tree-optimization/57157
11845 * tree-ssa-forwprop.c (simplify_rotate): New function.
11846 (ssa_forward_propagate_and_combine): Call it.
11847
afb119be
RB
118482013-05-10 Richard Biener <rguenther@suse.de>
11849
11850 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
11851 disable peeling when we version for aliasing.
11852 (vector_alignment_reachable_p): Honor explicit user alignment.
11853 (vect_supportable_dr_alignment): Likewise.
11854 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
11855 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
11856 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
11857 then peeling to arrange for the cost-model check to come first.
11858
01ae4861
AM
118592013-05-10 Alan Modra <amodra@gmail.com>
11860
11861 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
11862 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
11863 * configure: Regenerate.
11864
ebc9a431
AM
118652013-05-10 Alan Modra <amodra@gmail.com>
11866
11867 PR target/55033
11868 * varasm.c (default_elf_select_section): Move !DECL_P check..
11869 (get_named_section): ..to here before calling get_section_name.
11870 Adjust assertion.
11871 (default_section_type_flags): Add DECL_P check.
11872 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
11873 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
11874
d4bca93c
JR
118752013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
11876
11877 * config/epiphany/epiphany.c (epiphany_expand_prologue):
11878 When using gen_stack_adjust_str with a register offset, add a
11879 REG_FRAME_RELATED_EXPR note.
11880
60ca9a65
UB
118812013-05-09 Uros Bizjak <ubizjak@gmail.com>
11882
11883 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
11884 (*vec_extractv4si_zext_mem): Ditto.
11885 (*vec_extractv2di): Add 0->x and x->x alternatives.
11886 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
11887 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
11888
a3409c02
JM
118892013-05-09 Jason Merrill <jason@redhat.com>
11890
0138d6b2
JM
11891 N3639 C++1y VLA support
11892 * gimplify.c (gimplify_vla_decl): Don't touch an existing
11893 DECL_VALUE_EXPR.
11894
a3409c02
JM
11895 * tree.c (build_constructor_va): New.
11896 * tree.h: Declare it.
11897
66e6b990
MJ
118982013-05-09 Martin Jambor <mjambor@suse.cz>
11899
11900 PR lto/57084
11901 * gimple-fold.c (canonicalize_constructor_val): Call
11902 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
11903
64cfa6c0
JH
119042013-05-09 Jan Hubicka <jh@suse.cz>
11905 Richard Biener <rguenther@suse.de>
11906
11907 PR lto/54095
11908 * symtab.c (symtab_make_decl_local): Do not add private names.
11909
c3167b00
JH
119102013-05-09 Jan Hubicka <jh@suse.cz>
11911
11912 PR lto/54095
11913 * symtab.c (insert_to_assembler_name_hash): Handle clones.
11914 (unlink_from_assembler_name_hash): Likewise.
11915 (symtab_prevail_in_asm_name_hash, symtab_register_node,
11916 symtab_unregister_node, symtab_initialize_asm_name_hash,
11917 change_decl_assembler_name): Update.
11918
12dc6974
SN
119192013-05-09 Sofiane Naci <sofiane.naci@arm.com>
11920
11921 * config/aarch64/aarch64.md: New movtf split.
11922 (*movtf_aarch64): Update.
11923 (aarch64_movdi_tilow): Handle TF modes and rename to
11924 aarch64_movdi_<mode>low.
11925 (aarch64_movdi_tihigh): Handle TF modes and rename to
11926 aarch64_movdi_<mode>high
11927 (aarch64_movtihigh_di): Handle TF modes and rename to
11928 aarch64_mov<mode>high_di
11929 (aarch64_movtilow_di): Handle TF modes and rename to
11930 aarch64_mov<mode>low_di
11931 (aarch64_movtilow_tilow): Remove spurious whitespace.
11932 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
11933 splits.
11934 (aarch64_print_operand): Update.
11935
227eb343
AM
119362013-05-09 Alan Modra <amodra@gmail.com>
11937
11938 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
11939 powerpc64le.
11940 * configure: Regenerate.
11941
0b013847
UB
119422013-05-08 Uros Bizjak <ubizjak@gmail.com>
11943
11944 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
11945 splitter preparation statements.
11946 * config/i386/sse.md (*vec_extract* splitters): Ditto.
11947 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
11948 adjust_address_nv.
11949
1dc3d6e9
BS
119502013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11951
11952 * gimple-ssa-strength-reduction.c (count_candidates): Change
11953 return value to int.
11954 (analyze_candidates_and_replace): Change type of length to int.
11955
e61e7d28
UB
119562013-05-08 Uros Bizjak <ubizjak@gmail.com>
11957
11958 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
11959 (*vec_extract<mode>): Use VI12_128 mode iterator.
11960 (*vec_extract<mode>_mem): Ditto.
11961 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
11962 attribute.
11963
4fbfcf44
DN
119642013-05-08 Diego Novillo <dnovillo@google.com>
11965
11966 PR bootstrap/54659
11967
11968 Revert:
4fbfcf44
DN
11969 2012-08-17 Diego Novillo <dnovillo@google.com>
11970
e61e7d28
UB
11971 PR bootstrap/54281
11972 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
11973 * config.in: Regenerate.
11974 * configure: Regenerate.
11975 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 11976
702d8703
JH
119772013-05-08 Jan Hubicka <jh@suse.cz>
11978
11979 PR lto/54095
11980 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
11981 * cgraph.h (symtab_node_base): Add unique_name.
11982 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11983 input_overwrite_node, input_varpool_node): Stream unique_name.
11984 * cgraphclones.c (cgraph_create_virtual_clone,
11985 cgraph_function_versioning): Set unique_name.
11986 * ipa.c (function_and_variable_visibility): Set unique_name.
11987
8b28cf47
BS
119882013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11989
11990 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
11991 (alloc_cand_and_find_basis): Restrict conditional candidate
11992 processing to CAND_MULTs.
11993
e86074fd
JH
119942013-05-08 Jan Hubicka <jh@suse.cz>
11995
11996 PR lto/54095
11997 lto-symtab.c (lto_symtab_symbol_p): New function.
11998 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
11999 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
12000 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
12001 Skip static symbols.
12002
44398cbe
PC
120032013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
12004
12005 PR tree-optimization/57200
12006 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
12007 Only call inform if the preceding warning_at returns true.
12008
f6bc1c4a
HS
120092013-05-07 Han Shen <shenhan@google.com>
12010
12011 * cfgexpand.c (record_or_union_type_has_array_p): New function.
12012 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
12013 * common.opt (fstack-protector-strong): New option.
12014 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
12015 * doc/invoke.texi (Optimization Options): Document
12016 "-fstack-protector-strong".
12017 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
12018
4ffecb1f
SB
120192013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12020
12021 * config/mips/mips.c (mips_machine_reorg2): Return 0.
12022
5a107a0f
VM
120232013-05-07 Vladimir Makarov <vmakarov@redhat.com>
12024
12025 * ira.c (update_equiv_regs): Add insn having equiv memory even if
12026 it is not lhs of the insn.
12027 (setup_reg_equiv): Remove insn having equiv memory which it is not
12028 lhs of the insn.
12029 * lra-constraints.c (process_address): Try to improve generation
12030 code for address base + disp.
12031 (lra_constraints): Make correct the code for checking insn setting
12032 up backward equivalence. Remove insn only if it is in the init
12033 insn list.
12034 * lra-eliminations.c (update_reg_eliminate): Change return value.
12035 (lra_eliminate): Use the result.
12036
3f5783ea
UB
120372013-05-07 Uros Bizjak <ubizjak@gmail.com>
12038
12039 * config/i386/sse.md (ssescalarnummask): New mode attribute.
12040 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
12041 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
12042 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
12043 register target operands.
12044 (*vec_extractv8hi_sse2): New pattern.
12045 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
12046 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
12047 (*vec_extract<mode>_mem): New insn and split pattern.
12048
8a5800b8
CL
120492013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
12050
12051 * config/arm/arm.c (arm_asan_shadow_offset): New function.
12052 (TARGET_ASAN_SHADOW_OFFSET): Define.
12053 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
12054 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
12055
7bf55a70
BS
120562013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12057
12058 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
12059 (incr_vec_index): Return -1 if increment not found.
12060 (create_add_on_incoming_edge): Assert if increment not found.
12061 (record_increment): Limit number of increments recorded.
12062 (all_phi_incrs_profitable): Return false if an increment not found.
12063 (replace_profitable_candidates): Don't process increments that were
12064 not recorded.
12065 (analyze_candidates_and_replace): Limit size of incr_vec.
12066
3f8825c0
RB
120672013-05-07 Richard Biener <rguenther@suse.de>
12068
12069 * calls.c (special_function_p): setjmp-like functions are leaf.
12070 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
12071 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
12072
fd4842cd
SN
120732013-05-07 Sofiane Naci <sofiane.naci@arm.com>
12074
12075 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
12076 (aarch64_simd_mov<mode>): New expander.
12077 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
12078 (aarch64_simd_mov_to_<mode>high): Likewise.
12079 (aarch64_simd_mov_from_<mode>low): Likewise.
12080 (aarch64_simd_mov_from_<mode>high): Likewise.
12081 (aarch64_dup_lane<mode>): Update.
12082 (aarch64_dup_lanedi): New instruction pattern.
12083 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
12084 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
12085
a7a7d10e
BS
120862013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12087
12088 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
12089 (replace_mult_candidate): Remove unnecessary argument; remove
12090 unnecessary parameter from call to introduce_cast_before_cand.
12091 (replace_unconditional_candidate): Remove unnecessary parameter
12092 from call to replace_mult_candidate.
12093 (replace_conditional_candidate): Likewise.
12094 (insert_initializers): Use make_temp_ssa_name.
12095 (introduce_cast_before_cand): Remove unnecessary argument; use
12096 make_temp_ssa_name.
12097 (replace_one_candidate): Remove unnecessary argument; remove
12098 unnecessary parameter from calls to introduce_cast_before_cand.
12099 (replace_profitable_candidates): Remove unnecessary parameters
12100 from calls to replace_one_candidate.
12101
29105868
BS
121022013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12103
12104 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
12105 phi def as possibly hiding a basis for a CAND_ADD whose operands
12106 have been commuted in the analysis.
12107 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
12108
4095f9fa
N
121092013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12110
12111 * config/aarch64/aarch64.md
12112 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
12113 shift value between 0-4.
12114
39e843e8
RB
121152013-05-07 Richard Biener <rguenther@suse.de>
12116
12117 * double-int.h (rshift): New overload.
12118 * double-int.c (rshift): New function.
12119 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
12120 (create_reference_ops_from_ref): Remove.
12121 (vn_reference_insert): Use shared ops for constructing the
12122 reference and copy it.
12123
0a1a83cb
RB
121242013-05-07 Richard Biener <rguenther@suse.de>
12125
12126 PR middle-end/57190
12127 * tree-eh.c (sink_clobbers): Properly propagate
12128 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
12129
43303d6f
JJ
121302013-05-07 Jakub Jelinek <jakub@redhat.com>
12131
ba7e83f8
JJ
12132 PR tree-optimization/57149
12133 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
12134 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
12135 collect_phi_def_edges, execute_late_warn_uninitialized): Use
12136 uninit_undefined_value_p instead of ssa_undefined_value_p.
12137
43303d6f
JJ
12138 PR debug/57184
12139 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
12140 for modifier == EXPAND_INITIALIZER.
12141
14523c25
AB
121422013-05-07 Anton Blanchard <anton@samba.org>
12143
12144 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
12145 for powerpc64 little endian.
12146 * configure: Regenerate.
12147
cb7c8be9
GS
121482013-05-06 Graham Stott <grahams@btinternet.com>
12149
12150 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
12151 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
12152 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
12153 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
12154
121552013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS
12156
12157 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
12158 codes which allow non-lvalues.
12159
9a0ee7b0
MG
121602013-05-06 Marc Glisse <marc.glisse@inria.fr>
12161
12162 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
12163 components are all 1s.
12164 (integer_minus_onep): New function.
12165 * tree.h (integer_minus_onep): Declare it.
12166 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
12167 integer_minus_onep instead of integer_all_onesp.
12168
f2c17ea9
OE
121692013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12170
12171 PR target/52933
12172 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
12173 variations of these patterns.
12174
f75e6a51
UB
121752013-05-06 Uros Bizjak <ubizjak@gmail.com>
12176
12177 * config/i386/i386.md (isa): Add x64_sse4 member.
12178 (enabled): Handle x64_sse4.
12179 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
12180 instruction for 64bit SSE4_1 targets. Update insn attributes.
12181 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
12182 instruction for SSE4_1 targets. Update insn attributes.
12183 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12184 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
12185 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
12186 const_1 selector.
12187 (*vec_extractv4si): Rename from *sse4_1_pextrd.
12188 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
12189 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
12190
a986d468
OE
121912013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12192
12193 PR target/57108
12194 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
12195
382522cb
MK
121962013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
12197
12198 * final.c (do_assembler_dialects): Don't handle curly braces and
12199 vertical bar escaped by % as dialect delimiters.
12200 (output_asm_insn): Print curly braces and vertical bar if escaped
12201 by % and ASSEMBLER_DIALECT defined.
12202 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
12203 * doc/tm.texi: Regenerated.
12204
fb0d5c60
SB
122052013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12206
fb0d5c60
SB
12207 * config/mips/mips.c: Include tree-pass.h.
12208 (mips_reorg): Split in pre- and post-dbr_schedule parts.
12209 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
12210 (pass_mips_machine_reorg2): New machine specific pass.
12211 (insert_pass_mips_machine_reorg2): New pass plugin definition.
12212 (mips_option_override): Register the new pass.
12213 * rtl.h (cleanup_barriers): Remove prototype.
12214 (dbr_schedule): Likewise.
12215 * jump.c (cleanup_barriers): Make static.
12216 * reorg.c (dbr_schedule): Likewise.
12217
aa06a978
RB
122182013-05-06 Richard Biener <rguenther@suse.de>
12219
12220 PR tree-optimization/57185
12221 * tree-parloops.c (add_field_for_reduction): Handle anonymous
12222 SSA names properly.
12223
0b953bec
UB
122242013-05-06 Uros Bizjak <ubizjak@gmail.com>
12225
12226 PR target/57106
12227 * config/i386/i386.c (add_parameter_dependencies): Add dependence
12228 between "first_arg" and "insn", not "last" and "insn".
12229
28708525
WS
122302013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12231
12232 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
12233 (find_candidates_in_block): Re-enable slsr_process_phi.
12234 (create_phi_basis): Fix double counting of candidate adjustment.
12235
0107dca2
RB
122362013-05-06 Richard Biener <rguenther@suse.de>
12237
12238 PR middle-end/57147
12239 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12240 the edge is also fallthru, preserve it and just clear the
12241 abnormal flag.
12242 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12243 also complex, preserve that and just clear the fallthru flag.
12244 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12245 update virtual operands.
12246
470d4d13
AM
122472013-05-06 Alan Modra <amodra@gmail.com>
12248
12249 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12250 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12251 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12252 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12253 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12254 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12255
8f1dbf8d
AM
122562013-05-06 Alan Modra <amodra@gmail.com>
12257
12258 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12259 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12260 (DEFAULT_ASM_ENDIAN): Define.
12261 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12262 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12263 Update -K PIC clause from sysv4.h.
12264 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12265 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12266
54890767
AM
122672013-05-06 Alan Modra <amodra@gmail.com>
12268
12269 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12270 twice for little-endian.
12271 (ashrdi3_no_power, ashrdi3): Support little-endian.
12272
2353515d
OE
122732013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12274
12275 PR target/55303
12276 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12277 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12278 related expanders.
12279 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12280 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12281 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12282 New predicates.
12283
33e67557
SB
122842013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12285 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12286
12287 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12288 * config/pa/pa.opt: Make mbig-switch a no-op.
12289 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12290 (CASE_VECTOR_MODE): Always return SImode.
12291 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12292 for the !TARGET_BIG_SWITCH case.
12293 * config/pa/pa-linux.h: Likewise.
12294 * config/pa/pa-openbsd.h: Likewise.
12295 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12296 * config/pa/pa.md (short_jump): Remove define_insn.
12297 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12298 (casesi0): Remove define_insn.
12299 (type): Remove btable_branch.
12300 (pa_combine_type): Likewise.
12301 (in_nullified_branch_delay): Likewise.
12302 (in_call_delay): Likewise.
12303 (define_delay): Likewise.
12304 (define_insn_reservation "Z3"): Likewise.
12305 (define_insn_reservation "Z4"): Likewise.
12306 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12307 (pa_adjust_insn_length): Remove adjustment for btable branches.
12308 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12309 and mno-big-switch
12310
3095685e
UB
123112013-05-05 Uros Bizjak <ubizjak@gmail.com>
12312
12313 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12314 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12315 Add m->r,x alternatives.
12316 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12317 splitters using SWI48x mode iterator.
12318 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12319 TARGET_64BIT. Add m->x alternative.
12320 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12321 Add o->x alternative. Enable for TARGET_SSE.
12322 (sse_storeq): Remove expander.
12323 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12324 with memory input operand.
12325 (*vec_extractv2di_1 splitter): New.
12326 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12327 * config/i386/i386.md (ssevecmodelower): New mode attribute.
12328
4b36ae28
SB
123292013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12330
12331 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
12332 (INT_LOWPART): Delete.
12333 (extract_MB): Adjust.
12334 (extract_ME): Adjust.
12335 (print_operand): Adjust.
12336
da226db2
SB
123372013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12338
12339 * config/rs6000/predicates.md (reg_or_add_cint_operand,
12340 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12341 (reg_or_logical_cint_operand, easy_fp_constant,
12342 logical_const_operand): Delete "CONST_DOUBLE" case.
12343 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
12344 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 12345 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB
12346 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
12347 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
12348 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
12349 test.
12350 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
12351 CONST_DOUBLE DImode/VOIDmode case.
12352 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
12353 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12354 CONST_DOUBLE VOIDmode case.
12355 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12356 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
12357 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
12358 Delete CONST_DOUBLE case.
12359 (splitters for mov FMOVE64 const_double): Delete
12360 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12361 "HOST_BITS_PER_WIDE_INT >= 64" test.
12362 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
12363 case.
12364 (mov DI const_double): Delete.
12365
40de22d6
JJ
123662013-05-04 Jakub Jelinek <jakub@redhat.com>
12367
12368 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
12369 on op shows all bits zero in mode of a lowpart subreg, return zero.
12370
5ec6aff2
MM
123712013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
12372
12373 PR target/57150
12374 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
12375 to save TFmode registers and DImode to save TImode registers for
12376 caller save operations.
12377 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
12378 mark being partially clobbered since they only use the first
12379 double word.
12380
12381 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
12382 and TDmode only use the upper 64-bits of each VSX register.
12383
2cefad90
BS
123842013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12385
12386 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
12387 (find_candidates_in_block): Disable slsr_process_phi.
12388
d6d7eee1
GW
123892013-05-03 Guozhi Wei <carrot@google.com>
12390
12391 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB
12392 constructor to ...
12393 (build_init_ctor): ... here.
d6d7eee1 12394
9b92d12b
BS
123952013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12396
12397 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12398 (slsr_cand_d): Redefine def_phi.
12399 (stride_status, phi_adjust_status, count_phis_status): New enums.
12400 (find_phi_def): New.
12401 (find_basis_for_base_expr): New.
12402 (find_basis_for_candidate): Handle hidden bases.
12403 (alloc_cand_and_find_basis): Handle phi candidates.
12404 (slsr_process_phi): New.
12405 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12406 (create_mul_imm_cand): Likewise.
12407 (create_add_ssa_cand): Exclude phi base candidates.
12408 (create_add_imm_cand): Likewise.
12409 (slsr_process_cast): Likewise.
12410 (slsr_process_copy): Likewise.
12411 (find_candidates_in_block): Handle phi candidates.
12412 (dump_candidate): Likewise.
12413 (unconditional_cands): Delete.
12414 (unconditional_cands_with_known_stride_p): Delete.
12415 (phi_dependent_cand_p): New.
12416 (cand_increment): Handle phi-dependent candidates.
12417 (replace_dependent): Delete.
12418 (replace_mult_candidate): New.
12419 (replace_unconditional_candidate): New.
12420 (incr_vec_index): Move to avoid forward reference.
12421 (create_add_on_incoming_edge): New.
12422 (create_phi_basis): New.
12423 (replace_dependents): Delete.
12424 (replace_conditional_candidate): New.
12425 (phi_add_costs): New.
12426 (replace_uncond_cands_and_profitable_phis): New.
12427 (record_increment): Handle phi adjustments.
12428 (record_phi_increments): New.
12429 (record_increments): Handle phi adjustments.
12430 (phi_incr_cost): New.
12431 (lowest_cost_path): Handle phis.
12432 (total_savings): Likewise.
12433 (analyze_increments): Likewise.
12434 (ncd_with_phi): New.
12435 (ncd_of_cand_and_phis): New.
12436 (nearest_common_dominator_for_cands): Handle phi increments.
12437 (all_phi_incrs_profitable): New.
12438 (replace_profitable_candidates): Handle phi-dependent candidates.
12439 (analyze_candidates_and_replace): Likewise.
12440
68f073d4
TJ
124412013-05-03 Teresa Johnson <tejohnson@google.com>
12442
12443 PR bootstrap/57154
12444 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12445 do not exceed REG_BR_PROB_BASE.
12446
a4ee7cb9
JL
124472013-05-03 Jeff Law <law@redhat.com>
12448
ade67f70 12449 PR tree-optimization/57144
a4ee7cb9
JL
12450 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12451 operand of the condition will bit into the new type when eliminating
12452 a cast feeding a condition.
12453
47954c4e
JJ
124542013-05-03 Jakub Jelinek <jakub@redhat.com>
12455
12456 PR rtl-optimization/57130
3095685e
UB
12457 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12458 of COMPARE as in_code to the recursive call if needed.
47954c4e 12459
3c21604f
UB
124602013-05-03 Uros Bizjak <ubizjak@gmail.com>
12461
12462 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12463 (enabled): Handle new members.
12464 * config/i386/sse.md (*vec_concatv2si): Merge from
12465 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12466 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12467
12211b99 124682013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR
12469
12470 PR tree-optimization/57027
12471 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12472 for fnms opportunity, check we got the prerequisite kind
12473 of tree / gimple before using accessor functions.
12474
07bfc9ec
RB
124752013-05-03 Richard Biener <rguenther@suse.de>
12476
12477 * double-int.h (lshift): New overload without precision
12478 and arith argument.
12479 (operator *=, operator +=, operator -=): Move ...
12480 * double-int.c (operator *=, operator +=, operator -=): ... here
12481 and implement more efficiently.
12482 (mul_double_with_sign): Remove.
12483 (lshift_double): Adjust to take unsinged shift argument, push
12484 dispatching code to callers.
12485 (mul_double_wide_with_sign): Add early out for callers that
12486 are not interested in high parts or overflow.
12487 (lshift): New function.
12488 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12489 dispatch code here.
12490 (lrotate, rrotate): Use logical shifts.
12491 * expr.c (get_inner_reference): Use lshift.
12492 * fixed-value.c (do_fixed_divide): Likewise.
12493 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12494 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12495 (indirect_refs_may_alias_p): Likewise.
12496 (stmt_kills_ref_p_1): Likewise.
12497
7769bb64
VP
124982013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12499
12500 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12501
1d0c8e5c
VP
125022013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12503
12504 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12505 scalar form of FABD instruction.
12506
f15feaf9
VM
125072013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12508
12509 * lra-constraints.c (process_alt_operands): Add checking alt
12510 number to choose the best alternative.
12511
d90e76d4
RB
125122013-05-02 Richard Biener <rguenther@suse.de>
12513
12514 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12515 bitmap and its handling.
12516 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12517
9f8e7a96
RB
125182013-05-02 Richard Biener <rguenther@suse.de>
12519
12520 PR middle-end/57140
12521 * tree-inline.c (copy_loops): Properly handle removed loops.
12522 (copy_cfg_body): Mark destination loops for fixup if source
12523 loops needed fixup.
12524
f3a81b39
GY
125252013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12526
12527 PR target/56732
12528 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12529 generating simple_return for naked functions.
12530
7b920a9a
MJ
125312013-05-02 Martin Jambor <mjambor@suse.cz>
12532
12533 PR middle-end/56988
12534 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12535 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12536 flags match.
12537 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12538 ipa_agg_replacement_value structures.
12539 (known_aggs_to_agg_replacement_list): Likewise.
12540 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12541 (read_agg_replacement_chain): Likewise.
12542 (ipcp_transform_function): Also check that by_ref flags match.
12543
2c41c19d
RB
125442013-05-02 Richard Biener <rguenther@suse.de>
12545
12546 * graphds.h (struct graph): Add obstack member.
12547 * graphds.c (new_graph): Initialize obstack and allocate
12548 vertices from it.
12549 (add_edge): Allocate edge from the obstack.
3c21604f 12550 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 12551
8b47039c
TJ
125522013-05-02 Teresa Johnson <tejohnson@google.com>
12553
12554 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12555 divides.
12556 * cfg.c (update_bb_profile_for_threading): Ditto.
12557 * tree-inline.c (copy_bb): Ditto.
12558 (copy_edges_for_bb): Ditto.
12559 (initialize_cfun): Ditto.
12560 (copy_cfg_body): Ditto.
12561 (expand_call_inline): Ditto.
12562 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12563 (estimate_node_size_and_time): Ditto.
12564 (inline_merge_summary): Ditto.
12565 * cgraphclones.c (cgraph_clone_edge): Ditto.
12566 (cgraph_clone_node): Ditto.
12567 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12568 (compute_trg_info): Ditto.
12569
da65928c
IB
125702013-05-02 Ian Bolton <ian.bolton@arm.com>
12571
12572 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12573 S reg when fp attribute set.
12574 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12575
add3c965
IB
125762013-05-02 Ian Bolton <ian.bolton@arm.com>
12577
12578 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12579 New pattern.
12580 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12581 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12582 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12583
fdd43ac4
RB
125842013-05-02 Richard Biener <rguenther@suse.de>
12585
12586 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12587 (struct instantiate_cache_entry): New type.
12588 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12589 (struct instantiate_cache_type): New type.
12590 (set_instantiated_value, get_instantiated_value): Remove.
12591 (get_instantiated_value_entry): New function.
12592 (instantiate_scev_name): Use the new cache and adjust.
12593 (instantiate_scev_poly): Adjust.
12594 (instantiate_scev_binary): Likewise.
12595 (instantiate_array_ref): Likewise.
12596 (instantiate_scev_convert): Likewise.
12597 (instantiate_scev_not): Likewise.
12598 (instantiate_scev_3): Likewise.
12599 (instantiate_scev_2): Likewise.
12600 (instantiate_scev_r): Likewise.
12601 (instantiate_scev): Likewise.
12602 (resolve_mixers): Likewise.
12603
36ff9dfb
VM
126042013-05-01 Vladimir Makarov <vmakarov@redhat.com>
12605
12606 PR target/57091
12607 * lra-constraints.c (best_small_class_operands_num): Remove.
12608 (process_alt_operands): Remove small_class_operands_num. Take
12609 small classes operands into losers and only if the operand is not
12610 matched. Modify debugging output.
12611 (curr_insn_transform): Remove best_small_class_operands_num.
12612 Print insn name.
12613
36054fab
JG
126142013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12615
12616 * config/aarch64/aarch64-builtins.c
12617 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
12618 * config/aarch64/aarch64-simd-builtins.def
12619 (reduc_splus_): Add new modes.
12620 (reduc_uplus_): New.
12621 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
12622 (reduc_uplus_v4sf): Likewise.
12623 (reduc_splus_v4sf): Likewise.
12624 (aarch64_addv<mode>): Likewise.
12625 (reduc_uplus_<mode>): Likewise.
12626 (reduc_splus_<mode>): Likewise.
12627 (aarch64_addvv2di): Likewise.
12628 (reduc_uplus_v2di): Likewise.
12629 (reduc_splus_v2di): Likewise.
12630 (aarch64_addvv2si): Likewise.
12631 (reduc_uplus_v2si): Likewise.
12632 (reduc_splus_v2si): Likewise.
12633 (reduc_<sur>plus_<mode>): New.
12634 (reduc_<sur>plus_v2di): Likewise.
12635 (reduc_<sur>plus_v2si): Likewise.
12636 (reduc_<sur>plus_v4sf): Likewise.
12637 (aarch64_addpv4sf): Likewise.
12638 * config/aarch64/arm_neon.h
12639 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
12640 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
12641 add UNSPEC_SADDV, UNSPEC_UADDV.
12642 (SUADDV): New.
12643 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
12644
6dce23a8
JG
126452013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12646
12647 * config/aarch64/arm_neon.h
12648 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
12649
1598945b
JG
126502013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12651
12652 * config/aarch64/aarch64-builtins
12653 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
12654
998eaf97
JG
126552013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12656
12657 * config/aarch64/aarch64-simd-builtins.def
12658 (reduc_smax_): New.
12659 (reduc_smin_): Likewise.
12660 (reduc_umax_): Likewise.
12661 (reduc_umin_): Likewise.
12662 (reduc_smax_nan_): Likewise.
12663 (reduc_smin_nan_): Likewise.
12664 (fmax): Remove.
12665 (fmin): Likewise.
12666 (smax): Update for V2SF, V4SF and V2DF modes.
12667 (smin): Likewise.
12668 (smax_nan): New.
12669 (smin_nan): Likewise.
12670 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
12671 (<su><maxmin><mode>3): ...This, refactor.
12672 (s<maxmin><mode>3): New.
12673 (<maxmin_uns><mode>3): Likewise.
12674 (reduc_<maxmin_uns>_<mode>): Refactor.
12675 (reduc_<maxmin_uns>_v4sf): Likewise.
12676 (reduc_<maxmin_uns>_v2si): Likewise.
12677 (aarch64_<fmaxmin><mode>: Remove.
12678 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
12679 new builtin names.
12680 (vmin<q>_f<32,64>): Likewise.
12681 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12682 (FMAXMIN): New.
12683 (su): Add mappings for smax, smin, umax, umin.
12684 (maxmin): New.
12685 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12686 (FMAXMIN): Rename as...
12687 (FMAXMIN_UNS): ...This.
12688 (maxminv): Remove.
12689 (fmaxminv): Likewise.
12690 (fmaxmin): Likewise.
12691 (maxmin_uns): New.
12692 (maxmin_uns_op): Likewise.
12693
bd11644e
JG
126942013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12695
12696 * config/aarch64/arm_neon.h
12697 (vac<ge, gt><sd>_f<32, 64>): Rename to...
12698 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
12699 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
12700
75dd5ace
JG
127012013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12702
12703 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
12704 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
12705
7c19979f
JG
127062013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12707
12708 * config/aarch64/aarch64-simd.md
12709 (vcond<mode>_internal): Handle special cases for constant masks.
12710 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
12711 (vcondu<mode><mode>): Likewise.
12712 (vcond<v_cmp_result><mode>): New.
12713
bb60efd9
JG
127142013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12715
12716 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
12717 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
12718 * config/aarch64/aarch64-simd-builtins.def
12719 (cmeq): Update to BUILTIN_VALLDI.
12720 (cmgt): Likewise.
12721 (cmge): Likewise.
12722 (cmle): Likewise.
12723 (cmlt): Likewise.
12724 * config/aarch64/arm_neon.h
12725 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
12726 to builtins or C as appropriate.
12727
889b9412
JG
127282013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12729
12730 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
12731 (cmgeu): ...This.
12732 (cmhi): Rename to...
12733 (cmgtu): ...This.
12734 * config/aarch64/aarch64-simd.md
12735 (simd_mode): Add SF.
12736 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
12737 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
12738 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
12739 (cstore<mode>_neg): ...This.
12740 * config/aarch64/iterators.md
12741 (VALLF): new.
12742 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
12743 (COMPARISONS): New.
12744 (UCOMPARISONS): Likewise.
12745 (optab): Add missing comparisons.
12746 (n_optab): New.
12747 (cmp_1): Likewise.
12748 (cmp_2): Likewise.
12749 (CMP): Likewise.
12750 (cmp): Remove.
12751 (VCMP_S): Likewise.
12752 (VCMP_U): Likewise.
12753 (V_cmp_result): Add DF, SF modes.
12754 (v_cmp_result): Likewise.
12755 (v): Likewise.
12756 (vmtype): Likewise.
12757 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
12758
0a7dbb76
GY
127592013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
12760
12761 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
12762 define_insn to define_insn_and_split.
12763 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
12764 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
12765 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
12766 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
12767 (thumb2_negscc): Likewise.
12768
fb614ca6
GY
127692013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12770
12771 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
12772
9e64a0bf
GY
127732013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12774
12775 * config/arm/thumb2.md: Remove trailing whitespaces.
12776
d6b28156
RS
127772013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12778
12779 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
12780 Use gen_int_mode rather than GEN_INT.
12781
f91674c3
L
127822013-04-30 H.J. Lu <hongjiu.lu@intel.com>
12783
b0dec607 12784 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L
12785 debug_gimple_stmt.
12786
3551257c
RB
127872013-04-30 Richard Biener <rguenther@suse.de>
12788
12789 PR middle-end/57122
3c21604f 12790 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 12791
4c1aff1c
RB
127922013-04-30 Richard Biener <rguenther@suse.de>
12793
12794 PR middle-end/57107
12795 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
12796
54b8379a
AB
127972013-04-30 Andrey Belevantsev <abel@ispras.ru>
12798
12799 PR rtl-optimization/56957
12800 PR rtl-optimization/57105
54b8379a
AB
12801 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
12802 variable. Use just INSN_UID for determining whether an insn
12803 should be only disconnected from the insn stream.
12804 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
12805
abc27962
JJ
128062013-04-30 Jakub Jelinek <jakub@redhat.com>
12807
12808 PR tree-optimization/57104
12809 * tsan.c (instrument_expr): Don't instrument accesses to
12810 DECL_HARD_REGISTER VAR_DECLs.
12811
0fc822d0
RB
128122013-04-30 Richard Biener <rguenther@suse.de>
12813
12814 * function.h (loops_for_fn): New inline function.
12815 (set_loops_for_fn): Likewise.
12816 * cfgloop.h (place_new_loop): Add struct function parameter.
12817 (get_loop): Likewise.
12818 (get_loops): Likewise.
12819 (number_of_loops): Likewise.
12820 (fel_next): Adjust.
12821 (fel_init): Likewise.
12822 * cfg.c (get_loop_copy): Adjust.
12823 * cfgloop.c (flow_loops_dump): Likewise.
12824 (record_loop_exits): Likewise.
12825 (verify_loop_structure): Likewise.
12826 * cfgloopanal.c (mark_irreducible_loops): Likewise.
12827 (estimate_reg_pressure_cost): Likewise.
12828 (mark_loop_exit_edges): Likewise.
12829 * cfgloopmanip.c (place_new_loop): Likewise.
12830 (add_loop): Likewise.
12831 (duplicate_loop): Likewise.
12832 * graph.c (draw_cfg_nodes): Likewise.
12833 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
12834 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
12835 (extract_affine_chrec): Likewise.
12836 (build_scop_iteration_domain): Likewise.
12837 * graphite.c (graphite_initialize): Likewise.
12838 * ira-build.c (create_loop_tree_nodes): Likewise.
12839 (more_one_region_p): Likewise.
12840 (rebuild_regno_allocno_maps): Likewise.
12841 (mark_loops_for_removal): Likewise.
12842 (mark_all_loops_for_removal): Likewise.
12843 (remove_unnecessary_regions): Likewise.
12844 (ira_build): Likewise.
12845 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
12846 * loop-init.c (fix_loop_structure): Likewise.
12847 (gate_rtl_move_loop_invariants): Likewise.
12848 (gate_rtl_unswitch): Likewise.
12849 (gate_rtl_unroll_and_peel_loops): Likewise.
12850 (rtl_doloop): Likewise.
12851 * lto-streamer-in.c (input_cfg): Likewise.
12852 * lto-streamer-out.c (output_cfg): Likewise.
12853 * modulo-sched.c (sms_schedule): Likewise.
12854 * predict.c (tree_estimate_probability): Likewise.
12855 (tree_estimate_probability_driver): Likewise.
12856 (estimate_loops): Likewise.
12857 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
12858 (move_sese_region_to_fn): Likewise.
12859 (debug_loop_num): Likewise.
12860 * tree-chrec.c (chrec_evaluate): Likewise.
12861 (hide_evolution_in_other_loops_than_loop): Likewise.
12862 (chrec_component_in_loop_num): Likewise.
12863 (reset_evolution_in_loop): Likewise.
12864 (evolution_function_is_invariant_rec_p): Likewise.
12865 * tree-if-conv.c (main_tree_if_conversion): Likewise.
12866 * tree-inline.c (copy_loops): Likewise.
12867 (copy_cfg_body): Likewise.
12868 (tree_function_versioning): Likewise.
12869 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
12870 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
12871 Likewise.
12872 (add_to_evolution_1): Likewise.
12873 (scev_const_prop): Likewise.
12874 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
12875 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
12876 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
12877 (tree_ssa_lim_initialize): Likewise.
12878 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
12879 (verify_loop_closed_ssa): Likewise.
12880 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
12881 (tree_ssa_loop_im): Likewise.
12882 (tree_ssa_loop_unswitch): Likewise.
12883 (tree_vectorize): Likewise.
12884 (check_data_deps): Likewise.
12885 (tree_ssa_loop_ivcanon): Likewise.
12886 (tree_ssa_loop_bounds): Likewise.
12887 (tree_complete_unroll): Likewise.
12888 (tree_complete_unroll_inner): Likewise.
12889 (tree_parallelize_loops): Likewise.
12890 (tree_ssa_loop_prefetch): Likewise.
12891 (tree_ssa_loop_ivopts): Likewise.
12892 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
12893 * tree-vectorizer.c (vectorize_loops): Likewise.
12894
37953bd3
MF
128952013-04-29 Mike Frysinger <vapier@gentoo.org>
12896
12897 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
12898 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
12899 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
12900 with EABI_LINK_SPEC.
12901
f9ed28db
UB
129022013-04-29 Uros Bizjak <ubizjak@gmail.com>
12903
12904 PR target/44578
12905 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
12906 alternative.
12907
deca73f5
VM
129082013-04-29 Vladimir Makarov <vmakarov@redhat.com>
12909
12910 PR target/57097
37953bd3 12911 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM
12912 using memory for pseudos. Print cost dump for alternatives.
12913 Modify cost values for conflicts with early clobbers.
12914 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
12915
fdca7d03
UB
129162013-04-29 Uros Bizjak <ubizjak@gmail.com>
12917
12918 PR target/57098
12919 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
12920
0b064172
IB
129212013-04-29 Ian Bolton <ian.bolton@arm.com>
12922
12923 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
12924 from/to S register.
12925 (movdi_aarch64): Support LDR/STR from/to D register.
12926
473cec55
IB
129272013-04-29 Ian Bolton <ian.bolton@arm.com>
12928
12929 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
12930 or higher by default.
12931
a6f30e66
RB
129322013-04-29 Richard Biener <rguenther@suse.de>
12933
12934 PR middle-end/57075
12935 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
12936 even if not adding abnormal edges for calls that can make
12937 abnormal gotos.
12938
0c2b2040
RB
129392013-04-29 Richard Biener <rguenther@suse.de>
12940
12941 PR middle-end/57103
12942 * tree-cfg.c (move_stmt_op): Fix condition under which to update
12943 TREE_BLOCK.
12944 (move_stmt_r): Remove redundant checking.
12945
f41f80f9
TJ
129462013-04-29 Teresa Johnson <tejohnson@google.com>
12947
12948 PR bootstrap/57077
12949 * basic-block.h (apply_scale): New function.
12950 (apply_probability): Use apply_scale.
12951 * gimple-streamer-in.c (input_bb): Ditto.
12952 * lto-streamer-in.c (input_cfg): Ditto.
12953 * lto-cgraph.c (merge_profile_summaries): Ditto.
12954 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 12955 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ
12956 (copy_edges_for_bb): Ditto.
12957 (copy_cfg_body): Ditto.
12958
315bbd2e
TV
129592013-04-29 Tom de Vries <tom@codesourcery.com>
12960
12961 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
12962 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
12963 (tail_merge_optimize): Handle current_loops == NULL.
12964
ebbd90d8
JL
129652013-04-26 Jeff Law <law@redhat.com>
12966
12967 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
12968 (simplify_cond_using_ranges): Generalize code to simplify
12969 COND_EXPRs where one argument is a constant and the other
12970 is an SSA_NAME created by an integral type conversion.
12971
8b9b57eb
KT
129722013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12973
12974 * config/arm/arm.md (store_minmaxsi): Use only when
12975 optimize_insn_for_size_p.
12976
9498e5dc
CB
129772013-04-29 Christian Bruel <christian.bruel@st.com>
12978
12979 PR target/57108
12980 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
12981
6093bc06
RB
129822013-04-29 Richard Biener <rguenther@suse.de>
12983
12984 PR middle-end/57089
fdca7d03
UB
12985 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
12986 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB
12987 (expand_omp_for_generic): Properly add loops.
12988 (expand_omp_for_static_nochunk): Likewise.
12989 (expand_omp_for_static_chunk): Likewise.
12990 (expand_omp_for): For the degenerate case fixup loops.
12991 (expand_omp_sections): Fix default bb placement in loops.
12992 (expand_omp_atomic_pipeline): Properly add loops.
12993
84aacbfd
KT
129942013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12995
12996 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
12997
8dee4479
TV
129982013-04-29 Tom de Vries <tom@codesourcery.com>
12999
13000 * tree-ssa-tail-merge.c: Update header comment.
13001
47934dc4
JG
130022013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13003
13004 * config/aarch64/arm_neon.h
13005 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
13006 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
13007 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
13008 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
13009 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
13010 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
13011 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
13012 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
13013
384be29f
JG
130142013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13015
13016 * config/aarch64/aarch64-simd.md
13017 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
13018 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
13019 fix_trunc, fixuns_trunc.
13020 (ftrunc<VDQF:mode>2): New.
13021 * config/aarch64/iterators.md (optab): Add fix, fixuns.
13022 (fix_trunc_optab): New.
13023
0386b123
JG
130242013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13025
13026 * config/aarch64/aarch64-builtins.c
13027 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
13028 iceilf, lround, iroundf.
13029
00fcb892
UB
130302013-04-29 Uros Bizjak <ubizjak@gmail.com>
13031
13032 PR target/54349
13033 * config/i386/i386.h (enum ix86_tune_indices)
13034 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
13035 New, split from X86_TUNE_INTER_UNIT_MOVES.
13036 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
13037 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
13038 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
13039 (TARGET_INTER_UNIT_MOVES): Remove.
13040 * config/i386/i386.c (initial_ix86_tune_features): Update.
13041 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
13042 (ix86_expand_convert_uns_didf_sse): Use
13043 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13044 (ix86_expand_vector_init_one_nonzero): Ditto.
13045 (ix86_expand_vector_init_interleave): Ditto.
13046 (inline_secondary_memory_needed): Return true for moves from SSE class
13047 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
13048 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13049 * config/i386/constraints.md (Yi, Ym): Depend on
13050 TARGET_INTER_UNIT_MOVES_TO_VEC.
13051 (Yj, Yn): New constraints.
13052 * config/i386/i386.md (*movdi_internal): Change constraints of
13053 operand 1 from Yi to Yj and from Ym to Yn.
13054 (*movsi_internal): Ditto.
13055 (*movdf_internal): Ditto.
13056 (*movsf_internal): Ditto.
13057 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
13058 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13059 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
13060 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
13061 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
13062 * config/i386/sse.md (movdi_to_sse): Ditto.
13063 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
13064 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
13065 TARGET_INTER_UNIT_MOVES.
13066 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
13067 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
13068 instead of TARGET_INTER_UNIT_MOVES.
13069 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
13070 operand 1 from Yi to Yj and from Ym to Yn.
13071
4c871069
JG
130722013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13073
13074 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
13075 (float_truncate_hi_): Likewise.
13076 (float_extend_lo_): Likewise.
13077 (float_truncate_lo_): Likewise.
13078 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
13079 (aarch64_float_extend_lo_v2df): Likewise.
13080 (vec_unpacks_hi_v4sf): Likewise.
13081 (aarch64_float_truncate_lo_v2sf): Likewise.
13082 (aarch64_float_truncate_hi_v4sf): Likewise.
13083 (vec_pack_trunc_v2df): Likewise.
13084 (vec_pack_trunc_df): Likewise.
13085
1709ff9b
JG
130862013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13087
13088 * config/aarch64/aarch64-builtins.c
13089 (aarch64_fold_builtin): Fold float conversions.
13090 * config/aarch64/aarch64-simd-builtins.def
13091 (floatv2si, floatv4si, floatv2di): New.
13092 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
13093 * config/aarch64/aarch64-simd.md
13094 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
13095 * config/aarch64/iterators.md (FLOATUORS): New.
13096 (optab): Add float, floatuns.
13097 (su_optab): Likewise.
13098
ce966824
JG
130992013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13100
13101 * config/aarch64/aarch64-builtins.c
13102 (aarch64_builtin_vectorized_function): Use new names for
13103 fcvt builtins.
13104 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
13105 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
13106 (fcvtzu): Split as...
13107 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
13108 (fcvtas): Split as...
13109 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
13110 (fcvtau): Split as...
13111 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
13112 (fcvtps): Split as...
13113 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
13114 (fcvtpu): Split as...
13115 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
13116 (fcvtms): Split as...
13117 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
13118 (fcvtmu): Split as...
13119 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
13120 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
13121 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
13122 (lfrintnusf, lfrintnudf): Likewise.
13123 * config/aarch64/aarch64-simd.md
13124 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
13125 define_insn.
13126 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
13127 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
13128 (fcvt_pattern): Likewise.
13129
b9de24fe
JG
131302013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13131
13132 * config/aarch64/aarch64-simd.md
13133 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
13134 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
13135
77a205be
JG
131362013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13137
13138 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
13139 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
13140 (vrnd<a,m,n,p>_f32): Implement using builtins.
13141 (vrnd<i,x><q>_f<32, 64>): New.
13142
0659ce6f
JG
131432013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13144
13145 * config/aarch64/aarch64-builtins.c
13146 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
13147 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
13148 (frintz): Rename to...
13149 (btrunc): ...this.
13150 (frintp): Rename to...
13151 (ceil): ...this.
13152 (frintm): Rename to...
13153 (floor): ...this.
13154 (frinti): Rename to...
13155 (nearbyint): ...this.
13156 (frintx): Rename to...
13157 (rint): ...this.
13158 (frinta): Rename to...
13159 (round): ...this.
13160 * config/aarch64/aarch64-simd.md
13161 (aarch64_frint<frint_suffix><mode>): Delete.
13162 (<frint_pattern><mode>2): Convert to insn.
13163 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
13164 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
13165 (frint_pattern): Likewise.
13166 (frint_suffix): Likewise.
13167
ea78906a
RB
131682013-04-29 Richard Biener <rguenther@suse.de>
13169
13170 PR tree-optimization/57081
13171 * loop-init.c: Include tree-flow.h.
13172 (loop_optimizer_finalize): Free number of iteration estimates.
13173 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
13174
baee1763
JJ
131752013-04-29 Jakub Jelinek <jakub@redhat.com>
13176
94dc5332
JJ
13177 PR tree-optimization/57083
13178 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
13179 non-singleton shift count range, zero extend low_bound for uns case.
13180
baee1763
JJ
13181 * config/i386/predicates.md (general_vector_operand): New predicate.
13182 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
13183 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
13184 if they aren't nonimmediate operands. If their original values
13185 satisfy const_vector_equal_evenodd_p, don't shift them.
13186 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
13187 predicates. For the SSE4.1 case force operands[{1,2}] into registers
13188 if not nonimmediate_operand.
13189 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
13190 instead of register_operand.
13191 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
13192
a9073727 131932013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB
13194
13195 * stor-layout.c (finalize_size_functions): Allocate a structure and
13196 reset cfun before dumping the functions.
13197
ba8011e6
JJ
131982013-04-27 Jakub Jelinek <jakub@redhat.com>
13199
d6fde69e
JJ
13200 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
13201
ba8011e6
JJ
13202 PR target/56866
13203 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
13204 use xop_pmacsdqh if uns_p.
13205 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
13206 the immediate rotate count.
13207
2c62cbaa
VM
132082013-04-26 Vladimir Makarov <vmakarov@redhat.com>
13209
13210 * rtl.h (struct rtx_def): Add comment for field jump.
13211 (LRA_SUBREG_P): New macro.
13212 * recog.c (register_operand): Check LRA_SUBREG_P.
13213 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13214 stack frame.
13215 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13216 (lra_final_code_change): Skip subreg change for operators.
13217 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13218 if there are no operand changes.
13219 * lra-constraints.c (curr_insn_set): New.
13220 (match_reload): Set LRA_SUBREG_P.
13221 (emit_spill_move): Ditto.
13222 (check_and_process_move): Use curr_insn_set. Process only single
13223 set insns. Don't initialize sec_mem_p and change_p.
13224 (simplify_operand_subreg): Use LRA_SUBREG_P.
13225 (reg_in_class_p): New function.
13226 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13227 of #ifdef. Add code to remove cycling.
13228 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13229 non-null disp. Reload inner instead of disp when base and index
13230 are null. Try to put lo_sum into register.
13231 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13232 (check_and_process_move): Move code for move cost check to
13233 simple_move_p. Remove equiv_substitution.
13234 (simple_move_p): New function.
13235 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13236 curr_insn_set. Call check_and_process_move only for single set
13237 insns. Use the new function. Move call of check_and_process_move
13238 after operand equiv substitution and address process.
13239
e7d764f3
JJ
132402013-04-26 Jakub Jelinek <jakub@redhat.com>
13241
13242 PR go/57045
13243 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13244 with nonlocal goto receivers or returns twice calls, ignore
13245 unininitialized values from abnormal edges to nl goto receiver
13246 or returns twice call.
13247
41e10689
JJ
132482013-04-26 Jakub Jelinek <jakub@redhat.com>
13249
13250 PR tree-optimization/57051
13251 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13252 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13253 bitsize.
13254
d7ed20db
RB
132552013-04-26 Richard Biener <rguenther@suse.de>
13256
13257 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13258 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13259 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13260 (fixup_loop_arrays_after_move): New function.
13261 (move_sese_region_to_fn): Properly outline the loop tree parts
13262 of the SESE region.
13263
df93505e
UB
132642013-04-26 Uros Bizjak <ubizjak@gmail.com>
13265
13266 * config/i386/i386.md (type, unit): Fix long lines.
13267
dd366ec3
RB
132682013-04-26 Richard Biener <rguenther@suse.de>
13269
13270 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13271 (lto-streamer-out.o): Likewise.
13272 * cfgloop.c (init_loops_structure): Export, add struct function
13273 argument and adjust.
13274 (flow_loops_find): Adjust.
13275 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13276 (init_loops_structure): Declare.
13277 * lto-streamer-in.c: Include cfgloop.h.
13278 (input_cfg): Input the loop tree.
13279 * lto-streamer-out.c: Include cfgloop.h.
13280 (output_cfg): Output the loop tree.
13281 (output_struct_function_base): Do not drop PROP_loops.
13282
a9e0d843
RB
132832013-03-26 Richard Biener <rguenther@suse.de>
13284
13285 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13286 (pass_build_cfg): Provide PROP_loops.
13287 (move_sese_region_to_fn): Remove loops that are outlined into fn
13288 for now.
13289 * tree-inline.c: Include cfgloop.h.
13290 (initialize_cfun): Do not drop PROP_loops.
13291 (copy_loops): New function.
13292 (copy_cfg_body): Copy loop structure.
13293 (tree_function_versioning): Initialize destination loop tree.
13294 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13295 (pass_parallelize_loops): Do IL verification.
13296 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13297 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13298 the CFG make sure we fixup loops as well.
13299 * tree-ssa-tail-merge.c: Include cfgloop.h.
13300 (replace_block_by): When merging loop latches mark loops for fixup.
13301 * lto-streamer-out.c (output_struct_function_base): Drop
13302 PROP_loops for now.
13303 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13304 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13305 * ipa-split.c: Include cfgloop.h.
13306 (split_function): Add the new return block to the loop tree root.
13307 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13308 whether we have removed the forwarder block.
13309 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13310 * cfgloop.h (place_new_loop): Declare.
13311 * cfgloopmanip.c (place_new_loop): Export.
13312 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13313 (tree-switch-conversion.o): Likewise.
13314 (tree-complex.o): Likewise.
13315 (tree-inline.o): Likewise.
13316 (tree-ssa-tailmerge.o): Likewise.
13317 (ipa-split.o): Likewise.
13318 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13319 (tree-ssa-copy.o): Likewise.
13320 * tree-switch-conversion.c: Include cfgloop.h
13321 (process_switch): If we emit a bit-test cascade, schedule loops
13322 for fixup.
13323 * tree-complex.c: Include cfgloop.h.
13324 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13325 * asan.c: Include cfgloop.h.
13326 (create_cond_insert_point): Properly add new basic-blocks to
13327 loops, schedule loop fixup.
13328 * cfgloop.c (verify_loop_structure): Check that looks are not
13329 marked for fixup.
13330 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
13331 to loops.
13332 (expand_omp_for_generic): Likewise.
13333 (expand_omp_sections): Likewise.
13334 (expand_omp_atomic_pipeline): Schedule loops for fixup.
13335 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
13336 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
13337 is initialized, not when loops are present.
13338 * tree-parloops.c (parallelize_loops): Remove checking here.
13339 * passes.c (init_optimization_passes): Schedule a copy-propagation
13340 pass before complete unrolling of inner loops.
13341
e78e8a0b
JJ
133422013-04-26 Jakub Jelinek <jakub@redhat.com>
13343
a2e836b2
JJ
13344 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
13345 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
13346 (colorize_init): Add argument to _WIN32 version.
13347 * toplev.c: Include diagnostic-color.h.
13348 (process_options): Default to -fdiagnostics-color=auto if
13349 GCC_COLORS env var is in the environment.
13350 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB
13351 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
13352 env var is in the environment, the default is auto rather than never.
a2e836b2 13353
e78e8a0b
JJ
13354 * diagnostic.h (file_name_as_prefix): Add context argument.
13355 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
13356 the string as locus.
13357 * langhooks.c (lhd_print_error_function): Adjust caller.
13358
013e5ef9
LC
133592013-04-25 Lawrence Crowl <crowl@google.com>
13360
13361 * var-tracking.c (shared_hash_def::htab):
13362 Change type to hash_table. Update dependent calls and types.
13363
4a8fb1a1
LC
133642013-04-25 Lawrence Crowl <crowl@google.com>
13365
13366 * Makefile.in: Update as needed below.
13367
13368 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
13369 Move declaration to after the type's method definitons.
13370
13371 * attribs.c (htab_t scoped_attributes::attribute_hash):
13372 Change type to hash_table. Update dependent calls and types.
13373
13374 * bitmap.c (htab_t bitmap_desc_hash):
13375 Change type to hash_table. Update dependent calls and types.
13376
13377 * cselib.c (htab_t cselib_hash_table):
13378 Change type to hash_table. Update dependent calls and types.
13379
13380 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
13381 (hash_string_slot_node): Move implementation into lto-streamer.h
13382 struct string_slot_hasher.
13383 (eq_string_slot_node): Likewise.
13384
13385 * data-streamer-out.c: Update output_block::string_hash_table
13386 dependent calls and types.
13387
13388 * dwarf2cfi.c (htab_t trace_index):
13389 Change type to hash_table. Update dependent calls and types.
13390
13391 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
13392 Change type to hash_table. Update dependent calls and types.
13393 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
13394 (htab_t optimize_external_refs::map): Likewise.
13395 (htab_t output_comp_unit::extern_map): Likewise.
13396 (htab_t output_comdat_type_unit::extern_map): Likewise.
13397 (htab_t output_macinfo::macinfo_htab): Likewise.
13398 (htab_t optimize_location_lists::htab): Likewise.
13399 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13400
13401 * except.c (htab_t ehspec_hash_type):
13402 Change type to hash_table. Update dependent calls and types.
13403 (assign_filter_values::ttypes): Likewise.
13404 (assign_filter_values::ehspec): Likewise.
13405 (sjlj_assign_call_site_values::ar_hash): Likewise.
13406 (convert_to_eh_region_ranges::ar_hash): Likewise.
13407
13408 * gcse.c (htab_t pre_ldst_table):
13409 Change type to hash_table. Update dependent calls and types.
13410
13411 * ggc-common.c (htab_t saving_htab):
13412 Change type to hash_table. Update dependent calls and types.
13413 (htab_t loc_hash): Likewise.
13414 (htab_t ptr_hash): Likewise.
13415 (call_count): Rename ggc_call_count.
13416 (call_alloc): Rename ggc_call_alloc.
13417 (loc_descriptor): Rename make_loc_descriptor.
13418 (add_statistics): Rename ggc_add_statistics.
13419
13420 * ggc-common.c (saving_htab):
13421 Change type to hash_table. Update dependent calls and types.
13422
13423 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13424 (push_gimplify_context): Likewise.
13425 (pop_gimplify_context): Likewise.
13426 (struct gimple_temp_hash_elt): Added.
13427 (struct gimplify_hasher): Likewise.
13428 (struct gimplify_ctx.temp_htab):
13429 Change type to hash_table. Update dependent calls and types.
13430
13431 * gimple-fold.c: Include gimplify-ctx.h.
13432
13433 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13434 Change type to hash_table. Update dependent calls and types.
13435 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13436 avoid potential global name collision.
13437
13438 * gimplify.c: Include gimplify-ctx.h.
13439 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13440 (htab_t gimplify_ctx::temp_htab):
13441 Update dependent calls and types for new type hash_table.
13442 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13443 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13444
13445 * gimplify-ctx.h: New.
13446 (struct gimple_temp_hash_elt): Move from gimplify.c.
13447 (class gimplify_hasher): New.
13448 (struct gimplify_ctx): Move from gimple.h.
13449 (htab_t gimplify_ctx::temp_htab):
13450 Change type to hash_table. Update dependent calls and types.
13451
13452 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13453 (htab_t ivs_params::newivs_index):
13454 Change type to hash_table. Update dependent calls and types.
13455 (htab_t ivs_params::params_index): Likewise.
13456 (htab_t print_generated_program::params_index): Likewise.
13457 (htab_t gloog::newivs_index): Likewise.
13458 (htab_t gloog::params_index): Likewise.
13459
13460 * graphite.c: Include graphite-htab.h.
13461 4htab_t graphite_transform_loops::bb_pbb_mapping):
13462 Change type to hash_table. Update dependent calls and types.
13463
13464 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13465 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13466 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13467
13468 * graphite-dependences.c: Include graphite-htab.h.
13469 (loop_is_parallel_p): Change hash table type of parameter.
13470
13471 * graphite-htab.h: New.
13472 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13473 (extern find_pbb_via_hash): Move from graphite-poly.h.
13474 (extern loop_is_parallel_p): Move from graphite-poly.h.
13475 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13476
13477 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13478 (extern loop_is_parallel_p): Move to graphite-htab.h.
13479 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13480
13481 * haifa-sched.c (htab_t delay_htab):
13482 Change type to hash_table. Update dependent calls and types.
13483 (htab_t delay_htab_i2): Likewise.
13484
13485 * ira-color.c (htab_t allocno_hard_regs_htab):
13486 Change type to hash_table. Update dependent calls and types.
13487
13488 * ira-costs.c (htab_t cost_classes_htab):
13489 Change type to hash_table. Update dependent calls and types.
13490
13491 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13492 Change type to hash_table. Update dependent calls and types.
13493
13494 * loop-iv.c (htab_t bivs):
13495 Change type to hash_table. Update dependent calls and types.
13496
13497 * loop-unroll.c (htab_t opt_info::insns_to_split):
13498 Change type to hash_table. Update dependent calls and types.
13499 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13500
13501 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13502 (struct string_slot_hasher): New.
13503 (htab_t output_block::string_hash_table):
13504 Change type to hash_table. Update dependent calls and types.
13505
13506 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13507 (htab_t file_name_hash_table):
13508 Change type to hash_table. Update dependent calls and types.
13509
13510 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13511 calls and types.
13512
13513 * lto-streamer.c (htab_t tree_htab):
13514 Change type to hash_table. Update dependent calls and types.
13515
13516 * omp-low.c: Include gimplify-ctx.h.
13517
13518 * passes.c (htab_t name_to_pass_map):
13519 Change type to hash_table. Update dependent calls and types.
13520 (pass_traverse): Rename to passes_pass_traverse.
13521
13522 * plugin.c (htab_t event_tab):
13523 Change type to hash_table. Update dependent calls and types.
13524
13525 * postreload-gcse.c (htab_t expr_table):
13526 Change type to hash_table. Update dependent calls and types.
13527 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13528
13529 * sese.c (debug_rename_map_1): Make extern.
13530 (htab_t copy_bb_and_scalar_dependences::rename_map):
13531 Change type to hash_table. Update dependent calls and types.
13532
13533 * sese.h (extern debug_rename_map): Move to .c file.
13534
13535 * store-motion.c (htab_t store_motion_mems_table):
13536 Change type to hash_table. Update dependent calls and types.
13537
13538 * trans-mem.c (htab_t tm_new_mem_hash):
13539 Change type to hash_table. Update dependent calls and types.
13540
13541 * tree-browser.c (htab_t TB_up_ht):
13542 Change type to hash_table. Update dependent calls and types.
13543
13544 * tree-cfg.c (htab_t discriminator_per_locus):
13545 Change type to hash_table. Update dependent calls and types.
13546
13547 * tree-complex.c: Include tree-hasher.h
13548 (htab_t complex_variable_components):
13549 Change type to hash_table. Update dependent calls and types.
13550
13551 * tree-eh.c (htab_t finally_tree):
13552 Change type to hash_table. Update dependent calls and types.
13553
13554 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13555 struct int_tree_hasher.
13556 (extern int_tree_map_eq): Likewise.
13557 (uid_decl_map_hash): Removed.
13558 (extern decl_tree_map_eq): Likewise.
13559
13560 * tree-hasher.h: New.
13561 (struct int_tree_hasher): New.
13562 (typedef int_tree_htab_type): New.
13563
13564 * tree-inline.c: Include gimplify-ctx.h.
13565
13566 * tree-mudflap.c: Include gimplify-ctx.h.
13567
13568 * tree-parloops.c: Include tree-hasher.h.
13569 (htab_t eliminate_local_variables_stmt::decl_address):
13570 Change type to hash_table. Update dependent calls and types.
13571 (htab_t separate_decls_in_region::decl_copies): Likewise.
13572
13573 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13574 Change type to hash_table. Update dependent calls and types.
13575
13576 * tree-sra.c (candidates):
13577 Change type to hash_table. Update dependent calls and types.
13578
13579 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13580 in tree-flow.h.
13581 (int_tree_map_hash): Likewise.
13582
13583 * tree-ssa-dom.c (htab_t avail_exprs):
13584 Change type to hash_table. Update dependent calls and types.
13585
13586 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13587 Change type to hash_table. Update dependent calls and types.
13588
13589 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
13590 Change type to hash_table. Update dependent calls and types.
13591
13592 * tree-ssa-phiopt.c (seen_ssa_names):
13593 Change type to hash_table. Update dependent calls and types.
13594
13595 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
13596 Change type to hash_table. Update dependent calls and types.
13597
13598 * tree-ssa-uncprop.c (equiv):
13599 Change type to hash_table. Update dependent calls and types.
13600
c5a44004
JJ
136012013-04-25 Jakub Jelinek <jakub@redhat.com>
13602
13603 PR rtl-optimization/57003
13604 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
13605 call note_stores with kill_clobbered_value callback again after
13606 killing regs_invalidated_by_call.
13607
09962a4a
JG
136082013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13609
13610 * config/aarch64/aarch64-simd.md
13611 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
13612 (aarch64_simd_bsl<mode>): Likewise.
13613 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
13614
ea28bb0b
MP
136152013-04-25 Marek Polacek <polacek@redhat.com>
13616
13617 PR tree-optimization/57066
3c21604f 13618 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 13619
96659611
JG
136202013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
13621
13622 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
13623
9697e620
JG
136242013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13625
13626 * config/aarch64/aarch64-builtins.c
13627 (aarch64_fold_builtin): New.
13628 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
13629 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
13630 * config/aarch64/aarch64-simd-builtins.def (abs): New.
13631 * config/aarch64/arm_neon.h
13632 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
13633
0ac198d3
JG
136342013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13635 Tejas Belagod <tejas.belagod@arm.com>
13636
13637 * config/aarch64/aarch64-builtins.c
13638 (aarch64_gimple_fold_builtin): New.
13639 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
13640 * config/aarch64/aarch64-simd-builtins.def (addv): New.
13641 * config/aarch64/aarch64-simd.md (addpv4sf): New.
13642 (addvv4sf): Update.
13643 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
13644
58cff58c
N
136452013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13646
df93505e 13647 * config/aarch64/aarch64.md
58cff58c
N
13648 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
13649
7e0228bf
N
136502013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13651
13652 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
13653 (*ngcsi_uxtw): New pattern.
13654
5819f96f 136552013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 13656 Julian Brown <julian@codesourcery.com>
5819f96f
KT
13657
13658 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
13659 (TB_DREG): Add T_V4HF.
13660 (v4hf_UP): New macro.
13661 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 13662 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
13663 Handle initialisation of V4HF. Adjust initialisation of reinterpret
13664 built-ins.
df93505e 13665 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
13666 (arm_vector_mode_supported_p): Handle V4HF.
13667 (arm_mangle_map): Handle V4HFmode.
13668 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
13669 * config/arm/arm_neon_builtins.def: Add entries for
13670 vcvtv4hfv4sf, vcvtv4sfv4hf.
13671 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
13672 (neon_vcvtv4hfv4sf): Likewise.
13673 * config/arm/neon-gen.ml: Handle half-precision floating point
13674 features.
13675 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
13676 * config/arm/arm_neon.h: Regenerate.
13677 * config/arm/neon.ml (type elts): Add F16.
13678 (type vectype): Add T_float16x4, T_floatHF.
13679 (type vecmode): Add V4HF.
13680 (type features): Add Requires_FP_bit feature.
13681 (elt_width): Handle F16.
13682 (elt_class): Likewise.
13683 (elt_of_class_width): Likewise.
13684 (mode_of_elt): Refactor.
13685 (type_for_elt): Handle F16, fix error messages.
13686 (vectype_size): Handle T_float16x4.
13687 (vcvt_sh): New function.
13688 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
13689 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
13690 (string_of_mode): Handle V4HF.
13691 * doc/arm-neon-intrinsics.texi: Regenerate.
13692
1ef395e4
JG
136932013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13694
13695 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
13696 format specifier in 'X' case.
13697
41c34e94
AM
136982013-04-25 Alan Modra <amodra@gmail.com>
13699
13700 PR target/57052
13701 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
13702 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
13703 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
13704 Repeat for many other rotate/shift and mask patterns using subregs.
13705 Name lshiftrt insns.
13706 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
13707 on WORDS_BIG_ENDIAN.
13708
b9a7eb5d
AM
137092013-04-25 Alan Modra <amodra@gmail.com>
13710
13711 * config.gcc: Support little-endian powerpc-linux targets.
13712 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
13713 (LINK_OS_LINUX_SPEC): Define.
13714 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
13715 Preserve MASK_LITTLE_ENDIAN.
13716 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
13717 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
13718 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
13719 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
13720 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
13721 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
13722 Correct fp word order for little-endian. Don't shift toc entries
13723 smaller than a word for little-endian.
13724 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
13725 (bswapdi2 splits): Correct low-part subreg for little-endian.
13726 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
13727 low/high where such is correct only for be.
13728 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
13729 little-endian for -mcall-aixdesc.
13730
87f73374
AM
137312013-04-25 Alan Modra <amodra@gmail.com>
13732
13733 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
13734 replace_equiv_address_nv.
13735
cabf91cd
AM
137362013-04-25 Alan Modra <amodra@gmail.com>
13737
13738 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
13739
0ae24cc8
VM
137402013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13741
13742 Revert:
13743 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13744 * rtl.h (struct rtx_def): ...
cabf91cd 13745
77bce07c
VM
137462013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13747
13748 PR rtl-optimizations/57046
13749 * lra-constraints (split_reg): Set up lra_risky_transformations_p
13750 for multi-reg splits.
13751
0db63e7f
L
137522013-04-24 H.J. Lu <hongjiu.lu@intel.com>
13753
13754 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
13755
3362b6b6
SA
137562013-04-24 Sterling Augustine <saugustine@google.com>
13757
13758 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
13759 (comp_dir_string, debug_str_dwo_section): New.
13760 (DEBUG_STR_DWO_SECTION): Rename to ...
13761 (DEBUG_DWO_STR_SECTION): ... this.
13762 (DEBUG_NORM_STR_SECTION): Delete.
13763 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
13764 (DEBUG_STR_DWO_SECTION_FLAGS): New.
13765 (find_AT_string): Move most logic to ...
13766 (find_AT_string_in_table): ... here. New.
13767 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
13768 add_skeleton_AT_string. Delete logic.
13769 (output_skeleton_debug_sections): Remove call to
13770 add_top_level_skeleton_die_attrs.
13771 (add_comp_dir_attribute): Move logic to comp_dir_string.
13772 (dwarf2out_init): Initialize debug_str_dwo_section.
13773 (output_indirect_string): Call find_string_form.
13774 (output_indirect_strings): Rewrite.
13775 (prune_unused_types): Empty skeleton_debug_str_hash.
13776 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
13777 (dwarf2out_finish): Call output_indirect_strings.
13778
e93e18e9
PC
137792013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13780
13781 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
13782
f6ce35ac
VM
137832013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13784
cabf91cd 13785 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM
13786 (LRA_SUBREG_P): New macro.
13787 * recog.c (register_operand): Check LRA_SUBREG_P.
13788 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13789 stack frame.
13790 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13791 (lra_final_code_change): Skip subreg change for operators.
13792 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13793 if there are no operand changes.
13794 * lra-constraints.c (curr_insn_set): New.
13795 (match_reload): Set LRA_SUBREG_P.
13796 (emit_spill_move): Ditto.
13797 (check_and_process_move): Use curr_insn_set. Process only single
13798 set insns. Don't initialize sec_mem_p and change_p.
13799 (simplify_operand_subreg): Use LRA_SUBREG_P.
13800 (reg_in_class_p): New function.
13801 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13802 of #ifdef. Add code to remove cycling.
13803 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13804 non-null disp. Reload inner instead of disp when base and index
13805 are null. Try to put lo_sum into register.
13806 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 13807 (check_and_process_move): Move code for move cost check to
f6ce35ac 13808 simple_move_p. Remove equiv_substitution.
cabf91cd 13809 (simple_move_p): New function.
f6ce35ac
VM
13810 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13811 curr_insn_set. Call check_and_process_move only for single set
13812 insns. Use the new function. Move call of check_and_process_move
13813 after operand equiv substitution and address process.
13814
38047d90
JG
138152013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
13816
13817 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
13818 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
13819 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
13820
13f39b2e
PC
138212013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13822
13823 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
13824
97c116dc
MP
138252013-04-24 Marek Polacek <polacek@redhat.com>
13826
13827 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
13828 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
13829 (select_loops_exit_conditions): Likewise.
13830 (number_of_iterations_for_all_loops): Likewise.
13831 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
13832 (scev_analysis): Likewise.
13833
83082391 138342013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 13835 Chao-ying Fu <fu@mips.com>
83082391 13836
cabf91cd
AM
13837 * config/mips/micromips.md (jraddiusp): New pattern.
13838 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
13839 instruction if possible.
83082391 13840
19e34aa2
AM
138412013-04-24 Alan Modra <amodra@gmail.com>
13842
13843 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
13844
fdb6603c
JB
138452013-04-24 Julian Brown <julian@codesourcery.com>
13846 Chung-Lin Tang <cltang@codesourcery.com>
13847
13848 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB
13849 dependency behavior in enumeration type DIE generation. Add TODO note
13850 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 13851
bf190e8d
LC
138522013-04-23 Lawrence Crowl <crowl@google.com>
13853
13854 * Makefile.in: Update as needed below.
13855
13856 * hash-table.h (class hash_table):
13857 Correct many methods with parameter types compare_type to the correct
13858 value_type. (Correct code was unlikely to notice the change.)
13859 (hash_table::elements_with_deleted) New.
13860 (class hashtable::iterator): New.
13861 (hashtable::begin()): New.
13862 (hashtable::end()): New.
13863 (FOR_EACH_HASH_TABLE_ELEMENT): New.
13864
13865 * statistics.c (statistics_hashes):
13866 Change type to hash_table. Update dependent calls and types.
13867
13868 * tree-into-ssa.c (var_infos):
13869 Change type to hash_table. Update dependent calls and types.
13870
13871 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
13872 Change type to hash_table. Update dependent calls and types.
13873
13874 * tree-ssa-loop-im.c (struct mem_ref.refs):
13875 Change type to hash_table. Update dependent calls and types.
13876
13877 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
13878 Change type to hash_table. Update dependent calls and types.
13879
13880 * tree-ssa-sccvn.c (vn_tables_s::nary):
13881 Change type to hash_table. Update dependent calls and types.
13882 (vn_tables_s::phis): Likewise.
13883 (vn_tables_s::references): Likewise.
13884
13885 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
13886 (vn_reference_eq): Update parameter and return types.
13887
13888 * tree-ssa-structalias.c (pointer_equiv_class_table):
13889 Change type to hash_table. Update dependent calls and types.
13890 (location_equiv_class_table): Likewise.
13891
13892 * tree-vect-data-refs.c: Consequential changes for making
13893 peeling a hash_table.
13894
13895 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
13896 (destroy_loop_vec_info): Dependent hash_table update.
13897
13898 * tree-vectorizer.h (peeling_htab):
13899 Change type to hash_table. Update dependent calls and types.
13900
d70a81dd
SC
139012013-04-23 Shiva Chen <shiva0217@gmail.com>
13902
cabf91cd
AM
13903 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
13904 to check the register content is equal or not.
13905 * lra-constraints.c (match_reload): Use lra_assign_reg_val
13906 to assign register content record.
13907 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 13908 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM
13909 * lra-int.h (struct lra_reg): Add offset member.
13910 (lra_reg_val_equal_p): New static inline function.
13911 (lra_update_reg_val_offset): New static inline function.
13912 (lra_assign_reg_val): New static inline function.
13913 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
13914 to assign register content record.
13915 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 13916
b894a1f3
CM
139172013-04-23 Catherine Moore <clm@codesourcery.com>
13918
13919 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
13920 operands. Record compression.
13921
ecd14de9
XDL
139222013-04-23 Xinliang David Li <davidxl@google.com>
13923
13924 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
13925
92e776e9
RB
139262013-04-23 Richard Biener <rguenther@suse.de>
13927
13928 PR middle-end/57036
13929 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
13930 parameter, only add abnormal goto edges from the copied body
13931 if the call could perform abnormal gotos.
13932 (copy_cfg_body): Adjust.
13933
a15ee567
SN
139342013-04-23 Sofiane Naci <sofiane.naci@arm.com>
13935
13936 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
13937
08c52234
AS
139382013-04-23 Andreas Schwab <schwab@linux-m68k.org>
13939
13940 * coretypes.h (gimple_stmt_iterator): Add struct to make
13941 compatible with C.
13942
999c1171
RB
139432013-04-23 Richard Biener <rguenther@suse.de>
13944
13945 PR tree-optimization/57026
13946 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
13947 from SSA names occuring in abnormal PHI nodes.
13948
53669259
AK
139492013-04-22 Andi Kleen <ak@linux.intel.com>
13950
13951 * lto/lto.c (print_lto_report_1): Fix LTO report names.
13952
1a0ad150
AK
139532013-04-22 Andi Kleen <ak@linux.intel.com>
13954
13955 * lto/lto.c (print_lto_report_1): Declare early.
13956 (read_cgraph_and_symbols): Call print_lto_report_1 early.
13957
057f8f20
AK
139582013-04-22 Andi Kleen <ak@linux.intel.com>
13959
13960 * common.opt (-flto-report-wpa): Add.
13961 * doc/invoke.texi (-flto-report-wpa): Add.
13962 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
13963 (lto_main): dito.
13964
473b1e05
XDL
139652013-04-22 Xinliang David Li <davidxl@google.com>
13966
13967 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
13968 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
13969 * Makefile.in: New dependency
13970
cabf91cd 13971 David Daney <ddaney.cavm@gmail.com>
b1485a33 13972
cabf91cd
AM
13973 * configure.ac (gcc_cv_as_micromips_support): Use the
13974 --fatal-warnings option.
13975 * configure: Regenerate.
b1485a33 13976
829d0168
MP
139772013-04-22 Marek Polacek <polacek@redhat.com>
13978
13979 PR sanitizer/56990
13980 * tsan.c (instrument_expr): Don't instrument expression
13981 in case its size is zero.
13982
6d9b7208
UB
139832013-04-22 Uros Bizjak <ubizjak@gmail.com>
13984
13985 PR target/57032
13986 Revert:
13987 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13988
13989 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13990
ea679d55
JG
139912013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13992
13993 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
13994 (gimple_stmt_iterator): New typedef.
13995 * gimple.h (gimple_stmt_iterator): Rename to...
13996 (gimple_stmt_iterator_d): ... This.
13997 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
13998 trees be valid for GIMPLE and GENERIC.
13999 (TARGET_GIMPLE_FOLD_BUILTIN): New.
14000 * gimple-fold.c (gimple_fold_call): Call target hook
14001 gimple_fold_builtin.
14002 * hooks.c (hook_bool_gsiptr_false): New.
14003 * hooks.h (hook_bool_gsiptr_false): New.
14004 * target.def (fold_stmt): New.
14005 * doc/tm.texi: Regenerate.
14006
88a581da
VM
140072013-04-22 Vladimir Makarov <vmakarov@redhat.com>
14008
14009 PR target/57018
14010 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
14011 a set sp if no stack realignment.
14012
92be22dc
NC
140132013-04-22 Nick Clifton <nickc@redhat.com>
14014
14015 * config.gcc (tilegx-linux): Extend extra_objs rather than
14016 overwriting it.
14017 (tilepro-linux): Likewise.
14018
0ddec79f
JG
140192013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14020
14021 * config/aarch64/aarch64-builtins.c
14022 (CF): Remove.
14023 (CF0, CF1, CF2, CF3, CF4, CF10): New.
14024 (VAR<1-12>): Add MAP parameter.
14025 (BUILTIN_*): Likewise.
14026 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
14027 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
14028 (aarch64_ushl_n<mode>): Likewise.
14029 (aarch64_sshr_n<mode>): Likewise.
14030 (aarch64_ushr_n<mode>): Likewise.
14031 (aarch64_<maxmin><mode>): Likewise.
14032 (aarch64_sqrt<mode>): Likewise.
14033 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
14034 (vshr<q>_n_*): Likewise.
14035
0050faf8
JG
140362013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14037
14038 * config/aarch64/aarch64-builtins.c
14039 (aarch64_simd_builtin_type_mode): Handle SF types.
14040 (sf_UP): Define.
14041 (BUILTIN_GPF): Define.
14042 (aarch64_init_simd_builtins): Handle SF types.
14043 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
14044 (frecps): Likewise.
14045 (frecpx): Likewise.
14046 * config/aarch64/aarch64-simd.md
14047 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
14048 (aarch64_frecpe<mode>): New.
14049 (aarch64_frecps<mode>): Likewise.
14050 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
14051 (v8type): Add frecp<esx>.
14052 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
14053 (aarch64_frecps<mode>): Likewise.
14054 * config/aarch64/iterators.md (FRECP): New.
14055 (frecp_suffix): Likewise.
14056 * config/aarch64/arm_neon.h
14057 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
14058
0fad3dbc 140592013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB
14060
14061 PR target/56995
14062 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
14063 (REG_CLASS_NAMES): Idem.
14064 (REG_CLASS_CONTENTS): Idem.
14065 (REGCLASS_HAS_FP_REG): Idem.
14066 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
14067 (sh_conditional_register_usage): Idem.
14068
3e8a33f9
JL
140692013-04-21 Jeff Law <law@redhat.com>
14070
14071 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
14072 (ssa_forward_propagate_and_combine): Use it.
14073
f38e1b0a
VM
140742013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14075
14076 * lra.c: Update the flow chart diagram.
14077
682303da
VM
140782013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14079
14080 PR rtl-optimization/56847
14081 * lra-constraints.c (process_alt_operands): Discourage alternative
14082 with non-matche doffsettable memory constraint fro memory with
14083 known offset.
14084
f6b64c35
RB
140852013-04-19 Richard Biener <rguenther@suse.de>
14086
14087 PR tree-optimization/56982
14088 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
14089 function.
14090 * gimplify.c (gimplify_call_expr): Notice special calls.
14091 (gimplify_modify_expr): Likewise.
14092 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
14093 abnormal control flow receivers.
14094 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
14095 in the same way as cfun->has_nonlocal_labels.
14096 (gimple_purge_dead_abnormal_call_edges): Likewise.
14097 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
14098 receivers start a basic-block.
14099
01d8bf07
RB
141002013-04-19 Richard Biener <rguenther@suse.de>
14101
14102 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
14103 member ...
14104 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
14105 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
14106 (SLP_TREE_LOAD_PERMUTATION): Add.
14107 (vect_transform_slp_perm_load): Adjust prototype.
14108 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
14109 (vect_free_slp_instance): Likewise.
14110 (vect_create_new_slp_node): Likewise.
14111 (vect_supported_slp_permutation_p): Remove.
14112 (vect_slp_rearrange_stmts): Adjust.
14113 (vect_supported_load_permutation_p): Likewise. Inline
14114 vect_supported_slp_permutation_p here.
14115 (vect_analyze_slp_instance): Compute load permutations per
14116 slp node instead of per instance.
14117 (vect_get_slp_defs): Adjust.
14118 (vect_transform_slp_perm_load): Likewise.
14119 (vect_schedule_slp_instance): Remove redundant code.
14120 (vect_schedule_slp): Remove hack for PR56270, add it ...
14121 * tree-vect-stmts.c (vectorizable_load): ... here, do not
14122 CSE loads for SLP. Adjust.
14123
ede22fc3
GY
141242013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14125
14126 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
14127 spelling in two comments.
14128
67bc84fb
GY
141292013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14130
14131 PR target/56797
14132 * config/arm/arm.c (load_multiple_sequence): Require SP
14133 as base register for loads if SP is in the register list.
14134
e248d83f
MJ
141352013-04-19 Martin Jambor <mjambor@suse.cz>
14136
14137 PR tree-optimization/56718
14138 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
14139 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
14140 and made public. Adjusted all callers.
14141 (ipa_intraprocedural_devirtualization): New function.
14142 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
14143 (ipa_intraprocedural_devirtualization): Likewise.
14144 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
14145
4891e8f8
RB
141462013-04-19 Richard Biener <rguenther@suse.de>
14147
14148 PR tree-optimization/57000
14149 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
14150
dad89f7c
TG
141512013-04-19 Terry Guo <terry.guo@arm.com>
14152
14153 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
14154 Replace with ...
14155 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
14156 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
14157 (cortex_m4_fmacs): Use new reservations.
14158 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
14159
72ea0d47
VM
141602013-04-18 Vladimir Makarov <vmakarov@redhat.com>
14161
f1e6512c 14162 PR rtl-optimization/56999
72ea0d47
VM
14163 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
14164 related code.
14165 (lra_coalesce): Remove split_origin_bitmap and related code.
14166 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
14167 ranges if necessary.
14168
780a5b71
UB
141692013-04-18 Uros Bizjak <ubizjak@gmail.com>
14170
14171 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
14172 New array.
14173 (ix86_expand_call): Remove clobbered_registers array and use
14174 x86_64_ms_sysv_extra_clobbered_registers instead.
14175 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
14176 Declare here.
14177 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
14178 predicate.
14179 * config/i386/i386.md (*call_rex64_ms_sysv): Use
14180 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
14181 (*call_value_rex64_ms_sysv): Ditto.
14182
6f5a366a
CC
141832013-04-18 Cary Coutant <ccoutant@google.com>
14184
14185 * dwarf2out.c (output_pubnames): Check die_perennial_p of
14186 parent instead of die_mark.
14187
475b8f37
DN
141882013-04-18 Diego Novillo <dnovillo@google.com>
14189
14190 * gimple.c (create_gimple_tmp): New.
14191 (get_expr_type): New.
14192 (build_assign): New.
14193 (build_type_cast): New.
14194 * gimple.h (enum ssa_mode): Define.
14195 (gimple_seq_set_location): New.
14196 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 14197 to use build_assign and build_type_cast.
475b8f37 14198
08940f33
RB
141992013-04-18 Richard Biener <rguenther@suse.de>
14200
14201 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
14202 handle negative step. Remove redundant checks.
14203 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
14204 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
14205 for negative step and grouped loads fail to vectorize.
14206
0e0f87d4
SB
142072013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14208
14209 * emit-rtl.c (reset_insn_used_flags): New function.
14210 (reset_all_used_flags): Use it.
14211 (verify_insn_sharing): New function.
14212 (verify_rtl_sharing): Fix verification for SEQUENCEs.
14213
4c445590
JJ
142142013-04-18 Jakub Jelinek <jakub@redhat.com>
14215
14216 PR tree-optimization/56984
14217 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
14218 and (x >> M) >= N don't register any assertion if N << M is the
14219 minimum value.
14220
6873ecab
SB
142212013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14222
14223 * lower-subreg.c (resolve_simple_move): If called self-recursive,
14224 do not delete_insn insns that have not yet been emitted, only
14225 unlink them with remove_insn.
14226 * df-scan.c (df_insn_delete): Revert r197492.
14227
3ccb989e
SB
142282013-04-17 Steven Bosscher <steven@gcc.gnu.org>
14229
14230 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
14231 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
14232
142332013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY
14234
14235 * config/arm/arm.md (movsicc_insn): Convert define_insn into
14236 define_insn_and_split.
14237 (and_scc,ior_scc,negscc): Likewise.
14238 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14239
3ccb989e 142402013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY
14241
14242 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14243 can benefit from using a sequence of LDRD instructions in epilogue
14244 instead of a single LDM instruction.
14245
6d10a203
MLI
142462013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14247
14248 PR 45688
14249 * doc/extend.texi: Fix typo.
14250
6983e6b5
RB
142512013-04-17 Richard Biener <rguenther@suse.de>
14252
14253 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14254 (vect_build_slp_tree): ... here.
14255 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14256 match. Remove special-casing of mismatched complex loads.
14257 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14258 re-try the match with swapped commutative operands.
14259 (vect_supported_load_permutation_p): Remove special-casing of
14260 mismatched complex loads.
14261 (vect_analyze_slp_instance): Adjust.
14262
ef23e6a2
RB
142632013-04-17 Richard Biener <rguenther@suse.de>
14264
14265 PR rtl-optimization/56921
14266 * cfgloop.h (struct loop): Add simple_loop_desc member.
14267 (struct niter_desc): Mark with GTY(()).
14268 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14269 * loop-iv.c (get_simple_loop_desc): Likewise.
14270 (free_simple_loop_desc): Likewise.
14271
14272 Revert
14273 2013-04-16 Richard Biener <rguenther@suse.de>
14274
14275 PR rtl-optimization/56921
14276 * loop-init.c (pass_rtl_move_loop_invariants): Add
14277 TODO_do_not_ggc_collect to todo_flags_finish.
14278 (pass_rtl_unswitch): Same.
14279 (pass_rtl_unroll_and_peel_loops): Same.
14280 (pass_rtl_doloop): Same.
14281
fc6f94f5
EB
142822013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14283
14284 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14285 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14286 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14287 references.
14288 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14289 * tree-streamer.c (record_common_node): Adjust reference in comment.
14290
10a88311
TG
142912013-04-17 Terry Guo <terry.guo@arm.com>
14292
14293 * config/arm/cortex-m4.md: Add a new bypass.
14294
6d9b7208 142952013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N
14296
14297 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14298 New pattern.
14299 (*subs_<optab><mode>_multp2): New pattern.
14300 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14301 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14302
6d9b7208 143032013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N
14304
14305 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14306 (*subs_mul_imm_<mode>): New pattern.
14307
18a6701e
DE
143082013-04-16 David Edelsohn <dje.gcc@gmail.com>
14309
14310 PR target/56948
14311 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14312 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14313 (vsx_movti_32bit): Change j->wa to O->wa.
14314
07c37b2f
RB
143152013-04-16 Richard Biener <rguenther@suse.de>
14316
14317 PR rtl-optimization/56921
14318 * loop-init.c (pass_rtl_move_loop_invariants): Add
14319 TODO_do_not_ggc_collect to todo_flags_finish.
14320 (pass_rtl_unswitch): Same.
14321 (pass_rtl_unroll_and_peel_loops): Same.
14322 (pass_rtl_doloop): Same.
14323
0e0f87d4 143242013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY
14325
14326 * config/arm/arm.c (emit_multi_reg_push): New declaration
14327 for an existing function.
14328 (arm_emit_strd_push): New function.
14329 (arm_expand_prologue): Used here.
14330 (arm_emit_ldrd_pop): New function.
14331 (arm_expand_epilogue): Used here.
14332 (arm_get_frame_offsets): Update condition.
14333 (arm_emit_multi_reg_pop): Add a special case for load of a single
14334 register with writeback.
14335
5e8e2af4
UB
143362013-04-16 Uros Bizjak <ubizjak@gmail.com>
14337
14338 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
14339 description.
14340
9fd9ccf7
RB
143412013-04-16 Richard Biener <rguenther@suse.de>
14342
14343 PR tree-optimization/56756
14344 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
14345 (first_mem_ref_loc): New.
14346 (execute_sm): Place the load temporarily before a previous
14347 access instead of in the latch edge to ensure its SSA dependencies
14348 are defined at points dominating the load.
14349
96fba521
SB
143502013-04-16 Steven Bosscher <steven@gcc.gnu.org>
14351
4c8af858
SB
14352 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
14353 correct fix by moving header and footer insn to the footer of
14354 the merged basic block. Clear BB_END of the merged-away block.
14355
96fba521
SB
14356 PR middle-end/43631
14357 * emit-rtl.c (make_note_raw): New function.
14358 (link_insn_into_chain): New static inline function.
14359 (add_insn): Use it.
14360 (add_insn_before, add_insn_after): Factor insn chain linking code...
14361 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
14362 using link_insn_into_chain.
14363 (note_outside_basic_block_p): New helper function for emit_note_after
14364 and emit_note_before.
14365 (emit_note_after): Use nobb variant of add_insn_after if the note
14366 should not be contained in a basic block.
14367 (emit_note_before): Use nobb variant of add_insn_before if the note
14368 should not be contained in a basic block.
14369 (emit_note_copy): Use make_note_raw.
14370 (emit_note): Likewise.
14371 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
14372 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
14373 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
14374 the moved barrier the tail of the basic block it follows.
14375 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
14376
7b8265ba
JJ
143772013-04-15 Jakub Jelinek <jakub@redhat.com>
14378
14379 PR tree-optimization/56962
14380 * gimple-ssa-strength-reduction.c (record_increment): Only set
14381 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
14382 either rhs1 or rhs2 is equal to c->base_expr.
14383
5185d248
RB
143842013-04-15 Richard Biener <rguenther@suse.de>
14385
14386 PR tree-optimization/56933
14387 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
14388 member.
14389 (GROUP_READ_WRITE_DEPENDENCE): Remove.
14390 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
14391 * tree-vect-data-refs.c (vect_analyze_group_access): Move
14392 dependence check ...
14393 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14394 ... here.
14395 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14396 GROUP_READ_WRITE_DEPENDENCE.
14397
a24243a0
AK
143982013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14399
14400 * emit-rtl.c (reset_all_used_flags): New function.
14401 (verify_rtl_sharing): Call reset_all_used_flags before and after
14402 performing the checks.
14403
1c50eada
KT
144042013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14405
14406 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14407 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14408 * config/arm/constraints.md (De): New constraint.
14409 * config/arm/neon.md (anddi3_neon): Delete.
14410 (neon_vand<mode>): Expand to standard anddi3 pattern.
14411 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14412 Move earlier in the file.
14413 (neon_inv_logic_op2): Likewise.
14414 (arm_anddi_operand_neon): New predicate.
14415
e927b6ad
RO
144162013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14417
14418 * configure.ac (gcc_cv_ld_as_needed): Set
14419 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14420 Use -z ignore, -z record on *-*-solaris2*.
14421 (HAVE_LD_AS_NEEDED): Update comment.
14422 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14423 * configure: Regenerate.
14424 * config.in: Regenerate.
14425 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14426 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14427 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14428 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14429 equivalents. Fix markup.
14430 * doc/tm.texi: Regenerate.
14431
e0ea8797
AH
144322013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14433
14434 * config/i386/i386.opt: New option mstack-protector-guard=.
14435 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14436 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14437 TARGET_SSP_TLS_GUARD.
14438 * config/i386/i386.c (ix86_option_override_internal): Set
14439 ix86_stack_protector_guard.
14440 * config/i386/i386.md (stack_protect_set): Enable for
14441 TARGET_SSP_TLS_GUARD only.
14442 (stack_protect_set_<mode>): Ditto.
14443 (stack_protect_test): Ditto.
14444 (stack_protect_test_<mode>): Ditto.
14445 * doc/invoke.texi (i386 Option): Document.
14446
811b72f9
EB
144472013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14448
14449 PR target/56890
14450 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14451 (S_MODES): Set H_MODE bit.
14452 (SF_MODES): Set only S_MODE and SF_MODE bits.
14453 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14454 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14455 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14456 <MODE_FLOAT>: Likewise.
14457
5529fdd6
JY
144582013-04-15 Joey Ye <joey.ye@arm.com>
14459
14460 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14461
517b1da2
JY
144622013-04-15 Joey Ye <joey.ye@arm.com>
14463
14464 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14465 for real far jump.
14466 (thumb_far_jump_used_p): Count instruction size and set
14467 far_jump_used.
14468
01007ae0
EB
144692013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14470
14471 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14472 * resource.c (mark_target_live_regs): Reformat conditional block.
14473
c46f6580
SB
144742013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14475
14476 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14477 notes, they are emitted only just before final.
14478 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14479
1f397f45
SB
144802013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14481
14482 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14483 * cfgrtl.c (delete_insn): Call it here instead.
14484 * lra-spills.c (lra_final_code_change): Use delete_insn.
14485 * haifa-sched.c (sched_remove_insn): Likewise.
14486 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14487 returning to the nop pool.
14488 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14489 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14490
58a51369
SB
144912013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14492
14493 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14494 * doc/tm.texi: Regenerated.
14495
33159866
UB
144962013-04-12 Uros Bizjak <ubizjak@gmail.com>
14497
14498 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14499 QImode checks.
14500
226e378f
SB
145012013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14502
14503 * df-core.c (df_find_def): Compare register numbers.
14504 (df_find_use): Likewise.
14505
fafb9b18
VM
145062013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14507
14508 PR target/56903
14509 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14510 lra_in_progress for return.
14511
9a946fd6
GY
145122013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14513
14514 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14515 define_insn into define_insn_and_split and emit movsicc patterns.
14516
41b83758
GY
145172013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14518
14519 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14520
d3afd9aa
RB
145212013-04-12 Richard Biener <rguenther@suse.de>
14522
14523 * tree-pass.h (TODO_do_not_ggc_collect): New.
14524 * passes.c (execute_one_ipa_transform_pass): Honor
14525 TODO_do_not_ggc_collect.
14526 (execute_one_pass): Likewise.
14527
14528 Revert
14529 2013-04-10 Richard Biener <rguenther@suse.de>
14530
14531 * passes.c (init_optimization_passes): Remove reload pass.
14532 * ira.c (do_reload): Merge into ...
14533 (ira): ... this.
14534 (rest_of_handle_reload): Remove.
14535 (pass_reload): Likewise.
14536 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14537 of reload for vzeroupper pass placement.
14538
06f9b387
JJ
145392013-04-12 Jakub Jelinek <jakub@redhat.com>
14540
14541 PR tree-optimization/56918
14542 PR tree-optimization/56920
14543 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14544 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14545 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14546 use wide_mul_with_sign method.
14547
953094d2
RB
145482013-04-12 Richard Biener <rguenther@suse.de>
14549
14550 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14551 not be considered a gimple constant.
14552
bb506982
MG
145532013-04-12 Marc Glisse <marc.glisse@inria.fr>
14554
14555 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14556 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14557 operations.
14558
4b84d650
JJ
145592013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14560 Jakub Jelinek <jakub@redhat.com>
14561
14562 * opts.c: Include diagnostic-color.h.
14563 (common_handle_option): Handle OPT_fdiagnostics_color_.
14564 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14565 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14566 (diagnostic-color.o): New.
14567 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14568 (diagnostic_color_rule): New enum.
14569 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14570 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14571 the location string.
14572 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14573 either NULL, or color kind.
14574 * diagnostic-color.c: New file.
14575 * diagnostic-color.h: New file.
14576 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14577 arguments.
14578 * doc/invoke.texi (-fdiagnostics-color): Document.
14579 * pretty-print.h (pp_show_color): Define.
14580 (struct pretty_print_info): Add show_color field.
14581 * diagnostic.c: Include diagnostic-color.h.
14582 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14583 macros. Colorize error:, warning: etc. strings and also the location
14584 string.
14585 (diagnostic_show_locus): Colorize the caret line.
14586 * pretty-print.c: Include diagnostic-color.h.
14587 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14588 inside of %< %> quotes or quoted through q format modifier.
14589
067a1e71
AK
145902013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14591
33159866 14592 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 14593
33159866 145942013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N
14595
14596 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
14597 code in CC_NZ mode.
14598 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
14599 pattern.
14600
7b55f98f
MP
146012013-04-11 Marek Polacek <polacek@redhat.com>
14602
14603 PR tree-optimization/48184
33159866 14604 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 14605
966b587e
EB
146062013-04-11 Eric Botcazou <ebotcazou@adacore.com>
14607
14608 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
14609 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
14610 (skip_simple_arithmetic): Tidy up.
14611 * tree.h (skip_simple_constant_arithmetic): Declare.
14612
33159866 146132013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N
14614
14615 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
14616
1d42c1ec
RB
146172013-04-11 Richard Biener <rguenther@suse.de>
14618
14619 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14620 generate vector constants.
14621
4ba5ea11
RB
146222013-04-11 Richard Biener <rguenther@suse.de>
14623
14624 PR tree-optimization/56878
14625 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
14626 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
14627 New function.
14628 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
14629 Prefer to align the DR with the most invariant base address.
14630
f0defe58
SKS
146312013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14632
14633 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
14634 comment.
14635
d07458be
JG
146362013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
14637
14638 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
14639 floating-point vector comparisons against 0.
14640
146b8692
JJ
146412013-04-11 Jakub Jelinek <jakub@redhat.com>
14642
14643 PR tree-optimization/56899
14644 * fold-const.c (extract_muldiv_1): Apply distributive law
14645 only if TYPE_OVERFLOW_WRAPS (ctype).
14646
b8578ff7
BC
146472013-04-11 Bin Cheng <bin.cheng@arm.com>
14648
14649 PR target/56124
14650 * ira-costs.c (scan_one_insn): Check whether the source rtx of
14651 loading has side effect.
14652
0ea8a6f9
SB
146532013-04-10 Steven Bosscher <steven@gcc.gnu.org>
14654
14655 * config/sparc/sparc.c: Include tree-pass.h.
14656 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
14657 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
14658 head of file. Change return type. Split off gate function.
14659 (sparc_gate_work_around_errata): New function.
14660 (pass_work_around_errata): New pass definition.
14661 (insert_pass_work_around_errata) New pass insert definition to
14662 insert pass_work_around_errata just after delayed-branch scheduling.
14663 (sparc_option_override): Insert the pass.
14664 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
14665
42e37616
DM
146662013-04-10 David S. Miller <davem@davemloft.net>
14667
89deeb3b
DM
14668 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
14669 or -mcpu=hypersparc.
14670
42e37616
DM
14671 * target.def (cstore_mode): New hook.
14672 * target.h: Include insn-codes.h
14673 * targhooks.c: Likewise.
14674 (default_cstore_mode): New function.
14675 * targhooks.h: Declare it.
14676 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
14677 * doc/tm.texi: Rebuild.
14678 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
14679 target hook, rather than inspecting the insn_data.
14680 * config/sparc/sparc.c (sparc_cstore_mode): New function.
14681 (TARGET_CSTORE_MODE): Redefine.
14682 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
14683 result patterns.
14684 * config/sparc/predicates.md (cstore_result_operand): New special
14685 predicate.
14686 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
14687 Use it for operand 0.
14688 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
14689 (*snesi_special): Likewise.
14690 (*snesi_zero): Likewise.
14691 (*seqsi_zero): Likewise.
14692 (*sltu_insn): Likewise.
14693 (*sgeu_insn): Likewise.
14694 (*seqdi_special): Make operand 0 and comparison operation be of
14695 DImode.
14696 (*snedi_special): Likewise.
14697 (*snedi_special_vis3): Likewise.
14698 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
14699 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
14700 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
14701 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
14702 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
14703 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
14704 (*sltu_extend_sp64): Likewise.
14705 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
14706 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
14707 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
14708 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
14709 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
14710
95ca411e
YZ
147112013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
14712
14713 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
14714 (aarch64_start_file): Use the new function.
14715
6782438d 147162013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 14717 Jason Merrill <jason@redhat.com>
6782438d
SKS
14718
14719 * common.opt: Add -gdwarf.
14720 * opts.c (common_handle_option): Handle it.
14721 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
14722
bb313b93
RB
147232013-04-10 Richard Biener <rguenther@suse.de>
14724
14725 * passes.c (execute_todo): Do not call ggc_collect conditional here.
14726 (execute_one_ipa_transform_pass): But unconditionally here.
14727 (execute_one_pass): And here.
14728 (init_optimization_passes): Remove reload pass.
14729 * tree-pass.h (TODO_ggc_collect): Remove.
14730 (pass_reload): Likewise.
14731 * ira.c (do_reload): Merge into ...
14732 (ira): ... this.
14733 (rest_of_handle_reload): Remove.
14734 (pass_reload): Likewise.
14735 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14736 of reload for vzeroupper pass placement.
14737 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
14738 and todo_flags_finish of all passes.
14739
793d9a16
RB
147402013-04-10 Richard Biener <rguenther@suse.de>
14741
14742 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
14743 first_const_oprnd field, rename first_def_type to first_op_type.
14744 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
14745 (vect_get_and_check_slp_defs): Always use the type of the
14746 operand. Allow mixed vect_external_def, vect_constant_def types.
14747 (vect_get_constant_vectors): Handle mixed vect_external_def,
14748 vect_constant_def types.
14749
12211b99 147502013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR
14751
14752 PR tree-optimization/55524
14753 * tree-ssa-math-opts.c
14754 (convert_mult_to_fma): Don't use an fms construct
14755 when we don't have an fms operation, but fnma, and it looks
14756 likely that we'll be able to use the latter.
14757
12211b99 147582013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ
14759
14760 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
14761 function.
14762 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
14763 inline fail caused by overwritable functions.
14764
34ab4a5b
CJW
147652013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
14766
14767 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
14768 unnecessary bits in the constant power of two case.
14769
abf9bfbc
RB
147702013-04-10 Richard Biener <rguenther@suse.de>
14771
14772 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
14773 broken code swapping operands.
14774 (vect_build_slp_tree): Do not compute load permutations here.
14775 (vect_analyze_slp_instance): Compute load permutations here,
14776 after building the SLP tree.
14777
f408477e
CB
147782013-04-09 Christian Bruel <christian.bruel@st.com>
14779
14780 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
14781 of next/prev_real_insn.
14782
5ac42672
JH
147832013-04-09 Jan Hubicka <jh@suse.cz>
14784
abf9bfbc
RB
14785 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14786 Drop aliased parameter.
5ac42672
JH
14787 (function_and_variable_visibility): Do not handle alias pairs.
14788 * cgraph.c (varpool_externally_visible_p): Update prototype.
14789 * varpool.c (varpool_add_new_variable): Update.
14790
5017f1d2
KT
147912013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14792
14793 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
14794
48eecbee
SB
147952013-04-09 Steven Bosscher <steven@gcc.gnu.org>
14796
a949cf1c
SB
14797 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
14798
48eecbee
SB
14799 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
14800
75ef8e3d
MP
148012013-04-09 Marek Polacek <polacek@redhat.com>
14802
14803 PR tree-optimization/48762
33159866 14804 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 14805
23847df4
RB
148062013-04-09 Richard Biener <rguenther@suse.de>
14807
14808 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
14809 dealing with cost.
14810 (vect_build_slp_tree): Likewise.
14811 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
14812 calculating the cost of a SLP instance.
14813 (vect_analyze_slp_instance): Use it from here, after building
14814 the SLP tree.
14815
ea3a0fde
JJ
148162013-04-09 Jakub Jelinek <jakub@redhat.com>
14817
14818 PR middle-end/56883
14819 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
14820 expand_omp_for_static_chunk): Use simple_p = true in
14821 force_gimple_operand_gsi calls when assigning to addressable decls.
14822
a32dfe9d
JL
148232013-04-09 Jeff Law <law@redhat.com>
14824
14825 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
14826 when the boolean was created by converting a wider object which
14827 had a boolean range.
14828
d755c7ef
RB
148292013-04-09 Richard Biener <rguenther@suse.de>
14830
14831 * tree-vectorizer.h (slp_void_p): Remove.
14832 (slp_tree): Typedef before _slp_tree declaration.
14833 (struct _slp_tree): Use a vector of slp_tree as children.
14834 (vect_get_place_in_interleaving_chain): Remove.
14835 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
14836 Move ...
14837 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
14838 and make static.
14839 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
14840 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
14841 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
14842 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
14843 Use slp_node instead of slp_void_p and adjust.
14844
3d741091
RB
148452013-04-09 Richard Biener <rguenther@suse.de>
14846
14847 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
14848 work that is not necessary.
14849
39307ba7
JJ
148502013-04-09 Jakub Jelinek <jakub@redhat.com>
14851
14852 PR tree-optimization/56854
14853 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
14854 forward into clobber stmts if it would change MEM_REF lhs into
14855 non-MEM_REF.
14856
343881fd
MK
148572013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
14858
14859 * tree.c (type_hash_lookup, type_hash_add): Make static.
14860 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
14861
3922658a
RB
148622013-04-09 Richard Biener <rguenther@suse.de>
14863
14864 * tree.h (unsave_expr_now): Remove.
14865 * tree-inline.c (mark_local_for_remap_r): Remove.
14866 (unsave_expr_1): Likewise.
14867 (unsave_r): Likewise.
14868 (unsave_expr_now): Likewise.
14869 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
14870 (propagate_tree_value): Likewise.
14871
9fb6b620
SB
148722013-04-08 Steven Bosscher <steven@gcc.gnu.org>
14873
14874 * doc/rtl.texi (sequence): Rewrite documentation to match the
14875 current use of SEQUENCE rtl objects.
14876 * rtl.def (SEQUENCE): Likewise.
14877
14878 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
14879 Update documentation.
14880 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
14881 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
14882
14883 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
14884
8ddb5a29
TJ
148852013-04-08 Teresa Johnson <tejohnson@google.com>
14886
14887 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
14888 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 14889 methods.
8ddb5a29
TJ
14890 (estimate_edge_size_and_time): Add comment to suggest using rounding
14891 methods.
14892 (estimate_node_size_and_time): Ditto.
14893 (remap_edge_change_prob): Use helper rounding divide methods.
14894 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
14895 (gimple_mod_pow2_value_transform): Ditto.
14896 (gimple_mod_subtract_transform): Ditto.
14897 (gimple_ic_transform): Ditto.
14898 (gimple_stringops_transform): Ditto.
14899 * stmt.c (conditional_probability): Ditto.
14900 (emit_case_dispatch_table): Ditto.
14901 * lto-cgraph.c (merge_profile_summaries): Ditto.
14902 * tree-optimize.c (execute_fixup_cfg): Ditto.
14903 * cfgcleanup.c (try_forward_edges): Ditto.
14904 * cfgloopmanip.c (scale_loop_profile): Ditto.
14905 (loopify): Ditto.
14906 (duplicate_loop_to_header_edge): Ditto.
14907 (lv_adjust_loop_entry_edge): Ditto.
14908 * tree-vect-loop.c (vect_transform_loop): Ditto.
14909 * profile.c (compute_branch_probabilities): Ditto.
14910 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
14911 * lto-streamer-in.c (input_cfg): Ditto.
14912 * gimple-streamer-in.c (input_bb): Ditto.
14913 * ipa-cp.c (update_profiling_info): Ditto.
14914 (update_specialized_profile): Ditto.
14915 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
14916 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 14917 rounding methods.
8ddb5a29
TJ
14918 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14919 (compute_trg_info): Ditto.
14920 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
14921 (purge_dead_edges): Ditto.
14922 * loop-unswitch.c (unswitch_loop): Ditto.
14923 * cgraphclones.c (cgraph_clone_edge): Ditto.
14924 (cgraph_clone_node): Ditto.
14925 * tree-inline.c (copy_bb): Ditto.
14926 (copy_edges_for_bb): Ditto.
14927 (initialize_cfun): Ditto.
14928 (copy_cfg_body): Ditto.
14929 (expand_call_inline): Ditto.
14930
661e6bd7
KT
149312013-04-08 Kai Tietz <ktietz@redhat.com>
14932
14933 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
14934 TARGET_CYGWIN64 by TARGET_64BIT.
14935
105766f3
JR
149362013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14937
14938 * config/epiphany/epiphany.md (GPR_1): New constant.
14939 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
14940 * config/epiphany/epiphany.c (gen_compare_reg):
14941 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
14942 is already in place.
14943 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
14944 Don't require being called during rtl expansion; If y operlaps r0,
14945 return 0.
14946 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
14947 (epiphany_expand_epilogue): Likewise.
14948
8afab237
JR
14949 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
14950 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 14951 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 14952
fa7707d6
JR
14953 * config/epiphany/constraints.md (CnL): New constraint.
14954 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
14955 * config/epiphany/predicates.md (add_operand): Allow 1024.
14956
5138e18d
JR
14957 * config/epiphany/epiphany.md (logical_op): New code iterator.
14958 (op_mnc): New code attribute.
14959 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
14960 (mov_f+1, mov_f+2): New peephole2 patterns.
14961
2ccc703d
JR
14962 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
14963 (cstoresi4): Also allow re-use of zero result when doing a NE
14964 comparison to a non-zero operand.
aefb0819 14965 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 14966
093ac0a5
JR
14967 * config/epiphany/epiphany.md (<insn_opname>v2si3):
14968 Use gen_addsi3_i / gen_subsi3_i.
14969
f223bb13
JJ
149702013-04-08 Jakub Jelinek <jakub@redhat.com>
14971
14972 PR c++/34949
14973 PR c++/50243
14974 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
14975 contain anything but clobbers, at most one __builtin_stack_restore,
14976 optionally debug stmts and final resx, and if it has at least one
14977 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
14978 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
14979 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
14980 which isn't defaut definition, remove them.
14981 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
14982 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
14983 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
14984 with MEM_REF LHS with SSA_NAME address.
14985
4481581f
JL
149862013-04-08 Jeff Law <law@redhat.com>
14987
14988 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 14989
451dabda
RB
149902013-04-08 Richard Biener <rguenther@suse.de>
14991
14992 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
14993 extra newline.
14994 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
14995 determined vector type.
14996 (vect_analyze_data_refs): Likewise.
14997 (vect_get_new_vect_var): Adjust.
14998 (vect_create_destination_var): Preserve SSA name versions.
14999 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
15000 not dump anything here.
15001
3b088b47
JR
150022013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15003
15004 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
15005 Add member lr_slot_known.
15006 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
15007 if necessary.
15008 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15009 Remove code that sets lr_slot_offset according to what a previous
15010 version of epiphany_emit_save_restore used to do.
15011 (epiphany_emit_save_restore): When doing an lr save or restore,
15012 set/verify lr_slot_known and lr_slot_offset.
15013
d8484d41
XQ
150142013-04-08 Xinyu Qi <xyqi@marvell.com>
15015
33159866
UB
15016 PR target/54338
15017 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ
15018 in ALL_REGS.
15019
481be1c4
RB
150202013-04-08 Richard Biener <rguenther@suse.de>
15021
15022 * alias.c (find_base_term): Fix thinko in previous change.
15023
401f3a81
JJ
150242013-04-08 Jakub Jelinek <jakub@redhat.com>
15025
15026 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
15027 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
15028 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
15029 if possible to compute val.
15030 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
15031 For QImode integers don't require anything about precision. Use
15032 const_with_all_bytes_same to find out if the constant doesn't have
15033 repeated bytes in it.
15034
03ed99a8
AK
150352013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15036
15037 * config/s390/s390.c (s390_expand_insv): Only accept insertions
15038 within mode size.
15039
781b2e62
MP
150402013-04-08 Marek Polacek <polacek@redhat.com>
15041
15042 PR rtl-optimization/48182
15043 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
15044 value to 1.
15045
27e430a2
JDA
150462013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15047
15048 PR target/55487
15049 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
15050 nuses, make sure we have a label.
15051
4902aa64
BS
150522013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15053
15054 PR target/56843
15055 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
15056 (rs6000_emit_swdiv_low_precision): Remove.
15057 (rs6000_emit_swdiv): Rewrite to handle between one and four
15058 iterations of Newton-Raphson generally; modify required number of
15059 iterations for some cases.
15060 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
15061
7bca81dc
SB
150622013-04-05 Steven Bosscher <steven@gcc.gnu.org>
15063
15064 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
15065 set-but-unused variable.
15066
15067 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
15068 basic blocks of released function bodies garbage-collectable.
15069
15070 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
15071 (struct rtl_opt_pass): Add TODO_df_finish.
15072
15073 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
15074
4542a38a
GY
150752013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15076
15077 * config/arm/constraints.md (q): New constraint.
15078 * config/arm/ldrdstrd.md: New file.
15079 * config/arm/arm.md (ldrdstrd.md) New include.
15080 (arm_movdi): Use "q" instead of "r" constraint
15081 for double-word memory access.
15082 (movdf_soft_insn): Likewise.
15083 * config/arm/vfp.md (movdi_vfp): Likewise.
15084 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 15085 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY
15086 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
15087 (mem_ok_for_ldrd_strd): Likewise.
15088 (output_move_double): Update assertion.
15089
2385b218
GY
150902013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15091
15092 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
15093
75fe1cb5
GY
150942013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15095
15096 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
15097 define_insn_and_split.
15098 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
15099
dd660e8e
GY
151002013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15101
15102 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
15103 define_insn_and_split.
33159866 15104 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY
15105 (shiftsi3_compare): New pattern.
15106 (rrx): New pattern.
15107 * config/arm/unspecs.md (UNSPEC_RRX): New.
15108
045e472c
GY
151092013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15110
15111 * config/arm/arm.md (negdi_extendsidi): New pattern.
15112 (negdi_zero_extendsidi): Likewise.
15113
3f3bf1a8
GY
151142013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15115
15116 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
15117 define_insn_and_split.
15118 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
15119 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
15120
b0b49556
GY
151212013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15122
15123 * config/arm/arm.md (arm_subdi3): Convert define_insn into
15124 define_insn_and_split.
15125 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
15126 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
15127
d633dd84
GY
151282013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15129
15130 * config/arm/arm.md (subsi3_carryin): New pattern.
15131 (subsi3_carryin_const): Likewise.
15132 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
15133 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
15134
f4499066
GY
151352013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15136
15137 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
15138
ceef6fd9
GY
151392013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15140
15141 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 15142 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 15143
ddbdd8a7
KT
151442013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15145
15146 * config/arm/arm.c (arm_expand_builtin): Change fcode
15147 type to unsigned int.
15148
8456d78a
RR
151492013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15150
15151 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
15152
526c230f
ILT
151532013-04-04 Ian Lance Taylor <iant@google.com>
15154
33159866
UB
15155 * doc/standards.texi (Standards): The Go frontend supports the Go 1
15156 language standard.
526c230f 15157
3cfbe04d
SB
151582013-04-04 Steven Bosscher <steven@gcc.gnu.org>
15159
15160 PR middle-end/56729
15161 * df-scan.c (df_insn_delete): Disable failing assert.
15162
dfa3f8d0
KT
151632013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15164
15165 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
15166 New function prototype.
15167 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
15168 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
15169 (arm_builtin_vectorized_function): New function.
15170
f7837758
KT
151712013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15172
15173 * config/arm/arm_neon_builtins.def: New file.
15174 * config/arm/arm.c (neon_builtin_data): Move contents to
15175 arm_neon_builtins.def.
15176 (enum arm_builtins): Include neon builtin definitions.
15177 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 15178 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 15179
39c1b6db
MP
151802013-04-04 Marek Polacek <polacek@redhat.com>
15181
15182 PR tree-optimization/48186
15183 * predict.c (maybe_hot_frequency_p): Return false if
15184 HOT_BB_FREQUENCY_FRACTION is 0.
15185 (cgraph_maybe_hot_edge_p): Likewise.
15186
314f64eb
RB
151872013-04-04 Richard Biener <rguenther@suse.de>
15188
15189 PR tree-optimization/56826
15190 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
15191 more accurately.
15192
51a905b2
RB
151932013-04-04 Richard Biener <rguenther@suse.de>
15194
15195 PR tree-optimization/56213
15196 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 15197 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 15198
f1bf4f3a
RB
151992013-04-04 Richard Biener <rguenther@suse.de>
15200
15201 PR tree-optimization/56837
15202 * tree-loop-distribution.c (classify_partition): For non-zero
15203 values require that the value has the same precision as its
15204 mode to be useful as memset value.
15205
0bca7ded
NC
152062013-04-03 Nick Clifton <nickc@redhat.com>
15207
33159866 15208 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC
15209 (fmssf4): Use fmsf.s on E3V5 architectures.
15210 (fnmasf4): Use fnmaf.s on E3V5 architectures.
15211 (fnmssf4): Use fnmsf.s on E3V5 architectures.
15212
b4019227
JL
152132013-04-03 Jeff Law <law@redhat.com>
15214
15215 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
15216 (lra-eliminations.o): Likewise.
15217
f57ddb5b
TJ
152182013-04-03 Teresa Johnson <tejohnson@google.com>
15219
15220 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 15221 compute_working_sets here from profile.c.
f57ddb5b 15222 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB
15223 (gcov_working_set_t): Moved typedef here from basic-block.h
15224 (compute_working_set): Declare.
f57ddb5b
TJ
15225 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
15226 (get_working_sets): Renamed from compute_working_set,
33159866 15227 replace most of body with call to new compute_working_sets.
f57ddb5b 15228 (get_exec_counts): Replace call to compute_working_sets
33159866
UB
15229 to get_working_sets.
15230 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 15231 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 15232 to get_working_sets.
f57ddb5b
TJ
15233 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
15234 * gcov-dump.c (dump_working_sets): New function.
15235
12211b99 152362013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ
15237
15238 * hwint.c (sext_hwi, zext_hwi): New functions.
15239 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15240 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15241 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15242 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15243 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15244 (sext_hwi, zext_hwi): New functions.
0bca7ded 15245
be672e08
JL
152462013-04-03 Jeff Law <law@redhat.com>
15247
15248 PR tree-optimization/56799
33159866
UB
15249 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15250 back test for widening conversion erroneously dropped in prior change.
be672e08 15251
9d821fa5
KT
152522013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15253
15254 PR target/56809
15255 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15256 instead of next_real_insn.
15257
71c581e7
MP
152582013-04-03 Marek Polacek <polacek@redhat.com>
15259
15260 PR sanitizer/55702
33159866 15261 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 15262
4a32ef80
KT
152632013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15264
15265 PR target/56809
15266 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15267 next_real_insn.
15268 (thumb1_output_casesi): Likewise.
15269 (thumb2_output_casesi): Likewise.
15270
1b2253d4
RB
152712013-04-03 Richard Biener <rguenther@suse.de>
15272
15273 PR tree-optimization/56817
15274 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15275 Split out ...
15276 (tree_unroll_loops_completely_1): ... new function to manually
15277 walk the loop tree, properly defering outer loops of unrolled
15278 loops to later iterations.
15279
38000232
MG
152802013-04-03 Marc Glisse <marc.glisse@inria.fr>
15281
15282 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15283 (vectorizable_load): Likewise.
15284 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15285 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15286
3fa35298
MG
152872013-04-03 Marc Glisse <marc.glisse@inria.fr>
15288
15289 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15290 BIT_FIELD_REF.
15291
b3d45ff0
UW
152922013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15293
15294 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15295
ec9202a8
BC
152962013-04-03 Bin Cheng <bin.cheng@arm.com>
15297
15298 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15299
6805bd36
MG
153002013-04-03 Marc Glisse <marc.glisse@inria.fr>
15301
15302 PR tree-optimization/56790
33159866
UB
15303 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15304 folding.
6805bd36 15305
da694a77
MG
153062013-04-03 Marc Glisse <marc.glisse@inria.fr>
15307
15308 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15309 Handle VEC_MERGE.
15310 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15311 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15312 equal arguments.
15313
4e7d7b3d
JJ
153142013-04-03 Jakub Jelinek <jakub@redhat.com>
15315
15316 PR c/19449
15317 * tree.h (force_folding_builtin_constant_p): New decl.
15318 * builtins.c (force_folding_builtin_constant_p): New variable.
15319 (fold_builtin_constant_p): Fold immediately also if
15320 force_folding_builtin_constant_p.
15321
e6c9d234
RB
153222013-04-03 Richard Biener <rguenther@suse.de>
15323
15324 PR tree-optimization/56812
15325 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15326 DRs of the same interleaving chain are independent.
15327
984d07dd
JM
153282013-04-02 Jason Merrill <jason@redhat.com>
15329
15330 * gdbinit.in (pbb): Use debug fn.
15331
622849c9
LC
153322013-04-02 Lawrence Crowl <crowl@google.com>
15333
15334 * sese.h (struct ivtype_map_elt_s): Remove unused.
15335 (extern debug_ivtype_map): Remove unused.
15336 (extern eq_ivtype_map_elts): Remove unused.
15337 * sese.c (debug_ivtype_map): Removed unused.
15338 (debug_ivtype_map_1): Removed unused.
15339 (debug_ivtype_elt): Remove unused.
15340 (eq_ivtype_map_elts): Remove unused.
15341
15342
82c0e1a0
KT
153432013-04-02 Kai Tietz <ktietz@redhat.com>
15344
15345 PR target/52790
15346 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
15347 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB
15348 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
15349 function.
82c0e1a0
KT
15350 (legitimize_pe_coff_symbol): Likewise.
15351 (is_imported_p): New helper-function.
15352 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
15353 for Windows x64 targets.
15354 (ix86_expand_prologue): Optimize for pe-coff targets.
15355 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
15356 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
15357 medium/large code-model.
15358 (legitimize_pic_address): Likewise.
15359 (legitimize_tls_address): Likewise.
15360 (ix86_expand_call): Likewise.
15361 (x86_output_mi_thunk): Likewise.
15362 (get_dllimport_decl): Add new beimport argument.
15363 (construct_plt_address): Don't assert for x64 pe-coff targets.
15364 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
15365 targets.
15366 (SYMBOL_FLAG_STUBVAR): New macro.
15367 (SYMBOL_REF_STUBVAR_P): Likewise.
15368 * config/i386/winnt.c (stub_list): New structure.
15369 (stub_head): New local variable.
15370 (i386_pe_record_stub): New function.
15371 (i386_pe_file_end): Emit refptr-stubs.
15372
5d751b0c
JJ
153732013-04-02 Jakub Jelinek <jakub@redhat.com>
15374
09bb4c99
JJ
15375 PR rtl-optimization/56745
15376 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
15377 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
15378
a1d8947a
JJ
15379 PR c++/34949
15380 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
15381 and both of them are MEM_REFs, just compare first argument for
15382 equality and attempt to deal even with differing offsets.
15383
5d751b0c
JJ
15384 PR c++/34949
15385 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
15386 of gimple_clobber_p to be MEM_REF.
15387 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
15388 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
15389 after gimplification.
15390 * asan.c (get_mem_ref_of_assignment): Don't instrument
15391 gimple_clobber_p stmts.
15392 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
15393 gimple_clobber_p stmt if they have MEM_REF lhs and
15394 are dead because of another gimple_clobber_p stmt.
15395 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15396 gimple_clobber_p stmts like debug stmts.
15397 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15398 that refer to unused VAR_DECLs or uninitialized values.
15399 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15400 gimple_clobber_p stmts if they refer to removed parameters.
15401 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15402 formatting.
15403
e594716a
UB
154042013-04-02 Uros Bizjak <ubizjak@gmail.com>
15405
15406 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15407 using SWI48 mode attribute.
15408
7a80735b
WM
154092013-04-02 Wei Mi <wmi@google.com>
15410
15411 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15412 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15413 *<rotate_insn><mode>3_mask in i386.md.
15414
f423a9e4
AI
154152013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15416
15417 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15418
90eb75f2
RB
154192013-04-02 Richard Biener <rguenther@suse.de>
15420
15421 PR tree-optimization/56778
15422 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15423 Runtime alias tests are not supported for gather loads.
15424 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15425 stmts referenced from SSA operands before updating SSA form.
15426
d8c69a92
IC
154272013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15428 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15429
15430 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB
15431 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15432 * config/arm/cortex-a53.md: New file.
15433 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15434 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15435 * config/arm/arm.c (arm_issue_rate): Likewise.
15436 * config/arm/arm-tune.md: Regenerate
15437 * config/arm/arm-tables.opt: Regenerate.
15438 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 15439
239eb04c
ZC
154402013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15441
15442 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15443 non-static link.
15444
c902d3c8
SN
154452013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15446
d8c69a92
IC
15447 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15448 scalar load/store operations using B/H registers.
c902d3c8
SN
15449 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15450
051d0e2f
SN
154512013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15452
15453 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15454 scalar move.
15455 * config/aarch64/aarch64.c
15456 (aarch64_simd_scalar_immediate_valid_for_move): New.
15457 * config/aarch64/aarch64-protos.h
15458 (aarch64_simd_scalar_immediate_valid_for_move): New.
15459 * config/aarch64/constraints.md (Dh, Dq): New.
15460 * config/aarch64/iterators.md (hq): New.
15461
0ee1e3d9
EB
154622013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15463
15464 * reorg.c (get_branch_condition): Deal with conditional returns.
15465 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15466
136fb3f7
WM
154672013-04-01 Wei Mi <wmi@google.com>
15468
15469 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15470 Truncate operand 2 using %b asm operand modifier.
15471 (*<shift_insn><mode>3_mask): Ditto.
15472 (*<rotate_insn><mode>3_mask): Ditto.
15473
6388c738
SB
154742013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15475
15476 PR middle-end/56798
15477 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15478
f7a4d826
KK
154792013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15480
15481 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15482 of next_real_insn.
15483 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15484
4b943a49
LC
154852013-03-30 Lawrence Crowl <crowl@google.com>
15486
15487 * dse.c (clear_alias_sets): Remove never set.
15488 (disqualified_clear_alias_sets): Remove never set.
15489 (clear_alias_mode_pool): Remove never set.
15490 (dse_step0): Remove condition that is never true.
15491 (canon_address): Remove condition that is never true.
15492 (dse_step7): Remove condition that is never true.
15493 (rest_of_handle_dse): Remove condition that is never true.
15494 (rest_of_handle_dse::did_global): Remove never read from above.
15495 (dse_step2_spill): Remove never called from above.
15496 (dse_step5_spill): Remove never called from above.
15497
39718607
SB
154982013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15499
da5c6bde
SB
15500 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15501 JUMP_TABLE_DATA changes.
15502 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15503 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15504 (Insns) <jump_table_data>: New entry.
15505 * doc/tm.texi: Regenerate.
15506
39718607
SB
15507 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15508
15509 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15510 for table jump at the end of a basic block using tablejump_p.
15511 * targhooks.c (default_invalid_within_doloop): Likewise.
15512 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15513 target hook implementation that is identical to the default hook.
15514 (rs6000_invalid_within_doloop): Remove.
15515
15516 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15517 unused variable from tablejump_p call.
15518
15519 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15520 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15521 (INSN_DELETED_P): Likewise.
15522 (emit_jump_table_data): New prototype.
15523 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15524 after 4th as unused.
15525 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15526 * sched-vis.c (print_insn): Likewise.
15527 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15528 insn for compatibility with back ends that use next_active_insn to
15529 identify jump table data.
15530 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15531 (remove_insn): Likewise.
15532 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15533 to be emitted.
15534 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15535 (emit_jump_table_data): New function.
15536
15537 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15538 basic block, a JUMP_TABLE_DATA never is.
15539 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15540 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15541 off from code handling real insns.
15542 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15543 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15544 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15545 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15546 is not a NONDEBUG_INSN_P.
15547 * ira-costs.c (scan_one_insn): Likewise.
15548 * jump.c (mark_all_labels): Likewise.
15549 (mark_jump_label_1): Likewise.
15550 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15551 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15552 a basic block.
15553 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15554 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15555 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15556 JUMP_TABLE_DATA_P insns.
15557 (calculate_elim_costs_all_insns): Likewise.
15558 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15559 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15560 (delete_output_reload): Code style fixups.
15561 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15562 insn flags on this non-insn.
15563 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15564 as scheduling barriers, for pre-change compatibility.
15565 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15566 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15567
15568 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15569 redundant JUMP_TABLE_DATA_P test.
15570 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15571 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15572 (frv_for_each_packet): Likewise.
15573 * config/i386/i386.c (min_insn_size): Likewise.
15574 (ix86_avoid_jump_mispredicts): Likewise.
15575 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15576 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15577 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15578 (mips16_insn_length): Robustify.
15579 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15580 (mips16_split_long_branches): Likewise.
15581 * config/pa/pa.c (pa_combine_instructions): Likewise.
15582 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15583 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15584 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15585 as contributing to pool range lengths.
15586 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15587 Remove redundant JUMP_TABLE_DATA_P test.
15588 (sh_loop_align): Likewise.
15589 (split_branches): Likewise.
15590 (sh_insn_length_adjustment): Likewise.
15591 * config/spu/spu.c (get_branch_target): Likewise.
15592
0208f7da
JH
155932013-03-29 Jan Hubicka <jh@suse.cz>
15594
15595 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
15596 gcov streaming; stream hot bb threshold to ltrans.
15597 * predict.c (get_hot_bb_threshold): Break out from ....
15598 (maybe_hot_count_p): ... here.
15599 (set_hot_bb_threshold): New function.
15600 * lto-section-in.c (lto_section_name): Add profile.
15601 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
15602 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
15603 and data-streamer.h
15604 (histogram_entry): New structure.
15605 (histogram, histogram_pool): New global vars.
15606 (histogram_hash): New structure.
15607 (histogram_hash::hash): New method.
15608 (histogram_hash::equal): Likewise.
15609 (account_time_size): New function.
15610 (cmp_counts): New function.
15611 (dump_histogram): New function.
15612 (ipa_profile_generate_summary): New function.
15613 (ipa_profile_write_summary): New function.
15614 (ipa_profile_read_summary): New function.
15615 (ipa_profile): Decide on threshold.
e594716a
UB
15616 (pass_ipa_profile): Add ipa_profile_write_summary and
15617 ipa_profile_read_summary.
0208f7da
JH
15618 * Makefile.in (ipa.o): Update dependencies.
15619 * lto-streamer.h (LTO_section_ipa_profile): New section.
15620
5a6ccc94
GDR
156212013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
15622
15623 * tree.h (VAR_P): New.
15624
39385fa6
PC
156252013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
15626
15627 PR lto/56777
15628 * doc/invoke.texi ([-fwhole-program]): Fix typo.
15629
34f0d87a
SB
156302013-03-29 Steven Bosscher <steven@gcc.gnu.org>
15631
15632 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
15633 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15634 (control_flow_insn_p): Likewise.
15635 * cfgrtl.c (duplicate_insn_chain): Likewise.
15636 * final.c (get_attr_length_1): Likewise.
15637 (shorten_branches): Likewise.
15638 (final_scan_insn): Likewise.
15639 * function.c (instantiate_virtual_regs): Likewise.
15640 * gcse.c (insert_insn_end_basic_block): Likewise.
15641 * ira-costs.c (scan_one_insn): Likewise.
15642 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15643 * lra.c (check_rtl): Likewise.
15644 * reload1.c (elimination_costs_in_insn): Likewise.
15645 * reorg.c (follow_jumps): Likewise.
15646
15647 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
15648 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15649 (thumb_far_jump_used_p): Likewise.
15650 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
15651 (workaround_speculation): Likewise.
15652 (add_sched_insns_for_speculation): Likewise.
15653 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
15654 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15655 (frv_for_each_packet): Likewise.
15656 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
15657 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
15658 (final_emit_insn_group_barriers): Likewise.
15659 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15660 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15661 (mips16_insn_length): Likewise.
15662 * config/pa/pa.c (pa_reorg): Likewise.
15663 (pa_combine_instructions): Likewise.
15664 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
15665 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
15666 (sh_reorg): Likewise.
15667 (split_branches): Likewise.
15668 * config/spu/spu.c (get_branch_target): Likewise.
15669
15670 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
15671 JUMP_TABLE_DATA_P.
15672
4ac761b0
KY
156732013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
15674
39385fa6 15675 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY
15676 Fix declaration name.
15677
58a49189
LC
156782013-03-28 Lawrence Crowl <crowl@google.com>
15679
15680 * graphds.h (struct graph.indicies): Remove unused.
15681 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
15682 (SCOP_ORIGINAL_PDDRS): Remove unused.
15683 * sese.h (extern insert_loop_close_phis): Removed unused.
15684 (extern insert_guard_phis): Removed unused.
15685 (extern ivtype_map_elt_info): Removed unused.
15686 (new_ivtype_map_elt): Removed unused.
15687 * sese.c (ivtype_map_elt_info): Removed unused.
15688
7b3b6ae4
LC
156892013-03-28 Lawrence Crowl <crowl@google.com>
15690
15691 * Makefile.in: Add several missing include dependences.
15692 (DUMPFILE_H): New.
15693 (test-dump.o): New. This object is not added to any executable,
15694 but is present for ad-hoc testing.
15695 * bitmap.c
15696 (debug (const bitmap_head_def &)): New.
15697 (debug (const bitmap_head_def *)): New.
15698 * bitmap.h
15699 (extern debug (const bitmap_head_def &)): New.
15700 (extern debug (const bitmap_head_def *)): New.
15701 * cfg.c
15702 (debug (edge_def &)): New.
15703 (debug (edge_def *)): New.
15704 * cfghooks.c
15705 (debug (basic_block_def &)): New.
15706 (debug (basic_block_def *)): New.
15707 * dumpfile.h
15708 (dump_node (const_tree, int, FILE *)): Correct source file.
15709 * dwarf2out.c
15710 (debug (die_struct &)): New.
15711 (debug (die_struct *)): New.
15712 * dwarf2out.h
15713 (extern debug (die_struct &)): New.
15714 (extern debug (die_struct *)): New.
15715 * gimple-pretty-print.c
15716 (debug (gimple_statement_d &)): New.
15717 (debug (gimple_statement_d *)): New.
15718 * gimple-pretty-print.h
15719 (extern debug (gimple_statement_d &)): New.
15720 (extern debug (gimple_statement_d *)): New.
15721 * ira-build.c
15722 (debug (ira_allocno_copy &)): New.
15723 (debug (ira_allocno_copy *)): New.
15724 (debug (ira_allocno &)): New.
15725 (debug (ira_allocno *)): New.
15726 * ira-int.h
15727 (extern debug (ira_allocno_copy &)): New.
15728 (extern debug (ira_allocno_copy *)): New.
15729 (extern debug (ira_allocno &)): New.
15730 (extern debug (ira_allocno *)): New.
15731 * ira-lives.c
15732 (debug (live_range &)): New.
15733 (debug (live_range *)): New.
15734 * lra-int.h
15735 (debug (lra_live_range &)): New.
15736 (debug (lra_live_range *)): New.
15737 * lra-lives.c
15738 (debug (lra_live_range &)): New.
15739 (debug (lra_live_range *)): New.
15740 * omega.c
15741 (debug (omega_pb_d &)): New.
15742 (debug (omega_pb_d *)): New.
15743 * omega.h
15744 (extern debug (omega_pb_d &)): New.
15745 (extern debug (omega_pb_d *)): New.
15746 * print-rtl.c
15747 (debug (const rtx_def &)): New.
15748 (debug (const rtx_def *)): New.
15749 * print-tree.c
15750 (debug_tree (tree): Move within file.
15751 (debug_raw (const tree_node &)): New.
15752 (debug_raw (const tree_node *)): New.
15753 (dump_tree_via_hooks (const tree_node *, int)): New.
15754 (debug (const tree_node &)): New.
15755 (debug (const tree_node *)): New.
15756 (debug_verbose (const tree_node &)): New.
15757 (debug_verbose (const tree_node *)): New.
15758 (debug_head (const tree_node &)): New.
15759 (debug_head (const tree_node *)): New.
15760 (debug_body (const tree_node &)): New.
15761 (debug_body (const tree_node *)): New.
15762 (debug_vec_tree (tree): Move and reimplement in terms of dump.
15763 (debug (vec<tree, va_gc> &)): New.
15764 (debug (vec<tree, va_gc> *)): New.
15765 * rtl.h
15766 (extern debug (const rtx_def &)): New.
15767 (extern debug (const rtx_def *)): New.
15768 * sbitmap.c
15769 (debug_raw (simple_bitmap_def &)): New.
15770 (debug_raw (simple_bitmap_def *)): New.
15771 (debug (simple_bitmap_def &)): New.
15772 (debug (simple_bitmap_def *)): New.
15773 * sbitmap.h
15774 (extern debug (simple_bitmap_def &)): New.
15775 (extern debug (simple_bitmap_def *)): New.
15776 (extern debug_raw (simple_bitmap_def &)): New.
15777 (extern debug_raw (simple_bitmap_def *)): New.
15778 * sel-sched-dump.c
15779 (debug (vinsn_def &)): New.
15780 (debug (vinsn_def *)): New.
15781 (debug_verbose (vinsn_def &)): New.
15782 (debug_verbose (vinsn_def *)): New.
15783 (debug (expr_def &)): New.
15784 (debug (expr_def *)): New.
15785 (debug_verbose (expr_def &)): New.
15786 (debug_verbose (expr_def *)): New.
15787 (debug (vec<rtx> &)): New.
15788 (debug (vec<rtx> *)): New.
15789 * sel-sched-dump.h
15790 (extern debug (vinsn_def &)): New.
15791 (extern debug (vinsn_def *)): New.
15792 (extern debug_verbose (vinsn_def &)): New.
15793 (extern debug_verbose (vinsn_def *)): New.
15794 (extern debug (expr_def &)): New.
15795 (extern debug (expr_def *)): New.
15796 (extern debug_verbose (expr_def &)): New.
15797 (extern debug_verbose (expr_def *)): New.
15798 (extern debug (vec<rtx> &)): New.
15799 (extern debug (vec<rtx> *)): New.
15800 * sel-sched-ir.h
15801 (_list_iter_cond_expr): Make inline instead of static.
15802 * sreal.c
15803 (debug (sreal &)): New.
15804 (debug (sreal *)): New.
15805 * sreal.h
15806 (extern debug (sreal &)): New.
15807 (extern debug (sreal *)): New.
15808 * tree.h
15809 (extern debug_raw (const tree_node &)): New.
15810 (extern debug_raw (const tree_node *)): New.
15811 (extern debug (const tree_node &)): New.
15812 (extern debug (const tree_node *)): New.
15813 (extern debug_verbose (const tree_node &)): New.
15814 (extern debug_verbose (const tree_node *)): New.
15815 (extern debug_head (const tree_node &)): New.
15816 (extern debug_head (const tree_node *)): New.
15817 (extern debug_body (const tree_node &)): New.
15818 (extern debug_body (const tree_node *)): New.
15819 (extern debug (vec<tree, va_gc> &)): New.
15820 (extern debug (vec<tree, va_gc> *)): New.
15821 * tree-cfg.c
15822 (debug (struct loop &)): New.
15823 (debug (struct loop *)): New.
15824 (debug_verbose (struct loop &)): New.
15825 (debug_verbose (struct loop *)): New.
15826 * tree-dump.c: Add header dependence.
15827 * tree-flow.h
15828 (extern debug (struct loop &)): New.
15829 (extern debug (struct loop *)): New.
15830 (extern debug_verbose (struct loop &)): New.
15831 (extern debug_verbose (struct loop *)): New.
15832 * tree-data-ref.c
15833 (debug (data_reference &)): New.
15834 (debug (data_reference *)): New.
15835 (debug (vec<data_reference_p> &)): New.
15836 (debug (vec<data_reference_p> *)): New.
15837 (debug (vec<ddr_p> &)): New.
15838 (debug (vec<ddr_p> *)): New.
15839 * tree-data-ref.h
15840 (extern debug (data_reference &)): New.
15841 (extern debug (data_reference *)): New.
15842 (extern debug (vec<data_reference_p> &)): New.
15843 (extern debug (vec<data_reference_p> *)): New.
15844 (extern debug (vec<ddr_p> &)): New.
15845 (extern debug (vec<ddr_p> *)): New.
15846 * tree-ssa-alias.c
15847 (debug (pt_solution &)): New.
15848 (debug (pt_solution *)): New.
15849 * tree-ssa-alias.h
15850 (extern debug (pt_solution &)): New.
15851 (extern debug (pt_solution *)): New.
15852 * tree-ssa-alias.c
15853 (debug (_var_map &)): New.
15854 (debug (_var_map *)): New.
15855 (debug (tree_live_info_d &)): New.
15856 (debug (tree_live_info_d *)): New.
15857 * tree-ssa-alias.h
15858 (extern debug (_var_map &)): New.
15859 (extern debug (_var_map *)): New.
15860 (extern debug (tree_live_info_d &)): New.
15861 (extern debug (tree_live_info_d *)): New.
15862
be77e1e5
JH
158632013-03-28 Jan Hubicka <jh@suse.cz>
15864
15865 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
15866
777e6976
IB
158672013-03-28 Ian Bolton <ian.bolton@arm.com>
15868
15869 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
15870 record only when desired or required.
15871
9cd347ae
UB
158722013-03-28 Uros Bizjak <ubizjak@gmail.com>
15873
15874 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
15875 *vec_extractv2di_1_rex64. Use x64 isa attribute.
15876
33159866 158772013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N
15878
15879 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
15880 (*andsi3_compare0_uxtw): New pattern.
15881 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
15882 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
15883
89ab31c1
JH
158842013-03-28 Jan Hubicka <jh@suse.cz>
15885
15886 * data-streamer-in.c (streamer_read_gcov_count): New function.
15887 * gimple-streamer-out.c: Include value-prof.h.
15888 (output_gimple_stmt): Output histogram.
15889 (output_bb): Use streamer_write_gcov_count.
15890 * value-prof.c: Include data-streamer.h
15891 (dump_histogram_value): Add HIST_TYPE_MAX.
15892 (stream_out_histogram_value): New function.
15893 (stream_in_histogram_value): New function.
15894 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
15895 (stream_out_histogram_value, stream_in_histogram_value): Declare.
15896 * data-streamer-out.c (streamer_write_gcov_count): New function.
15897 (streamer_write_gcov_count_stream): New function.
15898 * lto-cgraph.c (lto_output_edge): Update counter streaming.
15899 (lto_output_node): Likewise.
15900 (input_node, input_edge): Likewise.
15901 * lto-streamer-out.c (output_cfg): Update streaming.
15902 * lto-streamer-in.c (input_cfg): Likewise.
15903 * data-streamer.h (streamer_write_gcov_count,
15904 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
15905 * gimple-streamer-in.c: Include value-prof.h
15906 (input_gimple_stmt): Input histograms.
15907 (input_bb): Update profile streaming.
15908
e594716a 159092013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 15910
33159866 15911 * genmodes.c (emit_max_int): New function.
8697be17 15912 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC
15913 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
15914 Added doc.
8697be17 15915 * machmode.def: Fixed comment.
89ab31c1 15916
e594716a 159172013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ
15918
15919 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ
15920 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
15921
74946978
MP
159222013-03-28 Marek Polacek <polacek@redhat.com>
15923 Richard Biener <rguenther@suse.de>
15924
15925 PR tree-optimization/56695
15926 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
15927 build signed result of a vector comparison.
15928 * tree-cfg.c (verify_gimple_comparison): Check that a result
15929 of a vector comparison has signed type.
15930
a64b9c26
RB
159312013-03-28 Richard Biener <rguenther@suse.de>
15932
15933 PR tree-optimization/37021
15934 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
15935 do not restrict gaps between groups.
15936 * tree-vect-stmts.c (vectorizable_load): Properly account for
15937 a gap between groups.
15938
a9dc2a2f
EB
159392013-03-28 Eric Botcazou <ebotcazou@adacore.com>
15940
15941 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
15942 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
15943 is not enabled.
15944
53cb97f9
GP
159452013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
15946
15947 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
15948 * doc/extend.texi (Named Address Spaces): Ditto.
15949 (Variable Attributes): Ditto.
15950
b802ae5c
KT
159512013-03-27 Kai Tietz <ktietz@redhat.com>
15952
eddae10a
KT
15953 * config.build: Add support for cygwin x64 target.
15954 * config.gcc: Likewise.
15955 * config.host: Likewise.
15956 * configure.ac: Likewise
15957 * configure: Regenerated.
b802ae5c 15958
371e77e3 159592013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT
15960
15961 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
15962 * config/i386/t-cygwin-w64: New file.
15963 * config/i386/cygwin-w64.h: New file.
15964 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
15965 and add support for x64-cygwin target.
15966 (CPP_SPEC): Likewise.
15967 (CXX_WRAP_SPEC_LIST): Undefine before define.
15968 (LIBGCJ_SONAME): Use 15 as version.
15969
f49b33cb
RB
159702013-03-27 Richard Biener <rguenther@suse.de>
15971
15972 PR tree-optimization/56716
15973 * tree-ssa-structalias.c (perform_var_substitution): Adjust
15974 dumping for ref nodes.
15975
b37a6ce5
MJ
159762013-03-27 Martin Jambor <mjambor@suse.cz>
15977
15978 PR tree-optimization/55334
15979 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
15980 restricted pointers to arrays.
15981
9469b9b2
GDR
159822013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
15983
15984 * Makefile.in (.SUFFIXES): Add .cc.
15985 (.c.o): Apply same recipe for implicit rule .cc.o.
15986
7d24f650
RB
159872013-03-27 Richard Biener <rguenther@suse.de>
15988
15989 PR tree-optimization/37021
15990 * tree-vect-data-refs.c (vect_check_strided_load): Allow
15991 REALPART/IMAGPART_EXPRs around the supported refs.
15992 * tree-ssa-structalias.c (find_func_aliases): Assume that
15993 floating-point values are not used to transfer pointers.
15994
2f251a05
AI
159952013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
15996
f43245d1
UB
15997 * target.def (TARGET_HAS_IFUNC_P): New target hook.
15998 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
15999 * doc/tm.texi: Regenerate.
16000 * targhooks.h (default_has_ifunc_p): New.
16001 * targhooks.c (default_has_ifunc_p): Ditto.
16002 * config/linux-protos.h: New file.
16003 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
16004 hook for linux which disables support of indirect functions in android.
16005 * config/linux-android.c: New file.
16006 * config/t-linux-android.c: Ditto.
16007 * config.gcc: Added new object file linux-android.o.
16008 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 16009 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 16010 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI
16011 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
16012 doesn't support indirect functions.
16013 * configure: Regenerate.
16014
78b4e425
BC
160152013-03-27 Bin Cheng <bin.cheng@arm.com>
16016
16017 PR target/56102
16018 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
16019 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
16020 mult-word mode.
16021
a5ba7b92
AK
160222013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16023
16024 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
16025
be624986
TG
160262013-03-27 Terry Guo <terry.guo@arm.com>
16027
16028 * config/arm/arm-cores.def: Added core cortex-r7.
16029 * config/arm/arm-tune.md: Regenerated.
16030 * config/arm/arm-tables.opt: Regenerated.
16031 * doc/invoke.texi: Added entry for core cortex-r7.
16032
0a514f47
WL
160332013-03-27 Walter Lee <walt@tilera.com>
16034
16035 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
16036 double-decrement of next_scratch_regno.
16037
9332b0d2
WL
160382013-03-27 Walter Lee <walt@tilera.com>
16039
801d9b2a 16040 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL
16041 input operands.
16042 (insn_v1mulus): Ditto.
16043 (insn_v2muls): Ditto.
16044
f54ea5dd
WL
160452013-03-27 Walter Lee <walt@tilera.com>
16046
e594716a 16047 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL
16048 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
16049
bf60f4ca
WL
160502013-03-27 Walter Lee <walt@tilera.com>
16051
e594716a 16052 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL
16053 (*sibcall_value): Ditto.
16054
e3b51eeb
WL
160552013-03-27 Walter Lee <walt@tilera.com>
16056
16057 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
16058 (insn_mnz_v8qi): ... this ...
16059 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
16060 vector equivalent.
16061 (insn_v<n>mnz): Replaced by ...
16062 (insn_v1mnz): ... this ...
16063 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
16064 equivalent.
16065 (insn_mz_<mode>): Replaced by ...
16066 (insn_mz_v8qi): ... this ...
16067 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
16068 vector equivalent.
16069 (insn_v<n>mz): Replaced by ...
16070 (insn_v1mz): ... this ...
16071 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
16072 equivalent.
16073
4fc7b145
EB
160742013-03-26 Eric Botcazou <ebotcazou@adacore.com>
16075
16076 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
16077
6f33817e
RM
160782013-03-26 Roland McGrath <mcgrathr@google.com>
16079
16080 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
16081 than fprintf with a non-constant, non-format string.
16082
93a4145b
UB
160832013-03-26 Uros Bizjak <ubizjak@gmail.com>
16084
16085 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
16086 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
16087 operand 0 predicate.
16088 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
16089 attribute. Use general_x64nomem_operand as operand 1 predicate.
16090 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
16091 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16092 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
16093 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16094 (mov<mode>_insv_1): Remove expander. Merge insn with
16095 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
16096 Use general_x64nomem_operand as operand 1 predicate.
16097 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
16098 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
16099 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16100 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
16101 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
16102 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
16103 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16104 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
16105 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
16106 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
16107 (general_x64nomem_operand): Ditto.
16108
c6a9ed5a 161092013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH
16110
16111 * config/rtems.opt: Add -pthread option.
16112
31b0a960
RB
161132013-03-26 Richard Biener <rguenther@suse.de>
16114
93a4145b
UB
16115 * alias.c (find_base_term): Avoid redundant and not used recursion.
16116 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB
16117 (true_dependence_1): Compute and pass base terms to base_alias_check.
16118 (write_dependence_p): Likewise.
16119 (may_alias_p): Likewise.
16120
79517551
SN
161212013-03-26 Sofiane Naci <sofiane.naci@arm.com>
16122
16123 * config/aarch64/aarch64.c (aarch64_classify_address): Support
16124 PC-relative load in SI modes and above only.
16125
a76213b9
XQ
161262013-03-26 Xinyu Qi <xyqi@marvell.com>
16127
16128 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB
16129 * config/arm/iwmmxt.md (WCGR0): Update.
16130 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 16131
37ff9355
UB
161322013-03-26 Uros Bizjak <ubizjak@gmail.com>
16133
16134 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
16135 Use x64 and nox64 isa attributes.
16136
30b0317c
RB
161372013-03-26 Richard Biener <rguenther@suse.de>
16138
16139 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
16140 alignment computations and rely on get_object_alignment_1
16141 for the !TYPE_P case.
16142 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
16143
cbcdb140
WL
161442013-03-26 Walter Lee <walt@tilera.com>
16145
16146 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
16147 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
16148
dd884525 161492013-03-25 Jeff Law <law@redhat.com>
be672e08 16150
33159866
UB
16151 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
16152 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL
16153 wrong version of prior patch.
16154
0051d3ec
WL
161552013-03-25 Walter Lee <walt@tilera.com>
16156
16157 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
16158 TILEGX_INSN_SHUFFLEBYTES1.
16159 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
16160 shufflebytes1.
16161 (tilegx_builtins): Ditto.
16162 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
16163
065a3605
WL
161642013-03-25 Walter Lee <walt@tilera.com>
16165
16166 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL
16167 (floatunssisf2): New pattern.
16168 (floatsidf2): New pattern.
16169 (floatunssidf2): New pattern.
065a3605 16170
5b2a3c82
WL
161712013-03-25 Walter Lee <walt@tilera.com>
16172
16173 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
16174 tests for constraint J, K, N, P.
16175
192ea533
WL
161762013-03-25 Walter Lee <walt@tilera.com>
16177
16178 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
16179 Use indirect/pcrel encoding.
16180 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
16181 Ditto.
16182
855e0d0b
SE
161832013-03-25 Steve Ellcey <sellcey@mips.com>
16184
16185 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
16186 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
16187 * config/mips/mips.c (mips_option_override): Set IMADD default.
16188 * config/mips/mips.h (PTF_AVOID_IMADD): New.
16189 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
16190 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
16191 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
16192 * doc/invoke.texi (-mimadd/-mno-imadd): New.
16193
39e45653
JL
161942013-03-25 Jeff Law <law@redhat.com>
16195
16196 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
16197 slightly to avoid creating and folding useless trees. Simplify
16198 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
16199
7482c470
UB
162002013-03-25 Uros Bizjak <ubizjak@gmail.com>
16201
16202 * config/i386/i386.md (*zero_extendsidi2): Merge with
16203 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
16204 * config/i386/predicates.md (x86_64_zext_operand): Rename from
16205 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
16206 targets. Clarify comment.
16207
4a53743e
MJ
162082013-03-25 Martin Jambor <mjambor@suse.cz>
16209
16210 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
16211 pass-through jump functions differently.
16212 (ipa_read_jump_function): Likewise. Also use setter functions to set
16213 up jump functions.
16214
162712de
MJ
162152013-03-25 Martin Jambor <mjambor@suse.cz>
16216
16217 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
16218 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
16219 process it.
16220 (ipa_get_indirect_edge_target): New function.
16221 (devirtualization_time_bonus): New parameter known_aggs, pass it to
16222 ipa_get_indirect_edge_target. Update all callers.
16223 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
16224 ipa_get_indirect_edge_target_1 instead of calling
16225 ipa_get_indirect_edge_target.
16226 (create_specialized_node): Pass aggvlas to
16227 ipcp_discover_new_direct_edges.
16228
a5a4c20a
KT
162292013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16230
16231 * config/arm/arm.md (f_sels, f_seld): New types.
16232 (*cmov<mode>): New pattern.
16233 * config/arm/predicates.md (arm_vsel_comparison_operator): New
16234 predicate.
16235
f992b9fc
KT
162362013-03-25 Kai Tietz <ktietz@redhat.com>
16237
16238 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16239 POSIX-printf for mingw-hosted builds.
16240
c16fd676
RB
162412013-03-25 Richard Biener <rguenther@suse.de>
16242
16243 PR middle-end/56694
16244 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16245 must-not-throw stmt location.
16246
5ad29f12
KT
162472013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16248
16249 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16250 Emit load-acquire versions when acq is true.
16251 (arm_emit_store_exclusive): Add rel parameter.
16252 Emit store-release versions when rel is true.
16253 (arm_split_compare_and_swap): Use acquire-release instructions
16254 instead.
16255 of barriers when appropriate.
16256 (arm_split_atomic_op): Likewise.
16257 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16258 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16259 (VUNSPEC_SLX): Likewise.
16260 (VUNSPEC_LDA): Likewise.
16261 (VUNSPEC_STL): Likewise.
16262 * config/arm/sync.md (atomic_load<mode>): New pattern.
16263 (atomic_store<mode>): Likewise.
16264 (arm_load_acquire_exclusive<mode>): Likewise.
16265 (arm_load_acquire_exclusivesi): Likewise.
16266 (arm_load_acquire_exclusivedi): Likewise.
16267 (arm_store_release_exclusive<mode>): Likewise.
16268
03a7dddb
CM
162692013-03-25 Catherine Moore <clm@codesourcery.com>
16270
16271 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16272 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16273 * config/mip/predicates.md (lwsp_swsp_operand,
16274 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16275 sb16_operand, db4_operand, db7_operand, ib3_operand,
16276 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16277 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16278 andi16_operand): New predicates.
16279 * config/mips/mips.md (compression): New attribute.
16280 (enabled): New attribute.
16281 (length): Consider compression in computing length.
16282 (shift_compression): New code attribute.
16283 (*add<mode>3): New operands. Record compression.
16284 (sub<mode>3): Likewise.
16285 (one_cmpl<mode>2): Likewise.
16286 (*and<mode>3): Likewise.
16287 (*ior<mode>3): Likewise.
16288 (unnamed pattern for xor): Likewise.
16289 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16290 (*<optab><mode>3): Likewise.
16291 (*mov<mode>_internal: Likewise.
16292 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16293 (mips_unsigned_immediate_p): New.
16294 (umips_lwsp_swsp_address_p): New.
16295 (m16_based_address_p): New.
16296 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16297 (mips_unsigned_immediate_p): New prototype.
16298 (lwsp_swsp_address_p): New prototype.
16299 (m16_based_address_p): New prototype.
16300 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16301 (mips_signed_immediate_p): New function.
16302 (m16_based_address_p): New function.
16303 (lwsp_swsp_address_p): New function.
16304 (mips_print_operand_punctuation): Recognize short delay slot insns
16305 for microMIPS.add<mode>3"
16306
f35c297f 163072013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 16308
f35c297f
KT
16309 PR target/56720
16310 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16311 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16312
051b9446
RB
163132013-03-25 Richard Biener <rguenther@suse.de>
16314
16315 PR tree-optimization/56689
16316 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16317 any edge.
16318
374001cb
RB
163192013-03-25 Richard Biener <rguenther@suse.de>
16320
16321 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16322 of bitmap.
16323 (memory_references): Likewise.
16324 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16325 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16326 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16327 (gather_mem_refs_in_loops): Fold into ...
16328 (analyze_memory_references): ... this. Move initialization
16329 to tree_ssa_lim_initialize.
16330 (fill_always_executed_in): Rename to ...
16331 (fill_always_executed_in_1): ... this.
16332 (fill_always_executed_in): Move contains_call computation to
16333 this new function from ...
16334 (tree_ssa_lim_initialize): ... here.
16335 (tree_ssa_lim): Call fill_always_executed_in.
16336
57895947
EB
163372013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16338
16339 * postreload.c (reload_combine): Fix code detecting returns.
16340
1e1b18c1
EB
163412013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16342
16343 * function.c (emit_use_return_register_into_block): On cc0 targets,
16344 do not emit the sequence between cc0 setter and user.
16345
9216baf1
KT
163462013-03-25 Kai Tietz <ktietz@redhat.com>
16347
1e1b18c1
EB
16348 * config/i386/predicates.md (local_symbolic_operand): Interpret
16349 dll-imported symbols as none-local.
9216baf1 16350
76421b44
RB
163512013-03-25 Richard Biener <rguenther@suse.de>
16352
16353 * tree-ssa-loop-im.c (struct depend): Remove.
16354 (struct lim_aux_data): Make depends a vec of gimples.
16355 (free_lim_aux_data): Adjust.
16356 (add_dependency): Likewise.
16357 (set_level): Likewise.
16358
d154bfa2
RB
163592013-03-25 Richard Biener <rguenther@suse.de>
16360
16361 PR middle-end/56434
16362 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
16363 the pointer returned by calls with ECF_MALLOC set.
16364
2cd9804e
UB
163652013-03-24 Uros Bizjak <ubizjak@gmail.com>
16366
acef5fe0
UB
16367 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
16368
163692013-03-24 Uros Bizjak <ubizjak@gmail.com>
16370
16371 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB
16372 using MMXMODE mode iterator.
16373 (*move<mode>_internal): Merge with *movv2sf_internal and
16374 *movv2sf_internal_rex64 using MMXMODE mode iterator.
16375
225ccc68
SB
163762013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16377
85c0f02d
SB
16378 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
16379 (record_last_mem_set_info): Likewise.
16380
225ccc68
SB
16381 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
16382 of XNEWVEC followed by memset.
16383 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
16384
b64925dc
SB
163852013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16386
16387 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
16388 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
16389 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
16390 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
16391 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
16392 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
16393 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16394 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16395 BARRIER_P instead of GET_CODE.
16396
b9aaf52e
EB
163972013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16398
16399 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16400 inaccuracy in the probing code.
16401
16402 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16403 (ctrapdi4): Likewise.
16404
66de4d7c
EB
164052013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16406
16407 * calls.c (expand_call): Add missing guard to code handling return
16408 of non-BLKmode structures in MSB.
16409 * function.c (expand_function_end): Likewise.
16410
4f25c8fb
EB
164112013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16412
16413 * combine.c (try_combine): Adjust comment. Do not add the set of
16414 insn #0 if the destination indirectly is set or dies in insn #2.
16415 Tidy up code to distribute a new note.
16416
23b7850d
UB
164172013-03-22 Uros Bizjak <ubizjak@gmail.com>
16418
16419 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16420 also for alternatives 16 and 17.
16421
bed852cf
UB
164222013-03-22 Uros Bizjak <ubizjak@gmail.com>
16423
16424 * config/i386/sse.md (*mov<mode>_internal): Merge with
16425 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16426 Emit insn template depending on type attribute. Use
16427 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16428 movd instead of movq mnemonic for interunit moves. Rewrite mode
16429 attribute calculation. Remove unit attribute calculation.
16430 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16431 Set prefix_data16 attribute for DImode ssemov types.
16432 Use Ym instead of y for SSE-MMX conversion alternatives.
16433 Reorder operand constraints.
16434
1e8a7937
SB
164352013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16436
16437 * df.h (df_insn_delete): Adjust prototype.
16438 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16439 and let it decide whether mark the basic block dirty.
16440 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16441 * df-scan.c (df_insn_info_delete): New helper function, split
16442 off from df_insn_delete.
16443 (df_scan_free_bb_info): Use it.
16444 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16445 Likewise.
16446 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16447 that the insn is actually an insn and it has a non-NULL basic block.
16448 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16449
e14d094c
RB
164502013-03-22 Richard Biener <rguenther@suse.de>
16451
16452 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16453 dep_ref members.
16454 (mem_ref_alloc): Do not allocate them.
16455 (refs_independent_p): Do not query or maintain a cache.
16456
e9cf7316
RB
164572013-03-22 Richard Biener <rguenther@suse.de>
16458
16459 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16460 (gather_mem_refs_in_loops): Do not compute it.
16461 (analyze_memory_references): Do not allocate it.
16462 (tree_ssa_lim_finalize): Do not free it.
16463 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16464
f046e81b
RB
164652013-03-22 Richard Biener <rguenther@suse.de>
16466
16467 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 16468
a05c0ddf
IB
164692013-03-22 Ian Bolton <ian.bolton@arm.com>
16470
16471 * config/aarch64/aarch64.c (aarch64_print_operand): New
16472 format specifier for printing a constant in hex.
16473 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16474 format specifier for printing second operand.
16475
c00217fc
RB
164762013-03-22 Richard Biener <rguenther@suse.de>
16477
16478 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16479 bitmaps.
16480 (gather_mem_refs_in_loops): Perform store accumulation here.
16481 (create_vop_ref_mapping_loop): Remove.
16482 (create_vop_ref_mapping): Likewise.
16483 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 16484 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB
16485 (record_indep_loop): Remove.
16486 (record_dep_loop): New function.
16487 (ref_indep_loop_p_1): Adjust to only walk over references
16488 in the loop, not its subloops.
16489 (ref_indep_loop_p): Rename to ...
16490 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16491 maintaining a more fine-grained cache.
16492 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16493 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16494
15d19bf8
RB
164952013-03-22 Richard Biener <rguenther@suse.de>
16496
16497 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16498 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16499 aggregate mem_ref_loc.
16500 (free_mem_ref_locs): Inline into ...
16501 (memref_free): ... this and adjust.
16502 (mem_ref_alloc): Adjust.
16503 (mem_ref_locs_alloc): Remove.
16504 (record_mem_ref_loc): Adjust.
16505 (get_all_locs_in_loop): Rewrite into ...
16506 (for_all_locs_in_loop): ... this iterator.
16507 (rewrite_mem_ref_loc): New functor.
16508 (rewrite_mem_refs): Use for_all_locs_in_loop.
16509 (sm_set_flag_if_changed): New functor.
16510 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16511 (ref_always_accessed): New functor.
16512 (ref_always_accessed_p): Use for_all_locs_in_loop.
16513
6f37411d
MG
165142013-03-21 Marc Glisse <marc.glisse@inria.fr>
16515
16516 * tree-pass.h (PROP_gimple_lvec): New.
16517 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16518 (init_optimization_passes): Move pass_lower_vector.
16519 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16520 PROP_gimple_lvec.
16521 (pass_lower_vector): Provide PROP_gimple_lvec.
16522 (pass_lower_vector_ssa): Likewise.
16523 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16524
511ef689
MW
165252013-03-21 Mark Wielaard <mjw@redhat.com>
16526
16527 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16528
cc1df30b
UB
165292013-03-21 Uros Bizjak <ubizjak@gmail.com>
16530
16531 * config/i386/i386.md (*movdi_internal): Disparage slightly
16532 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16533 conversion alternatives.
16534
c845cfe1
JJ
165352013-03-21 Jakub Jelinek <jakub@redhat.com>
16536
37e99116
JJ
16537 PR middle-end/48087
16538 * diagnostic.def (DK_WERROR): New kind.
16539 * diagnostic.h (werrorcount): Define.
16540 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16541 promoted to DK_ERROR, increment DK_WERROR counter instead of
16542 DK_ERROR counter.
16543 * toplev.c (toplev_main): Call print_ignored_options even if
16544 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16545 even if just werrorcount is non-zero.
16546
c845cfe1
JJ
16547 PR debug/55608
16548 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16549 on failure.
16550 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16551 (string_cst_pool_decl): New function.
16552 (optimize_one_addr_into_implicit_ptr): New function.
16553 (resolve_addr_in_expr): Optimize DWARF location expression
16554 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16555 which doesn't live in memory, but has DW_AT_location or
16556 DW_AT_const_value, or refers to a string literal, into
16557 DW_OP_GNU_implicit_pointer.
16558 (optimize_location_into_implicit_ptr): New function.
16559 (resolve_addr): If removing DW_AT_location of a variable because
16560 it was DW_OP_addr of address of the variable, but the variable doesn't
16561 live in memory, try to emit const value attribute for the initializer.
16562
08e0cda6
MG
165632013-03-21 Marc Glisse <marc.glisse@inria.fr>
16564
16565 * tree.h (VECTOR_TYPE_P): New macro.
16566 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16567 TYPE_MODE): Use it.
16568 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16569 VEC_COND_EXPR cannot be lvalues.
16570 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16571
d08633b4
MG
165722013-03-21 Marc Glisse <marc.glisse@inria.fr>
16573
16574 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16575 Restrict the transformation to equal modes.
16576
e6647190
RB
165772013-03-21 Richard Biener <rguenther@suse.de>
16578
16579 PR tree-optimization/39326
16580 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16581 (MEM_ANALYZABLE): Adjust.
16582 (record_mem_ref_loc): Move bitmap ops ...
16583 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16584 unanalyzable refs, do not record locations for it.
16585 (analyze_memory_references): Allocate ref zero as shared
16586 unanalyzable ref.
16587 (refs_independent_p): Do not test for unanalyzed mems here.
16588 (ref_indep_loop_p_1): Special-case disambiguation against
16589 the unanalyzed ref.
cc1df30b 16590 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 16591
65074f54
CL
165922013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
16593
16594 * config/arm/arm-protos.h (tune_params): Add
16595 prefer_neon_for_64bits field.
16596 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
16597 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
16598 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
16599 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
16600 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
16601 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
16602 (arm_option_override): Handle -mneon-for-64bits new option.
16603 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
16604 (prefer_neon_for_64bits): Declare new variable.
16605 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 16606 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL
16607 (arch_enabled): Handle new arch types. Remove support for onlya8
16608 and nota8.
16609 (one_cmpldi2): Use new arch names.
bc5faa5b 16610 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL
16611 * config/arm/arm.opt (mneon-for-64bits): Add option.
16612 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
16613 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
16614 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
16615 of onlya8.
16616 * doc/invoke.texi (-mneon-for-64bits): Document.
16617
5a2d2a79
RB
166182013-03-21 Richard Biener <rguenther@suse.de>
16619
16620 PR tree-optimization/39326
16621 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
16622 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 16623 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 16624
5abe1e05
RB
166252013-03-21 Richard Biener <rguenther@suse.de>
16626
16627 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
16628 (vect_insert_into_interleaving_chain): Likewise.
16629 (vect_drs_dependent_in_basic_block): Inline ...
16630 (vect_slp_analyze_data_ref_dependence): ... here. New function,
16631 split out from ...
16632 (vect_analyze_data_ref_dependence): ... here. Simplify.
16633 (vect_check_interleaving): Simplify.
16634 (vect_analyze_data_ref_dependences): Likewise. Split out ...
16635 (vect_slp_analyze_data_ref_dependences): ... this new function.
16636 (dr_group_sort_cmp): New function.
16637 (vect_analyze_data_ref_accesses): Compute data-reference groups
16638 here instead of in vect_analyze_data_ref_dependence. Use
16639 a more efficient algorithm.
16640 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
16641 vect_slp_analyze_data_ref_dependences. Call
16642 vect_analyze_data_ref_accesses earlier.
16643 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
16644 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
16645 (vect_slp_analyze_data_ref_dependences): New prototype.
16646
cad1735b
RB
166472013-03-21 Richard Biener <rguenther@suse.de>
16648
16649 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
16650 ref is stored in the loop.
16651 (find_refs_for_sm): Walk only over all stores.
16652 (store_motion_loop): Allocate from lim_bitmap_obstack.
16653 (store_motion): Likewise.
16654
141310ef
RB
166552013-03-21 Richard Biener <rguenther@suse.de>
16656
16657 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
16658 Update virtual SSA form.
16659
5022315a
RO
166602013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16661
16662 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
16663 * configure: Regenerate.
16664 * config.in: Regenerate.
16665 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
16666 if !HAVE_LD_EH_FRAME_CIEV3.
16667
50d4421c
RB
166682013-03-21 Richard Biener <rguenther@suse.de>
16669
16670 * tree-cfg.c (verify_expr_no_block): New function.
16671 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
16672 nor DECL_VALUE_EXPR have locations with associated blocks.
16673 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
16674 (clear_unused_block_pointer): Remove code dealing with
16675 blocks in DECL_DEBUG_EXPR locations.
16676
839b422f
RB
166772013-03-21 Richard Biener <rguenther@suse.de>
16678
16679 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
16680 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
16681 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
16682 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
16683 instead of DECL_DEBUG_EXPR_IS_FROM.
16684 * gimplify.c (gimplify_modify_expr): Likewise.
16685 * tree-cfg.c (verify_expr_location_1): Likewise.
16686 * tree-complex.c (create_one_component_var): Likewise.
16687 * tree-sra.c (create_access_replacement): Likewise.
16688 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
16689 (clear_unused_block_pointer): Likewise.
16690 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
16691 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
16692 * var-tracking.c (var_debug_decl): Likewise.
16693 (track_expr_p): Likewise.
16694 * tree-inline.c (add_local_variables): Likewise. Set
16695 DECL_HAS_DEBUG_EXPR_P after copying it.
16696 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
16697 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
16698
fe04878d
UB
166992013-03-21 Uros Bizjak <ubizjak@gmail.com>
16700
16701 PR bootstrap/56656
16702 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
16703 * configure: Regenerate.
16704 * config.in: Regenerate.
16705 * config/i386/i386.md (*movdf_internal): Use
16706 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16707 movd instead of movq mnemonic for interunit moves.
16708 (*movdi_internal): Ditto.
16709
bd059b26 167102013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N
16711
16712 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
16713 (abd<mode>_3): New pattern.
16714 (aba<mode>_3): New pattern.
16715 (fabd<mode>_3): New pattern.
16716
bd059b26 167172013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N
16718
16719 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
16720 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
16721 occurrence of REGISTER_PREFIX as its empty string.
16722
4f1f78b9
JL
167232013-03-20 Jeff Law <law@redhat.com>
16724
16725 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
16726 addititional equivalences for equality comparisons between an SSA_NAME
16727 and a constant where the SSA_NAME was set from a widening conversion.
16728
327a1118
WL
167292013-03-20 Walter Lee <walt@tilera.com>
16730
16731 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
16732
813e0036
UB
167332013-03-20 Uros Bizjak <ubizjak@gmail.com>
16734
16735 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
16736 depending on type attribute.
16737 (*movti_internal): Ditto.
16738 (*movtf_internal): Ditto.
16739 (*movxf_internal): Ditto.
16740 (*movdf_internal): Ditto.
16741 (*movsf_internal): Ditto.
16742
7cf34aae
UB
167432013-03-20 Uros Bizjak <ubizjak@gmail.com>
16744
16745 * config/i386/i386.md (*movti_internal): Set prefix attribute to
16746 maybe_vex for sselog1 and ssemov types.
16747 (*movdi_internal): Reorder operand constraints.
16748 (*movsi_internal): Ditto. Set prefix attribute to
16749 maybe_vex for sselog1 and ssemov types.
16750 (*movtf_internal): Set prefix attribute to maybe_vex
16751 for sselog1 and ssemov types.
16752 (*movdf_internal): Ditto. Set prefix_data16 attribute for
16753 DImode ssemov types. Reorder operand constraints.
16754 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
16755 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
16756 attribute for SImode ssemov types. Reorder operand constraints.
16757
19321415
MJ
167582013-03-20 Martin Jambor <mjambor@suse.cz>
16759
16760 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
16761 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
16762
12211b99 167632013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM
16764
16765 * config/rs6000/predicates.md (indexed_address, update_address_mem
16766 update_indexed_address_mem): New predicates.
16767 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
16768 attribute for load/store instructions.
16769 * config/rs6000/dfp.md (movsd_store): Likewise.
16770 (movsd_load): Likewise.
16771 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
16772 (unnamed HI->DI extend define_insn): Likewise.
16773 (unnamed SI->DI extend define_insn): Likewise.
16774 (unnamed QI->SI extend define_insn): Likewise.
16775 (unnamed QI->HI extend define_insn): Likewise.
16776 (unnamed HI->SI extend define_insn): Likewise.
16777 (unnamed HI->SI extend define_insn): Likewise.
16778 (extendsfdf2_fpr): Likewise.
16779 (movsi_internal1): Likewise.
16780 (movsi_internal1_single): Likewise.
16781 (movhi_internal): Likewise.
16782 (movqi_internal): Likewise.
16783 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
16784 attribute for load/store instructions.
16785 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
16786 instructions.
16787 (mov<mode>_softfloat): Likewise.
16788 (mov<mode>_hardfloat32): Likewise.
16789 (mov<mode>_hardfloat64): Likewise.
16790 (mov<mode>_softfloat64): Likewise.
16791 (movdi_internal32): Likewise.
16792 (movdi_internal64): Likewise.
16793 (probe_stack_<mode>): Likewise.
16794
167952013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
16796
16797 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
16798 floating point, and decimal floating point to reload iterator.
16799
16800 * config/rs6000/constraints.md (wl constraint): New constraints to
16801 return FLOAT_REGS if certain options are used to reduce the number
16802 of separate patterns that exist in the file.
16803 (wx constraint): Likewise.
16804 (wz constraint): Likewise.
16805
16806 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
16807 -mdebug=reg, print wg, wl, wx, and wz constraints.
16808 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
16809 Initialize the reload functions for 64-bit binary/decimal floating
16810 point types.
16811 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
16812 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
16813 create the buffer on the stack to overcome not having a 32-bit
16814 load and store.
16815 (rs6000_emit_move): Likewise.
16816 (rs6000_secondary_memory_needed_rtx): Likewise.
16817 (rs6000_alloc_sdmode_stack_slot): Likewise.
16818 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
16819 via xxlxor, just like DFmode 0.0.
16820
16821 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
16822 define as 1 if we are running on a power7 or newer.
16823 (enum r6000_reg_class_enum): Add new constraints.
16824
16825 * config/rs6000/dfp.md (movsd): Delete, combine with binary
16826 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
16827 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 16828 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM
16829 (movsd splitter): Likewise.
16830 (movsd_hardfloat): Likewise.
16831 (movsd_softfloat): Likewise.
16832
16833 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
16834 binary and decimal floating point moves.
16835 (fmove_ok): New attributes to combine binary and decimal floating
16836 point moves, and to combine power6x (mfpgpr) moves along normal
16837 floating moves.
16838 (real_value_to_target): Likewise.
16839 (f32_lr): Likewise.
16840 (f32_lm): Likewise.
16841 (f32_li): Likewise.
16842 (f32_sr): Likewise.
16843 (f32_sm): Likewise.
16844 (f32_si): Likewise.
16845 (movsf): Combine binary and decimal floating point moves. Combine
16846 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 16847 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM
16848 (mov<mode> for SFmode/SDmode); Likewise.
16849 (SFmode/SDmode splitters): Likewise.
16850 (movsf_hardfloat): Likewise.
16851 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
16852 (movsf_softfloat): Likewise.
16853 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
16854
16855 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
16856 wx and wz constraints.
16857
16858 * config/rs6000/constraints.md (wg constraint): New constraint to
16859 return FLOAT_REGS if -mmfpgpr (power6x) was used.
16860
16861 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
16862 constraint.
16863
16864 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
16865 -mdebug=reg, print wg, wl, wx, and wz constraints.
16866 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
16867 Initialize the reload functions for 64-bit binary/decimal floating
16868 point types.
16869 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
16870 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
16871 create the buffer on the stack to overcome not having a 32-bit
16872 load and store.
16873 (rs6000_emit_move): Likewise.
16874 (rs6000_secondary_memory_needed_rtx): Likewise.
16875 (rs6000_alloc_sdmode_stack_slot): Likewise.
16876 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
16877 via xxlxor, just like DFmode 0.0.
16878
c6d5ff83
MM
16879 * config/rs6000/dfp.md (movdd): Delete, combine with binary
16880 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
16881 with other moves by using conditional constraits (wg). Use LFIWZX
16882 and STFIWX for loading SDmode on power7.
16883 (movdd splitters): Likewise.
16884 (movdd_hardfloat32): Likewise.
16885 (movdd_softfloat32): Likewise.
16886 (movdd_hardfloat64_mfpgpr): Likewise.
16887 (movdd_hardfloat64): Likewise.
16888 (movdd_softfloat64): Likewise.
16889
16890 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
16891 64-bit binary and decimal floating point moves.
16892 (FMOVE64X): Likewise.
16893 (movdf): Combine 64-bit binary and decimal floating point moves.
16894 Combine power6x (mfpgpr) moves with other moves by using
16895 conditional constraits (wg).
16896 (mov<mode> for DFmode/DDmode): Likewise.
16897 (DFmode/DDmode splitters): Likewise.
16898 (movdf_hardfloat32): Likewise.
16899 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
16900 (movdf_softfloat32): Likewise.
16901 (movdf_hardfloat64_mfpgpr): Likewise.
16902 (movdf_hardfloat64): Likewise.
16903 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
16904 (movdf_softfloat64): Likewise.
16905 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
16906 (reload_<mode>_load): Move to later in the file so they aren't in
16907 the middle of the floating point move insns.
16908 (reload_<mode>_store): Likewise.
16909
16910 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
16911 constraint.
16912
16913 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
16914 constraint if -mdebug=reg.
bd059b26
UB
16915 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
16916 Enable using dd reload support if needed.
c6d5ff83
MM
16917
16918 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
16919 binary and decimal floating point moves in rs6000.md.
16920 (movtd_internal): Likewise.
16921
16922 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
16923 decimal floating point moves.
16924 (movtf): Likewise.
16925 (movtf_internal): Likewise.
16926 (mov<mode>_internal, TDmode/TFmode): Likewise.
16927 (movtf_softfloat): Likewise.
16928 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
16929
16930 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
16931 movdi_internal64, using wg constraint for move direct operations.
16932 (movdi_internal64): Likewise.
16933
16934 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
16935 MODES_TIEABLE_P for selected modes. Print the numerical value of
16936 the various virtual registers. Use GPR/FPR first/last values,
16937 instead of hard coding the register numbers. Print which modes
16938 have reload functions registered.
bd059b26
UB
16939 (rs6000_option_override_internal): If -mdebug=reg, trace the options
16940 settings before/after setting cpu, target and subtarget settings.
16941 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
16942 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM
16943 (rs6000_secondary_reload_fail): Likewise.
16944 (rs6000_secondary_reload_inner): Likewise.
16945
16946 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
16947 macros for first/last GPR and FPR registers.
16948 (LAST_GPR_REGNO): Likewise.
16949 (FIRST_FPR_REGNO): Likewise.
16950 (LAST_FPR_REGNO): Likewise.
16951
16952 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
16953 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
16954 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
16955 (vcond<mode><mode>): Likewise.
16956 (vcondu<mode><mode>): Likewise.
16957 (vector_gtu<mode>): Likewise.
16958 (vector_gte<mode>): Likewise.
16959 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 16960 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM
16961 (ior<mode>3): Likewise.
16962 (and<mode>3): Likewise.
16963 (one_cmpl<mode>2): Likewise.
16964 (nor<mode>3): Likewise.
16965 (andc<mode>3): Likewise.
16966
16967 * config/rs6000/constraints.md (wt constraint): New constraint
16968 that returns VSX_REGS if TImode is allowed in VSX registers.
16969
16970 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
16971 constant under VSX.
16972
16973 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
16974 similar to TImode, but it is restricted to being in the GPRs.
16975
16976 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
16977 TImode to occupy a single VSX register.
16978
16979 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
16980 -mvsx-timode for power7/power8.
16981 (power7 cpu): Likewise.
16982 (power8 cpu): Likewise.
16983
16984 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
16985 sure that TFmode/TDmode take up two registers if they are ever
16986 allowed in the upper VSX registers.
16987 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
16988 registers.
16989 (rs6000_init_hard_regno_mode_ok): Likewise.
16990 (rs6000_debug_reg_global): Add debugging for PTImode and wt
16991 constraint. Print if LRA is turned on.
16992 (rs6000_option_override_internal): Give an error if -mvsx-timode
16993 and VSX is not enabled.
16994 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
16995 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
16996 to reg+offset addressing. Use PTImode when checking offset
16997 addresses for validity.
16998 (reg_offset_addressing_ok_p): Likewise.
16999 (rs6000_legitimate_offset_address_p): Likewise.
17000 (rs6000_legitimize_address): Likewise.
17001 (rs6000_legitimize_reload_address): Likewise.
17002 (rs6000_legitimate_address_p): Likewise.
17003 (rs6000_eliminate_indexed_memrefs): Likewise.
17004 (rs6000_emit_move): Likewise.
17005 (rs6000_secondary_reload): Likewise.
17006 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
17007 reloads to fpr registers to continue to use reg+offset addressing,
17008 but 64-bit reloads to altivec registers need reg+reg addressing.
17009 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
17010 it. Treat LO_SUM like a PLUS operation.
17011 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 17012 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM
17013 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
17014 registers to share a register with a smaller sized type, since VSX
17015 puts scalars in the upper 64-bits.
17016 (print_operand): Add support for PTImode.
17017 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
17018 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
17019 registers, but don't have arithmetic support.
17020 (rs6000_memory_move_cost): Add test for VSX.
17021 (rs6000_opt_masks): Add -mvsx-timode.
17022
17023 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
17024 for TImode.
17025 (VSs): Likewise.
17026 (VSr): Use wt constraint for TImode.
17027 (VSv): Drop TImode support.
17028 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
17029 (vsx_movti_64bit): Likewise.
17030 (vsx_movti_32bit): Likewise.
17031 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
17032 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
17033 one '?' on the appropriate output constraint. Do not allow TImode
17034 logical operations on 32-bit systems.
17035 (vsx_ior<mode>3): Likewise.
17036 (vsx_xor<mode>3): Likewise.
17037 (vsx_one_cmpl<mode>2): Likewise.
17038 (vsx_nor<mode>3): Likewise.
17039 (vsx_andc<mode>3): Likewise.
17040 (vsx_concat_<mode>): Likewise.
17041 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
17042
17043 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
17044 OPTION_MASK_VSX_TIMODE.
17045 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
17046 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
17047
17048 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
17049 (TI2 iterator): New iterator for TImode, PTImode.
17050 (wd mode attribute): Add values for vector types.
bd059b26
UB
17051 (movti_string): Replace TI move operations with operations for TImode
17052 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM
17053 (mov<mode>_string, TImode/PTImode): Likewise.
17054 (movti_ppc64): Likewise.
17055 (mov<mode>_ppc64, TImode/PTImode): Likewise.
17056 (TI mode splitters): Likewise.
17057
17058 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
17059 constraint.
17060
1fc5eced
MG
170612013-03-20 Marc Glisse <marc.glisse@inria.fr>
17062
17063 PR tree-optimization/56355
17064 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
17065 Also handle integers with undefined overflow.
17066
22c4c869
CM
170672013-03-20 Catherine Moore <clm@codesourcery.com>
17068 Maciej W. Rozycki <macro@codesourcery.com>
17069 Tom de Vries <tom@codesourcery.com>
12211b99 17070 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM
17071 Iain Sandoe <iain@codesourcery.com>
17072 Nathan Froyd <froydnj@codesourcery.com>
12211b99 17073 Chao-ying Fu <fu@mips.com>
22c4c869
CM
17074
17075 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 17076 Document new function attributes.
22c4c869
CM
17077 * doc/invoke.texi (minterlink-compressed, mmicromips,
17078 m14k, m14ke, m14kec): Document new options.
17079 (minterlink-mips16): Update documentation.
17080 * doc/md.texi (ZC, ZD): Document new constraints.
17081 * configure.ac (gcc_cv_as_micromips): Check if linker
17082 supports the .set micromips directive.
17083 * configure: Regenerate.
17084 * config.in: Regenerate.
17085 * config/mips/mips-tables.opt: Regenerate.
17086 * config/mips/micromips.md: New file.
17087 * constraints.md (ZC, ZD): New constraints.
17088 * config/mips/predicates.md (movep_src_register): New predicate.
17089 (movep_src_operand): New predicate.
17090 (non_volatile_mem_operand): New predicate.
17091 * config/mips/mips.md (multimem): New type.
17092 (length): Differentiate between 17-bit and 18-bit branch offsets.
17093 (MOVEP1, MOVEP2): New mode iterator.
33159866 17094 (mov_<load>l): Use ZC constraint.
22c4c869
CM
17095 (mov_<load>r): Likewise.
17096 (mov_<store>l): Likewise.
17097 (mov_<store>r): Likewise.
17098 (*branch_equality<mode>_inverted): Add microMIPS support.
17099 (*branch_equality<mode>): Likewise.
17100 (*jump_absolute): Likewise.
17101 (indirect_jump_<mode>): Likewise.
17102 (tablejump_<mode>): Likewise.
17103 (<optab>_internal): Likewise.
17104 (sibcall_internal): Likewise.
17105 (sibcall_value_internal): Likewise.
17106 (prefetch): Use constraint ZD.
17107 * config/mips/mips.opt (minterlink-compressed): New option.
17108 (minterlink-mips16): Now an alias for minterlink-compressed.
17109 (mmicromips): New option.
17110 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17111 (compare_and_swap_12): Likewise.
17112 (sync_add<mode>): Likewise.
17113 (sync_<optab>_12): Likewise.
17114 (sync_old_<optab>_12): Likewise.
17115 (sync_new_<optab>_12): Likewise.
17116 (sync_nand_12): Likewise.
17117 (sync_old_nand_12): Likewise.
17118 (sync_new_nand_12): Likewise.
17119 (sync_sub<mode>): Likewise.
17120 (sync_old_add<mode>): Likewise.
17121 (sync_old_sub<mode>): Likewise.
17122 (sync_new_add<mode>): Likewise.
17123 (sync_new_sub<mode>): Likewise.
17124 (sync_<optab><mode>): Likewise.
17125 (sync_old_<optab><mode>): Likewise.
17126 (sync_new_<optab><mode>): Likewise.
17127 (sync_nand<mode>): Likewise.
17128 (sync_old_nand<mode>): Likewise.
17129 (sync_new_nand<mode>): Likewise.
17130 (sync_lock_test_and_set<mode>): Likewise.
17131 (test_and_set_12): Likewise.
17132 (atomic_compare_and_swap<mode>): Likewise.
17133 (atomic_exchange<mode>_llsc): Likewise.
17134 (atomic_fetch_add<mode>_llsc): Likewise.
17135 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17136 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17137 (umips_save_restore_pattern_p): Likewise.
17138 (umips_load_store_pair_p): Likewise.
17139 (umips_output_load_store_pair): Likewise.
17140 (umips_movep_target_p): Likewise.
17141 (umips_12bit_offset_address_p): Likewise.
17142 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17143 (mips_base_mips16): Rename this...
17144 (mips_base_compression_flags): ...to this. Update all uses.
17145 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17146 (mips_mips16_decl_p): Delete.
17147 (mips_nomips16_decl_p): Delete.
33159866 17148 (mips_get_compress_on_flags): New function.
22c4c869
CM
17149 (mips_get_compress_off_flags): New function.
17150 (mips_get_compress_mode): New function.
17151 (mips_get_compress_on_name): New function.
17152 (mips_get_compress_off_name): New function.
17153 (mips_insert_attributes): Support multiple compression types.
17154 (mips_merge_decl_attributes): Likewise.
17155 (umips_12bit_offset_address_p): New function.
17156 (mips_start_function_definition): Emit .set micromips directive.
17157 (mips_call_may_need_jalx_p): New function.
17158 (mips_function_ok_for_sibcall): Add microMIPS support.
17159 (mips_print_operand_punctuation): Support short delay slots and
17160 compact jumps.
17161 (umips_swm_mask, umips_swm_encoding): New.
17162 (umips_build_save_restore): New function.
17163 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17164 (was_mips16_p): Remove.
17165 (old_compression_mode): New.
17166 (mips_set_compression_mode): New function.
17167 (mips_set_current_function): Add microMIPS support.
17168 (mips_option_override): Likewise.
17169 (umips_save_restore_pattern_p): New function.
17170 (umips_output_save_restore): New function.
17171 (umips_load_store_pair_p_1): New function.
17172 (umips_load_store_pair_p): New function.
17173 (umips_output_load_store_pair_1): New function.
17174 (umips_output_load_store_pair): New function.
17175 (umips_movep_target_p) New function.
17176 (mips_prepare_pch_save): Add microMIPS support.
17177 * config/mips/mips.h (TARGET_COMPRESSION): New.
17178 (TARGET_CPU_CPP_BUILTINS): Update macro
17179 to use new compression flags and to support microMIPS.
17180 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17181 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17182 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17183 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17184 (ASM_SPEC): Support mmicromips and mno-micromips.
17185 (M16STORE_REG_P): New macro.
17186 (MIPS_CALL): Support TARGET_MICROMIPS.
17187 (MICROMIPS_J): New macro.
17188 (mips_base_mips16): Rename this...
17189 (mips_base_compression_flags): ...to this.
17190 (UMIPS_12BIT_OFFSET_P): New macro.
17191 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17192 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB
171932013-03-20 Richard Biener <rguenther@suse.de>
17194
17195 PR tree-optimization/56661
17196 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
17197 the result does not have to be distinct.
17198
54714c68
RB
171992013-03-20 Richard Biener <rguenther@suse.de>
17200
17201 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
17202 remap_gimple_op_r.
17203
cca1130d 172042013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 17205 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS
17206
17207 PR rtl-optimization/56605
17208 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
17209
2a930377
UB
172102013-03-20 Uros Bizjak <ubizjak@gmail.com>
17211
17212 PR bootstrap/56656
17213 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
17214 that require movd instead of movq.
17215
d6d305fe
RB
172162013-03-20 Richard Biener <rguenther@suse.de>
17217
17218 * tree-ssa-structalias.c (struct variable_info): Add pointer
17219 to the first field of an aggregate with sub-vars. Make
17220 this and the pointer to the next subfield its ID.
17221 (vi_next): New function.
17222 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
17223 storedanything_id, integer_id): Increment by one.
17224 (new_var_info, get_call_vi, lookup_call_clobber_vi,
17225 get_call_clobber_vi): Adjust.
17226 (solution_set_expand): Simplify and speedup.
17227 (solution_set_add): Inline into ...
17228 (set_union_with_increment): ... this. Adjust accordingly.
17229 (do_sd_constraint): Likewise.
17230 (do_ds_constraint): Likewise.
17231 (do_complex_constraint): Simplify.
17232 (build_pred_graph): Adjust.
17233 (solve_graph): Likewise. Simplify and speedup.
17234 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
17235 get_constraint_for_component_ref, get_constraint_for_1,
17236 first_vi_for_offset, first_or_preceding_vi_for_offset,
17237 create_function_info_for, create_variable_info_for_1,
17238 create_variable_info_for, intra_create_variable_infos): Adjust.
17239 (init_base_vars): Push NULL for ID zero.
17240 (compute_points_to_sets): Adjust.
17241
a271b42d
RB
172422013-03-20 Richard Biener <rguenther@suse.de>
17243
17244 * cfgloop.c (verify_loop_structure): Streamline and avoid
17245 ICEing on corrupt loop tree.
17246 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17247 loop tree.
17248
ebd65954
RB
172492013-03-20 Richard Biener <rguenther@suse.de>
17250
17251 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17252 check whether an SSA update is needed.
17253
4547b7ee
RS
172542013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17255
17256 * config/mips/constraints.md (T): Rename to...
17257 (Yf): ...this.
17258 (U): Rename to...
17259 (Yd): ...this.
17260 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17261 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17262
0b8cdc58
IB
172632013-03-19 Ian Bolton <ian.bolton@arm.com>
17264
17265 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17266 (*subsi3_carryin_uxtw): Likewise.
17267
5977a10d
IB
172682013-03-19 Ian Bolton <ian.bolton@arm.com>
17269
17270 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17271 (*rorsi3_insn_uxtw): Likewise.
17272
bd83ff2c
IB
172732013-03-19 Ian Bolton <ian.bolton@arm.com>
17274
17275 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17276 (*extrsi5_insn_uxtw): Likewise.
17277
ba65123c
RB
172782013-03-19 Richard Biener <rguenther@suse.de>
17279
17280 PR tree-optimization/56273
17281 * passes.c (init_optimization_passes): Move second VRP after DOM.
17282
2eac0476
UB
172832013-03-19 Uros Bizjak <ubizjak@gmail.com>
17284
17285 * config/i386/i386.md (*movti_internal): Merge from
17286 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17287 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17288 nox64 isa attributes.
17289
6e55eda7
RB
172902013-03-18 Richard Biener <rguenther@suse.de>
17291
17292 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17293 (unite): Likewise.
17294 (merge_node_constraints): Likewise.
17295 (build_succ_graph): Likewise.
17296 (valid_graph_edge): Inline into single caller.
17297 (unify_nodes): Likewise. Use bitmap_set_bit return value
17298 and cache varinfo.
17299 (scc_visit): Fix formatting and variable use.
17300 (do_sd_constraint): Use gcc_checking_assert.
17301 (do_ds_constraint): Likewise.
17302 (do_complex_constraint): Likewise.
17303 (condense_visit): Likewise. Cleanup.
17304 (dump_pred_graph): New function.
17305 (perform_var_substitution): Dump the pred-graph before
17306 variable substitution.
17307 (find_equivalent_node): Use gcc_checking_assert.
17308 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17309
4bdd44c4
RB
173102013-03-18 Richard Biener <rguenther@suse.de>
17311
17312 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17313 Remove cond_expr_stmt_list argument and do not gimplify the
17314 built expression.
17315 (vect_loop_versioning): Adjust.
17316 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17317 Cleanup to use less temporaries.
17318 (vect_create_data_ref_ptr): Cleanup.
17319
38c56a5b
JJ
173202013-03-18 Jakub Jelinek <jakub@redhat.com>
17321
17322 PR tree-optimization/56635
17323 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17324 require types_compatible_p types.
17325
20b2e6a0
NC
173262013-03-18 Nick Clifton <nickc@redhat.com>
17327
a6178a25
NC
17328 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
17329 spurious backslash.
17330
20b2e6a0
NC
17331 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
17332 Add missing line to comment describing function.
17333
92608d0e
RB
173342013-03-18 Richard Biener <rguenther@suse.de>
17335
17336 PR tree-optimization/56210
17337 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17338 Handle string / character search functions.
17339 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
17340
31348d52
RB
173412013-03-18 Richard Biener <rguenther@suse.de>
17342
17343 PR middle-end/56483
17344 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
17345 and implement properly.
17346 * gimple.h (gimple_cond_single_var_p): Remove.
17347
fcac74a1
RB
173482013-03-18 Richard Biener <rguenther@suse.de>
17349
17350 * tree-data-ref.h (find_data_references_in_loop): Declare.
17351 * tree-data-ref.c (get_references_in_stmt): Use a stack
17352 vector pre-allocated in the callers.
17353 (find_data_references_in_stmt): Adjust.
17354 (graphite_find_data_references_in_stmt): Likewise.
17355 (create_rdg_vertices): Likewise.
17356 (find_data_references_in_loop): Export.
17357 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
17358 Compute dependences here...
17359 (vect_analyze_data_refs): ...not here. When we encounter
17360 a non-vectorizable data reference in basic-block vectorization
17361 truncate the data reference vector. Do not bother to
17362 fixup data-dependence information for gather loads.
17363 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
17364 of data references, as reported.
17365
0d5a1b56
RB
173662013-03-18 Richard Biener <rguenther@suse.de>
17367
17368 PR tree-optimization/3713
17369 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
17370 has_constants and expr.
17371 (stmt_has_constants): Properly valueize SSA names when deciding
17372 whether the stmt has constants.
17373
789c34e3
RB
173742013-03-18 Richard Biener <rguenther@suse.de>
17375
17376 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
17377 whole function when there is nothing to do.
17378 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
17379 * tree-vectorizer.c (vectorize_loops): Update virtual and
17380 loop-closed SSA once.
17381 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
17382
076b4605
RB
173832013-03-18 Richard Biener <rguenther@suse.de>
17384
17385 PR middle-end/56113
17386 * domwalk.c (bb_postorder): New global static.
17387 (cmp_bb_postorder): New function.
17388 (walk_dominator_tree): Replace scheme imposing an order for
17389 visiting dominator sons by one sorting them at the time they
17390 are pushed on the stack.
17391
bdb01696
RB
173922013-03-18 Richard Biener <rguenther@suse.de>
17393
17394 PR tree-optimization/39326
17395 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17396 (struct mem_ref): Replace mem member with ao_ref typed member.
17397 (MEM_ANALYZABLE): Adjust.
17398 (memref_eq): Likewise.
17399 (mem_ref_alloc): Likewise.
17400 (gather_mem_refs_stmt): Likewise.
17401 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17402 (execute_sm_if_changed_flag_set): Adjust.
17403 (execute_sm): Likewise.
17404 (ref_always_accessed_p): Likewise.
17405 (refs_independent_p): Likewise.
17406 (can_sm_ref_p): Likewise.
17407
12d2dc5e
JJ
174082013-03-18 Jakub Jelinek <jakub@redhat.com>
17409
17410 PR c/56566
17411 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17412 return 1 even for !unsignedp.
17413
286fb677
UB
174142013-03-17 Uros Bizjak <ubizjak@gmail.com>
17415
17416 * config/i386/i386.md (isa): Add x64 and nox64.
17417 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17418 (*pushtf): Enable *roF alternative for x64 isa only.
17419 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17420 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17421 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17422 (*movtf_internal): Merge from *movtf_internal_rex64 and
17423 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17424 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17425 nox64 isa attributes.
17426 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17427 nox64 isa attributes.
17428 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17429
88b97037
UB
174302013-03-17 Uros Bizjak <ubizjak@gmail.com>
17431
17432 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17433
9f4f1735
JJ
174342013-03-17 Jakub Jelinek <jakub@redhat.com>
17435
17436 PR target/56640
17437 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17438 class names. Remove trailing comma after "ALL_REGS".
17439
48f4a6fa
JH
174402013-03-16 Jan Hubicka <jh@suse.cz>
17441
17442 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17443 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17444 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17445 of cgraph_get_create_node.
17446 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17447
98e81123
JM
174482013-03-16 Jason Merrill <jason@redhat.com>
17449
17450 PR debug/49090
17451 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17452 with DW_AT_default_value.
17453
8a298c04
JJ
174542013-03-16 Jakub Jelinek <jakub@redhat.com>
17455
17456 * BASE-VER: Set to 4.9.0.
17457
4323afa0
AK
174582013-03-14 Andi Kleen <ak@linux.intel.com>
17459
17460 PR target/56619
17461 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17462 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17463 Document _x* TSX intrinsics.
17464
b3c0d469
JJ
174652013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17466 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI
17467
17468 * configure.ac: Add MicroBlaze TLS support detection.
17469 * configure: Regenerate.
88b97037
UB
17470 * config/microblaze/microblaze-protos.h
17471 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17472 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17473 * config/microblaze/microblaze.c (microblaze_address_type): Add
17474 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI
17475 (microblaze_address_info): Add tls_reloc.
17476 (TARGET_HAVE_TLS): Define.
17477 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17478 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB
17479 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17480 load_tls_operand, microblaze_call_tls_get_addr,
17481 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI
17482 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17483 (get_base_reg): Use microblaze_tls_symbol_p.
17484 (microblaze_classify_address): Handle TLS.
88b97037
UB
17485 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17486 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI
17487 (microblaze_legitimize_address): Handle TLS.
17488 (microblaze_address_insns): Handle ADDRESS_TLS.
17489 (pic_address_needs_scratch): Handle TLS.
17490 (print_operand_address): Handle TLS.
17491 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17492 (microblaze_expand_move): Handle TLS.
88b97037
UB
17493 (microblaze_legitimate_constant_p): Check
17494 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI
17495 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17496 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17497 (PIC_OFFSET_TABLE_REGNUM): Set.
17498 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17499 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17500 (addsi3, movsi_internal2, movdf_internal): Update constraints
17501 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB
17502 (move_operand): Redefine as move_src_operand,
17503 check microblaze_tls_referenced_p.
8cc9a5a5 17504
d803a491
IB
175052013-03-14 Ian Bolton <ian.bolton@arm.com>
17506
17507 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17508 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17509
a8504f22
IB
175102013-03-14 Ian Bolton <ian.bolton@arm.com>
17511
17512 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17513 CC mode for AND.
17514
df2dfaea
JJ
175152013-03-14 Jakub Jelinek <jakub@redhat.com>
17516
fbd28bc3
JJ
17517 PR tree-optimization/53265
17518 * common.opt (Waggressive-loop-optimizations): New option.
17519 * tree-ssa-loop-niter.c: Include tree-pass.h.
17520 (do_warn_aggressive_loop_optimizations): New function.
17521 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17522 if number_of_latch_executions returned constant.
17523 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17524 early. If number_of_latch_executions returned constant, set
17525 nb_iterations_upper_bound back to it.
17526 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17527 field.
17528 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17529 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17530
df2dfaea
JJ
17531 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17532 (MULTILIB_OSDIRNAMES): Set.
17533 * genmultilib: If defaultosdirname doesn't start with :: , set
17534 defaultosdirname2 instead, clear it and emit two . multilib_raw
17535 entries instead of just one.
17536
ee0d2b68
KK
175372013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17538
17539 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17540 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17541 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17542 (SUBTARGET_OVERRIDE_OPTIONS): New.
17543
decc676e
OE
175442013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17545
17546 PR target/49880
17547 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17548 (musermode): Convert to Var(TARGET_USERMODE).
17549 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17550 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17551 * config/sh/sh.c (sh_option_override): Use
17552 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17553 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17554 condition.
17555 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17556 TARGET_SH4.
17557 (udivsi3_i4_single, divsi3_i4_single): Use
17558 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17559
f4b719c7
DK
175602013-03-13 Dave Korn <dave.korn.cygwin@....>
17561
17562 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17563 default setting.
17564
c40eced0
RB
175652013-03-13 Richard Biener <rguenther@suse.de>
17566
17567 PR tree-optimization/56608
17568 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17569 calls when vectorizing basic-blocks.
17570
1bfa5973
JJ
175712013-03-13 Jakub Jelinek <jakub@redhat.com>
17572
17573 PR plugins/45078
17574 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17575 tm_file.
17576
6fcf5434
JJ
175772013-03-12 Jakub Jelinek <jakub@redhat.com>
17578
17579 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17580
4f38fa8c
JH
175812013-03-11 Jan Hubicka <jh@suse.cz>
17582
17583 PR lto/56557
17584 * lto-streamer-out.c (output_symbol_p): Skip references from
17585 constructors of external variables.
17586
c5c5ba89
JH
175872013-03-11 Jan Hubicka <jh@suse.cz>
17588
17589 PR middle-end/56571
17590 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
17591 from pseudos.
17592 * emit-rtl.c (verify_rtx_sharing): Likewise.
17593 (copy_insn_1): Likewise.
17594 * rtl.c (copy_rtx): Likewise.
17595
c2a939b1
GJL
175962013-03-11 Georg-Johann Lay <avr@gjlay.de>
17597
17598 PR target/56591
17599 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
17600 output_operand_lossage message.
88b97037 17601
3c3279fb
RE
176022013-03-11 Richard Earnshaw <rearnsha@arm.com>
17603
17604 PR target/56470
17605 * arm.c (shift_op): Validate RTL pattern on the fly.
17606 (arm_print_operand, case 'S'): Don't use shift_operator to validate
17607 the RTL.
17608
aef5ef9d
JDA
176092013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17610
17611 PR target/56347
17612 * config/pa/pa.md (call_value): Check for calls to powf and direct to
17613 new call patterns that clobber %fr12.
17614 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
17615 split and postreload patterns.
17616 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
17617 registers %fr12 and %fr12R as call used.
17618
456610d3
SB
176192013-03-09 Steven Bosscher <steven@gcc.gnu.org>
17620
17621 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
17622 (canon_address, record_store, replace_read, check_mem_read_rtx,
17623 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
17624 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
17625 rest_of_handle_dse): Likewise.
17626
4b1baac8
RS
176272013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
17628
17629 PR middle-end/56524
17630 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
17631 Add base_optabs.
17632 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
17633 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
17634 (save_optabs_if_changed): Replace with...
17635 (init_tree_optimization_optabs): ...this.
17636 * optabs.c (save_optabs_if_changed): Rename to...
17637 (init_tree_optimization_optabs): ...this. Take the optimization node
17638 as argument. Do nothing if the base optabs are already correct.
17639 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
17640 to recompute optabs.
17641 * function.h (function): Remove optabs field.
17642 * function.c (invoke_set_current_function_hook): Call
17643 init_tree_optimization_optabs. Use the result to initialize
17644 this_fn_optabs.
17645
b7a78683
AH
176462013-02-27 Aldy Hernandez <aldyh@redhat.com>
17647
17648 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
17649 if GTMA_HAS_NO_INSTRUMENTATION.
17650 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
17651 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
17652 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
17653 * gimple-pretty-print.c (dump_gimple_transaction): Handle
17654 GTMA_HAS_NO_INSTRUMENTATION.
17655
6384c29b
JJ
176562013-03-08 Jakub Jelinek <jakub@redhat.com>
17657
17658 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
17659 libasan_preinit.o.
17660
ecd4f20a
MP
176612013-03-08 Marek Polacek <polacek@redhat.com>
17662 Jakub Jelinek <jakub@redhat.com>
17663
17664 PR tree-optimization/56478
17665 * predict.c (is_comparison_with_loop_invariant_p): Change the
17666 type of loop_step to tree.
17667 (predict_loops): Adjust.
17668 (predict_iv_comparison): Perform the computations on double_ints.
17669
64366d35
RB
176702013-03-08 Richard Biener <rguenther@suse.de>
17671
17672 PR tree-optimization/56570
17673 * tree-cfg.c (verify_expr_location_1): Verify locations for
17674 DECL_DEBUG_EXPR.
17675 * tree-sra.c (create_access_replacement): Strip locations
17676 from DECL_DEBUG_EXPRs.
17677
a9d5a059
RB
176782013-03-08 Richard Biener <rguenther@suse.de>
17679
17680 * tree-inline.c (expand_call_inline): Do not associate
17681 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
17682 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
17683
b7aa4e9a
RB
176842013-03-08 Richard Biener <rguenther@suse.de>
17685
17686 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
17687 or block changes with -Og. Fix for location / block encoding
17688 changes and PHI arguments with locations.
17689
c4c2f9fa
SB
176902013-03-07 Steven Bosscher <steven@gcc.gnu.org>
17691
17692 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
17693 for all counters.
17694 (struct output_info): Likewise.
17695 (register_overhead): Remove bad gcc_assert.
17696 (bitmap_find_bit): If there is only a single bitmap element, do not
17697 count a miss as a search.
17698 (print_statistics): Update for counter type changes.
17699 (dump_bitmap_statistics): Likewise. Print headers such that they
17700 are properly lined up with the printed counters.
17701
5bf6606a
JJ
177022013-03-07 Jakub Jelinek <jakub@redhat.com>
17703
17704 PR tree-optimization/56559
17705 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
17706 check that it has only a single use.
17707
2c653d46
RB
177082013-03-07 Richard Biener <rguenther@suse.de>
17709
17710 * doc/invoke.texi (fwhole-program): Discourage use in combination
17711 with -flto.
17712
a72d8780
JJ
177132013-03-06 Jakub Jelinek <jakub@redhat.com>
17714
01a454df
JJ
17715 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
17716
c1781047
JJ
17717 PR tree-optimization/56539
17718 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
17719 instead of GSI_CONTINUE_LINKING as last argument to
17720 force_gimple_operand_gsi. Adjust function comment.
17721
9772c47a
JJ
17722 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
17723 aarch64-cores.def.
17724
a72d8780
JJ
17725 PR middle-end/56548
17726 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
17727 promoted mode, convert the result back to the original mode.
17728
fa5556de
RB
177292013-03-06 Richard Biener <rguenther@suse.de>
17730
17731 PR middle-end/56294
17732 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
17733 (insert_updated_phi_nodes_compare_uids): New function.
17734 (update_ssa): Sort symbols_to_rename after UID before
17735 traversing it to insert PHI nodes.
17736
010403d1
RB
177372013-03-06 Richard Biener <rguenther@suse.de>
17738
17739 PR middle-end/50494
17740 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
17741 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
17742
17743 Revert
17744 2013-02-13 Richard Biener <rguenther@suse.de>
17745
17746 PR lto/50494
17747 * varasm.c (output_constant_def_1): Get the decl representing
17748 the constant as argument.
17749 (output_constant_def): Wrap output_constant_def_1.
17750 (make_decl_rtl): Use output_constant_def_1 with the decl
17751 representing the constant.
17752 (build_constant_desc): Optionally re-use a decl already
17753 representing the constant.
17754 (tree_output_constant_def): Adjust.
17755
3c27ce4c
JY
177562013-03-06 Joey Ye <joey.ye@arm.com>
17757
17758 PR lto/50293
17759 * gcc.c (convert_white_space): New function.
17760 (main): Handles white space in function name.
17761
8f6d1c86
OE
177622013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
17763
17764 PR target/56529
17765 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
17766 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
17767 to SH_DIV_CALL_TABLE for TARGET_SH2.
17768 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
17769 list.
17770 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
17771 call-table options.
17772
798a209f
SA
177732013-03-05 Sterling Augustine <saugustine@google.com>
17774 Cary Coutant <ccoutant@google.com>
17775
17776 PR debug/55364
17777 * dwarf2out.c (resolve_addr): Don't call
17778 remove_loc_list_addr_table_entries a second time for the same
17779 expression.
17780
6cfa417f 177812013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 17782
6cfa417f
JJ
17783 PR debug/56510
17784 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
17785 (avoid_complex_debug_insns): New function.
17786 (expand_debug_locations): Call it.
17787
17788 PR rtl-optimization/56484
17789 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
17790 lifetimes of hard registers on small register class machines.
17791
177922013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
17793
17794 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 17795 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 17796 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH
17797 fast_interrupt.
17798 (microblaze_fast_interrupt_function_p): New function.
17799 (microblaze_is_interrupt_handler): Rename to
17800 microblaze_is_interrupt_variant and add fast_interrupt check.
17801 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
17802 (save_restore_insns): Likewise.
17803 (compute_frame_size): Likewise.
17804 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
17805 (microblaze_globalize_label): Likewise.
6cfa417f
JJ
17806 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
17807 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH
17808 microblaze_is_interrupt_variant.
17809
731aefac
KT
178102013-03-05 Kai Tietz <ktietz@redhat.com>
17811
17812 * sdbout.c (sdbout_one_type): Switch to current function's section
17813 supporting cold/hot.
17814
a72d8780 178152013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH
17816
17817 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
17818 -mxl-reorder.
17819
334e71e8
JJ
178202013-03-05 Jakub Jelinek <jakub@redhat.com>
17821
0b50e654
JJ
17822 PR middle-end/56461
17823 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
17824 if VALGRIND_GET_VBITS is defined, temporarily make object
17825 memory all defined, and restore previous valgrind addressability
17826 and definability afterwards. Free this_object at the end.
17827
4ccf8f43
JJ
17828 PR middle-end/56461
17829 * lra.c (lra): Call lra_clear_live_ranges if live_p,
17830 right before calling lra_create_live_ranges, also call it
17831 when clearing live_p. Only call lra_clear_live_ranges
17832 at the end if live_p.
17833
334e71e8
JJ
17834 PR middle-end/56461
17835 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
17836
9ca966ca
RB
178372013-03-05 Richard Biener <rguenther@suse.de>
17838
17839 PR tree-optimization/56521
17840 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
17841 value-id.
17842
d273b176
SB
178432013-03-05 Steven Bosscher <steven@gcc.gnu.org>
17844
17845 PR c++/55135
17846 * except.h (remove_unreachable_eh_regions): New prototype.
17847 * except.c (remove_eh_handler_splicer): New function, split out
17848 of remove_eh_handler.
17849 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
17850 warning about running it on many EH regions one at a time.
17851 (remove_unreachable_eh_regions_worker): New function, walk the
17852 EH tree in depth-first order and remove non-marked regions.
17853 (remove_unreachable_eh_regions): New function.
17854 * tree-eh.c (mark_reachable_handlers): New function, split out
17855 from remove_unreachable_handlers.
17856 (remove_unreachable_handlers): Use mark_reachable_handlers and
17857 remove_unreachable_eh_regions.
17858 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
17859 and remove_unreachable_eh_regions.
17860
5e7f1aef
RB
178612013-03-05 Richard Biener <rguenther@suse.de>
17862
17863 PR middle-end/56525
17864 * loop-init.c (fix_loop_structure): Remove loops in two stages,
17865 not freeing them until the end.
17866
f276b762
AK
178672013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17868
17869 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
17870
12a43ab8
RB
178712013-03-05 Richard Biener <rguenther@suse.de>
17872
17873 PR tree-optimization/56270
17874 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
17875 of loads after scheduling an SLP instance.
17876
93675444
JJ
178772013-03-05 Jakub Jelinek <jakub@redhat.com>
17878
db4138e3
JJ
17879 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
17880 tic6x.exp.
17881 (check_gcc_parallelize): Run guality.exp as a separate job from
17882 vect.exp with unsorted.exp and $(dg_target_exps) separately from
17883 struct-layout-1.exp with stackalign.exp.
17884
dd3d1ec0
JJ
17885 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
17886
b8d381a3
JJ
17887 PR middle-end/56461
17888 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
17889 load_index sbitmap even if some bit in it isn't set.
17890
b4f9786b
JJ
17891 PR middle-end/56461
17892 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
17893 (discover_iteration_bound_by_body_walk): Change queues to
17894 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
17895 spelling in comment. Call safe_push on queues[bound_index] directly.
17896 Release queues[queue_index] in every iteration unconditionally.
17897 Release bounds vector.
17898
93675444
JJ
17899 PR middle-end/56461
17900 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
17901 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
17902 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
17903 inner_phis vector.
17904
3e492e9c
RB
179052013-03-05 Richard Biener <rguenther@suse.de>
17906
17907 PR lto/56515
17908 * tree-inline.c (remap_blocks_to_null): New function.
17909 (expand_call_inline): When expanding a call stmt without
17910 an associated block inline remap all callee blocks to NULL.
17911
a5d1569a
JJ
179122013-03-05 Jakub Jelinek <jakub@redhat.com>
17913
86efb5cd
JJ
17914 PR rtl-optimization/56494
17915 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
17916 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
17917 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
17918
85f5dbea
JJ
17919 PR middle-end/56461
17920 * sel-sched-ir.c (free_sched_pools): Release
17921 succs_info_pool.stack[succs_info_pool.max_top] vectors too
17922 if succs_info_pool.max_top isn't -1.
17923
a5d1569a
JJ
17924 PR bootstrap/56509
17925 * opts.c (opts_obstack, opts_concat): Moved to...
17926 * opts-common.c (opts_obstack, opts_concat): ... here.
17927
4432aa6c
JJ
179282013-03-04 Jakub Jelinek <jakub@redhat.com>
17929
17930 PR middle-end/56461
17931 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
17932
f3a1fb91
MJ
179332013-03-04 Martin Jambor <mjambor@suse.cz>
17934
17935 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
17936 all appropriate places.
17937
22110e6c
EB
179382013-01-04 Eric Botcazou <ebotcazou@adacore.com>
17939
17940 PR tree-optimization/56424
17941 * ipa-split.c (split_function): Do not set the RSO flag if result is
17942 not by reference and its type is a register type.
17943
a72d8780 179442013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 17945
88b97037 17946 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 17947 (microblaze_legitimate_pic_operand): Likewise
88b97037 17948 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH
17949 new function microblaze_legitimate_pic_operand
17950 * config/microblaze/microblaze-protos.h
17951 (microblaze_legitimate_pic_operand): Declare.
17952
a72d8780 179532013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 17954
a72d8780 17955 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 17956 New predicate for supported rtx code types.
a72d8780 17957 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI
17958 call_insn_simple_operand predicate.
17959
541d9ac8
JJ
179602013-03-04 Jakub Jelinek <jakub@redhat.com>
17961
5eb010bc
JJ
17962 PR middle-end/56461
17963 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
17964 partitions.ordered_remove.
17965
30862efc
JJ
17966 PR middle-end/56461
17967 * tree-vect-stmts.c (vectorizable_conversion): Don't call
17968 vec_oprnds0.create (1) for modifier == NONE.
17969
8930f723
JJ
17970 PR middle-end/56461
17971 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
17972 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
17973 vec_oprnds1 right before pushing anything to it for
17974 scalar_shift_arg.
17975
541d9ac8
JJ
17976 PR middle-end/56461
17977 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
17978 set nbbs to 0 instead of having separate code path.
17979 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
17980 instead of false as last argument if returning NULL.
17981
7aa7f2e3
SL
179822013-03-03 Sandra Loosemore <sandra@codesourcery.com>
17983
17984 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
17985 the attribute is now called "target" instead of "option".
17986 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
17987 * doc/tm.texi.in (Target Attributes): Likewise document the correct
17988 attribute/pragma name for TARGET_OPTION_VALID_P and
17989 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
17990 * doc/tm.texi: Regenerated.
17991
8930f723 179922013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 17993
8930f723 17994 * config/microblaze/microblaze.c:
cb8a1637 17995 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ
17996 * config/microblaze/microblaze.h: Add -mxl-reorder to
17997 DRIVER_SELF_SPECS.
17998 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
17999 instructions emitted if TARGET_REORDER.
88b97037
UB
18000 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
18001 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 18002 separately.
cb8a1637 18003
4dc1d68c
XDL
180042013-03-01 Xinliang David Li <davidxl@google.com>
18005
18006 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
18007 walk length.
18008
689eaba3
JJ
180092013-03-01 Jakub Jelinek <jakub@redhat.com>
18010
9d676bf6
JJ
18011 PR middle-end/56461
18012 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
18013 vector even when returning true. Fix up function comment formatting.
18014
caff7edf
JJ
18015 PR middle-end/56461
18016 * ira-build.c (ira_loop_nodes_count): New variable.
18017 (create_loop_tree_nodes): Initialize it.
18018 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
18019
b6b9227d
JJ
18020 PR middle-end/56461
18021 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
18022 method on dr_chain and result_chain.
18023 * tree-vect-stmts.c (vectorizable_store): Only call
18024 result_chain.create if j == 0.
18025
689eaba3
JJ
18026 PR middle-end/56461
18027 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
18028 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
18029 before overwriting it.
18030
40bf31ed
TB
180312013-03-01 Tobias Burnus <burnus@net-b.de>
18032
18033 * doc/extended.texi (C Extensions): Change order in @menu
18034 to match @node.
18035 (Other MIPS Built-in Functions): Move last MIPS entry before
18036 "picoChip Built-in Functions".
18037 (SH Built-in Functions): Move after RX Built-in Functions.
18038 * doc/gcc.texi (Introduction): Change order in @menu
18039 to match @node.
18040 * doc/md.texi (Constraints): Ditto.
18041 * gty.texi (Type Information): Ditto.
18042 (User-provided marking routines for template types): Make
18043 subsection.
18044 * doc/invoke.texi (AArch64 Options): Move before
18045 "Adapteva Epiphany Options".
18046
e664c61c
KS
180472013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
18048 Jakub Jelinek <jakub@redhat.com>
18049
18050 PR sanitizer/56454
18051 * asan.c (gate_asan): Lookup no_sanitize_address instead of
18052 no_address_safety_analysis attribute.
18053 * doc/extend.texi (no_address_safety_attribute): Rename to
18054 no_sanitize_address attribute, mention no_address_safety_analysis
18055 attribute as deprecated alias.
18056
37b5ec8f
JJ
180572013-02-28 Jakub Jelinek <jakub@redhat.com>
18058
18059 PR middle-end/56461
18060 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
18061 type to vec<vec<tree> > *.
18062 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
18063 to be vec<tree> instead of vec<tree> *, set vec_defs
18064 to vNULL and call vec_defs.create (number_of_vects), adjust other
18065 uses of vec_defs.
18066 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
18067 vectorizable_condition): Adjust vect_get_slp_defs callers.
18068
ba96cdfb
JG
180692013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18070
18071 * config/aarch64/aarch64.c
18072 (aarch64_float_const_representable): Remove unused variable.
18073
6f549691
JG
180742013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18075
18076 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
18077
af55e82d
JG
180782013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18079
18080 * config/aarch64/aarch64-builtins.c
18081 (aarch64_init_simd_builtins): Make static.
18082
1df3f464
JG
180832013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18084
18085 * config/aarch64/aarch64.c
18086 (aarch64_simd_make_constant): Make static.
18087
f8f42513
MJ
180882013-02-28 Martin Jambor <mjambor@suse.cz>
18089
18090 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
18091 with no initialization to the RHS of debug statements.
18092
b48b3fc4
MJ
180932013-02-28 Martin Jambor <mjambor@suse.cz>
18094
18095 PR tree-optimization/56294
18096 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
18097 Adjust dumping.
18098 (get_access_replacement): Do not call create_access_replacement.
18099 Assert a replacement exists.
18100 (get_repl_default_def_ssa_name): Create the replacement declaration
18101 itself.
18102
c3ae224c
RR
181032013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18104
18105 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
18106 final_end_function.
18107
45fa0eef
MP
181082013-02-28 Marek Polacek <polacek@redhat.com>
18109
18110 PR rtl-optimization/56466
18111 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
18112 if we're changing a loop.
18113 (peel_loops_completely): Likewise.
18114
502c067d
PC
181152013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
18116
18117 PR c++/55813
18118 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
18119
ff24afc8
GJL
181202013-02-28 Georg-Johann Lay <avr@gjlay.de>
18121
18122 PR target/56445
18123 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
18124 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
18125 INTX_FTYPE_FX, FX_FTYPE_INTX.
18126 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
18127
c7b1fc1b
GJL
181282013-02-28 Georg-Johann Lay <avr@gjlay.de>
18129
18130 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
18131 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18132 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
18133 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
18134 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
18135 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
18136 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
18137 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
18138 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
18139 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
18140 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
18141 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
18142 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
18143 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
18144 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
18145 (avrxmega6): Increase max flash segments from 5 to 6.
18146 * config/avr/t-multilib: Regenerate.
18147 * config/avr/avr-tables.opt: Regenerate.
18148 * doc/avr-mmcu.texi: Regenerate.
18149
4a0e3cfe
GJL
181502013-02-28 Georg-Johann Lay <avr@gjlay.de>
18151
18152 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
18153 (avr_device_to_arch): Rename to avr_device_to_ld.
18154 (avr_device_to_as): New prototype.
18155 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
18156 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
18157 * config/avr/driver-avr.c (avr_device_to_as): New.
18158 (avr_device_to_arch): Rename to avr_device_to_ld.
18159
97785e52
JJ
181602013-02-27 Jakub Jelinek <jakub@redhat.com>
18161
3f292312
JJ
18162 PR middle-end/56461
18163 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
18164 method on dr_chain and result_chain.
18165
a344216b
JJ
18166 PR middle-end/56461
18167 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
18168 pointer_set_destroy on not_executed_last_iteration.
18169
f121ad02 18170 PR middle-end/56461
88b97037 18171 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 18172
307f83a3
JJ
18173 PR middle-end/56461
18174 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
18175 FOR_EACH_DEFINED_FUNCTION when freeing state.
18176
e19624ee
JJ
18177 PR middle-end/56461
18178 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
18179 pool_free.
18180 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
18181 overwriting it.
18182
90e709fd
JJ
18183 PR middle-end/56461
18184 * ipa-cp.c (decide_whether_version_node): Call vec_free on
18185 known_aggs[i].items and release known_aggs vector.
18186
97785e52
JJ
18187 PR middle-end/56461
18188 * ipa-reference.c (propagate): Free node_info even for alias nodes.
18189
181902013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 18191
97785e52
JJ
18192 * config/microblaze/microblaze.c (microblaze_emit_compare):
18193 Use xor for EQ/NE comparisions.
18194 * config/microblaze/microblaze.md (cstoresf4): Add constraints
18195 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 18196
6fa5e0ed
JJ
181972013-02-27 Jakub Jelinek <jakub@redhat.com>
18198
18199 PR middle-end/56461
18200 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
18201 vector.
18202 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
18203 vec_safe_push, always update *slot.
18204 (redirect_edge_var_map_clear): Use vec_free.
18205 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
18206 (free_var_map_entry): Use vec_free.
18207 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
18208 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
18209
436a956a
AB
182102013-02-27 Andrey Belevantsev <abel@ispras.ru>
18211
18212 PR middle-end/45472
436a956a
AB
18213 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
18214 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB
18215 Reorder tests for speculativeness in the logical and operator.
18216
f5c2caca
JJ
182172013-02-27 Jakub Jelinek <jakub@redhat.com>
18218
0fcb564b
JJ
18219 * incpath.c (add_standard_paths): Use reconcat instead of concat
18220 where appropriate and avoid leaking memory.
18221
dc357798
JJ
18222 * opts.h: Include obstack.h.
18223 (opts_concat): New prototype.
18224 (opts_obstack): New declaration.
18225 * opts.c (opts_concat): New function.
18226 (opts_obstack): New variable.
18227 (init_options_struct): Call gcc_init_obstack on opts_obstack.
18228 (finish_options): Use opts_concat instead of concat
18229 and XOBNEWVEC instead of XNEWVEC.
18230 * opts-common.c (generate_canonical_option, decode_cmdline_option,
18231 generate_option): Likewise.
18232 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
18233 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
18234
f5c2caca
JJ
18235 PR target/56455
18236 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18237 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18238
d0163673
JJ
182392013-02-26 Jakub Jelinek <jakub@redhat.com>
18240
18241 PR middle-end/56461
18242 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18243
b5ad2b8e
JR
182442013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18245
18246 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18247 (arm_block_move_unaligned_straight): Likewise.
18248 (arm_adjust_block_mem): Likewise.
18249
9b639e2c
JR
182502013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18251
18252 PR target/48901
18253 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18254 temp, cond and label.
18255 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18256
e13a0ccb
JR
18257 PR target/52500
18258 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18259 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18260
fbe4f171
JR
18261 PR target/52501
18262 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18263 of prologue/epilogue functions.
18264
ae006543
JR
18265 PR target/52550
18266 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18267 Remove unused variable cfa_offset.
18268 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18269
c81369fa
JR
18270 PR target/54639
18271 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18272 type promotion to unsigned.
18273
f8a8fea7
JR
18274 PR target/54640
18275 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18276 for HOST_WIDE_INT of 32 bit / same size as int.
18277 (arm_block_move_unaligned_straight): Likewise.
18278 (arm_adjust_block_mem): Likewise.
18279
f8be5169
JR
18280 PR target/54662
18281 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18282 ALL_CFLAGS.
18283
f1ad3354
MP
182842013-02-26 Marek Polacek <polacek@redhat.com>
18285
18286 PR tree-optimization/56426
88b97037 18287 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 18288
a6af0f42
RB
182892013-02-26 Richard Biener <rguenther@suse.de>
18290
18291 PR target/56444
18292 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18293 unused variable loops.
18294
aca43c6c
JJ
182952013-02-26 Jakub Jelinek <jakub@redhat.com>
18296
cecbe5d9
JJ
18297 PR tree-optimization/56448
18298 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18299 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18300 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18301 later operands of the references, or even first operand for
18302 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18303
aca43c6c
JJ
18304 PR tree-optimization/56443
18305 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18306 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18307 to type_for_mode langhook.
18308
53e2e141
MT
183092013-02-25 Matt Turner <mattst88@gmail.com>
18310
18311 * doc/invoke.texi: Document r4700.
18312
259ee451
RB
183132013-02-25 Richard Biener <rguenther@suse.de>
18314
18315 PR tree-optimization/56175
18316 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18317 split out from ...
18318 (simplify_bitwise_binary): ... here. Also guard the conversion
18319 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18320
6d65e8f1
CM
183212013-02-25 Catherine Moore <clm@codesourcery.com>
18322
18323 Revert:
18324 2013-02-24 Catherine Moore <clm@codesourcery.com>
18325 Maciej W. Rozycki <macro@codesourcery.com>
18326 Tom de Vries <tom@codesourcery.com>
a72d8780 18327 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM
18328 Iain Sandoe <iain@codesourcery.com>
18329 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 18330 Chao-ying Fu <fu@mips.com>
88b97037 18331
6d65e8f1 18332 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 18333 Document new function attributes.
6d65e8f1
CM
18334 * doc/invoke.texi (minterlink-compressed, mmicromips,
18335 m14k, m14ke, m14kec): Document new options.
18336 (minterlink-mips16): Update documentation.
18337 * doc/md.texi (ZC, ZD): Document new constraints.
18338 * configure.ac (gcc_cv_as_micromips): Check if linker
18339 supports the .set micromips directive.
18340 * configure: Regenerate.
18341 * config.in: Regenerate.
18342 * config/mips/mips-tables.opt: Regenerate.
18343 * config/mips/micromips.md: New file.
18344 * constraints.md (ZC, AD): New constraints.
18345 * config/mips/predicates.md (movep_src_register): New predicate.
18346 (movep_src_operand): New predicate.
18347 (non_volatile_mem_operand): New predicate.
18348 * config/mips/mips.md (multimem): New type.
18349 (length): Differentiate between 17-bit and 18-bit branch offsets.
18350 (MOVEP1, MOVEP2): New mode iterator.
18351 (mov_<load>l): Use ZC constraint.
18352 (mov_<load>r): Likewise.
18353 (mov_<store>l): Likewise.
18354 (mov_<store>r): Likewise.
18355 (*branch_equality<mode>_inverted): Add microMIPS support.
18356 (*branch_equality<mode>): Likewise.
18357 (*jump_absolute): Likewise.
18358 (indirect_jump_<mode>): Likewise.
18359 (tablejump_<mode>): Likewise.
18360 (<optab>_internal): Likewise.
18361 (sibcall_internal): Likewise.
18362 (sibcall_value_internal): Likewise.
18363 (prefetch): Use constraint ZD.
18364 * config/mips/mips.opt (minterlink-compressed): New option.
18365 (minterlink-mips16): Now an alias for minterlink-compressed.
18366 (mmicromips): New option.
18367 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18368 (compare_and_swap_12): Likewise.
18369 (sync_add<mode>): Likewise.
18370 (sync_<optab>_12): Likewise.
18371 (sync_old_<optab>_12): Likewise.
18372 (sync_new_<optab>_12): Likewise.
18373 (sync_nand_12): Likewise.
18374 (sync_old_nand_12): Likewise.
18375 (sync_new_nand_12): Likewise.
18376 (sync_sub<mode>): Likewise.
18377 (sync_old_add<mode>): Likewise.
18378 (sync_old_sub<mode>): Likewise.
18379 (sync_new_add<mode>): Likewise.
18380 (sync_new_sub<mode>): Likewise.
18381 (sync_<optab><mode>): Likewise.
18382 (sync_old_<optab><mode>): Likewise.
18383 (sync_new_<optab><mode>): Likewise.
18384 (sync_nand<mode>): Likewise.
18385 (sync_old_nand<mode>): Likewise.
18386 (sync_new_nand<mode>): Likewise.
18387 (sync_lock_test_and_set<mode>): Likewise.
18388 (test_and_set_12): Likewise.
18389 (atomic_compare_and_swap<mode>): Likewise.
18390 (atomic_exchange<mode>_llsc): Likewise.
18391 (atomic_fetch_add<mode>_llsc): Likewise.
18392 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18393 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18394 (umips_save_restore_pattern_p): Likewise.
18395 (umips_load_store_pair_p): Likewise.
18396 (umips_output_load_store_pair): Likewise.
18397 (umips_movep_target_p): Likewise.
18398 (umips_12bit_offset_address_p): Likewise.
18399 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18400 (mips_base_mips16): Rename this...
18401 (mips_base_compression_flags): ...to this. Update all uses.
18402 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18403 (mips_mips16_decl_p): Delete.
18404 (mips_nomips16_decl_p): Delete.
18405 (mips_get_compress_on_flags): New function.
18406 (mips_get_compress_off_flags): New function.
18407 (mips_get_compress_mode): New function.
18408 (mips_get_compress_on_name): New function.
18409 (mips_get_compress_off_name): New function.
18410 (mips_insert_attributes): Support multiple compression types.
18411 (mips_merge_decl_attributes): Likewise.
18412 (umips_12bit_offset_address_p): New function.
18413 (mips_start_function_definition): Emit .set micromips directive.
18414 (mips_call_may_need_jalx_p): New function.
18415 (mips_function_ok_for_sibcall): Add microMIPS support.
18416 (mips_print_operand_punctuation): Support short delay slots and
18417 compact jumps.
18418 (umips_swm_mask, umips_swm_encoding): New.
18419 (umips_build_save_restore): New function.
18420 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18421 (was_mips16_p): Remove.
18422 (old_compression_mode): New.
18423 (mips_set_compression_mode): New function.
18424 (mips_set_current_function): Add microMIPS support.
18425 (mips_option_override): Likewise.
18426 (umips_save_restore_pattern_p): New function.
18427 (umips_output_save_restore): New function.
18428 (umips_load_store_pair_p_1): New function.
18429 (umips_load_store_pair_p): New function.
18430 (umips_output_load_store_pair_1): New function.
18431 (umips_output_load_store_pair): New function.
18432 (umips_movep_target_p) New function.
18433 (mips_prepare_pch_save): Add microMIPS support.
18434 * config/mips/mips.h (TARGET_COMPRESSION): New.
18435 (TARGET_CPU_CPP_BUILTINS): Update macro
18436 to use new compression flags and to support microMIPS.
18437 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18438 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18439 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18440 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18441 (ASM_SPEC): Support mmicromips and mno-micromips.
18442 (M16STORE_REG_P): New macro.
18443 (MIPS_CALL): Support TARGET_MICROMIPS.
18444 (MICROMIPS_J): New macro.
18445 (mips_base_mips16): Rename this...
18446 (mips_base_compression_flags): ...to this.
18447 (UMIPS_12BIT_OFFSET_P): New macro.
18448 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18449 (MULTILIB_DIRNAMES): Likewise.
18450
5e5df392
TV
184512013-02-25 Tom de Vries <tom@codesourcery.com>
18452
18453 PR rtl-optimization/56131
18454 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18455 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18456 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18457
3fdb53c1
TB
184582013-02-25 Tobias Burnus <burnus@net-b.de>
18459
18460 * doc/invoke.texi (-fsanitize=): Move from optimization
18461 to debugging options.
18462
ed358aea
AB
184632013-02-25 Andrey Belevantsev <abel@ispras.ru>
18464
18465 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18466
f45e9053
AB
184672013-02-25 Andrey Belevantsev <abel@ispras.ru>
18468 Alexander Monakov <amonakov@ispras.ru>
18469
18470 PR middle-end/56077
18471 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18472 flush pending lists also on non-jumps. Adjust comment.
18473
6941b508
CM
184742013-02-24 Catherine Moore <clm@codesourcery.com>
18475 Maciej W. Rozycki <macro@codesourcery.com>
18476 Tom de Vries <tom@codesourcery.com>
a72d8780 18477 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM
18478 Iain Sandoe <iain@codesourcery.com>
18479 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 18480 Chao-ying Fu <fu@mips.com>
6941b508
CM
18481
18482 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 18483 Document new function attributes.
6941b508
CM
18484 * doc/invoke.texi (minterlink-compressed, mmicromips,
18485 m14k, m14ke, m14kec): Document new options.
18486 (minterlink-mips16): Update documentation.
18487 * doc/md.texi (ZC, ZD): Document new constraints.
18488 * configure.ac (gcc_cv_as_micromips): Check if linker
18489 supports the .set micromips directive.
18490 * configure: Regenerate.
18491 * config.in: Regenerate.
18492 * config/mips/mips-tables.opt: Regenerate.
18493 * config/mips/micromips.md: New file.
18494 * constraints.md (ZC, AD): New constraints.
18495 * config/mips/predicates.md (movep_src_register): New predicate.
18496 (movep_src_operand): New predicate.
18497 (non_volatile_mem_operand): New predicate.
18498 * config/mips/mips.md (multimem): New type.
18499 (length): Differentiate between 17-bit and 18-bit branch offsets.
18500 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 18501 (mov_<load>l): Use ZC constraint.
6941b508
CM
18502 (mov_<load>r): Likewise.
18503 (mov_<store>l): Likewise.
18504 (mov_<store>r): Likewise.
18505 (*branch_equality<mode>_inverted): Add microMIPS support.
18506 (*branch_equality<mode>): Likewise.
18507 (*jump_absolute): Likewise.
18508 (indirect_jump_<mode>): Likewise.
18509 (tablejump_<mode>): Likewise.
18510 (<optab>_internal): Likewise.
18511 (sibcall_internal): Likewise.
18512 (sibcall_value_internal): Likewise.
18513 (prefetch): Use constraint ZD.
18514 * config/mips/mips.opt (minterlink-compressed): New option.
18515 (minterlink-mips16): Now an alias for minterlink-compressed.
18516 (mmicromips): New option.
18517 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18518 (compare_and_swap_12): Likewise.
18519 (sync_add<mode>): Likewise.
18520 (sync_<optab>_12): Likewise.
18521 (sync_old_<optab>_12): Likewise.
18522 (sync_new_<optab>_12): Likewise.
18523 (sync_nand_12): Likewise.
18524 (sync_old_nand_12): Likewise.
18525 (sync_new_nand_12): Likewise.
18526 (sync_sub<mode>): Likewise.
18527 (sync_old_add<mode>): Likewise.
18528 (sync_old_sub<mode>): Likewise.
18529 (sync_new_add<mode>): Likewise.
18530 (sync_new_sub<mode>): Likewise.
18531 (sync_<optab><mode>): Likewise.
18532 (sync_old_<optab><mode>): Likewise.
18533 (sync_new_<optab><mode>): Likewise.
18534 (sync_nand<mode>): Likewise.
18535 (sync_old_nand<mode>): Likewise.
18536 (sync_new_nand<mode>): Likewise.
18537 (sync_lock_test_and_set<mode>): Likewise.
18538 (test_and_set_12): Likewise.
18539 (atomic_compare_and_swap<mode>): Likewise.
18540 (atomic_exchange<mode>_llsc): Likewise.
18541 (atomic_fetch_add<mode>_llsc): Likewise.
18542 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18543 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18544 (umips_save_restore_pattern_p): Likewise.
18545 (umips_load_store_pair_p): Likewise.
18546 (umips_output_load_store_pair): Likewise.
18547 (umips_movep_target_p): Likewise.
18548 (umips_12bit_offset_address_p): Likewise.
18549 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18550 (mips_base_mips16): Rename this...
18551 (mips_base_compression_flags): ...to this. Update all uses.
18552 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18553 (mips_mips16_decl_p): Delete.
18554 (mips_nomips16_decl_p): Delete.
3fdb53c1 18555 (mips_get_compress_on_flags): New function.
6941b508
CM
18556 (mips_get_compress_off_flags): New function.
18557 (mips_get_compress_mode): New function.
18558 (mips_get_compress_on_name): New function.
18559 (mips_get_compress_off_name): New function.
18560 (mips_insert_attributes): Support multiple compression types.
18561 (mips_merge_decl_attributes): Likewise.
18562 (umips_12bit_offset_address_p): New function.
18563 (mips_start_function_definition): Emit .set micromips directive.
18564 (mips_call_may_need_jalx_p): New function.
18565 (mips_function_ok_for_sibcall): Add microMIPS support.
18566 (mips_print_operand_punctuation): Support short delay slots and
18567 compact jumps.
18568 (umips_swm_mask, umips_swm_encoding): New.
18569 (umips_build_save_restore): New function.
18570 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18571 (was_mips16_p): Remove.
18572 (old_compression_mode): New.
18573 (mips_set_compression_mode): New function.
18574 (mips_set_current_function): Add microMIPS support.
18575 (mips_option_override): Likewise.
18576 (umips_save_restore_pattern_p): New function.
18577 (umips_output_save_restore): New function.
18578 (umips_load_store_pair_p_1): New function.
18579 (umips_load_store_pair_p): New function.
18580 (umips_output_load_store_pair_1): New function.
18581 (umips_output_load_store_pair): New function.
18582 (umips_movep_target_p) New function.
18583 (mips_prepare_pch_save): Add microMIPS support.
18584 * config/mips/mips.h (TARGET_COMPRESSION): New.
18585 (TARGET_CPU_CPP_BUILTINS): Update macro
18586 to use new compression flags and to support microMIPS.
18587 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18588 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18589 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18590 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18591 (ASM_SPEC): Support mmicromips and mno-micromips.
18592 (M16STORE_REG_P): New macro.
18593 (MIPS_CALL): Support TARGET_MICROMIPS.
18594 (MICROMIPS_J): New macro.
18595 (mips_base_mips16): Rename this...
18596 (mips_base_compression_flags): ...to this.
18597 (UMIPS_12BIT_OFFSET_P): New macro.
18598 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18599 (MULTILIB_DIRNAMES): Likewise.
18600
cdebbc6d
JJ
186012013-02-24 Jakub Jelinek <jakub@redhat.com>
18602
18603 PR target/52555
18604 * target-globals.c (save_target_globals): For init_reg_sets and
18605 target_reinit remporarily set this_fn_optabs to this_target_optabs.
18606
18c63565
JG
186072013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
18608
18609 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
18610 * config/aarch64/t-aarch64
18611 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
18612
2194f7a2
VM
186132013-02-22 Vladimir Makarov <vmakarov@redhat.com>
18614
18615 PR inline-asm/56148
18616 * lra-constraints.c (process_alt_operands): Reload operand
18617 conflicting with earlier clobber only if no more other conflicting
18618 operands.
18619
7d613735
JJ
186202013-02-22 Jakub Jelinek <jakub@redhat.com>
18621
18622 PR sanitizer/56393
18623 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
18624 if not linking a shared library.
18625
ac8d93a7
SL
186262013-02-22 Seth LaForge <sethml@google.com>
18627
18628 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
18629
e0237780
GY
186302013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
18631
3fdb53c1
TB
18632 * config/arm/arm.md (split for extendsidi): Update condition.
18633 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
18634 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
18635 (qhs_zextenddi_cstr): Likewise.
e0237780 18636
d7fde18c
JJ
186372013-02-21 Jakub Jelinek <jakub@redhat.com>
18638
be63b77d
JJ
18639 PR middle-end/56420
18640 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
18641 avoid signed wrapping.
18642 (expand_mult): Handle properly multiplication by
18643 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
18644 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
18645 in the compiler if coeff is HOST_WIDE_INT_MIN.
18646 (expand_divmod): Don't make ext_op1 static, change it's type to
18647 uhwi. Avoid undefined behavior in -INTVAL (op1).
18648
d7fde18c
JJ
18649 PR rtl-optimization/50339
18650 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
18651 field.
18652 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
18653 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
18654 into splitting_ashiftrt field.
18655 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
18656 ASHIFTRT.
18657 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
18658 choices.
18659
6aad4455
AH
186602013-02-20 Aldy Hernandez <aldyh@redhat.com>
18661
18662 PR middle-end/56108
18663 * trans-mem.c (execute_tm_mark): Do not expand transactions that
18664 are sure to go irrevocable.
18665
38fe784d
HPN
186662013-02-21 Hans-Peter Nilsson <hp@axis.com>
18667
18668 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
18669 scalars are valid operands.
18670
0fd44da3
MJ
186712013-02-21 Martin Jambor <mjambor@suse.cz>
18672
18673 PR tree-optimization/56310
18674 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
18675 only matching indices and non-negative final offsets.
18676 (intersect_aggregates_with_edge): Pass src_idx to
18677 agg_replacements_to_vector. Pass src_idx insstead of index to
18678 intersect_with_agg_replacements.
18679
7a92038b
MJ
186802013-02-21 Martin Jambor <mjambor@suse.cz>
18681
18682 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
18683 instead of hard-wired defaults.
18684
c0da9c37
MR
186852013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
18686
18687 * doc/invoke.texi (MIPS Options): Update documentation of the
18688 floating-point multiply-accumulate instruction restrictions.
18689
d247ea0c 186902013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS
18691
18692 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
18693 asan_shadow_offset on x86_64 linux.
18694
22deefcb
RB
186952013-02-21 Richard Biener <rguenther@suse.de>
18696
18697 PR tree-optimization/56415
18698 Revert
18699 2013-02-11 Richard Biener <rguenther@suse.de>
18700
18701 PR tree-optimization/56273
18702 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18703 first VRP run.
18704
7bcc6e75
JJ
187052013-02-21 Jakub Jelinek <jakub@redhat.com>
18706
a0ad148f
JJ
18707 PR bootstrap/56258
18708 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
18709 instead of @itemx.
18710
7bcc6e75
JJ
18711 PR inline-asm/56405
18712 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
18713 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
18714
a0a7b611
JH
187152013-02-20 Jan Hubicka <jh@suse.cz>
18716
18717 PR tree-optimization/56265
88b97037
UB
18718 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
18719 when target is referenced for first time.
a0a7b611 18720
c0e50f72
RB
187212013-02-20 Richard Biener <rguenther@suse.de>
18722
18723 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
18724 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
18725 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
18726 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
18727 not return anything.
18728 (rename_ssa_copies): Do not remove unused locals.
18729 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 18730 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB
18731 * passes.c (execute_function_todo): Do not schedule unused locals
18732 removal if cleanup_tree_cfg did something.
18733 * tree-ssa-live.c (remove_unused_locals): Dump statistics
18734 about the number of removed locals.
18735
a52ca739
RB
187362013-02-20 Richard Biener <rguenther@suse.de>
18737
18738 PR tree-optimization/56398
88b97037 18739 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 18740
ec9c9d1b
MJ
187412013-02-20 Martin Jambor <mjambor@suse.cz>
18742
18743 PR tree-optimization/55334
18744 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
18745 restricted pointers to arrays.
18746
e91c8ed6 187472013-02-20 Richard Biener <rguenther@suse.de>
88b97037 18748 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB
18749
18750 PR tree-optimization/56396
18751 * tree-ssa-ccp.c (n_const_val): New static variable.
18752 (get_value): Return NULL for SSA names we don't have a lattice
18753 entry for.
18754 (ccp_initialize): Initialize n_const_val.
18755 * tree-ssa-copy.c (n_copy_of): New static variable.
18756 (init_copy_prop): Initialize n_copy_of.
18757 (get_value): Return NULL_TREE for SSA names we don't have a
18758 lattice entry for.
18759
3d916479
MJ
187602013-02-20 Martin Jambor <mjambor@suse.cz>
18761
18762 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
18763
71a86758
RB
187642013-02-20 Richard Biener <rguenther@suse.de>
18765
18766 * genpreds.c (write_lookup_constraint): Do not compare first
18767 letter of the constraint again.
18768
79836a12
RB
187692013-02-20 Richard Biener <rguenther@suse.de>
18770
18771 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
18772 and ceil_log2.
18773 (get_use_iv_cost): Terminate hashtable walk when coming across
18774 an empty entry.
18775
bbe4fb2c
IZ
187762013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
18777
18778 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
18779 reassociation for avx2 targets.
18780
a72d8780 187812012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 18782
c77f83d5
EI
18783 * config/microblaze/microblaze.c: microblaze_has_clz = 0
18784 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 18785 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI
18786 version and TARGET_PATTERN_COMPARE check
18787 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 18788
a72d8780 187892012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 18790
a72d8780 18791 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI
18792 function before branching.
18793
242387fa
AB
187942012-02-19 Andrey Belevantsev <abel@ispras.ru>
18795
18796 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
18797 DUMP_INSN_RTX_UID.
18798 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
18799
b4979ab9
AB
188002012-02-19 Andrey Belevantsev <abel@ispras.ru>
18801
18802 PR middle-end/55889
b4979ab9
AB
18803 * sel-sched.c: Include ira.h.
18804 (implicit_clobber_conflict_p): New function.
18805 (moveup_expr): Use it.
88b97037 18806 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 18807
24d63016
RB
188082013-02-19 Richard Biener <rguenther@suse.de>
18809
18810 PR tree-optimization/56384
18811 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
18812 (vn_hash_type): Split out from ...
18813 (vn_hash_constant_with_type): ... here.
18814 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
18815 (vn_phi_eq): Compare types from vn_phi_s structure.
18816 (vn_phi_lookup): Populate vn_phi_s type.
18817 (vn_phi_insert): Likewise.
18818
a475fd3d
JJ
188192013-02-19 Jakub Jelinek <jakub@redhat.com>
18820
47cc28f5
JJ
18821 PR tree-optimization/56350
18822 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
18823 if haven't found reduction or nested cycle operand, rather than
18824 asserting we must find it.
18825
a475fd3d
JJ
18826 PR tree-optimization/56381
18827 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
18828 to fold_build3.
18829
188302013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH
18831 Jakub Jelinek <jakub@redhat.com>
18832
18833 PR target/52555
18834 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
18835 (swap_optab_enable): Same.
18836 (init_all_optabs): Use argument instead of global.
88b97037 18837 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH
18838 * expr.h (init_all_optabs): Add argument to prototype.
18839 (TREE_OPTIMIZATION_OPTABS): New.
18840 (save_optabs_if_changed): Protoize.
18841 * optabs.h: Declare this_fn_optabs.
18842 * optabs.c (save_optabs_if_changed): New.
18843 Declare this_fn_optabs.
18844 (init_optabs): Add argument to init_all_optabs() call.
18845 * function.c (invoke_set_current_function_hook): Handle per
18846 function optabs.
18847 * function.h (struct function): New field optabs.
18848 * config/mips/mips.c (mips_set_mips16_mode): Handle when
18849 optimization_current_node has changed.
18850 * target-globals.h (save_target_globals_default_opts): Protoize.
18851 * target-globals.c (save_target_globals_default_opts): New.
18852
3f587ca3
JDA
188532013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18854
18855 PR target/56347
18856 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
18857 registers %fr12 and %fr12R as call used.
ceaca33e
JDA
18858
18859 PR target/56214
18860 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB
18861 and HImode, require all displacements to be an integer multiple of
18862 their mode size.
ceaca33e
JDA
18863 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
18864 only allow QImode and HImode when reload is in progress and strict is
18865 true. Likewise for symbolic addresses. Use base14_operand to check
18866 displacements in REG+BASE addresses.
18867
fe0b4796
RB
188682013-02-18 Richard Biener <rguenther@suse.de>
18869
18870 PR tree-optimization/56366
18871 * tree-vect-loop.c (get_initial_def_for_induction): Properly
18872 handle sign-conversion of outer-loop initial induction value.
18873
6aaf596b
RB
188742013-02-18 Richard Biener <rguenther@suse.de>
18875
73db8ff1 18876 PR middle-end/56349
6aaf596b
RB
18877 * cfghooks.c (merge_blocks): If we merge a latch into another
18878 block adjust references to it.
18879 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
18880 (verify_loop_structure): Verify that a recorded latch is in fact
18881 a latch.
18882
5e97dfb6
RB
188832013-02-18 Richard Biener <rguenther@suse.de>
18884
18885 PR tree-optimization/56321
18886 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
18887 order SSA name release and virtual operand unlinking.
18888
825527e8
EI
188892013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18890
18891 * config/microblaze/microblaze.md (save_stack_block): Define.
18892 (restore_stack_block): Likewise.
18893
debd11d9
EI
188942013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18895
18896 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
18897 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
18898 * config/microblaze/microblaze.c (microblaze_option_override):
18899 Bail out early for PIC modes when target does not support PIC.
18900
8ec77be0
EI
189012013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18902
88b97037 18903 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI
18904 Replace with a microblaze version.
18905 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 18906 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI
18907 microblaze.
18908
b41288b3
JJ
189092013-02-16 Jakub Jelinek <jakub@redhat.com>
18910 Dodji Seketeli <dodji@redhat.com>
18911
18912 PR asan/56330
88b97037 18913 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ
18914 (instrument_mem_region_access): Do not forget to always put
18915 instrumentation of the of 'base' and 'base + len' in a "if (len !=
18916 0) statement, even for cases where either 'base' or 'base + len'
18917 are not instrumented -- because they have been previously
18918 instrumented. Simplify the logic by putting all the statements
18919 instrument 'base + len' inside a sequence, and then insert that
18920 sequence right before the current insertion point. Then, to
18921 instrument 'base + len', just get an iterator on that statement.
18922 And do not forget to update the pointer to iterator the function
18923 received as argument.
18924
47918951
VM
189252013-02-15 Vladimir Makarov <vmakarov@redhat.com>
18926
18927 PR rtl-optimization/56348
18928 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
18929
3fb7c699
SB
189302013-02-15 Steven Bosscher <steven@gcc.gnu.org>
18931
18932 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
18933 (clean_graph_dump_file): Pass base to start_graph_dump.
18934
576fe41a
RH
189352013-02-14 Richard Henderson <rth@redhat.com>
18936
18937 PR target/55941
18938 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
18939
bb3accfa
SB
189402013-02-14 Steven Bosscher <steven@gcc.gnu.org>
18941
18942 * collect2-aix.h: Define F_LOADONLY.
18943
03143140
RB
189442013-02-14 Richard Biener <rguenther@suse.de>
18945
18946 PR lto/50494
18947 * varasm.c (output_constant_def_1): Get the decl representing
18948 the constant as argument.
18949 (output_constant_def): Wrap output_constant_def_1.
18950 (make_decl_rtl): Use output_constant_def_1 with the decl
18951 representing the constant.
18952 (build_constant_desc): Optionally re-use a decl already
18953 representing the constant.
18954 (tree_output_constant_def): Adjust.
18955
8fb06726
DS
189562013-02-14 Dodji Seketeli <dodji@redhat.com>
18957
18958 Fix an asan crash
18959 * asan.c (instrument_builtin_call): Really put the length of the
18960 second source argument into src1_len.
18961
c4bfe8bf
JJ
189622013-02-13 Jakub Jelinek <jakub@redhat.com>
18963
18964 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
18965 argument. If it is false, don't create edge from then_bb to
18966 fallthru_bb.
18967 (insert_if_then_before_iter): Pass true to it.
18968 (build_check_stmt): Pass false to it.
18969 (transform_statements): Flush hash table only on extended basic
18970 block boundaries, rather than at the beginning of every bb.
18971 Don't flush hash table on nonfreeing_call_p calls.
18972 * tree-flow.h (nonfreeing_call_p): New prototype.
18973 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
18974
7afe2801
DM
189752013-02-13 David S. Miller <davem@davemloft.net>
18976
18977 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
18978
70cc3288
VM
189792013-02-13 Vladimir Makarov <vmakarov@redhat.com>
18980
18981 PR target/56184
18982 * ira.c (max_regno_before_ira): Move from ...
18983 (ira): ... here.
18984 (fix_reg_equiv_init): Use max_regno_before_ira instead of
18985 vec_safe_length.
18986
6422242b
JJ
189872013-02-13 Jakub Jelinek <jakub@redhat.com>
18988
18989 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
18990
976a81ee
RB
189912013-02-13 Richard Biener <rguenther@suse.de>
18992
18993 PR lto/56295
18994 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
18995 globals in MEM_REFs.
18996
c1874a87
RB
189972013-02-13 Richard Biener <rguenther@suse.de>
18998
18999 * loop-init.c (loop_optimizer_init): Clear loop state when
19000 re-initializing preserved loops.
19001 * loop-unswitch.c (unswitch_single_loop): Return whether
19002 we unswitched the loop. Do not verify loop state here.
88b97037 19003 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 19004
b4ab7d34
KS
190052013-02-13 Kostya Serebryany <kcc@google.com>
19006
88b97037
UB
19007 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
19008 on x86_64 linux.
b4ab7d34
KS
19009 * sanitizer.def: Rename __asan_init to __asan_init_v1.
19010
bdcbe80c
DS
190112013-02-12 Dodji Seketeli <dodji@redhat.com>
19012
19013 Avoid instrumenting duplicated memory access in the same basic block
19014 * Makefile.in (asan.o): Add new dependency on hash-table.h
19015 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
19016 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
19017 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
19018 (free_mem_ref_resources, has_mem_ref_been_instrumented)
19019 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
19020 (get_mem_ref_of_assignment): New functions.
19021 (get_mem_refs_of_builtin_call): Extract from
19022 instrument_builtin_call and tweak a little bit to make it fit with
19023 the new signature.
19024 (instrument_builtin_call): Use the new
19025 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
19026 of is_gimple_builtin_call.
19027 (instrument_derefs, instrument_mem_region_access): Insert the
19028 instrumented memory reference into the hash table.
19029 (maybe_instrument_assignment): Renamed instrument_assignment into
19030 this, and change it to advance the iterator when instrumentation
19031 actually happened and return true in that case. This makes it
19032 homogeneous with maybe_instrument_assignment, and thus give a
19033 chance to callers to be more 'regular'.
19034 (transform_statements): Clear the memory reference hash table
19035 whenever we enter a new BB, when we cross a function call, or when
19036 we are done transforming statements. Use
19037 maybe_instrument_assignment instead of instrumentation. No more
19038 need to special case maybe_instrument_assignment and advance the
19039 iterator after calling it; it's now handled just like
19040 maybe_instrument_call. Update comment.
19041
4861a1f7
RB
190422013-02-13 Richard Biener <rguenther@suse.de>
19043
19044 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
19045 Fix loop discovery code.
19046
1c86bd80
VM
190472013-02-12 Vladimir Makarov <vmakarov@redhat.com>
19048
19049 PR inline-asm/56148
19050 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 19051 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM
19052 the operand is not reloaded. Prefer to reload conflicting operand
19053 if earlyclobber and matching operands are the same.
19054
7cbda518
RB
190552013-02-12 Richard Biener <rguenther@suse.de>
19056
19057 PR lto/56297
19058 * lto-streamer-out.c (write_symbol): Do not output symbols
19059 for hard register variables.
19060
e68a4ef6
GJL
190612013-02-12 Georg-Johann Lay <avr@gjlay.de>
19062
19063 PR target/54222
19064 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
19065 (umulsidi3_insn, mulsidi3_insn): New insns.
19066
a72d8780 190672013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL
19068
19069 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
19070 (struct tune_params): Add vec_costs field.
19071 * config/arm/arm.c (arm_builtin_vectorization_cost)
19072 (arm_add_stmt_cost): New functions.
19073 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
19074 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
19075 (arm_default_vec_cost): New struct of type cpu_vec_costs.
19076 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
19077 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
19078 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
19079 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
19080
43320568
RB
190812013-02-12 Richard Biener <rguenther@suse.de>
19082
19083 PR lto/56295
19084 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
19085 decls again if possible.
19086
a011aa39
RB
190872013-02-12 Richard Biener <rguenther@suse.de>
19088
19089 PR middle-end/56288
19090 * tree-ssa.c (verify_ssa_name): Fix check, move
19091 SSA_NAME_IN_FREE_LIST check up.
19092
6da26889
JJ
190932013-02-12 Jakub Jelinek <jakub@redhat.com>
19094 Steven Bosscher <steven@gcc.gnu.org>
19095
19096 PR rtl-optimization/56151
19097 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
19098 equal to op0 or op1, and last_insn pattern is CODE operation
19099 with MEM dest and one of the operands matches that MEM.
19100
f80e0faf
ST
191012013-02-11 Sriraman Tallam <tmsriramgoogle.com>
19102
19103 * doc/extend.texi: Document Function Multiversioning and "default"
19104 parameter string to target attribute.
19105 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
19106 target attribute parameter is "default".
19107 (ix86_compare_version_priority): Remove checks for target attribute.
19108 (ix86_mangle_function_version_assembler_name): Change error to sorry.
19109 Remove check for target attribute equal to NULL. Add assert.
19110 (ix86_generate_version_dispatcher_body): Change error to sorry.
19111
6c59ffd1
IS
191122013-02-11 Iain Sandoe <iain@codesourcery.com>
19113 Jack Howarth <howarth@bromo.med.uc.edu>
19114 Patrick Marlier <patrick.marlier@gmail.com>
19115
19116 PR libitm/55693
19117 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
19118 define ENDFILE_SPEC as TM_DESTRUCTOR.
19119 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
19120
a72d8780 191212013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP
19122 Jack Howarth <howarth@bromo.med.uc.edu>
19123 Jakub Jelinek <jakub@redhat.com>
19124
19125 PR sanitizer/55617
19126 * config/darwin.c (cdtor_record): Rename ctor_record.
19127 (sort_cdtor_records): Rename sort_ctor_records.
19128 (finalize_dtors): New routine to sort destructors by
19129 priority before use in assemble_integer.
19130 (machopic_asm_out_destructor): Use finalize_dtors if needed.
19131
b63fe007
UB
191322013-02-11 Uros Bizjak <ubizjak@gmail.com>
19133
19134 PR rtl-optimization/56275
19135 * simplify-rtx.c (avoid_constant_pool_reference): Check that
19136 offset is non-negative and less than cmode size before
19137 calling simplify_subreg.
19138
8e89b5b5
RB
191392013-02-11 Richard Biener <rguenther@suse.de>
19140
19141 PR tree-optimization/56264
19142 * cfgloop.h (fix_loop_structure): Adjust prototype.
19143 * loop-init.c (fix_loop_structure): Return the number of
19144 newly discovered loops.
19145 * tree-cfgcleanup.c (repair_loop_structures): When new loops
19146 are discovered, do a full loop-closed SSA rewrite.
19147
b4a4b56d
RB
191482013-02-11 Richard Biener <rguenther@suse.de>
19149
19150 PR tree-optimization/56273
19151 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19152 first VRP run.
19153 (check_array_ref): Fix missing newline in dumps.
19154 (search_for_addr_array): Likewise.
19155
0c885229
DE
191562013-02-09 David Edelsohn <dje.gcc@gmail.com>
19157
19158 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
19159
59ac9a55
JJ
191602013-02-09 Jakub Jelinek <jakub@redhat.com>
19161
19162 PR target/56256
19163 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
19164
25bb0bb5
VM
191652013-02-08 Vladimir Makarov <vmakarov@redhat.com>
19166
19167 PR rtl-optimization/56246
0c885229 19168 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM
19169 reload pseudo.
19170 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
19171 constraints are satisfied.
19172
a698cc03
JL
191732013-02-08 Jeff Law <law@redhat.com>
19174
19175 PR debug/53948
19176 * emit-rtl.c (reg_is_parm_p): New function.
19177 * regs.h (reg_is_parm_p): New prototype.
19178 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
19179 callee-clobbered registers.
19180
e1122ddd
MM
191812013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
19182
19183 PR target/56043
19184 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
19185 If there is no implicit builtin declaration, just return NULL.
19186
19c5f6e6
UB
191872013-02-08 Uros Bizjak <ubizjak@gmail.com>
19188
19189 * config/i386/sse.md (FMAMODEM): New mode iterator.
19190 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
19191 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
19192
2480f2ca 191932013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 19194
2480f2ca
UB
19195 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
19196 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
19197 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
19198
191992013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19200
19201 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
19202 (microblaze*-*-elf): Likewise.
19203 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
19204 LINK_SPEC.
19205 * config/microblaze/microblaze-c.c: Add builtin defines for
19206 _LITTLE_ENDIAN and _BIG_ENDIAN.
19207 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
19208 add to TARGET_DEFAULT flags.
76ef61fb 19209 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB
19210 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
19211 * config/microblaze/microblaze.md: Update extendsidi2 and
19212 movdi_internal instructions to use low-order / high-order reg
19213 print_operands.
19214 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
19215 options and inversemask / mask of LITTLE_ENDIAN.
19216 * config/microblaze/t-microblaze: Expand multilib options to
19217 include mlittle-endian (le) and update exceptions patterns.
19218
600a5961
JJ
192192013-02-08 Jakub Jelinek <jakub@redhat.com>
19220
5df81313
JJ
19221 PR rtl-optimization/56195
19222 * lra-constraints.c (get_reload_reg): Don't reuse regs
19223 if they have smaller mode than requested, if they have
19224 wider mode than requested, try to return a SUBREG.
19225
600a5961
JJ
19226 PR tree-optimization/56250
19227 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
19228 if type is unsigned and code isn't MULT_EXPR.
19229
ff544649
GJL
192302013-02-08 Georg-Johann Lay <avr@gjlay.de>
19231
19232 PR tree-optimization/56064
19233 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
19234 bits according to mode.
19235 * fixed-value.h (fixed_from_double_int)
19236 (const_fixed_from_double_int): Adjust comments.
19237
e45cde98
RB
192382013-02-08 Richard Biener <rguenther@suse.de>
19239
19240 PR lto/56231
19241 * lto-streamer.h (struct data_in): Remove current_file, current_line
19242 and current_col members.
19243 * lto-streamer-out.c (lto_output_location): Stream changed bits
19244 en-block for efficiency.
19245 * lto-streamer-in.c (clear_line_info): Remove.
19246 (lto_input_location): Cache current file, line and column
19247 globally via local statics. Read changed bits en-block.
19248 (input_function): Do not call clear_line_info.
19249 (lto_read_body): Likewise.
19250 (lto_input_toplevel_asms): Likewise.
19251
c1ca73d8
MM
192522013-02-08 Michael Matz <matz@suse.de>
19253
19254 PR tree-optimization/52448
19255 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19256 (nt_call_phase): New static.
19257 (add_or_mark_expr): Only mark accesses with newer phase than any
19258 call seen.
19259 (nonfreeing_call_p): New.
19260 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19261 (nt_fini_block): Keep blocks marked as visited.
19262 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19263
57548aa2
RB
192642013-02-08 Richard Biener <rguenther@suse.de>
19265
19266 * ira.c (ira): Free broken dominator information.
19267
8e10366f
UB
192682013-02-08 Uros Bizjak <ubizjak@gmail.com>
19269
19270 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19271
f64fb0fa
MP
192722013-02-08 Marek Polacek <polacek@redhat.com>
19273
8e10366f 19274 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 19275
0375167b
RB
192762013-02-08 Richard Biener <rguenther@suse.de>
19277
19278 PR middle-end/56181
19279 * cfgloop.h (flow_loops_find): Adjust.
19280 (bb_loop_header_p): Declare.
19281 * cfgloop.c (bb_loop_header_p): New function split out from ...
19282 (flow_loops_find): ... here. Adjust function signature,
19283 support incremental loop structure update.
19284 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19285 * cfgloopmanip.c (fix_loop_structure): Move ...
19286 * loop-init.c (fix_loop_structure): ... here.
19287 (apply_loop_flags): Split out from ...
19288 (loop_optimizer_init): ... here.
19289 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19290 in incremental mode, only remove dead loops here.
19291
85d768f3
GJL
192922013-02-08 Georg-Johann Lay <avr@gjlay.de>
19293
19294 PR target/54222
19295 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19296 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19297 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19298 (*round<mode>3.libgcc): New insns for fixed-modes.
19299 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19300 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19301 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19302 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19303 implementations. Define to __builtin_avr_absFX,
19304 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19305 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19306 __builtin_avr_countlsFX, respectively.
19307 * config/avr/avr-c.c (target.h): Include it.
19308 (enum avr_builtin_id): New enum.
19309 (avr_resolve_overloaded_builtin): New static function.
19310 (avr_register_target_pragmas): Use it to set
19311 targetm.resolve_overloaded_builtin.
19312 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19313 tree nodes used by DEF_BUILTIN.
19314 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19315 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19316 <AVR_BUILTIN_xxBITS>: Same.
19317
661bc682
RB
193182013-02-08 Richard Biener <rguenther@suse.de>
19319
19320 * cfgloop.c (verify_loop_structure): Properly handle
19321 a loop exiting to another loop header.
19322 * ira-int.h (ira_loops): Remove.
19323 * ira.c (ira_loops): Remove.
19324 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19325 (do_reload): Use loop_optimizer_finalize.
19326 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19327 number_of_loops to access the loop tree.
19328 (more_one_region_p): Likewise.
19329 (finish_loop_tree_nodes): Likewise.
19330 (rebuild_regno_allocno_maps): Likewise.
19331 (mark_loops_for_removal): Likewise.
19332 (mark_all_loops_for_removal): Likewise.
19333 (remove_unnecessary_regions): Likewise.
19334 (ira_build): Likewise.
19335 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19336
0d5049b2
RB
193372013-02-08 Richard Biener <rguenther@suse.de>
19338
19339 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
19340 * ipa-pure-const.c (analyze_function): Avoid calling
19341 mark_irreducible_loops twice.
8e10366f 19342 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 19343
499f32e8
DM
193442013-02-07 David S. Miller <davem@davemloft.net>
19345
19346 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
19347 on 'reg'.
19348 * var-tracking.c (vt_add_function_parameter): Test the presence of
19349 HAVE_window_save properly and do not remap argument registers when
19350 we have a leaf function.
19351
6edc3e32
UB
193522013-02-07 Uros Bizjak <ubizjak@gmail.com>
19353
19354 PR bootstrap/56227
19355 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
19356 instead of "ll".
19357 * config/i386/i386.c (ix86_print_operand): Ditto.
19358
5306401f
VM
193592013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19360
6edc3e32 19361 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 19362
027ece11
VM
193632013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19364
19365 PR rtl-optimization/56225
19366 * lra-constraints.c (process_alt_operands): Check that reload hard
19367 reg can hold value for strict_low_part.
19368
f980dfdb
JJ
193692013-02-07 Jakub Jelinek <jakub@redhat.com>
19370
19371 PR debug/56154
19372 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
19373 dwarf2out_end_function.
19374 (in_first_function_p, maybe_at_text_label_p,
19375 first_loclabel_num_not_at_text_label): New variables.
19376 (dwarf2out_var_location): In the first function find out
6edc3e32 19377 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ
19378 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
19379 functions.
19380
73dd3123
EB
193812013-02-07 Eric Botcazou <ebotcazou@adacore.com>
19382
19383 PR rtl-optimization/56178
19384 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
19385 SUBREG of a register. Tidy up related block of code.
19386 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
19387 note if the source is a register or a SUBREG of a register.
19388
e3936f47
JJ
193892013-02-07 Jakub Jelinek <jakub@redhat.com>
19390
19391 PR target/56228
19392 * config/rs6000/rs6000.md (ptrm): New mode attr.
19393 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19394 call_value_indirect_aix<pttrsize>,
19395 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19396 m in constraints.
19397
d96d674b
MH
193982013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19399
19400 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19401 if -bnortl. Convert to strcmp and strncmp.
19402
921f2dee
AM
194032013-02-07 Alan Modra <amodra@gmail.com>
19404
19405 PR target/54009
19406 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19407 addresses won't wrap when offsetting.
19408 (rs6000_secondary_reload): Provide secondary reloads needed for
19409 wrapping LO_SUM addresses.
19410
d09c7dba
TS
194112013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19412
19413 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19414 MACH, just __MACH__.
19415
a44bbd48
RB
194162013-02-06 Richard Biener <rguenther@suse.de>
19417
19418 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19419 instead of calling fix_loop_structure.
19420
6d840d99
JJ
194212013-02-06 Jakub Jelinek <jakub@redhat.com>
19422
19423 PR middle-end/56217
19424 * omp-low.c (use_pointer_for_field): Return false if
19425 lower_send_shared_vars doesn't generate any copy-out code.
19426
0f33baa9
TV
194272013-02-06 Tom de Vries <tom@codesourcery.com>
19428
19429 PR rtl-optimization/56131
19430 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19431 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19432 of the label is NULL. Add comment.
19433
11f1e3ab
JJ
194342013-02-05 Jakub Jelinek <jakub@redhat.com>
19435
31502f9f
JJ
19436 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19437
a74db9bd
JJ
19438 PR sanitizer/55374
19439 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19440 (STATIC_LIBTSAN_LIBS): Likewise.
19441 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19442 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19443 is defined, don't add anything else beyond that.
19444 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19445 (LINK_COMMAND_SPEC): Use them.
19446
11f1e3ab
JJ
19447 PR tree-optimization/56205
19448 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19449 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19450 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19451
14ac6aa2
RB
194522013-02-05 Richard Biener <rguenther@suse.de>
19453
19454 PR tree-optimization/53342
19455 PR tree-optimization/53185
19456 * tree-vectorizer.h (vect_check_strided_load): Remove.
19457 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19458 not disallow peeling for vectorized strided loads.
19459 (vect_check_strided_load): Make static and simplify.
19460 (vect_analyze_data_refs): Adjust.
19461 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19462 correctly when vectorizing strided loads.
19463
6f22445a
RB
194642013-02-05 Richard Biener <rguenther@suse.de>
19465
19466 * doc/install.texi: Refer to ISL, not PPL.
19467
39f9719e
JH
194682013-02-05 Jan Hubicka <jh@suse.cz>
19469
ec4224ac
JH
19470 PR tree-optimization/55789
19471 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19472
194732013-02-05 Jan Hubicka <jh@suse.cz>
19474
19475 PR tree-optimization/55789
39f9719e
JH
19476 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19477 the dead call anyway.
19478
956d3b33
EB
194792013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19480
19481 PR sanitizer/55374
19482 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19483
194842013-02-04 Alexander Potapenko <glider@google.com>
19485 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP
19486 Jakub Jelinek <jakub@redhat.com>
19487
19488 PR sanitizer/55617
19489 * config/darwin.c (sort_ctor_records): Stabilized qsort
19490 on constructor priority by using original position.
19491 (finalize_ctors): New routine to sort constructors by
19492 priority before use in assemble_integer.
19493 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19494
7ac3af38
JJ
194952013-02-04 Jakub Jelinek <jakub@redhat.com>
19496
19497 PR libstdc++/54314
19498 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19499 about visibility on artificial decls.
19500 * config/sol2.c (solaris_assemble_visibility): Likewise.
19501
152689dc
KT
195022013-02-04 Kai Tietz <ktietz@redhat.com>
19503
19504 PR target/56186
19505 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19506 argument and improve checking of return-argument types for 16-byte
19507 modes.
19508 (ix86_function_value_1): Add additional valtype argument on call
19509 of function_value_64.
19510 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19511 handling infunction_value_64 function.
19512
2b5987b5
MGD
195132013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19514
7ac3af38 19515 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 19516
14597080
RB
195172013-02-04 Richard Biener <rguenther@suse.de>
19518
19519 PR tree-optimization/56188
19520 * tree-ssa-structalias.c (label_visit): Consider case with
19521 initially non-empty points-to set.
19522 (perform_var_substitution): Dump node mapping and clean up.
19523
ed73881e
RG
195242013-02-04 Richard Guenther <rguenther@suse.de>
19525
19526 PR lto/56168
19527 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19528 node prevail as last resort.
19529 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 19530 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 19531
139a0707
RB
195322013-02-04 Richard Biener <rguenther@suse.de>
19533
19534 PR tree-optimization/56113
19535 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19536 Merge into ...
19537 (equiv_class_lookup_or_add): ... this.
19538 (label_visit): Adjust and fix error in previous patch.
19539 (perform_var_substitution): Adjust.
19540
50fe8924
OE
195412013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19542
19543 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19544 * config/sh/sh4-300.md: Likewise.
19545 * config/sh/sh4a.md: Likewise.
19546 * config/sh/constraints.md: Likewise.
19547 * config/sh/sh.md: Likewise.
19548 * config/sh/netbsd-elf.h: Likewise.
19549 * config/sh/predicates.md: Likewise.
19550 * config/sh/sh-protos.h: Likewise.
19551 * config/sh/ushmedia.h: Likewise.
19552 * config/sh/linux.h: Likewise.
19553 * config/sh/sh.c: Likewise.
19554 * config/sh/superh.h: Likewise.
19555 * config/sh/elf.h: Likewise.
19556 * config/sh/sh4.md: Likewise.
19557 * config/sh/sh.h: Likewise.
19558
1a04ac2b
JDA
195592013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19560
19561 * config/pa/constraints.md: Adjust unused letters. Change "T"
19562 constraint to match_test floating_point_store_memory_operand().
19563 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19564 (base14_operand): New.
19565 (floating_point_store_memory_operand): New.
19566 (integer_store_memory_operand): Revise to use base14_operand and
19567 reg_plus_base_memory_operand.
19568 (move_dest_operand): Allow symbolic_memory_operands.
19569 (symbolic_memory_operand): Check for LO_SOM.
19570 (symbolic_operand): Change default case to break.
19571 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19572 CONST_DOUBLE values to be reloaded by putting them into memory when
19573 the destination is a floating point register.
19574 (movdf): Remove code to handle CONST_DOUBLE.
19575 (movsf): Likewise.
19576 (reload_indf_r1): New.
19577 (reload_insf_r1): New.
19578 Consistently use "Q" and "T" constraints with integer and floating
19579 point move instructions, respectively.
19580 (movdi): Remove FAIL.
19581 Change predicate for source operand unamed DImode move from
19582 general_operand to move_src_operand.
19583 (umulsidi3): Change predicate for destination operand to
19584 register_operand.
19585 Likewise for similar unamed patterns.
19586 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19587 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19588 (hppa_legitimize_address): Simplify mask calculation.
19589 (pa_emit_move_sequence): Revised handling of secondary reloads from
19590 REG+D addresses for floating point loads and stores. Directly handle
19591 loading CONST0_RTX (mode) to a floating point register.
19592 (pa_secondary_reload): Handle reloading DF and SFmode constant values
19593 to floating point registers. Don't restrict secondary reloads to
19594 floating point registers to integer modes. Revise some comments and
19595 cleanup some code.
19596 (TARGET_LEGITIMATE_ADDRESS_P): Define.
19597 (pa_legitimate_address_p): New.
19598 (pa_legitimize_reload_address): New.
19599 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
19600 (STRICT_REG_OK_FOR_BASE_P): New.
19601 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
19602 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
19603
3bf65591
DE
196042013-02-03 David Edelsohn <dje.gcc@gmail.com>
19605 Andrew Dixie <andrewd@gentrack.com>
19606
19607 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
19608 flag set.
19609
c0a8a3e6
RS
196102013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
19611
19612 * expmed.c (extract_bit_field_1): Pass the full width of the
19613 structure to get_best_reg_extraction_insn.
19614
99113dff
DE
196152013-02-01 David Edelsohn <dje.gcc@gmail.com>
19616
19617 PR target/54601
19618 * configure.ac (use_cxa_atexit): Add AIX.
19619 * configure: Regenerate.
19620
19621 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
19622
ddd84654
JJ
196232013-02-01 Jakub Jelinek <jakub@redhat.com>
19624
19625 PR debug/54793
19626 * final.c (need_profile_function): New variable.
19627 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
19628 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
19629 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
19630 notes, targetm.asm_out.function_prologue doesn't emit anything,
19631 HAVE_prologue and profiler should be emitted before prologue,
19632 set need_profile_function instead of emitting it.
19633 (final_scan_insn): If need_profile_function, emit
19634 profile_function on the first NOTE_INSN_BASIC_BLOCK or
19635 NOTE_INSN_FUNCTION_BEG note.
19636
196372013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH
19638
19639 * config/rs6000/rs6000.md (smulditi3): New.
19640 (umulditi3): New.
19641
ff2a9d88
RH
19642 * config/alpha/alpha.md (umulditi3): New.
19643
14d52b90
DE
196442013-02-01 David Edelsohn <dje.gcc@gmail.com>
19645
19646 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
19647 (ASM_OUTPUT_ALIGNED_LOCAL): New.
19648
8c7ca45c
RB
196492013-02-01 Richard Biener <rguenther@suse.de>
19650
19651 PR tree-optimization/56113
19652 * tree-ssa-structalias.c (label_visit): Reduce work for
19653 single-predecessor nodes.
19654
9f419393
EB
196552013-02-01 Eric Botcazou <ebotcazou@adacore.com>
19656
19657 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
19658 range isn't testing for zero.
19659
c5f4be84
SB
196602013-01-31 Steven Bosscher <steven@gcc.gnu.org>
19661
19662 PR middle-end/56113
19663 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
19664
dbdbd982
NC
196652013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
19666 Nick Clifton <nickc@redhat.com>
19667
19668 * config/v850/constraints.md (Q): Define as a memory constraint.
19669 * config/v850/predicates.md (label_ref_operand): New predicate.
19670 (e3v5_shift_operand): New predicate.
19671 (ior_operator): New predicate.
19672 * config/v850/t-v850: Add e3v5 multilib.
19673 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
19674 (v850_gen_movdi): Prototype.
19675 * config/v850/v850.c: Add support for e3v5 architecture.
19676 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
19677 TARGET_V850E_UP.
19678 (construct_save_jarl): Add e3v5 long JARL support.
19679 (v850_adjust_insn_length): New function. Adjust length of call
19680 insns when using e3v5 instructions.
19681 (v850_gen_movdi): New function: Generate instructions to move a
19682 DImode value.
19683 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
19684 (CPP_SPEC): Define __v850e3v5__ as appropriate.
19685 (TARGET_USE_FPU): Enable for e3v5.
19686 (CONST_OK_FOR_W): New macro.
19687 (ADJUST_INSN_LENGTH): Define.
19688 * config/v850/v850.md (UNSPEC_LOOP): Define.
19689 (attr cpu): Add v850e3v5.
19690 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
19691 (movdi): New pattern.
19692 (movdi_internal): New pattern.
19693 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
19694 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
19695 (cstoresf4): Likewise.
19696 (cstoredf4): Likewise.
19697 (insv): New pattern.
19698 (rotlso3_a): New pattern.
19699 (rotlsi3_b): New pattern
19700 (rotlsi3_v850e3v5): New pattern.
19701 (doloop_begin): New pattern.
19702 (fix_loop_counter): New pattern.
19703 (doloop_end): New pattern.
19704 (branch_normal): Add e3v5 long branch support.
19705 (branch_invert): Likewise.
19706 (branch_z_normal): Likewise.
19707 (branch_z_invert): Likewise.
19708 (branch_nz_normal): Likewise.
19709 (branch_nz_invert): Likewise.
19710 (call_internal_short): Add e3v5 register-indirect JARL support.
19711 (call_internal_long): Likewise.
19712 (call_value_internal_short): Likewise.
19713 (call_value_internal_long): Likewise.
19714 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
19715 (mloop): New option.
19716 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 19717 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 19718
73861a41
PK
197192013-01-31 Paul Koning <ni1d@arrl.net>
19720
19721 PR debug/55059
19722 PR debug/54508
19723 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
19724 children if parent is a class.
19725 (prune_unused_types_prune): Don't add DW_AT_declaration.
19726
e44978dc
RB
197272013-01-31 Richard Biener <rguenther@suse.de>
19728
19729 PR tree-optimization/56157
19730 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
19731 match up operand with SLP child.
19732
90b10dec
JM
197332013-01-31 Jason Merrill <jason@redhat.com>
19734
04d2dadd 19735 PR debug/54410
90b10dec
JM
19736 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
19737 parameters the first time.
19738 (gen_scheduled_generic_parms_dies): Check completeness here.
19739
6e616110
RB
197402013-01-31 Richard Biener <rguenther@suse.de>
19741
19742 PR middle-end/53073
19743 * common.opt (faggressive-loop-optimizations): New flag,
19744 enabled by default.
19745 * doc/invoke.texi (faggressive-loop-optimizations): Document.
19746 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
19747 infer_loop_bounds_from_undefined by it.
19748
636f59cf
RB
197492013-01-31 Richard Biener <rguenther@suse.de>
19750
19751 PR tree-optimization/56150
19752 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
19753 visit virtual operands.
19754 (find_uses_to_rename_bb): Likewise.
19755
197562013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB
19757
19758 PR tree-optimization/56150
19759 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
19760 mixed store non-store stmts.
19761
32887460
JJ
197622013-01-30 Jakub Jelinek <jakub@redhat.com>
19763
e60e09a0
JJ
19764 PR sanitizer/55374
19765 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
19766 LIBASAN_EARLY_SPEC is defined.
19767 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
19768 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
19769 before %o.
19770 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
19771
32887460
JJ
19772 PR c++/55742
19773 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
19774 invalid args instead of ICEing on it.
19775 (ix86_valid_target_attribute_tree): Return error_mark_node if
19776 ix86_valid_target_attribute_inner_p failed.
19777 (ix86_valid_target_attribute_p): Return false only if
19778 ix86_valid_target_attribute_tree returned error_mark_node. Allow
19779 target("default") attribute.
19780 (sorted_attr_string): Change argument from const char * to tree,
19781 merge in all target attribute arguments rather than just one.
19782 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
19783 instead of free. Avoid using strcat.
19784 (ix86_mangle_function_version_assembler_name): Mangle
19785 target("default") as if no target attribute is present. Adjust
19786 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
19787 instead of xmalloc and XDELETEVEC instead of free.
19788 (ix86_function_versions): Don't return true if one of the decls
19789 doesn't have target attribute. If they don't and one of the decls
19790 is DECL_FUNCTION_VERSIONED, report an error. Adjust
19791 sorted_attr_string caller. Use XDELETEVEC instead of free.
19792 (ix86_supports_function_versions): Remove.
19793 (make_name): Fix up formatting.
19794 (make_dispatcher_decl): Remove resolver_name and its initialization.
19795 Avoid leaking memory.
19796 (is_function_default_version): Return true if there is
19797 target("default") attribute rather than no target attribute at all.
19798 (make_resolver_func): Avoid leaking memory.
19799 (ix86_generate_version_dispatcher_body): Likewise.
19800 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
19801 * target.def (supports_function_versions): Remove.
19802 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
19803 * doc/tm.texi: Regenerated.
19804
73cca0cc
VM
198052013-01-30 Vladimir Makarov <vmakarov@redhat.com>
19806
19807 PR rtl-optimization/56144
19808 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
19809 for values with side effects.
19810
11452e7b
RB
198112013-01-30 Richard Biener <rguenther@suse.de>
19812
19813 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
19814 (sparseset_pop): Likewise.
19815 * cfganal.c (compute_idf): Likewise. Increase work-stack size
19816 to be able to use quick_push in the worker loop.
19817
01cb1ef5
MP
198182013-01-30 Marek Polacek <polacek@redhat.com>
19819
19820 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
19821
e892936e
RB
198222013-01-30 Richard Biener <rguenther@suse.de>
19823
19824 PR lto/56147
6edc3e32 19825 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 19826
cc06c01d
GJL
198272013-01-30 Georg-Johann Lay <avr@gjlay.de>
19828
19829 PR tree-optimization/56064
19830 * fixed-value.c (fixed_from_double_int): New function.
19831 * fixed-value.h (fixed_from_double_int): New prototype.
19832 (const_fixed_from_double_int): New static inline function.
19833 * fold-const.c (native_interpret_fixed): New static function.
19834 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
19835 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
19836 (native_encode_fixed): New static function.
19837 (native_encode_expr) <FIXED_CST>: Use it.
19838 (native_interpret_int): Move double_int worker code to...
19839 * double-int.c (double_int::from_buffer): ...this new static method.
19840 * double-int.h (double_int::from_buffer): Prototype it.
19841
d394a308
RB
198422013-01-30 Richard Biener <rguenther@suse.de>
19843
19844 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
19845 New pointer-map and obstack.
19846 (init_alias_vars): Allocate pointer-map and obstack.
19847 (delete_points_to_sets): Free them.
19848 (find_what_var_points_to): Cache result.
19849 (find_what_p_points_to): Adjust for changed interface of
19850 find_what_var_points_to.
19851 (compute_points_to_sets): Likewise.
19852 (ipa_pta_execute): Likewise.
19853
20804d96
RO
198542013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19855
19856 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
19857 * configure: Regenerate.
19858 * config.in: Regenerate.
19859 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
19860 #nobits/#progbits if supported.
19861
a7ad88a2
OE
198622013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
19863
19864 PR target/56121
19865 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
19866 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
19867 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
19868
7af79f92
GY
198692013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19870
91bfca59
OE
19871 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
19872 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 19873
753bcf7b
GY
198742013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19875
91bfca59
OE
19876 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
19877 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 19878
8cbc2ea8
GY
198792013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19880
91bfca59
OE
19881 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
19882 declaration.
19883 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
19884 * config/arm/cortex-a7.md: New bypasses using
19885 arm_mac_accumulator_is_result.
8cbc2ea8 19886
697a3325
GY
198872013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19888
91bfca59 19889 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 19890 (cortex_a7_neon_mla): Likewise.
91bfca59
OE
19891 (cortex_a7_fpfmad): New reservation.
19892 (cortex_a7_fpmacs): Use ffmas and update required units.
19893 (cortex_a7_fpmuld): Update required units and latency.
19894 (cortex_a7_fpmacd): Likewise.
19895 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
19896 (cortex_a7_neon). Likewise.
19897 (bypass) Update participating units.
697a3325 19898
29637783
GY
198992013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19900
91bfca59
OE
19901 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
19902 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
19903 from fmac to ffma.
19904 * config/arm/vfp11.md (vfp_farith): Use ffmas.
19905 (vfp_fmul): Use ffmad.
19906 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
19907 (cortex_r4_fmacd): Use ffmad.
19908 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
19909 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
19910 (cortex_a9_fmacd): Use ffmad.
19911 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
19912 (cortex_a8_vfp_macd): Use ffmad.
19913 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
19914 (cortex_a5_fpmacd): Use ffmad.
19915 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
19916 (cortex_a15_vfp_macd): Use ffmad.
19917 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 19918
2ee8a2d5
JM
199192013-01-29 Jason Merrill <jason@redhat.com>
19920
19921 PR libstdc++/54314
19922 * varasm.c (default_assemble_visibility): Don't warn about
19923 visibility on artificial decls.
19924
78d087bc
RB
199252013-01-29 Richard Biener <rguenther@suse.de>
19926
19927 PR tree-optimization/56113
19928 * tree-ssa-structalias.c (equiv_class_lookup): Also return
19929 the bitmap leader.
19930 (label_visit): Free duplicate bitmaps and record the leader instead.
19931 (perform_var_substitution): Adjust.
19932
83ba4d6f
RB
199332013-01-29 Richard Biener <rguenther@suse.de>
19934
19935 PR tree-optimization/55270
19936 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
19937 the CFG, schedule loops for fixup.
19938
66dfe13f
NC
199392013-01-29 Nick Clifton <nickc@redhat.com>
19940
19941 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
19942 SP_REG.
19943
5a579c3b
LE
199442013-01-28 Leif Ekblad <leif@rdos.net>
19945
19946 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
19947 * config/i386/i386.h (TARGET_RDOS): New macro.
19948 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
19949 * config/i386/i386.c (ix86_option_override_internal): For 64bit
19950 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
19951 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
19952 DEFAULT_LARGE_SECTION_THRESHOLD.
19953 * config/i386/i386.md (R14_REG, R15_REG): New constants.
19954 * config/i386/rdos.h: New file.
19955 * config/i386/rdos64.h: New file.
19956
51e44392
BS
199572013-01-28 Bernd Schmidt <bernds@codesourcery.com>
19958
19959 PR other/54814
19960 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
19961 TEST_HARD_REG_BIT.
19962
db1fb332
JJ
199632013-01-28 Jakub Jelinek <jakub@redhat.com>
19964
19965 PR rtl-optimization/56117
19966 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
19967 call cselib_lookup_from_insn on the MEM before calling
19968 add_insn_mem_dependence.
19969
16917761
RB
199702013-01-28 Richard Biener <rguenther@suse.de>
19971
19972 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
19973 to a stmt that didn't have one.
19974 (copy_phis_for_bb): Likewise for PHI arguments.
19975 (copy_debug_stmt): Likewise for debug stmts.
19976
b9fc0497
RB
199772013-01-28 Richard Biener <rguenther@suse.de>
19978
19979 PR tree-optimization/56034
6edc3e32 19980 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB
19981 (partition_builtin_p): Adjust.
19982 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
19983 it is the last partition.
19984 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
19985 up the vertex for the definition.
19986 (classify_partition): Classify whether a partition is a
19987 PKIND_REDUCTION, thus has uses outside of the loop.
19988 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
19989 Merge all PKIND_REDUCTION partitions into the last partition.
19990 (tree_loop_distribution): Seed partitions from reductions as well.
19991
aa710d25
JJ
199922013-01-28 Jakub Jelinek <jakub@redhat.com>
19993
0bfbca58
JJ
19994 PR tree-optimization/56125
19995 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
19996 pow(x,c) into sqrt(x) * powi(x, n/2) or
19997 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
19998 optimizing for size.
19999 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
20000 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
20001 integer.
20002
aa710d25
JJ
20003 PR tree-optimization/56094
20004 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
20005 to UNKNOWN_LOCATION while gimplifying expr.
20006
77dc5297
UB
200072013-01-27 Uros Bizjak <ubizjak@gmail.com>
20008
20009 PR target/56114
20010 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
20011 operand 0 in movabs insn template for -masm=intel asm alternative.
20012 (*movabs<mode>_2): Ditto for operand 1.
20013
0bfbca58 200142013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH
20015
20016 PR target/54663
20017 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
20018 of microblaze-c.o
20019
0bfbca58 200202013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI
20021
20022 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
20023 tm_file.
20024
0bfbca58 200252013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N
20026
20027 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
20028 Undef to avoid warning.
20029
478f60f9
MH
200302013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20031
20032 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
20033 * configure: Regenerate.
20034
d7fa6ee2
JJ
200352013-01-25 Jakub Jelinek <jakub@redhat.com>
20036
20037 PR tree-optimization/56098
20038 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
20039 for stmts with volatile ops.
20040 (cond_store_replacement): Don't optimize if assign has volatile ops.
20041 (cond_if_else_store_replacement_1): Don't optimize if either
20042 then_assign or else_assign have volatile ops.
20043 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
20044 volatile ops.
20045
f8fe87bd
GJL
200462013-01-25 Georg-Johann Lay <avr@gjlay.de>
20047
20048 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
20049
28398d0d
GJL
200502013-01-25 Georg-Johann Lay <avr@gjlay.de>
20051
20052 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
20053 missing ':' in asm example.
20054
b7d7d917
TB
200552013-01-25 Tejas Belagod <tejas.belagod@arm.com>
20056
20057 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
20058 entries into lane and laneq entries.
77dc5297
UB
20059 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
20060 Remove AdvSIMD scalar modes.
b7d7d917
TB
20061 (aarch64_sq<r>dmulh_laneq<mode>): New.
20062 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
20063 modes.
20064 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
20065 builtin implementations to relfect changes in RTL in aarch64-simd.md.
20066 * config/aarch64/iterators.md (VCOND): New.
20067 (VCONQ): New.
20068
556f9906
GJL
200692013-01-25 Georg-Johann Lay <avr@gjlay.de>
20070
20071 PR target/54222
20072 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
20073 Add NULL LIBNAME argument to existing definitions.
20074 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
20075 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
20076 * config/avr/avr.c (DEF_BUILTIN): Same.
20077 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
20078 (avr_expand_builtin): Expand to a vanilla call if a libgcc
20079 implementation is available (DECL_ASSEMBLER_NAME is set).
20080 (avr_fold_absfx): New static function.
20081 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
20082 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
20083 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
20084 AVR_BUILTIN_ABSLLK.
20085 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
20086 (abshk, absk, abslk, absllk): Provide as static inline functions.
20087
1f546bbb
MP
200882013-01-25 Marek Polacek <polacek@redhat.com>
20089
20090 PR tree-optimization/56035
20091 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
20092
7e184bd7
UB
200932012-01-24 Uros Bizjak <ubizjak@gmail.com>
20094
20095 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
20096 (*movtf_internal_rex64): Add (!o,C) alternative
20097 (*movxf_internal_rex64): Ditto.
20098 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
20099
f5ba49ac
SM
201002013-01-24 Shenghou Ma <minux.ma@gmail.com>
20101
20102 * doc/invoke.texi: fix typo.
20103 * doc/objc.texi: fix typo.
20104
a6343728
RS
201052013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
20106
20107 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
20108 for the first two alternatives.
20109
cd030c07
DN
201102013-01-24 Diego Novillo <dnovillo@google.com>
20111
77dc5297 20112 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN
20113 (ggc-zone.o): Remove.
20114 * configure.ac: Remove option --with-gc.
20115 * configure: Re-generate.
20116 * doc/install.texi: Remove documentation for --with-gc.
20117 * gengtype.c (write_enum_defn): Remove. Update all users.
20118 (write_Types_process_field): Remove generation of gt_e_* argument.
20119 (output_type_enum): Remove. Update all users.
20120 (write_enum_defn): Remove. Update all users.
20121 (enum alloc_zone): Remove. Update all users.
77dc5297 20122 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN
20123 * ggc-common.c (ggc_splay_alloc): Remove first argument.
20124 Update all callers.
20125 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 20126 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN
20127 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
20128 Update all users.
aaf1e810 20129 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN
20130 (struct alloc_zone): Remove.
20131 (ggc_internal_alloc_zone_stat): Remove.
20132 (ggc_internal_cleared_alloc_zone_stat): Remove.
20133 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB
20134 (ggc_pch_count_object): Remove last argument. Update all users.
20135 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN
20136 (struct alloc_zone): Remove.
20137 * ggc-zone.c: Remove.
77dc5297 20138 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN
20139 (struct alloc_zone): Remove.
20140 (ggc_alloc_typed_stat): Remove.
20141 (ggc_alloc_typed): Remove.
20142 (ggc_splay_alloc): Remove first argument.
20143 (rtl_zone): Remove. Update all users.
20144 (tree_zone): Remove. Update all users.
20145 (tree_id_zone): Remove. Update all users.
20146 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 20147 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 20148 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 20149 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 20150
a861ffa4
GJL
201512013-01-24 Georg-Johann Lay <avr@gjlay.de>
20152
20153 * config/avr/avr.c (avr_out_fract): Make register numbers that
20154 might be outside of source operand signed.
20155
593c0ddd
UB
201562013-01-24 Uros Bizjak <ubizjak@gmail.com>
20157
20158 * config/i386/constraints.md (Yf): New constraint.
20159 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
20160 of f constraint to conditionaly disable x87 register preferences.
20161 (*movdf_internal): Ditto.
20162 (*movsf_internal): Ditto.
20163
e86c0101
SB
201642013-01-24 Steven Bosscher <steven@gcc.gnu.org>
20165
20166 PR inline-asm/55934
20167 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
20168 that have operands with impossible constraints.
20169 Add a FIXME for a speed-up opportunity.
20170 * lra-constraints.c (process_alt_operands): Verify that a class
20171 selected from constraints on asms is valid for the operand mode.
20172 (curr_insn_transform): Remove incorrect comment.
20173
f6fee35f
DE
201742013-01-23 David Edelsohn <dje.gcc@gmail.com>
20175
20176 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
20177 TOC operand is a valid symbol ref in the constant pool.
20178
201792013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 20180
aaf1e810 20181 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 20182
dc62d7d1
GJL
201832013-01-23 Georg-Johann Lay <avr@gjlay.de>
20184
20185 PR target/54222
20186 * config/avr/stdfix.h: New file.
20187 * t-avr (stdfix-gcc.h): New rule to build it.
20188 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
20189
2660d12d
KS
201902013-01-23 Kostya Serebryany <kcc@google.com>
20191
77dc5297
UB
20192 * config/darwin.h: remove dependency on
20193 CoreFoundation (asan on Mac OS).
2660d12d 20194
a70418fc
JJ
201952013-01-23 Jakub Jelinek <jakub@redhat.com>
20196
20197 PR target/49069
20198 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
20199 instead of cmpdi_operand for first comparison operand.
20200 Don't assert that comparison operands aren't both constants.
20201
47876a2a
JW
202022013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20203
20204 * doc/install.texi (Downloading the Source): Update references to
20205 downloading separate components.
20206
202072013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW
20208
20209 * doc/extend.texi (__int128): Improve grammar.
20210
47876a2a 202112013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB
20212
20213 PR target/56028
20214 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
20215 alternative to (o,r).
20216 (*movdi_internal_rex64): Remove (!o,n) alternative.
20217 (DImode immediate->memory splitter): Remove.
20218 (DImode immediate->memory peephole2): Remove.
20219 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
20220 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
20221 alternative to (!o,*r).
20222 (*movtf_internal_sse): New pattern.
20223 (*movxf_internal_rex64): New pattern.
20224 (*movxf_internal): Disable for TARGET_64BIT.
20225 (*movdf_internal_rex64): Remove (!o,F) alternative.
20226
3a984f10
JJ
202272013-01-22 Jakub Jelinek <jakub@redhat.com>
20228
502498d5
JJ
20229 PR middle-end/56074
20230 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
20231 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
20232 * tree-vect-loop-manip.c (find_loop_location): Also ignore
20233 stmt locations where LOCATION_LOCUS of the stmt location is
20234 UNKNOWN_LOCATION or BUILTINS_LOCATION.
20235
3a984f10
JJ
20236 PR target/55686
20237 * config/i386/i386.md (UNSPEC_STOS): New.
20238 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20239 *strsetqi_1): Add UNSPEC_STOS.
20240
fa817f7f
PC
202412013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20242
20243 PR c++/56067
20244 * doc/invoke.texi: Remove left over -Wsynth example.
20245
8f498c1b
JJ
202462013-01-21 Jakub Jelinek <jakub@redhat.com>
20247
20248 PR tree-optimization/56051
20249 * fold-const.c (fold_binary_loc): Don't fold
20250 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20251 a narrowing conversion, or widening conversion from signed
20252 to unsigned.
20253
47876a2a 202542013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB
20255
20256 PR rtl-optimization/56023
20257 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20258 dependent on debug instruction.
20259
5b9db1bc
MJ
202602013-01-21 Martin Jambor <mjambor@suse.cz>
20261
20262 PR middle-end/56022
20263 * function.c (allocate_struct_function): Call
20264 invoke_set_current_function_hook earlier.
20265
e8bb7d68
JJ
202662013-01-21 Jakub Jelinek <jakub@redhat.com>
20267
20268 * reload1.c (init_reload): Only initialize reload_obstack
20269 during the first call.
20270
616a4e32
MP
202712013-01-21 Marek Polacek <polacek@redhat.com>
20272
20273 * cfgloop.c (verify_loop_structure): Fix up grammar.
20274
4401981b
YHH
202752013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20276
20277 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20278 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20279
8e87740b
RR
202802013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20281
20282 PR target/56058
20283 * config/arm/marvell-pj4.md: Update copyright year.
20284 Fix up use of alu to alu_reg and simple_alu_imm.
20285
47876a2a 202862013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB
20287
20288 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20289
89d56d79
VM
202902013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20291
20292 PR target/55433
20293 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 20294 insn for secondary memory move when memory mode should be different.
89d56d79 20295
fe603553
JDA
202962013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20297
20298 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20299 atomic_storedi_1): New patterns.
20300
01284895
VK
203012013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20302
20303 btver2 pipeline descriptions.
20304 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20305 descriptions.
20306 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 20307 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK
20308 type attributes.
20309 * config/i386/btver2.md: New file describing btver2 pipelines.
20310
5630e3e1
JL
203112013-01-19 Andrew Pinski <apinski@cavium.com>
20312
20313 PR tree-optimization/52631
20314 * tree-ssa-sccvn (visit_use): Before looking up the original
20315 statement, try looking up the simplified expression.
20316
650ae806
AG
203172013-01-19 Anthony Green <green@moxielogic.com>
20318
20319 * config/moxie/moxie.c (moxie_expand_prologue): Set
20320 current_function_static_stack_size.
20321
e300ec2d
JJ
203222013-01-18 Jakub Jelinek <jakub@redhat.com>
20323
20324 PR tree-optimization/56029
20325 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20326 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20327
a3d7af04
SS
203282013-01-18 Sharad Singhai <singhai@google.com>
20329
20330 PR tree-optimization/55995
20331 * dumpfile.c (dump_loc): Print location only if available.
20332 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
20333
66aa7879
VM
203342013-01-18 Vladimir Makarov <vmakarov@redhat.com>
20335
20336 PR target/55433
20337 * lra-constraints.c (curr_insn_transform): Reuse original insn for
20338 secondary memory move.
20339 (inherit_reload_reg): Use rclass instead of cl for
20340 check_secondary_memory_needed_p.
20341
3f0fee7b
JJ
203422013-01-18 Jakub Jelinek <jakub@redhat.com>
20343
20344 PR middle-end/56015
20345 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 20346 the case where writing real complex part of target modifies op1.
3f0fee7b 20347
70c67693
JG
203482013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
20349
20350 * config/aarch64/aarch64-simd.md
20351 (aarch64_vcond_internal<mode>): Handle unordered cases.
20352 * config/aarch64/iterators.md (v_cmp_result): New.
20353
df8de9b3
YHH
203542013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
20355 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20356
20357 * config/arm/marvell-pj4.md: New file.
20358 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
20359 * config/arm/arm.md (generic_sched): Add marvell_pj4.
20360 (generic_vfp): Likewise.
20361 * config/arm/arm-cores.def: Add marvell-pj4.
20362 * config/arm/arm-tune.md: Regenerate.
20363 * config/arm/arm-tables.opt: Regenerate.
20364 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
20365 * doc/invoke.texi: Document marvell-pj4.
20366
be30c356
TB
203672013-01-18 Tejas Belagod <tejas.belagod@arm.com>
20368
20369 * config/aarch64/arm_neon.h: Map scalar types to standard types.
20370
0bfbca58 203712013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO
20372
20373 PR debug/54114
20374 PR debug/54402
20375 PR debug/49888
20376 * var-tracking.c (negative_power_of_two_p): New.
20377 (global_get_addr_cache, local_get_addr_cache): New.
20378 (get_addr_from_global_cache, get_addr_from_local_cache): New.
20379 (vt_canonicalize_addr): Rewrite using the above. Adjust the
20380 heading comment.
20381 (vt_stack_offset_p): Remove.
20382 (vt_canon_true_dep): Always canonicalize loc's address.
20383 (clobber_overlapping_mems): Make sure we have a MEM.
20384 (local_get_addr_clear_given_value): New.
20385 (val_reset): Clear local cached entries.
20386 (compute_bb_dataflow): Create and release the local cache.
20387 Disable duplicate MEMs clobbering.
20388 (emit_notes_in_bb): Clobber MEMs likewise.
20389 (vt_emit_notes): Create and release the local cache.
20390 (vt_initialize, vt_finalize): Create and release the global
20391 cache, respectively.
1f6bc337 20392 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 20393
0bfbca58 203942013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO
20395
20396 PR libmudflap/53359
20397 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20398 not found in the symtab.
20399
0bfbca58 204002013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 20401
c350ba53 20402 PR debug/56006
3aa03517
AO
20403 PR rtl-optimization/55547
20404 PR rtl-optimization/53827
20405 PR debug/53671
20406 PR debug/49888
20407 * alias.c (offset_overlap_p): New, factored out of...
20408 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20409 the conservative special case for symbolic constants. Don't
20410 adjust zero sizes on alignment.
20411
c664546f
JL
204122013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20413
20414 PR rtl-optimization/52573
20415 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20416 REG_UNUSED for the same register.
20417
1bd3f750
MP
204182013-01-17 Richard Biener <rguenther@suse.de>
20419 Marek Polacek <polacek@redhat.com>
20420
20421 PR rtl-optimization/55833
20422 * loop-unswitch.c (unswitch_loops): Move loop verification...
20423 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20424 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20425 Set it to true when we're removing a loop from hierarchy tree in
20426 an irreducible region.
20427 (fix_bb_placements): Adjust caller.
20428 (fix_loop_placements): Likewise.
20429
e52a8b71
GJL
204302013-01-17 Georg-Johann Lay <avr@gjlay.de>
20431
20432 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20433 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20434 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20435 Remove ID. Adjust comments.
20436 * config/avr/avr-c.c (avr_builtin_name): Remove.
20437 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20438 * config/avr/avr.c (avr_tolower): New static function.
20439 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20440 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20441 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20442 default expansion.
20443
8386a7ea
JH
204442013-01-17 Jan Hubicka <jh@suse.cz>
20445
610fb637 20446 PR tree-optimization/55273
8386a7ea
JH
20447 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20448
47876a2a 204492013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB
20450
20451 PR target/55981
20452 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20453 store through atomic_store<mode>_1.
20454 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20455
8222c37e
MJ
204562013-01-17 Martin Jambor <mjambor@suse.cz>
20457
20458 PR tree-optimizations/55264
20459 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20460 for virtual methods.
20461 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20462 virtual methods before inlining is over.
20463 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20464 virtual functions.
20465 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20466 non-virtual.
20467
79f01c76
VM
204682013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20469
20470 PR rtl-optimization/56005
20471 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20472 pending reads for prefetch.
20473
d0b6bb1b
IB
204742013-01-16 Ian Bolton <ian.bolton@arm.com>
20475
aaf1e810 20476 * config/aarch64/aarch64.md
d0b6bb1b
IB
20477 (*cstoresi_neg_uxtw): New pattern.
20478 (*cmovsi_insn_uxtw): New pattern.
20479 (*<optab>si3_uxtw): New pattern.
20480 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20481 (*<optab>si3_insn_uxtw): New pattern.
20482 (*bswapsi2_uxtw): New pattern.
20483
cb9cf03b
RB
204842013-01-16 Richard Biener <rguenther@suse.de>
20485
20486 * tree-inline.c (tree_function_versioning): Remove set but
20487 never used variable.
20488
2cfc56b9
RB
204892013-01-16 Richard Biener <rguenther@suse.de>
20490
20491 PR tree-optimization/55964
20492 * tree-flow.h (rename_variables_in_loop): Remove.
20493 (rename_variables_in_bb): Likewise.
20494 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20495 (copy_loop_before): Adjust and delete update-ssa status.
20496 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 20497 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB
20498 (rename_variables_in_loop): Remove.
20499 (slpeel_update_phis_for_duplicate_loop): Likewise.
20500 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20501 use available cfg machinery instead of duplicating it.
20502 Update PHI nodes and perform poor-mans SSA update here.
20503 (slpeel_tree_peel_loop_to_edge): Adjust.
20504
c25a0c60
RB
205052013-01-16 Richard Biener <rguenther@suse.de>
20506
20507 PR tree-optimization/54767
20508 PR tree-optimization/53465
20509 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20510 (vrp_visit_phi_node): For PHI arguments coming via backedges
20511 drop all symbolical range information.
20512 (execute_vrp): Compute backedges.
20513
04b535af
RB
205142013-01-16 Richard Biener <rguenther@suse.de>
20515
20516 * doc/install.texi: Update CLooG and ISL requirements to
20517 0.18.0 and 0.11.1.
20518
8b0a1e0b
CB
205192013-01-16 Christian Bruel <christian.bruel@st.com>
20520
20521 PR target/55301
20522 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20523 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20524 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20525
205262013-01-16 DJ Delorie <dj@redhat.com>
20527
20528 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20529 (UNSPECV_SP_SWITCH_E): New.
20530 (sp_switch_1): Change to an unspec.
20531 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20532 replace $r15.
20533
205342013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB
20535
20536 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20537 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20538 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20539 (expand_mem_thread_fence): Ditto.
20540 (expand_mem_signal_fence): Ditto.
20541 (expand_atomic_load): Ditto.
20542 (expand_atomic_store): Ditto.
20543
0bfbca58 205442013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO
20545
20546 PR rtl-optimization/55547
20547 PR rtl-optimization/53827
20548 PR debug/53671
20549 PR debug/49888
20550 * alias.c (memrefs_conflict_p): Set sizes to negative after
20551 AND adjustments.
20552
305e3ac1
JJ
205532013-01-15 Jakub Jelinek <jakub@redhat.com>
20554
20555 PR target/55940
20556 * function.c (thread_prologue_and_epilogue_insns): Always
20557 add crtl->drap_reg to set_up_by_prologue.set, even if
20558 stack_realign_drap is false.
20559
f78ac4f2
JBG
205602013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20561
20562 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20563 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20564 *call): Fix indention.
20565
a78a8cc4
TV
205662013-01-15 Tom de Vries <tom@codesourcery.com>
20567
20568 PR target/55876
20569 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20570 Update comment.
20571
0e80383f
VM
205722013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20573
305e3ac1 20574 PR rtl-optimization/55153
0e80383f
VM
20575 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20576
207b5956
MJ
205772013-01-15 Martin Jambor <mjambor@suse.cz>
20578
20579 PR tree-optimization/55920
20580 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20581 accesses as grp_to_be_debug_replaced.
20582
a7818b54
JJ
205832013-01-15 Jakub Jelinek <jakub@redhat.com>
20584
20585 PR tree-optimization/55920
20586 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20587 there is non-useless type conversion needed from debug rhs to lhs,
20588 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20589
b0fe107e
JM
205902013-01-15 Joseph Myers <joseph@codesourcery.com>
20591 Mikael Pettersson <mikpe@it.uu.se>
20592
20593 PR target/43961
20594 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
20595 Thumb.
20596 (ASM_OUTPUT_CASE_LABEL): Remove.
20597 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
20598 * final.c (shorten_branches): Update alignment of labels before
20599 jump tables if CASE_VECTOR_SHORTEN_MODE.
20600
34ab62ee
RB
206012013-01-15 Richard Biener <rguenther@suse.de>
20602
20603 PR bootstrap/55961
20604 * system.h: Do not include gmp.h for building host tools.
20605
783a3a05
RB
206062013-01-15 Richard Biener <rguenther@suse.de>
20607
20608 PR middle-end/55882
20609 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
20610 account for bitpos when computing alignment.
20611
3a579e09
VY
206122013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
20613
20614 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
20615 (ix86_target_macros_internal): Likewise.
20616
20617 * config/i386/i386.c (m_CORE2I7): Removed.
20618 (m_CORE_HASWELL): New macro.
20619 (m_CORE_ALL): Likewise.
20620 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
20621 (initial_ix86_arch_features): Likewise.
20622 (processor_target_table): Initializations for Core avx2.
20623 (cpu_names): New names "core-avx2".
20624 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
20625 PROCESSOR_CORE_HASWELL.
20626 (ix86_issue_rate): New case.
20627 (ia32_multipass_dfa_lookahead): Likewise.
20628 (ix86_sched_init_global): Likewise.
20629
20630 * config/i386/i386.h (TARGET_HASWELL): New macro.
20631 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
20632 (processor_type): New PROCESSOR_HASWELL.
20633
955f5a07
JJ
206342013-01-15 Jakub Jelinek <jakub@redhat.com>
20635
ff784829
JJ
20636 PR tree-optimization/55955
20637 * tree-vect-loop.c (vectorizable_reduction): Give up early on
20638 *SHIFT_EXPR and *ROTATE_EXPR codes.
20639
955f5a07
JJ
20640 PR tree-optimization/48766
20641 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
20642 -ftrapv disable -fwrapv.
20643
aeb8b4e9
GJL
206442013-01-14 Georg-Johann Lay <avr@gjlay.de>
20645
20646 PR target/55974
20647 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
20648 etc. to 1 and not to __flash.
20649 Use LL suffix for __INT24_MAX__ with -mint8.
20650 Use ULL suffix for __UINT24_MAX__ with -mint8.
20651
1c494c6a
GJL
206522013-01-14 Georg-Johann Lay <avr@gjlay.de>
20653
20654 * config/avr/avr-arch.h
20655 (struct base_arch_s): Use typedef avr_arch_t instead.
20656 (struct arch_info_s): Use typedef avr_arch_info_t instead.
20657 (struct mcu_type_s): Use typedef avr_mcu_t instead.
20658 * config/avr/avr.c: Same.
20659 * config/avr/avr-devices.c: Same.
20660 * config/avr/driver-avr.c: Same.
20661 * config/avr/gen-avr-mmcu-texi.c: Same.
20662 * config/avr/avr-mcus.def: Adjust comment.
20663
a50344cb
TB
206642013-01-14 Tejas Belagod <tejas.belagod@arm.com>
20665
88e784e6
UB
20666 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
20667 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 20668
47876a2a 206692013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB
20670 Andi Kleen <ak@linux.intel.com>
20671
20672 PR target/55948
20673 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
20674 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
20675 memmodel flag.
20676
00892272
GJL
206772013-01-14 Georg-Johann Lay <avr@gjlay.de>
20678
20679 * config/avr/avr-stdint.h: Remove trailing blanks.
20680 * config/avr/avr-log.h: Same.
20681 * config/avr/avr-arch.h: Same.
20682 * config/avr/avr-devices.c: Same.
20683 * config/avr/avr-dimode.md: Same.
20684 * config/avr/predicates.md: Same.
20685 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 20686
00892272
GJL
20687 * config/avr/avr-protos.h: Same. And:
20688 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
20689 (init_cumulative_args): Rename to avr_init_cumulative_args.
20690 (expand_prologue): Rename to avr_expand_prologue.
20691 (expand_epilogue): Rename to avr_expand_epilogue.
20692 (adjust_insn_length): Rename to avr_adjust_insn_length.
20693 (notice_update_cc): Rename to avr_notice_update_cc.
20694 (final_prescan_insn): Rename to avr_final_prescan_insn.
20695 * config/avr/avr.c: Same.
20696 * config/avr/avr.h: Same.
20697 * config/avr/avr.md: Remove trailing blanks.
20698 (prologue): Use avr_expand_prologue.
20699 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
20700
7bb01996
RB
207012013-01-14 Richard Biener <rguenther@suse.de>
20702
20703 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
20704 verify_location, collect_subblocks): New functions.
20705 (verify_gimple_in_cfg): Verify that locations only reference
20706 BLOCKs in the functions BLOCK tree.
20707
2724573f
RB
207082013-01-14 Richard Biener <rguenther@suse.de>
20709
20710 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
20711 PHI argument.
20712 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
20713 unshare reference.
20714 (insert_out_of_ssa_copy_on_edge): Likewise.
20715 (rewrite_close_phi_out_of_ssa): Likewise.
20716 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
20717 debug expressions.
20718 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
20719 propagated constants.
20720 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
20721 can not be shared.
20722
9a0bbab6
GJL
207232013-01-14 Georg-Johann Lay <avr@gjlay.de>
20724
20725 * config/avr/avr-modes.def: Add GPL copyright notice.
20726
45805f17
UB
207272013-01-13 Uros Bizjak <ubizjak@gmail.com>
20728
20729 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
20730 MEMMODEL_MASK to determine memory model.
20731 (atomic_store<mode>): Ditto from operands[2].
20732 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
20733
9d60be38
JJ
207342013-01-13 Jakub Jelinek <jakub@redhat.com>
20735
20736 PR fortran/55935
45805f17 20737 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ
20738 (fold_gimple_assign): Don't call unshare_expr here.
20739 (fold_ctor_reference): Call unshare_expr.
20740
e7f49d92
TG
207412013-01-13 Terry Guo <terry.guo@arm.com>
20742
aaf1e810
EB
20743 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
20744 * doc/fragments.texi: Document MULTILIB_REUSE.
20745 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 20746 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 20747 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG
20748 (tmpmultilib4): Ditto.
20749 (multilib_reuse): New multilib argument.
20750
fbd03a27
RS
207512013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
20752
20753 * Makefile.in: Update copyright.
20754
b3681f13
TV
207552013-01-12 Tom de Vries <tom@codesourcery.com>
20756
20757 PR middle-end/55890
20758 * calls.c (expand_call): Check if arg_nr is valid.
20759
3f287e4b
MM
207602013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
20761
20762 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
20763 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM
20764 documentation. Add missing '__' in front of
20765 __builtin_ia32_packssdw256.
3f287e4b 20766
1abcd5eb
AK
207672013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20768
20769 PR target/55719
20770 * config/s390/s390.c (s390_preferred_reload_class): Do not return
20771 NO_REGS for larl operands.
20772 (s390_reload_larl_operand): Use s390_load_address instead of
20773 emit_move_insn.
20774
980d0812
RB
207752013-01-11 Richard Biener <rguenther@suse.de>
20776
20777 * tree-cfg.c (verify_node_sharing_1): Split out from ...
20778 (verify_node_sharing): ... here.
20779 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
20780
7cb9fd07
EB
207812013-01-11 Eric Botcazou <ebotcazou@adacore.com>
20782
20783 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
20784 Substitute TREECHECKING.
20785 * configure: Regenerate.
20786 * Makefile.in (TREECHECKING): New.
20787
47876a2a 207882013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG
20789
20790 PR tree-optimization/44061
20791 * tree-vrp.c (extract_range_basic): Compute zero as
20792 value-range for __builtin_constant_p of function parameters.
20793
d1e082c2
RS
207942013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
20795
45805f17 20796 Update copyright years.
d1e082c2 20797
f85021b0
VM
207982013-01-10 Vladimir Makarov <vmakarov@redhat.com>
20799
9d60be38 20800 PR rtl-optimization/55672
0160442c
VM
20801 * lra-eliminations.c (mark_not_eliminable): Permit addition with
20802 const to be eliminable.
f85021b0 20803
7a8b1ec4
DE
208042013-01-10 David Edelsohn <dje.gcc@gmail.com>
20805
20806 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
20807 * configure: Regenerate.
20808
ad2e5b71
RB
208092013-01-10 Richard Biener <rguenther@suse.de>
20810
ddf9322d 20811 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 20812
ffc5b2bb
RB
208132013-01-10 Richard Biener <rguenther@suse.de>
20814
20815 PR bootstrap/55792
20816 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
20817 locations for virtual PHI arguments.
20818 (rewrite_update_phi_arguments): Likewise.
20819
e1f674e4
JS
208202013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
20821
20822 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
20823 on to assembler.
20824
a57fc743
JJ
208252013-01-10 Jakub Jelinek <jakub@redhat.com>
20826
20827 PR tree-optimization/55921
20828 * tree-complex.c (expand_complex_asm): New function.
20829 (expand_complex_operations_1): Call it for GIMPLE_ASM.
20830
0ff4390d
AK
208312013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20832
20833 PR target/55718
20834 * config/s390/s390.c (s390_symref_operand_p)
20835 (s390_loadrelative_operand_p): Merge the two functions.
20836 (s390_check_qrst_address, print_operand_address): Add parameters
20837 to s390_loadrelative_operand_p invokation.
20838 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
20839 (s390_reload_larl_operand, s390_secondary_reload): Use
20840 s390_loadrelative_operand_p instead of s390_symref_operand_p.
20841 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
20842
6f557e0e
MS
208432013-01-09 Mike Stump <mikestump@comcast.net>
20844
20845 * dse.c (record_store): Remove unnecessary assert.
20846
7770c9e9
JH
208472013-01-09 Jan Hubicka <jh@suse.cz>
20848
20849 PR tree-optimization/55569
20850 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
20851 * cfgloop.h (scale_loop_profile): Likewise.
20852
a19b1432
JH
208532013-01-09 Jan Hubicka <jh@suse.cz>
20854
20855 PR lto/45375
ddf9322d
UB
20856 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
20857 functions.
a19b1432
JH
20858 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
20859
9a002da8
RS
208602013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
20861
20862 PR middle-end/55114
20863 * expr.h (maybe_emit_group_store): Declare.
20864 * expr.c (maybe_emit_group_store): New function.
20865 * builtins.c (expand_builtin_int_roundingfn): Call it.
20866 (expand_builtin_int_roundingfn_2): Likewise.
20867
511dcace
VM
208682013-01-09 Vladimir Makarov <vmakarov@redhat.com>
20869
e1f2b729 20870 PR rtl-optimization/55829
511dcace
VM
20871 * lra-constraints.c (match_reload): Add code for absent output.
20872 (curr_insn_transform): Add code for reloads of matched inputs
20873 without output.
20874
7b0fe4f4
UB
208752013-01-09 Uros Bizjak <ubizjak@gmail.com>
20876
20877 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
20878 attribute of movddup insn to DF.
20879 (*vec_interleave_lowv2df): Ditto.
20880 (vec_dupv2df): Ditto.
20881
870ca331
JH
208822013-01-09 Jan Hubicka <jh@suse.cz>
20883
20884 PR tree-optimiation/55875
20885 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
20886 EVERY_ITERATION parameter.
7b0fe4f4 20887 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH
20888 (idx_infer_loop_bounds): Similarly here.
20889 (n_of_executions_at_most): Simplify
20890 to only test for cases where statement is dominated by the
7b0fe4f4 20891 particular bound; handle correctly the "postdominance" test.
870ca331
JH
20892 (scev_probably_wraps_p): Use max loop iterations info
20893 as a global bound first.
20894
6f575fe4 208952013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD
20896 Nick Clifton <nickc@redhat.com>
20897
20898 * config/v850/v850.md (cbranchsf4): New pattern.
20899 (cstoresf4): New pattern.
20900 (cbranchdf4): New pattern.
20901 (cstoredf4): New pattern.
20902 (movsicc): Disallow floating point comparisons.
20903 (cmpsf_le_insn): Fix order of operators.
20904 (cmpsf_lt_insn): Likewise.
20905 (cmpsf_eq_insn): Likewise.
20906 (cmpdf_le_insn): Likewise.
20907 (cmpdf_lt_insn): Likewise.
20908 (cmpdf_eq_insn): Likewise.
20909 (cmpsf_ge_insn): Use LE comparison.
20910 (cmpdf_ge_insn): Likewise.
20911 (cmpsf_gt_insn): Use LT comparison.
20912 (cmpdf_gt_insn): Likewise.
20913 (cmpsf_ne_insn): Delete pattern.
20914 (cmpdf_ne_insn): Delete pattern.
20915 * config/v850/v850.c (v850_gen_float_compare): Use
20916 gen_cmpdf_eq_insn for NE comparison.
20917 (v850_float_z_comparison_operator)
20918 (v850_float_nz_comparison_operator): Move from here ...
20919 * config/v850/predicates.md: ... to here. Move GT and GE
20920 comparisons into v850_float_z_comparison_operator.
20921 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
20922 Delete prototype.
20923 (v850_float_nz_comparison_operator): Likewise.
20924
f0d54148
JDA
209252013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20926
20927 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
20928 with calls to gen_insvsi/gen_insvdi.
20929
8f01beca
VK
209302013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20931
20932 * config/i386/i386.c (initial_ix86_tune_features): Set up
20933 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
20934
2aa579ad
JJ
209352013-01-09 Steven Bosscher <steven@gcc.gnu.org>
20936 Jakub Jelinek <jakub@redhat.com>
20937
20938 PR tree-optimization/48189
20939 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
20940 If nitercst is 0, don't predict the exit edge.
20941
6edc3e32 209422013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N
20943
20944 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
20945 in asm_fprintf with reg_names.
20946 (aarch64_print_operand_address): Likewise.
20947 (aarch64_return_addr): Likewise.
20948 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
20949
f6f94d94
JDA
209502013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20951
20952 * config/pa/pa.h (VAL_U6_BITS_P): Define.
20953 (INT_U6_BITS): Likewise.
20954 * config/pa/predicates.md (uint6_operand): New predicate.
20955 (shift5_operand, shift6_operand): Likewise.
20956 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
20957 arith32_operand.
20958 (lshrdi3): Use shift6_operand.
20959 (shrpsi4, shrpdi4): New insn patterns.
20960 (extzv): Delete expander.
20961 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
20962 predicates in unamed zero extract patterns. Tighten common constraint.
20963 (extv): Delete expander.
20964 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
20965 predicates in unamed sign extract patterns. Tighten common constraint.
20966 (insv): Delete expander.
20967 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
20968 predicates in unamed insert patterns. Tighten common constraint.
20969 Change uint32_operand predicate to uint6_operand predicate in unamed
20970 DImode pattern to insert constant values of type 1...1xxxx.
20971
36b72910
JH
209722013-01-04 Jan Hubicka <jh@suse.cz>
20973
20974 PR tree-optimization/55823
7b0fe4f4
UB
20975 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
20976 issue.
36b72910 20977
47876a2a 209782013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ
20979 Uros Bizjak <ubizjak@gmail.com>
20980
20981 PR rtl-optimization/55845
20982 * df-problems.c (can_move_insns_across): Stop scanning at
20983 volatile_insn_p source instruction or give up if
20984 across_from .. across_to range contains any volatile_insn_p
20985 instructions.
20986
4369c11e
TB
209872013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20988
7b0fe4f4
UB
20989 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
20990 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
20991 Declare.
4369c11e 20992 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 20993 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 20994
aeb7e7c1
JJ
209952013-01-08 Jakub Jelinek <jakub@redhat.com>
20996
20997 PR fortran/55341
20998 * asan.c (asan_clear_shadow): New function.
20999 (asan_emit_stack_protection): Use it.
21000
a02ad1aa
TB
210012013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21002
21003 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
21004 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
21005 with tab instead of space.
21006
f90d87f5
NC
210072013-01-08 Nick Clifton <nickc@redhat.com>
21008
21009 * config/rl78/rl78.c (rl78_expand_prologue): Always select
21010 register bank 0 at the start of an interrupt handler.
83ffd964
NC
21011 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
21012 MDBH registers.
f90d87f5 21013
385eb93d
JG
210142013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21015
21016 * config/aarch64/aarch64-simd.md
21017 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
21018 (aarch64_simd_bsl): Likewise.
21019 (aarch64_vcond_internal<mode>): Likewise.
21020 (vcond<mode><mode>): Likewise.
21021 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
21022 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
21023
4dcd1054
JG
210242013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21025
21026 * config/aarch64/aarch64-builtins.c
21027 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
21028
4267a4a6
MJ
210292013-01-08 Martin Jambor <mjambor@suse.cz>
21030
21031 PR debug/55579
21032 * tree-sra.c (analyze_access_subtree): Return true also after
21033 potentially creating a debug-only replacement.
21034
5f4e6de3
JJ
210352013-01-08 Jakub Jelinek <jakub@redhat.com>
21036
3138f224
JJ
21037 PR middle-end/55890
21038 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
21039
5f4e6de3
JJ
21040 PR tree-optimization/54120
21041 * tree-vrp.c (range_fits_type_p): Don't allow
21042 src_precision < precision from signed vr to unsigned_p
21043 if vr->min or vr->max is negative.
21044 (simplify_float_conversion_using_ranges): Test can_float_p
21045 against CODE_FOR_nothing.
21046
f3ef18ff
JJ
210472013-01-08 Jakub Jelinek <jakub@redhat.com>
21048 Richard Biener <rguenther@suse.de>
21049
21050 PR middle-end/55851
21051 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
21052 types instead of just INTEGER_TYPE types.
21053
25c210f9
MK
210542013-01-07 Mark Kettenis <kettenis@openbsd.org>
21055
21056 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
21057 TF_SIZE): Define.
f90d87f5 21058
81826a7b
SE
210592013-01-07 Steve Ellcey <sellcey@mips.com>
21060
21061 PR target/42661
21062 * config/mips/mips.opt: Change mad to mmad to match documentation.
21063
35678514
GJL
210642013-01-07 Georg-Johann Lay <avr@gjlay.de>
21065
21066 PR target/55897
21067 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
21068 .progmemx.data now.
21069
e5669488
GJL
210702013-01-07 Georg-Johann Lay <avr@gjlay.de>
21071
21072 PR target/55897
21073 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
21074 (avr_addrspace_t): Add .section_name field.
21075 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
21076 array size.
21077 (avr_addrspace): Same. Initialize .section_name. Remove last
21078 NULL entry. Put __memx into .progmemx.data.
21079 (progmem_section_prefix): Remove.
21080 (avr_asm_init_sections): No need to initialize progmem_section.
21081 (avr_asm_named_section): Use avr_addrspace[].section_name to get
21082 section name prefix.
21083 (avr_asm_select_section): Ditto. And use get_unnamed_section to
21084 retrieve the progmem section.
21085 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
21086 boundary to run over avr_addrspace[].
21087 (avr_register_target_pragmas): Ditto.
21088
656e6f37
JJ
210892013-01-06 Jakub Jelinek <jakub@redhat.com>
21090
21091 * varasm.c (output_constant_def_contents): For asan_protect_global
21092 protected strings, adjust DECL_ALIGN if needed, before testing for
21093 anchored symbols.
21094 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
21095 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
21096 normal decls.
21097 (output_object_block): For asan protected decls, emit asan padding
21098 after their contents.
21099 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
21100 (asan_finish_file): Test it here instead.
21101
6edc3e32
UB
211022013-01-07 Nick Clifton <nickc@redhat.com>
21103 Matthias Klose <doko@debian.org>
21104 Doug Kwan <dougkwan@google.com>
21105 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC
21106
21107 PR driver/55470
21108 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
21109
21110 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
21111
21112 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
21113
7b0fe4f4 21114 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC
21115
21116 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
21117
2a095093
GJL
211182013-01-07 Georg-Johann Lay <avr@gjlay.de>
21119
f3b503f4 21120 PR target/54461
2a095093
GJL
21121 * doc/install.texi (Cross-Compiler-Specific Options): Document
21122 --with-avrlibc.
21123
383f9b34
TB
211242013-01-07 Tejas Belagod <tejas.belagod@arm.com>
21125
21126 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
21127 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
21128 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
21129 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
21130 vqmovun_high_s64): Fix source operand number and update copyright.
21131
3626621a
RB
211322013-01-07 Richard Biener <rguenther@suse.de>
21133
21134 PR middle-end/55890
21135 * gimple.h (gimple_call_builtin_p): New overload.
21136 * gimple.c (validate_call): New function.
21137 (gimple_call_builtin_p): Likewise.
21138 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21139 Use gimple_call_builtin_p.
21140 (find_func_clobbers): Likewise.
21141 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
21142 (strlen_optimize_stmt): Likewise.
21143
8b2ea410
JG
211442013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21145
21146 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
21147 (vld1q_dup_*): Likewise.
21148 (vld1_*): Likewise.
21149 (vld1q_*): Likewise.
21150 (vld1_lane_*): Likewise.
21151 (vld1q_lane_*): Likewise.
21152
9713d329
RB
211532013-01-07 Richard Biener <rguenther@suse.de>
21154
21155 * lto-streamer.h (LTO_minor_version): Bump to 2.
21156
3520f7cc
JG
211572013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21158
21159 * config/aarch64/aarch64-protos.h
21160 (aarch64_const_double_zero_rtx_p): Rename to...
21161 (aarch64_float_const_zero_rtx_p): ...this.
21162 (aarch64_float_const_representable_p): New.
21163 (aarch64_output_simd_mov_immediate): Likewise.
21164 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
21165 move immediate case.
21166 * config/aarch64/aarch64.c
21167 (aarch64_const_double_zero_rtx_p): Rename to...
21168 (aarch64_float_const_zero_rtx_p): ...this.
21169 (aarch64_print_operand): Allow printing of new constants.
21170 (aarch64_valid_floating_const): New.
21171 (aarch64_legitimate_constant_p): Check for valid floating-point
21172 constants.
21173 (aarch64_simd_valid_immediate): Likewise.
21174 (aarch64_vect_float_const_representable_p): New.
21175 (aarch64_float_const_representable_p): Likewise.
21176 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
21177 (aarch64_output_simd_mov_immediate): New.
21178 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
21179 (*movdf_aarch64): Likewise.
21180 * config/aarch64/constraints.md (Ufc): New.
21181 (Y): call aarch64_float_const_zero_rtx.
21182 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
21183
e167c04d
RB
211842013-01-07 Richard Biener <rguenther@suse.de>
21185
21186 PR tree-optimization/55888
21187 PR tree-optimization/55862
21188 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
21189 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
21190 not if it is contained therein.
21191
0139aaab
GJL
211922013-01-07 Georg-Johann Lay <avr@gjlay.de>
21193
21194 * config/avr/t-avr: Typo.
21195
4a176b23
GJL
211962013-01-07 Georg-Johann Lay <avr@gjlay.de>
21197
21198 PR55243
21199 * config/avr/t-avr: Don't automatically rebuild
21200 $(srcdir)/config/avr/t-multilib
21201 $(srcdir)/config/avr/avr-tables.opt
21202 $(srcdir)/doc/avr-mmcu.texi
21203 (avr-mcus): New phony target to build them on request.
21204 (s-avr-mlib, s-avr-mmcu-texi): Remove.
21205 * avr/avr-mcus.def: Adjust comments.
21206
c7afdc98
UB
212072013-01-07 Uros Bizjak <ubizjak@gmail.com>
21208
21209 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
21210
1ab05c31
RS
212112013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21212
21213 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
21214
488e3acc
RS
212152013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21216
21217 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
21218
a89599d2
DE
212192013-01-05 David Edelsohn <dje.gcc@gmail.com>
21220
21221 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
21222 to generate profiling.
21223 * config/rs6000/aix64.h (LIB_SPEC): Same.
21224
70f09188
AP
212252013-01-04 Andrew Pinski <apinski@cavium.com>
21226
21227 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
21228 New function.
21229 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
21230
918d445f
UB
212312013-01-04 Uros Bizjak <ubizjak@gmail.com>
21232
21233 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
21234 unconditionally.
21235 (ix86_expand_move): Ditto.
21236 (ix86_zero_extend_to_Pmode): Ditto.
21237 (ix86_expand_call): Ditto.
21238 (ix86_expand_special_args_builtin): Ditto.
21239 (ix86_expand_builtin): Ditto.
21240
361618ec
RB
212412013-01-04 Richard Biener <rguenther@suse.de>
21242
21243 PR tree-optimization/55862
21244 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21245 translating them through PHI nodes.
21246
87eab554
MJ
212472013-01-04 Martin Jambor <mjambor@suse.cz>
21248
21249 PR tree-optimization/55755
21250 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21251 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21252
33879b9f
MP
212532013-01-04 Marek Polacek <polacek@redhat.com>
21254
21255 PR middle-end/55859
21256 * opts.c (default_options_optimization): Clarify error message.
21257
3068819a
RB
212582013-01-04 Richard Biener <rguenther@suse.de>
21259
21260 PR middle-end/55863
21261 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21262 reassociation.
21263
0e4ae794
JDA
212642013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21265
21266 PR target/53789
21267 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA
21268 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21269 references.
0e4ae794 21270
65c88cef
RH
212712013-01-03 Richard Henderson <rth@redhat.com>
21272
21273 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21274 after eliminating TLS symbols.
21275
8ac16127
MG
212762013-01-03 Marc Glisse <marc.glisse@inria.fr>
21277
21278 PR bootstrap/50167
21279 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21280 * graphite-poly.c (debug_gmp_value): Likewise.
21281
bb664f09
UB
212822013-01-03 Uros Bizjak <ubizjak@gmail.com>
21283
21284 PR target/55712
21285 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21286 selected code model, define __code_mode_small__, __code_model_medium__,
21287 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21288 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21289 xchg temporary register with %k. Declare temporary register as
21290 early clobbered.
21291 [__x86_64__]: For medium and large code models, preserve %rbx register.
21292
bcf1ef00
RB
212932013-01-03 Richard Biener <rguenther@suse.de>
21294
0506634a 21295 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB
21296 (dump_subscript): Adjust.
21297 (finalize_ddr_dependent): Do not dump redundant info.
21298 (analyze_siv_subscript): Adjust.
21299 (subscript_dependence_tester): Likewise.
21300 (compute_affine_dependence): Likewise.
21301
59fd17e3
RB
213022013-01-03 Richard Biener <rguenther@suse.de>
21303
21304 Revert
21305 2013-01-03 Richard Biener <rguenther@suse.de>
21306
21307 PR tree-optimization/55857
21308 * tree-vect-stmts.c (vectorizable_load): Do not setup
21309 re-alignment for invariant loads.
21310
21311 2013-01-02 Richard Biener <rguenther@suse.de>
21312
21313 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 21314 invariant load do not generate a vector load from the scalar location.
59fd17e3 21315
595c2679
RB
213162013-01-03 Richard Biener <rguenther@suse.de>
21317
21318 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21319 for not vectorizing.
21320 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21321 not build INDIRECT_REFs, call get_name once only.
21322 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21323 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21324
90bb7d7a
RB
213252013-01-03 Richard Biener <rguenther@suse.de>
21326
21327 PR tree-optimization/55857
21328 * tree-vect-stmts.c (vectorizable_load): Do not setup
21329 re-alignment for invariant loads.
21330
f09b77ca
RB
213312013-01-03 Richard Biener <rguenther@suse.de>
21332
21333 PR lto/55848
21334 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
21335 prefer a built-in decl.
21336
3461a16e
JJ
213372013-01-03 Jakub Jelinek <jakub@redhat.com>
21338
df8e2b4f
JJ
21339 * gcc.c (process_command): Update copyright notice dates.
21340 * gcov.c (print_version): Likewise.
21341 * gcov-dump.c (print_version): Likewise.
21342
3461a16e
JJ
21343 PR rtl-optimization/55838
21344 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
21345 iv0.step, iv1.step and step.
21346
8b5546d6
JJ
213472013-01-03 Jakub Jelinek <jakub@redhat.com>
21348 Marc Glisse <marc.glisse@inria.fr>
21349
21350 PR tree-optimization/55832
21351 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
21352 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
21353 integer_{one,zero}_node.
21354
8ab1d2e9
JJ
213552013-01-03 Jakub Jelinek <jakub@redhat.com>
21356
21357 PR debug/54402
21358 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
21359 * var-tracking.c (reverse_op): Don't add reverse ops to
21360 VALUEs that have already
0506634a 21361 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 21362
5b9ad1d4
GP
213632013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
21364
21365 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
21366
e25a6711
TJ
213672013-01-02 Teresa Johnson <tejohnson@google.com>
21368
21369 * dumpfile.c (dump_loc): Print filename with location.
21370 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
21371 new location_t parameter to emit complete unroll message with
21372 new dump framework.
21373 (canonicalize_loop_induction_variables): Compute loops location
21374 and pass to try_unroll_loop_completely.
21375 * loop-unroll.c (report_unroll_peel): New function.
21376 (peel_loops_completely): Use new dump format with location
21377 for main dumpfile message, and invoke report_unroll_peel on success.
21378 (decide_unrolling_and_peeling): Ditto.
21379 (decide_peel_once_rolling): Remove old dumpfile message subsumed
21380 by report_unroll_peel.
21381 (decide_peel_completely): Ditto.
21382 (decide_unroll_constant_iterations): Ditto.
21383 (decide_unroll_runtime_iterations): Ditto.
21384 (decide_peel_simple): Ditto.
21385 (decide_unroll_stupid): Ditto.
21386 * cfgloop.c (get_loop_location): New function.
21387 * cfgloop.h (get_loop_location): Declare.
21388
77878621
ST
213892013-01-02 Sriraman Tallam <tmsriram@google.com>
21390
21391 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21392 NULL.
21393
9e65d03e
JDA
213942013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21395
21396 PR middle-end/55198
21397 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21398 BLKmode objects when EXPAND_MEMORY is specified.
21399
6a7da30f
ST
214002013-01-02 Sriraman Tallam <tmsriram@google.com>
21401
21402 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21403 in loop predicate.
21404 (fold_builtin_cpu): Do not share cpu model decls across statements.
21405
e78167f2
JM
214062013-01-02 Jason Merrill <jason@redhat.com>
21407
21408 PR c++/55804
21409 * tree.c (build_array_type_1): Revert earlier change.
21410
8c075fb4
YZ
214112013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21412
21413 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21414 "cortex-a57".
21415 * config/aarch64/aarch64-tune.md: Re-generate.
21416
0682ed3e
RB
214172013-01-02 Richard Biener <rguenther@suse.de>
21418
21419 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 21420 invariant load do not generate a vector load from the scalar location.
0682ed3e 21421
b8f6a302
RB
214222013-01-02 Richard Biener <rguenther@suse.de>
21423
21424 PR bootstrap/55784
21425 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21426 * configure: Regenerate.
21427
04b80dbb
RS
214282013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21429
21430 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21431 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21432 (expand_builtin_int_roundingfn_2): Keep the original target around
21433 for the fallback case.
21434
635b0b0c
RS
214352013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21436
21437 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21438 to be clear for sign changes.
21439
42d57399
JH
214402013-01-01 Jan Hubicka <jh@suse.cz>
21441
21442 * ipa-inline-analysis.c: Fix formatting.
21443
5bb6669d
JJ
214442013-01-01 Jakub Jelinek <jakub@redhat.com>
21445
21446 PR tree-optimization/55831
21447 * tree-vect-loop.c (get_initial_def_for_induction): Use
21448 gsi_after_labels instead of gsi_start_bb.
ad41bd84 21449\f
86a2db33 21450Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM
21451
21452Copying and distribution of this file, with or without modification,
21453are permitted in any medium without royalty provided the copyright
21454notice and this notice are preserved.