1 2013-09-12 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/58404
4 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5 propagate non-invariant addresses into dereferences wrapped
6 in component references.
8 2013-09-12 Richard Biener <rguenther@suse.de>
10 PR tree-optimization/58402
11 * passes.def: Move pass_late_warn_uninitialized later.
13 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
15 * tree-ssa.h: New. Move content from tree-flow.h and
17 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
18 Move prototypes belonging to tree-ssa.c.
19 * tree-flow-inline.h (redirect_edge_var_map_def,
20 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
22 * gimple.h: Adjust prototypes.
23 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move to...
24 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
25 * tree.h: Move prototype to tree-ssa.h.
26 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
27 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
28 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
29 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
30 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
31 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
32 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
33 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
34 graphite-blocking.c, graphite-clast-to-gimple.c, graphite-dependences.c,
35 graphite-interchange.c, graphite-optimize-isl.c, graphite-poly.c,
36 graphite-scop-detection.c, graphite-sese-to-poly.c, graphite.c,
37 ipa-cp.c, ipa-inline-analysis.c, ipa-inline-transform.c, ipa-inline.c,
38 ipa-prop.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa-utils.c,
39 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
40 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
41 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
42 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
43 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
44 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
45 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
46 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
47 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
48 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
49 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
50 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
51 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
52 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
53 value-prof.c, var-tracking.c,
54 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
56 2013-09-12 Richard Biener <rguenther@suse.de>
58 PR tree-optimization/58396
59 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
60 (build_rdg): Take a loop-nest parameter, fix memleaks.
61 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
63 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
65 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
68 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
70 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
71 The FMA instruction names should have a 'v' prefix.
73 2013-09-12 Richard Biener <rguenther@suse.de>
75 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
76 (dot_rdg): Use popen instead of system in optional code.
77 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
78 (already_processed_vertex_p): Adjust.
79 (has_anti_or_output_dependence, predecessor_has_mem_write,
80 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
81 rdg_flag_uses): Remove.
82 (rdg_flag_vertex): Simplify.
83 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
85 (build_rdg_partition_for_component): Process the first vertex
87 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
89 2013-09-12 Alan Modra <amodra@gmail.com>
91 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
93 2013-09-11 DJ Delorie <dj@redhat.com>
94 Nick Clifton <nickc@redhat.com>
96 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
97 (rl78_stack_based_mem): New.
98 * config/rl78/constraints.md (Iv08): New.
107 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
109 (movsi): Change from expand to insn-and-split.
110 (ashrsi3): Clobber AX.
114 * config/rl78/rl78.md (CC_REG): Fix.
115 (addsi3): Allow memory and immediate operands.
116 (addsi3_internal): Split into...
117 (addsi3_internal_virt): ...new, and ...
118 (addsi3_internal_real): ...new.
120 (subsi3_internal_virt): New.
121 (subsi3_internal_real): New.
122 (mulsi3): Add memory operand.
123 (mulsi3_rl78): Likewise.
124 (mulsi3_g13): Likewise.
125 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
126 (cbranchqi4_real): Add more constraint options.
127 (cbranchhi4_real): Expand pattern.
128 (cbranchhi4_real_signed): New.
129 (cbranchhi4_real_inverted): New.
130 (cbranchsi4_real_lt): New.
131 (cbranchsi4_real_ge): New.
132 (cbranchsi4_real_signed): New.
133 (cbranchsi4_real): New.
135 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
137 (lshrsi3_virt): Likewise.
138 (ashlsi3_virt): Likewise.
139 (cbranchqi4_virt_signed): New.
140 (cbranchhi4_virt_signed): New.
141 (cbranchsi4_virt): New.
142 * config/rl78/rl78.c: Whitespace fixes throughout.
143 (move_elim_pass): New.
144 (pass_data_rl78_move_elim): New.
145 (pass_rl78_move_elim): New.
146 (make_pass_rl78_move_elim): New.
147 (rl78_devirt_info): Run devirt earlier.
148 (rl78_move_elim_info): New.
149 (rl78_asm_file_start): Register it.
150 (rl78_split_movsi): New.
151 (rl78_as_legitimate_address): Allow virtual base registers when
153 (rl78_addr_space_convert): Remove spurious debug stuff.
154 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
155 (rl78_print_operand): More cases for not printing '#'.
156 (rl78_expand_compare): Remove most of the logic.
157 (content_memory): New.
158 (clear_content_memory): New.
159 (get_content_index): New.
160 (get_content_name): New.
161 (display_content_memory): New.
162 (update_content): New.
163 (record_content): New.
164 (already_contains): New.
165 (insn_ok_now): Re-recog insns with virtual registers.
166 (add_postponed_content_update): New.
167 (process_postponed_content_update): New.
168 (gen_and_emit_move): New.
169 (transcode_memory_rtx): Record new location content.
170 Use gen_and_emit_move.
171 (force_into_acc): New.
172 (move_to_acc): Use gen_and_emit_move.
173 (move_from_acc): Likewise.
174 (move_acc_to_reg): Likewise.
175 (move_to_x): Likewise.
176 (move_to_hl): Likewise.
177 (move_to_de): Likewise.
178 (rl78_alloc_physical_registers_op1): Record location content.
179 (has_constraint): New.
180 (rl78_alloc_physical_registers_op2): Record location content.
182 (rl78_alloc_physical_registers_ro1): Likewise.
183 (rl78_alloc_physical_registers_cmp): Likewise.
184 (rl78_alloc_physical_registers_umul): Likewise.
185 (rl78_alloc_address_registers_macax): New.
186 (rl78_alloc_physical_registers): Initialize and set location
187 content memory as needed.
188 (rl78_reorg): Make sure split2 is called.
189 (rl78_rtx_costs): New.
191 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
193 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
194 for (not (neg ...)) and (neg (not ...)) cases.
196 2013-09-11 Richard Biener <rguenther@suse.de>
199 * passes.def: Split critical edges before late uninit warning passes.
200 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
202 2013-09-11 Jakub Jelinek <jakub@redhat.com>
204 PR tree-optimization/58385
205 * fold-const.c (build_range_check): If both low and high are NULL,
206 use omit_one_operand_loc to preserve exp side-effects.
208 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
210 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
212 2013-09-11 Richard Biener <rguenther@suse.de>
214 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
215 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
216 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
217 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
218 stmts_from_loop, known_dependences_p, build_empty_rdg,
219 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
220 * tree-loop-distribution.c: ... here.
221 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
222 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
223 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
224 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
225 * tree-loop-distribution.c: ... here.
226 * tree-loop-distribution.c: Include gimple-pretty-print.h.
227 (struct partition_s): Add loops member.
228 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
229 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
230 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
232 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
233 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
234 Sergey Lega <sergey.s.lega@intel.com>
235 Anna Tikhonova <anna.tikhonova@intel.com>
236 Ilya Tocar <ilya.tocar@intel.com>
237 Andrey Turetskiy <andrey.turetskiy@intel.com>
238 Ilya Verbin <ilya.verbin@intel.com>
239 Kirill Yukhin <kirill.yukhin@intel.com>
240 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
242 * config/i386/constraints.md (k): New.
244 * config/i386/i386.c (const regclass_map): Add new mask registers.
245 (dbx_register_map): Ditto.
246 (dbx64_register_map): Ditto.
247 (svr4_dbx_register_map): Ditto.
248 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
250 (ix86_preferred_reload_class): Disable constants for mask registers.
251 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
252 (ix86_hard_regno_mode_ok): Support new mask registers.
253 (x86_order_regs_for_local_alloc): Ditto.
254 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
255 (FIXED_REGISTERS): Add new mask registers.
256 (CALL_USED_REGISTERS): Ditto.
257 (REG_ALLOC_ORDER): Ditto.
258 (VALID_MASK_REG_MODE): New.
259 (FIRST_MASK_REG): Ditto.
260 (LAST_MASK_REG): Ditto.
261 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
262 (MAYBE_MASK_CLASS_P): New.
263 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
264 (REG_CLASS_CONTENTS): Ditto.
266 (ANY_MASK_REG_P): Ditto.
267 (HI_REGISTER_NAMES): Add new mask registers.
268 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
269 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
271 (attribute "type"): Add mskmov, msklog.
272 (attribute "length_immediate"): Support them.
273 (attribute "memory"): Ditto.
274 (attribute "prefix_0f"): Ditto.
275 (*movhi_internal): Support new mask registers.
276 (*movqi_internal): Ditto.
277 (define_split): Split out clobber pattern is a logic
278 insn on mask registers.
279 (*k<logic><mode>): New.
280 (*andhi_1): Extend to support mask regs.
281 (*andqi_1): Extend to support mask regs.
283 (define_split): Split and-not to and and not if operands
285 (*<code><mode>_1): Separate HI mode to new pattern...
287 (*<code>qi_1): Extend to support mask regs.
292 (*one_cmpl<mode>2_1): Remove HImode and handle it...
293 (*one_cmplhi2_1): ...Here, now with mask registers support.
294 (*one_cmplqi2_1): Support new mask registers.
295 (HI/QImode arithmetics splitter): Don't split if mask registers
297 (HI/QImode not splitter): Ditto.
298 * config/i386/predicated.md (mask_reg_operand): New.
299 (general_reg_operand): Ditto.
301 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
303 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
304 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
306 2013-09-10 Jeff Law <law@redhat.com>
308 PR tree-optimization/58380
309 * tree-ssa-threadupdate.c (thread_block): Recognize another case
310 of threading through a buried loop header.
312 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
313 return value for single successor case.
315 2013-09-10 Jan Hubicka <jh@suse.cz>
317 * ipa-devirt.c (ipa_devirt): Enable with LTO.
319 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
322 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
323 support conditional execution.
324 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
326 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
328 * lra.c (lra): Clear lra_optional_reload_pseudos before every
330 * lra-constraints.c (curr_insn_transform): Switch on optional
331 reloads. Check destination too to check move insn.
332 (undo_optional_reloads): Add check that the original peudo did not
333 changed its allocation and the optional reload was inherited on
334 last inheritance pass. Break loop after deciding to keep optional
336 (lra_undo_inheritance): Add check that inherited pseudo still in
339 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
341 * config/aarch64/aarch64.md (generic_sched): New.
342 * config/aarch64/aarch64-generic.md (load): Make conditional
343 on generic_sched attribute.
346 2013-09-10 Jan Hubicka <jh@suse.cz>
348 * lto-cgraph.c: Include ipa-utils.h.
349 (compute_ltrans_boundary): Also add possible targets into the boundary.
351 2013-09-10 Jan Hubicka <jh@suse.cz>
353 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
354 VAR_DECL of vtable rather than full expression.
356 2013-09-10 Jan Hubicka <jh@suse.cz>
357 Paolo Carlini <paolo.carlini@oracle.com>
359 * cgraphunit.c (analyze_functions): Save input_location, set it
360 to UNKNOWN_LOCATION and restore it at the end.
362 2013-09-10 Martin Jambor <mjambor@suse.cz>
364 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
365 represented by a thunk.
367 2013-09-10 Jeff Law <law@redhat.com>
369 PR tree-optimization/58343
370 * tree-ssa-threadupdate.c (thread_block): Identify and disable
371 jump threading requests through loop headers buried in the middle
372 of a jump threading path.
374 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
375 in return value/type.
377 2013-09-10 Jakub Jelinek <jakub@redhat.com>
379 PR rtl-optimization/58365
380 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
381 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
384 2013-09-10 Richard Biener <rguenther@suse.de>
386 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
387 * tree-data-ref.c (create_rdg_vertices): Collect all data
388 references, signal failure to the caller, use data-ref API.
389 (build_rdg): Compute data references only once. Maintain lifetime
390 of data references and data dependences from within RDG.
391 (free_rdg): Free dependence relations.
392 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
393 inventing extra dependences.
394 (distribute_loop): Update for RDG API changes.
396 2013-09-10 Kai Tietz <ktietz@redhat.com>
398 * doc/invoke.texi (fms-extensions): Document changed
399 behavior for ms-abi targets.
400 * config/i386/i386.c (ix86_option_override_internal):
401 Set default value of option -fms-extension for ms-abi targets.
403 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
405 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
407 2013-09-10 Alan Modra <amodra@gmail.com>
410 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
412 2013-09-10 Alan Modra <amodra@gmail.com>
414 * config/rs6000/predicates.md (add_cint_operand): New.
415 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
416 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
417 using add_cint_operand.
418 (largetoc_high_plus_aix): Likewise.
420 2013-09-09 Jakub Jelinek <jakub@redhat.com>
422 PR tree-optimization/58364
423 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
424 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
425 the current range can't be an unconditional true or false.
427 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
429 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
431 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
433 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
435 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
438 * doc/invoke.texi (-Wdelete-incomplete): Document it.
440 2013-09-09 Ian Bolton <ian.bolton@arm.com>
442 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
443 NO_REGS for immediate that can't be moved directly into FP_REGS.
445 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
447 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
448 comparison with negated operand.
449 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
452 2013-09-09 Richard Biener <rguenther@suse.de>
455 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
456 of a subloop record all its block as affecting loop-closed SSA form.
458 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
460 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
461 of an rtx/bitpos pair.
462 (store_fixed_bit_field): Update accordingly.
464 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
466 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
468 * builtins.c (expand_errno_check): Likewise.
469 * dwarf2cfi.c (init_return_column_size): Likewise.
470 * except.c (sjlj_mark_call_sites): Likewise.
471 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
472 * lra-constraints.c (emit_inc): Likewise.
473 * ree.c (combine_set_extension): Likewise.
474 * regmove.c (fixup_match_2): Likewise.
475 * reload1.c (inc_for_reload): Likewise.
477 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
479 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
480 (force_to_mode, simplify_shift_const_1, simplify_comparison):
481 Use gen_int_mode with the mode of the associated simplify_* call.
482 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
483 * expmed.c (expand_shift_1): Likewise.
484 * function.c (instantiate_virtual_regs_in_insn): Likewise.
485 * loop-iv.c (iv_number_of_iterations): Likewise.
486 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
487 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
489 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
491 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
492 of the associated expand_* call.
493 (asan_emit_stack_protection): Likewise.
494 * builtins.c (round_trampoline_addr): Likewise.
495 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
496 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
497 (emit_store_flag): Likewise.
498 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
499 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
501 * function.c (instantiate_virtual_regs_in_insn): Likewise.
502 * ifcvt.c (noce_try_store_flag_constants): Likewise.
503 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
504 * modulo-sched.c (generate_prolog_epilog): Likewise.
505 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
506 (expand_ctz, expand_ffs, expand_unop): Likewise.
508 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
510 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
511 of the associated gen_rtx_* call.
512 * caller-save.c (init_caller_save): Likewise.
513 * combine.c (find_split_point, make_extraction): Likewise.
514 (make_compound_operation): Likewise.
515 * dwarf2out.c (mem_loc_descriptor): Likewise.
516 * explow.c (plus_constant, probe_stack_range): Likewise.
517 * expmed.c (expand_mult_const): Likewise.
518 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
519 * reload1.c (init_reload): Likewise.
520 * valtrack.c (cleanup_auto_inc_dec): Likewise.
521 * var-tracking.c (adjust_mems): Likewise.
522 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
523 rather than gen_rtx_fmt_ee.
525 2013-09-09 Jan Hubicka <jh@suse.cz>
528 * value-prof.c (gimple_ic): Copy also abnormal edges.
530 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
532 * asan.c (asan_shadow_cst): Use gen_int_mode.
534 2013-09-08 Jan Hubicka <jh@suse.cz>
536 * ipa-profile.c: Add toplevel comment.
537 (ipa_propagate_frequency_1): Be more conservative when profile is read.
538 (contains_hot_call_p): New function.
539 (ipa_propagate_frequency): Set frequencies based on counts when
541 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
542 profile; do not tamper with profile after inlining if it is read.
544 2013-09-08 Jan Hubicka <jh@suse.cz>
546 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
549 2013-09-08 Jan Hubicka <jh@suse.cz>
551 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
554 2013-09-08 Jeff Law <law@redhat.com>
557 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
560 2013-09-08 Andi Kleen <ak@linux.intel.com>
562 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
564 2013-09-08 Jan Hubicka <jh@suse.cz>
566 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
568 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
570 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
572 * lra.c (new_insn_reg): Take the containing insn as a parameter.
573 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
574 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
577 2013-09-08 Jan Hubicka <jh@suse.cz>
579 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
580 targets and devirtualize to BUILT_IN_UNREACHABLE.
581 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
582 * ipa.c (walk_polymorphic_call_targets): New function.
583 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
584 functions; use the new timevar.
585 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
586 was removed from the program.
587 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
588 not consider it in the walk when its vtable is dead.
589 (possible_polymorphic_call_targets_1): Pass anonymous flag to
591 (devirt_variable_node_removal_hook): New function.
592 (possible_polymorphic_call_targets): Also register
593 devirt_variable_node_removal_hook.
594 (ipa_devirt): Do not do non-speculative devirtualization.
595 (gate_ipa_devirt): One execute if devirtualizing speculatively.
597 2013-09-08 Jan Hubicka <jh@suse.cz>
599 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
600 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
601 varpool_remove_variable_insertion_hook): Declare.
602 * varpool.c (varpool_node_hook_list): New structure.
603 (first_varpool_node_removal_hook,
604 first_varpool_variable_insertion_hook): New variables.
605 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
606 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
607 varpool_remove_variable_insertion_hook,
608 varpool_call_variable_insertion_hooks): New functions.
609 (varpool_remove_node): Use it.
611 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
614 * diagnostic.c (diagnostic_build_prefix): When s.file is
615 "<built-in>" don't output line and column numbers.
617 2013-09-06 Jan Hubicka <jh@suse.cz>
619 * cgraphunit.c (expand_thunk): Get body before touching arguments.
620 * lto-streamer-out.c: Stream thunks, too.
621 * lto-streamer-in.c (input_function): Pop cfun here
622 (lto_read_body): Instead of here.
624 2013-09-06 Caroline Tice <cmtice@google.com>
626 * doc/install.texi: Add documentation for the --enable-vtable-verify
627 and the --disable-libvtv configure options.
629 2013-09-06 Jeff Law <law@redhat.com>
631 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
632 edge implied equivalences into successor phis.
634 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
636 * resource.c (mark_referenced_resources): Handle COND_EXEC.
638 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
640 * resource.c (mark_target_live_regs): Compute resources taking
641 into account if a call is predicated or not.
643 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
645 * toplev.c (output_stack_usage): Be prepared for suffixes created by
646 the compiler in the function names.
648 2013-09-06 Jan Hubicka <jh@suse.cz>
651 * ipa-inline.c (has_caller_p): New function.
652 (want_inline_function_to_all_callers_p): Use it.
653 (sum_callers, inline_to_all_callers): Break out from ...
654 (ipa_inline): ... here.
656 2013-09-06 Jan Hubicka <jh@suse.cz>
658 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
659 only when AVX is enabled.
661 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
663 * config/aarch64/aarch64.md
664 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
665 is fpsimd_<load/store>2.
666 (load_pair<mode>): Likewise.
667 (store_pair<mode>): Likewise.
669 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
671 * config/arm/types.md (type): Add "mrs" type.
672 * config/aarch64/aarch64.md
673 (aarch64_load_tp_hard): Make type "mrs".
675 (load_tp_hard): Make type "mrs".
676 * config/arm/cortex-a15.md: Update with new attributes.
677 * config/arm/cortex-a5.md: Update with new attributes.
678 * config/arm/cortex-a53.md: Update with new attributes.
679 * config/arm/cortex-a7.md: Update with new attributes.
680 * config/arm/cortex-a8.md: Update with new attributes.
681 * config/arm/cortex-a9.md: Update with new attributes.
682 * config/arm/cortex-m4.md: Update with new attributes.
683 * config/arm/cortex-r4.md: Update with new attributes.
684 * config/arm/fa526.md: Update with new attributes.
685 * config/arm/fa606te.md: Update with new attributes.
686 * config/arm/fa626te.md: Update with new attributes.
687 * config/arm/fa726te.md: Update with new attributes.
689 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
691 * config/aarch64/aarch64.md
692 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
693 (*movtf_aarch64): Likewise.
695 (thumb1_movdi_insn): Use "multiple" for type where more than one
696 instruction is used for a move.
697 (*arm32_movhf): Likewise.
698 (*thumb_movdf_insn): Likewise.
700 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
702 * config/arm/types.md (type): Rename fcpys to fmov.
704 (*arm_movsi_vfp): Rename type fcpys as fmov.
705 (*thumb2_movsi_vfp): Likewise
706 (*movhf_vfp_neon): Likewise
707 (*movhf_vfp): Likewise
708 (*movsf_vfp): Likewise
709 (*thumb2_movsf_vfp): Likewise
710 (*movsfcc_vfp): Likewise
711 (*thumb2_movsfcc_vfp): Likewise
712 * config/aarch64/aarch64-simd.md
713 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
714 * config/aarch64/aarch64.md
715 (*movsi_aarch64): Replace type mov_reg with fmovs.
716 (*movdi_aarch64): Likewise
717 (*movsf_aarch64): Likewise
718 (*movdf_aarch64): Likewise
720 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
721 * config/arm/iwmmxt.md
722 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
723 * config/arm/arm1020e.md: Update with new attributes.
724 * config/arm/cortex-a15-neon.md: Update with new attributes.
725 * config/arm/cortex-a5.md: Update with new attributes.
726 * config/arm/cortex-a53.md: Update with new attributes.
727 * config/arm/cortex-a7.md: Update with new attributes.
728 * config/arm/cortex-a8-neon.md: Update with new attributes.
729 * config/arm/cortex-a9.md: Update with new attributes.
730 * config/arm/cortex-m4-fpu.md: Update with new attributes.
731 * config/arm/cortex-r4f.md: Update with new attributes.
732 * config/arm/marvell-pj4.md: Update with new attributes.
733 * config/arm/vfp11.md: Update with new attributes.
735 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
737 * config/aarch64/aarch64.md
738 (*madd<mode>): Fix type attribute.
739 (*maddsi_uxtw): Likewise.
740 (*msub<mode>): Likewise.
741 (*msubsi_uxtw): Likewise.
742 (<su_optab>maddsidi4): Likewise.
743 (<su_optab>msubsidi4): Likewise.
745 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
747 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
748 * config/arm/arm.md (core_cycles): Remove fdiv.
750 (*sqrtsf2_vfp): Update for attribute changes.
751 (*sqrtdf2_vfp): Likewise.
752 * config/aarch64/aarch64.md:
753 (sqrt<mode>2): Update for attribute changes.
754 * config/arm/arm1020e.md: Update with new attributes.
755 * config/arm/cortex-a15-neon.md: Update with new attributes.
756 * config/arm/cortex-a5.md: Update with new attributes.
757 * config/arm/cortex-a53.md: Update with new attributes.
758 * config/arm/cortex-a7.md: Update with new attributes.
759 * config/arm/cortex-a8-neon.md: Update with new attributes.
760 * config/arm/cortex-a9.md: Update with new attributes.
761 * config/arm/cortex-m4-fpu.md: Update with new attributes.
762 * config/arm/cortex-r4f.md: Update with new attributes.
763 * config/arm/marvell-pj4.md: Update with new attributes.
764 * config/arm/vfp11.md: Update with new attributes.
766 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
768 * config/arm/types.md
769 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
770 * config/aarch64/aarch64.md
771 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
773 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
774 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
775 (float<GPI:mode><GPF:mode>2): Likewise.
777 (*truncsisf2_vfp): Update with new attributes.
778 (*truncsidf2_vfp): Likewise.
779 (fixuns_truncsfsi2): Likewise.
780 (fixuns_truncdfsi2): Likewise.
781 (*floatsisf2_vfp): Likewise.
782 (*floatsidf2_vfp): Likewise.
783 (floatunssisf2): Likewise.
784 (floatunssidf2): Likewise.
785 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
786 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
787 * config/arm/arm1020e.md: Update with new attributes.
788 * config/arm/cortex-a15-neon.md: Update with new attributes.
789 * config/arm/cortex-a5.md: Update with new attributes.
790 * config/arm/cortex-a53.md: Update with new attributes.
791 * config/arm/cortex-a7.md: Update with new attributes.
792 * config/arm/cortex-a8-neon.md: Update with new attributes.
793 * config/arm/cortex-a9.md: Update with new attributes.
794 * config/arm/cortex-m4-fpu.md: Update with new attributes.
795 * config/arm/cortex-r4f.md: Update with new attributes.
796 * config/arm/marvell-pj4.md: Update with new attributes.
797 * config/arm/vfp11.md: Update with new attributes.
799 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
801 * config/aarch64/arm_neon.h
802 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
803 (vqtbx<1,2,3,4><q>_s8): Likewise.
805 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
807 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
808 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
809 (add<mode>3): Add type attribute.
810 (add<mode>3): Likewise.
811 (usadd<mode>3): Likewise.
812 (ssadd<mode>3): Likewise.
813 (sub<mode>3): Likewise.
814 (sub<mode>3): Likewise.
815 (ussub<mode>3): Likewise.
816 (sssub<mode>3): Likewise.
817 (ssmulsa3): Likewise.
818 (usmulusa3): Likewise.
819 (arm_usatsihi): Likewise.
821 (*movdi_vfp): Add types for all instructions.
822 (*movdi_vfp_cortexa8): Likewise.
823 (*movhf_vfp_neon): Likewise.
824 (*movhf_vfp): Likewise.
825 (*movdf_vfp): Likewise.
826 (*thumb2_movdf_vfp): Likewise.
827 (*thumb2_movdfcc_vfp): Likewise.
828 * config/arm/arm.md: Add type attribute to all insn patterns.
829 (*thumb1_adddi3): Add type attribute.
830 (*arm_adddi3): Likewise.
831 (*adddi_sesidi_di): Likewise.
832 (*adddi_zesidi_di): Likewise.
833 (*thumb1_addsi3): Likewise.
834 (addsi3_compare0): Likewise.
835 (*addsi3_compare0_scratch): Likewise.
836 (*compare_negsi_si): Likewise.
837 (cmpsi2_addneg): Likewise.
838 (*addsi3_carryin_<optab>): Likewise.
839 (*addsi3_carryin_alt2_<optab>): Likewise.
840 (*addsi3_carryin_clobercc_<optab>): Likewise.
841 (*subsi3_carryin): Likewise.
842 (*subsi3_carryin_const): Likewise.
843 (*subsi3_carryin_compare): Likewise.
844 (*subsi3_carryin_compare_const): Likewise.
845 (*arm_subdi3): Likewise.
846 (*thumb_subdi3): Likewise.
847 (*subdi_di_zesidi): Likewise.
848 (*subdi_di_sesidi): Likewise.
849 (*subdi_zesidi_di): Likewise.
850 (*subdi_sesidi_di): Likewise.
851 (*subdi_zesidi_ze): Likewise.
852 (thumb1_subsi3_insn): Likewise.
853 (*arm_subsi3_insn): Likewise.
854 (*anddi3_insn): Likewise.
855 (*anddi_zesidi_di): Likewise.
856 (*anddi_sesdi_di): Likewise.
857 (*ne_zeroextracts): Likewise.
858 (*ne_zeroextracts): Likewise.
859 (*ite_ne_zeroextr): Likewise.
860 (*ite_ne_zeroextr): Likewise.
861 (*anddi_notdi_di): Likewise.
862 (*anddi_notzesidi): Likewise.
863 (*anddi_notsesidi): Likewise.
864 (andsi_notsi_si): Likewise.
865 (thumb1_bicsi3): Likewise.
866 (*iordi3_insn): Likewise.
867 (*iordi_zesidi_di): Likewise.
868 (*iordi_sesidi_di): Likewise.
869 (*thumb1_iorsi3_insn): Likewise.
870 (*xordi3_insn): Likewise.
871 (*xordi_zesidi_di): Likewise.
872 (*xordi_sesidi_di): Likewise.
873 (*arm_xorsi3): Likewise.
874 (*andsi_iorsi3_no): Likewise.
876 (*smax_m1): Likewise.
877 (*arm_smax_insn): Likewise.
879 (*arm_smin_insn): Likewise.
880 (*arm_umaxsi3): Likewise.
881 (*arm_uminsi3): Likewise.
882 (*minmax_arithsi): Likewise.
883 (*minmax_arithsi_): Likewise.
884 (*satsi_<SAT:code>): Likewise.
885 (arm_ashldi3_1bit): Likewise.
886 (arm_ashrdi3_1bit): Likewise.
887 (arm_lshrdi3_1bit): Likewise.
888 (*arm_negdi2): Likewise.
889 (*thumb1_negdi2): Likewise.
890 (*arm_negsi2): Likewise.
891 (*thumb1_negsi2): Likewise.
892 (*negdi_extendsid): Likewise.
893 (*negdi_zero_extend): Likewise.
894 (*arm_abssi2): Likewise.
895 (*thumb1_abssi2): Likewise.
896 (*arm_neg_abssi2): Likewise.
897 (*thumb1_neg_abss): Likewise.
898 (one_cmpldi2): Likewise.
899 (extend<mode>di2): Likewise.
900 (*compareqi_eq0): Likewise.
901 (*arm_extendhisi2addsi): Likewise.
902 (*arm_movdi): Likewise.
903 (*thumb1_movdi_insn): Likewise.
904 (*arm_movt): Likewise.
905 (*thumb1_movsi_insn): Likewise.
906 (pic_add_dot_plus_four): Likewise.
907 (pic_add_dot_plus_eight): Likewise.
908 (tls_load_dot_plus_eight): Likewise.
909 (*thumb1_movhi_insn): Likewise.
910 (*thumb1_movsf_insn): Likewise.
911 (*movdf_soft_insn): Likewise.
912 (*thumb_movdf_insn): Likewise.
913 (cbranchsi4_insn): Likewise.
914 (cbranchsi4_scratch): Likewise.
915 (*negated_cbranchsi4): Likewise.
916 (*tbit_cbranch): Likewise.
917 (*tlobits_cbranch): Likewise.
918 (*tstsi3_cbranch): Likewise.
919 (*cbranchne_decr1): Likewise.
920 (*addsi3_cbranch): Likewise.
921 (*addsi3_cbranch_scratch): Likewise.
922 (*arm_cmpdi_insn): Likewise.
923 (*arm_cmpdi_unsig): Likewise.
924 (*arm_cmpdi_zero): Likewise.
925 (*thumb_cmpdi_zero): Likewise.
926 (*deleted_compare): Likewise.
927 (*mov_scc): Likewise.
928 (*mov_negscc): Likewise.
929 (*mov_notscc): Likewise.
930 (*cstoresi_eq0_thumb1_insn): Likewise.
931 (cstoresi_nltu_thumb1): Likewise.
932 (cstoresi_ltu_thu): Likewise.
933 (thumb1_addsi3_addgeu): Likewise.
934 (*arm_jump): Likewise.
935 (*thumb_jump): Likewise.
936 (*check_arch2): Likewise.
937 (arm_casesi_internal): Likewise.
938 (thumb1_casesi_dispatch): Likewise.
939 (*arm_indirect_jump): Likewise.
940 (*thumb1_indirect_jump): Likewise.
942 (*and_scc): Likewise.
943 (*ior_scc): Likewise.
944 (*compare_scc): Likewise.
945 (*cond_move): Likewise.
946 (*cond_arith): Likewise.
947 (*cond_sub): Likewise.
948 (*cmp_ite0): Likewise.
949 (*cmp_ite1): Likewise.
950 (*cmp_and): Likewise.
951 (*cmp_ior): Likewise.
952 (*ior_scc_scc): Likewise.
953 (*ior_scc_scc_cmp): Likewise.
954 (*and_scc_scc): Likewise.
955 (*and_scc_scc_cmp): Likewise.
956 (*and_scc_scc_nod): Likewise.
958 (movcond_addsi): Likewise.
960 (*ifcompare_plus_move): Likewise.
961 (*if_plus_move): Likewise.
962 (*ifcompare_move_plus): Likewise.
963 (*if_move_plus): Likewise.
964 (*ifcompare_arith_arith): Likewise.
965 (*if_arith_arith): Likewise.
966 (*ifcompare_arith_move): Likewise.
967 (*if_arith_move): Likewise.
968 (*ifcompare_move_arith): Likewise.
969 (*if_move_arith): Likewise.
970 (*ifcompare_move_not): Likewise.
971 (*if_move_not): Likewise.
972 (*ifcompare_not_move): Likewise.
973 (*if_not_move): Likewise.
974 (*ifcompare_shift_move): Likewise.
975 (*if_shift_move): Likewise.
976 (*ifcompare_move_shift): Likewise.
977 (*if_move_shift): Likewise.
978 (*ifcompare_shift_shift): Likewise.
979 (*ifcompare_not_arith): Likewise.
980 (*ifcompare_arith_not): Likewise.
981 (*if_arith_not): Likewise.
982 (*ifcompare_neg_move): Likewise.
983 (*if_neg_move): Likewise.
984 (*ifcompare_move_neg): Likewise.
985 (*if_move_neg): Likewise.
986 (prologue_thumb1_interwork): Likewise.
987 (*cond_move_not): Likewise.
988 (*sign_extract_onebit): Likewise.
989 (*not_signextract_onebit): Likewise.
990 (stack_tie): Likewise.
993 (consttable_end): Likewise.
994 (consttable_1): Likewise.
995 (consttable_2): Likewise.
996 (consttable_4): Likewise.
997 (consttable_8): Likewise.
998 (consttable_16): Likewise.
999 (*thumb1_tablejump): Likewise.
1000 (prefetch): Likewise.
1001 (force_register_use): Likewise.
1002 (thumb_eh_return): Likewise.
1003 (load_tp_hard): Likewise.
1004 (load_tp_soft): Likewise.
1005 (tlscall): Likewise.
1006 (*arm_movtas_ze): Likewise.
1007 (*arm_rev): Likewise.
1008 (*arm_revsh): Likewise.
1009 (*arm_rev16): Likewise.
1010 * config/arm/thumb2.md
1011 (*thumb2_smaxsi3): Likewise.
1012 (*thumb2_sminsi3): Likewise.
1013 (*thumb32_umaxsi3): Likewise.
1014 (*thumb2_uminsi3): Likewise.
1015 (*thumb2_negdi2): Likewise.
1016 (*thumb2_abssi2): Likewise.
1017 (*thumb2_neg_abss): Likewise.
1018 (*thumb2_movsi_insn): Likewise.
1019 (tls_load_dot_plus_four): Likewise.
1020 (*thumb2_movhi_insn): Likewise.
1021 (*thumb2_mov_scc): Likewise.
1022 (*thumb2_mov_negs): Likewise.
1023 (*thumb2_mov_negs): Likewise.
1024 (*thumb2_mov_nots): Likewise.
1025 (*thumb2_mov_nots): Likewise.
1026 (*thumb2_movsicc_): Likewise.
1027 (*thumb2_movsfcc_soft_insn): Likewise.
1028 (*thumb2_indirect_jump): Likewise.
1029 (*thumb2_and_scc): Likewise.
1030 (*thumb2_ior_scc): Likewise.
1031 (*thumb2_ior_scc_strict_it): Likewise.
1032 (*thumb2_cond_move): Likewise.
1033 (*thumb2_cond_arith): Likewise.
1034 (*thumb2_cond_ari): Likewise.
1035 (*thumb2_cond_sub): Likewise.
1036 (*thumb2_negscc): Likewise.
1037 (*thumb2_movcond): Likewise.
1038 (thumb2_casesi_internal): Likewise.
1039 (thumb2_casesi_internal_pic): Likewise.
1040 (*thumb2_alusi3_short): Likewise.
1041 (*thumb2_mov<mode>_shortim): Likewise.
1042 (*thumb2_addsi_short): Likewise.
1043 (*thumb2_subsi_short): Likewise.
1044 (thumb2_addsi3_compare0): Likewise.
1045 (*thumb2_cbz): Likewise.
1046 (*thumb2_cbnz): Likewise.
1047 (*thumb2_one_cmplsi2_short): Likewise.
1048 (*thumb2_negsi2_short): Likewise.
1049 (*orsi_notsi_si): Likewise.
1050 * config/arm/arm1020e.md: Update with new attributes.
1051 * config/arm/arm1026ejs.md: Update with new attributes.
1052 * config/arm/arm1136jfs.md: Update with new attributes.
1053 * config/arm/arm926ejs.md: Update with new attributes.
1054 * config/arm/cortex-a15.md: Update with new attributes.
1055 * config/arm/cortex-a5.md: Update with new attributes.
1056 * config/arm/cortex-a53.md: Update with new attributes.
1057 * config/arm/cortex-a7.md: Update with new attributes.
1058 * config/arm/cortex-a8.md: Update with new attributes.
1059 * config/arm/cortex-a9.md: Update with new attributes.
1060 * config/arm/cortex-m4.md: Update with new attributes.
1061 * config/arm/cortex-r4.md: Update with new attributes.
1062 * config/arm/fa526.md: Update with new attributes.
1063 * config/arm/fa606te.md: Update with new attributes.
1064 * config/arm/fa626te.md: Update with new attributes.
1065 * config/arm/fa726te.md: Update with new attributes.
1067 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1069 * config/aarch64/aarch64-simd.md
1070 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
1071 <vwx> iterator to ensure correct register choice.
1072 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1073 (aarch64_sqdmull_n<mode>): Likewise.
1074 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1075 * config/aarch64/arm_neon.h
1076 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
1077 (vml<as><q>_n_<su>16): Likewise.
1078 (vml<as>l_high_lane<q>_<su>16): Likewise.
1079 (vml<as>l_high_n_<su>16): Likewise.
1080 (vml<as>l_lane<q>_<su>16): Likewise.
1081 (vml<as>l_n_<su>16): Likewise.
1082 (vmul<q>_lane<q>_<su>16): Likewise.
1083 (vmul<q>_n_<su>16): Likewise.
1084 (vmull_lane<q>_<su>16): Likewise.
1085 (vmull_n_<su>16): Likewise.
1086 (vmull_high_lane<q>_<su>16): Likewise.
1087 (vmull_high_n_<su>16): Likewise.
1088 (vqrdmulh<q>_n_s16): Likewise.
1090 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
1092 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
1093 have the correct lane parameter.
1095 2013-09-06 Richard Biener <rguenther@suse.de>
1097 * cfganal.c (control_dependences::~control_dependences):
1098 Properly free all of the vector.
1100 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
1103 * config/i386/i386.c (ix86_conditional_register_usage):
1104 Proper initialize extended SSE registers.
1106 2013-09-06 Jan Hubicka <jh@suse.cz>
1108 PR tree-optimization/58311
1109 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
1111 2013-09-06 Jan Hubicka <jh@suse.cz>
1113 * Makefile.in (tree-sra.o): Update dependencies.
1114 * tree-sra.c: Include ipa-utils.h
1115 (scan_function): Use recursive_call_p.
1116 (has_caller_p): New function.
1117 (cgraph_for_node_and_aliases): Count also callers of aliases.
1119 2013-09-06 Jan Hubicka <jh@suse.cz>
1122 * cgraph.h (symtab_semantically_equivalent_p): Declare.
1123 * tree-tailcall.c: Include ipa-utils.h.
1124 (find_tail_calls): Use it.
1125 * ipa-pure-const.c (check_call): Likewise.
1126 * ipa-utils.c (recursive_call_p): New function.
1127 * ipa-utils.h (recursive_call_p): Dclare.
1128 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
1129 (symtab_semantically_equivalent_p): New function.
1130 * Makefile.in (tree-tailcall.o): Update dependencies.
1132 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1134 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
1137 2013-09-06 Richard Biener <rguenther@suse.de>
1139 * lto-streamer.h (lto_global_var_decls): Remove.
1140 * Makefile.in (OBJS): Remove lto-symtab.o.
1141 (lto-symtab.o): Remove.
1142 (GTFILES): Remove lto-symtab.c
1143 * lto-symtab.c: Move to lto/
1145 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1147 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
1148 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
1149 (UNSPEC_FPINT_RINT): New constant definitions.
1150 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
1151 definition with 2 attributes.
1152 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
1153 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
1156 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1158 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
1159 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
1160 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
1161 attribute to "z196_alone".
1162 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
1165 2013-09-06 Richard Biener <rguenther@suse.de>
1167 * basic-block.h (class control_dependences): New.
1168 * tree-ssa-dce.c (control_dependence_map): Remove.
1170 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
1171 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
1172 find_pdom, find_control_dependence, find_all_control_dependences):
1174 (mark_control_dependent_edges_necessary,
1175 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
1176 tree_dce_done, perform_tree_ssa_dce): Adjust.
1177 * cfganal.c (set_control_dependence_map_bit,
1178 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
1179 find_all_control_dependences): Move from tree-ssa-dce.c and
1180 implement as methods of control_dependences class.
1181 (control_dependences::control_dependences): New.
1182 (control_dependences::~control_dependences): Likewise.
1183 (control_dependences::get_edges_dependent_on): Likewise.
1184 (control_dependences::get_edge): Likewise.
1186 2013-09-04 Jan Hubicka <jh@suse.cz>
1188 * tree.c (types_same_for_odr): Drop overactive check.
1189 * ipa-devirt.c (hash_type_name): Likewise.
1191 2013-09-04 Jan Hubicka <jh@suse.cz>
1193 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
1194 (analyze_functions): ... here.
1196 2013-09-04 Jan Hubicka <jh@suse.cz>
1199 * cgraphunit.c (analyze_functions): Clear AUX fields
1200 after processing; initialize assembler name has.
1202 2013-09-05 Jeff Law <law@redhat.com>
1204 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
1205 from thread_around_empty_block. Record threading path into PATH.
1206 Recurse if threading through the initial block is successful.
1207 (thread_across_edge): Corresponding changes to slightly simplify.
1209 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1211 * config/aarch64/aarch64.md
1212 (type): Remove frecpe, frecps, frecpx.
1213 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
1214 fix to be a TARGET_SIMD instruction.
1215 (aarch64_frecps): Remove.
1216 * config/aarch64/aarch64-simd.md
1217 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
1218 (aarch64_frecps<mode>): Handle all float/vector of float modes.
1220 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1221 Sofiane Naci <sofiane.naci@arm.com>
1223 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
1224 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
1225 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
1226 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
1227 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
1228 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
1229 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
1230 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
1231 into "clz, "rbit". Rename "shift" to "shift_imm".
1232 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
1233 changes. Update for attribute changes all occurrences of arlo_* and
1235 * config/arm/arm-fixed.md: Update for attribute changes
1236 all occurrences of arlo_* types.
1237 * config/arm/thumb2.md: Update for attribute changes all occurrences
1239 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
1240 (cortexa7_older_only): Likewise.
1241 (cortexa7_younger): Likewise.
1242 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
1243 (1020alu_shift_op): Likewise.
1244 (1020alu_shift_reg_op): Likewise.
1245 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
1246 (alu_shift_op): Likewise.
1247 (alu_shift_reg_op): Likewise.
1248 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
1249 (11_alu_shift_op): Likewise.
1250 (11_alu_shift_reg_op): Likewise.
1251 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
1252 (9_alu_shift_reg_op): Likewise.
1253 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
1255 (cortex_a15_alu_shift): Likewise.
1256 (cortex_a15_alu_shift_reg): Likewise.
1257 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
1259 (cortex_a5_alu_shift): Likewise.
1260 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
1262 (cortex_a53_alu_shift): Likewise.
1263 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
1265 (cortex_a7_alu_reg): Likewise.
1266 (cortex_a7_alu_shift): Likewise.
1267 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
1269 (cortex_a8_alu_shift): Likewise.
1270 (cortex_a8_alu_shift_reg): Likewise.
1271 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
1272 (cortex_a9_dp_shift): Likewise.
1273 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
1275 * config/arm/cortex-r4.md
1276 (cortex_r4_alu): Update for attribute changes.
1277 (cortex_r4_mov): Likewise.
1278 (cortex_r4_alu_shift_reg): Likewise.
1279 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
1280 (526_alu_shift_op): Likewise.
1281 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
1282 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
1283 (626te_alu_shift_op): Likewise.
1284 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
1285 (726te_alu_shift_op): Likewise.
1286 (726te_alu_shift_reg_op): Likewise.
1287 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
1288 (mp626_alu_shift_op): Likewise.
1289 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
1290 (pj4_alu_conds): Likewise.
1291 (pj4_shift): Likewise.
1292 (pj4_shift_conds): Likewise.
1293 (pj4_alu_shift): Likewise.
1294 (pj4_alu_shift_conds): Likewise.
1295 * config/aarch64/aarch64.md: Update for attribute change
1296 all occurrences of arlo_* and shift* types.
1298 2013-09-05 Mike Stump <mikestump@comcast.net>
1300 * tree.h: Move documentation for tree_function_decl to tree-core.h
1301 with the declaration.
1303 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
1306 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
1307 looking for widest mode.
1309 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
1311 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
1313 2013-09-05 Richard Biener <rguenther@suse.de>
1315 PR tree-optimization/58137
1316 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
1317 Do not create vectors of pointers.
1318 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
1319 types for the components of the vector initializer.
1320 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
1321 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
1323 2013-09-05 Martin Jambor <mjambor@suse.cz>
1325 * ipa-prop.c (remove_described_reference): Accept missing references,
1326 return false if that hppens, otherwise return true.
1327 (cgraph_node_for_jfunc): New function.
1328 (try_decrement_rdesc_refcount): Likewise.
1329 (try_make_edge_direct_simple_call): Use them.
1330 (ipa_edge_removal_hook): Remove references from rdescs.
1331 (ipa_edge_duplication_hook): Clone rdescs and their references
1332 when the new edge has the same caller as the old one.
1333 * cgraph.c (cgraph_resolve_speculation): Remove speculative
1334 reference before removing any edges.
1336 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
1338 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
1340 * thumb2.md (thumb2_storewb_parisi): New pattern.
1342 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
1344 * config/aarch64/aarch64-option-extensions.def: Add
1345 AARCH64_OPT_EXTENSION of 'crc'.
1346 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
1347 (AARCH64_ISA_CRC): Ditto.
1348 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
1349 description of the CRC extension.
1351 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
1353 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
1354 * config/rs6000/linux.h: Ditto.
1355 * alpha/linux.h: Ditto.
1356 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
1357 no_c99_libc_has_function.
1358 * config/c6x/uclinux-elf.h: Ditto.
1359 * config/lm32/uclinux-elf.h: Ditto.
1360 * config/m68k/uclinux.h: Ditto.
1361 * config/moxie/uclinux.h: Ditto.
1362 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
1363 (crisv32-*-linux*, cris-*-linux*): Ditto.
1364 * config/bfin/bfin.c: Include "tm_p.h".
1366 2013-09-05 Richard Biener <rguenther@suse.de>
1368 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
1369 check for a definition without a basic-block.
1371 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1372 Sofiane Naci <sofiane.naci@arm.com>
1374 * config/aarch64/aarch64.md
1375 (*movti_aarch64): Rename r_2_f and f_2_r.
1376 (*movsf_aarch64): Likewise.
1377 (*movdf_aarch64): Likewise.
1378 (*movtf_aarch64): Likewise.
1379 (aarch64_movdi_<mode>low): Likewise.
1380 (aarch64_movdi_<mode>high): Likewise.
1381 (aarch64_mov<mode>high_di): Likewise.
1382 (aarch64_mov<mode>low_di): Likewise.
1383 (aarch64_movtilow_tilow): Likewise.
1384 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
1385 values to config/arm/types.md
1386 (attribute "conds"): Update for attribute change.
1387 (anddi3_insn): Likewise.
1388 (iordi3_insn): Likewise.
1389 (xordi3_insn): Likewise.
1390 (one_cmpldi2): Likewise.
1391 * config/arm/types.md (type): Add Neon types.
1392 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
1393 use "type" attribute.
1394 (movmisalign<mode>_neon_store): Likewise.
1395 (movmisalign<mode>_neon_load): Likewise.
1396 (vec_set<mode>_internal): Likewise.
1397 (vec_setv2di_internal): Likewise.
1398 (vec_extract<mode>): Likewise.
1399 (vec_extractv2di): Likewise.
1400 (add<mode>3_neon): Likewise.
1401 (adddi3_neon): Likewise.
1402 (sub<mode>3_neon): Likewise.
1403 (subdi3_neon): Likewise.
1404 (mul<mode>3_neon): Likewise.
1405 (mul<mode>3add<mode>_neon): Likewise.
1406 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
1407 (fma<VCVTF:mode>4)): Likewise.
1408 (fma<VCVTF:mode>4_intrinsic): Likewise.
1409 (fmsub<VCVTF:mode>4)): Likewise.
1410 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1411 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1412 (ior<mode>3): Likewise.
1413 (and<mode>3): Likewise.
1414 (anddi3_neon): Likewise.
1415 (orn<mode>3_neon): Likewise.
1416 (orndi3_neon): Likewise.
1417 (bic<mode>3_neon): Likewise.
1418 (bicdi3_neon): Likewise.
1419 (xor<mode>3): Likewise.
1420 (one_cmpl<mode>2): Likewise.
1421 (abs<mode>2): Likewise.
1422 (neg<mode>2): Likewise.
1423 (umin<mode>3_neon): Likewise.
1424 (umax<mode>3_neon): Likewise.
1425 (smin<mode>3_neon): Likewise.
1426 (smax<mode>3_neon): Likewise.
1427 (vashl<mode>3): Likewise.
1428 (vashr<mode>3_imm): Likewise.
1429 (vlshr<mode>3_imm): Likewise.
1430 (ashl<mode>3_signed): Likewise.
1431 (ashl<mode>3_unsigned): Likewise.
1432 (neon_load_count): Likewise.
1433 (ashldi3_neon_noclobber): Likewise.
1434 (signed_shift_di3_neon): Likewise.
1435 (unsigned_shift_di3_neon): Likewise.
1436 (ashrdi3_neon_imm_noclobber): Likewise.
1437 (lshrdi3_neon_imm_noclobber): Likewise.
1438 (widen_ssum<mode>3): Likewise.
1439 (widen_usum<mode>3): Likewise.
1440 (quad_halves_<code>v4si): Likewise.
1441 (quad_halves_<code>v4sf): Likewise.
1442 (quad_halves_<code>v8hi): Likewise.
1443 (quad_halves_<code>v16qi): Likewise.
1444 (reduc_splus_v2di): Likewise.
1445 (neon_vpadd_internal<mode>): Likewise.
1446 (neon_vpsmin<mode>): Likewise.
1447 (neon_vpsmax<mode>): Likewise.
1448 (neon_vpumin<mode>): Likewise.
1449 (neon_vpumax<mode>): Likewise.
1450 (ss_add<mode>_neon): Likewise.
1451 (us_add<mode>_neon): Likewise.
1452 (ss_sub<mode>_neon): Likewise.
1453 (us_sub<mode>_neon): Likewise.
1454 (neon_vadd<mode>_unspec): Likewise.
1455 (neon_vaddl<mode>): Likewise.
1456 (neon_vaddw<mode>): Likewise.
1457 (neon_vhadd<mode>): Likewise.
1458 (neon_vqadd<mode>): Likewise.
1459 (neon_vaddhn<mode>): Likewise.
1460 (neon_vmul<mode>): Likewise.
1461 (neon_vmla<mode>): Likewise.
1462 (neon_vmlal<mode>): Likewise.
1463 (neon_vmls<mode>): Likewise.
1464 (neon_vmlsl<mode>): Likewise.
1465 (neon_vqdmulh<mode>): Likewise.
1466 (neon_vqdmlal<mode>): Likewise.
1467 (neon_vqdmlsl<mode>): Likewise.
1468 (neon_vmull<mode>): Likewise.
1469 (neon_vqdmull<mode>): Likewise.
1470 (neon_vsub<mode>_unspec): Likewise.
1471 (neon_vsubl<mode>): Likewise.
1472 (neon_vsubw<mode>): Likewise.
1473 (neon_vqsub<mode>): Likewise.
1474 (neon_vhsub<mode>): Likewise.
1475 (neon_vsubhn<mode>): Likewise.
1476 (neon_vceq<mode>): Likewise.
1477 (neon_vcge<mode>): Likewise.
1478 (neon_vcgeu<mode>): Likewise.
1479 (neon_vcgt<mode>): Likewise.
1480 (neon_vcgtu<mode>): Likewise.
1481 (neon_vcle<mode>): Likewise.
1482 (neon_vclt<mode>): Likewise.
1483 (neon_vcage<mode>): Likewise.
1484 (neon_vcagt<mode>): Likewise.
1485 (neon_vtst<mode>): Likewise.
1486 (neon_vabd<mode>): Likewise.
1487 (neon_vabdl<mode>): Likewise.
1488 (neon_vaba<mode>): Likewise.
1489 (neon_vabal<mode>): Likewise.
1490 (neon_vmax<mode>): Likewise.
1491 (neon_vmin<mode>): Likewise.
1492 (neon_vpaddl<mode>): Likewise.
1493 (neon_vpadal<mode>): Likewise.
1494 (neon_vpmax<mode>): Likewise.
1495 (neon_vpmin<mode>): Likewise.
1496 (neon_vrecps<mode>): Likewise.
1497 (neon_vrsqrts<mode>): Likewise.
1498 (neon_vqabs<mode>): Likewise.
1499 (neon_vqneg<mode>): Likewise.
1500 (neon_vcls<mode>): Likewise.
1501 (clz<mode>2): Likewise.
1502 (popcount<mode>2): Likewise.
1503 (neon_vrecpe): Likewise.
1504 (neon_vrsqrte): Likewise.
1505 (neon_vget_lane<mode>_sext_internal): Likewise.
1506 (neon_vget_lane<mode>_zext_internal): Likewise.
1507 (neon_vdup_n<mode>): Likewise.
1508 (neon_vdup_nv2di): Likewise.
1509 (neon_vdpu_lane<mode>_internal): Likewise.
1510 (neon_vswp<mode>): Likewise.
1511 (float<mode><V_cvtto>2): Likewise.
1512 (floatuns<mode><V_cvtto>2): Likewise.
1513 (fix_trunc<mode><V_cvtto>)2): Likewise
1514 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1515 (neon_vcvt<mode>): Likewise.
1516 (neon_vcvtv4sfv4hf): Likewise.
1517 (neon_vcvtv4hfv4sf): Likewise.
1518 (neon_vcvt_n<mode>): Likewise.
1519 (neon_vmovn<mode>): Likewise.
1520 (neon_vqmovn<mode>): Likewise.
1521 (neon_vqmovun<mode>): Likewise.
1522 (neon_vmovl<mode>): Likewise.
1523 (neon_vmul_lane<mode>): Likewise.
1524 (neon_vmull_lane<mode>): Likewise.
1525 (neon_vqdmull_lane<mode>): Likewise.
1526 (neon_vqdmulh_lane<mode>): Likewise.
1527 (neon_vmla_lane<mode>): Likewise.
1528 (neon_vmlal_lane<mode>): Likewise.
1529 (neon_vqdmlal_lane<mode>): Likewise.
1530 (neon_vmls_lane<mode>): Likewise.
1531 (neon_vmlsl_lane<mode>): Likewise.
1532 (neon_vqdmlsl_lane<mode>): Likewise.
1533 (neon_vext<mode>): Likewise.
1534 (neon_vrev64<mode>): Likewise.
1535 (neon_vrev32<mode>): Likewise.
1536 (neon_vrev16<mode>): Likewise.
1537 (neon_vbsl<mode>_internal): Likewise.
1538 (neon_vshl<mode>): Likewise.
1539 (neon_vqshl<mode>): Likewise.
1540 (neon_vshr_n<mode>): Likewise.
1541 (neon_vshrn_n<mode>): Likewise.
1542 (neon_vqshrn_n<mode>): Likewise.
1543 (neon_vqshrun_n<mode>): Likewise.
1544 (neon_vshl_n<mode>): Likewise.
1545 (neon_vqshl_n<mode>): Likewise.
1546 (neon_vqshlu_n<mode>): Likewise.
1547 (neon_vshll_n<mode>): Likewise.
1548 (neon_vsra_n<mode>): Likewise.
1549 (neon_vsri_n<mode>): Likewise.
1550 (neon_vsli_n<mode>): Likewise.
1551 (neon_vtbl1v8qi): Likewise.
1552 (neon_vtbl2v8qi): Likewise.
1553 (neon_vtbl3v8qi): Likewise.
1554 (neon_vtbl4v8qi): Likewise.
1555 (neon_vtbx1v8qi): Likewise.
1556 (neon_vtbx2v8qi): Likewise.
1557 (neon_vtbx3v8qi): Likewise.
1558 (neon_vtbx4v8qi): Likewise.
1559 (neon_vtrn<mode>_internal): Likewise.
1560 (neon_vzip<mode>_internal): Likewise.
1561 (neon_vuzp<mode>_internal): Likewise.
1562 (neon_vld1<mode>): Likewise.
1563 (neon_vld1_lane<mode>): Likewise.
1564 (neon_vld1_dup<mode>): Likewise.
1565 (neon_vld1_dupv2di): Likewise.
1566 (neon_vst1<mode>): Likewise.
1567 (neon_vst1_lane<mode>): Likewise.
1568 (neon_vld2<mode>): Likewise.
1569 (neon_vld2_lane<mode>): Likewise.
1570 (neon_vld2_dup<mode>): Likewise.
1571 (neon_vst2<mode>): Likewise.
1572 (neon_vst2_lane<mode>): Likewise.
1573 (neon_vld3<mode>): Likewise.
1574 (neon_vld3qa<mode>): Likewise.
1575 (neon_vld3qb<mode>): Likewise.
1576 (neon_vld3_lane<mode>): Likewise.
1577 (neon_vld3_dup<mode>): Likewise.
1578 (neon_vst3<mode>): Likewise.
1579 (neon_vst3qa<mode>): Likewise.
1580 (neon_vst3qb<mode>): Likewise.
1581 (neon_vst3_lane<mode>): Likewise.
1582 (neon_vld4<mode>): Likewise.
1583 (neon_vld4qa<mode>): Likewise.
1584 (neon_vld4qb<mode>): Likewise.
1585 (neon_vld4_lane<mode>): Likewise.
1586 (neon_vld4_dup<mode>): Likewise.
1587 (neon_vst4<mode>): Likewise.
1588 (neon_vst4qa<mode>): Likewise.
1589 (neon_vst4qb<mode>): Likewise.
1590 (neon_vst4_lane<mode>): Likewise.
1591 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1592 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1593 (neon_vec_<US>mult_lo_<mode>): Likewise.
1594 (neon_vec_<US>mult_hi_<mode>): Likewise.
1595 (neon_vec_<US>shiftl_<mode>): Likewise.
1596 (neon_unpack<US>_<mode>): Likewise.
1597 (neon_vec_<US>mult_<mode>): Likewise.
1598 (vec_pack_trunc_<mode>): Likewise.
1599 (neon_vec_pack_trunk_<mode>): Likewise.
1600 (neon_vabd<mode>_2): Likewise.
1601 (neon_vabd<mode>_3): Likewise.
1602 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1603 (thumb2_movsi_vfp): Likewise.
1604 (movdi_vfp): Likewise.
1605 (movdi_vfp_cortexa8): Likewise.
1606 (movhf_vfp_neon): Likewise.
1607 (movhf_vfp): Likewiwse.
1608 (movsf_vfp): Likewiwse.
1609 (thumb2_movsf_vfp): Likewiwse.
1610 (movdf_vfp): Likewise.
1611 (thumb2_movdf_vfp): Likewise.
1612 (movsfcc_vfp): Likewise.
1613 (thumb2_movsfcc_vfp): Likewise.
1614 (movdfcc_vfp): Likewise.
1615 (thumb2_movdfcc_vfp): Likewise.
1616 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1617 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1618 (v10_v2c): Likewise.
1619 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1621 (cortex_a15_neon_int_2): Likewise.
1622 (cortex_a15_neon_int_3): Likewise.
1623 (cortex_a15_neon_int_4): Likewise.
1624 (cortex_a15_neon_int_5): Likewise.
1625 (cortex_a15_neon_vqneg_vqabs): Likewise.
1626 (cortex_a15_neon_vmov): Likewise.
1627 (cortex_a15_neon_vaba): Likewise.
1628 (cortex_a15_neon_vaba_qqq): Likewise.
1629 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1630 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1631 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1632 scalar_64_32_long_scalar): Likewise.
1633 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1634 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1635 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1636 lotype_qdd_64_32_long): Likewise.
1637 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1638 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1639 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1640 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1641 (cortex_a15_neon_shift_1): Likewise.
1642 (cortex_a15_neon_shift_2): Likewise.
1643 (cortex_a15_neon_shift_3): Likewise.
1644 (cortex_a15_neon_vshl_ddd): Likewise.
1645 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1646 (cortex_a15_neon_vsra_vrsra): Likewise.
1647 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1648 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1649 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1650 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1651 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1652 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1653 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1654 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1655 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1656 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1657 (cortex_a15_neon_bp_simple): Likewise.
1658 (cortex_a15_neon_bp_2cycle): Likewise.
1659 (cortex_a15_neon_bp_3cycle): Likewise.
1660 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1661 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1662 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1663 (cortex_a15_neon_vld2_4_regs): Likewise.
1664 (cortex_a15_neon_vld3_vld4): Likewise.
1665 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1666 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1667 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1668 (cortex_a15_neon_vst3_vst4): Likewise.
1669 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1670 (cortex_a15_neon_vld3_vld4_lane" 10
1671 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1672 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1673 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1674 (cortex_a15_neon_ldm_2): Likewise.0
1675 (cortex_a15_neon_stm_2): Likewise.
1676 (cortex_a15_neon_mcr): Likewise.
1677 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1678 (cortex_a15_neon_mrc): Likewise.
1679 (cortex_a15_neon_mrrc): Likewise.
1680 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1682 (cortex_a15_alu_shift): Likewise.
1683 (cortex_a15_alu_shift_reg): Likewise.
1684 (cortex_a15_mult32): Likewise.
1685 (cortex_a15_mult64): Likewise.
1686 (cortex_a15_block): Likewise.
1687 (cortex_a15_branch): Likewise.
1688 (cortex_a15_load1): Likewise.
1689 (cortex_a15_load3): Likewise.
1690 (cortex_a15_store1): Likewise.
1691 (cortex_a15_store3): Likewise.
1692 (cortex_a15_call): Likewise.
1693 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
1694 (cortex_a5_f2r): Likewise.
1695 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1697 (cortex_a53_f2r): Likewise.
1698 * config/arm/cortex-a7.md
1699 (cortex_a7_branch): Update for attribute change.
1700 (cortex_a7_call): Likewise.
1701 (cortex_a7_alu_imm): Likewise.
1702 (cortex_a7_alu_reg): Likewise.
1703 (cortex_a7_alu_shift): Likewise.
1704 (cortex_a7_mul): Likewise.
1705 (cortex_a7_load1): Likewise.
1706 (cortex_a7_store1): Likewise.
1707 (cortex_a7_load2): Likewise.
1708 (cortex_a7_store2): Likewise.
1709 (cortex_a7_load3): Likewise.
1710 (cortex_a7_store3): Likewise.
1711 (cortex_a7_load4): Likewise.
1712 (cortex_a7_store4): Likewise.
1713 (cortex_a7_fpalu): Likewise.
1714 (cortex_a7_fconst): Likewise.
1715 (cortex_a7_fpmuls): Likewise.
1716 (cortex_a7_neon_mul): Likewise.
1717 (cortex_a7_fpmacs): Likewise.
1718 (cortex_a7_neon_mla: Likewise.
1719 (cortex_a7_fpmuld: Likewise.
1720 (cortex_a7_fpmacd: Likewise.
1721 (cortex_a7_fpfmad: Likewise.
1722 (cortex_a7_fdivs: Likewise.
1723 (cortex_a7_fdivd: Likewise.
1724 (cortex_a7_r2f: Likewise.
1725 (cortex_a7_f2r: Likewise.
1726 (cortex_a7_f_flags: Likewise.
1727 (cortex_a7_f_loads: Likewise.
1728 (cortex_a7_f_loadd: Likewise.
1729 (cortex_a7_f_stores: Likewise.
1730 (cortex_a7_f_stored: Likewise.
1731 (cortex_a7_neon): Likewise.
1732 * config/arm/cortex-a8-neon.md
1733 (cortex_a8_neon_mrc): Update for attribute change.
1734 (cortex_a8_neon_mrrc): Likewise.
1735 (cortex_a8_neon_int_1): Likewise.
1736 (cortex_a8_neon_int_2): Likewise.
1737 (cortex_a8_neon_int_3): Likewise.
1738 (cortex_a8_neon_int_4): Likewise.
1739 (cortex_a8_neon_int_5): Likewise.
1740 (cortex_a8_neon_vqneg_vqabs): Likewise.
1741 (cortex_a8_neon_vmov): Likewise.
1742 (cortex_a8_neon_vaba): Likewise.
1743 (cortex_a8_neon_vaba_qqq): Likewise.
1744 (cortex_a8_neon_vsma): Likewise.
1745 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1746 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1747 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1748 long_scalar): Likewise.
1749 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1750 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1751 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1752 long_scalar_qdd_64_32_long): Likewise.
1753 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1754 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1755 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1756 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1757 (cortex_a8_neon_shift_1): Likewise.
1758 (cortex_a8_neon_shift_2): Likewise.
1759 (cortex_a8_neon_shift_3): Likewise.
1760 (cortex_a8_neon_vshl_ddd): Likewise.
1761 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1762 (cortex_a8_neon_vsra_vrsra): Likewise.
1763 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1764 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1765 (cortex_a8_neon_fp_vsum): Likewise.
1766 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1767 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1768 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1769 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1770 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1771 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1772 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1773 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1774 (cortex_a8_neon_bp_simple): Likewise.
1775 (cortex_a8_neon_bp_2cycle): Likewise.
1776 (cortex_a8_neon_bp_3cycle): Likewise.
1777 (cortex_a8_neon_ldr): Likewise.
1778 (cortex_a8_neon_str): Likewise.
1779 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1780 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1781 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1782 (cortex_a8_neon_vld2_4_regs): Likewise.
1783 (cortex_a8_neon_vld3_vld4): Likewise.
1784 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1785 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1786 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1787 (cortex_a8_neon_vst3_vst4): Likewise.
1788 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1789 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1790 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1791 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1792 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1793 (cortex_a8_neon_mcr): Likewise.
1794 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1795 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
1796 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
1798 (ca9_neon_mrrc): Likewise.
1799 (cortex_a9_neon_int_1): Likewise.
1800 (cortex_a9_neon_int_2): Likewise.
1801 (cortex_a9_neon_int_3): Likewise.
1802 (cortex_a9_neon_int_4): Likewise.
1803 (cortex_a9_neon_int_5): Likewise.
1804 (cortex_a9_neon_vqneg_vqabs): Likewise.
1805 (cortex_a9_neon_vmov): Likewise.
1806 (cortex_a9_neon_vaba): Likewise.
1807 (cortex_a9_neon_vaba_qqq): Likewise.
1808 (cortex_a9_neon_vsma): Likewise.
1809 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1810 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1811 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1812 long_scalar): Likewise.
1813 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1814 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1815 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1816 long_scalar_qdd_64_32_long): Likewise.
1817 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1818 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1819 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1820 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1821 (cortex_a9_neon_shift_1): Likewise.
1822 (cortex_a9_neon_shift_2): Likewise.
1823 (cortex_a9_neon_shift_3): Likewise.
1824 (cortex_a9_neon_vshl_ddd): Likewise.
1825 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1826 (cortex_a9_neon_vsra_vrsra): Likewise.
1827 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1828 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1829 (cortex_a9_neon_fp_vsum): Likewise.
1830 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1831 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1832 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1833 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1834 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1835 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1836 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1837 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1838 (cortex_a9_neon_bp_simple): Likewise.
1839 (cortex_a9_neon_bp_2cycle): Likewise.
1840 (cortex_a9_neon_bp_3cycle): Likewise.
1841 (cortex_a9_neon_ldr): Likewise.
1842 (cortex_a9_neon_str): Likewise.
1843 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1844 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1845 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1846 (cortex_a9_neon_vld2_4_regs): Likewise.
1847 (cortex_a9_neon_vld3_vld4): Likewise.
1848 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1849 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1850 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1851 (cortex_a9_neon_vst3_vst4): Likewise.
1852 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1853 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1854 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1855 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1856 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1857 (cortex_a9_neon_mcr): Likewise.
1858 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1859 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
1860 (cortex_a9_fps): Likewise.
1861 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
1863 (cortex_m4_fmuls): Likewise.
1864 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
1866 (cortex_r4_mrc): Likewise.
1867 * config/arm/iterators.md: Update comment referring to neon_type.
1868 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
1869 (iwmmxt_movsi_insn): Likewise.
1870 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
1872 (pj4_core_to_vfp): Likewise.
1873 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
1875 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
1876 (vfp_fstore): Likewise.
1877 * doc/md.texi: Change references to neon_type to refer to type.
1879 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1881 * tree.h (DECL_BUILT_IN): Fix typo in comment.
1883 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
1885 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
1888 2013-09-04 Easwaran Raman <eraman@google.com>
1891 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
1892 (build_and_add_sum): Use it.
1893 (appears_later_in_bb): Simplify code.
1895 2013-09-04 Teresa Johnson <tejohnson@google.com>
1897 * dumpfile.c (dump_finish): Don't close stderr/stdout.
1899 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
1901 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
1903 2013-09-04 Jan Hubicka <jh@suse.cz>
1905 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
1906 * ipa-devirt.c: Include diganostic.h
1907 (odr_type_d): Add types and types_set.
1908 (hash_type_name): Work for types with vtables during LTO.
1909 (odr_hasher::remove): Fix comment; destroy types_set.
1910 (add_type_duplicate): New function,
1911 (get_odr_type): Use it.
1912 (dump_type_inheritance_graph): Dump type duplicates.
1913 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
1915 * tree.c (types_same_for_odr): Give exact answers on types with
1918 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1920 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
1921 explaining their differences.
1923 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
1925 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
1927 2013-09-03 Jeff Law <law@redhat.com>
1929 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
1930 when not threading through a joiner block. Pass joiner/no joiner
1931 state to register_jump_thread.
1932 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
1933 state from argument rather than implying on path length.
1934 Dump the entire jump thread path into debugging dump.
1935 * tree-flow.h (register_jump_thread): Update prototype.
1937 2013-08-29 Xinliang David Li <davidxl@google.com>
1939 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1940 Remove a trivial gcc_assert.
1942 2013-08-29 Xinliang David Li <davidxl@google.com>
1944 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
1945 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
1946 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1947 Delay base decl alignment adjustment.
1948 * tree-vectorizer.c (vect_destroy_datarefs): New function.
1949 * tree-vectorizer.h: New data structure.
1950 (set_dr_misalignment): New function.
1951 (dr_misalignment): Ditto.
1952 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
1953 (vectorizable_load): Ditto.
1954 (ensure_base_align): New function.
1955 (vectorize_loops): Add dbg_cnt support.
1956 (execute_vect_slp): Ditto.
1957 * dbgcnt.def: New debug counter.
1958 * Makefile: New dependency.
1960 2013-09-03 Meador Inge <meadori@codesourcery.com>
1964 2013-08-30 Meador Inge <meadori@codesourcery.com>
1966 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1968 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
1970 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
1971 function descriptor.
1973 2013-09-03 Richard Biener <rguenther@suse.de>
1975 * tree-affine.c (add_elt_to_tree): Fix association issue,
1976 avoid useless converts and make sure to always return a
1977 properly typed result.
1979 2013-09-03 Richard Biener <rguenther@suse.de>
1982 * fold-const.c (negate_expr_p): Fix division case.
1983 (negate_expr): Likewise.
1985 2013-09-03 Richard Biener <rguenther@suse.de>
1988 * tree-streamer-out.c: Include tm.h.
1989 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
1991 2013-09-03 Jan Hubicka <jh@suse.cz>
1993 * tree-profile.c (tree_profiling): Cleanup CFG when done.
1995 2013-09-03 Alan Modra <amodra@gmail.com>
1997 * config.gcc (powerpc*-*-linux*): Add support for little-endian
1998 multilibs to big-endian target and vice versa.
1999 * config/rs6000/t-linux64: Use := assignment on all vars.
2000 (MULTILIB_EXTRA_OPTS): Remove fPIC.
2001 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
2002 * config/rs6000/t-linux64le: New file.
2003 * config/rs6000/t-linux64bele: New file.
2004 * config/rs6000/t-linux64lebe: New file.
2006 2013-09-02 Jan Hubicka <jh@suse.cz>
2008 * ipa-inline-transform.c (inline_transform): Do not
2009 optimize_inline_calls when not optimizing.
2011 2013-09-02 Jan Hubicka <jh@suse.cz>
2013 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
2014 duplicated nodes for assembler names.
2015 * symtab.c (symtab_unregister_node): Do not attempt to unlink
2016 hard registers from assembler name hash.
2018 2013-09-02 Jan Hubicka <jh@suse.cz>
2020 * ipa-split.c (execute_split_functions): Split externally visible
2021 functions called once.
2023 2013-09-02 Martin Jambor <mjambor@suse.cz>
2026 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
2027 linked list. When finding the correct duplicate, also consider also
2028 the caller in additon to its inlined_to node.
2030 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
2032 * config/aarch64/aarch64-simd-builtins.def
2033 (dup_lane_scalar): Remove.
2034 * config/aarch64/aarch64-simd.md
2035 (aarch64_simd_dup): Add 'w->w' alternative.
2036 (aarch64_dup_lane<mode>): Allow for VALL.
2037 (aarch64_dup_lane_scalar<mode>): Remove.
2038 (aarch64_dup_lane_<vswap_width_name><mode>): New.
2039 (aarch64_get_lane_signed<mode>): Add w->w altenative.
2040 (aarch64_get_lane_unsigned<mode>): Likewise.
2041 (aarch64_get_lane<mode>): Likewise.
2042 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
2043 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
2044 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
2045 (VCON): Change container of V2SF.
2046 (vswap_width_name): Likewise.
2047 * config/aarch64/arm_neon.h
2048 (__aarch64_vdup_lane_any): New.
2049 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2050 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
2051 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2053 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
2056 * expr.c (emit_move_complex): Do not move complex FP values as parts if
2057 the source or the destination is a single hard register.
2059 2013-09-02 Richard Biener <rguenther@suse.de>
2062 * tree-scalar-evolution.c (instantiate_scev_name): Allow
2065 2013-09-02 Richard Biener <rguenther@suse.de>
2067 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
2068 arithmetic to sizetype.
2070 2013-09-02 Bin Cheng <bin.cheng@arm.com>
2072 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
2073 Find auto-increment use both before and after candidate.
2075 2013-09-02 Marek Polacek <polacek@redhat.com>
2077 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
2079 2013-09-01 Jan Hubicka <jh@suse.cz>
2081 * Makefile.in: Add ipa-profile.o
2082 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
2083 * cgraph.c (struct cgraph_propagate_frequency_data,
2084 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
2085 ipa-profile.c; replace cgraph_ by ipa_ prefix.
2086 * cgraph.h (cgraph_propagate_frequency): Remove.
2087 * ipa-inline-analysis.c: Include ipa-utils.h;
2088 drop duplicated cfgloop.h.
2089 (inline_update_callee_summaries): Update.
2090 * ipa-profile.c: New file.
2091 * ipa-utils.h (ipa_propagate_frequency): Declare.
2092 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
2093 data-streamer.h, value-prof.h.
2094 (symtab_remove_unreachable_nodes): Update profile.
2095 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
2096 account_time_size, cmp_counts, dump_histogram,
2097 ipa_profile_generate_summary, ipa_profile_write_summary,
2098 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
2099 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
2100 Move to ipa-profile.c.
2102 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
2104 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
2106 2013-09-01 Jan Hubicka <jh@suse.cz>
2108 * common.opt (fdevirtualize-speculatively): New function.
2109 * invoke.texi (fdevirtualize-speculatively): Document.
2110 * ipa-devirt.c: Include ipa-inline.h
2111 (likely_target_p): New function.
2112 (ipa_devirt): New function.
2113 (gate_ipa_devirt): New function.
2114 (pass_data_ipa_devirt): New static var.
2115 (pass_ipa_devirt): Likewise.
2116 (make_pass_ipa_devirt): New function.
2117 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
2118 (common_handle_option): Disable devirtualization when
2119 value range profiling is available.
2120 * passes.def (pass_ipa_devirt): Add.
2121 * timever.def (TV_IPA_DEVIRT): New timevar.
2122 * tree-pass.h (make_pass_ipa_devirt):
2124 2013-09-01 Iain Sandoe <iain@codesourcery.com>
2126 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
2127 include sanitize(undefined).
2129 2013-08-31 Diego Novillo <dnovillo@google.com>
2131 * Makefile.in (TREE_CORE_H): Define.
2133 (GTFILES): Add tree-core.h.
2134 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
2136 * tree-core.h: New file.
2137 Move all data structures, enum, typedefs, global
2138 declarations and constants from ...
2141 2013-08-31 Jan Hubicka <jh@suse.cz>
2143 * bulitins.c (expand_builtin): Do not early exit for gcov
2144 instrumented functions.
2146 2013-08-31 Marek Polacek <polacek@redhat.com>
2148 * ubsan.c: Include tm_p.h.
2150 2013-08-31 Jan Hubicka <jh@suse.cz>
2152 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
2155 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
2156 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
2157 * tree-cfg.c (verify_gimple_label): ... here.
2158 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
2159 (ipa_merge_profiles): New function.
2160 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
2161 (lto_input_function_body): Likewise.
2162 * ipa-utils.h (ipa_merge_profiles): Declare.
2163 * lto-streamer.h (lto_input_function_body): Update prototype.
2164 (emit_label_in_global_context_p): Remove.
2165 * lto-symtab.c: Include ipa-utils.h
2166 (lto_cgraph_replace_node): Use ipa_merge_profiles.
2168 2013-08-31 Jan Hubicka <jh@suse.cz>
2170 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
2172 2013-08-31 Jan Hubicka <jh@suse.cz>
2174 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
2176 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
2178 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
2179 "cmp" RTX before signed_comparison_operator check to account
2182 2013-08-30 Jan Hubicka <jh@suse.cz>
2184 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
2185 (detect_type_change_1): Rename to ...
2186 (detect_type_change): ... this one; early return on non-polymorphic
2188 (detect_type_change_ssa): Add comp_type parameter; update
2189 use of detect_type_change.
2190 (compute_complex_assign_jump_func): Add param_type parameter;
2191 update use of detect_type_change_ssa.
2192 (compute_complex_ancestor_jump_func): Likewise.
2193 (ipa_get_callee_param_type): New function.
2194 (ipa_compute_jump_functions_for_edge): Compute parameter type;
2195 update calls to the jump function computation functions.
2197 2013-08-30 Teresa Johnson <tejohnson@google.com>
2198 Steven Bosscher <steven@gcc.gnu.org>
2200 * cfgrtl.c (fixup_new_cold_bb): New routine.
2201 (commit_edge_insertions): Invoke fixup_partitions.
2202 (find_partition_fixes): New routine.
2203 (fixup_partitions): Ditto.
2204 (verify_hot_cold_block_grouping): Update comments.
2205 (rtl_verify_edges): Invoke find_partition_fixes.
2206 (rtl_verify_bb_pointers): Update comments.
2207 (rtl_verify_bb_layout): Ditto.
2208 * basic-block.h (probably_never_executed_edge_p): Declare.
2209 (fixup_partitions): Ditto.
2210 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
2211 * bb-reorder.c (sanitize_hot_paths): New function.
2212 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
2214 * predict.c (probably_never_executed_edge_p): New routine.
2215 * cfg.c (check_bb_profile): Add partition insanity warnings.
2217 2013-08-30 Meador Inge <meadori@codesourcery.com>
2219 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2221 2013-08-30 Marek Polacek <polacek@redhat.com>
2223 * Makefile.in (ubsan.o): Add.
2224 (c-family/c-ubsan.o): Add.
2225 (builtins.o): Add ubsan.h dependency.
2226 * ubsan.h: New file.
2227 * ubsan.c: New file.
2228 * common.opt: Add -fsanitize=undefined option.
2229 (flag_sanitize): Add variable.
2230 (fsanitize=): Add option. Add Driver.
2231 (fsanitize=thread): Remove option.
2232 (fsanitize=address): Likewise.
2233 (static-libubsan): New option.
2234 * doc/invoke.texi: Document the new flag and -static-libubsan.
2235 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
2236 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
2237 * builtin-attrs.def (ATTR_COLD): Define.
2238 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2239 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
2240 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
2241 * flag-types.h (sanitize_code): New enum.
2242 * opts.c (common_handle_option): Parse command line arguments
2243 of -fsanitize=. Add -fsanitize=unreachable option.
2244 * varasm.c (get_variable_section): Adjust.
2245 (assemble_noswitch_variable): Likewise.
2246 (assemble_variable): Likewise.
2247 (output_constant_def_contents): Likewise.
2248 (categorize_decl_for_section): Likewise.
2249 (place_block_symbol): Likewise.
2250 (output_object_block): Likewise.
2251 * builtins.def: Likewise.
2252 * toplev.c (compile_file): Likewise.
2253 (process_options): Likewise.
2254 * cppbuiltin.c: Likewise.
2255 * tsan.c (tsan_pass): Likewise.
2256 (tsan_gate): Likewise.
2257 (tsan_gate_O0): Likewise.
2258 * cfgexpand.c (partition_stack_vars): Likewise.
2259 (expand_stack_vars): Likewise.
2260 (defer_stack_allocation): Likewise.
2261 (expand_used_vars): Likewise.
2262 * cfgcleanup.c (old_insns_match_p): Likewise.
2263 * asan.c (asan_finish_file): Likewise.
2264 (asan_instrument): Likewise.
2265 (gate_asan): Likewise.
2266 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
2267 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2268 (asan_global_struct): Use pointer_sized_int_node instead
2269 calling build_nonstandard_integer_type.
2270 (initialize_sanitizer_builtins): Likewise.
2271 (asan_finish_file): Likewise.
2272 * gcc.c: Document %{%:function(args):X}.
2273 (static_spec_functions): Add sanitize.
2274 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
2275 store funcval != NULL there.
2276 (do_spec_1): Adjust handle_spec_function caller.
2277 (handle_braces): Allow %:function(args) as condition.
2278 (sanitize_spec_function): New function.
2279 (ADD_STATIC_LIBUBSAN_LIBS): Define.
2280 (LIBUBSAN_SPEC): Likewise.
2281 (LIBUBSAN_EARLY_SPEC): Likewise.
2282 (SANITIZER_SPEC): Handle libubsan.
2283 (SANITIZER_EARLY_SPEC): Likewise.
2284 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
2285 instead of fsanitize=address.
2286 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
2287 instead of fsanitize=address*.
2288 * builtins.c: Include ubsan.h.
2289 (fold_builtin_0): Instrument __builtin_unreachable.
2290 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
2291 instead of flag_asan.
2292 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
2293 (pointer_sized_int_node): Define.
2294 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
2296 2013-08-30 Mike Stump <mikestump@comcast.net>
2298 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
2301 2013-08-29 Jan Hubicka <jh@suse.cz>
2303 * cgraph.c (cgraph_function_body_availability): Handle weakref
2305 * passes.def: Remove pass_fixup_cfg.
2306 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
2307 track when we need to remove functions.
2308 (gate_ipa_inline): Execute inlining always; add comment why.
2309 (pass_data_ipa_inline): Remove TODO_remove_functions.
2310 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
2311 do not produce summaries.
2312 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
2313 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
2314 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
2315 constant pool and vtable.
2317 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
2319 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
2320 New arm_neon.h's internal macros to specify 64-bit constants.
2321 Avoid using stdint.h's macros.
2323 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
2325 * recog.c (verify_changes): Verify that changes[i].old is non-zero
2326 before applying REG_P.
2328 2013-08-30 Jakub Jelinek <jakub@redhat.com>
2330 PR tree-optimization/58277
2331 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
2332 after seeing too many stmts with vdef in between dombb and current
2333 bb, invalidate everything.
2335 2013-08-30 Richard Biener <rguenther@suse.de>
2337 * fold-const.c (fold_single_bit_test): Fix overflow test.
2339 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
2341 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
2342 and which can live in a register, always retrieve the value on entry.
2343 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
2344 passed by invisible reference specially.
2345 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
2346 (vt_add_function_parameter): Correctly deal with a parameter passed by
2347 invisible reference.
2349 2013-08-30 Jan Hubicka <jh@suse.cz>
2351 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
2353 2013-08-30 Richard Biener <rguenther@suse.de>
2355 PR tree-optimization/58228
2356 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
2357 allow invariant loads in nested loop vectorization.
2359 2013-08-30 Richard Biener <rguenther@suse.de>
2361 PR tree-optimization/58223
2362 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
2363 (has_anti_or_output_dependence): ... this and adjust to also
2364 look for output dependences.
2365 (mark_nodes_having_upstream_mem_writes): Adjust.
2366 (rdg_flag_uses): Likewise.
2368 2013-08-30 Richard Biener <rguenther@suse.de>
2370 PR tree-optimization/58010
2371 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
2372 assert that we have a loop-closed PHI.
2374 2013-08-29 Jan Hubicka <jh@suse.cz>
2376 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
2377 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
2378 * lto-section-in.c (lto_free_function_in_decl_state): New function.
2379 (lto_free_function_in_decl_state_for_node): New function.
2381 2013-08-29 Xinliang David Li <davidxl@google.com>
2383 * loop-unroll.c (report_unroll_peel): Minor message change.
2384 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
2385 Emit alignment peeling message with default -fopt-info.
2386 (vect_loop_versioning): Emit loop version info message.
2387 * tree-vectorizer.c (vectorize_loops): Minor message change.
2388 (execute_vect_slp): Ditto.
2390 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
2392 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
2393 of the clone from the DECL_NAME of the original function.
2395 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
2397 * passes.c (register_pass): Add overload.
2398 * tree-pass.h (register_pass): Forward declare it. Add comment.
2400 2013-08-29 Jan Hubicka <jh@suse.cz>
2402 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
2403 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
2404 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
2405 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2406 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
2407 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
2408 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2409 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
2411 2013-08-29 Teresa Johnson <tejohnson@google.com>
2413 * dumpfile.c (dump_loc): Output column number.
2414 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
2415 * doc/invoke.texi: Document optall -fopt-info flag.
2416 * profile.c (read_profile_edge_counts): Use new dump framework.
2417 (compute_branch_probabilities): Ditto.
2418 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
2419 when pass not in any opt group.
2420 * pass_manager.h (pass_manager::get_pass_profile): New method.
2421 * value-prof.c (check_counter): Use new dump framework.
2422 (check_ic_target): Ditto.
2423 * coverage.c (get_coverage_counts): Ditto.
2424 (coverage_init): Setup new dump framework.
2426 2013-08-29 Richard Biener <rguenther@suse.de>
2428 PR tree-optimization/58246
2429 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
2430 handle the dominance check inside a basic-block.
2432 2013-08-29 Richard Biener <rguenther@suse.de>
2435 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
2436 of default defs that appear in abnormal PHI nodes.
2438 2013-08-29 Richard Biener <rguenther@suse.de>
2440 PR tree-optimization/57685
2441 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2442 single-use operands to avoid exponential complexity.
2444 2013-08-28 Dehao Chen <dehao@google.com>
2446 * ipa-inline.c (edge_badness): Fix integer underflow.
2448 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
2450 * gtm-builtins.def (_ITM_free): Declare leaf.
2452 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2455 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2456 (*tls_local_dynamic_base_64_largepic): Likewise.
2457 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2458 Remove predicate from call operand.
2459 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2460 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2462 2013-08-28 Jeff Law <law@redhat.com>
2464 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2465 checks for the number of predecessors and successors allowed.
2466 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2467 which require copying a joiner block if there is a request which
2468 is a subpath that requires no joiner block copying.
2470 2013-08-28 Jan Hubicka <jh@suse.cz>
2472 * lto-streamer-out.c (DFS_write_tree_body): Drop
2473 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2474 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2475 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2476 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2477 Do not read DECL_ERROR_ISSUED.
2478 (unpack_ts_decl_with_vis_value_fields): Do not read
2480 (lto_input_ts_binfo_tree_pointers): Do not read
2481 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
2482 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2483 write DECL_ERROR_ISSUED..
2484 (pack_ts_decl_with_vis_value_fields): Do not write
2486 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
2487 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
2488 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2489 * tree.h (tree_decl_common): Update comment.
2490 (DECL_ERROR_ISSUED): Remove.
2492 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2495 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2497 2013-08-28 Jan Hubicka <jh@suse.cz>
2499 * builtins.def (free): Declare leaf.
2501 2013-08-27 David Malcolm <dmalcolm@redhat.com>
2504 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2505 * configure: Regenerate.
2507 2013-08-27 Martin Jambor <mjambor@suse.cz>
2509 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2510 (ipa_ancestor_jf_data): Likewise.
2511 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2512 (ipa_get_jf_pass_through_type_preserved): New function.
2513 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2514 (ipa_get_jf_ancestor_type_preserved): New function.
2515 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
2516 (ipa_get_jf_ancestor_result): Likewise.
2517 (propagate_vals_accross_pass_through): Use
2518 ipa_get_jf_pass_through_result to do all the value mappings.
2519 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2520 type_preserved flag.
2521 (ipa_set_jf_cst_copy): New function.
2522 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2523 (ipa_set_jf_arith_pass_through): Likewise.
2524 (ipa_set_ancestor_jf): Likewise.
2525 (compute_complex_assign_jump_func): Set type_preserved instead of
2527 (ipa_compute_jump_functions_for_edge): Likewise.
2528 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2529 (update_jump_functions_after_inlining): Update type_preserved.
2530 Explicitely create jump functions when combining one with pass_through.
2531 (ipa_write_jump_function): Stream the type_preserved flags.
2532 (ipa_read_jump_function): Likewise.
2534 2013-08-27 Jakub Jelinek <jakub@redhat.com>
2535 Aldy Hernandez <aldyh@redhat.com>
2537 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
2538 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
2539 * function.h (struct function): Add has_force_vect_loops and
2541 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
2542 * gimple.c (gimple_build_omp_critical): Add KIND argument and
2544 * gimple.def: Update CLAUSES comments.
2545 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
2546 (gimple_build_omp_for): Add argument to prototype.
2547 (gimple_omp_for_kind): New.
2548 (gimple_omp_for_set_kind): New.
2549 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
2550 GOVD_DATA_SHARE_CLASS.
2551 (enum omp_region_type): Add ORT_SIMD.
2552 (gimple_add_tmp_var): Handle ORT_SIMD.
2553 (gimplify_var_or_parm_decl): Same.
2554 (is_gimple_stmt): Same.
2555 (omp_firstprivatize_variable): Same.
2556 (omp_add_variable): Only use splay_tree_insert if lookup failed.
2557 (omp_notice_variable): Handle ORT_SIMD.
2558 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
2559 (omp_check_private): Handle ORT_SIMD.
2560 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2562 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
2563 Handle OMP_CLAUSE_LASTPRIVATE.
2564 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2566 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
2567 (gimplify_expr): Handle OMP_SIMD.
2568 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
2569 (expand_GOMP_SIMD_VF): New.
2570 (expand_GOMP_SIMD_LAST_LANE): New.
2571 * internal-fn.def (GOMP_SIMD_LANE): New.
2572 (GOMP_SIMD_VF): New.
2573 (GOMP_SIMD_LAST_LANE): New.
2574 * omp-low.c: Include target.h.
2575 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
2577 (check_omp_nesting_restrictions): Same.
2579 (lower_rec_simd_input_clauses): New.
2580 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
2582 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
2583 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
2584 (expand_omp_build_assign): New.
2585 (expand_omp_for_init_counts): New.
2586 (expand_omp_for_init_vars): New.
2587 (extract_omp_for_update_vars): New.
2588 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
2589 and rewrite accordingly.
2590 (expand_omp_simd): New.
2591 (expand_omp_for): Use expand_omp_simd.
2592 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2593 (lower_omp_for): Do not lower the body.
2594 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2596 * tree-flow.h (find_omp_clause): Remove prototype.
2597 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2598 forcing vectorization of the loop, or if flag_tree_vectorize.
2599 (gate_tree_if_conversion): Similarly.
2600 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2601 gimple_build_omp_for.
2602 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2603 * tree-parloops (create_parallel_loop): Pass kind argument to
2604 gimple_build_omp_for.
2605 * tree-pretty-print.c (dump_omp_clause): Add cases for
2606 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2607 OMP_CLAUSE__SIMDUID_.
2608 (dump_generic_node): Handle OMP_SIMD.
2609 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2610 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2611 unroll OMP_SIMD loops here.
2612 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
2613 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
2615 (vect_analyze_data_refs): Handle simd loops.
2616 * tree-vect-loop.c (vectorizable_live_operation): Handle
2618 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
2619 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2620 (vectorizable_load): Same.
2621 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
2622 (struct simduid_to_vf): New.
2623 (simduid_to_vf::hash): New.
2624 (simduid_to-vf::equal): New.
2625 (struct simd_array_to_simduid): New.
2626 (simd_array_to_simduid::hash): New.
2627 (simd_array_to_simduid::equal): New.
2628 (adjust_simduid_builtins): New.
2629 (struct note_simd_array_uses_struct): New.
2630 (note_simd_array_uses_cb): New.
2631 (note_simd_array_uses): New.
2632 (vectorize_loops): Handle simd hints and adjust simd builtins
2634 * tree-vectorizer.h (struct _stmt_vec_info): Add
2635 simd_lane_access_p field.
2636 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2637 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2638 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2639 (omp_clause_code_name): Same.
2640 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2641 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2642 * tree.def (OMP_SIMD): New entry.
2643 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2644 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
2645 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2646 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2647 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2648 (OMP_CLAUSE_LINEAR_STEP): New.
2649 (OMP_CLAUSE_SAFELEN_EXPR): New.
2650 (OMP_CLAUSE__SIMDUID__DECL): New.
2651 (find_omp_clause): New prototype.
2653 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2655 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2656 Haswell processor detection.
2658 2013-08-27 Christian Widmer <shadow@umbrox.de>
2661 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2662 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2663 AVX2 capable processors.
2665 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2667 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2668 of vget_low_* with implementations in terms of other intrinsics.
2670 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
2673 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2674 values to -1, 0 and 1.
2676 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2678 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2679 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2680 (<optab><mode>3_insn): Remove.
2681 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2682 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2683 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2684 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2685 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2686 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2687 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2688 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2689 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2690 (ror<mode>3_insn): Likewise.
2691 * config/aarch64/predicates.md (aarch64_simd_register): New.
2693 2013-08-27 Richard Biener <rguenther@suse.de>
2695 PR tree-optimization/57521
2696 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2697 one edge is non-critical.
2698 (find_phi_replacement_condition): Make sure to use a non-critical
2699 edge. Cleanup and remove old bug workarounds.
2700 (bb_postdominates_preds): Remove.
2701 (if_convertible_loop_p_1): Do not compute post-dominators.
2702 (combine_blocks): Do not free post-dominators.
2703 (main_tree_if_conversion): Likewise.
2704 (pass_data_if_conversion): Add TODO_verify_ssa.
2706 2013-08-27 DJ Delorie <dj@redhat.com>
2708 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2710 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2712 * function.c (assign_parm_find_data_types): Set passed_mode and
2713 nominal_mode to the TYPE_MODE of nominal_type for the built
2714 pointer type in case of the struct-pass-by-reference.
2716 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2718 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2719 (UINT16_TYPE): Change default to "unsigned int".
2721 * config/avr/avr.opt (mfract-convert-truncate): New option.
2722 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2723 is set, round negative fractional integers according to n1169
2724 when converting to integer types.
2726 2013-08-26 Jan Hubicka <jh@suse.cz>
2728 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2729 methods can not be called indirectly when their address is not taken.
2731 2013-08-26 Jan Hubicka <jh@suse.cz>
2733 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2736 2013-08-26 Jan Hubicka <jh@suse.cz>
2738 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2741 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2743 * reload.c (find_valid_class): Allow classes that do not include
2744 FIRST_PSEUDO_REGISTER - 1.
2746 2013-08-26 Jan Hubicka <jh@suse.cz>
2748 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2749 fix edge count/frequency when speculation failed; fix type check
2750 for the direct call.
2752 2013-08-26 Jan Hubicka <jh@suse.cz>
2754 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2756 2013-08-26 Jan Hubicka <jh@suse.cz>
2758 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2759 to be changed by edge redirection.
2761 2013-08-26 Jan Hubicka <jh@suse.cz>
2763 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2764 formating; add sanity check.
2765 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2766 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2767 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2768 (resolve_noninline_speculation): Update callee keys, too.
2770 2013-08-26 Jan Hubicka <jh@suse.cz>
2772 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2773 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2775 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2777 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2780 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
2782 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2783 (x86_64_elf_select_section): Ditto.
2784 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2786 (ix86_pass_by_reference): Ditto.
2787 (output_set_got): Ditto.
2788 (ix86_unary_operator_ok): Ditto.
2789 (ix86_expand_builtin): Ditto.
2791 2013-08-23 Jan Hubicka <jh@suse.cz>
2793 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
2795 2013-08-23 Jan Hubicka <jh@suse.cz>
2797 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
2798 (tree_decl_with_vis): Add FINAL field.
2800 2013-08-23 Jeff Law <law@redhat.com>
2802 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
2803 the debugging dump when the expression is fully redundant.
2805 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2807 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
2808 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
2809 * pretty-print.c (pp_formatted_text_data): Likewise.
2810 (pp_write_text_to_stream): Likewise.
2811 (pp_write_text_as_dot_label_to_stream): Likewise.
2812 (pp_append_r): Likewise.
2813 (pp_format): Likewise.
2814 (pp_flush): Likewise.
2815 (pp_clear_output_area): Likewise.
2816 (pp_append_text): Likewise.
2817 (pp_formatted_text): Likewise.
2818 (pp_remaining_character_count_for_line): Likewise.
2819 (pp_newline): Likewise.
2820 (pp_character): Likewise.
2821 (output_buffer::~output_buffer): Define.
2822 (pretty_printer::~pretty_printer): Destruct output buffer.
2823 * pretty-print.h (output_buffer::~output_buffer): Declare.
2824 (pretty_printer::~pretty_printer): Declare virtual.
2826 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
2829 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
2830 HOST_WIDE_INT_M1U): New macros.
2831 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
2832 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
2833 unsigned -1 for lshift.
2834 * cse.c (cse_insn): Likewise.
2835 * double-int.c (rshift_double, lshift_double): Likewise.
2836 * builtins.c (fold_builtin_bitop): Likewise.
2837 * combine.c (force_to_mode): Likewise.
2838 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
2839 * simplify-rtx.c (simplify_const_unary_operation,
2840 simplify_const_binary_operation): Likewise.
2841 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
2842 check_va_list_escapes): Likewise.
2843 * rtlanal.c (nonzero_bits1): Likewise.
2844 * expmed.c (expand_smod_pow2): Likewise.
2845 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
2847 2013-08-23 Jan Hubicka <jh@suse.cz>
2849 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
2850 as having address taken.
2852 2013-08-23 Jan Hubicka <jh@suse.cz>
2854 * ipa-utils.h (method_class_type): Declare.
2855 * ipa-devirt.c (method_class_type): Export.
2857 * cgraphunit.c (analyze_functions): Do basic devirtualization;
2858 do not walk base classes of anonymous types.
2860 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
2862 PR rtl-optimization/58220
2864 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
2865 handle SEQUENCE insns properly.
2867 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2869 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
2871 (pp_newline_and_indent): Likewise.
2872 (pp_separate_with): Likewise.
2873 * pretty-print.c (pp_newline_and_flush): Define.
2874 (pp_newline_and_indent): Likewise.
2875 (pp_separate_with): Likewise.
2877 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2880 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
2881 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
2883 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
2885 * config/i386/predicates.md (ext_sse_reg_operand): New.
2886 * config/i386/i386.md (*movti_internal): Use
2887 predicate to determine if EVEX is needed.
2888 (*movsi_internal): Ditto.
2889 (*movdf_internal): Ditto.
2890 (*movsf_internal): Ditto.
2891 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2893 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2895 PR tree-optimization/58209
2896 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
2897 (find_tail_calls): Give up for pointer result types if m is non-NULL.
2898 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
2899 emit POINTER_PLUS_EXPR.
2900 (create_tailcall_accumulator): For pointer result type accumulate in
2903 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
2905 * configure.ac: Add backslashes missing from the last change.
2906 * configure: Regenerate.
2908 2013-08-22 Jan Hubicka <jh@susue.cz>
2910 * ipa.c (function_and_variable_visibility): First remember function
2911 was global and then make it local.
2913 2013-08-22 Julian Brown <julian@codesourcery.com>
2915 * configure.ac: Add aarch64 to list of arches which use "nop" in
2917 * configure: Regenerate.
2919 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2921 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
2922 gnu_libc_has_function.
2923 * config/s390/tpf.h: Likewise.
2925 2013-08-22 Jan Hubicka <jh@susue.cz>
2927 * timevar.c (validate_phases): Add cast.
2929 2013-08-22 Jan Hubicka <jh@susue.cz>
2931 * timevar.c (validate_phases): Use size_t for memory.
2932 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
2934 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
2936 * pretty-print.h (output_buffer::output_buffer): Declare.
2937 (pretty_printer::pretty_printer): Likewise.
2938 (pp_construct): Remove.
2939 * pretty-print.c (output_buffer::output_buffer): Define.
2940 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
2941 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
2942 (print_gimple_expr): Likewise.
2943 (print_gimple_seq): Likewise.
2944 (gimple_dump_bb): Likewise.
2945 * sched-vis.c (dump_value_slim): Likewise.
2946 (dump_insn_slim): Likewise.
2947 (dump_rtl_slim): Likewise.
2948 (str_pattern_slim): Likewise.
2949 * tree-mudflap.c (mf_varname_tree): Likewise.
2950 * graph.c (print_graph_cfg): Likewise.
2951 (start_graph_dump): Likewise.
2952 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
2954 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
2955 pretty printer initialization.
2956 * coretypes.h (diagnostic_context): Remove superflous type alias
2958 (pretty_printer): Likewise. Declare directly as a class.
2959 (pretty_print_info): Remove declaration as class.
2960 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
2961 and pp_clear_output_area.
2962 (asan_add_global): Likewise.
2964 2013-08-22 Jan Hubicka <jh@suse.cz>
2966 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
2967 * ipa-utils.h (update_type_inheritance_graph): Declare.
2968 (possible_polymorphic_call_target_p): Declare.
2969 (possible_polymorphic_call_target_p): New.
2970 * ipa-devirt.c: Update toplevel comments.
2971 (cached_polymorphic_call_targets): Move up.
2972 (odr_type_d): Move ID down.
2973 (polymorphic_type_binfo_p): Update comment.
2974 (odr_hasher::remove): Likewise;
2975 (get_odr_type): Set anonymous_namespace.
2976 (dump_odr_type): Dump it.
2977 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
2978 (maybe_record_node): Record node in cached_polymorphic_call_targets.
2979 (record_binfo): Add comment.
2980 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
2982 (devirt_node_removal_hook): Do not iCE when cache is freed.
2983 (possible_polymorphic_call_target_p): New predicate.
2984 (update_type_inheritance_graph): New function.
2986 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
2987 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2988 Sergey Lega <sergey.s.lega@intel.com>
2989 Anna Tikhonova <anna.tikhonova@intel.com>
2990 Ilya Tocar <ilya.tocar@intel.com>
2991 Andrey Turetskiy <andrey.turetskiy@intel.com>
2992 Ilya Verbin <ilya.verbin@intel.com>
2993 Kirill Yukhin <kirill.yukhin@intel.com>
2994 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2996 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
2997 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
2998 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
2999 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
3000 (OPTION_MASK_ISA_AVX2_UNSET): Update.
3001 (OPTION_MASK_ISA_AVX512F_UNSET): New.
3002 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
3003 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
3004 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
3005 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
3006 OPT_mavx512pf, OPT_mavx512er cases.
3007 * config/i386/constraints.md (v): New constraint.
3008 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
3009 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
3010 (bit_AVX512CD): New.
3011 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3012 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
3013 * config/i386/i386-c.c (ix86_target_macros_internal):
3014 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
3016 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
3017 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
3018 * config/i386/i386.c (regclass_map, dbx_register_map)
3019 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
3020 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
3021 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
3023 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
3024 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
3025 -mavx512pf options. Fix formatting.
3026 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
3027 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
3028 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
3029 -mavx512cd, -mavx512pf options.
3030 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
3031 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
3032 (ix86_preferred_output_reload_class): Replace SSE_REGS with
3034 (ix86_hard_regno_mode_ok): Support 512-bit registers.
3035 (ix86_set_reg_reg_cost): Ditto.
3036 (x86_order_regs_for_local_alloc): Ditto.
3037 (MAX_VECT_LEN): Extend to 64-byte.
3038 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
3039 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
3040 (TARGET_AVX512ER, TARGET_AVX512CD): New.
3041 (BIGGEST_ALIGNMENT): Extend to 512-bits.
3042 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
3043 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
3044 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
3045 (SSE_REG_MODE_P): Support new modes.
3046 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
3047 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
3048 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
3049 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
3050 (REG_CLASS_CONTENTS): Add new registers.
3051 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
3052 (EXT_REX_SSE_REGNO_P): New.
3053 (HI_REGISTER_NAMES): Add new registers.
3054 * config/i386/i386.md: Define constants for new registers.
3055 (mode): Add new 512-bit modes.
3056 (prefix): Support evex prefix.
3057 (isa): Support avx512f, noavx512f, fma_avx512f.
3058 (ssemodesuffix): Add new 512-bit modes.
3060 (*movxi_internal_avx512f): Ditto.
3061 (*movdi_internal): Replace constraint "x" with the new constraint "v".
3063 (*movsi_internal): Likewise.
3064 (*movdf_internal): Likewise.
3065 (*movsf_internal): Likewise.
3066 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
3067 (<code><mode>3): Likewise.
3068 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
3070 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
3071 with the new constraint "v".
3072 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
3074 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
3075 with the new constraint "v".
3076 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
3077 (<sse2>_storedqu<avxsizesuffix>): Likewise.
3078 (*<plusminus_insn><mode>3): Likewise.
3079 (<sse>_vm<plusminus_insn><mode>3): Likewise.
3080 (*mul<mode>3): Likewise.
3081 (<sse>_vmmul<mode>3): Likewise.
3082 (<sse>_div<mode>3): Likewise.
3083 (<sse>_vmdiv<mode>3): Likewise.
3084 (<sse>_sqrt<mode>2): Likewise.
3085 (<sse>_vmsqrt<mode>2): Likewise.
3086 (*<code><mode>3_finite): Likewise.
3087 (*<code><mode>3) <smaxmin>: Likewise.
3088 (<sse>_vm<code><mode>3): Likewise.
3089 (*<code><mode>3) <any_logic>: Likewise.
3090 (*fma_fmadd_<mode>): Likewise.
3091 (*fma_fmsub_<mode>): Likewise.
3092 (*fma_fnmadd_<mode>): Likewise.
3093 (*fma_fnmsub_<mode>): Likewise.
3094 (*fma_fmaddsub_<mode>): Likewise.
3095 (*fma_fmsubadd_<mode>): Likewise.
3096 (*fmai_fmadd_<mode>): Likewise.
3097 (*fmai_fmsub_<mode>): Likewise.
3098 (*fmai_fnmadd_<mode>): Likewise.
3099 (*fmai_fnmsub_<mode>): Likewise.
3100 (sse_cvtsi2ss): Likewise.
3101 (sse_cvtsi2ssq): Likewise.
3102 (sse_cvtss2si): Likewise.
3103 (sse_cvtss2si_2): Likewise.
3104 (sse_cvtss2siq): Likewise.
3105 (sse_cvtss2siq_2): Likewise.
3106 (sse_cvttss2si): Likewise.
3107 (sse_cvtss2siq_2): Likewise.
3108 (float<sseintvecmodelower><mode>2): Likewise.
3109 (sse2_cvtsd2si_2): Likewise.
3110 (sse2_cvtsd2siq_2): Likewise.
3111 (*<plusminus_insn><mode>3): Likewise.
3112 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
3113 (*<sse4_1_avx2>_mul<mode>3): Likewise.
3114 (ashr<mode>3): Likewise.
3115 (<shift_insn><mode>3): Likewise.
3116 (avx2_<code><mode>3): Likewise.
3117 (*avx2_<code><mode>3): Likewise.
3118 (*andnot<mode>3): Likewise.
3119 (*<code><mode>3) <any_logic>: Likewise.
3120 (abs<mode>2): Likewise.
3121 (avx2_permvar<mode>): Likewise.
3122 (avx2_perm<mode>_1): Likewise.
3123 (*avx_vpermilp<mode>): Likewise.
3124 (avx_vpermilvar<mode>3): Likewise.
3125 (avx2_ashrv<mode>): Likewise.
3126 (avx2_<shift_insn>v<mode>): Likewise.
3127 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
3129 * doc/rtl.texi: Document XImode.
3131 2013-08-21 Jeff Law <law@redhat.com>
3133 * tree-flow.h (register_jump_thread): Pass vector of edges
3134 instead of each important edge.
3135 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
3136 thread path into a vector and pass that to register_jump_thread.
3137 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
3138 passed in edge vector to the current 3-edge form.
3141 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3143 * dce.c (fini_dce): Call df_analyze again just in case
3144 delete_unmarked_insns removed anything.
3146 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
3148 * reload.h (struct reg_equivs): Rename to ..
3149 (struct reg_equivs_s): .. this.
3151 2013-08-20 Martin Liska <marxin.liska@gmail.com>
3153 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
3155 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3157 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
3159 2013-08-21 Jeff Law <law@redhat.com>
3161 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
3162 simplify assignments too. If the RHS collapses to a singleton
3163 range, then return the value for the range.
3165 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
3167 * config/i386/sse.md (V16): Rename to...
3169 (mov<mode>): Update iterator name.
3170 (*mov<mode>_internal): Ditto.
3171 (push<mode>1): Ditto.
3172 (movmisalign<mode>): Ditto.
3174 2013-08-20 Jan Hubicka <jh@suse.cz>
3177 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
3178 entry for direct edges.
3179 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
3181 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3183 Revert my last two changes, r201865 and r201864:
3186 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3188 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3189 instances can own GC refs.
3191 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3192 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3193 (gcc::context::gt_pch_nx): Likewise.
3194 (gcc::context::gt_pch_nx): Likewise.
3195 * ggc.h (gt_ggc_mx <T>): New.
3196 (gt_pch_nx_with_op <T>): New.
3197 (gt_pch_nx <T>): New.
3198 * passes.c (opt_pass::gt_ggc_mx): New.
3199 (opt_pass::gt_pch_nx): New.
3200 (opt_pass::gt_pch_nx_with_op): New.
3201 (pass_manager::gt_ggc_mx): New.
3202 (pass_manager::gt_pch_nx): New.
3203 (pass_manager::gt_pch_nx_with_op): New.
3204 (pass_manager::operator new): Use
3205 ggc_internal_cleared_alloc_stat rather than xcalloc.
3206 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3207 (pass_manager::gt_ggc_mx): New.
3208 (pass_manager::gt_pch_nx): New.
3209 (pass_manager::gt_pch_nx_with_op): New.
3210 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3211 (opt_pass::operator new): New.
3212 (opt_pass::gt_ggc_mx): New.
3213 (opt_pass::gt_pch_nx): New.
3214 (opt_pass::gt_pch_nx_with_op): New.
3217 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3219 * Makefile.in (GTFILES): Add context.h.
3220 * context.c (gcc::context::operator new): New.
3221 (gcc::context::gt_ggc_mx): New.
3222 (gcc::context::gt_pch_nx): New.
3223 (gcc::context::gt_pch_nx): New.
3224 * context.h (gcc::context): Add GTY((user)) marking.
3225 (gcc::context::operator new): New.
3226 (gcc::context::gt_ggc_mx): New.
3227 (gcc::context::gt_pch_nx): New.
3228 (gcc::context::gt_pch_nx): New.
3229 (g): Add GTY marking.
3230 (gt_ggc_mx (gcc::context *)): New.
3231 (gt_pch_nx (gcc::context *)): New.
3232 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3233 void *cookie)): New.
3234 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3236 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3238 * dce.c (fini_dce): Call df_analyze again just in case
3239 delete_unmarked_insns removed anything.
3241 2013-08-20 Teresa Johnson <tejohnson@google.com>
3243 PR rtl-optimizations/57451
3244 * final.c (reemit_insn_block_notes): Prevent lexical blocks
3245 from crossing split section boundaries.
3247 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
3249 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
3250 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
3251 with MULTLIB_DEFAULTS.
3253 2013-08-20 Nick Clifton <nickc@redhat.com>
3255 * target.def (narrow_volatile_bitfield): Note that the default
3256 value is false, not !TARGET_STRICT_ALIGN.
3257 * doc/tm.texi: Regenerate.
3259 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
3261 Fix LIB_SPEC for systems without libpthread.
3263 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
3264 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
3266 * config/i386/linux-common.h: Likewise.
3267 * config/mips/linux-common.h: Likewise.
3269 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
3271 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
3274 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3276 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3277 instances can own GC refs.
3279 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3280 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3281 (gcc::context::gt_pch_nx): Likewise.
3282 (gcc::context::gt_pch_nx): Likewise.
3283 * ggc.h (gt_ggc_mx <T>): New.
3284 (gt_pch_nx_with_op <T>): New.
3285 (gt_pch_nx <T>): New.
3286 * passes.c (opt_pass::gt_ggc_mx): New.
3287 (opt_pass::gt_pch_nx): New.
3288 (opt_pass::gt_pch_nx_with_op): New.
3289 (pass_manager::gt_ggc_mx): New.
3290 (pass_manager::gt_pch_nx): New.
3291 (pass_manager::gt_pch_nx_with_op): New.
3292 (pass_manager::operator new): Use
3293 ggc_internal_cleared_alloc_stat rather than xcalloc.
3294 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3295 (pass_manager::gt_ggc_mx): New.
3296 (pass_manager::gt_pch_nx): New.
3297 (pass_manager::gt_pch_nx_with_op): New.
3298 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3299 (opt_pass::operator new): New.
3300 (opt_pass::gt_ggc_mx): New.
3301 (opt_pass::gt_pch_nx): New.
3302 (opt_pass::gt_pch_nx_with_op): New.
3304 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3306 * Makefile.in (GTFILES): Add context.h.
3307 * context.c (gcc::context::operator new): New.
3308 (gcc::context::gt_ggc_mx): New.
3309 (gcc::context::gt_pch_nx): New.
3310 (gcc::context::gt_pch_nx): New.
3311 * context.h (gcc::context): Add GTY((user)) marking.
3312 (gcc::context::operator new): New.
3313 (gcc::context::gt_ggc_mx): New.
3314 (gcc::context::gt_pch_nx): New.
3315 (gcc::context::gt_pch_nx): New.
3316 (g): Add GTY marking.
3317 (gt_ggc_mx (gcc::context *)): New.
3318 (gt_pch_nx (gcc::context *)): New.
3319 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3320 void *cookie)): New.
3321 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3323 2013-08-20 Alan Modra <amodra@gmail.com>
3326 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
3327 (rs6000_emit_epilogue): Likewise.
3329 2013-08-19 Dehao Chen <dehao@google.com>
3331 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
3333 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
3334 Jakub Jelinek <jakub@redhat.com>
3336 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
3337 (BUILT_IN_FABSD64): Likewise.
3338 (BUILT_IN_FABSD128): Likewise.
3339 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
3340 (fold_builtin_1): Likewise.
3341 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
3342 destination and source operands.
3343 (*abstd2_fpr): Likewise.
3344 (*nabstd2_fpr): Likewise.
3346 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
3348 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
3351 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
3353 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
3356 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
3358 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
3359 * builtins.c (default_libc_has_function): New.
3360 (gnu_libc_has_function): Ditto.
3361 (no_c99_libc_has_function): Ditto.
3362 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
3363 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
3364 (fold_builtin_sincos): Likewise.
3365 (fold_builtin_cexp): Likewise.
3366 * builtins.def (DEF_C94_BUILTIN): Likewise.
3367 (DEF_C99_BUILTIN): Likewise.
3368 (DEF_C99_C90RES_BUILTIN): Likewise.
3369 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
3370 definitions to using this define.
3371 * config/darwin-protos.h (darwin_libc_has_function): New.
3372 * config/darwin.c (darwin_libc_has_function): Ditto.
3373 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
3374 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
3375 * config/darwin.h: Ditto.
3376 * config/elfos.h: Ditto.
3377 * config/freebsd.h: Ditto.
3378 * config/i386/cygming.h: Ditto.
3379 * config/i386/djgpp.h: Ditto.
3380 * config/i386/i386-interix.h: Ditto.
3381 * config/microblaze/microblaze.h: Ditto.
3382 * config/mmix/mmix.h: Ditto.
3383 * config/gnu-user.h: Ditto.
3384 * config/ia64/hpux.h: Ditto.
3385 * config/pa/pa-hpux.h: Ditto.
3386 * config/pdp11/pdp11.h: Ditto.
3387 * config/picochip/picochip.h: Ditto.
3388 * config/linux.h: Ditto.
3389 * config/netbsd.h: Ditto.
3390 * config/openbsd.h: Ditto.
3391 * config/rs6000/aix43.h: Ditto.
3392 * config/rs6000/aix51.h: Ditto.
3393 * config/rs6000/aix52.h: Ditto.
3394 * config/rs6000/aix53.h: Ditto.
3395 * config/rs6000/aix61.h: Ditto.
3396 * config/rs6000/darwin.h: Ditto.
3397 * config/rs6000/linux.h: Ditto.
3398 * config/rs6000/linux64.h: Ditto.
3399 * config/s390/tpf.h: Ditto.
3400 * config/sol2-10.h: Ditto.
3401 * config/sol2.h: Ditto.
3402 * config/vms/vms.h: Ditto.
3403 * config/vxworks.h: Ditto.
3404 * config/linux-android.c (linux_android_libc_has_function):
3405 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
3406 * config/linux-protos.h (linux_android_libc_has_function):
3408 * config/i386/i386.c (ix86_libc_has_function): New.
3409 * config/i386/i386-protos.h
3410 (ix86_libc_has_function): New declaration.
3411 * config/i386/i386.md
3412 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
3413 ("isinf<mode>2): Likewise.
3414 * convert.c (convert_to_integer): Using new target hook
3415 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
3416 TARGET_C99_FUNCTIONS.
3417 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
3418 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
3419 * coretypes.h (function_class): New enum for different
3420 classes of functions.
3421 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
3422 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
3423 (TARGET_HAS_SINCOS): Likewise.
3424 (TARGET_LIBC_HAS_FUNCTION): New.
3425 * doc/tm.texi: Regenerated.
3426 * targhooks.h (default_libc_has_function): New declaration.
3427 (no_c99_libc_has_function): Ditto.
3428 (gnu_libc_has_function): Ditto.
3429 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
3430 and TARGET_HAS_SINCOS.
3432 2013-08-18 Jan Hubicka <jh@suse.cz>
3434 * Makeifle-in (ipa-devirt.o): New.
3435 (GTFILES): Add ipa-utils.h and ipa-devirt.c
3436 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
3437 (analyze_functions): Look into possible targets of polymorphic call.
3438 * dumpfile.c (dump_files): Add type-inheritance dump.
3439 * dumpfile.h (TDI_inheritance): New.
3440 * ipa-devirt.c: New file.
3441 * ipa-utils.h (odr_type_d): Forward declare.
3442 (odr_type): New type.
3443 (build_type_inheritance_graph): Declare.
3444 (possible_polymorphic_call_targets): Declare and introduce inline
3445 variant when only edge is pased.
3446 (dump_possible_polymorphic_call_targets): Likewise.
3447 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3448 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3449 (types_same_for_odr): ... here.
3450 * tree.h (type_in_anonymous_namespace_p): Declare.
3452 2013-08-18 Jakub Jelinek <jakub@redhat.com>
3454 PR tree-optimization/58006
3455 * tree-parloops.c (take_address_of): Don't ICE if get_name
3457 (eliminate_local_variables_stmt): Remove clobber stmts.
3459 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3461 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3462 error message is issued for an alias to undefined symbol.
3464 2013-08-18 Jan Hubicka <jh@suse.cz>
3466 * cgraph.c (cgraph_create_indirect_edge): Discover
3467 polymorphic calls and record basic info into indirect_info.
3468 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3469 devirtualization, ignore objc function calls.
3470 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3471 call with no parm index info.
3472 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3473 * tree.c (virtual_method_call_p): New function.
3474 * tree.h (virtual_method_call_p): Declare.
3476 2013-08-16 Jan Hubicka <jh@suse.cz>
3479 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3481 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
3483 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3486 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3488 * gengtype.c (type_for_name): Add special-case support for
3489 locating types within the "gcc::" namespace.
3490 (open_base_files): Emit a "using namespace gcc" directive.
3492 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3495 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3496 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3498 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3499 array instead of each individual operand as a separate argument.
3500 (emit_fusion_gpr_load): Likewise.
3501 (expand_fusion_gpr_load): Add new function declaration.
3503 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3504 signature to have the operands passed as an array, instead of as
3505 separate arguments. Allow ZERO_EXTEND to be in the memory
3506 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3507 depend on the register live/dead flags when peepholes are run.
3508 (expand_fusion_gpr_load): New function to be called from the
3509 peephole2 pass, to change the register that addis sets to be the
3511 (emit_fusion_gpr_load): Change the calling signature to have the
3512 operands passed as an array, instead of as separate arguments.
3513 Allow ZERO_EXTEND to be in the memory address, and also
3514 SIGN_EXTEND if -mpower8-fusion-sign.
3516 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3518 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3519 adjust the register addis loads up in the peephole2 pass. Do not
3520 depend on the register live/dead state when the peephole pass is done.
3522 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3524 * gengtype.c (create_user_defined_type): Ensure that the kind
3525 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3526 declaration is seen before the GTY((user)) marking.
3528 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
3531 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
3533 2013-08-16 Jan Hubicka <jh@suse.cz>
3535 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
3536 arugment expected_type.
3537 (gimple_fold_call): Use it.
3538 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
3539 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
3540 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
3541 (try_make_edge_direct_virtual_call): Likewise.
3542 * tree.c (obj_type_ref_class): New.
3543 * tree.h (obj_type_ref_class): Use it.
3545 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
3547 * sched-vis.c (rtl_slim_pp_initialized): Remove.
3548 (rtl_slim_pp): Likewise.
3549 (init_rtl_slim_pretty_print): Likewise.
3550 (dump_value_slim): Don't call it. Use local pretty printer.
3551 (dump_insn_slim): Likewise.
3552 (dump_rtl_slim): Likewise.
3553 (str_pattern_slim): Likewise.
3554 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
3557 2013-08-16 Jakub Jelinek <jakub@redhat.com>
3559 PR tree-optimization/58164
3560 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
3561 walk gimple_goto_dest of GIMPLE_GOTO.
3563 PR tree-optimization/58165
3564 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
3565 bi_call must be the last stmt in a bb, don't split_block, instead
3566 use fallthru edge from it and give up if there is none.
3567 Release conds vector when returning early.
3569 2013-08-14 Xinliang David Li <davidxl@google.com>
3571 * config/i386/i386.c (ix86_option_override_internal):
3572 Remove unused variable and field.
3574 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3577 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
3578 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
3579 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
3580 and Linux, correct BLKmode alignment when 128-bit alignment is
3581 required and compatibility flag is not set.
3582 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
3583 for zero-size arguments when compatibility flag is not set.
3585 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3587 PR tree-optimization/58145
3588 * tree-sra.c (build_ref_for_offset): If prev_base has
3589 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
3591 2013-08-14 Xinliang David Li <davidxl@google.com>
3593 * config/i386/i386.c (ix86_option_override_internal):
3594 Fix uninitialized variable error.
3596 2013-08-14 Xinliang David Li <davidxl@google.com>
3598 * config/i386/i386.opt: Define two new options.
3599 * config/i386/x86-tune.def: Add arch selector field in macros.
3600 * config/i386/i386.h: Adjust macro definition.
3601 * config/i386/i386.c (ix86_option_override_internal):
3603 (parse_mtune_ctrl_str): New function.
3604 (set_ix86_tune_features): New function.
3605 (ix86_function_specific_restore): Call the new helper function.
3607 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
3609 PR rtl-optimization/57662
3610 * sel-sched.c (code_motion_process_successors): When the current insn
3611 is removed after the recursive traversal, break from the loop.
3612 Add comments and debug printouts.
3614 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3615 Alexandre Oliva <aoliva@redhat.com>
3618 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3619 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3620 there also UNSPEC_PLTOFF.
3622 2013-08-14 Marek Polacek <polacek@redhat.com>
3624 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3625 NUM_CONDITIONS bit positions.
3627 2013-08-13 Cary Coutant <ccoutant@google.com>
3629 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3630 (attr_checksum): Hash vector contents instead of pointer.
3631 (attr_checksum_ordered): Likewise.
3633 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
3635 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3636 when Pmode != word_mode. Add length_address attribute.
3637 (sse3_monitor_<mode>): Merge from sse3_monitor and
3638 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3639 Pmode != word_mode. Update insn length attribute.
3640 * config/i386/i386.c (ix86_option_override_internal): Update
3641 ix86_gen_monitor selection for merged sse3_monitor insn.
3643 2013-08-13 Julian Brown <julian@codesourcery.com>
3645 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3646 perform invalid legitimization on greater-than-word-size modes for
3649 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3651 * ira.c (setup_class_translate_array): Use aclass instead of cl
3652 for classes not fully covered by allocno classes.
3654 2013-08-13 Jakub Jelinek <jakub@redhat.com>
3656 PR tree-optimization/57661
3657 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3658 * tree-inline.c (tree_function_versioning): Initialize it.
3659 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3660 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3661 that is not being copied.
3664 * asan.c (instrument_strlen_call): Fix typo in comment.
3665 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3667 2013-08-13 Steve Ellcey <sellcey@mips.com>
3669 * config/mips/mips.md (prefetch): Use lw instead of ld on
3670 loongson in 32bit mode.
3672 2013-08-13 Nick Clifton <nickc@redhat.com>
3674 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3676 2013-08-13 Jan Hubicka <jh@suse.cz>
3678 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3679 introduced edge; fix typo in sanity check.
3680 (cgraph_resolve_speculation): Export; improve diagnostic.
3681 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3682 speculation at type mismatch.
3683 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3684 (cgraph_resolve_speculation): Declare.
3685 (symtab_can_be_discarded): New function.
3686 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3687 * ipa-inline-transform.c (speculation_removed): New global var.
3688 (clone_inlined_nodes): See if speculation can be removed.
3689 (inline_call): If speculations was removed, we growths may not match.
3690 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3691 (speculation_useful_p): New function.
3692 (resolve_noninline_speculation): New function.
3693 (inline_small_functions): Resolve useless speculations.
3694 * ipa-inline.h (speculation_useful_p): Declare
3695 * ipa.c (can_replace_by_local_alias): Simplify.
3696 (ipa_profile): Produce speculative calls in non-lto, too;
3697 add simple cost model; produce local aliases.
3699 2013-08-13 David Malcolm <dmalcolm@redhat.com>
3701 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
3704 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3706 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3708 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
3710 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3711 iterator instead of X87MODEF.
3713 2013-08-12 Perez Read <netfirewall@gmail.com>
3716 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3717 operand 0 for intel asm alternative.
3718 (*movabs<mode>_2): Ditto for operand 1.
3720 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3722 * config/aarch64/arm_none.h
3723 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3725 2013-08-12 Nick Clifton <nickc@redhat.com>
3727 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3729 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3731 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3732 expand for QI/HImode operand to produce more effictive code for
3733 unsigned char(short) --> float(double) conversion.
3735 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
3737 * doc/invoke.texi: Mention that -ftls-model does not force the final
3740 2013-08-12 Marek Polacek <polacek@redhat.com>
3741 Marc Glisse <marc.glisse@inria.fr>
3743 PR tree-optimization/57980
3744 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3745 when creating -1 constant.
3747 2013-08-10 Jan Hubicka <jh@suse.cz>
3749 Workaround binutils PR14342.
3750 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3751 (gimple_init_edge_profiler): Likewise.
3752 (gimple_gen_ic_func_profiler): Likewise.
3754 2013-08-09 Jan Hubicka <jh@suse.cz>
3756 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3758 2013-08-09 Xinliang David Li <davidxl@google.com>
3760 * config/i386/stringop.def: New file.
3761 * config/i386/stringop.opt: New file.
3762 * config/i386/i386-opts.h: Include stringopt.def.
3763 * config/i386/i386.opt: Include stringopt.opt.
3764 * config/i386/i386.c (ix86_option_override_internal):
3765 Override default size based stringop inline strategies with options.
3766 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3769 2013-08-09 Jan Hubicka <jh@suse.cz>
3771 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
3773 2013-08-09 Jan Hubicka <jh@suse.cz>
3775 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3777 (dump_cgraph_node): Dump profile-id.
3778 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3779 and common_target_probability.
3780 * lto-cgraph.c (lto_output_edge): Stream common targets.
3781 (lto_output_node): Stream profile ids.
3782 (input_node): Stream profile ids.
3783 (input_edge): Stream common targets.
3784 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3785 * ipa.c: Include value-prof.h
3786 (ipa_profile_generate_summary): Turn indirect call statement histograms
3787 into common targets.
3788 (ipa_profile): Turn common targets into speculative edges.
3790 2013-08-09 Jan Hubicka <jh@suse.cz>
3792 * cgraph.h (cgraph_node): Add profile_id.
3793 * value-prof.c (cgraph_node_map): Turn into pointer_map.
3794 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
3795 (del_node_map): Update.
3796 (find_func_by_funcdef_no): Replace by ...
3797 (find_func_by_profile_id): ... this one.
3798 (gimple_ic_transform): Do not remove useful histograms when
3799 speculation is not done; dump info when indirect call removal
3801 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
3802 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
3803 (__gcov_indirect_call_profiler_v2): .. this one.
3804 * profile.h (init_node_map): Update.
3805 * coverage.c (coverage_compute_profile_id): New function.
3806 * coverage.h (coverage_compute_profile_id): Declare.
3807 * tree-profile.c (init_ic_make_global_vars): Make
3808 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
3809 (gimple_init_edge_profiler): Update prototype of
3810 __gcov_indirect_call_profiler.
3811 (gimple_gen_ic_func_profiler): Simplify.
3812 (tree_profiling): Use init_node_map
3814 2013-08-09 Jan Hubicka <jh@suse.cz>
3816 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
3817 non-speculative refs.
3818 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
3819 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
3820 (cgraph_set_call_stmt): Likewise.
3821 (cgraph_create_edge_1): Fix release checking compilatoin;
3823 (cgraph_free_edge): Free indirect info.
3824 (cgraph_turn_edge_to_speculative): New function.
3825 (cgraph_speculative_call_info): New function.
3826 (cgraph_make_edge_direct): Return direct edge; handle speculation.
3827 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
3828 (dump_cgraph_node): Dump speculation.
3829 (verify_edge_count_and_frequency): Accept speculative edges.
3830 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
3831 (verify_cgraph_node): Handle speculation.
3832 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
3833 (cgraph_set_call_stmt): Update prototype.
3834 (cgraph_make_edge_direct): Update prototype.
3835 (cgraph_speculative_call_info): Declare.
3836 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
3837 to change; update call of ipa_find_references.
3838 * ipa-ref.c (ipa_record_reference): Fix return value; clear
3839 lto_stmt_uid and speculative flags.
3840 (ipa_dump_references): Dump speculation.
3841 (ipa_clone_references): Clone speculative flag.
3842 (ipa_clone_referring): Likewise.
3843 (ipa_clone_ref): New function.
3844 (ipa_find_reference): Look into lto_stmt_uids
3845 (ipa_clear_stmts_in_references): Do not clear speculative calls.
3846 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
3847 (ipa_find_reference): Update declaration.
3848 (ipa_clone_ref): Declare.
3849 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
3850 stream speculative flag.
3851 (lto_output_ref): Stream statements uids and speculation.
3852 (input_ref): Likewise.
3853 (input_edge): Stream speuclation.
3854 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
3855 (cgraph_set_call_stmt_including_clones): Handle speculation.
3856 * ipa-inline.c (heap_edge_removal_hook): New function.
3857 (inline_small_functions): Register it.
3858 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
3859 also initialize refs.
3860 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
3862 (try_make_edge_direct_simple_call): Likewise.
3863 (try_make_edge_direct_simple_call): Likewise.
3864 (update_indirect_edges_after_inlining): Likewise.
3865 (remove_described_reference): Look proper lto_stmt_uid.
3866 (propagate_controlled_uses): Likewise.
3867 (propagate_controlled_uses): Liekwise.
3868 * tree-inline.c (copy_bb): Copy speculative edges.
3869 (redirect_all_calls): New function.
3870 (copy_cfg_body): Do redirection after loop info is updated.
3871 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
3873 2013-08-09 Jan Hubicka <jh@suse.cz>
3875 * lto-streamer-out.c (output_function): Renumber PHIs.
3876 * lto-streamer-in.c (input_function): Likewise.
3878 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
3880 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
3881 (get_lane_unsigned): Likewise.
3882 (dup_lane_scalar): Likewise.
3883 (get_lane): enable for VALL.
3884 * config/aarch64/aarch64-simd.md
3885 (aarch64_dup_lane_scalar<mode>): Remove.
3886 (aarch64_get_lane_signed<mode>): Likewise.
3887 (aarch64_get_lane_unsigned<mode>): Likewise.
3888 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
3889 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
3890 (aarch64_get_lane<mode>): Enable for all vector modes.
3891 (aarch64_get_lanedi): Remove misleading constraints.
3892 * config/aarch64/arm_neon.h
3893 (__aarch64_vget_lane_any): Define.
3894 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
3895 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
3896 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
3897 * config/aarch64/iterators.md (VDQQH): New.
3901 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
3903 * configure.ac: Add GAS check for LEON instructions on SPARC.
3904 * configure: Regenerate.
3905 * config.in: Likewise.
3906 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
3908 * config/sparc/sparc.opt (LEON, LEON3): New masks.
3909 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
3911 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
3912 (AS_LEON_FLAG): New macro.
3913 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
3914 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
3915 Deal with LEON and LEON3 for the memory model.
3916 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
3917 (atomic_compare_and_swap<mode>_1): Likewise.
3918 (*atomic_compare_and_swap<mode>_1): Likewise.
3920 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3922 * config/arm/neon.md (vcond): Fix floating-point vector
3923 comparisons against 0.
3925 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
3927 * lra-constraints.c (emit_spill_move): Remove assert.
3928 (process_alt_operands): Add more debugging
3929 output. Increase reject for spilling into memory. Decrease
3930 reject for reloading scratch.
3931 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
3933 2013-08-08 Steve Ellcey <sellcey@mips.com>
3935 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
3936 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
3937 micromips incompatible. Add nan2008.
3938 (MULTILIB_DIRNAMES): Add nan2008.
3939 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3940 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
3941 and micromips incompatible. Add nan2008.
3942 (MULTILIB_DIRNAMES): Add nan2008.
3943 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3945 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
3947 PR rtl-optimization/58079
3948 * combine.c (combine_simplify_rtx): Avoid using SUBST if
3949 simplify_comparison has widened a comparison with an integer.
3951 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3953 * config/arm/neon.md (movmisalign<mode>): Disable when we
3954 don't allow unaligned accesses.
3955 (*movmisalign<mode>_neon_store): Likewise.
3956 (*movmisalign<mode>_neon_load): Likewise.
3957 (*movmisalign<mode>_neon_store): Likewise.
3958 (*movmisalign<mode>_neon_load): Likewise.
3960 2013-08-08 Jan Hubicka <jh@suse.cz>
3962 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
3963 (make_pass_rebuild_cgraph_edges): Also clear references.
3964 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
3965 * ipa-inline-transform.c (inline_transform): Remove all references
3967 * cgraphunit.c (expand_function): Remove all references after
3969 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
3970 (ipa_find_reference): Rewrite to iterator.
3971 (remove_stmt_references): Likewise.
3972 (ipa_clear_stmts_in_references): New function.
3973 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
3974 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
3976 * ipa-split.c (split_function): Remove references in split function.
3978 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
3981 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
3982 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
3984 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
3987 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
3988 suggested mode for the assignment isn't compatible with the
3991 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
3994 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
3996 2013-08-07 Xinliang David Li <davidxl@google.com>
3998 * config/i386/i386.opt: New option -mtune-ctrl=.
3999 * config/i386/x86-tune.def: New file.
4000 * config/i386/i386.h: include x86-tune.def.
4001 * config/i386/i386.c (ix86_option_override_internal):
4002 Parsing -mtune-ctrl= option and set tune features.
4004 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
4007 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
4008 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
4011 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4013 * rtl.h (update_alignments): Declare.
4014 * final.c (grow_label_align): New function extracted from...
4015 (shorten_branches): ...here. Call it.
4016 (update_alignments): New function.
4017 * reorg.c (sibling_labels): New variable.
4018 (get_label_before): Add SIBLING parameter. If it is non-zero, push
4019 the new label along with it onto the sibling_labels vector.
4020 (fill_simple_delay_slots): Adjust call to get_label_before.
4021 (fill_slots_from_thread): Likewise.
4022 (relax_delay_slots): Likewise.
4023 (make_return_insns): Likewise.
4024 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
4026 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4028 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
4029 document its semantics.
4030 (diagnostic_report_diagnostic): Adjust accordingly.
4032 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4034 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
4035 (sparc_option_override): ...and port to new C++ pass API.
4036 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
4038 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
4040 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
4042 2013-08-06 Caroline Tice <cmtice@google.com>
4044 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4045 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
4046 * tree-pass.h: Add pass_vtable_verify.
4047 * varasm.c (assemble_variable): Add code to properly set the comdat
4048 section and name for the .vtable_map_vars section.
4049 (assemble_vtyv_preinit_initializer): New function.
4050 (default_sectin_type_flags): Make sure .vtable_map_vars section has
4052 * output.h: Add function decl for assemble_vtv_preinit_initializer.
4053 * vtable-verify.c: New file.
4054 * vtable-verify.h: New file.
4055 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
4056 initialiation levels.
4057 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
4058 * passes.def: Insert pass_vtable_verify.
4059 * aclocal.m4: Reorder includes.
4060 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
4061 -fvtv-counts options.
4062 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
4063 as appropriate, if -fvtable-verify=... is used.
4064 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
4065 -fvtable-verify=... is used.
4066 * Makefile.in (OBJS): Add vtable-verify.o to list.
4067 (vtable-verify.o): Add new build rule.
4068 (GTFILES): Add vtable-verify.c to list.
4069 * common.opt (fvtable-verify=): New flag.
4070 (vtv_priority): Values for fvtable-verify= flag.
4071 (fvtv-counts): New flag.
4072 (fvtv-debug): New flag.
4073 * tree.h (save_vtable_map_decl): New extern function decl.
4075 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4077 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
4078 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
4079 (pass_data_rl78_devirt): ...new pass_data instance and...
4080 (make_pass_rl78_devirt): ...new function.
4081 (rl78_asm_file_start): Port pass registration to new C++ API.
4083 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4085 * coretypes.h (rtl_opt_pass): Add.
4086 (gcc::context): Add.
4087 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
4088 (epiphany_init): Port to new C++ pass API.
4089 (epiphany_optimize_mode_switching): Likewise.
4090 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
4091 (pass_manager::get_pass_mode_switching): New.
4092 (pass_manager::get_pass_peephole2): New.
4093 * mode-switching.c (pass_mode_switching): Add clone method.
4094 * recog.c (pass_peephole2): Add clone method.
4095 (pass_split_all_insns): Add clone method.
4097 2013-08-06 David Malcolm <dmalcolm@redhat.com>
4099 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
4100 (mips_option_override): ...here, porting to new C++ API for passes.
4102 2013-08-06 Jan Hubicka <jh@suse.cz>
4104 * cgraph.c (cgraph_get_body): New function based on lto.c
4106 * cgraph.h (cgraph_get_body): Declare.
4107 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
4109 * cgraphunit.c (expand_function): Get body prior expanding.
4110 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
4111 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
4113 * passes.c (do_per_function_toporder): Get body.
4114 * tree-inline.c (expand_call_inline): Get body prior inlining it.
4115 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
4117 2013-08-06 Martin Jambor <mjambor@suse.cz>
4120 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
4121 is not re-finalized. Rename second parameter to no_collect.
4123 2013-08-06 Martin Jambor <mjambor@suse.cz>
4126 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
4127 MEM_REF has proper alignment information.
4129 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
4132 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
4134 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
4135 rtx (*) (rtx, ...) with insn_gen_fn.
4136 * genoutput.c (output_insn_data): Cast gen_? function pointers to
4137 insn_gen_fn::stored_funcptr. Add initializer braces.
4139 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4141 Rewrite how instances of passes are cloned to remove assumptions
4142 about their sizes (thus allowing pass subclasses to have
4143 additional data fields, albeit non-GC-managed ones at this point).
4145 * passes.c (make_pass_instance): Now that passes have clone
4146 methods, rewrite this function to eliminate XNEW and memcpy
4147 calls that used hardcoded sizes. Since this function no longer
4148 creates pass instances, rename it to...
4149 (add_pass_instance): ...this. Document the old way that passes were
4150 numbered and flagged, and rework this function to continue using it.
4151 (next_pass_1): Add an initial_pass argument for use by
4153 (position_pass): When adding multiple instances of a pass, use
4154 the pass's clone method, rather than relying on the XNEW/memcpy
4155 within the former make_pass_instance (now add_pass_instance).
4156 (pass_manager::pass_manager): When invoking next_pass_1, also supply
4157 the initial instance of the current pass within the pass manager.
4159 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4161 This is the automated part of the conversion of passes from C
4162 structs to C++ classes.
4164 Patch autogenerated by refactor_passes.py from
4165 https://github.com/davidmalcolm/gcc-refactoring-scripts
4166 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
4168 * asan.c (pass_asan): Convert from a global struct to a subclass of
4169 gimple_opt_pass along with...
4170 (pass_data_asan): ...new pass_data instance and...
4171 (make_pass_asan): ...new function.
4172 (pass_asan_O0): Convert from a global struct to a subclass of
4173 gimple_opt_pass along with...
4174 (pass_data_asan_O0): ...new pass_data instance and...
4175 (make_pass_asan_O0): ...new function.
4176 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
4177 subclass of rtl_opt_pass along with...
4178 (pass_data_inc_dec): ...new pass_data instance and...
4179 (make_pass_inc_dec): ...new function.
4180 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
4181 a subclass of rtl_opt_pass along with...
4182 (pass_data_reorder_blocks): ...new pass_data instance and...
4183 (make_pass_reorder_blocks): ...new function.
4184 (pass_duplicate_computed_gotos): Convert from a global struct to a
4185 subclass of rtl_opt_pass along with...
4186 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
4187 (make_pass_duplicate_computed_gotos): ...new function.
4188 (pass_partition_blocks): Convert from a global struct to a subclass of
4189 rtl_opt_pass along with...
4190 (pass_data_partition_blocks): ...new pass_data instance and...
4191 (make_pass_partition_blocks): ...new function.
4192 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
4193 struct to a subclass of rtl_opt_pass along with...
4194 (pass_data_branch_target_load_optimize1): ...new pass_data instance
4196 (make_pass_branch_target_load_optimize1): ...new function.
4197 (pass_branch_target_load_optimize2): Convert from a global struct to a
4198 subclass of rtl_opt_pass along with...
4199 (pass_data_branch_target_load_optimize2): ...new pass_data instance
4201 (make_pass_branch_target_load_optimize2): ...new function.
4202 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
4203 of rtl_opt_pass along with...
4204 (pass_data_jump): ...new pass_data instance and...
4205 (make_pass_jump): ...new function.
4206 (pass_jump2): Convert from a global struct to a subclass of
4207 rtl_opt_pass along with...
4208 (pass_data_jump2): ...new pass_data instance and...
4209 (make_pass_jump2): ...new function.
4210 * cfgexpand.c (pass_expand): Convert from a global struct to a
4211 subclass of rtl_opt_pass along with...
4212 (pass_data_expand): ...new pass_data instance and...
4213 (make_pass_expand): ...new function.
4214 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
4215 of rtl_opt_pass along with...
4216 (pass_data_free_cfg): ...new pass_data instance and...
4217 (make_pass_free_cfg): ...new function.
4218 (pass_into_cfg_layout_mode): Convert from a global struct to a
4219 subclass of rtl_opt_pass along with...
4220 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
4221 (make_pass_into_cfg_layout_mode): ...new function.
4222 (pass_outof_cfg_layout_mode): Convert from a global struct to a
4223 subclass of rtl_opt_pass along with...
4224 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
4225 (make_pass_outof_cfg_layout_mode): ...new function.
4226 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
4227 struct to a subclass of gimple_opt_pass along with...
4228 (pass_data_build_cgraph_edges): ...new pass_data instance and...
4229 (make_pass_build_cgraph_edges): ...new function.
4230 (pass_rebuild_cgraph_edges): Convert from a global struct to a
4231 subclass of gimple_opt_pass along with...
4232 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
4233 (make_pass_rebuild_cgraph_edges): ...new function.
4234 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
4235 subclass of gimple_opt_pass along with...
4236 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
4238 (make_pass_remove_cgraph_callee_edges): ...new function.
4239 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
4240 struct to a subclass of rtl_opt_pass along with...
4241 (pass_data_stack_adjustments): ...new pass_data instance and...
4242 (make_pass_stack_adjustments): ...new function.
4243 * combine.c (pass_combine): Convert from a global struct to a subclass
4244 of rtl_opt_pass along with...
4245 (pass_data_combine): ...new pass_data instance and...
4246 (make_pass_combine): ...new function.
4247 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
4248 global struct to a subclass of rtl_opt_pass along with...
4249 (pass_data_compare_elim_after_reload): ...new pass_data instance
4251 (make_pass_compare_elim_after_reload): ...new function.
4252 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
4253 of rtl_opt_pass along with...
4254 (pass_data_rtl_cprop): ...new pass_data instance and...
4255 (make_pass_rtl_cprop): ...new function.
4256 * cse.c (pass_cse): Convert from a global struct to a subclass of
4257 rtl_opt_pass along with...
4258 (pass_data_cse): ...new pass_data instance and...
4259 (make_pass_cse): ...new function.
4260 (pass_cse2): Convert from a global struct to a subclass of
4261 rtl_opt_pass along with...
4262 (pass_data_cse2): ...new pass_data instance and...
4263 (make_pass_cse2): ...new function.
4264 (pass_cse_after_global_opts): Convert from a global struct to a
4265 subclass of rtl_opt_pass along with...
4266 (pass_data_cse_after_global_opts): ...new pass_data instance and...
4267 (make_pass_cse_after_global_opts): ...new function.
4268 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
4269 of rtl_opt_pass along with...
4270 (pass_data_ud_rtl_dce): ...new pass_data instance and...
4271 (make_pass_ud_rtl_dce): ...new function.
4272 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
4273 rtl_opt_pass along with...
4274 (pass_data_fast_rtl_dce): ...new pass_data instance and...
4275 (make_pass_fast_rtl_dce): ...new function.
4276 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
4277 a subclass of rtl_opt_pass along with...
4278 (pass_data_df_initialize_opt): ...new pass_data instance and...
4279 (make_pass_df_initialize_opt): ...new function.
4280 (pass_df_initialize_no_opt): Convert from a global struct to a
4281 subclass of rtl_opt_pass along with...
4282 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
4283 (make_pass_df_initialize_no_opt): ...new function.
4284 (pass_df_finish): Convert from a global struct to a subclass of
4285 rtl_opt_pass along with...
4286 (pass_data_df_finish): ...new pass_data instance and...
4287 (make_pass_df_finish): ...new function.
4288 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
4289 rtl_opt_pass along with...
4290 (pass_data_rtl_dse1): ...new pass_data instance and...
4291 (make_pass_rtl_dse1): ...new function.
4292 (pass_rtl_dse2): Convert from a global struct to a subclass of
4293 rtl_opt_pass along with...
4294 (pass_data_rtl_dse2): ...new pass_data instance and...
4295 (make_pass_rtl_dse2): ...new function.
4296 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
4297 subclass of rtl_opt_pass along with...
4298 (pass_data_dwarf2_frame): ...new pass_data instance and...
4299 (make_pass_dwarf2_frame): ...new function.
4300 * except.c (pass_set_nothrow_function_flags): Convert from a global
4301 struct to a subclass of rtl_opt_pass along with...
4302 (pass_data_set_nothrow_function_flags): ...new pass_data instance
4304 (make_pass_set_nothrow_function_flags): ...new function.
4305 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
4306 subclass of rtl_opt_pass along with...
4307 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
4309 (make_pass_convert_to_eh_region_ranges): ...new function.
4310 * final.c (pass_compute_alignments): Convert from a global struct to a
4311 subclass of rtl_opt_pass along with...
4312 (pass_data_compute_alignments): ...new pass_data instance and...
4313 (make_pass_compute_alignments): ...new function.
4314 (pass_final): Convert from a global struct to a subclass of
4315 rtl_opt_pass along with...
4316 (pass_data_final): ...new pass_data instance and...
4317 (make_pass_final): ...new function.
4318 (pass_shorten_branches): Convert from a global struct to a subclass of
4319 rtl_opt_pass along with...
4320 (pass_data_shorten_branches): ...new pass_data instance and...
4321 (make_pass_shorten_branches): ...new function.
4322 (pass_clean_state): Convert from a global struct to a subclass of
4323 rtl_opt_pass along with...
4324 (pass_data_clean_state): ...new pass_data instance and...
4325 (make_pass_clean_state): ...new function.
4326 * function.c (pass_instantiate_virtual_regs): Convert from a global
4327 struct to a subclass of rtl_opt_pass along with...
4328 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
4329 (make_pass_instantiate_virtual_regs): ...new function.
4330 (pass_leaf_regs): Convert from a global struct to a subclass of
4331 rtl_opt_pass along with...
4332 (pass_data_leaf_regs): ...new pass_data instance and...
4333 (make_pass_leaf_regs): ...new function.
4334 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
4335 subclass of rtl_opt_pass along with...
4336 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
4338 (make_pass_thread_prologue_and_epilogue): ...new function.
4339 (pass_match_asm_constraints): Convert from a global struct to a
4340 subclass of rtl_opt_pass along with...
4341 (pass_data_match_asm_constraints): ...new pass_data instance and...
4342 (make_pass_match_asm_constraints): ...new function.
4343 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
4344 subclass of rtl_opt_pass along with...
4345 (pass_data_rtl_fwprop): ...new pass_data instance and...
4346 (make_pass_rtl_fwprop): ...new function.
4347 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
4348 rtl_opt_pass along with...
4349 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
4350 (make_pass_rtl_fwprop_addr): ...new function.
4351 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
4352 rtl_opt_pass along with...
4353 (pass_data_rtl_pre): ...new pass_data instance and...
4354 (make_pass_rtl_pre): ...new function.
4355 (pass_rtl_hoist): Convert from a global struct to a subclass of
4356 rtl_opt_pass along with...
4357 (pass_data_rtl_hoist): ...new pass_data instance and...
4358 (make_pass_rtl_hoist): ...new function.
4359 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
4360 subclass of gimple_opt_pass along with...
4361 (pass_data_lower_cf): ...new pass_data instance and...
4362 (make_pass_lower_cf): ...new function.
4363 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
4364 from a global struct to a subclass of gimple_opt_pass along with...
4365 (pass_data_strength_reduction): ...new pass_data instance and...
4366 (make_pass_strength_reduction): ...new function.
4367 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
4368 of rtl_opt_pass along with...
4369 (pass_data_rtl_ifcvt): ...new pass_data instance and...
4370 (make_pass_rtl_ifcvt): ...new function.
4371 (pass_if_after_combine): Convert from a global struct to a subclass of
4372 rtl_opt_pass along with...
4373 (pass_data_if_after_combine): ...new pass_data instance and...
4374 (make_pass_if_after_combine): ...new function.
4375 (pass_if_after_reload): Convert from a global struct to a subclass of
4376 rtl_opt_pass along with...
4377 (pass_data_if_after_reload): ...new pass_data instance and...
4378 (make_pass_if_after_reload): ...new function.
4379 * init-regs.c (pass_initialize_regs): Convert from a global struct to
4380 a subclass of rtl_opt_pass along with...
4381 (pass_data_initialize_regs): ...new pass_data instance and...
4382 (make_pass_initialize_regs): ...new function.
4383 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
4384 of ipa_opt_pass_d along with...
4385 (pass_data_ipa_cp): ...new pass_data instance and...
4386 (make_pass_ipa_cp): ...new function.
4387 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
4388 global struct to a subclass of gimple_opt_pass along with...
4389 (pass_data_inline_parameters): ...new pass_data instance and...
4390 (make_pass_inline_parameters): ...new function.
4391 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
4392 subclass of gimple_opt_pass along with...
4393 (pass_data_early_inline): ...new pass_data instance and...
4394 (make_pass_early_inline): ...new function.
4395 (pass_ipa_inline): Convert from a global struct to a subclass of
4396 ipa_opt_pass_d along with...
4397 (pass_data_ipa_inline): ...new pass_data instance and...
4398 (make_pass_ipa_inline): ...new function.
4399 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
4400 struct to a subclass of gimple_opt_pass along with...
4401 (pass_data_local_pure_const): ...new pass_data instance and...
4402 (make_pass_local_pure_const): ...new function.
4403 (pass_ipa_pure_const): Convert from a global struct to a subclass of
4404 ipa_opt_pass_d along with...
4405 (pass_data_ipa_pure_const): ...new pass_data instance and...
4406 (make_pass_ipa_pure_const): ...new function.
4407 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
4408 to a subclass of ipa_opt_pass_d along with...
4409 (pass_data_ipa_reference): ...new pass_data instance and...
4410 (make_pass_ipa_reference): ...new function.
4411 * ipa-split.c (pass_split_functions): Convert from a global struct to
4412 a subclass of gimple_opt_pass along with...
4413 (pass_data_split_functions): ...new pass_data instance and...
4414 (make_pass_split_functions): ...new function.
4415 (pass_feedback_split_functions): Convert from a global struct to a
4416 subclass of gimple_opt_pass along with...
4417 (pass_data_feedback_split_functions): ...new pass_data instance and...
4418 (make_pass_feedback_split_functions): ...new function.
4419 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
4420 global struct to a subclass of simple_ipa_opt_pass along with...
4421 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
4423 (make_pass_ipa_function_and_variable_visibility): ...new function.
4424 (pass_ipa_free_inline_summary): Convert from a global struct to a
4425 subclass of simple_ipa_opt_pass along with...
4426 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
4427 (make_pass_ipa_free_inline_summary): ...new function.
4428 (pass_ipa_whole_program_visibility): Convert from a global struct to a
4429 subclass of ipa_opt_pass_d along with...
4430 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
4432 (make_pass_ipa_whole_program_visibility): ...new function.
4433 (pass_ipa_profile): Convert from a global struct to a subclass of
4434 ipa_opt_pass_d along with...
4435 (pass_data_ipa_profile): ...new pass_data instance and...
4436 (make_pass_ipa_profile): ...new function.
4437 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
4438 ipa_opt_pass_d along with...
4439 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4440 (make_pass_ipa_cdtor_merge): ...new function.
4441 * ira.c (pass_ira): Convert from a global struct to a subclass of
4442 rtl_opt_pass along with...
4443 (pass_data_ira): ...new pass_data instance and...
4444 (make_pass_ira): ...new function.
4445 (pass_reload): Convert from a global struct to a subclass of
4446 rtl_opt_pass along with...
4447 (pass_data_reload): ...new pass_data instance and...
4448 (make_pass_reload): ...new function.
4449 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4450 subclass of rtl_opt_pass along with...
4451 (pass_data_cleanup_barriers): ...new pass_data instance and...
4452 (make_pass_cleanup_barriers): ...new function.
4453 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4454 of rtl_opt_pass along with...
4455 (pass_data_loop2): ...new pass_data instance and...
4456 (make_pass_loop2): ...new function.
4457 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4458 rtl_opt_pass along with...
4459 (pass_data_rtl_loop_init): ...new pass_data instance and...
4460 (make_pass_rtl_loop_init): ...new function.
4461 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4462 rtl_opt_pass along with...
4463 (pass_data_rtl_loop_done): ...new pass_data instance and...
4464 (make_pass_rtl_loop_done): ...new function.
4465 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4466 subclass of rtl_opt_pass along with...
4467 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4468 (make_pass_rtl_move_loop_invariants): ...new function.
4469 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4470 rtl_opt_pass along with...
4471 (pass_data_rtl_unswitch): ...new pass_data instance and...
4472 (make_pass_rtl_unswitch): ...new function.
4473 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4474 subclass of rtl_opt_pass along with...
4475 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4477 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4478 (pass_rtl_doloop): Convert from a global struct to a subclass of
4479 rtl_opt_pass along with...
4480 (pass_data_rtl_doloop): ...new pass_data instance and...
4481 (make_pass_rtl_doloop): ...new function.
4482 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4483 a subclass of rtl_opt_pass along with...
4484 (pass_data_lower_subreg): ...new pass_data instance and...
4485 (make_pass_lower_subreg): ...new function.
4486 (pass_lower_subreg2): Convert from a global struct to a subclass of
4487 rtl_opt_pass along with...
4488 (pass_data_lower_subreg2): ...new pass_data instance and...
4489 (make_pass_lower_subreg2): ...new function.
4490 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4491 struct to a subclass of ipa_opt_pass_d along with...
4492 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4493 (make_pass_ipa_lto_gimple_out): ...new function.
4494 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4495 of ipa_opt_pass_d along with...
4496 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4497 (make_pass_ipa_lto_finish_out): ...new function.
4498 * mode-switching.c (pass_mode_switching): Convert from a global struct
4499 to a subclass of rtl_opt_pass along with...
4500 (pass_data_mode_switching): ...new pass_data instance and...
4501 (make_pass_mode_switching): ...new function.
4502 * modulo-sched.c (pass_sms): Convert from a global struct to a
4503 subclass of rtl_opt_pass along with...
4504 (pass_data_sms): ...new pass_data instance and...
4505 (make_pass_sms): ...new function.
4506 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4507 subclass of gimple_opt_pass along with...
4508 (pass_data_expand_omp): ...new pass_data instance and...
4509 (make_pass_expand_omp): ...new function.
4510 (pass_lower_omp): Convert from a global struct to a subclass of
4511 gimple_opt_pass along with...
4512 (pass_data_lower_omp): ...new pass_data instance and...
4513 (make_pass_lower_omp): ...new function.
4514 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4515 of gimple_opt_pass along with...
4516 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4517 (make_pass_diagnose_omp_blocks): ...new function.
4518 * passes.c (pass_early_local_passes): Convert from a global struct to
4519 a subclass of simple_ipa_opt_pass along with...
4520 (pass_data_early_local_passes): ...new pass_data instance and...
4521 (make_pass_early_local_passes): ...new function.
4522 (pass_all_early_optimizations): Convert from a global struct to a
4523 subclass of gimple_opt_pass along with...
4524 (pass_data_all_early_optimizations): ...new pass_data instance and...
4525 (make_pass_all_early_optimizations): ...new function.
4526 (pass_all_optimizations): Convert from a global struct to a subclass
4527 of gimple_opt_pass along with...
4528 (pass_data_all_optimizations): ...new pass_data instance and...
4529 (make_pass_all_optimizations): ...new function.
4530 (pass_all_optimizations_g): Convert from a global struct to a subclass
4531 of gimple_opt_pass along with...
4532 (pass_data_all_optimizations_g): ...new pass_data instance and...
4533 (make_pass_all_optimizations_g): ...new function.
4534 (pass_rest_of_compilation): Convert from a global struct to a subclass
4535 of rtl_opt_pass along with...
4536 (pass_data_rest_of_compilation): ...new pass_data instance and...
4537 (make_pass_rest_of_compilation): ...new function.
4538 (pass_postreload): Convert from a global struct to a subclass of
4539 rtl_opt_pass along with...
4540 (pass_data_postreload): ...new pass_data instance and...
4541 (make_pass_postreload): ...new function.
4542 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
4543 subclass of rtl_opt_pass along with...
4544 (pass_data_gcse2): ...new pass_data instance and...
4545 (make_pass_gcse2): ...new function.
4546 * postreload.c (pass_postreload_cse): Convert from a global struct to
4547 a subclass of rtl_opt_pass along with...
4548 (pass_data_postreload_cse): ...new pass_data instance and...
4549 (make_pass_postreload_cse): ...new function.
4550 * predict.c (pass_profile): Convert from a global struct to a subclass
4551 of gimple_opt_pass along with...
4552 (pass_data_profile): ...new pass_data instance and...
4553 (make_pass_profile): ...new function.
4554 (pass_strip_predict_hints): Convert from a global struct to a subclass
4555 of gimple_opt_pass along with...
4556 (pass_data_strip_predict_hints): ...new pass_data instance and...
4557 (make_pass_strip_predict_hints): ...new function.
4558 * recog.c (pass_peephole2): Convert from a global struct to a subclass
4559 of rtl_opt_pass along with...
4560 (pass_data_peephole2): ...new pass_data instance and...
4561 (make_pass_peephole2): ...new function.
4562 (pass_split_all_insns): Convert from a global struct to a subclass of
4563 rtl_opt_pass along with...
4564 (pass_data_split_all_insns): ...new pass_data instance and...
4565 (make_pass_split_all_insns): ...new function.
4566 (pass_split_after_reload): Convert from a global struct to a subclass
4567 of rtl_opt_pass along with...
4568 (pass_data_split_after_reload): ...new pass_data instance and...
4569 (make_pass_split_after_reload): ...new function.
4570 (pass_split_before_regstack): Convert from a global struct to a
4571 subclass of rtl_opt_pass along with...
4572 (pass_data_split_before_regstack): ...new pass_data instance and...
4573 (make_pass_split_before_regstack): ...new function.
4574 (pass_split_before_sched2): Convert from a global struct to a subclass
4575 of rtl_opt_pass along with...
4576 (pass_data_split_before_sched2): ...new pass_data instance and...
4577 (make_pass_split_before_sched2): ...new function.
4578 (pass_split_for_shorten_branches): Convert from a global struct to a
4579 subclass of rtl_opt_pass along with...
4580 (pass_data_split_for_shorten_branches): ...new pass_data instance
4582 (make_pass_split_for_shorten_branches): ...new function.
4583 * ree.c (pass_ree): Convert from a global struct to a subclass of
4584 rtl_opt_pass along with...
4585 (pass_data_ree): ...new pass_data instance and...
4586 (make_pass_ree): ...new function.
4587 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
4588 subclass of rtl_opt_pass along with...
4589 (pass_data_stack_regs): ...new pass_data instance and...
4590 (make_pass_stack_regs): ...new function.
4591 (pass_stack_regs_run): Convert from a global struct to a subclass of
4592 rtl_opt_pass along with...
4593 (pass_data_stack_regs_run): ...new pass_data instance and...
4594 (make_pass_stack_regs_run): ...new function.
4595 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4596 subclass of rtl_opt_pass along with...
4597 (pass_data_cprop_hardreg): ...new pass_data instance and...
4598 (make_pass_cprop_hardreg): ...new function.
4599 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4600 subclass of rtl_opt_pass along with...
4601 (pass_data_reginfo_init): ...new pass_data instance and...
4602 (make_pass_reginfo_init): ...new function.
4603 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4604 of rtl_opt_pass along with...
4605 (pass_data_regmove): ...new pass_data instance and...
4606 (make_pass_regmove): ...new function.
4607 * regrename.c (pass_regrename): Convert from a global struct to a
4608 subclass of rtl_opt_pass along with...
4609 (pass_data_regrename): ...new pass_data instance and...
4610 (make_pass_regrename): ...new function.
4611 * reorg.c (pass_delay_slots): Convert from a global struct to a
4612 subclass of rtl_opt_pass along with...
4613 (pass_data_delay_slots): ...new pass_data instance and...
4614 (make_pass_delay_slots): ...new function.
4615 (pass_machine_reorg): Convert from a global struct to a subclass of
4616 rtl_opt_pass along with...
4617 (pass_data_machine_reorg): ...new pass_data instance and...
4618 (make_pass_machine_reorg): ...new function.
4619 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4620 of rtl_opt_pass along with...
4621 (pass_data_sched): ...new pass_data instance and...
4622 (make_pass_sched): ...new function.
4623 (pass_sched2): Convert from a global struct to a subclass of
4624 rtl_opt_pass along with...
4625 (pass_data_sched2): ...new pass_data instance and...
4626 (make_pass_sched2): ...new function.
4627 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4628 to a subclass of rtl_opt_pass along with...
4629 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4630 (make_pass_stack_ptr_mod): ...new function.
4631 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4632 to a subclass of rtl_opt_pass along with...
4633 (pass_data_rtl_store_motion): ...new pass_data instance and...
4634 (make_pass_rtl_store_motion): ...new function.
4635 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4636 of gimple_opt_pass along with...
4637 (pass_data_tracer): ...new pass_data instance and...
4638 (make_pass_tracer): ...new function.
4639 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4640 to a subclass of gimple_opt_pass along with...
4641 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4642 (make_pass_diagnose_tm_blocks): ...new function.
4643 (pass_lower_tm): Convert from a global struct to a subclass of
4644 gimple_opt_pass along with...
4645 (pass_data_lower_tm): ...new pass_data instance and...
4646 (make_pass_lower_tm): ...new function.
4647 (pass_tm_init): Convert from a global struct to a subclass of
4648 gimple_opt_pass along with...
4649 (pass_data_tm_init): ...new pass_data instance and...
4650 (make_pass_tm_init): ...new function.
4651 (pass_tm_mark): Convert from a global struct to a subclass of
4652 gimple_opt_pass along with...
4653 (pass_data_tm_mark): ...new pass_data instance and...
4654 (make_pass_tm_mark): ...new function.
4655 (pass_tm_edges): Convert from a global struct to a subclass of
4656 gimple_opt_pass along with...
4657 (pass_data_tm_edges): ...new pass_data instance and...
4658 (make_pass_tm_edges): ...new function.
4659 (pass_tm_memopt): Convert from a global struct to a subclass of
4660 gimple_opt_pass along with...
4661 (pass_data_tm_memopt): ...new pass_data instance and...
4662 (make_pass_tm_memopt): ...new function.
4663 (pass_ipa_tm): Convert from a global struct to a subclass of
4664 simple_ipa_opt_pass along with...
4665 (pass_data_ipa_tm): ...new pass_data instance and...
4666 (make_pass_ipa_tm): ...new function.
4667 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4668 subclass of gimple_opt_pass along with...
4669 (pass_data_call_cdce): ...new pass_data instance and...
4670 (make_pass_call_cdce): ...new function.
4671 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4672 subclass of gimple_opt_pass along with...
4673 (pass_data_build_cfg): ...new pass_data instance and...
4674 (make_pass_build_cfg): ...new function.
4675 (pass_split_crit_edges): Convert from a global struct to a subclass of
4676 gimple_opt_pass along with...
4677 (pass_data_split_crit_edges): ...new pass_data instance and...
4678 (make_pass_split_crit_edges): ...new function.
4679 (pass_warn_function_return): Convert from a global struct to a
4680 subclass of gimple_opt_pass along with...
4681 (pass_data_warn_function_return): ...new pass_data instance and...
4682 (make_pass_warn_function_return): ...new function.
4683 (pass_warn_function_noreturn): Convert from a global struct to a
4684 subclass of gimple_opt_pass along with...
4685 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4686 (make_pass_warn_function_noreturn): ...new function.
4687 (pass_warn_unused_result): Convert from a global struct to a subclass
4688 of gimple_opt_pass along with...
4689 (pass_data_warn_unused_result): ...new pass_data instance and...
4690 (make_pass_warn_unused_result): ...new function.
4691 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4692 a subclass of gimple_opt_pass along with...
4693 (pass_data_merge_phi): ...new pass_data instance and...
4694 (make_pass_merge_phi): ...new function.
4695 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4696 a subclass of gimple_opt_pass along with...
4697 (pass_data_lower_complex): ...new pass_data instance and...
4698 (make_pass_lower_complex): ...new function.
4699 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4700 gimple_opt_pass along with...
4701 (pass_data_lower_complex_O0): ...new pass_data instance and...
4702 (make_pass_lower_complex_O0): ...new function.
4703 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4704 subclass of gimple_opt_pass along with...
4705 (pass_data_lower_eh): ...new pass_data instance and...
4706 (make_pass_lower_eh): ...new function.
4707 (pass_refactor_eh): Convert from a global struct to a subclass of
4708 gimple_opt_pass along with...
4709 (pass_data_refactor_eh): ...new pass_data instance and...
4710 (make_pass_refactor_eh): ...new function.
4711 (pass_lower_resx): Convert from a global struct to a subclass of
4712 gimple_opt_pass along with...
4713 (pass_data_lower_resx): ...new pass_data instance and...
4714 (make_pass_lower_resx): ...new function.
4715 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4716 of gimple_opt_pass along with...
4717 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4718 (make_pass_lower_eh_dispatch): ...new function.
4719 (pass_cleanup_eh): Convert from a global struct to a subclass of
4720 gimple_opt_pass along with...
4721 (pass_data_cleanup_eh): ...new pass_data instance and...
4722 (make_pass_cleanup_eh): ...new function.
4723 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4724 to a subclass of simple_ipa_opt_pass along with...
4725 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4726 (make_pass_ipa_lower_emutls): ...new function.
4727 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4728 a subclass of gimple_opt_pass along with...
4729 (pass_data_if_conversion): ...new pass_data instance and...
4730 (make_pass_if_conversion): ...new function.
4731 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4732 subclass of gimple_opt_pass along with...
4733 (pass_data_build_ssa): ...new pass_data instance and...
4734 (make_pass_build_ssa): ...new function.
4735 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4736 global struct to a subclass of gimple_opt_pass along with...
4737 (pass_data_loop_distribution): ...new pass_data instance and...
4738 (make_pass_loop_distribution): ...new function.
4739 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4740 subclass of gimple_opt_pass along with...
4741 (pass_data_mudflap_1): ...new pass_data instance and...
4742 (make_pass_mudflap_1): ...new function.
4743 (pass_mudflap_2): Convert from a global struct to a subclass of
4744 gimple_opt_pass along with...
4745 (pass_data_mudflap_2): ...new pass_data instance and...
4746 (make_pass_mudflap_2): ...new function.
4747 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4748 subclass of gimple_opt_pass along with...
4749 (pass_data_mudflap_1): ...new pass_data instance and...
4750 (make_pass_mudflap_1): ...new function.
4751 (pass_mudflap_2): Convert from a global struct to a subclass of
4752 gimple_opt_pass along with...
4753 (pass_data_mudflap_2): ...new pass_data instance and...
4754 (make_pass_mudflap_2): ...new function.
4755 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4756 gimple_opt_pass along with...
4757 (pass_data_nrv): ...new pass_data instance and...
4758 (make_pass_nrv): ...new function.
4759 (pass_return_slot): Convert from a global struct to a subclass of
4760 gimple_opt_pass along with...
4761 (pass_data_return_slot): ...new pass_data instance and...
4762 (make_pass_return_slot): ...new function.
4763 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4764 to a subclass of gimple_opt_pass along with...
4765 (pass_data_object_sizes): ...new pass_data instance and...
4766 (make_pass_object_sizes): ...new function.
4767 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4768 global struct to a subclass of gimple_opt_pass along with...
4769 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4771 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4772 (pass_fixup_cfg): Convert from a global struct to a subclass of
4773 gimple_opt_pass along with...
4774 (pass_data_fixup_cfg): ...new pass_data instance and...
4775 (make_pass_fixup_cfg): ...new function.
4776 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4777 (make_pass_mudflap_1): ...new function.
4778 (pass_mudflap_2): Replace declaration with that of...
4779 (make_pass_mudflap_2): ...new function.
4780 (pass_asan): Replace declaration with that of...
4781 (make_pass_asan): ...new function.
4782 (pass_asan_O0): Replace declaration with that of...
4783 (make_pass_asan_O0): ...new function.
4784 (pass_tsan): Replace declaration with that of...
4785 (make_pass_tsan): ...new function.
4786 (pass_tsan_O0): Replace declaration with that of...
4787 (make_pass_tsan_O0): ...new function.
4788 (pass_lower_cf): Replace declaration with that of...
4789 (make_pass_lower_cf): ...new function.
4790 (pass_refactor_eh): Replace declaration with that of...
4791 (make_pass_refactor_eh): ...new function.
4792 (pass_lower_eh): Replace declaration with that of...
4793 (make_pass_lower_eh): ...new function.
4794 (pass_lower_eh_dispatch): Replace declaration with that of...
4795 (make_pass_lower_eh_dispatch): ...new function.
4796 (pass_lower_resx): Replace declaration with that of...
4797 (make_pass_lower_resx): ...new function.
4798 (pass_build_cfg): Replace declaration with that of...
4799 (make_pass_build_cfg): ...new function.
4800 (pass_early_tree_profile): Replace declaration with that of...
4801 (make_pass_early_tree_profile): ...new function.
4802 (pass_cleanup_eh): Replace declaration with that of...
4803 (make_pass_cleanup_eh): ...new function.
4804 (pass_sra): Replace declaration with that of...
4805 (make_pass_sra): ...new function.
4806 (pass_sra_early): Replace declaration with that of...
4807 (make_pass_sra_early): ...new function.
4808 (pass_early_ipa_sra): Replace declaration with that of...
4809 (make_pass_early_ipa_sra): ...new function.
4810 (pass_tail_recursion): Replace declaration with that of...
4811 (make_pass_tail_recursion): ...new function.
4812 (pass_tail_calls): Replace declaration with that of...
4813 (make_pass_tail_calls): ...new function.
4814 (pass_tree_loop): Replace declaration with that of...
4815 (make_pass_tree_loop): ...new function.
4816 (pass_tree_loop_init): Replace declaration with that of...
4817 (make_pass_tree_loop_init): ...new function.
4818 (pass_lim): Replace declaration with that of...
4819 (make_pass_lim): ...new function.
4820 (pass_tree_unswitch): Replace declaration with that of...
4821 (make_pass_tree_unswitch): ...new function.
4822 (pass_predcom): Replace declaration with that of...
4823 (make_pass_predcom): ...new function.
4824 (pass_iv_canon): Replace declaration with that of...
4825 (make_pass_iv_canon): ...new function.
4826 (pass_scev_cprop): Replace declaration with that of...
4827 (make_pass_scev_cprop): ...new function.
4828 (pass_empty_loop): Replace declaration with that of...
4829 (make_pass_empty_loop): ...new function.
4830 (pass_record_bounds): Replace declaration with that of...
4831 (make_pass_record_bounds): ...new function.
4832 (pass_graphite): Replace declaration with that of...
4833 (make_pass_graphite): ...new function.
4834 (pass_graphite_transforms): Replace declaration with that of...
4835 (make_pass_graphite_transforms): ...new function.
4836 (pass_if_conversion): Replace declaration with that of...
4837 (make_pass_if_conversion): ...new function.
4838 (pass_loop_distribution): Replace declaration with that of...
4839 (make_pass_loop_distribution): ...new function.
4840 (pass_vectorize): Replace declaration with that of...
4841 (make_pass_vectorize): ...new function.
4842 (pass_slp_vectorize): Replace declaration with that of...
4843 (make_pass_slp_vectorize): ...new function.
4844 (pass_complete_unroll): Replace declaration with that of...
4845 (make_pass_complete_unroll): ...new function.
4846 (pass_complete_unrolli): Replace declaration with that of...
4847 (make_pass_complete_unrolli): ...new function.
4848 (pass_parallelize_loops): Replace declaration with that of...
4849 (make_pass_parallelize_loops): ...new function.
4850 (pass_loop_prefetch): Replace declaration with that of...
4851 (make_pass_loop_prefetch): ...new function.
4852 (pass_iv_optimize): Replace declaration with that of...
4853 (make_pass_iv_optimize): ...new function.
4854 (pass_tree_loop_done): Replace declaration with that of...
4855 (make_pass_tree_loop_done): ...new function.
4856 (pass_ch): Replace declaration with that of...
4857 (make_pass_ch): ...new function.
4858 (pass_ccp): Replace declaration with that of...
4859 (make_pass_ccp): ...new function.
4860 (pass_phi_only_cprop): Replace declaration with that of...
4861 (make_pass_phi_only_cprop): ...new function.
4862 (pass_build_ssa): Replace declaration with that of...
4863 (make_pass_build_ssa): ...new function.
4864 (pass_build_alias): Replace declaration with that of...
4865 (make_pass_build_alias): ...new function.
4866 (pass_build_ealias): Replace declaration with that of...
4867 (make_pass_build_ealias): ...new function.
4868 (pass_dominator): Replace declaration with that of...
4869 (make_pass_dominator): ...new function.
4870 (pass_dce): Replace declaration with that of...
4871 (make_pass_dce): ...new function.
4872 (pass_dce_loop): Replace declaration with that of...
4873 (make_pass_dce_loop): ...new function.
4874 (pass_cd_dce): Replace declaration with that of...
4875 (make_pass_cd_dce): ...new function.
4876 (pass_call_cdce): Replace declaration with that of...
4877 (make_pass_call_cdce): ...new function.
4878 (pass_merge_phi): Replace declaration with that of...
4879 (make_pass_merge_phi): ...new function.
4880 (pass_split_crit_edges): Replace declaration with that of...
4881 (make_pass_split_crit_edges): ...new function.
4882 (pass_pre): Replace declaration with that of...
4883 (make_pass_pre): ...new function.
4884 (pass_profile): Replace declaration with that of...
4885 (make_pass_profile): ...new function.
4886 (pass_strip_predict_hints): Replace declaration with that of...
4887 (make_pass_strip_predict_hints): ...new function.
4888 (pass_lower_complex_O0): Replace declaration with that of...
4889 (make_pass_lower_complex_O0): ...new function.
4890 (pass_lower_complex): Replace declaration with that of...
4891 (make_pass_lower_complex): ...new function.
4892 (pass_lower_vector): Replace declaration with that of...
4893 (make_pass_lower_vector): ...new function.
4894 (pass_lower_vector_ssa): Replace declaration with that of...
4895 (make_pass_lower_vector_ssa): ...new function.
4896 (pass_lower_omp): Replace declaration with that of...
4897 (make_pass_lower_omp): ...new function.
4898 (pass_diagnose_omp_blocks): Replace declaration with that of...
4899 (make_pass_diagnose_omp_blocks): ...new function.
4900 (pass_expand_omp): Replace declaration with that of...
4901 (make_pass_expand_omp): ...new function.
4902 (pass_expand_omp_ssa): Replace declaration with that of...
4903 (make_pass_expand_omp_ssa): ...new function.
4904 (pass_object_sizes): Replace declaration with that of...
4905 (make_pass_object_sizes): ...new function.
4906 (pass_strlen): Replace declaration with that of...
4907 (make_pass_strlen): ...new function.
4908 (pass_fold_builtins): Replace declaration with that of...
4909 (make_pass_fold_builtins): ...new function.
4910 (pass_stdarg): Replace declaration with that of...
4911 (make_pass_stdarg): ...new function.
4912 (pass_early_warn_uninitialized): Replace declaration with that of...
4913 (make_pass_early_warn_uninitialized): ...new function.
4914 (pass_late_warn_uninitialized): Replace declaration with that of...
4915 (make_pass_late_warn_uninitialized): ...new function.
4916 (pass_cse_reciprocals): Replace declaration with that of...
4917 (make_pass_cse_reciprocals): ...new function.
4918 (pass_cse_sincos): Replace declaration with that of...
4919 (make_pass_cse_sincos): ...new function.
4920 (pass_optimize_bswap): Replace declaration with that of...
4921 (make_pass_optimize_bswap): ...new function.
4922 (pass_optimize_widening_mul): Replace declaration with that of...
4923 (make_pass_optimize_widening_mul): ...new function.
4924 (pass_warn_function_return): Replace declaration with that of...
4925 (make_pass_warn_function_return): ...new function.
4926 (pass_warn_function_noreturn): Replace declaration with that of...
4927 (make_pass_warn_function_noreturn): ...new function.
4928 (pass_cselim): Replace declaration with that of...
4929 (make_pass_cselim): ...new function.
4930 (pass_phiopt): Replace declaration with that of...
4931 (make_pass_phiopt): ...new function.
4932 (pass_forwprop): Replace declaration with that of...
4933 (make_pass_forwprop): ...new function.
4934 (pass_phiprop): Replace declaration with that of...
4935 (make_pass_phiprop): ...new function.
4936 (pass_tree_ifcombine): Replace declaration with that of...
4937 (make_pass_tree_ifcombine): ...new function.
4938 (pass_dse): Replace declaration with that of...
4939 (make_pass_dse): ...new function.
4940 (pass_nrv): Replace declaration with that of...
4941 (make_pass_nrv): ...new function.
4942 (pass_rename_ssa_copies): Replace declaration with that of...
4943 (make_pass_rename_ssa_copies): ...new function.
4944 (pass_sink_code): Replace declaration with that of...
4945 (make_pass_sink_code): ...new function.
4946 (pass_fre): Replace declaration with that of...
4947 (make_pass_fre): ...new function.
4948 (pass_check_data_deps): Replace declaration with that of...
4949 (make_pass_check_data_deps): ...new function.
4950 (pass_copy_prop): Replace declaration with that of...
4951 (make_pass_copy_prop): ...new function.
4952 (pass_vrp): Replace declaration with that of...
4953 (make_pass_vrp): ...new function.
4954 (pass_uncprop): Replace declaration with that of...
4955 (make_pass_uncprop): ...new function.
4956 (pass_return_slot): Replace declaration with that of...
4957 (make_pass_return_slot): ...new function.
4958 (pass_reassoc): Replace declaration with that of...
4959 (make_pass_reassoc): ...new function.
4960 (pass_rebuild_cgraph_edges): Replace declaration with that of...
4961 (make_pass_rebuild_cgraph_edges): ...new function.
4962 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
4963 (make_pass_remove_cgraph_callee_edges): ...new function.
4964 (pass_build_cgraph_edges): Replace declaration with that of...
4965 (make_pass_build_cgraph_edges): ...new function.
4966 (pass_local_pure_const): Replace declaration with that of...
4967 (make_pass_local_pure_const): ...new function.
4968 (pass_tracer): Replace declaration with that of...
4969 (make_pass_tracer): ...new function.
4970 (pass_warn_unused_result): Replace declaration with that of...
4971 (make_pass_warn_unused_result): ...new function.
4972 (pass_diagnose_tm_blocks): Replace declaration with that of...
4973 (make_pass_diagnose_tm_blocks): ...new function.
4974 (pass_lower_tm): Replace declaration with that of...
4975 (make_pass_lower_tm): ...new function.
4976 (pass_tm_init): Replace declaration with that of...
4977 (make_pass_tm_init): ...new function.
4978 (pass_tm_mark): Replace declaration with that of...
4979 (make_pass_tm_mark): ...new function.
4980 (pass_tm_memopt): Replace declaration with that of...
4981 (make_pass_tm_memopt): ...new function.
4982 (pass_tm_edges): Replace declaration with that of...
4983 (make_pass_tm_edges): ...new function.
4984 (pass_split_functions): Replace declaration with that of...
4985 (make_pass_split_functions): ...new function.
4986 (pass_feedback_split_functions): Replace declaration with that of...
4987 (make_pass_feedback_split_functions): ...new function.
4988 (pass_strength_reduction): Replace declaration with that of...
4989 (make_pass_strength_reduction): ...new function.
4990 (pass_ipa_lower_emutls): Replace declaration with that of...
4991 (make_pass_ipa_lower_emutls): ...new function.
4992 (pass_ipa_function_and_variable_visibility): Replace declaration with
4994 (make_pass_ipa_function_and_variable_visibility): ...new function.
4995 (pass_ipa_tree_profile): Replace declaration with that of...
4996 (make_pass_ipa_tree_profile): ...new function.
4997 (pass_early_local_passes): Replace declaration with that of...
4998 (make_pass_early_local_passes): ...new function.
4999 (pass_ipa_whole_program_visibility): Replace declaration with that
5001 (make_pass_ipa_whole_program_visibility): ...new function.
5002 (pass_ipa_lto_gimple_out): Replace declaration with that of...
5003 (make_pass_ipa_lto_gimple_out): ...new function.
5004 (pass_ipa_increase_alignment): Replace declaration with that of...
5005 (make_pass_ipa_increase_alignment): ...new function.
5006 (pass_ipa_inline): Replace declaration with that of...
5007 (make_pass_ipa_inline): ...new function.
5008 (pass_ipa_free_lang_data): Replace declaration with that of...
5009 (make_pass_ipa_free_lang_data): ...new function.
5010 (pass_ipa_free_inline_summary): Replace declaration with that of...
5011 (make_pass_ipa_free_inline_summary): ...new function.
5012 (pass_ipa_cp): Replace declaration with that of...
5013 (make_pass_ipa_cp): ...new function.
5014 (pass_ipa_reference): Replace declaration with that of...
5015 (make_pass_ipa_reference): ...new function.
5016 (pass_ipa_pure_const): Replace declaration with that of...
5017 (make_pass_ipa_pure_const): ...new function.
5018 (pass_ipa_pta): Replace declaration with that of...
5019 (make_pass_ipa_pta): ...new function.
5020 (pass_ipa_lto_finish_out): Replace declaration with that of...
5021 (make_pass_ipa_lto_finish_out): ...new function.
5022 (pass_ipa_tm): Replace declaration with that of...
5023 (make_pass_ipa_tm): ...new function.
5024 (pass_ipa_profile): Replace declaration with that of...
5025 (make_pass_ipa_profile): ...new function.
5026 (pass_ipa_cdtor_merge): Replace declaration with that of...
5027 (make_pass_ipa_cdtor_merge): ...new function.
5028 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
5030 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5031 (pass_init_datastructures): Replace declaration with that of...
5032 (make_pass_init_datastructures): ...new function.
5033 (pass_fixup_cfg): Replace declaration with that of...
5034 (make_pass_fixup_cfg): ...new function.
5035 (pass_expand): Replace declaration with that of...
5036 (make_pass_expand): ...new function.
5037 (pass_instantiate_virtual_regs): Replace declaration with that of...
5038 (make_pass_instantiate_virtual_regs): ...new function.
5039 (pass_rtl_fwprop): Replace declaration with that of...
5040 (make_pass_rtl_fwprop): ...new function.
5041 (pass_rtl_fwprop_addr): Replace declaration with that of...
5042 (make_pass_rtl_fwprop_addr): ...new function.
5043 (pass_jump): Replace declaration with that of...
5044 (make_pass_jump): ...new function.
5045 (pass_jump2): Replace declaration with that of...
5046 (make_pass_jump2): ...new function.
5047 (pass_lower_subreg): Replace declaration with that of...
5048 (make_pass_lower_subreg): ...new function.
5049 (pass_cse): Replace declaration with that of...
5050 (make_pass_cse): ...new function.
5051 (pass_fast_rtl_dce): Replace declaration with that of...
5052 (make_pass_fast_rtl_dce): ...new function.
5053 (pass_ud_rtl_dce): Replace declaration with that of...
5054 (make_pass_ud_rtl_dce): ...new function.
5055 (pass_rtl_dce): Replace declaration with that of...
5056 (make_pass_rtl_dce): ...new function.
5057 (pass_rtl_dse1): Replace declaration with that of...
5058 (make_pass_rtl_dse1): ...new function.
5059 (pass_rtl_dse2): Replace declaration with that of...
5060 (make_pass_rtl_dse2): ...new function.
5061 (pass_rtl_dse3): Replace declaration with that of...
5062 (make_pass_rtl_dse3): ...new function.
5063 (pass_rtl_cprop): Replace declaration with that of...
5064 (make_pass_rtl_cprop): ...new function.
5065 (pass_rtl_pre): Replace declaration with that of...
5066 (make_pass_rtl_pre): ...new function.
5067 (pass_rtl_hoist): Replace declaration with that of...
5068 (make_pass_rtl_hoist): ...new function.
5069 (pass_rtl_store_motion): Replace declaration with that of...
5070 (make_pass_rtl_store_motion): ...new function.
5071 (pass_cse_after_global_opts): Replace declaration with that of...
5072 (make_pass_cse_after_global_opts): ...new function.
5073 (pass_rtl_ifcvt): Replace declaration with that of...
5074 (make_pass_rtl_ifcvt): ...new function.
5075 (pass_into_cfg_layout_mode): Replace declaration with that of...
5076 (make_pass_into_cfg_layout_mode): ...new function.
5077 (pass_outof_cfg_layout_mode): Replace declaration with that of...
5078 (make_pass_outof_cfg_layout_mode): ...new function.
5079 (pass_loop2): Replace declaration with that of...
5080 (make_pass_loop2): ...new function.
5081 (pass_rtl_loop_init): Replace declaration with that of...
5082 (make_pass_rtl_loop_init): ...new function.
5083 (pass_rtl_move_loop_invariants): Replace declaration with that of...
5084 (make_pass_rtl_move_loop_invariants): ...new function.
5085 (pass_rtl_unswitch): Replace declaration with that of...
5086 (make_pass_rtl_unswitch): ...new function.
5087 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
5088 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5089 (pass_rtl_doloop): Replace declaration with that of...
5090 (make_pass_rtl_doloop): ...new function.
5091 (pass_rtl_loop_done): Replace declaration with that of...
5092 (make_pass_rtl_loop_done): ...new function.
5093 (pass_web): Replace declaration with that of...
5094 (make_pass_web): ...new function.
5095 (pass_cse2): Replace declaration with that of...
5096 (make_pass_cse2): ...new function.
5097 (pass_df_initialize_opt): Replace declaration with that of...
5098 (make_pass_df_initialize_opt): ...new function.
5099 (pass_df_initialize_no_opt): Replace declaration with that of...
5100 (make_pass_df_initialize_no_opt): ...new function.
5101 (pass_reginfo_init): Replace declaration with that of...
5102 (make_pass_reginfo_init): ...new function.
5103 (pass_inc_dec): Replace declaration with that of...
5104 (make_pass_inc_dec): ...new function.
5105 (pass_stack_ptr_mod): Replace declaration with that of...
5106 (make_pass_stack_ptr_mod): ...new function.
5107 (pass_initialize_regs): Replace declaration with that of...
5108 (make_pass_initialize_regs): ...new function.
5109 (pass_combine): Replace declaration with that of...
5110 (make_pass_combine): ...new function.
5111 (pass_if_after_combine): Replace declaration with that of...
5112 (make_pass_if_after_combine): ...new function.
5113 (pass_ree): Replace declaration with that of...
5114 (make_pass_ree): ...new function.
5115 (pass_partition_blocks): Replace declaration with that of...
5116 (make_pass_partition_blocks): ...new function.
5117 (pass_match_asm_constraints): Replace declaration with that of...
5118 (make_pass_match_asm_constraints): ...new function.
5119 (pass_regmove): Replace declaration with that of...
5120 (make_pass_regmove): ...new function.
5121 (pass_split_all_insns): Replace declaration with that of...
5122 (make_pass_split_all_insns): ...new function.
5123 (pass_fast_rtl_byte_dce): Replace declaration with that of...
5124 (make_pass_fast_rtl_byte_dce): ...new function.
5125 (pass_lower_subreg2): Replace declaration with that of...
5126 (make_pass_lower_subreg2): ...new function.
5127 (pass_mode_switching): Replace declaration with that of...
5128 (make_pass_mode_switching): ...new function.
5129 (pass_sms): Replace declaration with that of...
5130 (make_pass_sms): ...new function.
5131 (pass_sched): Replace declaration with that of...
5132 (make_pass_sched): ...new function.
5133 (pass_ira): Replace declaration with that of...
5134 (make_pass_ira): ...new function.
5135 (pass_reload): Replace declaration with that of...
5136 (make_pass_reload): ...new function.
5137 (pass_clean_state): Replace declaration with that of...
5138 (make_pass_clean_state): ...new function.
5139 (pass_branch_prob): Replace declaration with that of...
5140 (make_pass_branch_prob): ...new function.
5141 (pass_value_profile_transformations): Replace declaration with that
5143 (make_pass_value_profile_transformations): ...new function.
5144 (pass_postreload_cse): Replace declaration with that of...
5145 (make_pass_postreload_cse): ...new function.
5146 (pass_gcse2): Replace declaration with that of...
5147 (make_pass_gcse2): ...new function.
5148 (pass_split_after_reload): Replace declaration with that of...
5149 (make_pass_split_after_reload): ...new function.
5150 (pass_branch_target_load_optimize1): Replace declaration with that
5152 (make_pass_branch_target_load_optimize1): ...new function.
5153 (pass_thread_prologue_and_epilogue): Replace declaration with that
5155 (make_pass_thread_prologue_and_epilogue): ...new function.
5156 (pass_stack_adjustments): Replace declaration with that of...
5157 (make_pass_stack_adjustments): ...new function.
5158 (pass_peephole2): Replace declaration with that of...
5159 (make_pass_peephole2): ...new function.
5160 (pass_if_after_reload): Replace declaration with that of...
5161 (make_pass_if_after_reload): ...new function.
5162 (pass_regrename): Replace declaration with that of...
5163 (make_pass_regrename): ...new function.
5164 (pass_cprop_hardreg): Replace declaration with that of...
5165 (make_pass_cprop_hardreg): ...new function.
5166 (pass_reorder_blocks): Replace declaration with that of...
5167 (make_pass_reorder_blocks): ...new function.
5168 (pass_branch_target_load_optimize2): Replace declaration with that
5170 (make_pass_branch_target_load_optimize2): ...new function.
5171 (pass_leaf_regs): Replace declaration with that of...
5172 (make_pass_leaf_regs): ...new function.
5173 (pass_split_before_sched2): Replace declaration with that of...
5174 (make_pass_split_before_sched2): ...new function.
5175 (pass_compare_elim_after_reload): Replace declaration with that of...
5176 (make_pass_compare_elim_after_reload): ...new function.
5177 (pass_sched2): Replace declaration with that of...
5178 (make_pass_sched2): ...new function.
5179 (pass_stack_regs): Replace declaration with that of...
5180 (make_pass_stack_regs): ...new function.
5181 (pass_stack_regs_run): Replace declaration with that of...
5182 (make_pass_stack_regs_run): ...new function.
5183 (pass_df_finish): Replace declaration with that of...
5184 (make_pass_df_finish): ...new function.
5185 (pass_compute_alignments): Replace declaration with that of...
5186 (make_pass_compute_alignments): ...new function.
5187 (pass_duplicate_computed_gotos): Replace declaration with that of...
5188 (make_pass_duplicate_computed_gotos): ...new function.
5189 (pass_variable_tracking): Replace declaration with that of...
5190 (make_pass_variable_tracking): ...new function.
5191 (pass_free_cfg): Replace declaration with that of...
5192 (make_pass_free_cfg): ...new function.
5193 (pass_machine_reorg): Replace declaration with that of...
5194 (make_pass_machine_reorg): ...new function.
5195 (pass_cleanup_barriers): Replace declaration with that of...
5196 (make_pass_cleanup_barriers): ...new function.
5197 (pass_delay_slots): Replace declaration with that of...
5198 (make_pass_delay_slots): ...new function.
5199 (pass_split_for_shorten_branches): Replace declaration with that of...
5200 (make_pass_split_for_shorten_branches): ...new function.
5201 (pass_split_before_regstack): Replace declaration with that of...
5202 (make_pass_split_before_regstack): ...new function.
5203 (pass_convert_to_eh_region_ranges): Replace declaration with that
5205 (make_pass_convert_to_eh_region_ranges): ...new function.
5206 (pass_shorten_branches): Replace declaration with that of...
5207 (make_pass_shorten_branches): ...new function.
5208 (pass_set_nothrow_function_flags): Replace declaration with that of...
5209 (make_pass_set_nothrow_function_flags): ...new function.
5210 (pass_dwarf2_frame): Replace declaration with that of...
5211 (make_pass_dwarf2_frame): ...new function.
5212 (pass_final): Replace declaration with that of...
5213 (make_pass_final): ...new function.
5214 (pass_rtl_seqabstr): Replace declaration with that of...
5215 (make_pass_rtl_seqabstr): ...new function.
5216 (pass_release_ssa_names): Replace declaration with that of...
5217 (make_pass_release_ssa_names): ...new function.
5218 (pass_early_inline): Replace declaration with that of...
5219 (make_pass_early_inline): ...new function.
5220 (pass_inline_parameters): Replace declaration with that of...
5221 (make_pass_inline_parameters): ...new function.
5222 (pass_update_address_taken): Replace declaration with that of...
5223 (make_pass_update_address_taken): ...new function.
5224 (pass_convert_switch): Replace declaration with that of...
5225 (make_pass_convert_switch): ...new function.
5226 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
5227 to a subclass of simple_ipa_opt_pass along with...
5228 (pass_data_ipa_tree_profile): ...new pass_data instance and...
5229 (make_pass_ipa_tree_profile): ...new function.
5230 * tree-sra.c (pass_sra_early): Convert from a global struct to a
5231 subclass of gimple_opt_pass along with...
5232 (pass_data_sra_early): ...new pass_data instance and...
5233 (make_pass_sra_early): ...new function.
5234 (pass_sra): Convert from a global struct to a subclass of
5235 gimple_opt_pass along with...
5236 (pass_data_sra): ...new pass_data instance and...
5237 (make_pass_sra): ...new function.
5238 (pass_early_ipa_sra): Convert from a global struct to a subclass of
5239 gimple_opt_pass along with...
5240 (pass_data_early_ipa_sra): ...new pass_data instance and...
5241 (make_pass_early_ipa_sra): ...new function.
5242 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
5243 subclass of gimple_opt_pass along with...
5244 (pass_data_ccp): ...new pass_data instance and...
5245 (make_pass_ccp): ...new function.
5246 (pass_fold_builtins): Convert from a global struct to a subclass of
5247 gimple_opt_pass along with...
5248 (pass_data_fold_builtins): ...new pass_data instance and...
5249 (make_pass_fold_builtins): ...new function.
5250 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
5251 subclass of gimple_opt_pass along with...
5252 (pass_data_copy_prop): ...new pass_data instance and...
5253 (make_pass_copy_prop): ...new function.
5254 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
5255 global struct to a subclass of gimple_opt_pass along with...
5256 (pass_data_rename_ssa_copies): ...new pass_data instance and...
5257 (make_pass_rename_ssa_copies): ...new function.
5258 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
5259 subclass of gimple_opt_pass along with...
5260 (pass_data_dce): ...new pass_data instance and...
5261 (make_pass_dce): ...new function.
5262 (pass_dce_loop): Convert from a global struct to a subclass of
5263 gimple_opt_pass along with...
5264 (pass_data_dce_loop): ...new pass_data instance and...
5265 (make_pass_dce_loop): ...new function.
5266 (pass_cd_dce): Convert from a global struct to a subclass of
5267 gimple_opt_pass along with...
5268 (pass_data_cd_dce): ...new pass_data instance and...
5269 (make_pass_cd_dce): ...new function.
5270 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
5271 subclass of gimple_opt_pass along with...
5272 (pass_data_dominator): ...new pass_data instance and...
5273 (make_pass_dominator): ...new function.
5274 (pass_phi_only_cprop): Convert from a global struct to a subclass of
5275 gimple_opt_pass along with...
5276 (pass_data_phi_only_cprop): ...new pass_data instance and...
5277 (make_pass_phi_only_cprop): ...new function.
5278 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
5279 subclass of gimple_opt_pass along with...
5280 (pass_data_dse): ...new pass_data instance and...
5281 (make_pass_dse): ...new function.
5282 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
5283 a subclass of gimple_opt_pass along with...
5284 (pass_data_forwprop): ...new pass_data instance and...
5285 (make_pass_forwprop): ...new function.
5286 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
5287 struct to a subclass of gimple_opt_pass along with...
5288 (pass_data_tree_ifcombine): ...new pass_data instance and...
5289 (make_pass_tree_ifcombine): ...new function.
5290 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
5291 subclass of gimple_opt_pass along with...
5292 (pass_data_ch): ...new pass_data instance and...
5293 (make_pass_ch): ...new function.
5294 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
5295 subclass of gimple_opt_pass along with...
5296 (pass_data_tree_loop): ...new pass_data instance and...
5297 (make_pass_tree_loop): ...new function.
5298 (pass_tree_loop_init): Convert from a global struct to a subclass of
5299 gimple_opt_pass along with...
5300 (pass_data_tree_loop_init): ...new pass_data instance and...
5301 (make_pass_tree_loop_init): ...new function.
5302 (pass_lim): Convert from a global struct to a subclass of
5303 gimple_opt_pass along with...
5304 (pass_data_lim): ...new pass_data instance and...
5305 (make_pass_lim): ...new function.
5306 (pass_tree_unswitch): Convert from a global struct to a subclass of
5307 gimple_opt_pass along with...
5308 (pass_data_tree_unswitch): ...new pass_data instance and...
5309 (make_pass_tree_unswitch): ...new function.
5310 (pass_predcom): Convert from a global struct to a subclass of
5311 gimple_opt_pass along with...
5312 (pass_data_predcom): ...new pass_data instance and...
5313 (make_pass_predcom): ...new function.
5314 (pass_vectorize): Convert from a global struct to a subclass of
5315 gimple_opt_pass along with...
5316 (pass_data_vectorize): ...new pass_data instance and...
5317 (make_pass_vectorize): ...new function.
5318 (pass_graphite): Convert from a global struct to a subclass of
5319 gimple_opt_pass along with...
5320 (pass_data_graphite): ...new pass_data instance and...
5321 (make_pass_graphite): ...new function.
5322 (pass_graphite_transforms): Convert from a global struct to a subclass
5323 of gimple_opt_pass along with...
5324 (pass_data_graphite_transforms): ...new pass_data instance and...
5325 (make_pass_graphite_transforms): ...new function.
5326 (pass_check_data_deps): Convert from a global struct to a subclass of
5327 gimple_opt_pass along with...
5328 (pass_data_check_data_deps): ...new pass_data instance and...
5329 (make_pass_check_data_deps): ...new function.
5330 (pass_iv_canon): Convert from a global struct to a subclass of
5331 gimple_opt_pass along with...
5332 (pass_data_iv_canon): ...new pass_data instance and...
5333 (make_pass_iv_canon): ...new function.
5334 (pass_scev_cprop): Convert from a global struct to a subclass of
5335 gimple_opt_pass along with...
5336 (pass_data_scev_cprop): ...new pass_data instance and...
5337 (make_pass_scev_cprop): ...new function.
5338 (pass_record_bounds): Convert from a global struct to a subclass of
5339 gimple_opt_pass along with...
5340 (pass_data_record_bounds): ...new pass_data instance and...
5341 (make_pass_record_bounds): ...new function.
5342 (pass_complete_unroll): Convert from a global struct to a subclass of
5343 gimple_opt_pass along with...
5344 (pass_data_complete_unroll): ...new pass_data instance and...
5345 (make_pass_complete_unroll): ...new function.
5346 (pass_complete_unrolli): Convert from a global struct to a subclass of
5347 gimple_opt_pass along with...
5348 (pass_data_complete_unrolli): ...new pass_data instance and...
5349 (make_pass_complete_unrolli): ...new function.
5350 (pass_parallelize_loops): Convert from a global struct to a subclass
5351 of gimple_opt_pass along with...
5352 (pass_data_parallelize_loops): ...new pass_data instance and...
5353 (make_pass_parallelize_loops): ...new function.
5354 (pass_loop_prefetch): Convert from a global struct to a subclass of
5355 gimple_opt_pass along with...
5356 (pass_data_loop_prefetch): ...new pass_data instance and...
5357 (make_pass_loop_prefetch): ...new function.
5358 (pass_iv_optimize): Convert from a global struct to a subclass of
5359 gimple_opt_pass along with...
5360 (pass_data_iv_optimize): ...new pass_data instance and...
5361 (make_pass_iv_optimize): ...new function.
5362 (pass_tree_loop_done): Convert from a global struct to a subclass of
5363 gimple_opt_pass along with...
5364 (pass_data_tree_loop_done): ...new pass_data instance and...
5365 (make_pass_tree_loop_done): ...new function.
5366 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
5367 struct to a subclass of gimple_opt_pass along with...
5368 (pass_data_cse_reciprocals): ...new pass_data instance and...
5369 (make_pass_cse_reciprocals): ...new function.
5370 (pass_cse_sincos): Convert from a global struct to a subclass of
5371 gimple_opt_pass along with...
5372 (pass_data_cse_sincos): ...new pass_data instance and...
5373 (make_pass_cse_sincos): ...new function.
5374 (pass_optimize_bswap): Convert from a global struct to a subclass of
5375 gimple_opt_pass along with...
5376 (pass_data_optimize_bswap): ...new pass_data instance and...
5377 (make_pass_optimize_bswap): ...new function.
5378 (pass_optimize_widening_mul): Convert from a global struct to a
5379 subclass of gimple_opt_pass along with...
5380 (pass_data_optimize_widening_mul): ...new pass_data instance and...
5381 (make_pass_optimize_widening_mul): ...new function.
5382 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
5383 subclass of gimple_opt_pass along with...
5384 (pass_data_phiopt): ...new pass_data instance and...
5385 (make_pass_phiopt): ...new function.
5386 (pass_cselim): Convert from a global struct to a subclass of
5387 gimple_opt_pass along with...
5388 (pass_data_cselim): ...new pass_data instance and...
5389 (make_pass_cselim): ...new function.
5390 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
5391 subclass of gimple_opt_pass along with...
5392 (pass_data_phiprop): ...new pass_data instance and...
5393 (make_pass_phiprop): ...new function.
5394 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
5395 subclass of gimple_opt_pass along with...
5396 (pass_data_pre): ...new pass_data instance and...
5397 (make_pass_pre): ...new function.
5398 (pass_fre): Convert from a global struct to a subclass of
5399 gimple_opt_pass along with...
5400 (pass_data_fre): ...new pass_data instance and...
5401 (make_pass_fre): ...new function.
5402 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
5403 subclass of gimple_opt_pass along with...
5404 (pass_data_reassoc): ...new pass_data instance and...
5405 (make_pass_reassoc): ...new function.
5406 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
5407 subclass of gimple_opt_pass along with...
5408 (pass_data_sink_code): ...new pass_data instance and...
5409 (make_pass_sink_code): ...new function.
5410 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
5411 subclass of gimple_opt_pass along with...
5412 (pass_data_strlen): ...new pass_data instance and...
5413 (make_pass_strlen): ...new function.
5414 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
5415 struct to a subclass of gimple_opt_pass along with...
5416 (pass_data_build_alias): ...new pass_data instance and...
5417 (make_pass_build_alias): ...new function.
5418 (pass_build_ealias): Convert from a global struct to a subclass of
5419 gimple_opt_pass along with...
5420 (pass_data_build_ealias): ...new pass_data instance and...
5421 (make_pass_build_ealias): ...new function.
5422 (pass_ipa_pta): Convert from a global struct to a subclass of
5423 simple_ipa_opt_pass along with...
5424 (pass_data_ipa_pta): ...new pass_data instance and...
5425 (make_pass_ipa_pta): ...new function.
5426 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
5427 subclass of gimple_opt_pass along with...
5428 (pass_data_uncprop): ...new pass_data instance and...
5429 (make_pass_uncprop): ...new function.
5430 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
5431 global struct to a subclass of gimple_opt_pass along with...
5432 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
5433 (make_pass_late_warn_uninitialized): ...new function.
5434 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
5435 to a subclass of gimple_opt_pass along with...
5436 (pass_data_init_datastructures): ...new pass_data instance and...
5437 (make_pass_init_datastructures): ...new function.
5438 (pass_early_warn_uninitialized): Convert from a global struct to a
5439 subclass of gimple_opt_pass along with...
5440 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5441 (make_pass_early_warn_uninitialized): ...new function.
5442 (pass_update_address_taken): Convert from a global struct to a
5443 subclass of gimple_opt_pass along with...
5444 (pass_data_update_address_taken): ...new pass_data instance and...
5445 (make_pass_update_address_taken): ...new function.
5446 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5447 struct to a subclass of gimple_opt_pass along with...
5448 (pass_data_release_ssa_names): ...new pass_data instance and...
5449 (make_pass_release_ssa_names): ...new function.
5450 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5451 subclass of gimple_opt_pass along with...
5452 (pass_data_stdarg): ...new pass_data instance and...
5453 (make_pass_stdarg): ...new function.
5454 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5455 global struct to a subclass of gimple_opt_pass along with...
5456 (pass_data_convert_switch): ...new pass_data instance and...
5457 (make_pass_convert_switch): ...new function.
5458 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5459 to a subclass of gimple_opt_pass along with...
5460 (pass_data_tail_recursion): ...new pass_data instance and...
5461 (make_pass_tail_recursion): ...new function.
5462 (pass_tail_calls): Convert from a global struct to a subclass of
5463 gimple_opt_pass along with...
5464 (pass_data_tail_calls): ...new pass_data instance and...
5465 (make_pass_tail_calls): ...new function.
5466 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5467 struct to a subclass of gimple_opt_pass along with...
5468 (pass_data_lower_vector): ...new pass_data instance and...
5469 (make_pass_lower_vector): ...new function.
5470 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5471 gimple_opt_pass along with...
5472 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5473 (make_pass_lower_vector_ssa): ...new function.
5474 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5475 to a subclass of gimple_opt_pass along with...
5476 (pass_data_slp_vectorize): ...new pass_data instance and...
5477 (make_pass_slp_vectorize): ...new function.
5478 (pass_ipa_increase_alignment): Convert from a global struct to a
5479 subclass of simple_ipa_opt_pass along with...
5480 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5481 (make_pass_ipa_increase_alignment): ...new function.
5482 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5483 gimple_opt_pass along with...
5484 (pass_data_vrp): ...new pass_data instance and...
5485 (make_pass_vrp): ...new function.
5486 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5487 subclass of simple_ipa_opt_pass along with...
5488 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5489 (make_pass_ipa_free_lang_data): ...new function.
5490 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5491 gimple_opt_pass along with...
5492 (pass_data_tsan): ...new pass_data instance and...
5493 (make_pass_tsan): ...new function.
5494 (pass_tsan_O0): Convert from a global struct to a subclass of
5495 gimple_opt_pass along with...
5496 (pass_data_tsan_O0): ...new pass_data instance and...
5497 (make_pass_tsan_O0): ...new function.
5498 * var-tracking.c (pass_variable_tracking): Convert from a global
5499 struct to a subclass of rtl_opt_pass along with...
5500 (pass_data_variable_tracking): ...new pass_data instance and...
5501 (make_pass_variable_tracking): ...new function.
5502 * web.c (pass_web): Convert from a global struct to a subclass of
5503 rtl_opt_pass along with...
5504 (pass_data_web): ...new pass_data instance and...
5505 (make_pass_web): ...new function.
5506 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5507 declaration with that of...
5508 (make_pass_mode_switch_use): ...new function.
5509 (pass_resolve_sw_modes): Replace declaration with that of...
5510 (make_pass_resolve_sw_modes): ...new function.
5511 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5512 from a global struct to a subclass of rtl_opt_pass along with...
5513 (pass_data_mode_switch_use): ...new pass_data instance and...
5514 (make_pass_mode_switch_use): ...new function.
5515 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5516 from a global struct to a subclass of rtl_opt_pass along with...
5517 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5518 (make_pass_resolve_sw_modes): ...new function.
5519 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5520 struct to a subclass of rtl_opt_pass along with...
5521 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5522 (make_pass_insert_vzeroupper): ...new function.
5523 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5524 global struct to a subclass of rtl_opt_pass along with...
5525 (pass_data_work_around_errata): ...new pass_data instance and...
5526 (make_pass_work_around_errata): ...new function.
5527 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5528 struct to a subclass of rtl_opt_pass along with...
5529 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
5530 (make_pass_mips_machine_reorg2): ...new function.
5532 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5534 * passes.c (pass_manager::operator new): New.
5536 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5538 Handwritten part of conversion of passes to C++ classes.
5540 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
5541 (toplev.o): Add dep on PASS_MANAGER_H.
5542 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
5543 of early local pases to reflect this moving from a global to a
5544 member of gcc::pass_manager.
5545 (cgraph_add_new_function): Likewise.
5546 * lto-cgraph.c (lto_output_node): Update for conversion of
5547 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
5548 * passes.c (opt_pass::clone): New.
5549 (opt_pass::gate): New.
5550 (opt_pass::execute): New.
5551 (opt_pass::opt_pass): New.
5552 (pass_manager::execute_early_local_passes): New.
5553 (pass_manager::execute_pass_mode_switching): new.
5554 (finish_optimization_passes): Convert to...
5555 (pass_manager::finish_optimization_passes): ...this.
5556 (finish_optimization_passes): Update for conversion of passes to
5558 (register_dump_files_1): Use has_gate since we cannot portably
5559 check a vtable entry against NULL.
5560 (dump_one_pass): Likewise.
5561 (ipa_write_summaries_2): Likewise.
5562 (ipa_write_optimization_summaries_1): Likewise.
5563 (ipa_read_summaries_1): Likewise.
5564 (ipa_read_optimization_summaries_1): Likewise.
5565 (execute_ipa_stmt_fixups): Likewise.
5566 (pass_manager::pass_manager): Rewrite pass-creation, invoking
5567 pass-creation functions rather than wiring up globals, and
5568 storing the results in fields of pass_manager generated using
5570 (pass_manager::dump_profile_report): Update for conversion of
5571 passes to C++ classes.
5572 (pass_manager::execute_ipa_summary_passes): Likewise.
5573 (execute_one_ipa_transform_pass): Likewise.
5574 (execute_one_pass): Use has_gate and has_execute since we cannot
5575 portably check a vtable entry against NULL.
5576 * pass_manager.h (pass_manager::finish_optimization_passes): New.
5577 (pass_manager): Use pass-instances.def to add fields for the
5578 various pass instances.
5579 * toplev.c (finalize): Update for move of
5580 finish_optimization_passes to a method of gcc::pass_manager.
5581 * toplev.h (finish_optimization_passes): Move to method of class
5583 * tree-pass.h (struct pass_data): New.
5584 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
5585 (opt_pass::gate): Convert to virtual function.
5586 (opt_pass::~opt_pass): New.
5587 (opt_pass::clone): New.
5588 (opt_pass::execute): Convert to virtual function.
5589 (opt_pass::opt_pass): New.
5590 (opt_pass::ctxt_): new.
5591 (gimple_opt_pass): Convert to subclass of opt_pass.
5592 (gimple_opt_pass::gimple_opt_pass): New.
5593 (rtl_opt_pass): Convert to subclass of opt_pass.
5594 (rtl_opt_pass::rtl_opt_pass): New.
5595 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5596 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5597 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5598 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5599 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5600 invocation of pass_mode_switching to reflect this moving from a
5601 global to a member of gcc::pass_manager.
5602 (ix86_option_override): Rework how pass_insert_vzeroupper is
5603 added to the pass_manager to reflect autogenerated changes.
5604 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
5606 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
5608 PR rtl-optimization/57708
5609 * recog.c (peep2_find_free_register): Validate all regs in a
5612 2013-08-05 Jan Hubicka <jh@suse.cz>
5615 * cgraph.c (verify_cgraph_node): Accept local flags from other
5617 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5618 (function_and_variable_visibility): Likewise.
5619 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5621 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5623 * graph.c (init_graph_slim_pretty_print): Remove.
5624 (print_graph_cfg): Do not call it. Use local pretty printer.
5625 (start_graph_dump): Likewise.
5627 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5629 * gimple-pretty-print.c (buffer): Remove.
5630 (initialized): Likewise.
5631 (maybe_init_pretty_print): Likewise.
5632 (print_gimple_stmt): Do not call it. Use non-static local
5633 pretty_printer variable.
5634 (print_gimple_expr): Likewise.
5635 (print_gimple_seq): Likewise.
5636 (gimple_dump_bb): Likewise.
5638 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5640 * asan.c (asan_pp): Remove.
5641 (asan_pp_initialized): Likewise.
5642 (asan_pp_initialize): Likewise.
5643 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5644 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5645 (asan_add_global): Likewise.
5647 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5649 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5650 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5651 * pretty-print.c (pp_formatted_text_data): Likewise.
5652 (pp_indent): Rename from pp_base_indent.
5653 (pp_format): Rename from pp_base_format.
5654 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5655 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5656 (pp_flush): Rename from pp_base_flush.
5657 (pp_set_line_maximum_length): Rename from
5658 pp_base_set_line_maximum_length.
5659 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5660 (pp_set_prefix): Rename from pp_base_set_prefix.
5661 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5662 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5663 (pp_append_text): Rename from pp_base_append_text.
5664 (pp_formatted_text): Rename from pp_base_formatted_text.
5665 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5666 (pp_remaining_character_count_for_line): Rename from
5667 pp_base_remaining_character_count_for_line.
5668 (pp_newline): Rename from pp_base_newline.
5669 (pp_character): Rename from pp_base_character.
5670 (pp_string): Rename from pp_base_string.
5671 (pp_maybe_space): Rename from pp_base_maybe_space.
5672 * asan.c (asan_pp_string): Adjust.
5673 (asan_emit_stack_protection): Likewise.
5674 (asan_add_global): Likewise.
5675 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5676 * tree-mudflap.c (mf_varname_tree): Likewise.
5677 * tree-pretty-print.c (pp_tree_identifier): Rename from
5678 pp_base_tree_identifier.
5679 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5680 Declare as function.
5682 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5684 * pretty-print.h (pp_bar_bar): New.
5685 (pp_ampersand_ampersand): Likewise.
5686 (pp_less_equal): Likewise.
5687 (pp_greater_equal): Likewise.
5688 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5689 printer functions instead of pp_string or operators and punctuators.
5690 (dump_gimple_call): Likewise.
5691 (dump_gimple_omp_for): Likewise.
5692 (dump_gimple_transaction): Likewise.
5693 (dump_gimple_phi): Likewise.
5694 (pp_gimple_stmt_1): Likewise.
5695 * sched-vis.c (print_insn): Likewise.
5696 * tree-mudflap.c (mf_varname_tree): Likewise.
5697 * tree-pretty-print.c (dump_block_node): Likewise.
5698 (dump_generic_node): Likewise.
5700 2013-08-02 Jan Hubicka <jh@suse.cz>
5702 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5704 * lto-streamer-out.c (tree_is_indexable): Results decls and
5705 parm decls are not indexable.
5706 (DFS_write_tree_body): Do not follow args and results.
5707 (hash_tree): Likewise.
5708 (output_functions): Rearrange so struct function is needed
5709 only when real body is output; be able to also ouptut abstract
5710 functions; output DECL_ARGUMENTS and DECL_RESULT.
5711 (lto_output): When not in WPA, ale store abstract functions.
5712 (write_symbol): Do not care about RESULT_DECL.
5713 (output_symbol_p): Handle correctly sbtract decls.
5714 * lto-streamer-in.c (input_function): Rearrange so struct
5715 function can be NULL at entry; allow streaming of
5716 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5717 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5718 sanity check during LTO.
5719 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5720 RESULT_DECl and DECL_ARGUMENTS.
5721 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5724 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5726 * pretty-print.h (pp_underscore): New.
5728 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5729 printer functions instead of pp_character.
5730 (dump_binary_rhs): Likewise.
5731 (dump_ternary_rhs): Likewise.
5732 (dump_gimple_call_args): Likewise.
5733 (pp_points_to_solution): Likewise.
5734 (dump_gimple_call): Likewise.
5735 (dump_gimple_switch): Likewise.
5736 (dump_gimple_cond): Likewise.
5737 (dump_gimple_bind): Likewise.
5738 (dump_gimple_try): Likewise.
5739 (dump_gimple_omp_for): Likewise.
5740 (dump_gimple_omp_continue): Likewise.
5741 (dump_gimple_omp_single): Likewise.
5742 (dump_gimple_omp_sections): Likewise.
5743 (dump_gimple_omp_block): Likewise.
5744 (dump_gimple_omp_critical): Likewise.
5745 (dump_gimple_transaction): Likewise.
5746 (dump_gimple_asm): Likewise.
5747 (dump_gimple_phi): Likewise.
5748 (dump_gimple_omp_parallel): Likewise.
5749 (dump_gimple_omp_task): Likewise.
5750 (dump_gimple_omp_atomic_load): Likewise.
5751 (dump_gimple_omp_atomic_store): Likewise.
5752 (dump_gimple_mem_ops): Likewise.
5753 (pp_gimple_stmt_1): Likewise.
5754 (pp_cfg_jump): Likewise.
5755 (dump_implicit_edges): Likewise.
5756 (gimple_dump_bb_for_graph): Likewise.
5757 * graph.c (draw_cfg_node): Likewise.
5758 * langhooks.c (lhd_print_error_function): Likewise.
5759 * sched-vis.c (print_exp): Likewise.
5760 (print_value): Likewise.
5761 (print_pattern): Likewise.
5762 (print_insn): Likewise.
5763 (rtl_dump_bb_for_graph): Likewise.
5764 * tree-pretty-print.c (dump_function_declaration): Likewise.
5765 (dump_array_domain): Likewise.
5766 (dump_omp_clause): Likewise.
5767 (dump_location): Likewise.
5768 (dump_generic_node): Likewise.
5769 (print_struct_decl): Likewise.
5770 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5772 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5774 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5775 candidate table when replacing a candidate statement.
5776 (replace_rhs_if_not_dup): Likewise.
5777 (replace_one_candidate): Likewise.
5779 2013-08-02 Jan Hubicka <jh@suse.cz>
5780 Martin Liska <marxin.liska@gmail.com>
5782 * cgraphunit.c (add_new_function): Fix logic when adding from
5784 (assemble_thunk): Rename to ...
5785 (expand_thunk); .. this one; export; get it working with
5786 general functions; make produced gimple valid.
5787 * cgraph.h (expand_thunk): Declare.
5789 2013-08-02 Jan Hubicka <jh@suse.cz>
5791 * ipa-cp.c (gather_context_independent_values): Use
5792 ipa_get_param_move_cost.
5793 (get_replacement_map): Remove PARAM; move parameter folding
5795 (create_specialized_node): Update.
5796 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
5797 assert that we have gimple body; update move_cost.
5798 (count_formal_params): Assert that we have gimple body.
5799 (ipa_dump_param): New function.
5800 (ipa_alloc_node_params): Break out from ...
5801 (ipa_initialize_node_params): ... here.
5802 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
5803 (ipa_write_node_info): Stream move costs.
5804 (ipa_read_node_info): Read move costs.
5805 (ipa_update_after_lto_read): Do not recompute node params.
5806 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
5807 (ipa_get_param): Check we are not in WPA.
5808 (ipa_get_param_move_cost): New.
5809 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
5810 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
5811 parm numbers to be present.
5813 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5815 PR rtl-optimization/58048
5816 * lra-constraints.c (process_alt_operands): Don't check asm
5817 operand on register.
5819 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
5821 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
5822 the implied StoreLoad barrier for atomic operations if before.
5824 2013-08-02 Jan Hubicka <jh@suse.cz>
5825 Martin Liska <marxin.liska@gmail.com>
5827 * cgraph.c (cgraph_function_body_availability): Do not check
5829 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
5830 symtab_node_availability): Declare.
5831 * ipa.c (can_replace_by_local_alias): New.
5832 (function_and_variable_visibility): Use it.
5833 * symtab.c (symtab_for_node_and_aliases,
5834 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
5836 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5838 PR rtl-optimization/57963
5839 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
5840 (lra_constraints): Use them.
5842 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
5844 * config/arm/types.md (define_attr "type"): Add "load_acq"
5846 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
5848 (cortex_a53_store1): Likewise.
5850 2013-08-01 Jan Hubicka <jh@suse.cz>
5852 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
5853 partitions are not needed.
5855 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
5857 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
5858 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
5859 MAYBE_NON_Q_CLASS_P where appropriate.
5861 2013-08-01 Jan Hubicka <jh@suse.cz>
5863 * cgraph.h (release_function_body): Declare.
5864 * tree.c (free_lang_data_in_decl): Free, parameters and return values
5865 of unused delcarations.
5867 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5869 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
5870 RTL form when subtracting a constant.
5872 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5874 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
5875 Generate canonical plus rtx with negated immediate instead of minus
5877 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
5879 2013-08-01 Jan Hubicka <jh@suse.cz>
5881 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
5882 (cgraph_release_function_body): Likewise.
5883 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5884 * cgraph.h (cgrpah_node): Rename abstract_and_needed
5885 to used_as_abstract_origin.
5886 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
5887 symbols used as abstract origins.
5888 * cgraphunit.c (analyze_functions): Update.
5889 * ipa.c (symtab_remove_unreachable_nodes): Recompute
5890 used_as_abstract_origin.
5891 * tree-inline.c (tree_function_versioning): Update
5892 used_as_abstract_origin; be ready for DECL_RESULT and
5893 DECL_ARGUMENTS to be NULL.
5895 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
5896 for abstract functions.
5897 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
5900 2013-08-01 Jan Hubicka <jh@suse.cz>
5902 * profile.c (compute_value_histograms): Fix thinko.
5904 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5906 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
5907 aarch-common-protos.h to extra_headers.
5908 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5909 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
5910 * config/aarch64/t-aarch64 (aarch-common.o): Define.
5912 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5914 * config/aarch64/aarch64.md (define_attr "type"): Delete.
5915 Include "../arm/types.md". Define "type" attribute for all patterns.
5916 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
5919 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
5921 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
5922 to support power8 load fusion.
5923 (fusion_gpr_mem_load): Likewise.
5925 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
5927 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
5928 declarations for power8 load fusion.
5929 (emit_fusion_gpr_load): Likewise.
5931 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
5932 tuning for power8, turn on fusion mode by default. Turn on sign
5933 extending fusion mode if normal fusion mode is on, and we are at
5935 (fusion_gpr_load_p): New function, return true if we can fuse an
5936 addis instruction with a dependent load to a GPR.
5937 (emit_fusion_gpr_load): Emit the instructions for power8 load
5940 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
5941 (VSX load fusion peepholes): New peepholes to fuse together an
5942 addi instruction with a VSX load instruction.
5944 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
5945 peepholes to fuse an addis instruction with a load to a GPR base
5946 register. If we are supporting sign extending fusions, convert
5947 sign extending loads to zero extending loads and add an explicit
5950 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5952 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
5953 aarch-common-protos.h to extra_headers.
5954 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5955 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
5956 (arm_early_store_addr_dep): Likewise.
5957 (arm_no_early_alu_shift_dep): Likewise.
5958 (arm_no_early_alu_shift_value_dep): Likewise.
5959 (arm_no_early_mul_dep): Likewise.
5960 (arm_no_early_store_addr_dep): Likewise.
5961 (arm_mac_accumulator_is_mul_result): Likewise.
5962 (arm_mac_accumulator_is_result): Likewise.
5963 * config/arm/aarch-common.c: ... here. New file.
5964 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
5966 (arm_early_store_addr_dep): Likewise.
5967 (arm_no_early_alu_shift_dep): Likewise.
5968 (arm_no_early_alu_shift_value_dep): Likewise.
5969 (arm_no_early_mul_dep): Likewise.
5970 (arm_no_early_store_addr_dep): Likewise.
5971 (arm_mac_accumulator_is_mul_result): Likewise.
5972 (arm_mac_accumulator_is_result): Likewise.
5973 * config/arm/aarch-common-protos.h: ... here. New file.
5974 * config/arm/t-arm (aarch-common.o): Define.
5976 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5978 * config/arm/arm.md: Include new file "types.md".
5979 (define_attr "type"): Move from here to ...
5980 (define_attr "mul32"): Likewise.
5981 (define_attr "mul64"): Likewise.
5982 * config/arm/types.md: ... here. New file.
5984 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
5986 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
5987 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
5989 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5991 * gen-pass-instances.awk: Fix offset of substr().
5993 2013-07-31 David Malcolm <dmalcolm@redhat.com>
5995 * Makefile.in (pass-instances.def): New.
5996 (passes.o): Replace dependency on passes.def with one on
5999 * gen-pass-instances.awk: New.
6001 * passes.c (pass_manager::pass_manager): Use pass-instances.def
6002 rather than passes.def, updating local definition of NEXT_PASS
6003 macro to add an extra NUM parameter (currently unused).
6005 2013-07-30 David Malcolm <dmalcolm@redhat.com>
6007 * Makefile.in (PASS_MANAGER_H): New.
6008 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
6009 (passes.o): Likewise.
6010 (statistics.o): Likewise.
6011 (cgraphunit.o): Likewise.
6012 (context.o): Depend on PASS_MANAGER_H.
6014 * pass_manager.h: New.
6016 * cgraphunit.c (cgraph_add_new_function): Update for moves
6017 of globals to fields of pass_manager.
6018 (analyze_function): Likewise.
6019 (expand_function): Likewise.
6020 (ipa_passes): Likewise.
6021 (compile): Likewise.
6023 * context.c (context::context): New.
6024 * context.h (context::context): New.
6025 (context::get_passes): New.
6026 (context::passes_): New.
6028 * lto-cgraph.c (input_node): Update for moves of globals to
6029 fields of pass_manager.
6031 * passes.c (all_passes): Remove, in favor of a field of the
6032 same name within the new class pass_manager.
6033 (all_small_ipa_passes): Likewise.
6034 (all_lowering_passes): Likewise.
6035 (all_regular_ipa_passes): Likewise.
6036 (all_late_ipa_passes): Likewise.
6037 (all_lto_gen_passes): Likewise.
6038 (passes_by_id): Likewise.
6039 (passes_by_id_size): Likewise.
6040 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6041 the new class pass_manager.
6042 (set_pass_for_id): Convert to...
6043 (pass_manager::set_pass_for_id): ...method.
6044 (get_pass_for_id): Convert to...
6045 (pass_manager::get_pass_for_id): ...method.
6046 (register_one_dump_file): Move body of implementation into...
6047 (pass_manager::register_one_dump_file): ...here.
6048 (register_dump_files_1): Convert to...
6049 (pass_manager::register_dump_files_1): ...method.
6050 (register_dump_files): Convert to...
6051 (pass_manager::register_dump_files): ...method.
6052 (create_pass_tab): Update for moves of globals to fields of
6054 (dump_passes): Move body of implementation into...
6055 (pass_manager::dump_passes): ...here.
6056 (register_pass): Move body of implementation into...
6057 (pass_manager::register_pass): ...here.
6058 (init_optimization_passes): Convert into...
6059 (pass_manager::pass_manager): ...constructor for new
6060 pass_manager class, and initialize the pass_lists array.
6061 (check_profile_consistency): Update for moves of globals to
6062 fields of pass_manager.
6063 (dump_profile_report): Move body of implementation into...
6064 (pass_manager::dump_profile_report): ...here.
6065 (ipa_write_summaries_1): Update for moves of pass lists from
6066 being globals to fields of pass_manager.
6067 (ipa_write_optimization_summaries): Likewise.
6068 (ipa_read_summaries): Likewise.
6069 (ipa_read_optimization_summaries): Likewise.
6070 (execute_all_ipa_stmt_fixups): Likewise.
6072 * statistics.c (statistics_fini): Update for moves of globals to
6073 fields of pass_manager.
6075 * toplev.c (general_init): Replace call to
6076 init_optimization_passes with construction of the pass_manager
6079 * tree-pass.h (all_passes): Remove, in favor of a field of the
6080 same name within the new class pass_manager.
6081 (all_small_ipa_passes): Likewise.
6082 (all_lowering_passes): Likewise.
6083 (all_regular_ipa_passes): Likewise.
6084 (all_lto_gen_passes): Likewise.
6085 (all_late_ipa_passes): Likewise.
6086 (passes_by_id): Likewise.
6087 (passes_by_id_size): Likewise.
6088 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6089 the new class pass_manager.
6090 (get_pass_for_id): Remove.
6092 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6094 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
6097 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6099 * arm.md (mulhi3): New expand pattern.
6101 2013-07-30 Jan Hubicka <jh@suse.cz>
6102 Martin Liska <marxin.liska@gmail.com>
6104 * profile.c (compute_value_histograms): Do not ICE when
6105 there is mismatch only on some counters.
6107 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6109 PR rtl-optimization/57637
6110 * function.c (move_insn_for_shrink_wrap): Also check the
6111 GEN set of the LIVE problem for the liveness analysis
6112 if it exists, otherwise give up.
6114 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6116 PR tree-optimization/57993
6117 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
6118 replaced statement in the candidate table.
6119 (phi_add_costs): Return infinite cost when the hidden basis does
6120 not dominate all phis on which the candidate is dependent.
6121 (replace_one_candidate): Record replaced statement in the
6124 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
6126 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
6127 (ashlv2si3): New expander.
6128 (*ashlv2si3_i): New define_insn_and_split.
6129 * predicates.md (float_operation): Allow patterns with three
6132 PR rtl-optimization/58021
6133 * mode-switching.c (create_pre_exit): Always split off preceding
6134 insns if we are not at the basic block head.
6136 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
6138 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
6139 (UCLIBC_DYNAMIC_LINKER): New macro.
6140 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
6142 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
6143 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
6145 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
6146 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
6147 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
6148 for SF and DF modes. Use ieee_quad_format for TF mode.
6149 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
6150 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
6151 (mips_option_override): Handle `-mnan=legacy'.
6152 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
6153 `-mabs=2008' and `-mnan=2008'.
6154 (OPTION_DEFAULT_SPECS): Add "nan" default.
6155 (ASM_SPEC): Handle `-mnan='.
6156 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
6157 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
6158 comment accordingly.
6159 (neg<mode>2): Likewise.
6160 * config/mips/mips.opt (mabs, mnan): New options.
6161 * doc/install.texi (Configuration): Document `--with-nan=' option.
6162 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
6164 (MIPS Options): Document them.
6165 * config.gcc <mips*-*-*>: Handle `--with-nan='.
6166 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
6167 * configure: Regenerate.
6168 * config.in: Regenerate.
6170 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6172 * config/i386/i386.md (float post-reload splitters): Do not check
6173 for subregs of SSE registers.
6175 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6176 H.J. Lu <hongjiu.lu@intel.com>
6180 * config/i386/i386.md (post-reload splitter
6181 to avoid partial SSE reg dependency stalls): New pattern.
6183 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
6185 * config/s390/s390.md ("movcc"): Swap load and store instructions.
6187 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
6189 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
6190 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
6192 2013-07-26 Cary Coutant <ccoutant@google.com>
6194 * dwarf2out.c (die_checksum_ordered): Don't include template
6195 instantiations in signature.
6196 (is_template_parameter): New function.
6197 (is_template_instantiation): New function.
6198 (generate_skeleton_bottom_up): Don't include template instantiations
6200 (generate_skeleton): Likewise.
6201 (break_out_comdat_types): Move recursive call to break out nested
6203 (prune_unused_types_mark_generic_parms_dies): Call
6204 is_template_parameter.
6206 2013-07-26 Ian Bolton <ian.bolton@arm.com>
6208 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
6209 uses vector registers.
6210 * config/aarch64/iterators.md: Add attributes rtn and vas.
6212 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6213 Richard Earnshaw <richard.earnshaw@arm.com>
6215 * combine.c (simplify_comparison): Re-canonicalize operands
6217 * config/arm/arm.md (movcond_addsi): New splitter.
6219 2013-07-25 Sterling Augustine <saugustine@google.com>
6221 * dwarf2out.c (size_of_pubnames): Move code to...
6222 (include_pubname_in_output): ...here. New.
6223 (want_pubnames): Rearrange.
6224 (output_pubnames): Call include_pubname_in_output. Move assertion.
6226 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6228 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
6230 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6233 * doc/extend.texi: Remove obsolete builtins. Fix
6234 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
6236 2013-07-25 Jan Hubicka <jh@suse.cz>
6238 * cgraph.c (release_function_body): Break out from ...
6239 (cgraph_release_function_body): ... this one; also release DECL_RESULT
6241 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
6242 old_tree in the map.
6243 (create_specialized_node): Update.
6244 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
6246 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
6247 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
6248 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
6249 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
6252 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6254 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
6255 addsi3_carryin_alt2_<optab>): Correct output template.
6257 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6259 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
6260 Adjust for arm_restrict_it.
6261 Remove trailing whitespace.
6263 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
6265 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
6266 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
6268 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
6270 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
6272 PR rtl-optimization/57960
6273 * lra-constraints.c (process_alt_operands): Use the right mode
6274 when checking strict_low.
6276 2013-07-25 Jan Hubicka <jh@suse.cz>
6278 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
6279 * cgraph.c (cgraph_remove_node): Do not release function body
6280 when in cgraph streaming.
6281 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
6282 in other partitions are not considered reachable; fix handling of
6285 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6287 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
6289 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6294 * config/arm/arm.md ("*sibcall_insn): Replace use of
6295 Ss with US. Adjust output for v5 and v4t.
6296 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
6298 * config/arm/constraints.md ("Ss"): Rename to US.
6300 2013-07-25 Terry Guo <terry.guo@arm.com>
6302 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
6303 shift_add/shift_sub0/shift_sub1 RTXs.
6305 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
6306 Anton Blanchard <anton@au1.ibm.com>
6308 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
6309 (altivec_vpks<VI_char>ss): Likewise.
6310 (altivec_vpks<VI_char>us): Likewise.
6311 (altivec_vpku<VI_char>us): Likewise.
6312 (altivec_vpku<VI_char>um): Likewise.
6314 2013-07-24 David Malcolm <dmalcolm@redhat.com>
6316 Introduce context class.
6318 * Makefile.in (CONTEXT_H): New.
6319 (OBJS): Add context.o.
6320 (toplev.o): Add CONTEXT_H to dependencies.
6323 * toplev.c (general_init): Create the singleton gcc::context instance.
6329 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
6331 PR rtl-optimization/57968
6332 * mode-switching.c (create_pre_exit): Allow instructions that
6333 don't set a return register to need a non-exit mode.
6335 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6336 Anton Blanchard <anton@au1.ibm.com>
6338 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
6339 operands to vperm for little endian.
6340 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
6341 of lvsl to create the control mask for a vperm for little endian.
6343 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6344 Anton Blanchard <anton@au1.ibm.com>
6346 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6347 two operands for little-endian.
6349 2013-07-23 Steve Ellcey <sellcey@mips.com>
6351 * config/mips/mips.c (mips_case_values_threshold): New.
6352 (TARGET_CASE_VALUES_THRESHOLD): Define.
6354 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6355 Anton Blanchard <anton@au1.ibm.com>
6357 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
6358 selection of field for vector splat in little endian mode.
6360 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6362 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
6363 expanders to rs6000.md.
6364 (ior<mode>3): Likewise.
6365 (and<mode>3): Likewise.
6366 (one_cmpl<mode>2): Likewise.
6367 (nor<mode>3): Likewise.
6368 (andc<mode>3): Likewise.
6369 (eqv<mode>3): Likewise.
6370 (nand<mode>3): Likewise.
6371 (orc<mode>3): Likewise.
6373 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
6376 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
6377 to split multi-word logical operations.
6378 (rs6000_split_logical_di): Likewise.
6379 (rs6000_split_logical): Likewise.
6381 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
6382 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
6383 and allow TImode operations in 32-bit.
6384 (vsx_and<mode>3_64bit): Likewise.
6385 (vsx_ior<mode>3_32bit): Likewise.
6386 (vsx_ior<mode>3_64bit): Likewise.
6387 (vsx_xor<mode>3_32bit): Likewise.
6388 (vsx_xor<mode>3_64bit): Likewise.
6389 (vsx_one_cmpl<mode>2_32bit): Likewise.
6390 (vsx_one_cmpl<mode>2_64bit): Likewise.
6391 (vsx_nor<mode>3_32bit): Likewise.
6392 (vsx_nor<mode>3_64bit): Likewise.
6393 (vsx_andc<mode>3_32bit): Likewise.
6394 (vsx_andc<mode>3_64bit): Likewise.
6395 (vsx_eqv<mode>3_32bit): Likewise.
6396 (vsx_eqv<mode>3_64bit): Likewise.
6397 (vsx_nand<mode>3_32bit): Likewise.
6398 (vsx_nand<mode>3_64bit): Likewise.
6399 (vsx_orc<mode>3_32bit): Likewise.
6400 (vsx_orc<mode>3_64bit): Likewise.
6402 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
6403 logical types in GPRs.
6405 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
6406 logical insns to rs6000.md, and allow TImode operations in
6408 (altivec_ior<mode>3): Likewise.
6409 (altivec_xor<mode>3): Likewise.
6410 (altivec_one_cmpl<mode>2): Likewise.
6411 (altivec_nor<mode>3): Likewise.
6412 (altivec_andc<mode>3): Likewise.
6414 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
6415 attributes for moving the 128-bit logical operations into
6417 (BOOL_REGS_OUTPUT): Likewise.
6418 (BOOL_REGS_OP1): Likewise.
6419 (BOOL_REGS_OP2): Likewise.
6420 (BOOL_REGS_UNARY): Likewise.
6421 (BOOL_REGS_AND_CR0): Likewise.
6422 (one_cmpl<mode>2): Add support for DI logical operations on
6423 32-bit, splitting the operations to 32-bit.
6427 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
6428 changes to combine the 32/64-bit code, allow logical operations on
6429 TI mode in 32-bit, and to use similar match_operator patterns like
6430 scalar mode uses. Combine the Altivec and VSX code for logical
6431 operations, and move it here.
6432 (ior<mode>3, 128-bit types): Likewise.
6433 (xor<mode>3, 128-bit types): Likewise.
6434 (one_cmpl<mode>3, 128-bit types): Likewise.
6435 (nor<mode>3, 128-bit types): Likewise.
6436 (andc<mode>3, 128-bit types): Likewise.
6437 (eqv<mode>3, 128-bit types): Likewise.
6438 (nand<mode>3, 128-bit types): Likewise.
6439 (orc<mode>3, 128-bit types): Likewise.
6440 (and<mode>3_internal): Likewise.
6441 (bool<mode>3_internal): Likewise.
6442 (boolc<mode>3_internal1): Likewise.
6443 (boolc<mode>3_internal2): Likewise.
6444 (boolcc<mode>3_internal1): Likewise.
6445 (boolcc<mode>3_internal2): Likewise.
6446 (eqv<mode>3_internal1): Likewise.
6447 (eqv<mode>3_internal2): Likewise.
6448 (one_cmpl1<mode>3_internal): Likewise.
6450 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6452 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6453 Rename flag_stack_usage to flag_stack_usage_info.
6455 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6457 * config/microblaze/sync.md: New file.
6458 * config/microblaze/microblaze.md: Include sync.md
6459 * config/microblaze/microblaze.c: Add print_operand 'y'.
6460 * config/microblaze/constraints.md: Add memory_contraint
6461 'Q' which is a single register.
6463 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6465 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6467 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6469 * reload.c (find_reloads): Exit loop once we find this operand
6470 cannot be reloaded somehow for this alternative.
6472 * reload.c (find_reloads): Exit loop once we find a hard register.
6474 * rtlanal.c (computed_jump_p): Exit loop once we find label
6477 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6479 * cfgloopmanip.c (remove_path): Exit loop after setting
6482 * gensupport.c (subst_dup): Avoid loop if code is not
6483 MATCH_DUP nor MATCH_OP_DUP.
6485 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6487 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6489 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6491 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6492 true for SP_REGNUM if mode == ptr_mode.
6493 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6494 with value R0_REGNUM + 31.
6496 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6498 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6499 pad pointer-typed argument downward.
6501 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6503 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6504 and __ILP32__ when the ILP32 model is in use.
6506 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6508 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6509 (aarch64_load_symref_appropriately): In the case of
6510 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6511 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6512 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6513 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6514 if 'mode' doesn't equal to 'ptr_mode'.
6515 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6516 'vcall_offset'; change to call aarch64_emit_move differently depending
6517 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6518 to calculate the upper bound of 'vcall_offset'.
6519 (aarch64_cannot_force_const_mem): Change to also return true if
6521 (aarch64_legitimize_reload_address): In the case of large
6522 displacements, add new local variable 'xmode' and an assertion
6523 based on it; change to use 'xmode' to generate the new rtx and
6525 (aarch64_asm_trampoline_template): Change to generate the template
6526 differently depending on TARGET_ILP32 or not; change to use
6527 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6528 (aarch64_trampoline_size): Removed.
6529 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
6530 and replace immediate literals with it. Change to use 'ptr_mode'
6531 instead of 'DImode' and call convert_memory_address if the mode
6532 of 'fnaddr' doesn't equal to 'ptr_mode'.
6533 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
6535 (aarch64_elf_asm_destructor): Likewise.
6536 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
6537 on TARGET_ILP32 instead of aarch64_trampoline_size.
6538 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
6539 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
6540 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
6541 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
6542 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
6543 (storewb_pair<GPI:mode>_<P:mode>): ... this.
6544 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
6545 depending on the value of 'mode'.
6546 (add_losym_<mode>): New.
6547 (ldr_got_small_<mode>): New, based on ldr_got_small.
6548 (ldr_got_small): Remove.
6549 (ldr_got_small_sidi): New.
6550 * config/aarch64/iterators.md (P): New.
6551 (PTR): Change to 'ptr_mode' in the condition.
6553 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6555 * config.gcc (aarch64*-*-*): Support --with-abi.
6556 (aarch64*-*-elf): Support --with-multilib-list.
6557 (aarch64*-*-linux*): Likewise.
6558 (supported_defaults): Add abi to aarch64*-*-*.
6559 * configure.ac: Mention AArch64 for --with-multilib-list.
6560 * configure: Re-generated.
6561 * config/aarch64/biarchilp32.h: New file.
6562 * config/aarch64/biarchlp64.h: New file.
6563 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
6565 (MULTILIB_DEFAULTS): Ditto.
6566 (DRIVER_SELF_SPECS): Ditto.
6567 (ASM_SPEC): Update to also substitute -mabi.
6568 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
6569 file whose name depends on -mabi= and -mbig-endian.
6570 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
6572 (POINTER_SIZE): New define.
6573 (POINTERS_EXTEND_UNSIGNED): Ditto.
6574 (enum aarch64_abi_type): New enumeration tag.
6575 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
6576 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
6577 (TARGET_ILP32): New define.
6578 * config/aarch64/aarch64.opt (mabi): New.
6580 (ilp32, lp64): New values for -mabi.
6581 * config/aarch64/t-aarch64 (comma): New define.
6582 (MULTILIB_OPTIONS): Ditto.
6583 (MULTILIB_DIRNAMES): Ditto.
6584 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
6585 * doc/invoke.texi: Document -mabi for AArch64.
6587 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
6589 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
6591 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6592 Anton Blanchard <anton@au1.ibm.com>
6594 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6595 endianness when selecting field to splat.
6597 2013-07-22 Eric Christopher <echristo@gmail.com>
6599 * dwarf2out.c (die_odr_checksum): New function to use
6600 CHECKSUM_ macros and ULEB128 for DIE tag.
6601 (generate_type_signature): Use.
6603 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6605 * config.gcc (sparc*-*-*): Accept leon3 processor.
6606 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6607 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6608 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6609 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6610 (mfix-ut699): Adjust comment.
6611 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6612 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6613 (CPP_CPU_SPEC): Likewise.
6614 (ASM_CPU_SPEC): Likewise.
6615 * config/sparc/sparc.c (leon3_cost): New constant.
6616 (sparc_option_override): Add leon3 support.
6617 (mem_ref): New function.
6618 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6619 (sparc_do_work_around_errata): Look into the instruction in the delay
6620 slot and adjust accordingly. Add fix for the data cache nullify issues
6621 of the UT699. Change insertion position for the NOP.
6622 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6623 (leon3_load): New reservation.
6624 (leon_store): Bump latency to 2.
6625 (grfpu): New automaton.
6626 (grfpu_alu): New unit.
6627 (grfpu_ds): Likewise.
6628 (leon_fp_alu): Adjust.
6629 (leon_fp_mult): Delete.
6630 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6631 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6632 * config/sparc/sparc.md (cpu): Add leon3.
6633 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6635 (atomic_test_and_set): Likewise.
6638 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6640 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6641 default for R5900 targets.
6642 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6643 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6644 * config/mips/mips.c (mips_option_override): Report an error for
6645 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6646 for -march=r5900 -mhard-float.
6648 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6650 * df-problems.c (can_move_insns_across): Exit loop once we
6651 find a non-fixed, non-global register.
6653 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6656 * omega.c (omega_eliminate_red): Break after setting red_found.
6657 (omega_problem_has_red_equations): Similarly after setting found.
6658 (omega_query_variable): Similarly after setting coupled.
6660 2013-07-22 Marek Polacek <polacek@redhat.com>
6662 * gimplify.c: Don't include gimple.h twice.
6664 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6666 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6667 instead of TARGET_THUMB1.
6668 (Pz): New constraint.
6669 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6671 (compare_negsi_si): Likewise.
6672 (compare_addsi2_op0): Likewise.
6673 (compare_addsi2_op1): Likewise.
6674 (addsi3_carryin_<optab>): Likewise.
6675 (addsi3_carryin_alt2_<optab>): Likewise.
6676 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6677 for arm_restrict_it.
6678 (subsi3_carryin): Likewise.
6679 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6680 (minmax_arithsi): Disable for arm_restrict_it.
6681 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6682 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6683 (satsi_<SAT:code>_shift): Likewise.
6684 (arm_shiftsi3): Add alternative for 16-bit encoding.
6685 (arm32_movhf): Disable for arm_restrict_it.
6686 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6687 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6689 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6691 * config/arm/arm.md (attribute "insn"): Delete.
6692 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6693 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6694 (not_shiftsi): Update for attribute change.
6695 (not_shiftsi_compare0): Likewise.
6696 (not_shiftsi_compare0_scratch): Likewise.
6697 (arm_one_cmplsi2): Likewise.
6698 (thumb1_one_cmplsi2): Likewise.
6699 (notsi_compare0): Likewise.
6700 (notsi_compare0_scratch): Likewise.
6701 (thumb1_movdi_insn): Likewise.
6702 (arm_movsi_insn): Likewise.
6703 (movhi_insn_arch4): Likewise.
6704 (movhi_bytes): Likewise.
6705 (arm_movqi_insn): Likewise.
6706 (thumb1_movqi_insn): Likewise.
6707 (arm32_movhf): Likewise.
6708 (thumb1_movhf): Likewise.
6709 (arm_movsf_soft_insn): Likewise.
6710 (thumb1_movsf_insn): Likewise.
6711 (thumb_movdf_insn): Likewise.
6712 (movsicc_insn): Likewise.
6713 (movsfcc_soft_insn): Likewise.
6714 (and_scc): Likewise.
6715 (cond_move): Likewise.
6716 (if_move_not): Likewise.
6717 (if_not_move): Likewise.
6718 (if_shift_move): Likewise.
6719 (if_move_shift): Likewise.
6720 (if_shift_shift): Likewise.
6721 (if_not_arith): Likewise.
6722 (if_arith_not): Likewise.
6723 (cond_move_not): Likewise.
6724 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6725 (neon_mov<mode>): Likewise.
6726 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6727 (thumb2_movsi_vfp): Likewise.
6728 (movsf_vfp): Likewise.
6729 (thumb2_movsf_vfp): Likewise.
6730 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6732 (cortexa7_older_only): Likewise.
6733 (cortexa7_younger): Likewise.
6734 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6735 (1020alu_shift_op): Likewise.
6736 (1020alu_shift_reg_op): Likewise.
6737 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6738 (alu_shift_op): Likewise.
6739 (alu_shift_reg_op): Likewise.
6740 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6741 (11_alu_shift_op): Likewise.
6742 (11_alu_shift_reg_op): Likewise.
6743 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6744 (9_alu_shift_reg_op): Likewise.
6745 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6747 (cortex_a15_alu_shift): Likewise.
6748 (cortex_a15_alu_shift_reg): Likewise.
6749 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6750 (cortex_a5_alu_shift): Likewise.
6751 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6753 (cortex_a53_alu_shift): Likewise.
6754 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6756 (cortex_a7_alu_reg): Likewise.
6757 (cortex_a7_alu_shift): Likewise.
6758 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6759 (cortex_a8_alu_shift): Likewise.
6760 (cortex_a8_alu_shift_reg): Likewise.
6761 (cortex_a8_mov): Likewise.
6762 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6763 (cortex_a9_dp_shift): Likewise.
6764 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6765 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6766 (cortex_r4_mov): Likewise.
6767 (cortex_r4_alu_shift): Likewise.
6768 (cortex_r4_alu_shift_reg): Likewise.
6769 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6770 (526_alu_shift_op): Likewise.
6771 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6772 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6773 (626te_alu_shift_op): Likewise.
6774 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6775 (726te_alu_op): Likewise.
6776 (726te_alu_shift_op): Likewise.
6777 (726te_alu_shift_reg_op): Likewise.
6778 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6779 (mp626_alu_shift_op): Likewise.
6780 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6781 (pj4_alu_e1_conds): Likewise.
6782 (pj4_alu): Likewise.
6783 (pj4_alu_conds): Likewise.
6784 (pj4_shift): Likewise.
6785 (pj4_shift_conds): Likewise.
6786 (pj4_alu_shift): Likewise.
6787 (pj4_alu_shift_conds): Likewise.
6789 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6791 * config/arm/predicates.md (shiftable_operator_strict_it):
6793 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
6794 Disable cond_exec version for arm_restrict_it.
6795 (thumb2_smaxsi3): Convert to generate cond_exec.
6796 (thumb2_sminsi3): Likewise.
6797 (thumb32_umaxsi3): Likewise.
6798 (thumb2_uminsi3): Likewise.
6799 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
6800 (thumb2_neg_abssi2): Likewise.
6801 (thumb2_mov_scc): Add alternative for 16-bit encoding.
6802 (thumb2_movsicc_insn): Adjust alternatives.
6803 (thumb2_mov_negscc): Disable for arm_restrict_it.
6804 (thumb2_mov_negscc_strict_it): New pattern.
6805 (thumb2_mov_notscc_strict_it): New pattern.
6806 (thumb2_mov_notscc): Disable for arm_restrict_it.
6807 (thumb2_ior_scc): Likewise.
6808 (thumb2_ior_scc_strict_it): New pattern.
6809 (thumb2_cond_move): Adjust for arm_restrict_it.
6810 (thumb2_cond_arith): Disable for arm_restrict_it.
6811 (thumb2_cond_arith_strict_it): New pattern.
6812 (thumb2_cond_sub): Adjust for arm_restrict_it.
6813 (thumb2_movcond): Likewise.
6814 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
6815 (thumb2_zero_extendhisi2_v6): Likewise.
6816 (thumb2_zero_extendqisi2_v6): Likewise.
6817 (orsi_notsi_si): Likewise.
6818 (orsi_not_shiftsi_si): Likewise.
6820 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
6822 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
6823 instruction sequence is 1 byte shorter.
6825 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
6827 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
6828 it is not needed after split.
6830 2013-07-20 Iain Sandoe <iain@codesourcery.com>
6833 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
6834 second label for nonlocal goto receivers. Don't output pic base labels
6835 unless we're producing PIC; mark that action unreachable().
6836 (ix86_save_reg): If the function contains a nonlocal label, save the
6838 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
6839 * config/darwin.c (emitted_pic_label_num): New GTY.
6840 (update_pic_label_number_if_needed): New.
6841 (machopic_output_function_base_name): Adjust for nonlocal receiver
6843 (machopic_should_output_picbase_label): New.
6844 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
6845 (nonlocal_goto_receiver): New insn and split.
6847 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
6849 * config/aarch64/aarch64-builtins.c
6850 (aarch64_fold_builtin): Fold abs in all modes.
6851 * config/aarch64/aarch64-simd-builtins.def
6852 (abs): Enable for all modes.
6853 * config/aarch64/arm_neon.h
6854 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
6855 (vabs_f64): Add missing intrinsic.
6857 2013-07-19 Ian Bolton <ian.bolton@arm.com>
6859 * config/aarch64/arm_neon.h (vabs_s64): New function
6861 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
6864 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
6865 * config/avr/avr.md (adjust_len): Add `round'.
6866 * config/avr/avr-protos.h (avr_out_round): New prototype.
6867 (avr_out_plus): Add `out_label' argument.
6868 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
6869 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
6870 Handle the case where `insn' is just a pattern.
6871 (avr_out_bitop): Handle the case where `insn' is just a pattern.
6872 (avr_out_round): New function.
6873 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
6875 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
6877 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6878 Add check for flag_stack_usage to handle -fstack-usage support
6880 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
6882 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
6883 interaction for new Power8 flags and VSX.
6885 2013-07-18 Sriraman Tallam <tmsriram@google.com>
6888 * tree-inline.c (expand_call_inline): Emit errors during
6889 early_inlining only if optimization is not turned on.
6891 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6895 * passes.c (init_optimization_passes): Move the construction of
6896 the pass hierarchy into a new passes.def file.
6898 * Makefile.in (passes.o): Add dependency on passes.def.
6900 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6902 * passes.c (init_optimization_passes): Introduce macros for
6903 constructing the tree of passes (INSERT_PASSES_AFTER,
6904 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
6905 TERMINATE_PASS_LIST).
6907 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
6908 Wei Mi <wmi@google.com>
6910 PR rtl-optimization/57878
6911 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
6913 (reload_pseudo_compare_func): Check nregs first for reload
6916 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6918 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
6920 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
6922 * read-rtl.c (validate_const_int): Once an invalid character is
6923 seen, quit the loop.
6925 * gengtype.c (write_roots): Similarly once we find the "deletable"
6926 or "if_marked" option.
6928 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6930 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
6931 "xtab" and "sat". Move value "clz" from here to ...
6932 (attriubte "type"): ... here.
6933 (satsi_<SAT:code>): Delete "insn" attribute.
6934 (satsi_<SAT:code>_shift): Likewise.
6935 (arm_zero_extendqisi2addsi): Likewise.
6936 (arm_extendqisi2addsi): Likewise.
6937 (clzsi2): Update for attribute changes.
6938 (rbitsi2): Likewise.
6939 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
6941 (arm_usatsihi): Likewise.
6942 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6944 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6946 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
6947 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
6948 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
6949 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
6950 in alphabetical order.
6951 (attribute "core_cycles"): Update for attribute changes.
6952 (arm_addsi3): Likewise.
6953 (addsi3_compare0): Likewise.
6954 (addsi3_compare0_scratch): Likewise.
6955 (addsi3_compare_op1): Likewise.
6956 (addsi3_compare_op2): Likewise.
6957 (compare_addsi2_op0): Likewise.
6958 (compare_addsi2_op1): Likewise.
6959 (addsi3_carryin_shift_<optab>): Likewise.
6960 (subsi3_carryin_shift): Likewise.
6961 (rsbsi3_carryin_shift): Likewise.
6962 (arm_subsi3_insn): Likewise.
6963 (subsi3_compare0): Likewise.
6964 (subsi3_compare): Likewise.
6965 (arm_andsi3_insn): Likewise.
6966 (thumb1_andsi3_insn): Likewise.
6967 (andsi3_compare0): Likewise.
6968 (andsi3_compare0_scratch): Likewise.
6969 (zeroextractsi_compare0_scratch
6970 (andsi_not_shiftsi_si): Likewise.
6971 (iorsi3_insn): Likewise.
6972 (iorsi3_compare0): Likewise.
6973 (iorsi3_compare0_scratch): Likewise.
6974 (arm_xorsi3): Likewise.
6975 (thumb1_xorsi3_insn): Likewise.
6976 (xorsi3_compare0): Likewise.
6977 (xorsi3_compare0_scratch): Likewise.
6978 (satsi_<SAT:code>_shift): Likewise.
6980 (arm_shiftsi3): Likewise.
6981 (shiftsi3_compare0): Likewise.
6982 (not_shiftsi): Likewise.
6983 (not_shiftsi_compare0): Likewise.
6984 (not_shiftsi_compare0_scratch): Likewise.
6985 (arm_one_cmplsi2): Likewise.
6986 (thumb_one_complsi2): Likewise.
6987 (notsi_compare0): Likewise.
6988 (notsi_compare0_scratch): Likewise.
6989 (thumb1_zero_extendhisi2): Likewise.
6990 (arm_zero_extendhisi2): Likewise.
6991 (arm_zero_extendhisi2_v6): Likewise.
6992 (arm_zero_extendhisi2addsi): Likewise.
6993 (thumb1_zero_extendqisi2): Likewise.
6994 (thumb1_zero_extendqisi2_v6): Likewise.
6995 (arm_zero_extendqisi2): Likewise.
6996 (arm_zero_extendqisi2_v6): Likewise.
6997 (arm_zero_extendqisi2addsi): Likewise.
6998 (thumb1_extendhisi2): Likewise.
6999 (arm_extendhisi2): Likewise.
7000 (arm_extendhisi2_v6): Likewise.
7001 (arm_extendqisi): Likewise.
7002 (arm_extendqisi_v6): Likewise.
7003 (arm_extendqisi2addsi): Likewise.
7004 (thumb1_extendqisi2): Likewise.
7005 (thumb1_movdi_insn): Likewise.
7006 (arm_movsi_insn): Likewise.
7007 (movsi_compare0): Likewise.
7008 (movhi_insn_arch4): Likewise.
7009 (movhi_bytes): Likewise.
7010 (arm_movqi_insn): Likewise.
7011 (thumb1_movqi_insn): Likewise.
7012 (arm32_movhf): Likewise.
7013 (thumb1_movhf): Likewise.
7014 (arm_movsf_soft_insn): Likewise.
7015 (thumb1_movsf_insn): Likewise.
7016 (movdf_soft_insn): Likewise.
7017 (thumb_movdf_insn): Likewise.
7018 (arm_cmpsi_insn): Likewise.
7019 (cmpsi_shiftsi): Likewise.
7020 (cmpsi_shiftsi_swp): Likewise.
7021 (arm_cmpsi_negshiftsi_si): Likewise.
7022 (movsicc_insn): Likewise.
7023 (movsfcc_soft_insn): Likewise.
7024 (arith_shiftsi): Likewise.
7025 (arith_shiftsi_compare0
7026 (arith_shiftsi_compare0_scratch
7027 (sub_shiftsi): Likewise.
7028 (sub_shiftsi_compare0
7029 (sub_shiftsi_compare0_scratch
7030 (and_scc): Likewise.
7031 (cond_move): Likewise.
7032 (if_plus_move): Likewise.
7033 (if_move_plus): Likewise.
7034 (if_move_not): Likewise.
7035 (if_not_move): Likewise.
7036 (if_shift_move): Likewise.
7037 (if_move_shift): Likewise.
7038 (if_shift_shift): Likewise.
7039 (if_not_arith): Likewise.
7040 (if_arith_not): Likewise.
7041 (cond_move_not): Likewise.
7042 (thumb1_ashlsi3): Set type attribute.
7043 (thumb1_ashrsi3): Likewise.
7044 (thumb1_lshrsi3): Likewise.
7045 (thumb1_rotrsi3): Likewise.
7046 (shiftsi3_compare0_scratch): Likewise.
7047 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
7048 (neon_mov<mode>): Likewise.
7049 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
7051 (thumb2_movsi_insn): Likewise.
7052 (thumb2_cmpsi_neg_shiftsi): Likewise.
7053 (thumb2_extendqisi_v6): Likewise.
7054 (thumb2_zero_extendhisi2_v6): Likewise.
7055 (thumb2_zero_extendqisi2_v6): Likewise.
7056 (thumb2_shiftsi3_short): Likewise.
7057 (thumb2_addsi3_compare0_scratch): Likewise.
7058 (orsi_not_shiftsi_si): Likewise.
7059 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7060 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
7062 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7063 (1020alu_shift_op): Likewise.
7064 (1020alu_shift_reg_op): Likewise.
7065 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7066 (alu_shift_op): Likewise.
7067 (alu_shift_reg_op): Likewise.
7068 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7069 (11_alu_shift_op): Likewise.
7070 (11_alu_shift_reg_op): Likewise.
7071 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7072 (9_alu_shift_reg_op): Likewise.
7073 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7075 (cortex_a15_alu_shift): Likewise.
7076 (cortex_a15_alu_shift_reg): Likewise.
7077 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
7079 (cortex_a5_alu_shift): Likewise.
7080 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
7082 (cortex_a53_alu_shift): Likewise.
7083 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7085 (cortex_a7_alu_reg): Likewise.
7086 (cortex_a7_alu_shift): Likewise.
7087 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
7089 (cortex_a8_alu_shift): Likewise.
7090 (cortex_a8_alu_shift_reg): Likewise.
7091 (cortex_a8_mov): Likewise.
7092 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7093 (cortex_a9_dp_shift): Likewise.
7094 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
7096 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
7098 (cortex_r4_mov): Likewise.
7099 (cortex_r4_alu_shift): Likewise.
7100 (cortex_r4_alu_shift_reg): Likewise.
7101 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7102 (526_alu_shift_op): Likewise.
7103 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7104 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7105 (626te_alu_shift_op): Likewise.
7106 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
7107 (726te_alu_op): Likewise.
7108 (726te_alu_shift_op): Likewise.
7109 (726te_alu_shift_reg_op): Likewise.
7110 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7111 (mp626_alu_shift_op): Likewise.
7112 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
7113 (pj4_alu_e1_conds): Likewise.
7114 (pj4_alu): Likewise.
7115 (pj4_alu_conds): Likewise.
7116 (pj4_shift): Likewise.
7117 (pj4_shift_conds): Likewise.
7118 (pj4_alu_shift): Likewise.
7119 (pj4_alu_shift_conds): Likewise.
7120 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7122 (cortexa7_older_only): Likewise.
7123 (cortexa7_younger): Likewise.
7125 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7127 * ipa-pure-const.c (generate_summary): Rename to...
7128 (pure_const_generate_summary): ... this.
7130 2013-07-17 Iain Sandoe <iain@codesourcery.com>
7132 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
7134 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
7137 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
7140 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7142 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
7143 enabled without -march=zEC12.
7144 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
7147 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7149 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
7150 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
7151 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
7152 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
7153 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
7154 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
7155 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
7157 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
7158 and ISA_HAS_NMADD3_NMSUB3.
7159 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
7160 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
7161 (nmsub4<mode>, nmsub3<mode>): Likewise.
7162 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
7164 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7166 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
7167 TARGET_MIPS5400 checking.
7169 2013-07-16 Jakub Jelinek <jakub@redhat.com>
7170 Peter Bergner <bergner@vnet.ibm.com>
7172 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
7173 registers in the comment.
7174 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
7175 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
7176 rather than FIRST_PSEUDO_REGISTERS.
7178 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
7180 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
7181 enable extra ISA flags with TARGET_HTM.
7183 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7185 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
7188 2013-07-15 Cong Hou <congh@google.com>
7190 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
7191 in compare function for sorting.
7193 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
7195 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
7196 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
7197 * config/rs6000/rs6000.opt: Add -mhtm option.
7198 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
7199 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
7200 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7201 __HTM__ if the HTM instructions are available.
7202 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
7203 htm_spr_reg_operand): New define_predicates.
7204 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
7205 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
7207 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
7208 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
7209 HTM builtin functions.
7210 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
7211 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
7212 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
7213 (rs6000_builtin_mask_calculate): Likewise.
7214 (rs6000_option_override_internal): Likewise.
7215 (bdesc_htm): Add new HTM builtin support.
7216 (htm_spr_num): New function.
7217 (htm_spr_regno): Likewise.
7218 (rs6000_htm_spr_icode): Likewise.
7219 (htm_expand_builtin): Likewise.
7220 (htm_init_builtins): Likewise.
7221 (rs6000_expand_builtin): Add support for HTM builtin functions.
7222 (rs6000_init_builtins): Likewise.
7223 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
7225 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
7226 (TARGET_HTM, MASK_HTM): Define macros.
7227 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
7228 (FIXED_REGISTERS): Likewise.
7229 (CALL_USED_REGISTERS): Likewise.
7230 (CALL_REALLY_USED_REGISTERS): Likewise.
7231 (REG_ALLOC_ORDER): Likewise.
7232 (enum reg_class): Likewise.
7233 (REG_CLASS_NAMES): Likewise.
7234 (REG_CLASS_CONTENTS): Likewise.
7235 (REGISTER_NAMES): Likewise.
7236 (ADDITIONAL_REGISTER_NAMES): Likewise.
7237 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
7238 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
7239 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
7240 * config/rs6000/htm.md: New file.
7241 * config/rs6000/htmintrin.h: New file.
7242 * config/rs6000/htmxlintrin.h: New file.
7244 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
7246 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7247 Define SYMBOL_TINY_GOT, update comment.
7248 * config/aarch64/aarch64.c
7249 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
7250 (aarch64_expand_mov_immediate): Likewise.
7251 (aarch64_print_operand): Likewise.
7252 (aarch64_classify_symbol): Likewise.
7253 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
7254 (ldr_got_tiny): Define.
7256 2013-07-13 Tobias Grosser <tobias@grosser.es>
7258 PR tree-optimization/54094
7259 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
7260 scheduling dimension for the parallelism check from the polyhedral
7261 information in the AST.
7262 * graphite-dependences.c (carries_deps): Do not assume the schedule is
7265 2013-07-13 Jason Merrill <jason@redhat.com>
7267 * print-tree.c (debug_vec_tree): Use debug_raw.
7268 (debug_raw (vec<tree, va_gc> &)): New.
7269 (debug_raw (vec<tree, va_gc> *)): New.
7270 * tree.h: Declare them.
7272 2013-07-13 Bin Cheng <bin.cheng@arm.com>
7274 * ifcvt.c (ifcvt_after_combine): New static variable.
7275 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
7277 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
7278 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
7279 rest_of_handle_if_after_reload): Pass new argument for if_convert.
7281 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
7283 * config/mips/mips.c (mips_expand_call): Remove empty statement.
7285 2013-07-12 Michael Matz <matz@suse.de>
7288 * convert.c (convert_to_real): Reject non-float inner types.
7290 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
7292 * config/aarch64/aarch64-protos.h
7293 (aarch64_simd_immediate_valid_for_move): Remove.
7294 * config/aarch64/aarch64.c (simd_immediate_info): New member.
7295 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
7297 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
7299 2013-07-11 Steve Ellcey <sellcey@mips.com>
7301 * config/mips/mips.c (mips_conditional_register_usage): Do not
7302 use t[0-7] registers in MIPS16 mode when optimizing for size.
7304 2013-07-11 Sriraman Tallam <tmsriram@google.com>
7306 * config/i386/i386.c (dispatch_function_versions): Fix array
7307 indexing of function_version_info to match actual_versions.
7309 2013-07-11 Teresa Johnson <tejohnson@google.com>
7311 * vec.h (struct va_gc): Move release out-of-line.
7312 (va_gc::release): Call ggc_free on released vec.
7314 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7316 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
7317 Require GOT register as additional operand in UNSPEC.
7318 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
7319 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7320 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7321 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
7322 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
7323 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
7324 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
7326 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
7329 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
7330 name seen by assembler/linker rather if available.
7332 2013-07-11 Andreas Schwab <schwab@suse.de>
7334 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
7336 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
7338 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
7340 2013-07-10 Joseph Myers <joseph@codesourcery.com>
7342 * doc/tm.texi.in: Move hook documentation to ....
7343 * target.def: ... here.
7345 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
7347 * doc/tm.texi: Regenerate.
7349 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
7352 * doc/invoke.texi: Document -Wconditionally-supported.
7354 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7357 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
7360 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7363 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
7364 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
7365 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
7366 Remove duplicate devices.
7367 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
7368 * config/avr/t-multilib: Regenerate.
7369 * config/avr/avr-tables.opt: Regenerate.
7370 * doc/avr-mmcu.texi: Regenerate.
7372 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7375 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
7377 2013-07-10 Graham Stott <graham.stott@btinternet.com>
7379 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
7380 the cost of MULT when optimizing for size.
7382 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7384 * config/cr16/cr16-protos.h: Don't include target.h.
7386 2013-07-09 Joseph Myers <joseph@codesourcery.com>
7388 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
7389 adjust register size for TDmode and TFmode for VSX registers.
7391 2013-07-08 Kai Tietz <ktietz@redhat.com>
7394 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
7395 hook_bool_const_tree_true.
7397 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7399 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
7400 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
7401 * config/s390/s390.md: Define FPR*_REGNUM constants.
7402 Fix FPR2_REGNUM constant (18 -> 17).
7403 ("*trunc<BFP:mode><DFP_ALL:mode>2")
7404 ("*trunc<DFP_ALL:mode><BFP:mode>2")
7405 ("trunc<BFP:mode><DFP_ALL:mode>2")
7406 ("trunc<DFP_ALL:mode><BFP:mode>2")
7407 ("*extend<BFP:mode><DFP_ALL:mode>2")
7408 ("*extend<DFP_ALL:mode><BFP:mode>2")
7409 ("extend<BFP:mode><DFP_ALL:mode>2")
7410 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
7413 2013-07-08 Graham Stott <graham.stott@btinternet.com>
7415 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
7417 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7419 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
7420 and cfun_fpr_bit_p to cfun_fpr_save_p.
7421 (s390_frame_area, s390_register_info, s390_frame_info)
7422 (s390_emit_prologue, s390_emit_epilogue)
7423 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
7425 * config/s390/s390.h: Define *_REGNUM macros for floating point
7428 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
7430 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
7432 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
7434 PR rtl-optimization/57786
7435 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
7436 and break out of the loop when it is set to false.
7438 2013-07-08 Jakub Jelinek <jakub@redhat.com>
7441 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7442 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7443 (const_int 63)) 0)).
7444 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7445 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7446 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7448 PR rtl-optimization/57829
7449 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7450 mask bits outside of mode are just sign-extension from mode to HWI.
7452 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7454 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7455 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7456 adjust_address instead of change_address to keep info about alignment.
7457 (emit_strmov): Remove.
7458 (emit_memmov): New function.
7459 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7460 (expand_movmem_epilogue): Likewise and return updated rtx for
7462 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7463 destination and source.
7464 (decide_alignment): Refactor, handle vector_loop.
7465 (ix86_expand_movmem): Likewise.
7466 (ix86_expand_setmem): Likewise.
7467 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7469 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
7471 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7472 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7474 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7476 * config/i386/sse.md (sse_movlhps): Change alternative 3
7477 of operand 2 to "m".
7479 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7482 * config/i386/sse.md (iptr): New mode attribute.
7483 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7484 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7485 (<sse>_vmmul<mode>3): Ditto.
7486 (<sse>_vmdiv<mode>3): Ditto.
7487 (sse_vmrcpv4sf2): Ditto.
7488 (<sse>_vmsqrt<mode>2): Ditto.
7489 (sse_vmrsqrtv4sf2): Ditto.
7490 (<sse>_vm<code><mode>3): Ditto.
7491 (avx_vmcmp<mode>3): Ditto.
7492 (<sse>_vmmaskcmp<mode>3): Ditto.
7493 (<sse>_comi): Ditto.
7494 (<sse>_ucomi): Ditto.
7495 (*xop_vmfrcz_<mode>): Ditto.
7496 (*fmai_fmadd_<mode>): Ditto.
7497 (*fmai_fmsub_<mode>): Ditto.
7498 (*fmai_fnmadd_<mode>): Ditto.
7499 (*fmai_fnmsub_<mode>): Ditto.
7500 (*fma4i_vmfmadd_<mode>): Ditto.
7501 (*fma4i_vmfmsub_<mode>): Ditto.
7502 (*fma4i_vmfnmadd_<mode>): Ditto.
7503 (*fma4i_vmfnmsub_<mode>): Ditto.
7504 (*xop_vmfrcz_<mode>): Ditto.
7505 (sse_cvtps2pi): Ditto.
7506 (sse_cvttps2pi): Ditto.
7507 (sse_cvtss2si): Ditto.
7508 (sse_cvtss2si_2): Ditto.
7509 (sse_cvtss2siq_2): Ditto.
7510 (sse_cvttss2si): Ditto.
7511 (sse_cvttss2siq): Ditto.
7512 (sse_cvtsd2si): Ditto.
7513 (sse_cvtsd2si_2): Ditto.
7514 (sse_cvtsd2siq_2): Ditto.
7515 (sse_cvttsd2si): Ditto.
7516 (sse_cvttsd2siq): Ditto.
7517 (sse_cvtsd2ss): Ditto.
7518 (sse_cvtss2sd): Ditto.
7519 (avx2_pbroadcast<mode>): Ditto.
7520 (avx2_pbroadcast<mode>_1): Ditto.
7521 (*avx_vperm_broadcast_v4sf): Ditto.
7523 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7524 (sse_movlhps): Ditto.
7525 (sse_storehps): Ditto.
7526 (sse_loadhps): Ditto.
7527 (sse_storelps): Ditto.
7528 (sse_loadlps): Ditto.
7529 (*vec_concatv4sf): Ditto.
7530 (*vec_interleave_highv2df): Ditto.
7531 (*vec_interleave_lowv2df): Ditto.
7532 (*vec_extractv2df_1_sse): Ditto.
7533 (*vec_extractv2df_0_sse): Ditto.
7534 (sse2_storelpd): Ditto.
7535 (sse2_loadlpd): Ditto.
7536 (sse2_movsd): Ditto.
7537 (*vec_concatv4si): Ditto.
7538 (vec_concatv2di): Ditto.
7540 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
7541 for Intel asm dialect.
7542 (mmx_punpcklwd): Ditto.
7543 (mmx_punpckldq): Ditto.
7545 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
7546 for intel assembler dialect.
7548 2013-07-06 Jakub Jelinek <jakub@redhat.com>
7551 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
7552 for BUILT_IN_C{LZ,LRSB}*.
7553 * tree.h (CASE_INT_FN): Add FN##IMAX case.
7554 * tree-vrp.c (extract_range_basic): Handle
7555 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
7556 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
7557 fall thru to code calling set_value*.
7558 * builtins.c (expand_builtin): Remove *IMAX cases.
7559 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
7560 if width is bigger than 2*HWI.
7562 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
7564 PR rtl-optimization/55342
7565 * lra-int.h (lra_subreg_reload_pseudos): New.
7566 * lra.c: Add undoing optional reloads to the block diagram.
7567 (lra_subreg_reload_pseudos): New.
7568 (lra_optional_reload_pseudos): Change comments.
7569 (lra): Init and clear lra_subreg_reload_pseudos. Clear
7570 lra_optional_reload_pseudos after undo transformations.
7571 * lra-assigns.c (pseudo_prefix_title): New.
7572 (lra_setup_reg_renumber): Use it.
7573 (spill_for): Ditto. Check subreg reload pseudos too.
7574 (assign_by_spills): Consider subreg reload pseudos too.
7575 * lra-constraints.c (simplify_operand_subreg): Use
7576 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
7577 (curr_insn_transform): Recognize and do optional reloads.
7578 (undo_optional_reloads): New.
7579 (lra_undo_inheritance): Call undo_optional_reloads.
7581 2013-07-05 Thomas Quinot <quinot@adacore.com>
7583 * tree-complex.c (expand_complex_operations_1): Fix typo.
7585 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
7587 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
7588 (tune_params): New member 'const vec_costs'.
7589 * config/aarch64/aarch64.c (generic_vector_cost): New.
7590 (generic_tunings): New member 'generic_vector_cost'.
7591 (aarch64_builtin_vectorization_cost): New.
7592 (aarch64_add_stmt_cost): New.
7593 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7594 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7596 2013-07-03 Jakub Jelinek <jakub@redhat.com>
7599 * config/i386/predicates.md (vsib_address_operand): Disallow
7600 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7602 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7605 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7606 expand_builtin_setjmp_receiver.
7607 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7608 with NULL for the label parameter.
7609 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7610 the frame-pointer. Adjust comments.
7611 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7612 only if LABEL is non-NULL.
7614 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7616 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7617 (ARM_ABI_AAPCS64): Ditto.
7619 (ARM_DEFAULT_ABI): Ditto.
7621 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7623 * config/aarch64/aarch64-builtins.c
7624 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7625 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7627 * config/aarch64/aarch64-simd.md
7628 (aarch64_ld1<VALL:mode>): New.
7629 (aarch64_st1<VALL:mode>): Likewise.
7630 * config/aarch64/arm_neon.h
7631 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7633 2013-07-02 Sriraman Tallam <tmsriram@google.com>
7635 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7637 (ix86_option_override_internal):Turn on all -mavx target flags by
7638 default as they are dependent on AVX anyway.
7640 2013-07-02 Cary Coutant <ccoutant@google.com>
7642 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7644 (loc_checksum_ordered): Likewise.
7645 (hash_loc_operands): Remove inline keyword.
7647 2013-07-02 Jakub Jelinek <jakub@redhat.com>
7649 PR tree-optimization/57741
7650 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7651 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7652 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7653 Allow REAL_CST step_exprs if flag_associative_math.
7654 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7656 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7658 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
7660 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7662 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7664 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7666 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7668 (iorsi3_insn): Likewise.
7669 (arm_xorsi3): Likewise.
7671 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7673 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7675 (attribute "type"): ... here, and prefix with "wmmx_".
7676 (attribute "core_cycles"): Update for attribute changes.
7677 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7678 (tbcstv4hi): Likewise.
7679 (tbcstv2si): Likewise.
7680 (iwmmxt_iordi3): Likewise.
7681 (iwmmxt_xordi3): Likewise.
7682 (iwmmxt_anddi3): Likewise.
7683 (iwmmxt_nanddi3): Likewise.
7684 (iwmmxt_arm_movdi): Likewise.
7685 (iwmmxt_movsi_insn): Likewise.
7686 (mov<mode>_internal): Likewise.
7687 (and<mode>3_iwmmxt): Likewise.
7688 (ior<mode>3_iwmmxt): Likewise.
7689 (xor<mode>3_iwmmxt): Likewise.
7690 (add<mode>3_iwmmxt): Likewise.
7691 (ssaddv8qi3): Likewise.
7692 (ssaddv4hi3): Likewise.
7693 (ssaddv2si3): Likewise.
7694 (usaddv8qi3): Likewise.
7695 (usaddv4hi3): Likewise.
7696 (usaddv2si3): Likewise.
7697 (sub<mode>3_iwmmxt): Likewise.
7698 (sssubv8qi3): Likewise.
7699 (sssubv4hi3): Likewise.
7700 (sssubv2si3): Likewise.
7701 (ussubv8qi3): Likewise.
7702 (ussubv4hi3): Likewise.
7703 (ussubv2si3): Likewise.
7704 (mulv4hi3_iwmmxt): Likewise.
7705 (smulv4hi3_highpart): Likewise.
7706 (umulv4hi3_highpart): Likewise.
7707 (iwmmxt_wmacs): Likewise.
7708 (iwmmxt_wmacsz): Likewise.
7709 (iwmmxt_wmacu): Likewise.
7710 (iwmmxt_wmacuz): Likewise.
7711 (iwmmxt_clrdi): Likewise.
7712 (iwmmxt_clrv8qi): Likewise.
7713 (iwmmxt_clr4hi): Likewise.
7714 (iwmmxt_clr2si): Likewise.
7715 (iwmmxt_uavgrndv8qi3): Likewise.
7716 (iwmmxt_uavgrndv4hi3): Likewise.
7717 (iwmmxt_uavgv8qi3): Likewise.
7718 (iwmmxt_uavgv4hi3): Likewise.
7719 (iwmmxt_tinsrb): Likewise.
7720 (iwmmxt_tinsrh): Likewise.
7721 (iwmmxt_tinsrw): Likewise.
7722 (iwmmxt_textrmub): Likewise.
7723 (iwmmxt_textrmsb): Likewise.
7724 (iwmmxt_textrmuh): Likewise.
7725 (iwmmxt_textrmsh): Likewise.
7726 (iwmmxt_textrmw): Likewise.
7727 (iwmxxt_wshufh): Likewise.
7728 (eqv8qi3): Likewise.
7729 (eqv4hi3): Likewise.
7730 (eqv2si3): Likewise.
7731 (gtuv8qi3): Likewise.
7732 (gtuv4hi3): Likewise.
7733 (gtuv2si3): Likewise.
7734 (gtv8qi3): Likewise.
7735 (gtv4hi3): Likewise.
7736 (gtv2si3): Likewise.
7737 (smax<mode>3_iwmmxt): Likewise.
7738 (umax<mode>3_iwmmxt): Likewise.
7739 (smin<mode>3_iwmmxt): Likewise.
7740 (umin<mode>3_iwmmxt): Likewise.
7741 (iwmmxt_wpackhss): Likewise.
7742 (iwmmxt_wpackwss): Likewise.
7743 (iwmmxt_wpackdss): Likewise.
7744 (iwmmxt_wpackhus): Likewise.
7745 (iwmmxt_wpackwus): Likewise.
7746 (iwmmxt_wpackdus): Likewise.
7747 (iwmmxt_wunpckihb): Likewise.
7748 (iwmmxt_wunpckihh): Likewise.
7749 (iwmmxt_wunpckihw): Likewise.
7750 (iwmmxt_wunpckilb): Likewise.
7751 (iwmmxt_wunpckilh): Likewise.
7752 (iwmmxt_wunpckilw): Likewise.
7753 (iwmmxt_wunpckehub): Likewise.
7754 (iwmmxt_wunpckehuh): Likewise.
7755 (iwmmxt_wunpckehuw): Likewise.
7756 (iwmmxt_wunpckehsb): Likewise.
7757 (iwmmxt_wunpckehsh): Likewise.
7758 (iwmmxt_wunpckehsw): Likewise.
7759 (iwmmxt_wunpckelub): Likewise.
7760 (iwmmxt_wunpckeluh): Likewise.
7761 (iwmmxt_wunpckeluw): Likewise.
7762 (iwmmxt_wunpckelsb): Likewise.
7763 (iwmmxt_wunpckelsh): Likewise.
7764 (iwmmxt_wunpckelsw): Likewise.
7765 (ror<mode>3): Likewise.
7766 (ashr<mode>3_iwmmxt): Likewise.
7767 (lshr<mode>3_iwmmxt): Likewise.
7768 (ashl<mode>3_iwmmxt): Likewise.
7769 (ror<mode>3_di): Likewise.
7770 (ashr<mode>3_di): Likewise.
7771 (lshr<mode>3_di): Likewise.
7772 (ashl<mode>3_di): Likewise.
7773 (iwmmxt_wmadds): Likewise.
7774 (iwmmxt_wmaddu): Likewise.
7775 (iwmmxt_tmia): Likewise.
7776 (iwmmxt_tmiaph): Likewise.
7777 (iwmmxt_tmiabb): Likewise.
7778 (iwmmxt_tmiatb): Likewise.
7779 (iwmmxt_tmiabt): Likewise.
7780 (iwmmxt_tmiatt): Likewise.
7781 (iwmmxt_tmovmskb): Likewise.
7782 (iwmmxt_tmovmskh): Likewise.
7783 (iwmmxt_tmovmskw): Likewise.
7784 (iwmmxt_waccb): Likewise.
7785 (iwmmxt_wacch): Likewise.
7786 (iwmmxt_waccw): Likewise.
7787 (iwmmxt_waligni): Likewise.
7788 (iwmmxt_walignr): Likewise.
7789 (iwmmxt_walignr0): Likewise.
7790 (iwmmxt_walignr1): Likewise.
7791 (iwmmxt_walignr2): Likewise.
7792 (iwmmxt_walignr3): Likewise.
7793 (iwmmxt_wsadb): Likewise.
7794 (iwmmxt_wsadh): Likewise.
7795 (iwmmxt_wsadbz): Likewise.
7796 (iwmmxt_wsadhz): Likewise.
7797 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
7798 (iwmmxt_wabsdiffb): Likewise.
7799 (iwmmxt_wabsdiffh): Likewise.
7800 (iwmmxt_wabsdiffw): Likewise.
7801 (iwmmxt_waddsubhx): Likewise
7802 (iwmmxt_wsubaddhx): Likewise.
7803 (addc<mode>3): Likewise.
7804 (iwmmxt_avg4): Likewise.
7805 (iwmmxt_avg4r): Likewise.
7806 (iwmmxt_wmaddsx): Likewise.
7807 (iwmmxt_wmaddux): Likewise.
7808 (iwmmxt_wmaddsn): Likewise.
7809 (iwmmxt_wmaddun): Likewise.
7810 (iwmmxt_wmulwsm): Likewise.
7811 (iwmmxt_wmulwum): Likewise.
7812 (iwmmxt_wmulsmr): Likewise.
7813 (iwmmxt_wmulumr): Likewise.
7814 (iwmmxt_wmulwsmr): Likewise.
7815 (iwmmxt_wmulwumr): Likewise.
7816 (iwmmxt_wmulwl): Likewise.
7817 (iwmmxt_wqmulm): Likewise.
7818 (iwmmxt_wqmulwm): Likewise.
7819 (iwmmxt_wqmulmr): Likewise.
7820 (iwmmxt_wqmulwmr): Likewise.
7821 (iwmmxt_waddbhusm): Likewise.
7822 (iwmmxt_waddbhusl): Likewise.
7823 (iwmmxt_wqmiabb): Likewise.
7824 (iwmmxt_wqmiabt): Likewise.
7825 (iwmmxt_wqmiatb): Likewise.
7826 (iwmmxt_wqmiatt): Likewise.
7827 (iwmmxt_wqmiabbn): Likewise.
7828 (iwmmxt_wqmiabtn): Likewise.
7829 (iwmmxt_wqmiatbn): Likewise.
7830 (iwmmxt_wqmiattn): Likewise.
7831 (iwmmxt_wmiabb): Likewise.
7832 (iwmmxt_wmiabt): Likewise.
7833 (iwmmxt_wmiatb): Likewise.
7834 (iwmmxt_wmiatt): Likewise.
7835 (iwmmxt_wmiabbn): Likewise.
7836 (iwmmxt_wmiabtn): Likewise.
7837 (iwmmxt_wmiatbn): Likewise.
7838 (iwmmxt_wmiattn): Likewise.
7839 (iwmmxt_wmiawbb): Likewise.
7840 (iwmmxt_wmiawbt): Likewise.
7841 (iwmmxt_wmiawtb): Likewise.
7842 (iwmmxt_wmiawtt): Likewise.
7843 (iwmmxt_wmiawbbn): Likewise.
7844 (iwmmxt_wmiawbtn): Likewise.
7845 (iwmmxt_wmiawtbn): Likewise.
7846 (iwmmxt_wmiawttn): Likewise.
7847 (iwmmxt_wmerge): Likewise.
7848 (iwmmxt_tandc<mode>3): Likewise.
7849 (iwmmxt_torc<mode>3): Likewise.
7850 (iwmmxt_torvsc<mode>3): Likewise.
7851 (iwmmxt_textrc<mode>3): Likewise.
7852 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
7853 (wmmxt_pack): Likewise.
7854 (wmmxt_mult_c1): Likewise.
7855 (wmmxt_mult_c2): Likewise.
7856 (wmmxt_alu_c1): Likewise.
7857 (wmmxt_alu_c2): Likewise.
7858 (wmmxt_alu_c3): Likewise.
7859 (wmmxt_transfer_c1): Likewise.
7860 (wmmxt_transfer_c2): Likewise.
7861 (wmmxt_transfer_c3): Likewise.
7862 (marvell_f_iwmmxt_wstr): Likewise.
7863 (marvell_f_iwmmxt_wldr): Likewise.
7865 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
7867 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
7869 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7872 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7873 * lra-constraints.c (need_for_split_p): Check call used hard regs
7874 living through calls.
7876 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
7877 call used regs for call insn.
7879 2013-06-28 Jakub Jelinek <jakub@redhat.com>
7882 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
7883 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
7885 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7887 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
7889 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7891 * lra-constraints.c (need_for_split_p): Check call used hard regs
7892 living through calls.
7894 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
7897 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
7898 to tie with any other modes. Eliminate Altivec vector mode tests,
7899 since these are a subset of ALTIVEC or VSX vector modes. Simplify
7900 code, to return 0 if testing MODE2 for a condition, if we've
7901 already tested MODE1 for the same condition.
7903 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7905 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
7908 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7910 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7911 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
7913 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7915 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
7917 (aarch64_symbolic_constant_p): Remove.
7918 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
7919 static. Fix line length and white space.
7920 (aarch64_symbolic_constant_p): Remove.
7921 * config/aarch64/predicates.md (aarch64_valid_symref):
7922 Use aarch64_classify_symbol_expression.
7924 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7926 * config/arm/constraints.md (Ts): New constraint.
7927 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
7929 (compare_scc): Use "Ts" constraint for operand 0.
7930 (ior_scc_scc): Likewise.
7931 (and_scc_scc): Likewise.
7932 (and_scc_scc_nodom): Likewise.
7933 (ior_scc_scc_cmp): Likewise for operand 7.
7934 (and_scc_scc_cmp): Likewise.
7935 * config/arm/thumb2.md (thumb2_movsi_insn):
7936 Add alternatives for 16-bit encodings.
7937 (thumb2_movhi_insn): Likewise.
7938 (thumb2_movsicc_insn): Likewise.
7939 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
7940 (thumb2_negscc): Use "Ts" constraint.
7941 Move mvn instruction outside cond_exec block.
7942 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
7943 for 16-bit encodings.
7945 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7947 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
7949 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
7950 (mulsi3subsi): Likewise.
7951 (mulsidi3adddi): Likewise.
7952 (mulsidi3_v6): Likewise.
7953 (umulsidi3_v6): Likewise.
7954 (umulsidi3adddi_v6): Likewise.
7955 (smulsi3_highpart_v6): Likewise.
7956 (umulsi3_highpart_v6): Likewise.
7957 (mulhisi3tb): Likewise.
7958 (mulhisi3bt): Likewise.
7959 (mulhisi3tt): Likewise.
7960 (maddhisi4): Likewise.
7961 (maddhisi4tb): Likewise.
7962 (maddhisi4tt): Likewise.
7963 (maddhidi4): Likewise.
7964 (maddhidi4tb): Likewise.
7965 (maddhidi4tt): Likewise.
7966 (zeroextractsi_compare0_scratch): Likewise.
7967 (insv_zero): Likewise.
7968 (insv_t2): Likewise.
7969 (anddi_notzesidi_di): Likewise.
7970 (anddi_notsesidi_di): Likewise.
7971 (andsi_notsi_si): Likewise.
7972 (iordi_zesidi_di): Likewise.
7973 (xordi_zesidi_di): Likewise.
7974 (andsi_iorsi3_notsi): Likewise.
7976 (smax_m1): Likewise.
7978 (not_shiftsi): Likewise.
7979 (unaligned_loadsi): Likewise.
7980 (unaligned_loadhis): Likewise.
7981 (unaligned_loadhiu): Likewise.
7982 (unaligned_storesi): Likewise.
7983 (unaligned_storehi): Likewise.
7984 (extv_reg): Likewise.
7985 (extzv_t2): Likewise.
7987 (udivsi3): Likewise.
7988 (arm_zero_extendhisi2addsi): Likewise.
7989 (arm_zero_extendqisi2addsi): Likewise.
7990 (compareqi_eq0): Likewise.
7991 (arm_extendhisi2_v6): Likewise.
7992 (arm_extendqisi2addsi): Likewise.
7993 (arm_movt): Likewise.
7994 (thumb2_ldrd): Likewise.
7995 (thumb2_ldrd_base): Likewise.
7996 (thumb2_ldrd_base_neg): Likewise.
7997 (thumb2_strd): Likewise.
7998 (thumb2_strd_base): Likewise.
7999 (thumb2_strd_base_neg): Likewise.
8000 (arm_negsi2): Add alternative for 16-bit encoding.
8001 (arm_one_cmplsi2): Likewise.
8003 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8005 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
8006 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
8007 (movdfcc): Likewise.
8008 * config/arm/vfp.md (*thumb2_movsf_vfp):
8009 Disable predication for arm_restrict_it.
8010 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
8011 (*thumb2_movdfcc_vfp): Likewise.
8012 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
8013 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
8014 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
8015 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
8016 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
8017 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
8018 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
8019 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
8020 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
8021 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
8022 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
8023 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
8024 Disable predication for arm_restrict_it.
8026 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
8028 * config/i386/bmiintrin.h (_bextr_u32): New.
8029 (_bextr_u64): Ditto.
8031 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
8033 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
8034 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
8035 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
8036 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
8037 * config/mips/n32-elf.h: ...this new file.
8039 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
8042 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
8043 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
8045 2013-06-27 Catherine Moore <clm@codesourcery.com>
8047 * config/mips/mips-tables.opt: Regenerate.
8048 * config/mips/mips-cpus.def: Add m14ke and m14kec.
8049 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8050 * doc/invoke.texi: Add -m14kc.
8052 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8055 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
8056 constraints of operand 1 and 2.
8059 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
8060 to match RTL canonicalization. Swap predicates and
8061 constraints of operand 1 and 2.
8063 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
8065 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
8066 Process OP_INOUT regs for splitting too.
8068 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8070 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
8071 decl before the loop, initialize to NULL.
8072 (vectorizable_load): Initialize ptr_incr to NULL.
8074 2013-06-27 Martin Jambor <mjambor@suse.cz>
8077 * ipa-ref.h (ipa_maybe_record_reference): Declare.
8078 * ipa-ref.c (ipa_maybe_record_reference): New function.
8079 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
8080 * ipa-cp.c (create_specialized_node): Record potential references from
8082 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
8084 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8086 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
8087 parameter 'mode' of type 'enum machine_mode mode'; change to pass
8088 'mode' to force_reg.
8089 (aarch64_add_offset): Update calls to aarch64_force_temporary.
8090 (aarch64_expand_mov_immediate): Likewise.
8092 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8094 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
8095 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
8097 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8099 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
8100 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8101 (struct machine_function): Add tbegin_p.
8102 (s390_canonicalize_comparison): Fold CC mode compares to
8103 conditional jump if possible.
8104 (s390_emit_jump): Return the emitted jump.
8105 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
8106 Handle CCRAWmode compares.
8107 (s390_option_override): Default to -mhtm if available.
8108 (s390_reg_clobbered_rtx): Handle floating point regs as well.
8109 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
8110 FPRs instead of df_regs_ever_live_p.
8111 (s390_optimize_nonescaping_tx): New function.
8112 (s390_init_frame_layout): Extend clobbered_regs array to cover
8114 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
8115 (s390_expand_tbegin): New function.
8116 (enum s390_builtin): New enum definition.
8117 (code_for_builtin): New array definition.
8118 (s390_init_builtins): New function.
8119 (s390_expand_builtin): New function.
8120 (TARGET_INIT_BUILTINS): Define.
8121 (TARGET_EXPAND_BUILTIN): Define.
8122 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
8123 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
8124 (s390_alc_comparison): Likewise.
8125 * config/s390/s390-modes.def: Add CCRAWmode.
8126 * config/s390/s390.h (processor_flags): Add PF_TX.
8127 (TARGET_CPU_HTM): Define macro.
8128 (TARGET_HTM): Define macro.
8129 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
8130 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
8131 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8132 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
8133 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
8134 (TBEGIN_MASK, TBEGINC_MASK): New constants.
8135 ("*cc_to_int"): Move up.
8136 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
8137 constants other than 0.
8138 ("*ccraw_to_int"): New insn and splitter definition.
8139 ("tbegin", "tbegin_nofloat", "tbegin_retry")
8140 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
8141 ("tx_assist"): New expander.
8142 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
8143 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
8144 * config/s390/s390.opt: Add -mhtm option.
8145 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
8146 * config/s390/htmxlintrin.h: New file.
8147 * config/s390/htmintrin.h: New file.
8148 * config/s390/s390intrin.h: New file.
8149 * doc/extend.texi: Document htm builtins.
8150 * config.gcc: Add the new header files to extra_headers.
8152 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
8154 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
8155 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
8157 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
8158 Pat Haugen <pthaugen@us.ibm.com>
8159 Peter Bergner <bergner@vnet.ibm.com>
8161 * config/rs6000/power8.md: New.
8162 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
8163 setting for power8 entry.
8164 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
8165 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
8166 test for Power4/Power5 only.
8167 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
8169 (force_new_group): Adjust comment.
8170 * config/rs6000/rs6000.md: Include power8.md.
8172 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
8174 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
8175 * config/arm/arm-protos.h (arm_max_conditional_execute): New
8177 (tune_params): Update comment.
8178 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
8179 (arm_max_conditional_execute): New function.
8180 (thumb2_final_prescan_insn): Use max_insn_skipped and
8181 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
8183 2013-06-25 Jakub Jelinek <jakub@redhat.com>
8185 PR tree-optimization/57705
8186 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
8187 SSA_NAME step, provided that it is not defined inside the loop.
8188 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
8189 (get_initial_def_for_induction): Handle SSA_NAME IV step.
8191 2013-06-25 Martin Jambor <mjambor@suse.cz>
8194 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
8195 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
8197 (ipa_note_param_call): Initialize member_ptr flag.
8198 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
8199 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
8200 (ipa_write_indirect_edge_info): Stream member_ptr flag.
8201 (ipa_read_indirect_edge_info): Likewise.
8203 2013-06-25 Richard Biener <rguenther@suse.de>
8206 * passes.c (init_optimization_passes): Move pass_fold_builtins
8207 and pass_dce earlier with -Og.
8209 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
8211 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
8212 <BIT_FIELD_REF>: Remove trailing TAB.
8213 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
8216 2013-06-24 Martin Jambor <mjambor@suse.cz>
8218 PR tree-optimization/57358
8219 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
8220 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
8221 (ipa_analyze_params_uses): Generate pessimistic info when true.
8223 2013-06-24 Martin Jambor <mjambor@suse.cz>
8225 PR tree-optimization/57539
8226 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
8227 global.inlined_to of the new node to it. All callers changed.
8228 * ipa-inline-transform.c (clone_inlined_nodes): New variable
8229 inlining_into, pass it to cgraph_clone_node.
8230 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
8231 ipa_free_edge_args_substructures.
8232 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
8233 rdesc linked list. Do not assert rdesc edges have inlined caller.
8234 Assert we have found an rdesc in the rdesc list.
8236 2013-06-24 Richard Biener <rguenther@suse.de>
8238 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
8239 (pointer_set_lookup): Declare.
8240 (class pointer_map): New template class implementing a
8241 generic pointer to T map.
8242 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
8243 pointer_map<T>::contains, pointer_map<T>::insert,
8244 pointer_map<T>::traverse): New functions.
8245 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
8246 (pointer_set_lookup): New function.
8247 (pointer_set_contains): Use pointer_set_lookup.
8248 (pointer_set_insert): Likewise.
8249 (insert_aux): Remove.
8250 (struct pointer_map_t): Embed a pointer_set_t.
8251 (pointer_map_create): Adjust.
8252 (pointer_map_destroy): Likewise.
8253 (pointer_map_contains): Likewise.
8254 (pointer_map_insert): Likewise.
8255 (pointer_map_traverse): Likewise.
8256 * tree-streamer.h (struct streamer_tree_cache_d): Use a
8257 pointer_map<unsigned> instead of a pointer_map_t.
8258 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
8259 (streamer_tree_cache_lookup): Likewise.
8260 (streamer_tree_cache_create): Likewise.
8261 (streamer_tree_cache_delete): Likewise.
8262 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
8263 pointer_map<unsigned> instead of a pointer_map_t.
8264 (lto_init_tree_ref_encoder): Adjust.
8265 (lto_destroy_tree_ref_encoder): Likewise.
8266 * lto-section-out.c (lto_output_decl_index): Likewise.
8267 (lto_record_function_out_decl_state): Likewise.
8268 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
8270 2013-06-24 Richard Biener <rguenther@suse.de>
8272 PR tree-optimization/57488
8273 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
8275 2013-06-24 Alan Modra <amodra@gmail.com>
8277 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
8278 (gen_easy_altivec_constant): Likewise.
8279 * config/rs6000/predicates.md (easy_vector_constant_add_self,
8280 easy_vector_constant_msb): Likewise.
8282 2013-06-23 Jakub Jelinek <jakub@redhat.com>
8285 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
8286 add missing return true.
8288 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
8291 * config/sh/predicates.md (general_extend_operand): Invoke
8292 general_movsrc_operand for memory operands.
8293 (general_movsrc_operand): Allow reg+reg addressing, do not use
8294 general_operand for memory operands.
8296 2013-06-23 Sriraman Tallam <tmsriram@google.com>
8298 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
8299 when current target options does not apply.
8300 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
8301 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
8302 * config/i386/bmiintrin.h: Pass appropriate target
8303 attributes to header.
8304 * config/i386/mmintrin.h: Ditto.
8305 * config/i386/nmmintrin.h: Ditto.
8306 * config/i386/avx2intrin.h: Ditto.
8307 * config/i386/fxsrintrin.h: Ditto.
8308 * config/i386/tbmintrin.h: Ditto.
8309 * config/i386/xsaveintrin.h: Ditto.
8310 * config/i386/f16cintrin.h: Ditto.
8311 * config/i386/xtestintrin.h: Ditto.
8312 * config/i386/xsaveoptintrin.h: Ditto.
8313 * config/i386/bmi2intrin.h: Ditto.
8314 * config/i386/lzcntintrin.h: Ditto.
8315 * config/i386/smmintrin.h: Ditto.
8316 * config/i386/wmmintrin.h: Ditto.
8317 * config/i386/x86intrin.h: Remove all header include guards.
8318 * config/i386/prfchwintrin.h: Ditto.
8319 * config/i386/pmmintrin.h: Ditto.
8320 * config/i386/tmmintrin.h: Ditto.
8321 * config/i386/xmmintrin.h: Ditto.
8322 * config/i386/popcntintrin.h: Ditto.
8323 * config/i386/rdseedintrin.h: Ditto.
8324 * config/i386/ammintrin.h: Ditto.
8325 * config/i386/emmintrin.h: Ditto.
8326 * config/i386/immintrin.h: Remove all header include guards.
8327 * config/i386/fma4intrin.h: Ditto.
8328 * config/i386/lwpintrin.h: Ditto.
8329 * config/i386/xopintrin.h: Ditto.
8330 * config/i386/ia32intrin.h: Ditto.
8331 * config/i386/avxintrin.h: Ditto.
8332 * config/i386/rtmintrin.h: Ditto.
8333 * config/i386/fmaintrin.h: Ditto.
8334 * config/i386/mm3dnow.h: Ditto.
8336 2013-06-22 Sriraman Tallam <tmsriram@google.com>
8338 * common/config/i386/i386-common.c: Handle LZCNT.
8340 2013-06-22 Andi Kleen <ak@linux.intel.com>
8342 * doc/extend.texi: Use __atomic_store_n instead of
8343 __atomic_store in HLE example.
8345 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
8347 * config/sh/sh.c: Remove <cstdlib> workaround.
8349 2013-06-21 Andi Kleen <ak@linux.intel.com>
8351 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
8353 2013-06-21 Andi Kleen <ak@linux.intel.com>
8355 * doc/extend.texi: Document that __atomic_clear and
8356 __atomic_test_and_set should only be used with bool.
8358 2013-06-20 Jan Hubicka <jh@suse.cz>
8360 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
8362 * tree.c (decls_same_for_odr): New function.
8363 (same_for_edr): New function.
8364 (types_same_for_odr): New function.
8365 (get_binfo_at_offset): Use it.
8366 * tree.h (types_same_for_odr): Declare.
8368 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
8369 Jason Merrill <jason@redhat.com>
8371 * system.h: Include <cstdlib> as well as <stdlib.h>.
8373 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
8376 * config/i386/i386.c (construct_container): Report error if
8377 long double is used with disabled x87 float returns.
8379 2013-06-20 Jan Hubicka <jh@suse.cz>
8381 * lto-cgraph.c (input_symtab): Do not set cgraph state.
8383 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
8385 PR rtl-optimization/57425
8386 PR rtl-optimization/57569
8387 * alias.c (write_dependence_p): Remove parameters mem_mode and
8388 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
8389 Changed all callers.
8390 (canon_anti_dependence): Get comments and semantics in sync.
8391 Add parameter mem_canonicalized. Changed all callers.
8392 * rtl.h (canon_anti_dependence): Update prototype.
8394 2013-06-20 Richard Biener <rguenther@suse.de>
8396 * data-streamer-in.c (streamer_read_uhwi): Optimize single
8397 byte case, inline streamer_read_uchar and defer section
8400 2013-06-20 Richard Biener <rguenther@suse.de>
8402 PR tree-optimization/57584
8403 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
8404 SSA names into the expanded expression that take part in
8405 abnormal coalescing.
8407 2013-06-19 Sharad Singhai <singhai@google.com>
8409 * gcov.c (print_usage): Handle new option.
8410 (process_args): Ditto.
8411 (get_gcov_intermediate_filename): New function.
8412 (output_intermediate_file): New function.
8413 (output_gcov_file): New function
8414 (generate_results): Handle new option.
8415 (release_function): Relase demangled name.
8416 (read_graph_file): Handle demangled name.
8417 (output_lines): Ditto.
8418 * doc/gcov.texi: Document gcov intermediate format.
8420 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
8423 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
8424 (lra_emit_add): Use the functions. Add comment about Y as an
8427 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
8430 * collect2.c (collect_atexit): New.
8431 (collect_exit): Delete.
8432 (main): Register collect_atexit with atexit.
8433 (collect_wait): Change collect_exit to exit.
8435 * collect2.h (collect_exit): Delete.
8436 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
8438 2013-06-19 Wei Mi <wmi@google.com>
8440 PR rtl-optimization/57518
8441 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8442 if regno is used in paradoxical subreg.
8443 (update_equiv_regs): Check pdx_subregs[regno] before
8444 set a reg to be equivalent with a mem.
8446 2013-06-19 Matthias Klose <doko@ubuntu.com>
8449 * file-find.h (find_a_file): Add a mode parameter.
8450 * file-find.c (find_a_file): Likewise.
8451 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8452 with X_OK for the executables.
8453 * collect2.c (main): Call find_a_file with X_OK.
8455 2013-06-19 Steve Ellcey <sellcey@mips.com>
8458 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8459 Use NEXT_INSN instead of next_real_insn.
8461 2013-06-19 Jan Hubicka <jh@suse.cz>
8463 * cgraph.h (const_value_known_p): Replace by ...
8464 (ctor_for_folding): .. this one.
8465 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8466 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8467 * expr.c (expand_expr_real_1): Likewise.
8468 (string_constant): Likewise.
8469 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8470 * ipa.c (process_references): Likewise.
8471 (symtab_remove_unreachable_nodes): Likewise.
8472 * ipa-inline-analysis.c (param_change_prob): Likewise.
8473 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8474 (get_base_constructor): Likwise.
8475 * varpool.c (varpool_remove_node): Likewise.
8476 (varpool_remove_initializer): LIkewise.
8477 (dump_varpool_node): LIkwise.
8478 (const_value_known_p): Rewrite to ...
8479 (ctor_for_folding): ... this one.
8481 2013-06-19 Jakub Jelinek <jakub@redhat.com>
8484 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8485 PERSONALITY in $PATH derived prefixes.
8487 2013-06-19 Jeff Law <law@redhat.com>
8489 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8492 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8493 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8496 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8498 * config/arm/vfp.md: Move VFP instruction classification documentation
8500 * config/arm/arm.md: ... here. Update instruction classification
8503 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
8505 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8506 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8507 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8509 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8512 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8513 with NEXT_INSN. Use tablejump_p to check for jump table data
8516 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8519 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8520 that now in C++ the value is correct per the C++ standards.
8522 2013-06-19 Richard Biener <rguenther@suse.de>
8524 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8527 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8530 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8533 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8534 with next_active_insn.
8536 2013-06-18 Sriraman Tallam <tmsriram@google.com>
8538 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
8539 functions are inlined during failures to flag an error.
8540 * tree-inline.c (expand_call_inline): Allow the error to be flagged
8541 in early inline pass.
8543 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
8545 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
8548 2013-06-18 Julian Brown <julian@codesourcery.com>
8550 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
8551 Permit virtual register pre-reload if !strict.
8552 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
8554 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
8556 * config/arm/neon.md (movmisalign<mode>): Use
8557 neon_perm_struct_or_reg_operand instead of
8558 neon_struct_or_register_operand.
8559 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
8560 neon_permissive_struct_operand instead of neon_struct_operand.
8561 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
8562 neon_vector_mem_operand.
8563 * config/arm/predicates.md (neon_struct_operand): Adjust call to
8564 neon_vector_mem_operand.
8565 (neon_permissive_struct_operand): New.
8566 (neon_struct_or_register_operand): Rename to...
8567 (neon_perm_struct_or_reg_operand): This. Adjust call to
8568 neon_vector_mem_operand.
8570 2013-06-18 Richard Biener <rguenther@suse.de>
8572 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
8573 * lto-streamer.h: Include pointer-set.h.
8574 (struct lto_decl_slot): Remove.
8575 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
8576 Remove next_index entry.
8577 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8578 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8579 (lto_init_tree_ref_encoder): Adjust.
8580 (lto_destroy_tree_ref_encoder): Likewise.
8581 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8582 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8583 (lto_output_decl_index): Adjust.
8584 (lto_new_out_decl_state): Likewise.
8585 (lto_record_function_out_decl_state): Likewise.
8586 * lto-streamer-out.c (copy_function): Likewise.
8588 2013-06-18 Richard Biener <rguenther@suse.de>
8590 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
8591 * cgraphunit.c: Include cfgloop.h.
8592 (init_lowered_empty_function): Initialize the loop tree.
8593 (assemble_thunk): Insert new BBs into loops.
8595 2013-06-18 Richard Biener <rguenther@suse.de>
8597 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8598 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8599 the map from cache entry to cache index optional.
8600 (streamer_tree_cache_replace_tree): Adjust accordingly.
8601 (streamer_tree_cache_append): Likewise.
8602 (streamer_tree_cache_delete): Likewise.
8603 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8604 streamer cache map from cache entry to cache index.
8605 * lto-streamer-out.c (create_output_block): Adjust.
8607 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8609 * config/arm/arm.md (attribute "insn"): Move multiplication and
8610 division attributes to...
8611 (attribute "type"): ... here. Remove mult.
8612 (attribute "mul32"): New attribute.
8613 (attribute "mul64"): Add umaal.
8614 (*arm_mulsi3): Update attributes.
8615 (*arm_mulsi3_v6): Likewise.
8616 (*thumb_mulsi3): Likewise.
8617 (*thumb_mulsi3_v6): Likewise.
8618 (*mulsi3_compare0): Likewise.
8619 (*mulsi3_compare0_v6): Likewise.
8620 (*mulsi_compare0_scratch): Likewise.
8621 (*mulsi_compare0_scratch_v6): Likewise.
8622 (*mulsi3addsi): Likewise.
8623 (*mulsi3addsi_v6): Likewise.
8624 (*mulsi3addsi_compare0): Likewise.
8625 (*mulsi3addsi_compare0_v6): Likewise.
8626 (*mulsi3addsi_compare0_scratch): Likewise.
8627 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8628 (*mulsi3subsi): Likewise.
8629 (*mulsidi3adddi): Likewise.
8630 (*mulsi3addsi_v6): Likewise.
8631 (*mulsidi3adddi_v6): Likewise.
8632 (*mulsidi3_nov6): Likewise.
8633 (*mulsidi3_v6): Likewise.
8634 (*umulsidi3_nov6): Likewise.
8635 (*umulsidi3_v6): Likewise.
8636 (*umulsidi3adddi): Likewise.
8637 (*umulsidi3adddi_v6): Likewise.
8638 (*smulsi3_highpart_nov6): Likewise.
8639 (*smulsi3_highpart_v6): Likewise.
8640 (*umulsi3_highpart_nov6): Likewise.
8641 (*umulsi3_highpart_v6): Likewise.
8642 (mulhisi3): Likewise.
8643 (*mulhisi3tb): Likewise.
8644 (*mulhisi3bt): Likewise.
8645 (*mulhisi3tt): Likewise.
8646 (maddhisi4): Likewise.
8647 (*maddhisi4tb): Likewise.
8648 (*maddhisi4tt): Likewise.
8649 (maddhidi4): Likewise.
8650 (*maddhidi4tb): Likewise.
8651 (*maddhidi4tt): Likewise.
8653 (udivsi3): Likewise.
8654 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8655 (thumb2_mulsi_short_compare0): Likewise.
8656 (thumb2_mulsi_short_compare0_scratch): Likewise.
8657 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8658 (1020mult2): Likewise.
8659 (1020mult3): Likewise.
8660 (1020mult4): Likewise.
8661 (1020mult5): Likewise.
8662 (1020mult6): Likewise.
8663 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8665 (cortex_a15_mult64): Likewise.
8666 (cortex_a15_sdiv): Likewise.
8667 (cortex_a15_udiv): Likewise.
8668 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8674 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8675 (pj4_ir_div): Likewise.
8676 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8677 (11_mult2): Likewise.
8678 (11_mult3): Likewise.
8679 (11_mult4): Likewise.
8680 (11_mult5): Likewise.
8681 (11_mult6): Likewise.
8682 (11_mult7): Likewise.
8683 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8684 (cortex_a8_mla): Likewise.
8685 (cortex_a8_mull): Likewise.
8686 (cortex_a8_smulwy): Likewise.
8687 (cortex_a8_smlald): Likewise.
8688 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8689 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8690 (cortex_r4_mul_3): Likewise.
8691 (cortex_r4_mla_4): Likewise.
8692 (cortex_r4_mla_3): Likewise.
8693 (cortex_r4_smlald): Likewise.
8694 (cortex_r4_mull): Likewise.
8695 (cortex_r4_sdiv): Likewise.
8696 (cortex_r4_udiv): Likewise.
8697 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8698 (cortex_a7_idiv): Likewise.
8699 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8700 (9_mult2): Likewise.
8701 (9_mult3): Likewise.
8702 (9_mult4): Likewise.
8703 (9_mult5): Likewise.
8704 (9_mult6): Likewise.
8705 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8706 (cortex_a53_sdiv): Likewise.
8707 (cortex_a53_udiv): Likewise.
8708 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8709 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8710 (mp626_mult2): Likewise.
8711 (mp626_mult3): Likewise.
8712 (mp626_mult4): Likewise.
8713 * config/arm/fa526.md (526_mult1): Update attribute change.
8714 (526_mult2): Likewise.
8715 * config/arm/arm-generic.md (mult): Update attribute change.
8716 (mult_ldsched_strongarm): Likewise.
8717 (mult_ldsched): Likewise.
8718 (multi_cycle): Likewise.
8719 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8720 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8721 (606te_mult2): Likewise.
8722 (606te_mult3): Likewise.
8723 (606te_mult4): Likewise.
8724 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8725 (cortex_a9_mac16): Likewise.
8726 (cortex_a9_multiply): Likewise.
8727 (cortex_a9_mac): Likewise.
8728 (cortex_a9_multiply_long): Likewise.
8729 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8730 (626te_mult2): Likewise.
8731 (626te_mult3): Likewise.
8732 (626te_mult4): Likewise.
8734 2013-06-18 Richard Biener <rguenther@suse.de>
8737 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8739 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8742 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8743 with next_active_insn.
8745 2013-06-18 Alan Modra <amodra@gmail.com>
8747 * config/rs6000/rs6000.h (enum data_align): New.
8748 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8749 (DATA_ABI_ALIGNMENT): Define.
8750 (CONSTANT_ALIGNMENT): Correct comment.
8751 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8752 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8754 2013-06-17 David Malcolm <dmalcolm@redhat.com>
8756 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8757 ATTRIBUTE_UNUSED marking.
8759 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8761 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8762 alternative and update.
8763 (aarch64_dup_lanedi): Delete.
8764 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8765 * config/aarch64/aarch64-simd-builtins.def: Update.
8767 2013-06-17 Richard Biener <rguenther@suse.de>
8769 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8770 (lto_input_scc): Declare.
8771 (lto_input_tree_1): Likewise.
8772 (struct lto_stats_d): Add num_tree_bodies_output and
8773 num_pickle_refs_output.
8774 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8775 (lto_read_tree_1): Split out from ...
8776 (lto_read_tree): ... this.
8777 (lto_input_scc): New function.
8778 (lto_input_tree_1): Split out from ...
8779 (lto_input_tree): ... this. Handle LTO_tree_scc.
8780 (lto_data_in_create): Create the streamer cache without hashes.
8781 * lto-streamer-out.c (create_output_block): Create the streamer
8782 cache with hashes when not doing WPA.
8783 (lto_write_tree_1): Split out from ...
8784 (lto_write_tree): ... this.
8785 (get_symbol_initial_value): New function.
8786 (lto_output_tree_1): Split out from ...
8787 (lto_output_tree): ... this. Write trees as series of SCCs
8788 using a DFS walk via DFS_write_tree.
8789 (struct sccs, struct scc_entry): New types.
8790 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
8791 (DFS_write_tree_body): New function.
8792 (DFS_write_tree): Likewise.
8793 (hash_tree): Likewise.
8794 (scc_entry_compare): Likewise.
8795 (hash_scc): Likewise.
8796 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
8797 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
8798 TREE_CHAIN as regular reference.
8799 (streamer_read_integer_cst): Remove.
8800 (streamer_get_pickled_tree): Adjust.
8801 * tree-streamer-out.c (streamer_write_chain): Disable streaming
8802 of DECL_EXTERNALs in BLOCK_VARS for now.
8803 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
8805 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
8806 Add hash value argument and record that if hashes are recorded
8808 (streamer_tree_cache_insert_1): Adjust.
8809 (streamer_tree_cache_insert): Likewise.
8810 (streamer_tree_cache_insert_at): Rename to ...
8811 (streamer_tree_cache_replace_tree): ... this and adjust.
8812 (streamer_tree_cache_append): Adjust.
8813 (record_common_node): Likewise.
8814 (streamer_tree_cache_create): Add argument whether to
8815 record hash values together with trees.
8816 (streamer_tree_cache_delete): Adjust.
8817 * tree-streamer.h (struct streamer_tree_cache_d): Add
8819 (streamer_read_integer_cst): Remove.
8820 (streamer_tree_cache_insert): Adjust.
8821 (streamer_tree_cache_append): Likewise.
8822 (streamer_tree_cache_insert_at): Rename to ...
8823 (streamer_tree_cache_replace_tree): ... this and adjust.
8824 (streamer_tree_cache_create): Add argument whether to record hashes.
8825 (streamer_tree_cache_get): Rename to ...
8826 (streamer_tree_cache_get_tree): ... this.
8827 (streamer_tree_cache_get_hash): New function.
8828 * tree.c (cache_integer_cst): New function.
8829 * tree.h (cache_integer_cst): Declare.
8830 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
8831 * lto-symtab.c (lto_varpool_replace_node): Only release
8832 DECL_INITIAL of non-prevailing decls.
8833 * varpool.c (varpool_remove_initializer): Do not release
8834 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
8836 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8838 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
8839 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
8840 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
8841 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
8842 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
8843 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
8844 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
8845 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
8846 instead of TARGET_64BIT.
8847 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
8848 Require ISA_HAS_<D>DIV.
8850 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
8852 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
8853 (mips*-*-linux*): Move default with_llsc setting to where other
8855 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
8857 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
8858 Likewise. Remove default with_tune setting. Move default float
8859 setting to its own block. Handle with_llsc in the same block as above.
8861 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
8863 PR rtl-optimization/57425
8864 PR rtl-optimization/57569
8865 * alias.c (write_dependence_p): Add new parameters mem_mode,
8866 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
8867 Changed all callers.
8868 (canon_anti_dependence): New function.
8869 * cse.c (check_dependence): Use canon_anti_dependence.
8870 * cselib.c (cselib_invalidate_mem): Likewise.
8871 * rtl.h (canon_anti_dependence): Declare.
8873 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8875 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
8876 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
8877 ".set mips3" for 64-bit targets.
8879 2013-06-15 Dehao Chen <dehao@google.com>
8881 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
8882 * gimple-low.c (gimple_check_call_matching_types): Likewise.
8883 (gimple_check_call_args): Likewise.
8884 * value-prof.c (check_ic_target): Likewise.
8885 * ipa-inline.c (early_inliner): Likewise.
8886 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
8887 * cgraph.c (cgraph_create_edge_1): Likewise.
8888 (cgraph_make_edge_direct): Likewise.
8890 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
8893 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
8894 rs6000_output_move_128bit to handle emitting quad memory
8895 operations. Set attribute length to 8 bytes.
8897 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
8899 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
8901 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
8902 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
8903 (aarch64_<su>mlsl<mode>): Likewise.
8905 2013-06-14 Mike Stump <mikestump@comcast.net>
8907 * Makefile.in (TARGET_H): Add insn-codes.h.
8909 2013-06-14 Alan Modra <amodra@gmail.com>
8913 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
8914 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
8915 bitfield expansion when EXPAND_MEMORY.
8916 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
8918 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
8920 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
8921 test for clearing quad memory on 32-bit later.
8923 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8925 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
8926 (fold_negate_expr): Likewise.
8927 (fold_real_zero_addition_p): Handle vectors.
8928 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
8930 2013-06-14 Alan Modra <amodra@gmail.com>
8932 * varasm.c (force_const_mem): Revert 2013-06-07 change.
8934 2013-06-13 Jan Hubicka <jh@suse.cz>
8936 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
8937 Local comdats are not externally visible.
8938 * symtab.c (dump_symtab_base): Dump externally visible.
8939 (verify_symtab_base): Verify back links in the symtab hash.
8941 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8943 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
8944 CONVERT_EXPR as equal nodes.
8946 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8948 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8950 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8952 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
8953 Generalize to complex and vector.
8954 * tree.c (build_all_ones_cst): New function.
8955 * tree.h (build_all_ones_cst): Declare it.
8957 2013-06-13 Alan Modra <amodra@gmail.com>
8959 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
8960 * config/rs6000/rs6000.md (signbittf2): New insn.
8961 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
8962 (abstf2_internal, cmptf_internal2): Likewise.
8963 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
8965 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8966 Pat Haugen <pthaugen@us.ibm.com>
8967 Peter Bergner <bergner@vnet.ibm.com>
8969 * config/rs6000/rs6000.c (emit_load_locked): Add support for
8970 power8 byte, half-word, and quad-word atomic instructions.
8971 (emit_store_conditional): Likewise.
8972 (rs6000_expand_atomic_compare_and_swap): Likewise.
8973 (rs6000_expand_atomic_op): Likewise.
8975 * config/rs6000/sync.md (larx): Add new modes for power8.
8977 (AINT): New mode iterator to include TImode as well as normal
8978 integer modes on power8.
8979 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
8980 that VSX registers are not considered. Use AINT mode iterator
8981 instead of INT1 to allow inclusion of quad word atomic operations
8983 (load_locked<mode>): Likewise.
8984 (store_conditional<mode>): Likewise.
8985 (atomic_compare_and_swap<mode>): Likewise.
8986 (atomic_exchange<mode>): Likewise.
8987 (atomic_nand<mode>): Likewise.
8988 (atomic_fetch_<fetchop_name><mode>): Likewise.
8989 (atomic_nand_fetch<mode>): Likewise.
8990 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
8992 (ATOMIC): On power8, add QImode, HImode modes.
8993 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
8994 modes that promote to SImode.
8995 (load_lockedti): Convert TImode arguments to PTImode, so that we
8996 get a guaranteed even/odd register pair.
8997 (load_lockedpti): Likewise.
8998 (store_conditionalti): Likewise.
8999 (store_conditionalpti): Likewise.
9001 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
9002 atomic load/store instructions.
9005 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
9007 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
9009 (insn_count): New attribute, with most cases extracted from...
9010 (length): ...here. Redefine most cases in terms of insn_count.
9011 (single_insn): Delete.
9012 (can_delay): Use insn_count to check for single instructions.
9013 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
9014 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
9015 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
9016 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
9017 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
9018 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
9019 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
9020 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
9021 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
9022 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
9023 rather than "length".
9024 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
9025 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
9026 Use "insn_count" rather than "length".
9027 * config/mips/mips-dsp.md
9028 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
9029 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
9032 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
9034 PR tree-optimization/57361
9035 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
9037 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
9039 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
9041 (aarch64_simd_combine<mode>): New instruction expansion.
9042 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
9044 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
9045 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
9047 2013-06-12 Jan Hubicka <jh@suse.cz>
9049 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
9050 decl has when in streaming stage.
9051 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
9052 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
9054 2013-06-12 Roland Stigge <stigge@antcom.de>
9057 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
9059 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9061 PR tree-optimization/57537
9062 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
9063 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
9065 2013-06-12 Richard Biener <rguenther@suse.de>
9067 * data-streamer.h (streamer_write_char_stream): CSE
9068 obs->current_pointer.
9069 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
9070 streamer_write_char_stream manually and optimize the resulting loop.
9071 (streamer_write_hwi_stream): Likewise.
9073 2013-06-12 Jan Hubicka <jh@suse.cz>
9075 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
9076 * cgraph.h (varpool_create_empty_node): Declare.
9077 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
9079 * symtab.c (symtab_unregister_node): Be lax about missin entries
9081 (symtab_get_node): Update comment.
9082 * varpool.c (varpool_create_empty_node): Break out from ...
9083 (varpool_node_for_decl): ... here.
9084 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
9086 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
9088 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
9089 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
9090 part. Use straight-line flow at the end.
9091 <COMPONENT_REF>: Remove superfluous else.
9092 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
9094 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9097 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
9098 target hook even for !TREE_PUBLIC decls. If no resolution info
9099 is available, return false for common and external decls.
9101 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9103 * config/rl78/constraints.md (U): New constraint.
9104 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
9107 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9110 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
9111 to allow returning address to AT_PLATFORM name.
9113 2013-06-11 Jan Hubicka <jh@suse.cz>
9115 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
9116 * cgraph.h (symtab_node_base): Add weakref flag.
9117 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
9118 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
9119 (output_weakrefs): Use weakref flag.
9120 * fold-const.c (simple_operand_p): Handle WEAK.
9121 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
9122 * ipa.c (varpool_externally_visible_p): Drop weakref.
9123 (function_and_variable_visibility): Update comment; fix weakref
9124 sanity checks; do not clear DECL_WEAK on them.
9125 * lto-cgraph.c (lto_output_node): update.
9126 (lto_output_varpool_node): Update.
9127 (input_overwrite_node): Update.
9128 (input_node): Update.
9129 (input_varpool_node): Update.
9130 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
9131 (lto_symtab_merge_symbols): Add sanity check.
9132 (lto_symtab_prevailing_decl): Do not special case weakrefs.
9133 * passes.c (rest_of_decl_compilation): Set static flag, too.
9134 * symtab.c (dump_symtab_base): Dump weakref.
9135 (verify_symtab_base): Sanity check weakrefs.
9136 (symtab_make_decl_local): Remove duplicated code.
9137 (symtab_alias_ultimate_target): Simplify.
9138 * varpool.c (varpool_create_variable_alias): Set weakref flag.
9140 2013-06-11 Tom de Vries <tom@codesourcery.com>
9142 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
9143 sequence_vect == NULL.
9145 2013-06-11 DJ Delorie <dj@redhat.com>
9147 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
9148 (rl78_unwind_word_mode): New.
9150 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9152 * final.c (debug_prefix_maps): Make static.
9154 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9156 * function.c (initial_trampoline): Remove stray copy.
9158 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
9160 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
9162 2013-06-11 Martin Jambor <mjambor@suse.cz>
9164 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
9165 within bounds at the beginning of the function.
9167 2013-06-11 Alan Modra <amodra@gmail.com>
9169 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
9171 (get_named_section): Don't NULL !DECL_P decl.
9173 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
9175 * doc/invoke.texi (core-avx2): Document.
9177 (atom): Updated with MOVBE.
9179 2013-06-11 Richard Biener <rguenther@suse.de>
9181 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
9183 2013-06-11 Anton Blanchard <anton@samba.org>
9185 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
9186 correct shift value in little-endian mode.
9188 2013-06-11 Jakub Jelinek <jakub@redhat.com>
9191 * varasm.c (get_variable_align): Move #endif to the right place.
9193 2013-06-10 Cary Coutant <ccoutant@google.com>
9195 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
9196 for hash so that hash table traversal order is deterministic.
9198 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
9199 Pat Haugen <pthaugen@us.ibm.com>
9200 Peter Bergner <bergner@vnet.ibm.com>
9202 * config/rs6000/vector.md (GPR move splitter): Do not split moves
9203 of vectors in GPRS if they are direct moves or quad word load or
9206 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
9208 (direct_move_p): Likewise.
9209 (quad_load_store_p): Likewise.
9211 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
9212 classes into bins based on the physical register type.
9213 (reg_class_to_reg_type): Likewise.
9214 (IS_STD_REG_TYPE): Likewise.
9215 (IS_FP_VECT_REG_TYPE): Likewise.
9216 (reload_fpr_gpr): Arrays to determine what insn to use if we can
9217 use direct move instructions.
9218 (reload_gpr_vsx): Likewise.
9219 (reload_vsx_gpr): Likewise.
9220 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
9221 information that is a simplification of register classes. Also
9222 precalculate direct move reload helpers.
9223 (direct_move_p): New function to return true if the operation can
9224 be done as a direct move instruciton.
9225 (quad_load_store_p): New function to return true if the operation
9226 is a quad memory operation.
9227 (rs6000_legitimize_address): If quad memory, only allow register
9228 indirect for TImode addresses.
9229 (rs6000_legitimate_address_p): Likewise.
9230 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
9231 (rs6000_reload_register_type): Likewise.
9232 (register_to_reg_type): Return register type.
9233 (rs6000_secondary_reload_simple_move): New helper function for
9234 secondary reload and secondary memory needed to identify anything
9235 that is a simple move, and does not need reloading.
9236 (rs6000_secondary_reload_direct_move): New helper function for
9237 secondary reload to identify cases that can be done with several
9238 instructions via the direct move instructions.
9239 (rs6000_secondary_reload_move): New helper function for secondary
9240 reload to identify moves between register types that can be done.
9241 (rs6000_secondary_reload): Add support for quad memory operations
9242 and for direct move.
9243 (rs6000_secondary_memory_needed): Likewise.
9244 (rs6000_debug_secondary_memory_needed): Change argument names.
9245 (rs6000_output_move_128bit): New function to return the move to
9246 use for 128-bit moves, including knowing about the various
9247 limitations of quad memory operations.
9249 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
9250 memory operations. call rs6000_output_move_128bit for the actual
9251 instruciton(s) to generate.
9252 (vsx_movti_64bit): Likewise.
9254 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
9255 (UNSPEC_P8V_MTVSRWZ): Likewise.
9256 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
9257 (UNSPEC_P8V_MTVSRD): Likewise.
9258 (UNSPEC_P8V_XXPERMDI): Likewise.
9259 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
9260 (UNSPEC_FUSION_GPR): Likewise.
9261 (FMOVE128_GPR): New iterator for direct move.
9262 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
9265 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
9266 loads and direct move instructions.
9267 (zero_extendsidi2_lfiwzx): Likewise.
9268 (extendsidi2_lfiwax): Likewise.
9269 (extendsidi2_nocell): Likewise.
9270 (floatsi<mode>2_lfiwax): Likewise.
9272 (floatunssi<mode>2_lfiwzx): Likewise.
9274 (fix_trunc<mode>_stfiwx): Likewise.
9275 (fixuns_trunc<mode>_stfiwx): Likewise.
9276 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
9277 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
9278 (parity<mode>2_cmpb): Set length/type attr.
9279 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
9280 for 'mr.' to fast_compare.
9281 (bpermd_<mode>): Change type attr to popcnt.
9282 (p8_fmrgow_<mode>): New insns for power8 direct move support.
9283 (p8_mtvsrwz_1): Likewise.
9284 (p8_mtvsrwz_2): Likewise.
9285 (reload_fpr_from_gpr<mode>): Likewise.
9286 (p8_mtvsrd_1): Likewise.
9287 (p8_mtvsrd_2): Likewise.
9288 (p8_xxpermdi_<mode>): Likewise.
9289 (reload_vsx_from_gpr<mode>): Likewise.
9290 (reload_vsx_from_gprsf): Likewise.
9291 (p8_mfvsrd_3_<mode>): LIkewise.
9292 (reload_gpr_from_vsx<mode>): Likewise.
9293 (reload_gpr_from_vsxsf): Likewise.
9294 (p8_mfvsrd_4_disf): Likewise.
9295 (multi-word GPR splits): Do not split direct moves or quad memory
9298 2013-06-10 David Malcolm <dmalcolm@redhat.com>
9300 * tree-into-ssa.c (interesting_blocks): Make static.
9302 2013-06-10 Jakub Jelinek <jakub@redhat.com>
9305 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
9306 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
9307 Use DATA_ABI_ALIGNMENT for that case instead if defined.
9308 (get_variable_align): New function.
9309 (get_variable_section, emit_bss, emit_common,
9310 assemble_variable_contents, place_block_symbol): Use
9311 get_variable_align instead of DECL_ALIGN.
9312 (assemble_noswitch_variable): Add align argument, use it
9313 instead of DECL_ALIGN.
9314 (assemble_variable): Adjust caller. Use get_variable_align
9315 instead of DECL_ALIGN.
9316 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
9318 (DATA_ABI_ALIGNMENT): Define.
9319 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
9320 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
9321 opt is false, only return the psABI mandated alignment increase.
9322 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
9323 (DATA_ABI_ALIGNMENT): ... this.
9324 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
9325 (DATA_ABI_ALIGNMENT): ... this.
9326 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
9327 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
9328 (DATA_ABI_ALIGNMENT): ... this.
9329 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
9330 * doc/tm.texi: Regenerated.
9332 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
9334 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
9335 cmp_code to construct REG_EQUAL note.
9337 2013-06-09 Jakub Jelinek <jakub@redhat.com>
9340 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
9341 that operands[2] doesn't overlap with operands[0].
9343 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
9344 Jan Hubicka <jh@suse.cz>
9346 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
9347 hack to mark symbols as used.
9349 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
9351 PR rtl-optimization/57559
9352 * lra-constraints.c (process_alt_operands): Don't discourage
9353 memory with known offset for offsetable memory constraint.
9354 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
9356 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
9358 * varasm.c (struct oc_local_state): Reorder fields.
9359 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
9360 and adjust accordingly.
9361 (output_constructor): Reorder initialization code and adjust call to
9362 output_constructor_bitfield.
9364 2013-06-07 Jan Hubicka <jh@suse.cz>
9366 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
9368 2013-06-07 David Malcolm <dmalcolm@redhat.com>
9370 * tree-object-size.c (unknown): Make const.
9372 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9374 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
9375 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
9376 for last alternative in the cpu_facility attribute.
9378 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9381 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
9382 (xordi3): Change operand 2 constraint to arm_xordi_operand.
9383 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
9384 * config/arm/constraints.md (Dg): New constraint.
9385 * config/arm/neon.md (xordi3_neon): Remove.
9386 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
9387 * config/arm/predicates.md (arm_xordi_operand): New predicate.
9389 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9391 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
9392 Clean up alternatives.
9394 2013-06-07 Alan Modra <amodra@gmail.com>
9396 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
9399 2013-06-07 Alan Modra <amodra@gmail.com>
9401 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
9403 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9405 * config/arm/constraints.md (Df): New constraint.
9406 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
9407 Correct length attribute for last two alternatives.
9409 2013-06-07 Alan Modra <amodra@gmail.com>
9411 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9412 override user -mfp-in-toc.
9413 (offsettable_ok_by_alignment): Consider just the current access
9414 rather than the whole object, unless BLKmode. Handle
9415 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
9416 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
9417 for -mcmodel=medium.
9418 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
9419 override user -mfp-in-toc or -msum-in-toc. Default to
9420 -mno-fp-in-toc for -mcmodel=medium.
9422 2013-06-06 DJ Delorie <dj@redhat.com>
9424 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
9425 TARGET_VALID_POINTER_MODE.
9427 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
9428 Pat Haugen <pthaugen@us.ibm.com>
9429 Peter Bergner <bergner@vnet.ibm.com>
9431 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9432 Document new power8 builtins.
9434 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
9435 condition code register, to allow 128-bit logical operations to be
9436 done in the VSX or GPR registers.
9437 (nor<mode>3): Use the canonical form for nor.
9438 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9439 vclz*, and vpopcnt* vector instructions.
9440 (nand<mode>3): Likewise.
9441 (orc<mode>3): Likewise.
9442 (clz<mode>2): LIkewise.
9443 (popcount<mode>2): Likewise.
9445 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9446 that only the GPRs are recognized.
9448 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9449 support for new power8 builtins.
9451 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9453 (xscvdpspn): Likewise.
9459 (vpopcnt): Likewise.
9460 (vpopcntb): Likewise.
9461 (vpopcnth): Likewise.
9462 (vpopcntw): Likewise.
9463 (vpopcntd): Likewise.
9468 (eqv_v16qi3): Likewise.
9469 (eqv_v8hi3): Likewise.
9470 (eqv_v4si3): Likewise.
9471 (eqv_v2di3): Likewise.
9472 (eqv_v4sf3): Likewise.
9473 (eqv_v2df3): Likewise.
9475 (nand_v16qi3): Likewise.
9476 (nand_v8hi3): Likewise.
9477 (nand_v4si3): Likewise.
9478 (nand_v2di3): Likewise.
9479 (nand_v4sf3): Likewise.
9480 (nand_v2df3): Likewise.
9482 (orc_v16qi3): Likewise.
9483 (orc_v8hi3): Likewise.
9484 (orc_v4si3): Likewise.
9485 (orc_v2di3): Likewise.
9486 (orc_v4sf3): Likewise.
9487 (orc_v2df3): Likewise.
9489 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9490 allow power8 quad mode in 64-bit.
9491 (rs6000_builtin_vectorized_function): Add support to vectorize
9492 ISA 2.07 count leading zeros, population count builtins.
9493 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9494 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9495 (builtin_function_type): Add vgbbd builtin function which takes an
9497 (altivec_expand_vec_perm_const): Add support for new power8 merge
9500 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9501 that does not include TImdoe for use with 32-bit.
9502 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9504 (UNSPEC_VSX_CVDPSPN): Likewise.
9505 (vsx_xscvdpspn): Likewise.
9506 (vsx_xscvspdpn): Likewise.
9507 (vsx_xscvdpspn_scalar): Likewise.
9508 (vsx_xscvspdpn_directmove): Likewise.
9509 (vsx_and<mode>3): Split logical operations into 32-bit and
9510 64-bit. Add support to do logical operations on TImode as well as
9511 VSX vector types. Allow logical operations to be done in either
9512 VSX registers or in general purpose registers in 64-bit mode. Add
9513 splitters if GPRs were used. For AND, add clobber of CCmode to
9514 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9516 (vsx_and<mode>3_32bit): Likewise.
9517 (vsx_and<mode>3_64bit): Likewise.
9518 (vsx_ior<mode>3): Likewise.
9519 (vsx_ior<mode>3_32bit): Likewise.
9520 (vsx_ior<mode>3_64bit): Likewise.
9521 (vsx_xor<mode>3): Likewise.
9522 (vsx_xor<mode>3_32bit): Likewise.
9523 (vsx_xor<mode>3_64bit): Likewise.
9524 (vsx_one_cmpl<mode>2): Likewise.
9525 (vsx_one_cmpl<mode>2_32bit): Likewise.
9526 (vsx_one_cmpl<mode>2_64bit): Likewise.
9527 (vsx_nor<mode>3): Likewise.
9528 (vsx_nor<mode>3_32bit): Likewise.
9529 (vsx_nor<mode>3_64bit): Likewise.
9530 (vsx_andc<mode>3): Likewise.
9531 (vsx_andc<mode>3_32bit): Likewise.
9532 (vsx_andc<mode>3_64bit): Likewise.
9533 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
9534 and xxlorc instructions.
9535 (vsx_eqv<mode>3_64bit): Likewise.
9536 (vsx_nand<mode>3_32bit): Likewise.
9537 (vsx_nand<mode>3_64bit): Likewise.
9538 (vsx_orc<mode>3_32bit): Likewise.
9539 (vsx_orc<mode>3_64bit): Likewise.
9541 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
9543 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
9545 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
9546 (p8_vmrgow): Likewise.
9547 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
9548 GPRs to be split under VSX.
9549 (p8v_clz<mode>2): Add power8 count leading zero support.
9550 (p8v_popcount<mode>2): Add power8 population count support.
9551 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
9554 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
9557 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
9559 (vec_nand): Likewise.
9560 (vec_vclz): Likewise.
9561 (vec_vclzb): Likewise.
9562 (vec_vclzd): Likewise.
9563 (vec_vclzh): Likewise.
9564 (vec_vclzw): Likewise.
9565 (vec_vgbbd): Likewise.
9566 (vec_vmrgew): Likewise.
9567 (vec_vmrgow): Likewise.
9568 (vec_vpopcnt): Likewise.
9569 (vec_vpopcntb): Likewise.
9570 (vec_vpopcntd): Likewise.
9571 (vec_vpopcnth): Likewise.
9572 (vec_vpopcntw): Likewise.
9574 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9576 PR rtl-optimization/57468
9577 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
9580 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9582 PR rtl-optimization/57459
9583 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
9584 type when setting live regs.
9586 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9588 * config/s390/s390.opt (mlra): New option.
9589 * config/s390/s390.c (s390_decompose_address): Check displacement
9590 for all registers for LRA.
9591 (s390_secondary_reload): Don't used secondary reloads for LRA.
9592 (s390_lra_p): New function.
9593 (TARGET_LRA_P): Define.
9594 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
9595 of attribute cpu_facility to zarch for the last alternative.
9596 (*cmpmem_short): Ditto.
9598 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9600 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9601 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9602 (arm_expand_prologue): Likewise.
9604 2013-06-06 Teresa Johnson <tejohnson@google.com>
9607 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9608 as this is now done by redirect_edge_and_branch_force.
9609 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9610 barriers, and fix interaction with splitting.
9611 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9612 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9613 reflect changes made in the routine.
9614 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9615 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9616 since this is called in cfglayout mode, and replace partition fixup
9617 with assert as that is now done by force_nonfallthru_and_redirect.
9618 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9619 already be marked with region crossing note.
9620 (insert_section_boundary_note): Make non-static, gate on flag
9621 has_bb_partition, rewrite to also check for multiple partitions.
9622 (rest_of_handle_reorder_blocks): Remove call to
9623 insert_section_boundary_note, now done later during free_cfg.
9624 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9625 * bb-reorder.h (insert_section_boundary_note): Declare.
9626 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9627 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9628 invoke insert_section_boundary_note.
9629 (try_redirect_by_replacing_jump): Remove unnecessary
9630 check for region crossing note.
9631 (fixup_partition_crossing): New function.
9632 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9633 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9634 in non-cfglayout mode.
9635 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9636 remove old code that tried to do this. Emit barrier correctly
9637 when we are in cfglayout mode.
9638 (last_bb_in_partition): New function.
9639 (rtl_split_edge): Correctly fixup partition boundaries.
9640 (commit_one_edge_insertion): Remove old code that tried to
9641 fixup region crossing edge since this is now handled in
9642 split_block, and set up insertion point correctly since
9643 block may now end in a jump.
9644 (verify_hot_cold_block_grouping): Guard against checking when not in
9645 linearized RTL mode.
9646 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9648 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9649 rtl_verify_flow_info, so not called in cfglayout mode.
9650 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9651 (fixup_reorder_chain): Remove old code that attempted to fixup region
9652 crossing note as this is now handled in force_nonfallthru_and_redirect.
9653 (duplicate_insn_chain): Don't duplicate switch section notes.
9654 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9656 * basic-block.h (emit_barrier_after_bb): Declare.
9658 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9660 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9661 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9662 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9664 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9666 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9668 * config/arm/ldmstm.md: Regenerate.
9670 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9672 * config/arm/sync.md (atomic_loaddi_1):
9673 Disable predication for arm_restrict_it.
9674 (arm_load_exclusive<mode>): Likewise.
9675 (arm_load_exclusivesi): Likewise.
9676 (arm_load_exclusivedi): Likewise.
9677 (arm_load_acquire_exclusive<mode>): Likewise.
9678 (arm_load_acquire_exclusivesi): Likewise.
9679 (arm_load_acquire_exclusivedi): Likewise.
9680 (arm_store_exclusive<mode>): Likewise.
9681 (arm_store_exclusive<mode>): Likewise.
9682 (arm_store_release_exclusivedi): Likewise.
9683 (arm_store_release_exclusive<mode>): Likewise.
9685 2013-06-06 Richard Biener <rguenther@suse.de>
9687 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9689 (lto_tag_is_tree_code_p): Adjust.
9690 (lto_tag_is_gimple_code_p): Likewise.
9691 (lto_gimple_code_to_tag): Likewise.
9692 (lto_tag_to_gimple_code): Likewise.
9693 (lto_tree_code_to_tag): Likewise.
9694 (lto_tag_to_tree_code): Likewise.
9695 * data-streamer.h (streamer_write_hwi_in_range): Use
9696 uhwi streaming to stream the normalized range.
9697 (streamer_read_hwi_in_range): Likewise.
9699 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9701 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9702 (predicable_short_it): Likewise.
9703 (predicated): Likewise.
9704 (enabled): Handle above.
9705 (define_cond_exec): Set predicated attribute to yes.
9707 2013-06-05 Mike Stump <mikestump@comcast.net>
9709 * gdbinit.in (__FUNCTION__): Add.
9711 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
9713 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9714 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9716 2013-06-05 Jan Hubicka <jh@suse.cz>
9718 * varasm.c (mark_decl_referenced): Revert the removal until targets
9721 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
9723 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9724 instead of mark_decl_referenced.
9726 2013-06-05 Jan Hubicka <jh@suse.cz>
9728 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9729 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9730 and symtab_used_from_object_file_p.
9731 (cgraph_make_node_local_1): Clear forced_by_abi.
9732 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9733 * cgraph.h (symtab_node_base): Add forced_by_abi.
9734 (decide_is_variable_needed): Remove.
9735 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9736 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9737 (decide_is_symbol_needed): ... this one; handle symbols in general;
9738 always analyze virtuals; honnor forced_by_abi.
9739 (cgraph_finalize_function): Update.
9740 (varpool_finalize_decl): Update.
9741 (symbol_defined_and_needed): Remove.
9742 (analyze_functions): Update.
9743 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9744 output_refs, input_overwrite_node): Handle forced_by_abi.
9745 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9746 (address_taken_from_non_vtable_p): ... this one.
9747 (comdat_can_be_unshared_p_1): New function.
9748 (cgraph_comdat_can_be_unshared_p): Rename to ...
9749 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9750 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9751 (function_and_variable_visibility): Clear forced_by_abi as needed.
9752 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9753 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
9754 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9755 * varpool.c (decide_is_variable_needed): Remove.
9757 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9759 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9760 (arm_option_override): Override arm_restrict_it where appropriate.
9761 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9762 * config/arm/arm.opt (mrestrict-it): New command-line option.
9763 * doc/invoke.texi: Document -mrestrict-it.
9765 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9767 * tsan.c (tsan_atomic_table): Make const.
9769 2013-06-05 Richard Biener <rguenther@suse.de>
9771 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9772 index associated with the tree we are supposed to replace.
9773 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9774 TREE_ASM_WRITTEN as zero for everything but SSA names.
9776 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9778 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9780 2013-06-04 Jan Hubicka <jh@suse.cz>
9782 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9783 (input_node, input_varpool_node): Handle correctly external same
9785 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9786 nodes at ltrans stage.
9788 2013-06-04 Jan Hubicka <jh@suse.cz>
9790 * ipa-inline.c (update_caller_keys): Fix availability test.
9791 (update_callee_keys): Likewise.
9792 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
9793 to follow ELF standard.
9795 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
9797 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
9798 (mips64r5900el-*-elf*): New configurations.
9799 * config/mips/mips-cpus.def (r5900): New processor.
9800 * config/mips/mips-tables.opt: Regenerate.
9801 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
9802 (mips_issue_rate): Handle PROCESSOR_R5900.
9803 (mips_reorg_process_insns): Force reorder mode for the R5900.
9804 * config/mips/mips.h (TARGET_MIPS5900): Define.
9805 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
9807 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
9809 * config/mips/mips.md (processor): Add r5900.
9810 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
9812 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9814 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
9815 into function to generate MOVI instruction.
9816 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
9817 (aarch64_preferred_simd_mode): Turn into wrapper.
9818 (aarch64_output_scalar_simd_mov_immediate): New function.
9819 * config/aarch64/aarch64-protos.h: Add prototype for above.
9821 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9823 * config/aarch64/aarch64.c (simd_immediate_info): Remove
9824 element_char member.
9825 (sizetochar): Return signed char.
9826 (aarch64_simd_valid_immediate): Remove elchar and other
9827 unnecessary variables.
9828 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
9829 Calculate element_char as required.
9830 * config/aarch64/aarch64-protos.h: Update and move prototype
9831 for aarch64_output_simd_mov_immediate.
9832 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
9835 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9837 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
9838 information completed by aarch64_simd_valid_immediate.
9839 (aarch64_legitimate_constant_p): Update arguments.
9840 (aarch64_simd_valid_immediate): Work with struct rather than many
9842 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
9843 (aarch64_simd_make_constant): Update arguments.
9844 (aarch64_output_simd_mov_immediate): Work with struct rather than
9845 many pointers. Output immediate directly rather than as operand.
9846 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9848 * config/aarch64/constraints.md (Dn): Update arguments.
9850 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9852 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
9854 (aarch64_simd_immediate_valid_for_move): Remove.
9855 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
9856 (aarch64_simd_make_constant): Update call.
9857 (aarch64_output_simd_mov_immediate): Update call.
9858 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9860 * config/aarch64/constraints.md (Dn): Update call.
9862 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9864 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
9865 return type to bool for prototype.
9866 (aarch64_legitimate_constant_p): Check for true instead of not -1.
9867 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
9868 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
9870 2013-06-04 Catherine Moore <clm@codesourcery.com>
9872 * config/mips/mips.opt (meva): New.
9873 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
9874 (ASM_SPEC): Handle -meva.
9875 * doc/invoke.texi (meva): Document.
9877 2013-06-04 Alan Modra <amodra@gmail.com>
9879 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
9882 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9884 * rtl.def: Add extra fourth optional field to define_cond_exec.
9885 * gensupport.c (process_one_cond_exec): Process attributes from
9887 * doc/md.texi: Document fourth field in define_cond_exec.
9889 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
9891 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
9892 out the processing order as in store_bit_field_1.
9894 2013-06-04 Jan Hubicka <jh@suse.cz>
9897 * cgraphunit.c (cgraph_process_same_body_aliases): Create
9898 non-VAR_DECL node if it does not exist yet.
9900 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
9902 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
9903 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
9904 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
9905 target_cpu_default setting.
9907 2013-06-03 Teresa Johnson <tejohnson@google.com>
9909 * dumpfile.c (opt_info_switch_p): Change -fopt-info
9910 default to -fopt-info=optimized instead of all.
9911 * doc/invoke.texi: Ditto.
9912 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
9913 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
9914 (execute_vect_slp): Emit BB vectorization success under
9915 MSG_OPTIMIZED_LOCATIONS.
9916 * tree-vect-slp.c (vect_slp_transform_bb): Change
9917 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
9918 * tree-vect-loop.c (vect_transform_loop): Ditto.
9920 2013-06-03 Jason Merrill <jason@redhat.com>
9923 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9924 Use TARGET_EXPR for C++.
9926 2013-06-03 Jakub Jelinek <jakub@redhat.com>
9928 PR rtl-optimization/57268
9929 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
9930 if DEBUG_INSN_P (insn).
9933 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9935 PR rtl-optimization/57268
9936 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9937 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9939 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9941 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
9942 (ix86_avoid_lea_for_addr): Likewise.
9943 (exact_dependency_1): Likewise.
9944 (ix86_adjust_cost): Likewise.
9945 (swap_top_of_ready_list): Fix formatting and !reload_completed check
9947 (do_reorder_for_imul): Fix typo, formatting and
9948 !reload_completed check removed.
9949 (ix86_sched_reorder): Fix typo and formatting.
9950 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
9953 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
9955 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
9957 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
9959 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
9960 <STRING_CST>: Likewise.
9961 <VECTOR_CST>: Likewise.
9963 2013-06-01 Janus Weil <janus@gcc.gnu.org>
9964 Mikael Morin <mikael@gcc.gnu.org>
9966 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
9967 * config.in: Regenerated.
9968 * configure: Regenerated.
9970 2013-06-01 Jan Hubicka <jh@suse.cz>
9973 * cgraphunit.c (compile): When weakref is not supported,
9974 set up transparent aliases before final output pass.
9975 * varasm.c (assemble_alias): Do not try to do it here.
9977 2013-06-01 Jan Hubicka <jh@suse.cz>
9980 * passes.c (for_per_function): Skip unanalyzed functions.
9982 2013-06-01 Jan Hubicka <jh@suse.cz>
9984 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
9985 (lto_symtab_merge_symbols_1): ... this one.
9986 (lto_symtab_merge_cgraph_nodes): Rename to ...
9987 (lto_symtab_merge_symbols): ... this one; simplify.
9988 * cgraph.c (same_body_aliases_done): Rename to ...
9989 (cpp_implicit_aliases_done): ... this one.
9990 (cgraph_create_function_alias): Update.
9991 (cgraph_same_body_alias): Update.
9992 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
9993 (verify_edge_corresponds_to_fndecl): Simplify.
9994 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
9995 (cgraph_node): Remove same_body_alias.
9996 (varpool_node): Remove alias_of and extra_name_alias.
9997 (same_body_aliases_done): Rename to ..
9998 (cpp_implicit_aliases_done): ... this one.
9999 (symtab_alias_ultimate_target): Add default parameter.
10000 (symtab_resolve_alias): New function.
10001 (fixup_same_cpp_alias_visibility): Declare.
10002 (cgraph_function_node): Add default parameter.
10003 (cgraph_node_asm_name): Likewise.
10004 (cgraph_function_or_thunk_node): Add default parameter; do
10005 not ICE when it is NULL.
10006 (varpool_variable_node): Likewise.
10007 * tree-emutls.c (create_emultls_var): Update.
10008 (ipa_lower_emutls): Update.
10009 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
10010 (cgraph_reset_node): Reset alias info.
10011 (cgraph_finalize_function): Update.
10012 (fixup_same_cpp_alias_visibility): Move to symtab.c.
10013 (analyze_function): Simplify.
10014 (cgraph_process_same_body_aliases): Simplify.
10015 (analyze_functions): Fixup same body aliases.
10016 (handle_alias_pairs): Simplify.
10017 (assemble_thunk): Update.
10018 (assemble_thunks_and_aliases): Update.
10019 (output_weakrefs): Rewrite.
10020 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
10021 (lto_output_varpool_node): Likewise.
10022 (compute_ltrans_boundary): Remve assert.
10023 (get_alias_symbol): New functoin.
10024 (input_node): Rewrite alias handling.
10025 (input_varpool_node): Likewise.
10026 * ipa-pure-const.c (propagate_pure_const): Fix formating.
10027 * ipa.c (process_references): Handle weakrefs correctly.
10028 (symtab_remove_unreachable_nodes): Likewise.
10029 * trans-mem.c (get_cg_data): Update.
10030 (ipa_tm_create_version_alias): Update.
10031 (ipa_tm_execute): Update.
10032 * symtab.c (dump_symtab_base): Dump aliases.
10033 (verify_symtab_base): Verify aliases.
10034 (symtab_node_availability): New function.
10035 (symtab_alias_ultimate_target): Simplify.
10036 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
10037 handle all the fixup cases.
10038 (symtab_resolve_alias): New function.
10039 * passes.c (ipa_write_summaries): Handle weakrefs.
10040 * varpool.c (varpool_analyze_node): Simplify.
10041 (assemble_aliases): Update.
10042 (varpool_create_variable_alias): Simplify.
10043 (varpool_extra_name_alias): Simplify.
10044 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
10045 (lto_symtab_merge_symbols): ... this one.
10047 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
10050 PR rtl-optimization/57268
10051 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10052 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10054 2013-06-01 Tobias Burnus <burnus@net-b.de>
10056 Partially reverted:
10057 2013-05-31 Tobias Burnus <burnus@net-b.de>
10059 PR middle-end/57073
10060 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
10063 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10065 PR rtl-optimization/57268
10066 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10067 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10069 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10071 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
10072 unordered comparison operators when -fno-trapping-math is in effect
10074 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
10075 and implement unordered comparison operators properly on the e500.
10077 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10079 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
10080 for constant scalar integers.
10081 (simplify_relational_operation_1): Likewise.
10083 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
10085 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
10086 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
10089 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10090 Igor Zamyatin <igor.zamyatin@intel.com>
10092 Silvermont (SLM) architecture performance tuning.
10093 * config/i386/i386.h (enum ix86_tune_indices): Add
10094 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
10095 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
10097 * config/i386/i386.c (initial_ix86_tune_features)
10098 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
10099 (ix86_lea_outperforms): Handle Silvermont tuning.
10100 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
10102 (ix86_use_lea_for_mov): Likewise.
10103 (ix86_avoid_lea_for_addr): Likewise.
10104 (ix86_lea_for_add_ok): Likewise.
10105 (exact_dependency_1): New function.
10106 (exact_store_load_dependency): Likewise.
10107 (ix86_adjust_cost): Handle Silvermont tuning.
10108 (do_reoder_for_imul): Likewise.
10109 (swap_top_of_ready_list): New function.
10110 (ix86_sched_reorder): Changed to handle Silvermont tuning.
10112 * config/i386/i386.md (peepholes that split memory operand in fp
10115 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10117 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
10118 Remove un-necessary braces.
10120 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10122 * config/aarch64/aarch64.c (aarch64_classify_symbol):
10123 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
10125 2013-05-31 Tobias Burnus <burnus@net-b.de>
10127 PR middle-end/57073
10128 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
10130 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10133 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
10134 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
10135 * config/arm/neon.md (iordi3_neon): Remove.
10136 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
10137 * config/arm/predicates.md (imm_for_neon_logic_operand):
10138 Move to earlier in the file.
10139 (neon_logic_op2): Likewise.
10140 (arm_iordi_operand_neon): New predicate.
10142 2013-05-31 Richard Biener <rguenther@suse.de>
10144 PR tree-optimization/57478
10145 PR tree-optimization/57453
10146 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
10149 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10151 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
10152 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
10154 2013-05-30 Tobias Burnus <burnus@net-b.de>
10155 Thomas Koenig <tkoenig@gcc.gnu.org>
10157 PR middle-end/57073
10158 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
10159 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
10161 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
10163 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
10165 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
10167 * target.def (register_usage_leveling_p): New hook.
10168 * targhooks.c (default_register_usage_leveling_p): New.
10169 * targhooks.h (default_register_usage_leveling_p): New prototype.
10170 * lra-assigns.c (register_usage_leveling_p): Use the hook.
10171 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
10172 * doc/tm.texi: Update.
10173 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
10175 2013-05-30 Ian Bolton <ian.bolton@arm.com>
10177 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
10178 (*insv_reg<mode>): New define_insn.
10180 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
10182 PR rtl-optimization/57439
10183 * postreload.c (move2add_valid_value_p): Check that we have
10184 a zero subreg_regno_offset when accessing the register in
10185 the requested mode.
10187 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10188 Igor Zamyatin <igor.zamyatin@intel.com>
10190 Silvermont (SLM) architecture pipeline model, tuning and
10192 * config.gcc: Add slm config options and target.
10194 * config/i386/slm.md: New.
10196 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
10198 * config/i386/i386-c.c (ix86_target_macros_internal): New case
10200 (ix86_target_macros_internal): Likewise.
10202 * config/i386/i386.c (slm_cost): New cost.
10203 (m_SLM): New macro flag.
10204 (initial_ix86_tune_features): Set m_SLM.
10205 (x86_accumulate_outgoing_args): Likewise.
10206 (x86_arch_always_fancy_math_387): Likewise.
10207 (processor_target_table): Add slm cost.
10208 (cpu_names): Add slm cpu name.
10209 (x86_option_override_internal): Set SLM ISA.
10210 (ix86_issue_rate): New case PROCESSOR_SLM.
10211 (ia32_multipass_dfa_lookahead): Likewise.
10212 (fold_builtin_cpu): Add slm.
10214 * config/i386/i386.h (TARGET_SLM): New target macro.
10215 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
10216 (processor_type): Add PROCESSOR_SLM.
10218 * config/i386/i386.md (cpu): Add new value "slm".
10219 (slm.md): Include slm.md.
10221 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
10222 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10224 * config/arm/arm-protos.h: Add and update function protos.
10225 * config/arm/arm.c (use_simple_return_p): New added.
10226 (thumb2_expand_return): Check simple_return flag.
10227 * config/arm/arm.md: Add simple_return and conditional simple_return.
10228 * config/arm/iterators.md: Add iterator for return and simple_return.
10230 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10232 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
10233 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
10234 (arm_emit_vfp_multi_reg_pop): Likewise.
10235 (thumb2_emit_ldrd_pop): Likewise.
10236 (arm_expand_epilogue): Add misc REG_CFA notes.
10237 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
10239 2013-05-29 Lawrence Crowl <crowl@google.com>
10241 * config/arm/t-arm: Update for below.
10243 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
10244 Change type to hash_table. Update dependent calls and types.
10246 * config/i386/t-cygming: Update for below.
10248 * config/i386/t-interix: Update for below.
10250 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
10251 Change type to hash_table. Update dependent calls and types.
10252 (i386_find_on_wrapper_list::wrappers): Likewise.
10254 * config/ia64/t-ia64: Update for below.
10256 * config/ia64/ia64.c (bundle_state_table):
10257 Change type to hash_table. Update dependent calls and types.
10259 * config/mips/mips.c (mips_reorg_process_insns::htab):
10260 Change type to hash_table. Update dependent calls and types.
10262 * config/sol2.c (solaris_comdat_htab):
10263 Change type to hash_table. Update dependent calls and types.
10265 * config/t-sol2: Update for above.
10267 2013-05-29 Teresa Johnson <tejohnson@google.com>
10269 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
10270 functions are not yet marked as defined.
10272 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
10273 Pat Haugen <pthaugen@us.ibm.com>
10274 Peter Bergner <bergner@vnet.ibm.com>
10276 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
10280 (vrotl<mode>3): Likewise.
10281 (vashl<mode>3): Likewise.
10282 (vlshr<mode>3): Likewise.
10283 (vashr<mode>3): Likewise.
10285 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10286 support for power8 V2DI builtins.
10288 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
10289 power8 V2DI builtins.
10290 (vupkhsw): Likewise.
10291 (vupklsw): Likewise.
10292 (vaddudm): Likewise.
10293 (vminsd): Likewise.
10294 (vmaxsd): Likewise.
10295 (vminud): Likewise.
10296 (vmaxud): Likewise.
10297 (vpkudum): Likewise.
10298 (vpksdss): Likewise.
10299 (vpkudus): Likewise.
10300 (vpksdus): Likewise.
10305 (vsubudm): Likewise.
10306 (vcmpequd): Likewise.
10307 (vcmpgtsd): Likewise.
10308 (vcmpgtud): Likewise.
10309 (vcmpequd_p): Likewise.
10310 (vcmpgtsd_p): Likewise.
10311 (vcmpgtud_p): Likewise.
10312 (vupkhsw): Likewise.
10313 (vupklsw): Likewise.
10314 (vaddudm): Likewise.
10315 (vmaxsd): Likewise.
10316 (vmaxud): Likewise.
10317 (vminsd): Likewise.
10318 (vminud): Likewise.
10319 (vpksdss): Likewise.
10320 (vpksdus): Likewise.
10321 (vpkudum): Likewise.
10322 (vpkudus): Likewise.
10327 (vsubudm): Likewise.
10329 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
10330 support for power8 V2DI instructions.
10332 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
10333 power8 V2DI instructions. Combine pack and unpack insns to use an
10334 iterator for each mode. Check whether a particular mode supports
10335 Altivec instructions instead of just checking TARGET_ALTIVEC.
10336 (UNSPEC_VPKUWUM): Likewise.
10337 (UNSPEC_VPKSHSS): Likewise.
10338 (UNSPEC_VPKSWSS): Likewise.
10339 (UNSPEC_VPKUHUS): Likewise.
10340 (UNSPEC_VPKSHUS): Likewise.
10341 (UNSPEC_VPKUWUS): Likewise.
10342 (UNSPEC_VPKSWUS): Likewise.
10343 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
10344 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
10345 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
10346 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
10347 (UNSPEC_VUPKHSB): Likewise.
10348 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
10349 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
10350 (UNSPEC_VUPKHSH): Likewise.
10351 (UNSPEC_VUPKLSB): Likewise.
10352 (UNSPEC_VUPKLSH): Likewise.
10354 (VI_char): Likewise.
10355 (VI_scalar): Likewise.
10356 (VI_unit): Likewise.
10358 (VP_small): Likewise.
10359 (VP_small_lc): Likewise.
10360 (VU_char): Likewise.
10361 (add<mode>3): Likewise.
10362 (altivec_vaddcuw): Likewise.
10363 (altivec_vaddu<VI_char>s): Likewise.
10364 (altivec_vadds<VI_char>s): Likewise.
10365 (sub<mode>3): Likewise.
10366 (altivec_vsubcuw): Likewise.
10367 (altivec_vsubu<VI_char>s): Likewise.
10368 (altivec_vsubs<VI_char>s): Likewise.
10369 (altivec_vavgs<VI_char>): Likewise.
10370 (altivec_vcmpbfp): Likewise.
10371 (altivec_eq<mode>): Likewise.
10372 (altivec_gt<mode>): Likewise.
10373 (altivec_gtu<mode>): Likewise.
10374 (umax<mode>3): Likewise.
10375 (smax<mode>3): Likewise.
10376 (umin<mode>3): Likewise.
10377 (smin<mode>3): Likewise.
10378 (altivec_vpkuhum): Likewise.
10379 (altivec_vpkuwum): Likewise.
10380 (altivec_vpkshss): Likewise.
10381 (altivec_vpkswss): Likewise.
10382 (altivec_vpkuhus): Likewise.
10383 (altivec_vpkshus): Likewise.
10384 (altivec_vpkuwus): Likewise.
10385 (altivec_vpkswus): Likewise.
10386 (altivec_vpks<VI_char>ss): Likewise.
10387 (altivec_vpks<VI_char>us): Likewise.
10388 (altivec_vpku<VI_char>us): Likewise.
10389 (altivec_vpku<VI_char>um): Likewise.
10390 (altivec_vrl<VI_char>): Likewise.
10391 (altivec_vsl<VI_char>): Likewise.
10392 (altivec_vsr<VI_char>): Likewise.
10393 (altivec_vsra<VI_char>): Likewise.
10394 (altivec_vsldoi_<mode>): Likewise.
10395 (altivec_vupkhsb): Likewise.
10396 (altivec_vupkhs<VU_char>): Likewise.
10397 (altivec_vupkls<VU_char>): Likewise.
10398 (altivec_vupkhsh): Likewise.
10399 (altivec_vupklsb): Likewise.
10400 (altivec_vupklsh): Likewise.
10401 (altivec_vcmpequ<VI_char>_p): Likewise.
10402 (altivec_vcmpgts<VI_char>_p): Likewise.
10403 (altivec_vcmpgtu<VI_char>_p): Likewise.
10404 (abs<mode>2): Likewise.
10405 (vec_unpacks_hi_v16qi): Likewise.
10406 (vec_unpacks_hi_v8hi): Likewise.
10407 (vec_unpacks_lo_v16qi): Likewise.
10408 (vec_unpacks_hi_<VP_small_lc>): Likewise.
10409 (vec_unpacks_lo_v8hi): Likewise.
10410 (vec_unpacks_lo_<VP_small_lc>): Likewise.
10411 (vec_pack_trunc_v8h): Likewise.
10412 (vec_pack_trunc_v4si): Likewise.
10413 (vec_pack_trunc_<mode>): Likewise.
10415 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
10417 (vec_vmaxsd): Likewise.
10418 (vec_vmaxud): Likewise.
10419 (vec_vminsd): Likewise.
10420 (vec_vminud): Likewise.
10421 (vec_vpksdss): Likewise.
10422 (vec_vpksdus): Likewise.
10423 (vec_vpkudum): Likewise.
10424 (vec_vpkudus): Likewise.
10425 (vec_vrld): Likewise.
10426 (vec_vsld): Likewise.
10427 (vec_vsrad): Likewise.
10428 (vec_vsrd): Likewise.
10429 (vec_vsubudm): Likewise.
10430 (vec_vupkhsw): Likewise.
10431 (vec_vupklsw): Likewise.
10433 2013-05-29 Jan Hubicka <jh@suse.cz>
10435 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
10436 flags; reorder rest of fields in more consistent way.
10437 (varpool_node): Remove analyzed, finalized and alias.
10438 (cgraph_ndoe): Likewise.
10439 (symtab_alias_ultimate_target): New function.
10440 (cgraph_function_node): Move offline.
10441 (cgraph_reset_node): Declare.
10442 (cgraph_comdat_can_be_unshared_p): Remove.
10443 (varpool_remove_initializer): Declare.
10444 (varpool_first_defined_variable, varpool_next_defined_variable
10445 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10446 (cgraph_function_with_gimple_body_p): Update.
10447 (varpool_all_refs_explicit_p): Update.
10448 (symtab_alias_target): New function.
10449 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10450 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
10451 (cgraph_function_or_thunk_node): Simplify using
10452 symtab_alias_ultimate_target.
10453 (varpool_variable_node): Likewise.
10454 * cgraph.c (cgraph_create_function_alias): Update.
10455 (cgraph_add_thunk): Update.
10456 (cgraph_remove_node): Update.
10457 (dump_cgraph_node): Do not dump removed flags.
10458 (cgraph_function_body_availability): Update.
10459 (cgraph_propagate_frequency): Update.
10460 (verify_cgraph_node): Check sanity of local flag.
10461 (cgraph_function_node): Move here from cgraph.h; revamp for
10462 cgraph_function_or_thunk_node.
10463 * lto-symtab.c (lto_varpool_replace_node): Update.
10464 (lto_symtab_resolve_can_prevail_p): Update.
10465 (lto_symtab_merge_cgraph_nodes): Update.
10466 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10467 propagate_constants_accross_call, devirtualization_time_bonus,
10468 ipcp_propagate_stage): Update.
10469 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
10470 * ipa-inline-transform.c (clone_inlined_nodes,
10471 preserve_function_body_p): Update.
10472 * ipa-reference.c (propagate): Update.
10473 (write_node_summary_p): Update.
10474 * toplev.c (wrapup_global_declaration_2): Update.
10475 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10476 (analyze_function) ... this one.
10477 (cgraph_process_new_functions): Update.
10478 (cgraph_reset_node): Export.
10479 (cgraph_finalize_function): Update.
10480 (cgraph_add_new_function): Update.
10481 (process_function_and_variable_attributes): Update.
10482 (varpool_finalize_decl): Update.
10483 (symbol_finalized): Remove.
10484 (symbol_finalized_and_needed): Rename to ...
10485 (symbol_defined_and_needed): ... update.
10486 (cgraph_analyze_functions): Update.
10487 (handle_alias_pairs): Update.
10488 (mark_functions_to_output): Update.
10489 (assemble_thunk): Update.
10490 (output_in_order): Update.
10491 (output_weakrefs): Update.
10492 (finalize_compilation_unit): Update.
10493 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10494 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10495 input_node, input_varpool_node): Update.
10496 * dbxout.c (dbxout_expand_expr): Update.
10497 * cgraphclones.c (cgraph_clone_node): Update.
10498 (cgraph_copy_node_for_versioning): Update.
10499 (cgraph_materialize_clone): Update.
10500 (cgraph_materialize_all_clones): Update.
10501 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10502 propagate_pure_const, propagate_nothrow): Update.
10503 * lto-streamer-out.c (lto_output, write_symbol): Update.
10504 * ipa-utils.c (ipa_reverse_postorder): Update.
10505 * ipa-inline.c (can_inline_edge_p): Update.
10506 (update_caller_keys, ipa_inline): Update.
10507 * dwarf2out.c (reference_to_unused,
10508 premark_types_used_by_global_vars_helper): Update.
10509 * tree-eh.c (tree_could_trap_p): Update.
10510 * ipa-split.c (consider_split, execute_split_functions): Update.
10511 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
10512 has_addr_references_p): Update; move ahead in file for better
10514 (process_references): Simplify.
10515 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10516 bodies are removed.
10517 (cgraph_comdat_can_be_unshared_p): Make static.
10518 (cgraph_externally_visible_p): Update.
10519 (varpool_externally_visible_p): Update.
10520 (function_and_variable_visibility): Update.
10521 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10522 ipa_tm_mark_force_output_node): Update.
10523 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10524 estimate_edge_devirt_benefit, inline_generate_summary,
10525 inline_write_summary): Update.
10526 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10527 * ipa-prop.c (ipa_compute_jump_functions): Update.
10528 (ipa_print_node_params, ipa_prop_read_section,
10529 ipa_update_after_lto_read, read_replacements_section): Update.
10530 * varasm.c (mark_decl_referenced): Update.
10531 (assemble_alias, dump_tm_clone_pairs): Update.
10532 * tree-inline.c (copy_bb): Update.
10533 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
10535 * symtab.c (dump_symtab_base): Print new flags.
10536 (verify_symtab_base): Verify new flags.
10537 (symtab_alias_ultimate_target): New function.
10538 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
10539 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
10541 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
10543 * i386.c (ix86_get_function_versions_dispatcher,
10544 ix86_generate_version_dispatcher_body): Update.
10545 (fold_builtin_cpu): Use varpool_add_new_variable.
10546 * varpool.c (varpool_remove_initializer): Break out from ...
10547 (varpool_remove_node): ... this one.
10548 (dump_varpool_node, varpool_node_for_asm,
10549 cgraph_variable_initializer_availability, varpool_analyze_node,
10550 varpool_assemble_decl, varpool_remove_unreferenced_decls,
10551 varpool_finalize_named_section_flags, varpool_create_variable_alias):
10554 2013-05-29 Jan Hubicka <jh@suse.cz>
10556 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
10558 2013-05-29 Easwaran Raman <eraman@google.com>
10560 PR tree-optimization/57442
10561 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
10562 when control exits the main loop.
10564 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
10566 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
10568 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
10569 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
10570 * rx/t-rx: Add rx100 under multi library matches option for nofpu
10573 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10575 PR tree-optimization/57441
10576 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
10577 Don't limit size of incr_vec to number of candidates.
10579 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
10581 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
10582 and mips16 directories.
10583 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
10584 (MULTILIB_DIRNAMES): Ditto.
10585 (MULTILIB_EXCEPTIONS): Add new exceptions.
10586 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
10587 (MULTILIB_DIRNAMES): Ditto.
10588 (MULTILIB_EXCEPTIONS): Add new exceptions.
10590 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10591 Marcus Shawcroft <marcus.shawcroft@arm.com>
10593 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
10594 SYMBOL_TINY_ABSOLUTE.
10595 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
10596 SYMBOL_TINY_ABSOLUTE.
10597 (aarch64_expand_mov_immediate): Likewise.
10598 (aarch64_classify_symbol): Likewise.
10599 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10600 Permit SYMBOL_TINY_ABSOLUTE.
10601 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10603 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10604 Marcus Shawcroft <marcus.shawcroft@arm.com>
10606 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10607 Refactor if/switch. Replace gcc_assert with if.
10609 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10611 * config/i386/i386.c (initial_ix86_tune_features): Enable
10612 FP Reassociation for AMD bdver1 and bdver2.
10614 2013-05-29 Martin Jambor <mjambor@suse.cz>
10616 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10617 and IMAGPART_EXPR do not occur within other handled_components.
10619 2013-05-29 Richard Biener <rguenther@suse.de>
10621 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10622 access on whether the use is in the BB we currently try to
10624 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10625 vectorize to vect_bb_slp_scalar_cost.
10627 2013-05-29 Richard Biener <rguenther@suse.de>
10629 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10630 computing scalar cost offsetted by stmts that are kept live
10632 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10633 for computation of scalar cost.
10635 2013-05-28 Steve Ellcey <sellcey@mips.com>
10637 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10639 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10641 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10642 array notation built-in reduction functions.
10643 * doc/passes.texi (Passes): Added documentation about changes done
10645 * doc/invoke.texi (C Dialect Options): Added documentation about
10646 the -fcilkplus flag.
10647 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10648 (BUILTINS_DEF): Depend on cilkplus.def.
10649 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10650 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10651 * cilkplus.def: New file.
10653 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
10655 PR rtl-optimization/57439
10656 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10658 2013-05-28 Easwaran Raman <eraman@google.com>
10660 PR tree-optimization/57337
10661 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10662 (find_insert_point): Correctly identify the insertion point
10663 when two statements with the same UID is compared.
10665 2013-05-28 Richard Biener <rguenther@suse.de>
10667 PR tree-optimization/56787
10668 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10669 from the list of data references.
10670 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10672 (vect_analyze_loop_operations): Likewise.
10673 (vect_transform_loop): Remove clobbers.
10675 2013-05-28 Martin Jambor <mjambor@suse.cz>
10677 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10678 and REALPART_EXPRs have scalar type.
10680 2013-05-28 Richard Biener <rguenther@suse.de>
10682 PR tree-optimization/57411
10683 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10685 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10686 virtual operand propagation.
10688 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10690 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10691 destination register for bmasksi_vis.
10692 (vector_init_bshuffle): Likewise.
10693 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10695 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10697 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10698 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10699 mode if the instruction isn't available in the original mode.
10700 * config/sparc/sparc.opt (mfix-ut699): New option.
10701 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10702 (divdf3): Turn into expander.
10703 (divdf3_nofix): New insn.
10704 (divdf3_fix): Likewise.
10705 (divsf3): Disable if -mfix-ut699.
10706 (sqrtdf2): Turn into expander.
10707 (sqrtdf2_nofix): New insn.
10708 (sqrtdf2_fix): Likewise.
10709 (sqrtsf2): Disable if -mfix-ut699.
10711 2013-05-27 Richard Biener <rguenther@suse.de>
10713 PR middle-end/57412
10714 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10715 block for the new loop.
10717 2013-05-27 Richard Biener <rguenther@suse.de>
10719 PR tree-optimization/57343
10720 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10721 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10722 (number_of_iterations_cond): Do not build the folded tree.
10724 2013-05-27 Richard Biener <rguenther@suse.de>
10727 PR middle-end/57381
10728 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10729 OEP_CONSTANT_ADDRESS_OF retained.
10731 PR tree-optimization/57417
10732 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10733 for unchanged base.
10734 (set_ssa_val_to): Compare addresses using
10735 get_addr_base_and_unit_offset.
10737 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
10739 PR rtl-optimization/56833
10740 * postreload.c (move2add_record_mode): New function.
10741 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10742 (move2add_use_add2_insn): Use move2add_record_sym_value.
10743 (move2add_use_add3_insn): Likewise.
10744 (reload_cse_move2add): Use move2add_valid_value_p and
10745 move2add_record_mode. Invalidate call-clobbered and REG_INC
10746 affected regs by setting reg_mode to VOIDmode.
10747 (move2add_note_store): Don't pretend the inside of a SUBREG is
10748 the actual destination. Invalidate single/leading registers by
10749 setting reg_mode to VOIDmode.
10750 Use move2add_record_sym_value, move2add_valid_value_p and
10751 move2add_record_mode.
10753 2013-05-27 Richard Biener <rguenther@suse.de>
10755 PR tree-optimization/57396
10756 * tree-affine.c (double_int_constant_multiple_p): Properly
10757 return false for val == 0 and div != 0.
10759 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10761 * config/mips/mips.h: Use #elif in preprocessor conditions.
10763 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10766 * config/mips/constraints.md (kl): New constraint.
10767 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10768 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10769 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10770 from LO for MIPS16.
10771 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10773 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10776 * config/mips/mips.c (mips_can_inline_p): New function.
10777 (TARGET_CAN_INLINE_P): Define.
10779 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10781 * sched-int.h (ds_t, dw_t): Make unsigned int.
10782 Fix documentation that describes how all the ds_t bits are used.
10783 Reserve the last bit for delayed-branch scheduling.
10784 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10785 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10786 (gen_dep_weak_1): Remove prototype.
10787 * sched-deps.c (get_dep_weak_1): Make static.
10788 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10789 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10790 * doc/tm.texi: Regenerate.
10791 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
10793 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
10796 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
10798 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
10799 Sandra Loosemore <sandra@codesourcery.com>
10801 * config.gcc (powerpc-*): Allow native for with-cpu.
10803 2013-05-24 Jeff Law <law@redhat.com>
10805 PR tree-optimization/57124
10806 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
10807 conversion feeding a condition if the range has an overflow
10808 if -fstrict-overflow. Add warnings for when we do make the
10811 2013-05-24 Dehao Chen <dehao@google.com>
10813 * tree-cfg.c (locus_discrim_map): Fix the typo.
10814 (locus_discrim_hasher): Likewise.
10815 (locus_discrim_hasher::hash): Likewise.
10816 (locus_discrim_hasher::equal): Likewise.
10818 2013-05-24 Martin Jambor <mjambor@suse.cz>
10820 PR tree-optimization/57294
10821 * cgraph.h (ipa_record_stmt_references): Declare.
10822 * cgraphbuild.c (ipa_record_stmt_references): New function.
10823 (build_cgraph_edges): Use ipa_record_stmt_references.
10824 (rebuild_cgraph_edges): Likewise.
10825 (cgraph_rebuild_references): Likewise.
10826 * ipa-prop.c (ipa_modify_call_arguments): Discard references
10827 associated with the old statement and build references from the
10828 newly built statements.
10829 * ipa-ref.c (ipa_remove_stmt_references): New function.
10830 * ipa-ref.h (ipa_remove_stmt_references): Declare.
10832 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
10834 * lra-constraints.c (emit_spill_move): Use smaller mode for
10836 (check_and_process_move): Consider mem-reg moves for secondary
10838 (curr_insn_transform): Don't lose insns emitted before for
10839 secondary memory moves.
10840 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
10841 reg set up in the current insn.
10843 2013-05-24 Dehao Chen <dehao@google.com>
10845 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
10847 (locus_descrim_hasher::equal): Likewise.
10848 (build_gimple_cfg): New discriminator assignment algorithm.
10849 (make_edges): Likewise.
10850 (next_discriminator_for_locus): Likewise.
10851 (same_line_p): Likewise.
10852 (assign_discriminators): Likewise.
10853 (make_cond_expr_edges): Likewise.
10854 (make_gimple_switch_edges): Likewise.
10855 (make_goto_expr_edges): Likewise.
10856 (make_gimple_asm_edges): Likewise.
10858 2013-05-24 Ian Bolton <ian.bolton@arm.com>
10860 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
10861 X format specifier to only display bottom 16 bits.
10862 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
10863 immediate to match for operand 2, since it will be masked.
10865 2013-05-24 Richard Biener <rguenther@suse.de>
10867 PR tree-optimization/57287
10868 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
10869 all SSA names that occur in abnormal PHIs.
10871 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
10874 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
10875 that index is not negative.
10877 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
10879 PR rtl-optimization/55177
10880 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
10881 (simplify_byte_swapping_operation): New.
10882 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
10883 (simplify_relational_operation_1): Deal with BSWAP.
10885 2013-05-23 Richard Henderson <rth@redhat.com>
10888 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
10889 (ix86_reorg): Call it.
10891 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
10894 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
10895 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
10896 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
10898 2013-05-23 Christian Bruel <christian.bruel@st.com>
10901 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
10903 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10904 Marcus Shawcroft <marcus.shawcroft@arm.com>
10906 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
10907 * config/aarch64/constraints.md (Usa): Remove.
10908 * doc/md.texi (AArch64 Usa): Remove.
10910 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10911 Marcus Shawcroft <marcus.shawcroft@arm.com>
10913 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
10914 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
10915 * config/aarch64/predicates.md (aarch64_const_address): Remove.
10916 (aarch64_mov_operand): Use aarch64_mov_operand_p.
10918 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
10920 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
10921 instruction (AdvSIMD).
10922 * config/aarch64/aarch64-builtins.c
10923 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
10924 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
10926 2013-05-23 Martin Jambor <mjambor@suse.cz>
10928 PR middle-end/57347
10929 * tree.h (contains_bitfld_component_ref_p): Declare.
10930 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
10931 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
10933 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
10934 not access a bit-field. Assert all final offsets are byte-aligned.
10936 2013-05-23 Richard Biener <rguenther@suse.de>
10938 PR tree-optimization/57380
10939 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
10940 least one invariant or re-used load.
10941 * passes.c (init_optimization_passes): Move pass_phiprop before
10944 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
10946 * config/aarch64/aarch64-simd.md
10947 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
10949 2013-05-23 Richard Biener <rguenther@suse.de>
10951 PR middle-end/57381
10952 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10953 OEP_CONSTANT_ADDRESS_OF retained.
10955 2013-05-23 Jakub Jelinek <jakub@redhat.com>
10957 PR middle-end/57344
10958 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
10959 don't lower unit. Handle unit not being always BITS_PER_WORD.
10961 2013-05-23 Richard Biener <rguenther@suse.de>
10963 PR rtl-optimization/57341
10964 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
10965 instead of true_dependence.
10967 2013-05-22 David Malcolm <dmalcolm@redhat.com>
10969 * bb-reorder.c (branch_threshold): Make const.
10970 (exec_threshold): Ditto.
10972 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
10973 Pat Haugen <pthaugen@us.ibm.com>
10974 Peter Bergner <bergner@vnet.ibm.com>
10976 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
10977 documentation for the power8 crypto builtins.
10979 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
10981 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
10982 macros for defining power8 builtin functions.
10983 (BU_P8V_AV_2): Likewise.
10984 (BU_P8V_AV_P): Likewise.
10985 (BU_P8V_VSX_1): Likewise.
10986 (BU_P8V_OVERLOAD_1): Likewise.
10987 (BU_P8V_OVERLOAD_2): Likewise.
10988 (BU_CRYPTO_1): Likewise.
10989 (BU_CRYPTO_2): Likewise.
10990 (BU_CRYPTO_3): Likewise.
10991 (BU_CRYPTO_OVERLOAD_1): Likewise.
10992 (BU_CRYPTO_OVERLOAD_2): Likewise.
10993 (XSCVSPDP): Fix typo, point to the correct instruction.
10994 (VCIPHER): Add power8 crypto builtins.
10995 (VCIPHERLAST): Likewise.
10996 (VNCIPHER): Likewise.
10997 (VNCIPHERLAST): Likewise.
10998 (VPMSUMB): Likewise.
10999 (VPMSUMH): Likewise.
11000 (VPMSUMW): Likewise.
11001 (VPERMXOR_V2DI): Likewise.
11002 (VPERMXOR_V4SI: Likewise.
11003 (VPERMXOR_V8HI: Likewise.
11004 (VPERMXOR_V16QI: Likewise.
11005 (VSHASIGMAW): Likewise.
11006 (VSHASIGMAD): Likewise.
11007 (VPMSUM): Likewise.
11008 (VPERMXOR): Likewise.
11009 (VSHASIGMA): Likewise.
11011 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11012 __CRYPTO__ if the crypto instructions are available.
11013 (altivec_overloaded_builtins): Add support for overloaded power8
11016 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
11017 support for power8 crypto builtins.
11018 (builtin_function_type): Likewise.
11019 (altivec_init_builtins): Add support for builtins that take vector
11020 long long (V2DI) arguments.
11022 * config/rs6000/crypto.md: New file, define power8 crypto
11025 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11026 Pat Haugen <pthaugen@us.ibm.com>
11027 Peter Bergner <bergner@vnet.ibm.com>
11029 * doc/invoke.texi (Option Summary): Add power8 options.
11030 (RS/6000 and PowerPC Options): Likewise.
11032 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
11033 constraints.md instead of rs6000.h. Reorder w* constraints. Add
11034 wm, wn, wr documentation.
11036 * config/rs6000/constraints.md (wm): New constraint for VSX
11037 registers if direct move instructions are enabled.
11038 (wn): New constraint for no registers.
11039 (wq): New constraint for quad word even GPR registers.
11040 (wr): New constraint if 64-bit instructions are enabled.
11041 (wv): New constraint if power8 vector instructions are enabled.
11042 (wQ): New constraint for quad word memory locations.
11044 * config/rs6000/predicates.md (const_0_to_15_operand): New
11045 constraint for 0..15 for crypto instructions.
11046 (gpc_reg_operand): If VSX allow registers in VSX registers as well
11047 as GPR and floating point registers.
11048 (int_reg_operand): New predicate to match only GPR registers.
11049 (base_reg_operand): New predicate to match base registers.
11050 (quad_int_reg_operand): New predicate to match even GPR registers
11051 for quad memory operations.
11052 (vsx_reg_or_cint_operand): New predicate to allow vector logical
11053 operations in both GPR and VSX registers.
11054 (quad_memory_operand): New predicate for quad memory operations.
11055 (reg_or_indexed_operand): New predicate for direct move support.
11057 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
11058 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
11059 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
11060 (POWERPC_MASKS): Add power8 options.
11061 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
11064 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11065 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
11067 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
11068 (-mpower8-fusion): New power8 options.
11069 (-mpower8-fusion-sign): Likewise.
11070 (-mpower8-vector): Likewise.
11071 (-mcrypto): Likewise.
11072 (-mdirect-move): Likewise.
11073 (-mquad-memory): Likewise.
11075 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
11076 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
11078 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
11079 (rs6000_debug_vector_unit): Add p8_vector.
11080 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
11081 definitions. Also print fusion state.
11082 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
11083 (rs6000_builtin_mask_calculate): Add power8 builtin support.
11084 (rs6000_option_override_internal): Add support for power8.
11085 (rs6000_common_init_builtins): Add debugging for skipped builtins
11086 if -mdebug=builtin.
11087 (rs6000_adjust_cost): Add power8 support.
11088 (rs6000_issue_rate): Likewise.
11089 (insn_must_be_first_in_group): Likewise.
11090 (insn_must_be_last_in_group): Likewise.
11091 (force_new_group): Likewise.
11092 (rs6000_register_move_cost): Likewise.
11093 (rs6000_opt_masks): Likewise.
11095 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
11096 power8 capable assembler, default to power7 options.
11097 (TARGET_DIRECT_MOVE): Likewise.
11098 (TARGET_CRYPTO): Likewise.
11099 (TARGET_P8_VECTOR): Likewise.
11100 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
11101 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
11102 (VECTOR_MEM_P8_VECTOR_P): Likewise.
11103 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
11104 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
11105 (TARGET_XSCVDPSPN): Likewise.
11106 (TARGET_XSCVSPDPN): Likewsie.
11107 (TARGET_SYNC_HI_QI): Likewise.
11108 (TARGET_SYNC_TI): Likewise.
11109 (MASK_CRYPTO): Likewise.
11110 (MASK_DIRECT_MOVE): Likewise.
11111 (MASK_P8_FUSION): Likewise.
11112 (MASK_P8_VECTOR): Likewise.
11113 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
11114 temporary used by some of the direct move instructions to get two FP
11115 temporary registers does not force creation of a stack frame.
11116 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
11117 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
11118 that any VSX registers are tieable, even if they are also an
11119 Altivec vector mode.
11120 (r6000_reg_class_enum): Add wm, wr, wv constraints.
11121 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
11122 (RS6000_BTM_CRYPTO): Likewise.
11123 (RS6000_BTM_COMMON): Likewise.
11125 * config/rs6000/rs6000.md (cpu attribute): Add power8.
11126 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
11127 (enum rs6000_vector): Add power8 vector support.
11129 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11133 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
11134 (any_sibcall_could_use_r3): this and handle indirect calls.
11135 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
11137 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11139 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
11141 2013-05-22 Richard Biener <rguenther@suse.de>
11143 PR middle-end/57349
11144 * profile.c (branch_prob): Do not split blocks that are
11145 abnormally receiving from ECF_RETURNS_TWICE functions.
11147 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11149 * recog.c (offsettable_address_addr_space_p): Fix calculation of
11150 address mode. Move pointer mode initialization to the same place.
11152 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11154 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
11155 while it has any effect.
11157 2013-05-21 Easwaran Raman <eraman@google.com>
11159 PR tree-optimization/57322
11160 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
11161 UID of the statement added to the BB to be 1.
11163 2013-05-21 Jakub Jelinek <jakub@redhat.com>
11165 PR tree-optimization/57331
11166 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
11167 of conversion from pointer type to integral type with integer.
11169 2013-05-21 Martin Jambor <mjambor@suse.cz>
11172 * ipa-prop.c (ipa_read_node_info): Process param_used and
11173 controlled_uses in the same order as when writing.
11175 2013-05-21 Magnus Granberg <baldrick@free.fr>
11178 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
11180 2013-05-21 Richard Biener <rguenther@suse.de>
11182 PR tree-optimization/57318
11183 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
11184 estimate stmts with side-effects as likely eliminated.
11186 2013-05-21 Richard Biener <rguenther@suse.de>
11188 PR tree-optimization/57330
11189 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
11190 preserve the call stmts fntype.
11192 2013-05-21 Richard Biener <rguenther@suse.de>
11194 PR tree-optimization/57303
11195 * tree-ssa-sink.c (statement_sink_location): Improve killing
11196 stmt detection and properly handle self-assignments.
11198 2013-05-21 Christian Bruel <christian.bruel@st.com>
11200 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
11201 spanning registers. LEAF_REG_REMAP is supported only for contiguous
11202 registers. Set register size out of the PARALLEL loop.
11204 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
11207 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
11208 (*fmasf4, *fmasf4_media): New insns.
11210 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11212 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
11213 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
11214 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
11215 (mips_idiv_insns): Update the comments to say that the returned
11216 instruction counts are in units of BASE_INSN_LENGTH.
11217 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
11218 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
11219 using 2 rather than 4 as the length of indirect MIPS16 and
11220 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
11221 length of a NOP. Don't divide MIPS16 lengths by 2.
11222 (mips16_split_long_branches): Assume a branch is long if the
11223 length is greater than 4 rather than 8.
11224 * config/mips/mips.md (length): Give MIPS16 lengths directly,
11225 rather than multiplying them by 2. Multiply instruction counts
11226 by BASE_INSN_LENGTH rather than 4.
11227 (*jump_mips16, tls_get_tp_mips16_<mode>)
11228 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
11230 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11232 * config/mips/mips.md (extended_mips16): Remove branch case.
11233 (length): Remove duplicated extended_mips16 test.
11235 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11237 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
11239 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
11241 * recog.h (Recog_data): Rename to...
11242 (recog_data_d): ...this.
11243 (recog_data): Update accordingly.
11244 * recog.c (recog_data): Likewise.
11245 * reload.c (save_recog_data): Likewise.
11246 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
11247 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
11249 2013-05-17 Julian Brown <julian@codesourcery.com>
11251 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
11252 found in a REG_EQUAL note, invalidate it.
11254 2013-05-17 Easwaran Raman <eraman@google.com>
11256 * tree-ssa-reassoc.c (find_insert_point): New function.
11257 (insert_stmt_after): Likewise.
11258 (get_def_stmt): Likewise.
11259 (ensure_ops_are_available): Likewise.
11260 (not_dominated_by): Likewise.
11261 (rewrite_expr_tree): Do not move statements beyond what is
11262 necessary. Remove call to swap_ops_for_binary_stmt...
11263 (reassociate_bb): ... and move it here.
11264 (build_and_add_sum): Assign UIDs for new statements.
11265 (linearize_expr): Likewise.
11266 (do_reassoc): Renumber gimple statement UIDs.
11268 2013-05-17 Jan Hubicka <jh@suse.cz>
11270 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
11272 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
11273 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
11275 (output_weakrefs): Update.
11277 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
11278 Martin Jambor <mjambor@suse.cz>
11280 PR middle-end/57276
11281 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
11282 value that corresponds to the given aggval is found in values vector.
11284 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
11286 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
11287 sse, sse2, sse3, ssse3 and sse4a flags to options.
11289 2013-05-17 David Malcolm <dmalcolm@redhat.com>
11291 * gengtype-state.c: (s_expr_writer): New class, to handle
11292 prettifying of output layout of s-expressions.
11293 (state_writer): New class, to write out gtype.state.
11294 (state_written_type_count): Move this variable into member data of
11296 (s_expr_writer::s_expr_writer): New code: constructor for new class
11297 (state_writer::state_writer(): ditto
11298 (s_expr_writer::write_new_line): New function
11299 (s_expr_writer::write_any_indent): ditto
11300 (s_expr_writer::begin_s_expr): ditto
11301 (s_expr_writer::end_s_expr): ditto
11302 (write_state_fileloc): convert to method of state_writer...
11303 (state_writer:: write_state_fileloc): ...and use methods of
11304 s_expr_writer to write indentation into the gtype.state output file
11305 to visually represent the hierarchical structure of the list
11307 (write_state_fields): ditto, renaming to...
11308 (state_writer::write_state_fields)
11309 (write_state_a_string): ditto, renaming to...
11310 (state_writer::write_state_a_string)
11311 (write_state_string_option): ditto, renaming to...
11312 (state_writer::write_state_string_option)
11313 (write_state_type_option): ditto, renaming to...
11314 (state_writer::write_state_type_option)
11315 (write_state_nested_option): ditto, renaming to...
11316 (state_writer::write_state_nested_option)
11317 (write_state_option): ditto, renaming to...
11318 (state_writer::write_state_option)
11319 (write_state_options): ditto, renaming to...
11320 (state_writer::write_state_options)
11321 (write_state_lang_bitmap): ditto, renaming to...
11322 (state_writer::write_state_lang_bitmap)
11323 (write_state_version): ditto, renaming to...
11324 (state_writer::write_state_version)
11325 (write_state_scalar_type): ditto, renaming to...
11326 (state_writer::write_state_scalar_type)
11327 (write_state_string_type): ditto, renaming to...
11328 (state_writer::write_state_string_type)
11329 (write_state_undefined_type): ditto, renaming to...
11330 (state_writer::write_state_undefined_type)
11331 (write_state_struct_union_type): ditto, renaming to...
11332 (state_writer::write_state_struct_union_type)
11333 (write_state_struct_type): ditto, renaming to...
11334 (state_writer::write_state_struct_type)
11335 (write_state_user_struct_type): ditto, renaming to...
11336 (state_writer::write_state_user_struct_type)
11337 (write_state_lang_struct_type): ditto, renaming to...
11338 (state_writer::write_state_lang_struct_type)
11339 (write_state_param_struct_type): ditto, renaming to...
11340 (state_writer::write_state_param_struct_type)
11341 (write_state_pointer_type): ditto, renaming to...
11342 (state_writer::write_state_pointer_type)
11343 (write_state_array_type): ditto, renaming to...
11344 (state_writer::write_state_array_type)
11345 (write_state_gc_used): ditto, renaming to...
11346 (state_writer::write_state_gc_used)
11347 (write_state_common_type_content): ditto, renaming to...
11348 (state_writer::write_state_common_type_content)
11349 (write_state_type): ditto, renaming to...
11350 (state_writer::write_state_type)
11351 (write_state_pair_list): ditto, renaming to...
11352 (state_writer::write_state_pair_list)
11353 (write_state_pair): ditto, renaming to...
11354 (state_writer::write_state_pair)
11355 (write_state_typedefs): ditto, renaming to...
11356 (state_writer::write_state_typedefs)
11357 (write_state_structures): ditto, renaming to...
11358 (state_writer::write_state_structures)
11359 (write_state_param_structs): ditto, renaming to...
11360 (state_writer::write_state_param_structs)
11361 (write_state_variables): ditto, renaming to...
11362 (state_writer::write_state_variables)
11363 (write_state_srcdir): ditto, renaming to...
11364 (state_writer::write_state_srcdir)
11365 (write_state_files_list): ditto, renaming to...
11366 (state_writer::write_state_files_list)
11367 (write_state_languages): ditto, renaming to...
11368 (state_writer::write_state_languages)
11369 (write_state): create a state_writer instance and use it when
11370 writing out the state file
11372 2013-05-17 Mike Stump <mikestump@comcast.net>
11374 PR rtl-optimization/57304
11375 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
11376 accessing DF_REF_REAL_LOC.
11378 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11380 PR rtl-optimization/57281
11381 PR rtl-optimization/57300
11382 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
11383 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
11384 what the other splitter did if the registers are dead.
11386 2013-05-17 Richard Biener <rguenther@suse.de>
11388 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
11391 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11393 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
11396 2013-05-17 Marek Polacek <polacek@redhat.com>
11398 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
11399 length when doing non-zero store of storing '\0' to '\0'.
11401 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11403 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
11404 vect_external_def oprnd1 with loop_vinfo, try to emit
11405 optional cast, negation and and stmts on the loop preheader
11406 edge instead of into the pattern def seq.
11408 PR tree-optimization/57051
11409 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
11410 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
11412 2013-05-16 Nick Clifton <nickc@redhat.com>
11414 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
11415 (rl78_is_naked_func): New function.
11416 (rl78_expand_prologue): Skip prologue generation for naked functions.
11417 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
11418 * doc/extend.texi (naked): Add RL78 to the list of processors
11419 that supports this attribute.
11421 2013-05-16 Jeff Law <law@redhat.com>
11423 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
11425 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11427 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
11428 cache parameters using detect_caches_amd also for CYRIX,
11429 NSC and TM2 signatures.
11431 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11432 Dzianis Kahanovich <mahatma@eu.by>
11436 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11437 VIA/Centaur processors and determine their cache parameters
11438 using detect_caches_amd.
11440 2013-05-16 Teresa Johnson <tejohnson@google.com>
11442 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11443 (rtl_verify_edges): New function.
11444 (rtl_verify_bb_insns): Ditto.
11445 (rtl_verify_bb_pointers): Ditto.
11446 (rtl_verify_bb_insn_chain): Ditto.
11447 (rtl_verify_fallthru): Ditto.
11448 (rtl_verify_bb_layout): Ditto.
11449 (rtl_verify_flow_info_1): Outline checks into new functions.
11450 (rtl_verify_flow_info): Ditto.
11452 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
11454 * cfghooks.c (copy_bbs): Add update_dominance argument.
11455 * cfghooks.h (copy_bbs): Update prototype.
11456 * tree-cfg.c (gimple_duplicate_sese_region):
11457 Add update_dominance argument.
11458 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11459 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11460 gimple_duplicate_sese_region call.
11461 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11462 Update copy_bbs call.
11463 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11464 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11466 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11468 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11469 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11470 vect_recog_rotate_pattern.
11471 (vect_recog_rotate_pattern): New function.
11473 2013-05-16 Jason Merrill <jason@redhat.com>
11475 * Makefile.in (LLINKER): New variable.
11476 (mostlyclean): Remove link mutex.
11477 * configure.ac: Handle --enable-link-mutex.
11478 * lock-and-run.sh: New script.
11480 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11483 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11486 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11488 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11490 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11492 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11493 * config/arm/arm.c (next_consecutive_mem): New function.
11494 (gen_movmem_ldrd_strd): Likewise.
11495 * config/arm/arm.md (movmemqi): Update condition and code.
11496 (unaligned_loaddi, unaligned_storedi): New patterns.
11498 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11500 * config.gcc: Obsolete *-*-solaris2.9*.
11501 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11503 2013-05-16 Richard Biener <rguenther@suse.de>
11505 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11506 earlier, after GRAPHITE transforms and IV canonicalization.
11508 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11510 * omp-low.c (extract_omp_for_data): For collapsed loops,
11511 if at least one of the loops is known at compile time to
11512 iterate zero times, set count to 0.
11513 (expand_omp_regimplify_p): New function.
11514 (expand_omp_for_generic): For collapsed loops, if at least
11515 one of the loops isn't known to iterate at least once,
11516 add runtime check with setting count to 0.
11517 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11518 For unsigned types if it isn't known at compile time that
11519 the loop will iterate at least once, add runtime check to bypass
11520 the whole loop if initial condition isn't true.
11522 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11524 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11526 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
11528 PR middle-end/57286
11529 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
11530 transformations to avoid an infinite loop.
11532 2013-05-16 Marek Polacek <polacek@redhat.com>
11534 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
11536 2013-05-15 Leif Ekblad <leif@rdos.net>
11538 * config/i386/i386.c (ix86_decompose_address): Use
11539 DEFAULT_TLS_SEG_REG to access TLS segment register.
11540 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
11541 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
11542 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
11544 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
11547 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
11548 sibling calls to functions that would normally be lazily bound,
11549 unless $gp is call-clobbered.
11551 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
11553 * config/i386/i386.c (ix86_option_override_internal): Update
11554 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
11555 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
11556 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
11557 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
11559 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
11561 2013-05-15 Andreas Schwab <schwab@suse.de>
11563 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
11564 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
11567 2013-05-15 Teresa Johnson <tejohnson@google.com>
11569 * loop-unroll.c (report_unroll_peel): Check decision before
11570 emitting unroll/peel message.
11572 2013-05-15 Teresa Johnson <tejohnson@google.com>
11574 * function.h (has_bb_partition): New rtl_data flag.
11575 (bb_reorder_complete): Ditto.
11576 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
11577 instead of flag_reorder_blocks_and_partition.
11578 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
11579 with some enhancements.
11580 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
11581 * bb-reorder.c (connect_traces): Check for has_bb_partition
11582 instead of flag_reorder_blocks_and_partition.
11583 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
11584 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
11585 verify_hot_cold_block_grouping.
11586 (partition_hot_cold_basic_blocks): Set has_bb_partition.
11588 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11591 * config/arm/predicates.md (call_insn_operand): New predicate.
11592 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
11593 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
11594 if insn is not a tail call.
11595 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11597 * config/arm/arm.h (enum reg_class): New caller save register class.
11598 (REG_CLASS_NAMES): Likewise.
11599 (REG_CLASS_CONTENTS): Likewise.
11600 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11603 2013-05-15 Richard Biener <rguenther@suse.de>
11605 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11606 of MSG_OPTIMIZED_LOCATIONS.
11607 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11608 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11610 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11611 of MSG_OPTIMIZED_LOCATIONS.
11612 (execute_vect_slp): Likewise.
11613 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11614 (vect_create_cond_for_alias_checks): Likewise.
11615 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11616 (vect_recog_widen_mult_pattern): Likewise.
11617 (vect_recog_widen_sum_pattern): Likewise.
11618 (vect_recog_over_widening_pattern): Likewise.
11619 (vect_recog_widen_shift_pattern): Likewise.
11620 (vect_recog_vector_vector_shift_pattern): Likewise.
11621 (vect_recog_divmod_pattern): Likewise.
11622 (vect_recog_mixed_size_cond_pattern): Likewise.
11623 (vect_recog_bool_pattern): Likewise.
11624 (vect_pattern_recog_1): Likewise.
11626 2013-05-15 Martin Jambor <mjambor@suse.cz>
11628 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11629 non-functions to builtin_unreachable.
11630 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11631 correct when new direct edges were discovered.
11633 2013-05-15 Martin Jambor <mjambor@suse.cz>
11635 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11636 header, print symbol order instead of node uid, print more information
11637 about indirect edge targets.
11638 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11640 (ipa_make_edge_direct_to_target): Likewise.
11641 (remove_described_reference): Likewise.
11642 (propagate_controlled_uses): Likewise.
11643 (ipa_print_node_params): Also print symbol order.
11644 (ipcp_transform_function): Print symbol order instead of node uids.
11645 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11646 (cgraph_get_create_real_symbol_node): Likewise.
11647 * ipa-cp.c (print_lattice): Likewise.
11648 (print_all_lattices): Likewise.
11649 (determine_versionability): Likewise.
11650 (initialize_node_lattices): Likewise.
11651 (estimate_local_effects): Likewise.
11652 (update_profiling_info): Likewise.
11653 (create_specialized_node): Likewise.
11654 (perhaps_add_new_callers): Likewise.
11655 (decide_about_value): Likewise.
11656 (decide_whether_version_node): Likewise.
11657 (identify_dead_nodes): Likewise.
11658 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11659 (dump_inline_summary): Likewise.
11660 (estimate_node_size_and_time): Likewise.
11661 (inline_analyze_function): Likewise.
11662 * ipa-inline.c (report_inline_failed_reason): Likewise.
11663 (want_early_inline_function_p): Likewise.
11664 (edge_badness): Likewise.
11665 (update_edge_key): Likewise.
11666 (inline_small_functions): Likewise. Add dumping of order to two other
11668 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11669 instead of node uids.
11670 (propagate_pure_const): Likewise.
11671 (propagate_pure_const): Likewise.
11672 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11673 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11674 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11676 * tree-pretty-print.c (dump_function_header): Likewise.
11677 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11678 Print symbol order instead of node uids.
11680 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11682 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11683 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11685 2013-05-15 Richard Biener <rguenther@suse.de>
11687 PR tree-optimization/57275
11688 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11689 return value for fail to do runtime alias checks for gather loads.
11691 2013-05-15 Jan Hubicka <jh@suse.cz>
11695 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11696 weakrefs are not external.
11697 (lto_symtab_merge_decls): Fix thinko when dealing with
11698 non-lto_symtab decls.
11699 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11700 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11701 * varpool.c (dump_varpool_node): Dump more flags.
11703 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11705 * config/i386/i386.c (processor_alias_table): Add instruction
11706 FSGSBASE for AMD bdver3 architecture.
11708 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11710 * tree.c (warn_deprecated_use): Print file:line using locus color.
11711 * diagnostic.c (diagnostic_report_current_module): Print file:line
11712 and file:line:column using locus color.
11714 2013-05-14 Mike Stump <mikestump@comcast.net>
11716 * gdbinit.in: Add __null.
11718 2013-05-14 Mike Stump <mikestump@comcast.net>
11720 * recog.h: Rename struct recog_data to Recog_data.
11721 * recog.c: Likewise.
11722 * reload.c (can_reload_into): Likewise.
11723 * config/picochip/picochip.c: Likewise.
11725 2013-05-14 Mike Stump <mikestump@comcast.net>
11727 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11729 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11731 * resource.h (struct resources): Remove unch_memory member.
11732 (CLEAR_RESOURCE): Don't clear unch_memory.
11733 * resource.c (mark_referenced_resources): Don't set it.
11734 (mark_set_resources): Likewise.
11735 (mark_target_live_regs): Don't clear it.
11736 (init_resource_info): Likewise.
11737 * reorg.c (resource_conflicts_p): Don't compare it.
11738 (redundant_insn): Don't set it.
11740 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11742 * emit-rtl.c (next_label): Remove unused function.
11743 (skip_consecutive_labels, link_cc0_insns): Move to ...
11744 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11745 only place where these functions are used, and make them static.
11747 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11749 * fold-const.c (fold_negate_expr): Handle vectors.
11750 (fold_truth_not_expr): Make it static.
11751 (fold_invert_truthvalue): New static function.
11752 (invert_truthvalue_loc): Handle vectors. Do not call
11753 fold_truth_not_expr directly.
11754 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11755 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11756 (fold_comparison): Handle vector constants.
11757 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11758 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11759 * tree.h (fold_truth_not_expr): Remove declaration.
11761 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11763 * config/aarch64/aarch64-simd.md
11764 (aarch64_vcond_internal<mode>): Rename to...
11765 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11766 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11767 float modes. Clarify all iterator modes.
11768 (vcond<mode><mode>): Use new name for vcond expanders.
11769 (vcond<v_cmp_result><mode>): Likewise.
11770 (vcondu<mode><mode>: Likewise.
11771 * config/aarch64/iterators.md (VDQF_COND): New.
11773 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11776 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11777 variable for the shift amount. Check that we shift by non-negative
11780 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11783 * config/arm/arm.h (EPILOGUE_USES): Only return true
11784 for LR_REGNUM after epilogue_completed.
11786 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
11788 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
11789 is error_mark_node.
11791 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11794 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
11795 and Solaris 11+/x86 with gld.
11796 * configure: Regenerate.
11798 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11800 * expmed.c (expand_shift_1): Canonicalize rotates by
11801 constant bitsize / 2 to bitsize - 1.
11802 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
11803 case ROTATERT>: Likewise.
11806 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11808 * config/i386/i386.md (rotateinv): New code attr.
11809 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11810 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11811 roll $31, %eax, etc.
11813 2013-05-14 Richard Biener <rguenther@suse.de>
11815 PR middle-end/57235
11816 * tree-eh.c (sink_clobbers): Give up for successors with
11817 multiple predecessors and no virtual uses.
11819 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
11821 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
11822 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
11824 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11826 PR middle-end/57251
11827 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
11828 the case when both op0 and op1 have VOIDmode.
11830 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11832 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
11833 in multiply-accumulate mode.
11835 2013-05-13 Guozhi Wei <carrot@google.com>
11837 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
11839 2013-05-13 Kai Tietz <ktietz@redhat.com>
11842 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
11843 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
11844 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
11845 * config/i386/i386.c (ix86_option_override_internal): Likewise.
11846 (ix86_expand_prologue): Likewise.
11847 (ix86_expand_split_stack_prologue): Likewise.
11848 (legitimate_pic_address_disp_p): Likewise.
11849 (legitimize_pic_address): Likewise.
11850 (legitimize_tls_address): Likewise.
11851 (legitimize_pe_coff_symbol): Likewise.
11852 (output_pic_addr_const): Likewise.
11853 (construct_plt_address): Likewise.
11854 (ix86_expand_call): Likewise.
11855 (x86_output_mi_thunk): Likewise.
11856 (x86_function_profiler): Likewise.
11858 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
11860 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
11861 similar switch cases.
11862 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
11863 (aarch64_simd_mov_to_<mode>low): Delete.
11864 (aarch64_simd_mov_to_<mode>high): Delete.
11865 (move_lo_quad_<mode>): Add w<-r alternative.
11866 (aarch64_simd_move_hi_quad_<mode>): Likewise.
11867 (aarch64_simd_mov_from_*): Update type attribute.
11868 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
11871 2013-05-13 Jan Hubicka <jh@suse.cz>
11873 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
11874 * config/i386/i386.c (ix86_compute_frame_layout,
11875 ix86_expand_epilogue, emit_i387_cw_initialization,
11876 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
11877 ix86_local_alignment): Fix use of size/speed predicates.
11879 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11881 PR tree-optimization/45216
11882 PR tree-optimization/57157
11883 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
11884 the (-Y) & (B - 1) variant if OP is |.
11885 * expmed.c (expand_shift_1): For rotations by const0_rtx just
11886 return shifted. Use (-op1) & (prec - 1) as other_amount
11887 instead of prec - op1.
11889 2013-05-13 Martin Jambor <mjambor@suse.cz>
11891 PR middle-end/42371
11892 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
11893 (ipa_constant_data): New type.
11894 (ipa_jump_func): Use ipa_constant_data to hold information about
11895 constant jump functions.
11896 (ipa_get_jf_constant): Adjust to jump function type changes.
11897 (ipa_get_jf_constant_rdesc): New function.
11898 (ipa_param_descriptor): New field controlled_uses.
11899 (ipa_get_controlled_uses): New function.
11900 (ipa_set_controlled_uses): Likewise.
11901 * ipa-ref.h (ipa_find_reference): Declare.
11902 * ipa-prop.c (ipa_cst_ref_desc): New type.
11903 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
11905 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
11906 New parameter cs. Adjust all callers.
11907 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
11908 (remove_described_reference): New function.
11909 (jfunc_rdesc_usable): Likewise.
11910 (try_make_edge_direct_simple_call): Decrement controlled use count,
11911 attempt to remove reference if it hits zero.
11912 (combine_controlled_uses_counters): New function.
11913 (propagate_controlled_uses): Likewise.
11914 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
11915 (ipa_edge_duplication_hook): Duplicate reference descriptions.
11916 (ipa_print_node_params): Print described use counter.
11917 (ipa_write_jump_function): Adjust to jump function type changes.
11918 (ipa_read_jump_function): New parameter CS, pass it to
11919 ipa_set_jf_constant. Adjust caller.
11920 (ipa_write_node_info): Stream controlled use count
11921 (ipa_read_node_info): Likewise.
11922 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
11924 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
11925 count. Remove cloning-added reference if it reaches zero.
11926 * ipa-ref.c (ipa_find_reference): New function.
11928 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11930 * config/i386/i386.c (processor_target_table): Modified default
11931 alignment values for AMD BD and BT architectures.
11933 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
11935 * tree-vect-generic.c (uniform_vector_p): Move ...
11936 * tree.c (uniform_vector_p): ... here.
11937 * tree.h (uniform_vector_p): Declare it.
11938 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
11941 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11943 PR tree-optimization/57230
11944 * tree-ssa-strlen.c (handle_char_store): Record length for
11945 array store from STRING_CST.
11947 PR tree-optimization/57230
11948 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
11951 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
11953 * config/epiphany/epiphany.c (epiphany_init): Check size of
11954 NUM_MODES_FOR_MODE_SWITCHING.
11955 (epiphany_expand_prologue):
11956 Remove CONFIG_REGNUM initial value handling code.
11957 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
11958 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
11959 (emit_set_fp_mode, epiphany_mode_after): Likewise.
11960 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
11961 Don't return 1 for FP_MODE_NONE.
11962 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
11963 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
11964 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
11965 * config/epiphany/epiphany.md (save_config): New pattern.
11967 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
11969 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
11971 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
11973 * config/i386/i386.md (memory): Handle sseishft1.
11974 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
11975 (*vec_extractv2di_1): Ditto.
11977 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
11979 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
11982 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
11984 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
11985 Add mthumb/march=armv7-a multilib.
11986 Add mthumb/march=armv7-r multilib.
11987 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
11989 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
11991 * config/v850/t-rtems: Add more multilibs.
11993 2013-05-10 Richard Biener <rguenther@suse.de>
11995 PR tree-optimization/57214
11996 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
11997 not propagate from SSA names that occur in abnormal PHI nodes.
11999 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
12001 * stor-layout.c (element_precision): New function.
12002 * machmode.h (element_precision): Declare it.
12003 * tree.c (build_minus_one_cst): New function.
12004 (element_precision): Likewise.
12005 * tree.h (build_minus_one_cst): Declare new function.
12006 (element_precision): Likewise.
12007 * fold-const.c (operand_equal_p): Use element_precision.
12008 (fold_binary_loc): Handle vector types.
12009 * convert.c (convert_to_integer): Use element_precision.
12010 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
12013 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
12015 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
12016 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
12017 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
12018 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12019 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
12020 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
12021 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
12022 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12023 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
12024 (Uuw8): New constraints.
12025 (Usb4): Move into alphabetical order.
12026 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
12027 (sd8_operand, ub8_operand, uw8_operand): New predicates.
12028 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
12029 previously unnamed patterns.
12030 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
12031 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
12032 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
12033 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
12034 of set_attr_alternative/if_then_else. Use extended_mips16 instead
12035 of specific lengths.
12037 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12039 * config/i386/i386.md (rotateinv): New code attr.
12040 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12041 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12042 roll $31, %eax, etc.
12044 PR tree-optimization/45216
12045 PR tree-optimization/57157
12046 * tree-ssa-forwprop.c (simplify_rotate): New function.
12047 (ssa_forward_propagate_and_combine): Call it.
12049 2013-05-10 Richard Biener <rguenther@suse.de>
12051 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
12052 disable peeling when we version for aliasing.
12053 (vector_alignment_reachable_p): Honor explicit user alignment.
12054 (vect_supportable_dr_alignment): Likewise.
12055 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
12056 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
12057 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
12058 then peeling to arrange for the cost-model check to come first.
12060 2013-05-10 Alan Modra <amodra@gmail.com>
12062 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
12063 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
12064 * configure: Regenerate.
12066 2013-05-10 Alan Modra <amodra@gmail.com>
12069 * varasm.c (default_elf_select_section): Move !DECL_P check..
12070 (get_named_section): ..to here before calling get_section_name.
12072 (default_section_type_flags): Add DECL_P check.
12073 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
12074 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
12076 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
12078 * config/epiphany/epiphany.c (epiphany_expand_prologue):
12079 When using gen_stack_adjust_str with a register offset, add a
12080 REG_FRAME_RELATED_EXPR note.
12082 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
12084 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
12085 (*vec_extractv4si_zext_mem): Ditto.
12086 (*vec_extractv2di): Add 0->x and x->x alternatives.
12087 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
12088 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
12090 2013-05-09 Jason Merrill <jason@redhat.com>
12092 N3639 C++1y VLA support
12093 * gimplify.c (gimplify_vla_decl): Don't touch an existing
12096 * tree.c (build_constructor_va): New.
12097 * tree.h: Declare it.
12099 2013-05-09 Martin Jambor <mjambor@suse.cz>
12102 * gimple-fold.c (canonicalize_constructor_val): Call
12103 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
12105 2013-05-09 Jan Hubicka <jh@suse.cz>
12106 Richard Biener <rguenther@suse.de>
12109 * symtab.c (symtab_make_decl_local): Do not add private names.
12111 2013-05-09 Jan Hubicka <jh@suse.cz>
12114 * symtab.c (insert_to_assembler_name_hash): Handle clones.
12115 (unlink_from_assembler_name_hash): Likewise.
12116 (symtab_prevail_in_asm_name_hash, symtab_register_node,
12117 symtab_unregister_node, symtab_initialize_asm_name_hash,
12118 change_decl_assembler_name): Update.
12120 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
12122 * config/aarch64/aarch64.md: New movtf split.
12123 (*movtf_aarch64): Update.
12124 (aarch64_movdi_tilow): Handle TF modes and rename to
12125 aarch64_movdi_<mode>low.
12126 (aarch64_movdi_tihigh): Handle TF modes and rename to
12127 aarch64_movdi_<mode>high
12128 (aarch64_movtihigh_di): Handle TF modes and rename to
12129 aarch64_mov<mode>high_di
12130 (aarch64_movtilow_di): Handle TF modes and rename to
12131 aarch64_mov<mode>low_di
12132 (aarch64_movtilow_tilow): Remove spurious whitespace.
12133 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
12135 (aarch64_print_operand): Update.
12137 2013-05-09 Alan Modra <amodra@gmail.com>
12139 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
12141 * configure: Regenerate.
12143 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12145 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
12146 splitter preparation statements.
12147 * config/i386/sse.md (*vec_extract* splitters): Ditto.
12148 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
12151 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12153 * gimple-ssa-strength-reduction.c (count_candidates): Change
12154 return value to int.
12155 (analyze_candidates_and_replace): Change type of length to int.
12157 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12159 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
12160 (*vec_extract<mode>): Use VI12_128 mode iterator.
12161 (*vec_extract<mode>_mem): Ditto.
12162 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
12165 2013-05-08 Diego Novillo <dnovillo@google.com>
12170 2012-08-17 Diego Novillo <dnovillo@google.com>
12173 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
12174 * config.in: Regenerate.
12175 * configure: Regenerate.
12176 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
12178 2013-05-08 Jan Hubicka <jh@suse.cz>
12181 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
12182 * cgraph.h (symtab_node_base): Add unique_name.
12183 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12184 input_overwrite_node, input_varpool_node): Stream unique_name.
12185 * cgraphclones.c (cgraph_create_virtual_clone,
12186 cgraph_function_versioning): Set unique_name.
12187 * ipa.c (function_and_variable_visibility): Set unique_name.
12189 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12191 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
12192 (alloc_cand_and_find_basis): Restrict conditional candidate
12193 processing to CAND_MULTs.
12195 2013-05-08 Jan Hubicka <jh@suse.cz>
12198 lto-symtab.c (lto_symtab_symbol_p): New function.
12199 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
12200 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
12201 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
12202 Skip static symbols.
12204 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
12206 PR tree-optimization/57200
12207 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
12208 Only call inform if the preceding warning_at returns true.
12210 2013-05-07 Han Shen <shenhan@google.com>
12212 * cfgexpand.c (record_or_union_type_has_array_p): New function.
12213 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
12214 * common.opt (fstack-protector-strong): New option.
12215 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
12216 * doc/invoke.texi (Optimization Options): Document
12217 "-fstack-protector-strong".
12218 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
12220 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12222 * config/mips/mips.c (mips_machine_reorg2): Return 0.
12224 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
12226 * ira.c (update_equiv_regs): Add insn having equiv memory even if
12227 it is not lhs of the insn.
12228 (setup_reg_equiv): Remove insn having equiv memory which it is not
12230 * lra-constraints.c (process_address): Try to improve generation
12231 code for address base + disp.
12232 (lra_constraints): Make correct the code for checking insn setting
12233 up backward equivalence. Remove insn only if it is in the init
12235 * lra-eliminations.c (update_reg_eliminate): Change return value.
12236 (lra_eliminate): Use the result.
12238 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
12240 * config/i386/sse.md (ssescalarnummask): New mode attribute.
12241 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
12242 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
12243 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
12244 register target operands.
12245 (*vec_extractv8hi_sse2): New pattern.
12246 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
12247 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
12248 (*vec_extract<mode>_mem): New insn and split pattern.
12250 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
12252 * config/arm/arm.c (arm_asan_shadow_offset): New function.
12253 (TARGET_ASAN_SHADOW_OFFSET): Define.
12254 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
12255 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
12257 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12259 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
12260 (incr_vec_index): Return -1 if increment not found.
12261 (create_add_on_incoming_edge): Assert if increment not found.
12262 (record_increment): Limit number of increments recorded.
12263 (all_phi_incrs_profitable): Return false if an increment not found.
12264 (replace_profitable_candidates): Don't process increments that were
12266 (analyze_candidates_and_replace): Limit size of incr_vec.
12268 2013-05-07 Richard Biener <rguenther@suse.de>
12270 * calls.c (special_function_p): setjmp-like functions are leaf.
12271 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
12272 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
12274 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
12276 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
12277 (aarch64_simd_mov<mode>): New expander.
12278 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
12279 (aarch64_simd_mov_to_<mode>high): Likewise.
12280 (aarch64_simd_mov_from_<mode>low): Likewise.
12281 (aarch64_simd_mov_from_<mode>high): Likewise.
12282 (aarch64_dup_lane<mode>): Update.
12283 (aarch64_dup_lanedi): New instruction pattern.
12284 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
12285 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
12287 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12289 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
12290 (replace_mult_candidate): Remove unnecessary argument; remove
12291 unnecessary parameter from call to introduce_cast_before_cand.
12292 (replace_unconditional_candidate): Remove unnecessary parameter
12293 from call to replace_mult_candidate.
12294 (replace_conditional_candidate): Likewise.
12295 (insert_initializers): Use make_temp_ssa_name.
12296 (introduce_cast_before_cand): Remove unnecessary argument; use
12297 make_temp_ssa_name.
12298 (replace_one_candidate): Remove unnecessary argument; remove
12299 unnecessary parameter from calls to introduce_cast_before_cand.
12300 (replace_profitable_candidates): Remove unnecessary parameters
12301 from calls to replace_one_candidate.
12303 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12305 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
12306 phi def as possibly hiding a basis for a CAND_ADD whose operands
12307 have been commuted in the analysis.
12308 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
12310 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12312 * config/aarch64/aarch64.md
12313 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
12314 shift value between 0-4.
12316 2013-05-07 Richard Biener <rguenther@suse.de>
12318 * double-int.h (rshift): New overload.
12319 * double-int.c (rshift): New function.
12320 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
12321 (create_reference_ops_from_ref): Remove.
12322 (vn_reference_insert): Use shared ops for constructing the
12323 reference and copy it.
12325 2013-05-07 Richard Biener <rguenther@suse.de>
12327 PR middle-end/57190
12328 * tree-eh.c (sink_clobbers): Properly propagate
12329 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
12331 2013-05-07 Jakub Jelinek <jakub@redhat.com>
12333 PR tree-optimization/57149
12334 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
12335 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
12336 collect_phi_def_edges, execute_late_warn_uninitialized): Use
12337 uninit_undefined_value_p instead of ssa_undefined_value_p.
12340 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
12341 for modifier == EXPAND_INITIALIZER.
12343 2013-05-07 Anton Blanchard <anton@samba.org>
12345 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
12346 for powerpc64 little endian.
12347 * configure: Regenerate.
12349 2013-05-06 Graham Stott <grahams@btinternet.com>
12351 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
12352 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
12353 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
12354 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
12356 2013-05-06 Graham Stott <grahams@btinternet.com>
12358 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
12359 codes which allow non-lvalues.
12361 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
12363 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
12364 components are all 1s.
12365 (integer_minus_onep): New function.
12366 * tree.h (integer_minus_onep): Declare it.
12367 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
12368 integer_minus_onep instead of integer_all_onesp.
12370 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12373 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
12374 variations of these patterns.
12376 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12378 * config/i386/i386.md (isa): Add x64_sse4 member.
12379 (enabled): Handle x64_sse4.
12380 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
12381 instruction for 64bit SSE4_1 targets. Update insn attributes.
12382 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
12383 instruction for SSE4_1 targets. Update insn attributes.
12384 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12385 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
12386 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
12388 (*vec_extractv4si): Rename from *sse4_1_pextrd.
12389 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
12390 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
12392 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12395 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
12397 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
12399 * final.c (do_assembler_dialects): Don't handle curly braces and
12400 vertical bar escaped by % as dialect delimiters.
12401 (output_asm_insn): Print curly braces and vertical bar if escaped
12402 by % and ASSEMBLER_DIALECT defined.
12403 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
12404 * doc/tm.texi: Regenerated.
12406 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12408 * config/mips/mips.c: Include tree-pass.h.
12409 (mips_reorg): Split in pre- and post-dbr_schedule parts.
12410 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
12411 (pass_mips_machine_reorg2): New machine specific pass.
12412 (insert_pass_mips_machine_reorg2): New pass plugin definition.
12413 (mips_option_override): Register the new pass.
12414 * rtl.h (cleanup_barriers): Remove prototype.
12415 (dbr_schedule): Likewise.
12416 * jump.c (cleanup_barriers): Make static.
12417 * reorg.c (dbr_schedule): Likewise.
12419 2013-05-06 Richard Biener <rguenther@suse.de>
12421 PR tree-optimization/57185
12422 * tree-parloops.c (add_field_for_reduction): Handle anonymous
12423 SSA names properly.
12425 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12428 * config/i386/i386.c (add_parameter_dependencies): Add dependence
12429 between "first_arg" and "insn", not "last" and "insn".
12431 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12433 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
12434 (find_candidates_in_block): Re-enable slsr_process_phi.
12435 (create_phi_basis): Fix double counting of candidate adjustment.
12437 2013-05-06 Richard Biener <rguenther@suse.de>
12439 PR middle-end/57147
12440 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12441 the edge is also fallthru, preserve it and just clear the
12443 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12444 also complex, preserve that and just clear the fallthru flag.
12445 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12446 update virtual operands.
12448 2013-05-06 Alan Modra <amodra@gmail.com>
12450 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12451 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12452 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12453 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12454 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12455 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12457 2013-05-06 Alan Modra <amodra@gmail.com>
12459 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12460 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12461 (DEFAULT_ASM_ENDIAN): Define.
12462 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12463 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12464 Update -K PIC clause from sysv4.h.
12465 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12466 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12468 2013-05-06 Alan Modra <amodra@gmail.com>
12470 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12471 twice for little-endian.
12472 (ashrdi3_no_power, ashrdi3): Support little-endian.
12474 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12477 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12478 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12480 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12481 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12482 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12485 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12486 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12488 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12489 * config/pa/pa.opt: Make mbig-switch a no-op.
12490 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12491 (CASE_VECTOR_MODE): Always return SImode.
12492 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12493 for the !TARGET_BIG_SWITCH case.
12494 * config/pa/pa-linux.h: Likewise.
12495 * config/pa/pa-openbsd.h: Likewise.
12496 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12497 * config/pa/pa.md (short_jump): Remove define_insn.
12498 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12499 (casesi0): Remove define_insn.
12500 (type): Remove btable_branch.
12501 (pa_combine_type): Likewise.
12502 (in_nullified_branch_delay): Likewise.
12503 (in_call_delay): Likewise.
12504 (define_delay): Likewise.
12505 (define_insn_reservation "Z3"): Likewise.
12506 (define_insn_reservation "Z4"): Likewise.
12507 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12508 (pa_adjust_insn_length): Remove adjustment for btable branches.
12509 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12512 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
12514 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12515 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12516 Add m->r,x alternatives.
12517 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12518 splitters using SWI48x mode iterator.
12519 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12520 TARGET_64BIT. Add m->x alternative.
12521 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12522 Add o->x alternative. Enable for TARGET_SSE.
12523 (sse_storeq): Remove expander.
12524 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12525 with memory input operand.
12526 (*vec_extractv2di_1 splitter): New.
12527 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12528 * config/i386/i386.md (ssevecmodelower): New mode attribute.
12530 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12532 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
12533 (INT_LOWPART): Delete.
12534 (extract_MB): Adjust.
12535 (extract_ME): Adjust.
12536 (print_operand): Adjust.
12538 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12540 * config/rs6000/predicates.md (reg_or_add_cint_operand,
12541 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12542 (reg_or_logical_cint_operand, easy_fp_constant,
12543 logical_const_operand): Delete "CONST_DOUBLE" case.
12544 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
12545 "HOST_BITS_PER_WIDE_INT == 64" test.
12546 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
12547 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
12548 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
12549 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
12551 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
12552 CONST_DOUBLE DImode/VOIDmode case.
12553 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
12554 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12555 CONST_DOUBLE VOIDmode case.
12556 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12557 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
12558 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
12559 Delete CONST_DOUBLE case.
12560 (splitters for mov FMOVE64 const_double): Delete
12561 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12562 "HOST_BITS_PER_WIDE_INT >= 64" test.
12563 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
12565 (mov DI const_double): Delete.
12567 2013-05-04 Jakub Jelinek <jakub@redhat.com>
12569 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
12570 on op shows all bits zero in mode of a lowpart subreg, return zero.
12572 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
12575 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
12576 to save TFmode registers and DImode to save TImode registers for
12577 caller save operations.
12578 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
12579 mark being partially clobbered since they only use the first
12582 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
12583 and TDmode only use the upper 64-bits of each VSX register.
12585 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12587 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
12588 (find_candidates_in_block): Disable slsr_process_phi.
12590 2013-05-03 Guozhi Wei <carrot@google.com>
12592 * coverage.c (coverage_obj_init): Move the construction of gcov
12594 (build_init_ctor): ... here.
12596 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12598 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12599 (slsr_cand_d): Redefine def_phi.
12600 (stride_status, phi_adjust_status, count_phis_status): New enums.
12601 (find_phi_def): New.
12602 (find_basis_for_base_expr): New.
12603 (find_basis_for_candidate): Handle hidden bases.
12604 (alloc_cand_and_find_basis): Handle phi candidates.
12605 (slsr_process_phi): New.
12606 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12607 (create_mul_imm_cand): Likewise.
12608 (create_add_ssa_cand): Exclude phi base candidates.
12609 (create_add_imm_cand): Likewise.
12610 (slsr_process_cast): Likewise.
12611 (slsr_process_copy): Likewise.
12612 (find_candidates_in_block): Handle phi candidates.
12613 (dump_candidate): Likewise.
12614 (unconditional_cands): Delete.
12615 (unconditional_cands_with_known_stride_p): Delete.
12616 (phi_dependent_cand_p): New.
12617 (cand_increment): Handle phi-dependent candidates.
12618 (replace_dependent): Delete.
12619 (replace_mult_candidate): New.
12620 (replace_unconditional_candidate): New.
12621 (incr_vec_index): Move to avoid forward reference.
12622 (create_add_on_incoming_edge): New.
12623 (create_phi_basis): New.
12624 (replace_dependents): Delete.
12625 (replace_conditional_candidate): New.
12626 (phi_add_costs): New.
12627 (replace_uncond_cands_and_profitable_phis): New.
12628 (record_increment): Handle phi adjustments.
12629 (record_phi_increments): New.
12630 (record_increments): Handle phi adjustments.
12631 (phi_incr_cost): New.
12632 (lowest_cost_path): Handle phis.
12633 (total_savings): Likewise.
12634 (analyze_increments): Likewise.
12635 (ncd_with_phi): New.
12636 (ncd_of_cand_and_phis): New.
12637 (nearest_common_dominator_for_cands): Handle phi increments.
12638 (all_phi_incrs_profitable): New.
12639 (replace_profitable_candidates): Handle phi-dependent candidates.
12640 (analyze_candidates_and_replace): Likewise.
12642 2013-05-03 Teresa Johnson <tejohnson@google.com>
12645 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12646 do not exceed REG_BR_PROB_BASE.
12648 2013-05-03 Jeff Law <law@redhat.com>
12650 PR tree-optimization/57144
12651 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12652 operand of the condition will bit into the new type when eliminating
12653 a cast feeding a condition.
12655 2013-05-03 Jakub Jelinek <jakub@redhat.com>
12657 PR rtl-optimization/57130
12658 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12659 of COMPARE as in_code to the recursive call if needed.
12661 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
12663 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12664 (enabled): Handle new members.
12665 * config/i386/sse.md (*vec_concatv2si): Merge from
12666 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12667 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12669 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
12671 PR tree-optimization/57027
12672 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12673 for fnms opportunity, check we got the prerequisite kind
12674 of tree / gimple before using accessor functions.
12676 2013-05-03 Richard Biener <rguenther@suse.de>
12678 * double-int.h (lshift): New overload without precision
12679 and arith argument.
12680 (operator *=, operator +=, operator -=): Move ...
12681 * double-int.c (operator *=, operator +=, operator -=): ... here
12682 and implement more efficiently.
12683 (mul_double_with_sign): Remove.
12684 (lshift_double): Adjust to take unsinged shift argument, push
12685 dispatching code to callers.
12686 (mul_double_wide_with_sign): Add early out for callers that
12687 are not interested in high parts or overflow.
12688 (lshift): New function.
12689 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12690 dispatch code here.
12691 (lrotate, rrotate): Use logical shifts.
12692 * expr.c (get_inner_reference): Use lshift.
12693 * fixed-value.c (do_fixed_divide): Likewise.
12694 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12695 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12696 (indirect_refs_may_alias_p): Likewise.
12697 (stmt_kills_ref_p_1): Likewise.
12699 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12701 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12703 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12705 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12706 scalar form of FABD instruction.
12708 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12710 * lra-constraints.c (process_alt_operands): Add checking alt
12711 number to choose the best alternative.
12713 2013-05-02 Richard Biener <rguenther@suse.de>
12715 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12716 bitmap and its handling.
12717 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12719 2013-05-02 Richard Biener <rguenther@suse.de>
12721 PR middle-end/57140
12722 * tree-inline.c (copy_loops): Properly handle removed loops.
12723 (copy_cfg_body): Mark destination loops for fixup if source
12724 loops needed fixup.
12726 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12729 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12730 generating simple_return for naked functions.
12732 2013-05-02 Martin Jambor <mjambor@suse.cz>
12734 PR middle-end/56988
12735 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12736 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12738 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12739 ipa_agg_replacement_value structures.
12740 (known_aggs_to_agg_replacement_list): Likewise.
12741 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12742 (read_agg_replacement_chain): Likewise.
12743 (ipcp_transform_function): Also check that by_ref flags match.
12745 2013-05-02 Richard Biener <rguenther@suse.de>
12747 * graphds.h (struct graph): Add obstack member.
12748 * graphds.c (new_graph): Initialize obstack and allocate
12750 (add_edge): Allocate edge from the obstack.
12751 (free_graph): Free the obstack instead of all edges and vertices.
12753 2013-05-02 Teresa Johnson <tejohnson@google.com>
12755 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12757 * cfg.c (update_bb_profile_for_threading): Ditto.
12758 * tree-inline.c (copy_bb): Ditto.
12759 (copy_edges_for_bb): Ditto.
12760 (initialize_cfun): Ditto.
12761 (copy_cfg_body): Ditto.
12762 (expand_call_inline): Ditto.
12763 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12764 (estimate_node_size_and_time): Ditto.
12765 (inline_merge_summary): Ditto.
12766 * cgraphclones.c (cgraph_clone_edge): Ditto.
12767 (cgraph_clone_node): Ditto.
12768 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12769 (compute_trg_info): Ditto.
12771 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12773 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12774 S reg when fp attribute set.
12775 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12777 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12779 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12781 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12782 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12783 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12785 2013-05-02 Richard Biener <rguenther@suse.de>
12787 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12788 (struct instantiate_cache_entry): New type.
12789 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12790 (struct instantiate_cache_type): New type.
12791 (set_instantiated_value, get_instantiated_value): Remove.
12792 (get_instantiated_value_entry): New function.
12793 (instantiate_scev_name): Use the new cache and adjust.
12794 (instantiate_scev_poly): Adjust.
12795 (instantiate_scev_binary): Likewise.
12796 (instantiate_array_ref): Likewise.
12797 (instantiate_scev_convert): Likewise.
12798 (instantiate_scev_not): Likewise.
12799 (instantiate_scev_3): Likewise.
12800 (instantiate_scev_2): Likewise.
12801 (instantiate_scev_r): Likewise.
12802 (instantiate_scev): Likewise.
12803 (resolve_mixers): Likewise.
12805 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
12808 * lra-constraints.c (best_small_class_operands_num): Remove.
12809 (process_alt_operands): Remove small_class_operands_num. Take
12810 small classes operands into losers and only if the operand is not
12811 matched. Modify debugging output.
12812 (curr_insn_transform): Remove best_small_class_operands_num.
12815 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12817 * config/aarch64/aarch64-builtins.c
12818 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
12819 * config/aarch64/aarch64-simd-builtins.def
12820 (reduc_splus_): Add new modes.
12821 (reduc_uplus_): New.
12822 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
12823 (reduc_uplus_v4sf): Likewise.
12824 (reduc_splus_v4sf): Likewise.
12825 (aarch64_addv<mode>): Likewise.
12826 (reduc_uplus_<mode>): Likewise.
12827 (reduc_splus_<mode>): Likewise.
12828 (aarch64_addvv2di): Likewise.
12829 (reduc_uplus_v2di): Likewise.
12830 (reduc_splus_v2di): Likewise.
12831 (aarch64_addvv2si): Likewise.
12832 (reduc_uplus_v2si): Likewise.
12833 (reduc_splus_v2si): Likewise.
12834 (reduc_<sur>plus_<mode>): New.
12835 (reduc_<sur>plus_v2di): Likewise.
12836 (reduc_<sur>plus_v2si): Likewise.
12837 (reduc_<sur>plus_v4sf): Likewise.
12838 (aarch64_addpv4sf): Likewise.
12839 * config/aarch64/arm_neon.h
12840 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
12841 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
12842 add UNSPEC_SADDV, UNSPEC_UADDV.
12844 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
12846 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12848 * config/aarch64/arm_neon.h
12849 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
12851 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12853 * config/aarch64/aarch64-builtins
12854 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
12856 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12858 * config/aarch64/aarch64-simd-builtins.def
12859 (reduc_smax_): New.
12860 (reduc_smin_): Likewise.
12861 (reduc_umax_): Likewise.
12862 (reduc_umin_): Likewise.
12863 (reduc_smax_nan_): Likewise.
12864 (reduc_smin_nan_): Likewise.
12867 (smax): Update for V2SF, V4SF and V2DF modes.
12870 (smin_nan): Likewise.
12871 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
12872 (<su><maxmin><mode>3): ...This, refactor.
12873 (s<maxmin><mode>3): New.
12874 (<maxmin_uns><mode>3): Likewise.
12875 (reduc_<maxmin_uns>_<mode>): Refactor.
12876 (reduc_<maxmin_uns>_v4sf): Likewise.
12877 (reduc_<maxmin_uns>_v2si): Likewise.
12878 (aarch64_<fmaxmin><mode>: Remove.
12879 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
12881 (vmin<q>_f<32,64>): Likewise.
12882 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12884 (su): Add mappings for smax, smin, umax, umin.
12886 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12887 (FMAXMIN): Rename as...
12888 (FMAXMIN_UNS): ...This.
12890 (fmaxminv): Likewise.
12891 (fmaxmin): Likewise.
12893 (maxmin_uns_op): Likewise.
12895 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12897 * config/aarch64/arm_neon.h
12898 (vac<ge, gt><sd>_f<32, 64>): Rename to...
12899 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
12900 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
12902 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12904 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
12905 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
12907 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12909 * config/aarch64/aarch64-simd.md
12910 (vcond<mode>_internal): Handle special cases for constant masks.
12911 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
12912 (vcondu<mode><mode>): Likewise.
12913 (vcond<v_cmp_result><mode>): New.
12915 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12917 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
12918 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
12919 * config/aarch64/aarch64-simd-builtins.def
12920 (cmeq): Update to BUILTIN_VALLDI.
12925 * config/aarch64/arm_neon.h
12926 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
12927 to builtins or C as appropriate.
12929 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12931 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
12933 (cmhi): Rename to...
12935 * config/aarch64/aarch64-simd.md
12936 (simd_mode): Add SF.
12937 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
12938 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
12939 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
12940 (cstore<mode>_neg): ...This.
12941 * config/aarch64/iterators.md
12943 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
12944 (COMPARISONS): New.
12945 (UCOMPARISONS): Likewise.
12946 (optab): Add missing comparisons.
12952 (VCMP_S): Likewise.
12953 (VCMP_U): Likewise.
12954 (V_cmp_result): Add DF, SF modes.
12955 (v_cmp_result): Likewise.
12957 (vmtype): Likewise.
12958 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
12960 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
12962 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
12963 define_insn to define_insn_and_split.
12964 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
12965 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
12966 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
12967 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
12968 (thumb2_negscc): Likewise.
12970 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12972 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
12974 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12976 * config/arm/thumb2.md: Remove trailing whitespaces.
12978 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12980 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
12981 Use gen_int_mode rather than GEN_INT.
12983 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
12985 * value-prof.c (stream_in_histogram_value): Remove the strayed
12988 2013-04-30 Richard Biener <rguenther@suse.de>
12990 PR middle-end/57122
12991 * cfghooks.c (split_edge): Properly check for the loop latch edge.
12993 2013-04-30 Richard Biener <rguenther@suse.de>
12995 PR middle-end/57107
12996 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
12998 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
13000 PR rtl-optimization/56957
13001 PR rtl-optimization/57105
13002 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
13003 variable. Use just INSN_UID for determining whether an insn
13004 should be only disconnected from the insn stream.
13005 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
13007 2013-04-30 Jakub Jelinek <jakub@redhat.com>
13009 PR tree-optimization/57104
13010 * tsan.c (instrument_expr): Don't instrument accesses to
13011 DECL_HARD_REGISTER VAR_DECLs.
13013 2013-04-30 Richard Biener <rguenther@suse.de>
13015 * function.h (loops_for_fn): New inline function.
13016 (set_loops_for_fn): Likewise.
13017 * cfgloop.h (place_new_loop): Add struct function parameter.
13018 (get_loop): Likewise.
13019 (get_loops): Likewise.
13020 (number_of_loops): Likewise.
13021 (fel_next): Adjust.
13022 (fel_init): Likewise.
13023 * cfg.c (get_loop_copy): Adjust.
13024 * cfgloop.c (flow_loops_dump): Likewise.
13025 (record_loop_exits): Likewise.
13026 (verify_loop_structure): Likewise.
13027 * cfgloopanal.c (mark_irreducible_loops): Likewise.
13028 (estimate_reg_pressure_cost): Likewise.
13029 (mark_loop_exit_edges): Likewise.
13030 * cfgloopmanip.c (place_new_loop): Likewise.
13031 (add_loop): Likewise.
13032 (duplicate_loop): Likewise.
13033 * graph.c (draw_cfg_nodes): Likewise.
13034 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
13035 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
13036 (extract_affine_chrec): Likewise.
13037 (build_scop_iteration_domain): Likewise.
13038 * graphite.c (graphite_initialize): Likewise.
13039 * ira-build.c (create_loop_tree_nodes): Likewise.
13040 (more_one_region_p): Likewise.
13041 (rebuild_regno_allocno_maps): Likewise.
13042 (mark_loops_for_removal): Likewise.
13043 (mark_all_loops_for_removal): Likewise.
13044 (remove_unnecessary_regions): Likewise.
13045 (ira_build): Likewise.
13046 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
13047 * loop-init.c (fix_loop_structure): Likewise.
13048 (gate_rtl_move_loop_invariants): Likewise.
13049 (gate_rtl_unswitch): Likewise.
13050 (gate_rtl_unroll_and_peel_loops): Likewise.
13051 (rtl_doloop): Likewise.
13052 * lto-streamer-in.c (input_cfg): Likewise.
13053 * lto-streamer-out.c (output_cfg): Likewise.
13054 * modulo-sched.c (sms_schedule): Likewise.
13055 * predict.c (tree_estimate_probability): Likewise.
13056 (tree_estimate_probability_driver): Likewise.
13057 (estimate_loops): Likewise.
13058 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
13059 (move_sese_region_to_fn): Likewise.
13060 (debug_loop_num): Likewise.
13061 * tree-chrec.c (chrec_evaluate): Likewise.
13062 (hide_evolution_in_other_loops_than_loop): Likewise.
13063 (chrec_component_in_loop_num): Likewise.
13064 (reset_evolution_in_loop): Likewise.
13065 (evolution_function_is_invariant_rec_p): Likewise.
13066 * tree-if-conv.c (main_tree_if_conversion): Likewise.
13067 * tree-inline.c (copy_loops): Likewise.
13068 (copy_cfg_body): Likewise.
13069 (tree_function_versioning): Likewise.
13070 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
13071 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
13073 (add_to_evolution_1): Likewise.
13074 (scev_const_prop): Likewise.
13075 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
13076 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
13077 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
13078 (tree_ssa_lim_initialize): Likewise.
13079 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
13080 (verify_loop_closed_ssa): Likewise.
13081 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
13082 (tree_ssa_loop_im): Likewise.
13083 (tree_ssa_loop_unswitch): Likewise.
13084 (tree_vectorize): Likewise.
13085 (check_data_deps): Likewise.
13086 (tree_ssa_loop_ivcanon): Likewise.
13087 (tree_ssa_loop_bounds): Likewise.
13088 (tree_complete_unroll): Likewise.
13089 (tree_complete_unroll_inner): Likewise.
13090 (tree_parallelize_loops): Likewise.
13091 (tree_ssa_loop_prefetch): Likewise.
13092 (tree_ssa_loop_ivopts): Likewise.
13093 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
13094 * tree-vectorizer.c (vectorize_loops): Likewise.
13096 2013-04-29 Mike Frysinger <vapier@gentoo.org>
13098 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
13099 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
13100 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
13101 with EABI_LINK_SPEC.
13103 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13106 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
13109 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
13112 * lra-constraints.c (process_alt_operands): Discourage a bit more
13113 using memory for pseudos. Print cost dump for alternatives.
13114 Modify cost values for conflicts with early clobbers.
13115 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
13117 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13120 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
13122 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13124 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
13125 from/to S register.
13126 (movdi_aarch64): Support LDR/STR from/to D register.
13128 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13130 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
13131 or higher by default.
13133 2013-04-29 Richard Biener <rguenther@suse.de>
13135 PR middle-end/57075
13136 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
13137 even if not adding abnormal edges for calls that can make
13140 2013-04-29 Richard Biener <rguenther@suse.de>
13142 PR middle-end/57103
13143 * tree-cfg.c (move_stmt_op): Fix condition under which to update
13145 (move_stmt_r): Remove redundant checking.
13147 2013-04-29 Teresa Johnson <tejohnson@google.com>
13150 * basic-block.h (apply_scale): New function.
13151 (apply_probability): Use apply_scale.
13152 * gimple-streamer-in.c (input_bb): Ditto.
13153 * lto-streamer-in.c (input_cfg): Ditto.
13154 * lto-cgraph.c (merge_profile_summaries): Ditto.
13155 * tree-optimize.c (execute_fixup_cfg): Ditto.
13156 * tree-inline.c (copy_bb): Update comment to use apply_scale.
13157 (copy_edges_for_bb): Ditto.
13158 (copy_cfg_body): Ditto.
13160 2013-04-29 Tom de Vries <tom@codesourcery.com>
13162 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
13163 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
13164 (tail_merge_optimize): Handle current_loops == NULL.
13166 2013-04-26 Jeff Law <law@redhat.com>
13168 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
13169 (simplify_cond_using_ranges): Generalize code to simplify
13170 COND_EXPRs where one argument is a constant and the other
13171 is an SSA_NAME created by an integral type conversion.
13173 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13175 * config/arm/arm.md (store_minmaxsi): Use only when
13176 optimize_insn_for_size_p.
13178 2013-04-29 Christian Bruel <christian.bruel@st.com>
13181 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
13183 2013-04-29 Richard Biener <rguenther@suse.de>
13185 PR middle-end/57089
13186 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
13187 loop tree make sure to schedule a fixup for the child as well.
13188 (expand_omp_for_generic): Properly add loops.
13189 (expand_omp_for_static_nochunk): Likewise.
13190 (expand_omp_for_static_chunk): Likewise.
13191 (expand_omp_for): For the degenerate case fixup loops.
13192 (expand_omp_sections): Fix default bb placement in loops.
13193 (expand_omp_atomic_pipeline): Properly add loops.
13195 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13197 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
13199 2013-04-29 Tom de Vries <tom@codesourcery.com>
13201 * tree-ssa-tail-merge.c: Update header comment.
13203 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13205 * config/aarch64/arm_neon.h
13206 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
13207 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
13208 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
13209 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
13210 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
13211 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
13212 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
13213 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
13215 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13217 * config/aarch64/aarch64-simd.md
13218 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
13219 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
13220 fix_trunc, fixuns_trunc.
13221 (ftrunc<VDQF:mode>2): New.
13222 * config/aarch64/iterators.md (optab): Add fix, fixuns.
13223 (fix_trunc_optab): New.
13225 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13227 * config/aarch64/aarch64-builtins.c
13228 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
13229 iceilf, lround, iroundf.
13231 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13234 * config/i386/i386.h (enum ix86_tune_indices)
13235 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
13236 New, split from X86_TUNE_INTER_UNIT_MOVES.
13237 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
13238 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
13239 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
13240 (TARGET_INTER_UNIT_MOVES): Remove.
13241 * config/i386/i386.c (initial_ix86_tune_features): Update.
13242 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
13243 (ix86_expand_convert_uns_didf_sse): Use
13244 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13245 (ix86_expand_vector_init_one_nonzero): Ditto.
13246 (ix86_expand_vector_init_interleave): Ditto.
13247 (inline_secondary_memory_needed): Return true for moves from SSE class
13248 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
13249 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13250 * config/i386/constraints.md (Yi, Ym): Depend on
13251 TARGET_INTER_UNIT_MOVES_TO_VEC.
13252 (Yj, Yn): New constraints.
13253 * config/i386/i386.md (*movdi_internal): Change constraints of
13254 operand 1 from Yi to Yj and from Ym to Yn.
13255 (*movsi_internal): Ditto.
13256 (*movdf_internal): Ditto.
13257 (*movsf_internal): Ditto.
13258 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
13259 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13260 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
13261 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
13262 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
13263 * config/i386/sse.md (movdi_to_sse): Ditto.
13264 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
13265 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
13266 TARGET_INTER_UNIT_MOVES.
13267 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
13268 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
13269 instead of TARGET_INTER_UNIT_MOVES.
13270 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
13271 operand 1 from Yi to Yj and from Ym to Yn.
13273 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13275 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
13276 (float_truncate_hi_): Likewise.
13277 (float_extend_lo_): Likewise.
13278 (float_truncate_lo_): Likewise.
13279 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
13280 (aarch64_float_extend_lo_v2df): Likewise.
13281 (vec_unpacks_hi_v4sf): Likewise.
13282 (aarch64_float_truncate_lo_v2sf): Likewise.
13283 (aarch64_float_truncate_hi_v4sf): Likewise.
13284 (vec_pack_trunc_v2df): Likewise.
13285 (vec_pack_trunc_df): Likewise.
13287 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13289 * config/aarch64/aarch64-builtins.c
13290 (aarch64_fold_builtin): Fold float conversions.
13291 * config/aarch64/aarch64-simd-builtins.def
13292 (floatv2si, floatv4si, floatv2di): New.
13293 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
13294 * config/aarch64/aarch64-simd.md
13295 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
13296 * config/aarch64/iterators.md (FLOATUORS): New.
13297 (optab): Add float, floatuns.
13298 (su_optab): Likewise.
13300 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13302 * config/aarch64/aarch64-builtins.c
13303 (aarch64_builtin_vectorized_function): Use new names for
13305 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
13306 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
13307 (fcvtzu): Split as...
13308 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
13309 (fcvtas): Split as...
13310 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
13311 (fcvtau): Split as...
13312 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
13313 (fcvtps): Split as...
13314 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
13315 (fcvtpu): Split as...
13316 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
13317 (fcvtms): Split as...
13318 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
13319 (fcvtmu): Split as...
13320 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
13321 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
13322 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
13323 (lfrintnusf, lfrintnudf): Likewise.
13324 * config/aarch64/aarch64-simd.md
13325 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
13327 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
13328 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
13329 (fcvt_pattern): Likewise.
13331 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13333 * config/aarch64/aarch64-simd.md
13334 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
13335 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
13337 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13339 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
13340 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
13341 (vrnd<a,m,n,p>_f32): Implement using builtins.
13342 (vrnd<i,x><q>_f<32, 64>): New.
13344 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13346 * config/aarch64/aarch64-builtins.c
13347 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
13348 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
13349 (frintz): Rename to...
13351 (frintp): Rename to...
13353 (frintm): Rename to...
13355 (frinti): Rename to...
13356 (nearbyint): ...this.
13357 (frintx): Rename to...
13359 (frinta): Rename to...
13361 * config/aarch64/aarch64-simd.md
13362 (aarch64_frint<frint_suffix><mode>): Delete.
13363 (<frint_pattern><mode>2): Convert to insn.
13364 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
13365 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
13366 (frint_pattern): Likewise.
13367 (frint_suffix): Likewise.
13369 2013-04-29 Richard Biener <rguenther@suse.de>
13371 PR tree-optimization/57081
13372 * loop-init.c: Include tree-flow.h.
13373 (loop_optimizer_finalize): Free number of iteration estimates.
13374 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
13376 2013-04-29 Jakub Jelinek <jakub@redhat.com>
13378 PR tree-optimization/57083
13379 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
13380 non-singleton shift count range, zero extend low_bound for uns case.
13382 * config/i386/predicates.md (general_vector_operand): New predicate.
13383 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
13384 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
13385 if they aren't nonimmediate operands. If their original values
13386 satisfy const_vector_equal_evenodd_p, don't shift them.
13387 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
13388 predicates. For the SSE4.1 case force operands[{1,2}] into registers
13389 if not nonimmediate_operand.
13390 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
13391 instead of register_operand.
13392 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
13394 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
13396 * stor-layout.c (finalize_size_functions): Allocate a structure and
13397 reset cfun before dumping the functions.
13399 2013-04-27 Jakub Jelinek <jakub@redhat.com>
13401 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
13404 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
13405 use xop_pmacsdqh if uns_p.
13406 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
13407 the immediate rotate count.
13409 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
13411 * rtl.h (struct rtx_def): Add comment for field jump.
13412 (LRA_SUBREG_P): New macro.
13413 * recog.c (register_operand): Check LRA_SUBREG_P.
13414 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13416 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13417 (lra_final_code_change): Skip subreg change for operators.
13418 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13419 if there are no operand changes.
13420 * lra-constraints.c (curr_insn_set): New.
13421 (match_reload): Set LRA_SUBREG_P.
13422 (emit_spill_move): Ditto.
13423 (check_and_process_move): Use curr_insn_set. Process only single
13424 set insns. Don't initialize sec_mem_p and change_p.
13425 (simplify_operand_subreg): Use LRA_SUBREG_P.
13426 (reg_in_class_p): New function.
13427 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13428 of #ifdef. Add code to remove cycling.
13429 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13430 non-null disp. Reload inner instead of disp when base and index
13431 are null. Try to put lo_sum into register.
13432 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13433 (check_and_process_move): Move code for move cost check to
13434 simple_move_p. Remove equiv_substitution.
13435 (simple_move_p): New function.
13436 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13437 curr_insn_set. Call check_and_process_move only for single set
13438 insns. Use the new function. Move call of check_and_process_move
13439 after operand equiv substitution and address process.
13441 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13444 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13445 with nonlocal goto receivers or returns twice calls, ignore
13446 unininitialized values from abnormal edges to nl goto receiver
13447 or returns twice call.
13449 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13451 PR tree-optimization/57051
13452 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13453 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13456 2013-04-26 Richard Biener <rguenther@suse.de>
13458 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13459 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13460 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13461 (fixup_loop_arrays_after_move): New function.
13462 (move_sese_region_to_fn): Properly outline the loop tree parts
13463 of the SESE region.
13465 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
13467 * config/i386/i386.md (type, unit): Fix long lines.
13469 2013-04-26 Richard Biener <rguenther@suse.de>
13471 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13472 (lto-streamer-out.o): Likewise.
13473 * cfgloop.c (init_loops_structure): Export, add struct function
13474 argument and adjust.
13475 (flow_loops_find): Adjust.
13476 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13477 (init_loops_structure): Declare.
13478 * lto-streamer-in.c: Include cfgloop.h.
13479 (input_cfg): Input the loop tree.
13480 * lto-streamer-out.c: Include cfgloop.h.
13481 (output_cfg): Output the loop tree.
13482 (output_struct_function_base): Do not drop PROP_loops.
13484 2013-03-26 Richard Biener <rguenther@suse.de>
13486 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13487 (pass_build_cfg): Provide PROP_loops.
13488 (move_sese_region_to_fn): Remove loops that are outlined into fn
13490 * tree-inline.c: Include cfgloop.h.
13491 (initialize_cfun): Do not drop PROP_loops.
13492 (copy_loops): New function.
13493 (copy_cfg_body): Copy loop structure.
13494 (tree_function_versioning): Initialize destination loop tree.
13495 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13496 (pass_parallelize_loops): Do IL verification.
13497 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13498 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13499 the CFG make sure we fixup loops as well.
13500 * tree-ssa-tail-merge.c: Include cfgloop.h.
13501 (replace_block_by): When merging loop latches mark loops for fixup.
13502 * lto-streamer-out.c (output_struct_function_base): Drop
13503 PROP_loops for now.
13504 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13505 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13506 * ipa-split.c: Include cfgloop.h.
13507 (split_function): Add the new return block to the loop tree root.
13508 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13509 whether we have removed the forwarder block.
13510 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13511 * cfgloop.h (place_new_loop): Declare.
13512 * cfgloopmanip.c (place_new_loop): Export.
13513 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13514 (tree-switch-conversion.o): Likewise.
13515 (tree-complex.o): Likewise.
13516 (tree-inline.o): Likewise.
13517 (tree-ssa-tailmerge.o): Likewise.
13518 (ipa-split.o): Likewise.
13519 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13520 (tree-ssa-copy.o): Likewise.
13521 * tree-switch-conversion.c: Include cfgloop.h
13522 (process_switch): If we emit a bit-test cascade, schedule loops
13524 * tree-complex.c: Include cfgloop.h.
13525 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13526 * asan.c: Include cfgloop.h.
13527 (create_cond_insert_point): Properly add new basic-blocks to
13528 loops, schedule loop fixup.
13529 * cfgloop.c (verify_loop_structure): Check that looks are not
13531 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
13533 (expand_omp_for_generic): Likewise.
13534 (expand_omp_sections): Likewise.
13535 (expand_omp_atomic_pipeline): Schedule loops for fixup.
13536 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
13537 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
13538 is initialized, not when loops are present.
13539 * tree-parloops.c (parallelize_loops): Remove checking here.
13540 * passes.c (init_optimization_passes): Schedule a copy-propagation
13541 pass before complete unrolling of inner loops.
13543 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13545 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
13546 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
13547 (colorize_init): Add argument to _WIN32 version.
13548 * toplev.c: Include diagnostic-color.h.
13549 (process_options): Default to -fdiagnostics-color=auto if
13550 GCC_COLORS env var is in the environment.
13551 * common.opt (fdiagnostics-color=): Add Var and Init.
13552 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
13553 env var is in the environment, the default is auto rather than never.
13555 * diagnostic.h (file_name_as_prefix): Add context argument.
13556 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
13557 the string as locus.
13558 * langhooks.c (lhd_print_error_function): Adjust caller.
13560 2013-04-25 Lawrence Crowl <crowl@google.com>
13562 * var-tracking.c (shared_hash_def::htab):
13563 Change type to hash_table. Update dependent calls and types.
13565 2013-04-25 Lawrence Crowl <crowl@google.com>
13567 * Makefile.in: Update as needed below.
13569 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
13570 Move declaration to after the type's method definitons.
13572 * attribs.c (htab_t scoped_attributes::attribute_hash):
13573 Change type to hash_table. Update dependent calls and types.
13575 * bitmap.c (htab_t bitmap_desc_hash):
13576 Change type to hash_table. Update dependent calls and types.
13578 * cselib.c (htab_t cselib_hash_table):
13579 Change type to hash_table. Update dependent calls and types.
13581 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
13582 (hash_string_slot_node): Move implementation into lto-streamer.h
13583 struct string_slot_hasher.
13584 (eq_string_slot_node): Likewise.
13586 * data-streamer-out.c: Update output_block::string_hash_table
13587 dependent calls and types.
13589 * dwarf2cfi.c (htab_t trace_index):
13590 Change type to hash_table. Update dependent calls and types.
13592 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
13593 Change type to hash_table. Update dependent calls and types.
13594 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
13595 (htab_t optimize_external_refs::map): Likewise.
13596 (htab_t output_comp_unit::extern_map): Likewise.
13597 (htab_t output_comdat_type_unit::extern_map): Likewise.
13598 (htab_t output_macinfo::macinfo_htab): Likewise.
13599 (htab_t optimize_location_lists::htab): Likewise.
13600 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13602 * except.c (htab_t ehspec_hash_type):
13603 Change type to hash_table. Update dependent calls and types.
13604 (assign_filter_values::ttypes): Likewise.
13605 (assign_filter_values::ehspec): Likewise.
13606 (sjlj_assign_call_site_values::ar_hash): Likewise.
13607 (convert_to_eh_region_ranges::ar_hash): Likewise.
13609 * gcse.c (htab_t pre_ldst_table):
13610 Change type to hash_table. Update dependent calls and types.
13612 * ggc-common.c (htab_t saving_htab):
13613 Change type to hash_table. Update dependent calls and types.
13614 (htab_t loc_hash): Likewise.
13615 (htab_t ptr_hash): Likewise.
13616 (call_count): Rename ggc_call_count.
13617 (call_alloc): Rename ggc_call_alloc.
13618 (loc_descriptor): Rename make_loc_descriptor.
13619 (add_statistics): Rename ggc_add_statistics.
13621 * ggc-common.c (saving_htab):
13622 Change type to hash_table. Update dependent calls and types.
13624 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13625 (push_gimplify_context): Likewise.
13626 (pop_gimplify_context): Likewise.
13627 (struct gimple_temp_hash_elt): Added.
13628 (struct gimplify_hasher): Likewise.
13629 (struct gimplify_ctx.temp_htab):
13630 Change type to hash_table. Update dependent calls and types.
13632 * gimple-fold.c: Include gimplify-ctx.h.
13634 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13635 Change type to hash_table. Update dependent calls and types.
13636 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13637 avoid potential global name collision.
13639 * gimplify.c: Include gimplify-ctx.h.
13640 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13641 (htab_t gimplify_ctx::temp_htab):
13642 Update dependent calls and types for new type hash_table.
13643 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13644 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13646 * gimplify-ctx.h: New.
13647 (struct gimple_temp_hash_elt): Move from gimplify.c.
13648 (class gimplify_hasher): New.
13649 (struct gimplify_ctx): Move from gimple.h.
13650 (htab_t gimplify_ctx::temp_htab):
13651 Change type to hash_table. Update dependent calls and types.
13653 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13654 (htab_t ivs_params::newivs_index):
13655 Change type to hash_table. Update dependent calls and types.
13656 (htab_t ivs_params::params_index): Likewise.
13657 (htab_t print_generated_program::params_index): Likewise.
13658 (htab_t gloog::newivs_index): Likewise.
13659 (htab_t gloog::params_index): Likewise.
13661 * graphite.c: Include graphite-htab.h.
13662 4htab_t graphite_transform_loops::bb_pbb_mapping):
13663 Change type to hash_table. Update dependent calls and types.
13665 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13666 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13667 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13669 * graphite-dependences.c: Include graphite-htab.h.
13670 (loop_is_parallel_p): Change hash table type of parameter.
13672 * graphite-htab.h: New.
13673 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13674 (extern find_pbb_via_hash): Move from graphite-poly.h.
13675 (extern loop_is_parallel_p): Move from graphite-poly.h.
13676 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13678 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13679 (extern loop_is_parallel_p): Move to graphite-htab.h.
13680 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13682 * haifa-sched.c (htab_t delay_htab):
13683 Change type to hash_table. Update dependent calls and types.
13684 (htab_t delay_htab_i2): Likewise.
13686 * ira-color.c (htab_t allocno_hard_regs_htab):
13687 Change type to hash_table. Update dependent calls and types.
13689 * ira-costs.c (htab_t cost_classes_htab):
13690 Change type to hash_table. Update dependent calls and types.
13692 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13693 Change type to hash_table. Update dependent calls and types.
13695 * loop-iv.c (htab_t bivs):
13696 Change type to hash_table. Update dependent calls and types.
13698 * loop-unroll.c (htab_t opt_info::insns_to_split):
13699 Change type to hash_table. Update dependent calls and types.
13700 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13702 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13703 (struct string_slot_hasher): New.
13704 (htab_t output_block::string_hash_table):
13705 Change type to hash_table. Update dependent calls and types.
13707 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13708 (htab_t file_name_hash_table):
13709 Change type to hash_table. Update dependent calls and types.
13711 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13714 * lto-streamer.c (htab_t tree_htab):
13715 Change type to hash_table. Update dependent calls and types.
13717 * omp-low.c: Include gimplify-ctx.h.
13719 * passes.c (htab_t name_to_pass_map):
13720 Change type to hash_table. Update dependent calls and types.
13721 (pass_traverse): Rename to passes_pass_traverse.
13723 * plugin.c (htab_t event_tab):
13724 Change type to hash_table. Update dependent calls and types.
13726 * postreload-gcse.c (htab_t expr_table):
13727 Change type to hash_table. Update dependent calls and types.
13728 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13730 * sese.c (debug_rename_map_1): Make extern.
13731 (htab_t copy_bb_and_scalar_dependences::rename_map):
13732 Change type to hash_table. Update dependent calls and types.
13734 * sese.h (extern debug_rename_map): Move to .c file.
13736 * store-motion.c (htab_t store_motion_mems_table):
13737 Change type to hash_table. Update dependent calls and types.
13739 * trans-mem.c (htab_t tm_new_mem_hash):
13740 Change type to hash_table. Update dependent calls and types.
13742 * tree-browser.c (htab_t TB_up_ht):
13743 Change type to hash_table. Update dependent calls and types.
13745 * tree-cfg.c (htab_t discriminator_per_locus):
13746 Change type to hash_table. Update dependent calls and types.
13748 * tree-complex.c: Include tree-hasher.h
13749 (htab_t complex_variable_components):
13750 Change type to hash_table. Update dependent calls and types.
13752 * tree-eh.c (htab_t finally_tree):
13753 Change type to hash_table. Update dependent calls and types.
13755 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13756 struct int_tree_hasher.
13757 (extern int_tree_map_eq): Likewise.
13758 (uid_decl_map_hash): Removed.
13759 (extern decl_tree_map_eq): Likewise.
13761 * tree-hasher.h: New.
13762 (struct int_tree_hasher): New.
13763 (typedef int_tree_htab_type): New.
13765 * tree-inline.c: Include gimplify-ctx.h.
13767 * tree-mudflap.c: Include gimplify-ctx.h.
13769 * tree-parloops.c: Include tree-hasher.h.
13770 (htab_t eliminate_local_variables_stmt::decl_address):
13771 Change type to hash_table. Update dependent calls and types.
13772 (htab_t separate_decls_in_region::decl_copies): Likewise.
13774 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13775 Change type to hash_table. Update dependent calls and types.
13777 * tree-sra.c (candidates):
13778 Change type to hash_table. Update dependent calls and types.
13780 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13782 (int_tree_map_hash): Likewise.
13784 * tree-ssa-dom.c (htab_t avail_exprs):
13785 Change type to hash_table. Update dependent calls and types.
13787 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13788 Change type to hash_table. Update dependent calls and types.
13790 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
13791 Change type to hash_table. Update dependent calls and types.
13793 * tree-ssa-phiopt.c (seen_ssa_names):
13794 Change type to hash_table. Update dependent calls and types.
13796 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
13797 Change type to hash_table. Update dependent calls and types.
13799 * tree-ssa-uncprop.c (equiv):
13800 Change type to hash_table. Update dependent calls and types.
13802 2013-04-25 Jakub Jelinek <jakub@redhat.com>
13804 PR rtl-optimization/57003
13805 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
13806 call note_stores with kill_clobbered_value callback again after
13807 killing regs_invalidated_by_call.
13809 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13811 * config/aarch64/aarch64-simd.md
13812 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
13813 (aarch64_simd_bsl<mode>): Likewise.
13814 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
13816 2013-04-25 Marek Polacek <polacek@redhat.com>
13818 PR tree-optimization/57066
13819 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
13821 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
13823 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
13825 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13827 * config/aarch64/aarch64-builtins.c
13828 (aarch64_fold_builtin): New.
13829 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
13830 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
13831 * config/aarch64/aarch64-simd-builtins.def (abs): New.
13832 * config/aarch64/arm_neon.h
13833 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
13835 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13836 Tejas Belagod <tejas.belagod@arm.com>
13838 * config/aarch64/aarch64-builtins.c
13839 (aarch64_gimple_fold_builtin): New.
13840 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
13841 * config/aarch64/aarch64-simd-builtins.def (addv): New.
13842 * config/aarch64/aarch64-simd.md (addpv4sf): New.
13843 (addvv4sf): Update.
13844 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
13846 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13848 * config/aarch64/aarch64.md
13849 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
13851 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13853 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
13854 (*ngcsi_uxtw): New pattern.
13856 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13857 Julian Brown <julian@codesourcery.com>
13859 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
13860 (TB_DREG): Add T_V4HF.
13861 (v4hf_UP): New macro.
13862 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13863 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13864 Handle initialisation of V4HF. Adjust initialisation of reinterpret
13866 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13867 (arm_vector_mode_supported_p): Handle V4HF.
13868 (arm_mangle_map): Handle V4HFmode.
13869 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
13870 * config/arm/arm_neon_builtins.def: Add entries for
13871 vcvtv4hfv4sf, vcvtv4sfv4hf.
13872 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
13873 (neon_vcvtv4hfv4sf): Likewise.
13874 * config/arm/neon-gen.ml: Handle half-precision floating point
13876 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
13877 * config/arm/arm_neon.h: Regenerate.
13878 * config/arm/neon.ml (type elts): Add F16.
13879 (type vectype): Add T_float16x4, T_floatHF.
13880 (type vecmode): Add V4HF.
13881 (type features): Add Requires_FP_bit feature.
13882 (elt_width): Handle F16.
13883 (elt_class): Likewise.
13884 (elt_of_class_width): Likewise.
13885 (mode_of_elt): Refactor.
13886 (type_for_elt): Handle F16, fix error messages.
13887 (vectype_size): Handle T_float16x4.
13888 (vcvt_sh): New function.
13889 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
13890 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
13891 (string_of_mode): Handle V4HF.
13892 * doc/arm-neon-intrinsics.texi: Regenerate.
13894 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13896 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
13897 format specifier in 'X' case.
13899 2013-04-25 Alan Modra <amodra@gmail.com>
13902 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
13903 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
13904 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
13905 Repeat for many other rotate/shift and mask patterns using subregs.
13906 Name lshiftrt insns.
13907 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
13908 on WORDS_BIG_ENDIAN.
13910 2013-04-25 Alan Modra <amodra@gmail.com>
13912 * config.gcc: Support little-endian powerpc-linux targets.
13913 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
13914 (LINK_OS_LINUX_SPEC): Define.
13915 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
13916 Preserve MASK_LITTLE_ENDIAN.
13917 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
13918 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
13919 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
13920 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
13921 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
13922 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
13923 Correct fp word order for little-endian. Don't shift toc entries
13924 smaller than a word for little-endian.
13925 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
13926 (bswapdi2 splits): Correct low-part subreg for little-endian.
13927 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
13928 low/high where such is correct only for be.
13929 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
13930 little-endian for -mcall-aixdesc.
13932 2013-04-25 Alan Modra <amodra@gmail.com>
13934 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
13935 replace_equiv_address_nv.
13937 2013-04-25 Alan Modra <amodra@gmail.com>
13939 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
13941 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13944 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13945 * rtl.h (struct rtx_def): ...
13947 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13949 PR rtl-optimizations/57046
13950 * lra-constraints (split_reg): Set up lra_risky_transformations_p
13951 for multi-reg splits.
13953 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
13955 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
13957 2013-04-24 Sterling Augustine <saugustine@google.com>
13959 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
13960 (comp_dir_string, debug_str_dwo_section): New.
13961 (DEBUG_STR_DWO_SECTION): Rename to ...
13962 (DEBUG_DWO_STR_SECTION): ... this.
13963 (DEBUG_NORM_STR_SECTION): Delete.
13964 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
13965 (DEBUG_STR_DWO_SECTION_FLAGS): New.
13966 (find_AT_string): Move most logic to ...
13967 (find_AT_string_in_table): ... here. New.
13968 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
13969 add_skeleton_AT_string. Delete logic.
13970 (output_skeleton_debug_sections): Remove call to
13971 add_top_level_skeleton_die_attrs.
13972 (add_comp_dir_attribute): Move logic to comp_dir_string.
13973 (dwarf2out_init): Initialize debug_str_dwo_section.
13974 (output_indirect_string): Call find_string_form.
13975 (output_indirect_strings): Rewrite.
13976 (prune_unused_types): Empty skeleton_debug_str_hash.
13977 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
13978 (dwarf2out_finish): Call output_indirect_strings.
13980 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13982 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
13984 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13986 * rtl.h (struct rtx_def): Add comment for field jump.
13987 (LRA_SUBREG_P): New macro.
13988 * recog.c (register_operand): Check LRA_SUBREG_P.
13989 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13991 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13992 (lra_final_code_change): Skip subreg change for operators.
13993 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13994 if there are no operand changes.
13995 * lra-constraints.c (curr_insn_set): New.
13996 (match_reload): Set LRA_SUBREG_P.
13997 (emit_spill_move): Ditto.
13998 (check_and_process_move): Use curr_insn_set. Process only single
13999 set insns. Don't initialize sec_mem_p and change_p.
14000 (simplify_operand_subreg): Use LRA_SUBREG_P.
14001 (reg_in_class_p): New function.
14002 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14003 of #ifdef. Add code to remove cycling.
14004 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14005 non-null disp. Reload inner instead of disp when base and index
14006 are null. Try to put lo_sum into register.
14007 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14008 (check_and_process_move): Move code for move cost check to
14009 simple_move_p. Remove equiv_substitution.
14010 (simple_move_p): New function.
14011 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14012 curr_insn_set. Call check_and_process_move only for single set
14013 insns. Use the new function. Move call of check_and_process_move
14014 after operand equiv substitution and address process.
14016 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
14018 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
14019 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
14020 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
14022 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14024 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
14026 2013-04-24 Marek Polacek <polacek@redhat.com>
14028 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
14029 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
14030 (select_loops_exit_conditions): Likewise.
14031 (number_of_iterations_for_all_loops): Likewise.
14032 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
14033 (scev_analysis): Likewise.
14035 2013-04-02 Catherine Moore <clm@codesourcery.com>
14036 Chao-ying Fu <fu@mips.com>
14038 * config/mips/micromips.md (jraddiusp): New pattern.
14039 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
14040 instruction if possible.
14042 2013-04-24 Alan Modra <amodra@gmail.com>
14044 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
14046 2013-04-24 Julian Brown <julian@codesourcery.com>
14047 Chung-Lin Tang <cltang@codesourcery.com>
14049 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
14050 dependency behavior in enumeration type DIE generation. Add TODO note
14051 to comments about future DW_FORM_sdata/udata re-work of related code.
14053 2013-04-23 Lawrence Crowl <crowl@google.com>
14055 * Makefile.in: Update as needed below.
14057 * hash-table.h (class hash_table):
14058 Correct many methods with parameter types compare_type to the correct
14059 value_type. (Correct code was unlikely to notice the change.)
14060 (hash_table::elements_with_deleted) New.
14061 (class hashtable::iterator): New.
14062 (hashtable::begin()): New.
14063 (hashtable::end()): New.
14064 (FOR_EACH_HASH_TABLE_ELEMENT): New.
14066 * statistics.c (statistics_hashes):
14067 Change type to hash_table. Update dependent calls and types.
14069 * tree-into-ssa.c (var_infos):
14070 Change type to hash_table. Update dependent calls and types.
14072 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
14073 Change type to hash_table. Update dependent calls and types.
14075 * tree-ssa-loop-im.c (struct mem_ref.refs):
14076 Change type to hash_table. Update dependent calls and types.
14078 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
14079 Change type to hash_table. Update dependent calls and types.
14081 * tree-ssa-sccvn.c (vn_tables_s::nary):
14082 Change type to hash_table. Update dependent calls and types.
14083 (vn_tables_s::phis): Likewise.
14084 (vn_tables_s::references): Likewise.
14086 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
14087 (vn_reference_eq): Update parameter and return types.
14089 * tree-ssa-structalias.c (pointer_equiv_class_table):
14090 Change type to hash_table. Update dependent calls and types.
14091 (location_equiv_class_table): Likewise.
14093 * tree-vect-data-refs.c: Consequential changes for making
14094 peeling a hash_table.
14096 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
14097 (destroy_loop_vec_info): Dependent hash_table update.
14099 * tree-vectorizer.h (peeling_htab):
14100 Change type to hash_table. Update dependent calls and types.
14102 2013-04-23 Shiva Chen <shiva0217@gmail.com>
14104 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
14105 to check the register content is equal or not.
14106 * lra-constraints.c (match_reload): Use lra_assign_reg_val
14107 to assign register content record.
14108 * lra-eliminations.c (update_reg_eliminate): Use
14109 lra_update_reg_val_offset to update register content offset.
14110 * lra-int.h (struct lra_reg): Add offset member.
14111 (lra_reg_val_equal_p): New static inline function.
14112 (lra_update_reg_val_offset): New static inline function.
14113 (lra_assign_reg_val): New static inline function.
14114 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
14115 to assign register content record.
14116 (initialize_lra_reg_info_element): Initial offset to zero.
14118 2013-04-23 Catherine Moore <clm@codesourcery.com>
14120 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
14121 operands. Record compression.
14123 2013-04-23 Xinliang David Li <davidxl@google.com>
14125 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
14127 2013-04-23 Richard Biener <rguenther@suse.de>
14129 PR middle-end/57036
14130 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
14131 parameter, only add abnormal goto edges from the copied body
14132 if the call could perform abnormal gotos.
14133 (copy_cfg_body): Adjust.
14135 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
14137 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
14139 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
14141 * coretypes.h (gimple_stmt_iterator): Add struct to make
14144 2013-04-23 Richard Biener <rguenther@suse.de>
14146 PR tree-optimization/57026
14147 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
14148 from SSA names occuring in abnormal PHI nodes.
14150 2013-04-22 Andi Kleen <ak@linux.intel.com>
14152 * lto/lto.c (print_lto_report_1): Fix LTO report names.
14154 2013-04-22 Andi Kleen <ak@linux.intel.com>
14156 * lto/lto.c (print_lto_report_1): Declare early.
14157 (read_cgraph_and_symbols): Call print_lto_report_1 early.
14159 2013-04-22 Andi Kleen <ak@linux.intel.com>
14161 * common.opt (-flto-report-wpa): Add.
14162 * doc/invoke.texi (-flto-report-wpa): Add.
14163 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
14166 2013-04-22 Xinliang David Li <davidxl@google.com>
14168 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
14169 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
14170 * Makefile.in: New dependency
14172 David Daney <ddaney.cavm@gmail.com>
14174 * configure.ac (gcc_cv_as_micromips_support): Use the
14175 --fatal-warnings option.
14176 * configure: Regenerate.
14178 2013-04-22 Marek Polacek <polacek@redhat.com>
14181 * tsan.c (instrument_expr): Don't instrument expression
14182 in case its size is zero.
14184 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
14188 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14190 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14192 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14194 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
14195 (gimple_stmt_iterator): New typedef.
14196 * gimple.h (gimple_stmt_iterator): Rename to...
14197 (gimple_stmt_iterator_d): ... This.
14198 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
14199 trees be valid for GIMPLE and GENERIC.
14200 (TARGET_GIMPLE_FOLD_BUILTIN): New.
14201 * gimple-fold.c (gimple_fold_call): Call target hook
14202 gimple_fold_builtin.
14203 * hooks.c (hook_bool_gsiptr_false): New.
14204 * hooks.h (hook_bool_gsiptr_false): New.
14205 * target.def (fold_stmt): New.
14206 * doc/tm.texi: Regenerate.
14208 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
14211 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
14212 a set sp if no stack realignment.
14214 2013-04-22 Nick Clifton <nickc@redhat.com>
14216 * config.gcc (tilegx-linux): Extend extra_objs rather than
14218 (tilepro-linux): Likewise.
14220 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14222 * config/aarch64/aarch64-builtins.c
14224 (CF0, CF1, CF2, CF3, CF4, CF10): New.
14225 (VAR<1-12>): Add MAP parameter.
14226 (BUILTIN_*): Likewise.
14227 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
14228 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
14229 (aarch64_ushl_n<mode>): Likewise.
14230 (aarch64_sshr_n<mode>): Likewise.
14231 (aarch64_ushr_n<mode>): Likewise.
14232 (aarch64_<maxmin><mode>): Likewise.
14233 (aarch64_sqrt<mode>): Likewise.
14234 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
14235 (vshr<q>_n_*): Likewise.
14237 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14239 * config/aarch64/aarch64-builtins.c
14240 (aarch64_simd_builtin_type_mode): Handle SF types.
14242 (BUILTIN_GPF): Define.
14243 (aarch64_init_simd_builtins): Handle SF types.
14244 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
14245 (frecps): Likewise.
14246 (frecpx): Likewise.
14247 * config/aarch64/aarch64-simd.md
14248 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
14249 (aarch64_frecpe<mode>): New.
14250 (aarch64_frecps<mode>): Likewise.
14251 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
14252 (v8type): Add frecp<esx>.
14253 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
14254 (aarch64_frecps<mode>): Likewise.
14255 * config/aarch64/iterators.md (FRECP): New.
14256 (frecp_suffix): Likewise.
14257 * config/aarch64/arm_neon.h
14258 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
14260 2013-04-22 Christian Bruel <christian.bruel@st.com>
14263 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
14264 (REG_CLASS_NAMES): Idem.
14265 (REG_CLASS_CONTENTS): Idem.
14266 (REGCLASS_HAS_FP_REG): Idem.
14267 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
14268 (sh_conditional_register_usage): Idem.
14270 2013-04-21 Jeff Law <law@redhat.com>
14272 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
14273 (ssa_forward_propagate_and_combine): Use it.
14275 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14277 * lra.c: Update the flow chart diagram.
14279 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14281 PR rtl-optimization/56847
14282 * lra-constraints.c (process_alt_operands): Discourage alternative
14283 with non-matche doffsettable memory constraint fro memory with
14286 2013-04-19 Richard Biener <rguenther@suse.de>
14288 PR tree-optimization/56982
14289 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
14291 * gimplify.c (gimplify_call_expr): Notice special calls.
14292 (gimplify_modify_expr): Likewise.
14293 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
14294 abnormal control flow receivers.
14295 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
14296 in the same way as cfun->has_nonlocal_labels.
14297 (gimple_purge_dead_abnormal_call_edges): Likewise.
14298 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
14299 receivers start a basic-block.
14301 2013-04-19 Richard Biener <rguenther@suse.de>
14303 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
14305 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
14306 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
14307 (SLP_TREE_LOAD_PERMUTATION): Add.
14308 (vect_transform_slp_perm_load): Adjust prototype.
14309 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
14310 (vect_free_slp_instance): Likewise.
14311 (vect_create_new_slp_node): Likewise.
14312 (vect_supported_slp_permutation_p): Remove.
14313 (vect_slp_rearrange_stmts): Adjust.
14314 (vect_supported_load_permutation_p): Likewise. Inline
14315 vect_supported_slp_permutation_p here.
14316 (vect_analyze_slp_instance): Compute load permutations per
14317 slp node instead of per instance.
14318 (vect_get_slp_defs): Adjust.
14319 (vect_transform_slp_perm_load): Likewise.
14320 (vect_schedule_slp_instance): Remove redundant code.
14321 (vect_schedule_slp): Remove hack for PR56270, add it ...
14322 * tree-vect-stmts.c (vectorizable_load): ... here, do not
14323 CSE loads for SLP. Adjust.
14325 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14327 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
14328 spelling in two comments.
14330 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14333 * config/arm/arm.c (load_multiple_sequence): Require SP
14334 as base register for loads if SP is in the register list.
14336 2013-04-19 Martin Jambor <mjambor@suse.cz>
14338 PR tree-optimization/56718
14339 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
14340 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
14341 and made public. Adjusted all callers.
14342 (ipa_intraprocedural_devirtualization): New function.
14343 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
14344 (ipa_intraprocedural_devirtualization): Likewise.
14345 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
14347 2013-04-19 Richard Biener <rguenther@suse.de>
14349 PR tree-optimization/57000
14350 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
14352 2013-04-19 Terry Guo <terry.guo@arm.com>
14354 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
14356 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
14357 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
14358 (cortex_m4_fmacs): Use new reservations.
14359 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
14361 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
14363 PR rtl-optimization/56999
14364 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
14366 (lra_coalesce): Remove split_origin_bitmap and related code.
14367 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
14368 ranges if necessary.
14370 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
14372 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
14374 (ix86_expand_call): Remove clobbered_registers array and use
14375 x86_64_ms_sysv_extra_clobbered_registers instead.
14376 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
14378 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
14380 * config/i386/i386.md (*call_rex64_ms_sysv): Use
14381 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
14382 (*call_value_rex64_ms_sysv): Ditto.
14384 2013-04-18 Cary Coutant <ccoutant@google.com>
14386 * dwarf2out.c (output_pubnames): Check die_perennial_p of
14387 parent instead of die_mark.
14389 2013-04-18 Diego Novillo <dnovillo@google.com>
14391 * gimple.c (create_gimple_tmp): New.
14392 (get_expr_type): New.
14393 (build_assign): New.
14394 (build_type_cast): New.
14395 * gimple.h (enum ssa_mode): Define.
14396 (gimple_seq_set_location): New.
14397 * asan.c (build_check_stmt): Change some gimple_build_* calls
14398 to use build_assign and build_type_cast.
14400 2013-04-18 Richard Biener <rguenther@suse.de>
14402 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
14403 handle negative step. Remove redundant checks.
14404 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
14405 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
14406 for negative step and grouped loads fail to vectorize.
14408 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14410 * emit-rtl.c (reset_insn_used_flags): New function.
14411 (reset_all_used_flags): Use it.
14412 (verify_insn_sharing): New function.
14413 (verify_rtl_sharing): Fix verification for SEQUENCEs.
14415 2013-04-18 Jakub Jelinek <jakub@redhat.com>
14417 PR tree-optimization/56984
14418 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
14419 and (x >> M) >= N don't register any assertion if N << M is the
14422 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14424 * lower-subreg.c (resolve_simple_move): If called self-recursive,
14425 do not delete_insn insns that have not yet been emitted, only
14426 unlink them with remove_insn.
14427 * df-scan.c (df_insn_delete): Revert r197492.
14429 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
14431 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
14432 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
14434 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14436 * config/arm/arm.md (movsicc_insn): Convert define_insn into
14437 define_insn_and_split.
14438 (and_scc,ior_scc,negscc): Likewise.
14439 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14441 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14443 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14444 can benefit from using a sequence of LDRD instructions in epilogue
14445 instead of a single LDM instruction.
14447 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14450 * doc/extend.texi: Fix typo.
14452 2013-04-17 Richard Biener <rguenther@suse.de>
14454 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14455 (vect_build_slp_tree): ... here.
14456 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14457 match. Remove special-casing of mismatched complex loads.
14458 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14459 re-try the match with swapped commutative operands.
14460 (vect_supported_load_permutation_p): Remove special-casing of
14461 mismatched complex loads.
14462 (vect_analyze_slp_instance): Adjust.
14464 2013-04-17 Richard Biener <rguenther@suse.de>
14466 PR rtl-optimization/56921
14467 * cfgloop.h (struct loop): Add simple_loop_desc member.
14468 (struct niter_desc): Mark with GTY(()).
14469 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14470 * loop-iv.c (get_simple_loop_desc): Likewise.
14471 (free_simple_loop_desc): Likewise.
14474 2013-04-16 Richard Biener <rguenther@suse.de>
14476 PR rtl-optimization/56921
14477 * loop-init.c (pass_rtl_move_loop_invariants): Add
14478 TODO_do_not_ggc_collect to todo_flags_finish.
14479 (pass_rtl_unswitch): Same.
14480 (pass_rtl_unroll_and_peel_loops): Same.
14481 (pass_rtl_doloop): Same.
14483 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14485 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14486 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14487 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14489 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14490 * tree-streamer.c (record_common_node): Adjust reference in comment.
14492 2013-04-17 Terry Guo <terry.guo@arm.com>
14494 * config/arm/cortex-m4.md: Add a new bypass.
14496 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14498 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14500 (*subs_<optab><mode>_multp2): New pattern.
14501 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14502 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14504 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14506 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14507 (*subs_mul_imm_<mode>): New pattern.
14509 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
14512 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14513 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14514 (vsx_movti_32bit): Change j->wa to O->wa.
14516 2013-04-16 Richard Biener <rguenther@suse.de>
14518 PR rtl-optimization/56921
14519 * loop-init.c (pass_rtl_move_loop_invariants): Add
14520 TODO_do_not_ggc_collect to todo_flags_finish.
14521 (pass_rtl_unswitch): Same.
14522 (pass_rtl_unroll_and_peel_loops): Same.
14523 (pass_rtl_doloop): Same.
14525 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
14527 * config/arm/arm.c (emit_multi_reg_push): New declaration
14528 for an existing function.
14529 (arm_emit_strd_push): New function.
14530 (arm_expand_prologue): Used here.
14531 (arm_emit_ldrd_pop): New function.
14532 (arm_expand_epilogue): Used here.
14533 (arm_get_frame_offsets): Update condition.
14534 (arm_emit_multi_reg_pop): Add a special case for load of a single
14535 register with writeback.
14537 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
14539 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
14542 2013-04-16 Richard Biener <rguenther@suse.de>
14544 PR tree-optimization/56756
14545 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
14546 (first_mem_ref_loc): New.
14547 (execute_sm): Place the load temporarily before a previous
14548 access instead of in the latch edge to ensure its SSA dependencies
14549 are defined at points dominating the load.
14551 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
14553 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
14554 correct fix by moving header and footer insn to the footer of
14555 the merged basic block. Clear BB_END of the merged-away block.
14557 PR middle-end/43631
14558 * emit-rtl.c (make_note_raw): New function.
14559 (link_insn_into_chain): New static inline function.
14560 (add_insn): Use it.
14561 (add_insn_before, add_insn_after): Factor insn chain linking code...
14562 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
14563 using link_insn_into_chain.
14564 (note_outside_basic_block_p): New helper function for emit_note_after
14565 and emit_note_before.
14566 (emit_note_after): Use nobb variant of add_insn_after if the note
14567 should not be contained in a basic block.
14568 (emit_note_before): Use nobb variant of add_insn_before if the note
14569 should not be contained in a basic block.
14570 (emit_note_copy): Use make_note_raw.
14571 (emit_note): Likewise.
14572 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
14573 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
14574 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
14575 the moved barrier the tail of the basic block it follows.
14576 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
14578 2013-04-15 Jakub Jelinek <jakub@redhat.com>
14580 PR tree-optimization/56962
14581 * gimple-ssa-strength-reduction.c (record_increment): Only set
14582 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
14583 either rhs1 or rhs2 is equal to c->base_expr.
14585 2013-04-15 Richard Biener <rguenther@suse.de>
14587 PR tree-optimization/56933
14588 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
14590 (GROUP_READ_WRITE_DEPENDENCE): Remove.
14591 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
14592 * tree-vect-data-refs.c (vect_analyze_group_access): Move
14593 dependence check ...
14594 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14596 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14597 GROUP_READ_WRITE_DEPENDENCE.
14599 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14601 * emit-rtl.c (reset_all_used_flags): New function.
14602 (verify_rtl_sharing): Call reset_all_used_flags before and after
14603 performing the checks.
14605 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14607 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14608 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14609 * config/arm/constraints.md (De): New constraint.
14610 * config/arm/neon.md (anddi3_neon): Delete.
14611 (neon_vand<mode>): Expand to standard anddi3 pattern.
14612 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14613 Move earlier in the file.
14614 (neon_inv_logic_op2): Likewise.
14615 (arm_anddi_operand_neon): New predicate.
14617 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14619 * configure.ac (gcc_cv_ld_as_needed): Set
14620 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14621 Use -z ignore, -z record on *-*-solaris2*.
14622 (HAVE_LD_AS_NEEDED): Update comment.
14623 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14624 * configure: Regenerate.
14625 * config.in: Regenerate.
14626 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14627 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14628 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14629 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14630 equivalents. Fix markup.
14631 * doc/tm.texi: Regenerate.
14633 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14635 * config/i386/i386.opt: New option mstack-protector-guard=.
14636 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14637 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14638 TARGET_SSP_TLS_GUARD.
14639 * config/i386/i386.c (ix86_option_override_internal): Set
14640 ix86_stack_protector_guard.
14641 * config/i386/i386.md (stack_protect_set): Enable for
14642 TARGET_SSP_TLS_GUARD only.
14643 (stack_protect_set_<mode>): Ditto.
14644 (stack_protect_test): Ditto.
14645 (stack_protect_test_<mode>): Ditto.
14646 * doc/invoke.texi (i386 Option): Document.
14648 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14651 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14652 (S_MODES): Set H_MODE bit.
14653 (SF_MODES): Set only S_MODE and SF_MODE bits.
14654 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14655 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14656 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14657 <MODE_FLOAT>: Likewise.
14659 2013-04-15 Joey Ye <joey.ye@arm.com>
14661 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14663 2013-04-15 Joey Ye <joey.ye@arm.com>
14665 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14667 (thumb_far_jump_used_p): Count instruction size and set
14670 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14672 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14673 * resource.c (mark_target_live_regs): Reformat conditional block.
14675 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14677 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14678 notes, they are emitted only just before final.
14679 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14681 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14683 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14684 * cfgrtl.c (delete_insn): Call it here instead.
14685 * lra-spills.c (lra_final_code_change): Use delete_insn.
14686 * haifa-sched.c (sched_remove_insn): Likewise.
14687 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14688 returning to the nop pool.
14689 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14690 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14692 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14694 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14695 * doc/tm.texi: Regenerated.
14697 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
14699 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14702 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14704 * df-core.c (df_find_def): Compare register numbers.
14705 (df_find_use): Likewise.
14707 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14710 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14711 lra_in_progress for return.
14713 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14715 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14716 define_insn into define_insn_and_split and emit movsicc patterns.
14718 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14720 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14722 2013-04-12 Richard Biener <rguenther@suse.de>
14724 * tree-pass.h (TODO_do_not_ggc_collect): New.
14725 * passes.c (execute_one_ipa_transform_pass): Honor
14726 TODO_do_not_ggc_collect.
14727 (execute_one_pass): Likewise.
14730 2013-04-10 Richard Biener <rguenther@suse.de>
14732 * passes.c (init_optimization_passes): Remove reload pass.
14733 * ira.c (do_reload): Merge into ...
14735 (rest_of_handle_reload): Remove.
14736 (pass_reload): Likewise.
14737 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14738 of reload for vzeroupper pass placement.
14740 2013-04-12 Jakub Jelinek <jakub@redhat.com>
14742 PR tree-optimization/56918
14743 PR tree-optimization/56920
14744 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14745 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14746 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14747 use wide_mul_with_sign method.
14749 2013-04-12 Richard Biener <rguenther@suse.de>
14751 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14752 not be considered a gimple constant.
14754 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
14756 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14757 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14760 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14761 Jakub Jelinek <jakub@redhat.com>
14763 * opts.c: Include diagnostic-color.h.
14764 (common_handle_option): Handle OPT_fdiagnostics_color_.
14765 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14766 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14767 (diagnostic-color.o): New.
14768 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14769 (diagnostic_color_rule): New enum.
14770 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14771 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14772 the location string.
14773 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14774 either NULL, or color kind.
14775 * diagnostic-color.c: New file.
14776 * diagnostic-color.h: New file.
14777 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14779 * doc/invoke.texi (-fdiagnostics-color): Document.
14780 * pretty-print.h (pp_show_color): Define.
14781 (struct pretty_print_info): Add show_color field.
14782 * diagnostic.c: Include diagnostic-color.h.
14783 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14784 macros. Colorize error:, warning: etc. strings and also the location
14786 (diagnostic_show_locus): Colorize the caret line.
14787 * pretty-print.c: Include diagnostic-color.h.
14788 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14789 inside of %< %> quotes or quoted through q format modifier.
14791 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14793 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
14795 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14797 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
14798 code in CC_NZ mode.
14799 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
14802 2013-04-11 Marek Polacek <polacek@redhat.com>
14804 PR tree-optimization/48184
14805 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
14807 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
14809 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
14810 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
14811 (skip_simple_arithmetic): Tidy up.
14812 * tree.h (skip_simple_constant_arithmetic): Declare.
14814 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14816 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
14818 2013-04-11 Richard Biener <rguenther@suse.de>
14820 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14821 generate vector constants.
14823 2013-04-11 Richard Biener <rguenther@suse.de>
14825 PR tree-optimization/56878
14826 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
14827 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
14829 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
14830 Prefer to align the DR with the most invariant base address.
14832 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14834 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
14837 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
14839 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
14840 floating-point vector comparisons against 0.
14842 2013-04-11 Jakub Jelinek <jakub@redhat.com>
14844 PR tree-optimization/56899
14845 * fold-const.c (extract_muldiv_1): Apply distributive law
14846 only if TYPE_OVERFLOW_WRAPS (ctype).
14848 2013-04-11 Bin Cheng <bin.cheng@arm.com>
14851 * ira-costs.c (scan_one_insn): Check whether the source rtx of
14852 loading has side effect.
14854 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
14856 * config/sparc/sparc.c: Include tree-pass.h.
14857 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
14858 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
14859 head of file. Change return type. Split off gate function.
14860 (sparc_gate_work_around_errata): New function.
14861 (pass_work_around_errata): New pass definition.
14862 (insert_pass_work_around_errata) New pass insert definition to
14863 insert pass_work_around_errata just after delayed-branch scheduling.
14864 (sparc_option_override): Insert the pass.
14865 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
14867 2013-04-10 David S. Miller <davem@davemloft.net>
14869 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
14870 or -mcpu=hypersparc.
14872 * target.def (cstore_mode): New hook.
14873 * target.h: Include insn-codes.h
14874 * targhooks.c: Likewise.
14875 (default_cstore_mode): New function.
14876 * targhooks.h: Declare it.
14877 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
14878 * doc/tm.texi: Rebuild.
14879 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
14880 target hook, rather than inspecting the insn_data.
14881 * config/sparc/sparc.c (sparc_cstore_mode): New function.
14882 (TARGET_CSTORE_MODE): Redefine.
14883 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
14885 * config/sparc/predicates.md (cstore_result_operand): New special
14887 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
14888 Use it for operand 0.
14889 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
14890 (*snesi_special): Likewise.
14891 (*snesi_zero): Likewise.
14892 (*seqsi_zero): Likewise.
14893 (*sltu_insn): Likewise.
14894 (*sgeu_insn): Likewise.
14895 (*seqdi_special): Make operand 0 and comparison operation be of
14897 (*snedi_special): Likewise.
14898 (*snedi_special_vis3): Likewise.
14899 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
14900 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
14901 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
14902 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
14903 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
14904 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
14905 (*sltu_extend_sp64): Likewise.
14906 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
14907 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
14908 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
14909 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
14910 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
14912 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
14914 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
14915 (aarch64_start_file): Use the new function.
14917 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14918 Jason Merrill <jason@redhat.com>
14920 * common.opt: Add -gdwarf.
14921 * opts.c (common_handle_option): Handle it.
14922 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
14924 2013-04-10 Richard Biener <rguenther@suse.de>
14926 * passes.c (execute_todo): Do not call ggc_collect conditional here.
14927 (execute_one_ipa_transform_pass): But unconditionally here.
14928 (execute_one_pass): And here.
14929 (init_optimization_passes): Remove reload pass.
14930 * tree-pass.h (TODO_ggc_collect): Remove.
14931 (pass_reload): Likewise.
14932 * ira.c (do_reload): Merge into ...
14934 (rest_of_handle_reload): Remove.
14935 (pass_reload): Likewise.
14936 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14937 of reload for vzeroupper pass placement.
14938 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
14939 and todo_flags_finish of all passes.
14941 2013-04-10 Richard Biener <rguenther@suse.de>
14943 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
14944 first_const_oprnd field, rename first_def_type to first_op_type.
14945 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
14946 (vect_get_and_check_slp_defs): Always use the type of the
14947 operand. Allow mixed vect_external_def, vect_constant_def types.
14948 (vect_get_constant_vectors): Handle mixed vect_external_def,
14949 vect_constant_def types.
14951 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
14953 PR tree-optimization/55524
14954 * tree-ssa-math-opts.c
14955 (convert_mult_to_fma): Don't use an fms construct
14956 when we don't have an fms operation, but fnma, and it looks
14957 likely that we'll be able to use the latter.
14959 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
14961 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
14963 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
14964 inline fail caused by overwritable functions.
14966 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
14968 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
14969 unnecessary bits in the constant power of two case.
14971 2013-04-10 Richard Biener <rguenther@suse.de>
14973 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
14974 broken code swapping operands.
14975 (vect_build_slp_tree): Do not compute load permutations here.
14976 (vect_analyze_slp_instance): Compute load permutations here,
14977 after building the SLP tree.
14979 2013-04-09 Christian Bruel <christian.bruel@st.com>
14981 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
14982 of next/prev_real_insn.
14984 2013-04-09 Jan Hubicka <jh@suse.cz>
14986 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14987 Drop aliased parameter.
14988 (function_and_variable_visibility): Do not handle alias pairs.
14989 * cgraph.c (varpool_externally_visible_p): Update prototype.
14990 * varpool.c (varpool_add_new_variable): Update.
14992 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14994 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
14996 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
14998 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
15000 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
15002 2013-04-09 Marek Polacek <polacek@redhat.com>
15004 PR tree-optimization/48762
15005 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
15007 2013-04-09 Richard Biener <rguenther@suse.de>
15009 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
15011 (vect_build_slp_tree): Likewise.
15012 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
15013 calculating the cost of a SLP instance.
15014 (vect_analyze_slp_instance): Use it from here, after building
15017 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15019 PR middle-end/56883
15020 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15021 expand_omp_for_static_chunk): Use simple_p = true in
15022 force_gimple_operand_gsi calls when assigning to addressable decls.
15024 2013-04-09 Jeff Law <law@redhat.com>
15026 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
15027 when the boolean was created by converting a wider object which
15028 had a boolean range.
15030 2013-04-09 Richard Biener <rguenther@suse.de>
15032 * tree-vectorizer.h (slp_void_p): Remove.
15033 (slp_tree): Typedef before _slp_tree declaration.
15034 (struct _slp_tree): Use a vector of slp_tree as children.
15035 (vect_get_place_in_interleaving_chain): Remove.
15036 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
15038 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
15040 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
15041 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
15042 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
15043 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
15044 Use slp_node instead of slp_void_p and adjust.
15046 2013-04-09 Richard Biener <rguenther@suse.de>
15048 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
15049 work that is not necessary.
15051 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15053 PR tree-optimization/56854
15054 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
15055 forward into clobber stmts if it would change MEM_REF lhs into
15058 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
15060 * tree.c (type_hash_lookup, type_hash_add): Make static.
15061 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
15063 2013-04-09 Richard Biener <rguenther@suse.de>
15065 * tree.h (unsave_expr_now): Remove.
15066 * tree-inline.c (mark_local_for_remap_r): Remove.
15067 (unsave_expr_1): Likewise.
15068 (unsave_r): Likewise.
15069 (unsave_expr_now): Likewise.
15070 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
15071 (propagate_tree_value): Likewise.
15073 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
15075 * doc/rtl.texi (sequence): Rewrite documentation to match the
15076 current use of SEQUENCE rtl objects.
15077 * rtl.def (SEQUENCE): Likewise.
15079 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
15080 Update documentation.
15081 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
15082 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
15084 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
15086 2013-04-08 Teresa Johnson <tejohnson@google.com>
15088 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
15089 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
15091 (estimate_edge_size_and_time): Add comment to suggest using rounding
15093 (estimate_node_size_and_time): Ditto.
15094 (remap_edge_change_prob): Use helper rounding divide methods.
15095 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
15096 (gimple_mod_pow2_value_transform): Ditto.
15097 (gimple_mod_subtract_transform): Ditto.
15098 (gimple_ic_transform): Ditto.
15099 (gimple_stringops_transform): Ditto.
15100 * stmt.c (conditional_probability): Ditto.
15101 (emit_case_dispatch_table): Ditto.
15102 * lto-cgraph.c (merge_profile_summaries): Ditto.
15103 * tree-optimize.c (execute_fixup_cfg): Ditto.
15104 * cfgcleanup.c (try_forward_edges): Ditto.
15105 * cfgloopmanip.c (scale_loop_profile): Ditto.
15107 (duplicate_loop_to_header_edge): Ditto.
15108 (lv_adjust_loop_entry_edge): Ditto.
15109 * tree-vect-loop.c (vect_transform_loop): Ditto.
15110 * profile.c (compute_branch_probabilities): Ditto.
15111 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
15112 * lto-streamer-in.c (input_cfg): Ditto.
15113 * gimple-streamer-in.c (input_bb): Ditto.
15114 * ipa-cp.c (update_profiling_info): Ditto.
15115 (update_specialized_profile): Ditto.
15116 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
15117 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
15119 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15120 (compute_trg_info): Ditto.
15121 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
15122 (purge_dead_edges): Ditto.
15123 * loop-unswitch.c (unswitch_loop): Ditto.
15124 * cgraphclones.c (cgraph_clone_edge): Ditto.
15125 (cgraph_clone_node): Ditto.
15126 * tree-inline.c (copy_bb): Ditto.
15127 (copy_edges_for_bb): Ditto.
15128 (initialize_cfun): Ditto.
15129 (copy_cfg_body): Ditto.
15130 (expand_call_inline): Ditto.
15132 2013-04-08 Kai Tietz <ktietz@redhat.com>
15134 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
15135 TARGET_CYGWIN64 by TARGET_64BIT.
15137 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15139 * config/epiphany/epiphany.md (GPR_1): New constant.
15140 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
15141 * config/epiphany/epiphany.c (gen_compare_reg):
15142 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
15143 is already in place.
15144 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
15145 Don't require being called during rtl expansion; If y operlaps r0,
15147 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
15148 (epiphany_expand_epilogue): Likewise.
15150 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
15151 Don't use CC_FPmode for ORDERED / UNORDERED.
15152 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
15154 * config/epiphany/constraints.md (CnL): New constraint.
15155 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
15156 * config/epiphany/predicates.md (add_operand): Allow 1024.
15158 * config/epiphany/epiphany.md (logical_op): New code iterator.
15159 (op_mnc): New code attribute.
15160 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
15161 (mov_f+1, mov_f+2): New peephole2 patterns.
15163 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
15164 (cstoresi4): Also allow re-use of zero result when doing a NE
15165 comparison to a non-zero operand.
15166 Use (clobber (scratch)) for first insn if the gpr output is not needed.
15168 * config/epiphany/epiphany.md (<insn_opname>v2si3):
15169 Use gen_addsi3_i / gen_subsi3_i.
15171 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15175 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
15176 contain anything but clobbers, at most one __builtin_stack_restore,
15177 optionally debug stmts and final resx, and if it has at least one
15178 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
15179 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
15180 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
15181 which isn't defaut definition, remove them.
15182 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
15183 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
15184 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
15185 with MEM_REF LHS with SSA_NAME address.
15187 2013-04-08 Jeff Law <law@redhat.com>
15189 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
15191 2013-04-08 Richard Biener <rguenther@suse.de>
15193 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
15195 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
15196 determined vector type.
15197 (vect_analyze_data_refs): Likewise.
15198 (vect_get_new_vect_var): Adjust.
15199 (vect_create_destination_var): Preserve SSA name versions.
15200 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
15201 not dump anything here.
15203 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15205 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
15206 Add member lr_slot_known.
15207 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
15209 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15210 Remove code that sets lr_slot_offset according to what a previous
15211 version of epiphany_emit_save_restore used to do.
15212 (epiphany_emit_save_restore): When doing an lr save or restore,
15213 set/verify lr_slot_known and lr_slot_offset.
15215 2013-04-08 Xinyu Qi <xyqi@marvell.com>
15218 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
15221 2013-04-08 Richard Biener <rguenther@suse.de>
15223 * alias.c (find_base_term): Fix thinko in previous change.
15225 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15227 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
15228 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
15229 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
15230 if possible to compute val.
15231 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
15232 For QImode integers don't require anything about precision. Use
15233 const_with_all_bytes_same to find out if the constant doesn't have
15234 repeated bytes in it.
15236 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15238 * config/s390/s390.c (s390_expand_insv): Only accept insertions
15241 2013-04-08 Marek Polacek <polacek@redhat.com>
15243 PR rtl-optimization/48182
15244 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
15247 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15250 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
15251 nuses, make sure we have a label.
15253 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15256 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
15257 (rs6000_emit_swdiv_low_precision): Remove.
15258 (rs6000_emit_swdiv): Rewrite to handle between one and four
15259 iterations of Newton-Raphson generally; modify required number of
15260 iterations for some cases.
15261 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
15263 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
15265 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
15266 set-but-unused variable.
15268 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
15269 basic blocks of released function bodies garbage-collectable.
15271 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
15272 (struct rtl_opt_pass): Add TODO_df_finish.
15274 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
15276 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15278 * config/arm/constraints.md (q): New constraint.
15279 * config/arm/ldrdstrd.md: New file.
15280 * config/arm/arm.md (ldrdstrd.md) New include.
15281 (arm_movdi): Use "q" instead of "r" constraint
15282 for double-word memory access.
15283 (movdf_soft_insn): Likewise.
15284 * config/arm/vfp.md (movdi_vfp): Likewise.
15285 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
15286 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
15287 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
15288 (mem_ok_for_ldrd_strd): Likewise.
15289 (output_move_double): Update assertion.
15291 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15293 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
15295 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15297 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
15298 define_insn_and_split.
15299 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
15301 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15303 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
15304 define_insn_and_split.
15305 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
15306 (shiftsi3_compare): New pattern.
15307 (rrx): New pattern.
15308 * config/arm/unspecs.md (UNSPEC_RRX): New.
15310 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15312 * config/arm/arm.md (negdi_extendsidi): New pattern.
15313 (negdi_zero_extendsidi): Likewise.
15315 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15317 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
15318 define_insn_and_split.
15319 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
15320 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
15322 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15324 * config/arm/arm.md (arm_subdi3): Convert define_insn into
15325 define_insn_and_split.
15326 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
15327 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
15329 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15331 * config/arm/arm.md (subsi3_carryin): New pattern.
15332 (subsi3_carryin_const): Likewise.
15333 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
15334 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
15336 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15338 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
15340 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15342 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
15343 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
15345 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15347 * config/arm/arm.c (arm_expand_builtin): Change fcode
15348 type to unsigned int.
15350 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15352 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
15354 2013-04-04 Ian Lance Taylor <iant@google.com>
15356 * doc/standards.texi (Standards): The Go frontend supports the Go 1
15359 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
15361 PR middle-end/56729
15362 * df-scan.c (df_insn_delete): Disable failing assert.
15364 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15366 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
15367 New function prototype.
15368 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
15369 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
15370 (arm_builtin_vectorized_function): New function.
15372 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15374 * config/arm/arm_neon_builtins.def: New file.
15375 * config/arm/arm.c (neon_builtin_data): Move contents to
15376 arm_neon_builtins.def.
15377 (enum arm_builtins): Include neon builtin definitions.
15378 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
15379 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
15381 2013-04-04 Marek Polacek <polacek@redhat.com>
15383 PR tree-optimization/48186
15384 * predict.c (maybe_hot_frequency_p): Return false if
15385 HOT_BB_FREQUENCY_FRACTION is 0.
15386 (cgraph_maybe_hot_edge_p): Likewise.
15388 2013-04-04 Richard Biener <rguenther@suse.de>
15390 PR tree-optimization/56826
15391 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
15394 2013-04-04 Richard Biener <rguenther@suse.de>
15396 PR tree-optimization/56213
15397 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
15398 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
15400 2013-04-04 Richard Biener <rguenther@suse.de>
15402 PR tree-optimization/56837
15403 * tree-loop-distribution.c (classify_partition): For non-zero
15404 values require that the value has the same precision as its
15405 mode to be useful as memset value.
15407 2013-04-03 Nick Clifton <nickc@redhat.com>
15409 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
15410 (fmssf4): Use fmsf.s on E3V5 architectures.
15411 (fnmasf4): Use fnmaf.s on E3V5 architectures.
15412 (fnmssf4): Use fnmsf.s on E3V5 architectures.
15414 2013-04-03 Jeff Law <law@redhat.com>
15416 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
15417 (lra-eliminations.o): Likewise.
15419 2013-04-03 Teresa Johnson <tejohnson@google.com>
15421 * gcov-io.c (compute_working_sets): Moved most of body of old
15422 compute_working_sets here from profile.c.
15423 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
15424 (gcov_working_set_t): Moved typedef here from basic-block.h
15425 (compute_working_set): Declare.
15426 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
15427 (get_working_sets): Renamed from compute_working_set,
15428 replace most of body with call to new compute_working_sets.
15429 (get_exec_counts): Replace call to compute_working_sets
15430 to get_working_sets.
15431 * profile.h (get_working_sets): Renamed from compute_working_set.
15432 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
15433 to get_working_sets.
15434 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
15435 * gcov-dump.c (dump_working_sets): New function.
15437 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
15439 * hwint.c (sext_hwi, zext_hwi): New functions.
15440 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15441 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15442 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15443 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15444 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15445 (sext_hwi, zext_hwi): New functions.
15447 2013-04-03 Jeff Law <law@redhat.com>
15449 PR tree-optimization/56799
15450 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15451 back test for widening conversion erroneously dropped in prior change.
15453 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15456 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15457 instead of next_real_insn.
15459 2013-04-03 Marek Polacek <polacek@redhat.com>
15462 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
15464 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15467 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15469 (thumb1_output_casesi): Likewise.
15470 (thumb2_output_casesi): Likewise.
15472 2013-04-03 Richard Biener <rguenther@suse.de>
15474 PR tree-optimization/56817
15475 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15477 (tree_unroll_loops_completely_1): ... new function to manually
15478 walk the loop tree, properly defering outer loops of unrolled
15479 loops to later iterations.
15481 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15483 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15484 (vectorizable_load): Likewise.
15485 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15486 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15488 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15490 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15493 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15495 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15497 2013-04-03 Bin Cheng <bin.cheng@arm.com>
15499 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15501 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15503 PR tree-optimization/56790
15504 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15507 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15509 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15511 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15512 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15515 2013-04-03 Jakub Jelinek <jakub@redhat.com>
15518 * tree.h (force_folding_builtin_constant_p): New decl.
15519 * builtins.c (force_folding_builtin_constant_p): New variable.
15520 (fold_builtin_constant_p): Fold immediately also if
15521 force_folding_builtin_constant_p.
15523 2013-04-03 Richard Biener <rguenther@suse.de>
15525 PR tree-optimization/56812
15526 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15527 DRs of the same interleaving chain are independent.
15529 2013-04-02 Jason Merrill <jason@redhat.com>
15531 * gdbinit.in (pbb): Use debug fn.
15533 2013-04-02 Lawrence Crowl <crowl@google.com>
15535 * sese.h (struct ivtype_map_elt_s): Remove unused.
15536 (extern debug_ivtype_map): Remove unused.
15537 (extern eq_ivtype_map_elts): Remove unused.
15538 * sese.c (debug_ivtype_map): Removed unused.
15539 (debug_ivtype_map_1): Removed unused.
15540 (debug_ivtype_elt): Remove unused.
15541 (eq_ivtype_map_elts): Remove unused.
15544 2013-04-02 Kai Tietz <ktietz@redhat.com>
15547 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
15548 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
15549 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
15551 (legitimize_pe_coff_symbol): Likewise.
15552 (is_imported_p): New helper-function.
15553 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
15554 for Windows x64 targets.
15555 (ix86_expand_prologue): Optimize for pe-coff targets.
15556 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
15557 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
15558 medium/large code-model.
15559 (legitimize_pic_address): Likewise.
15560 (legitimize_tls_address): Likewise.
15561 (ix86_expand_call): Likewise.
15562 (x86_output_mi_thunk): Likewise.
15563 (get_dllimport_decl): Add new beimport argument.
15564 (construct_plt_address): Don't assert for x64 pe-coff targets.
15565 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
15567 (SYMBOL_FLAG_STUBVAR): New macro.
15568 (SYMBOL_REF_STUBVAR_P): Likewise.
15569 * config/i386/winnt.c (stub_list): New structure.
15570 (stub_head): New local variable.
15571 (i386_pe_record_stub): New function.
15572 (i386_pe_file_end): Emit refptr-stubs.
15574 2013-04-02 Jakub Jelinek <jakub@redhat.com>
15576 PR rtl-optimization/56745
15577 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
15578 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
15581 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
15582 and both of them are MEM_REFs, just compare first argument for
15583 equality and attempt to deal even with differing offsets.
15586 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
15587 of gimple_clobber_p to be MEM_REF.
15588 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
15589 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
15590 after gimplification.
15591 * asan.c (get_mem_ref_of_assignment): Don't instrument
15592 gimple_clobber_p stmts.
15593 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
15594 gimple_clobber_p stmt if they have MEM_REF lhs and
15595 are dead because of another gimple_clobber_p stmt.
15596 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15597 gimple_clobber_p stmts like debug stmts.
15598 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15599 that refer to unused VAR_DECLs or uninitialized values.
15600 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15601 gimple_clobber_p stmts if they refer to removed parameters.
15602 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15605 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
15607 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15608 using SWI48 mode attribute.
15610 2013-04-02 Wei Mi <wmi@google.com>
15612 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15613 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15614 *<rotate_insn><mode>3_mask in i386.md.
15616 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15618 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15620 2013-04-02 Richard Biener <rguenther@suse.de>
15622 PR tree-optimization/56778
15623 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15624 Runtime alias tests are not supported for gather loads.
15625 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15626 stmts referenced from SSA operands before updating SSA form.
15628 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15629 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15631 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
15632 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15633 * config/arm/cortex-a53.md: New file.
15634 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15635 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15636 * config/arm/arm.c (arm_issue_rate): Likewise.
15637 * config/arm/arm-tune.md: Regenerate
15638 * config/arm/arm-tables.opt: Regenerate.
15639 * config/arm/arm-cores.def: Add cortex-a53.
15641 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15643 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15646 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15648 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15649 scalar load/store operations using B/H registers.
15650 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15652 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15654 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15656 * config/aarch64/aarch64.c
15657 (aarch64_simd_scalar_immediate_valid_for_move): New.
15658 * config/aarch64/aarch64-protos.h
15659 (aarch64_simd_scalar_immediate_valid_for_move): New.
15660 * config/aarch64/constraints.md (Dh, Dq): New.
15661 * config/aarch64/iterators.md (hq): New.
15663 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15665 * reorg.c (get_branch_condition): Deal with conditional returns.
15666 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15668 2013-04-01 Wei Mi <wmi@google.com>
15670 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15671 Truncate operand 2 using %b asm operand modifier.
15672 (*<shift_insn><mode>3_mask): Ditto.
15673 (*<rotate_insn><mode>3_mask): Ditto.
15675 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15677 PR middle-end/56798
15678 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15680 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15682 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15684 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15686 2013-03-30 Lawrence Crowl <crowl@google.com>
15688 * dse.c (clear_alias_sets): Remove never set.
15689 (disqualified_clear_alias_sets): Remove never set.
15690 (clear_alias_mode_pool): Remove never set.
15691 (dse_step0): Remove condition that is never true.
15692 (canon_address): Remove condition that is never true.
15693 (dse_step7): Remove condition that is never true.
15694 (rest_of_handle_dse): Remove condition that is never true.
15695 (rest_of_handle_dse::did_global): Remove never read from above.
15696 (dse_step2_spill): Remove never called from above.
15697 (dse_step5_spill): Remove never called from above.
15699 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15701 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15702 JUMP_TABLE_DATA changes.
15703 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15704 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15705 (Insns) <jump_table_data>: New entry.
15706 * doc/tm.texi: Regenerate.
15708 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15710 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15711 for table jump at the end of a basic block using tablejump_p.
15712 * targhooks.c (default_invalid_within_doloop): Likewise.
15713 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15714 target hook implementation that is identical to the default hook.
15715 (rs6000_invalid_within_doloop): Remove.
15717 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15718 unused variable from tablejump_p call.
15720 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15721 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15722 (INSN_DELETED_P): Likewise.
15723 (emit_jump_table_data): New prototype.
15724 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15725 after 4th as unused.
15726 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15727 * sched-vis.c (print_insn): Likewise.
15728 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15729 insn for compatibility with back ends that use next_active_insn to
15730 identify jump table data.
15731 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15732 (remove_insn): Likewise.
15733 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15735 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15736 (emit_jump_table_data): New function.
15738 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15739 basic block, a JUMP_TABLE_DATA never is.
15740 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15741 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15742 off from code handling real insns.
15743 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15744 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15745 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15746 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15747 is not a NONDEBUG_INSN_P.
15748 * ira-costs.c (scan_one_insn): Likewise.
15749 * jump.c (mark_all_labels): Likewise.
15750 (mark_jump_label_1): Likewise.
15751 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15752 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15754 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15755 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15756 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15757 JUMP_TABLE_DATA_P insns.
15758 (calculate_elim_costs_all_insns): Likewise.
15759 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15760 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15761 (delete_output_reload): Code style fixups.
15762 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15763 insn flags on this non-insn.
15764 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15765 as scheduling barriers, for pre-change compatibility.
15766 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15767 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15769 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15770 redundant JUMP_TABLE_DATA_P test.
15771 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15772 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15773 (frv_for_each_packet): Likewise.
15774 * config/i386/i386.c (min_insn_size): Likewise.
15775 (ix86_avoid_jump_mispredicts): Likewise.
15776 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15777 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15778 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15779 (mips16_insn_length): Robustify.
15780 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15781 (mips16_split_long_branches): Likewise.
15782 * config/pa/pa.c (pa_combine_instructions): Likewise.
15783 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15784 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15785 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15786 as contributing to pool range lengths.
15787 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15788 Remove redundant JUMP_TABLE_DATA_P test.
15789 (sh_loop_align): Likewise.
15790 (split_branches): Likewise.
15791 (sh_insn_length_adjustment): Likewise.
15792 * config/spu/spu.c (get_branch_target): Likewise.
15794 2013-03-29 Jan Hubicka <jh@suse.cz>
15796 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
15797 gcov streaming; stream hot bb threshold to ltrans.
15798 * predict.c (get_hot_bb_threshold): Break out from ....
15799 (maybe_hot_count_p): ... here.
15800 (set_hot_bb_threshold): New function.
15801 * lto-section-in.c (lto_section_name): Add profile.
15802 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
15803 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
15804 and data-streamer.h
15805 (histogram_entry): New structure.
15806 (histogram, histogram_pool): New global vars.
15807 (histogram_hash): New structure.
15808 (histogram_hash::hash): New method.
15809 (histogram_hash::equal): Likewise.
15810 (account_time_size): New function.
15811 (cmp_counts): New function.
15812 (dump_histogram): New function.
15813 (ipa_profile_generate_summary): New function.
15814 (ipa_profile_write_summary): New function.
15815 (ipa_profile_read_summary): New function.
15816 (ipa_profile): Decide on threshold.
15817 (pass_ipa_profile): Add ipa_profile_write_summary and
15818 ipa_profile_read_summary.
15819 * Makefile.in (ipa.o): Update dependencies.
15820 * lto-streamer.h (LTO_section_ipa_profile): New section.
15822 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
15824 * tree.h (VAR_P): New.
15826 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
15829 * doc/invoke.texi ([-fwhole-program]): Fix typo.
15831 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
15833 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
15834 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15835 (control_flow_insn_p): Likewise.
15836 * cfgrtl.c (duplicate_insn_chain): Likewise.
15837 * final.c (get_attr_length_1): Likewise.
15838 (shorten_branches): Likewise.
15839 (final_scan_insn): Likewise.
15840 * function.c (instantiate_virtual_regs): Likewise.
15841 * gcse.c (insert_insn_end_basic_block): Likewise.
15842 * ira-costs.c (scan_one_insn): Likewise.
15843 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15844 * lra.c (check_rtl): Likewise.
15845 * reload1.c (elimination_costs_in_insn): Likewise.
15846 * reorg.c (follow_jumps): Likewise.
15848 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
15849 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15850 (thumb_far_jump_used_p): Likewise.
15851 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
15852 (workaround_speculation): Likewise.
15853 (add_sched_insns_for_speculation): Likewise.
15854 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
15855 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15856 (frv_for_each_packet): Likewise.
15857 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
15858 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
15859 (final_emit_insn_group_barriers): Likewise.
15860 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15861 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15862 (mips16_insn_length): Likewise.
15863 * config/pa/pa.c (pa_reorg): Likewise.
15864 (pa_combine_instructions): Likewise.
15865 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
15866 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
15867 (sh_reorg): Likewise.
15868 (split_branches): Likewise.
15869 * config/spu/spu.c (get_branch_target): Likewise.
15871 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
15874 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
15876 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
15877 Fix declaration name.
15879 2013-03-28 Lawrence Crowl <crowl@google.com>
15881 * graphds.h (struct graph.indicies): Remove unused.
15882 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
15883 (SCOP_ORIGINAL_PDDRS): Remove unused.
15884 * sese.h (extern insert_loop_close_phis): Removed unused.
15885 (extern insert_guard_phis): Removed unused.
15886 (extern ivtype_map_elt_info): Removed unused.
15887 (new_ivtype_map_elt): Removed unused.
15888 * sese.c (ivtype_map_elt_info): Removed unused.
15890 2013-03-28 Lawrence Crowl <crowl@google.com>
15892 * Makefile.in: Add several missing include dependences.
15894 (test-dump.o): New. This object is not added to any executable,
15895 but is present for ad-hoc testing.
15897 (debug (const bitmap_head_def &)): New.
15898 (debug (const bitmap_head_def *)): New.
15900 (extern debug (const bitmap_head_def &)): New.
15901 (extern debug (const bitmap_head_def *)): New.
15903 (debug (edge_def &)): New.
15904 (debug (edge_def *)): New.
15906 (debug (basic_block_def &)): New.
15907 (debug (basic_block_def *)): New.
15909 (dump_node (const_tree, int, FILE *)): Correct source file.
15911 (debug (die_struct &)): New.
15912 (debug (die_struct *)): New.
15914 (extern debug (die_struct &)): New.
15915 (extern debug (die_struct *)): New.
15916 * gimple-pretty-print.c
15917 (debug (gimple_statement_d &)): New.
15918 (debug (gimple_statement_d *)): New.
15919 * gimple-pretty-print.h
15920 (extern debug (gimple_statement_d &)): New.
15921 (extern debug (gimple_statement_d *)): New.
15923 (debug (ira_allocno_copy &)): New.
15924 (debug (ira_allocno_copy *)): New.
15925 (debug (ira_allocno &)): New.
15926 (debug (ira_allocno *)): New.
15928 (extern debug (ira_allocno_copy &)): New.
15929 (extern debug (ira_allocno_copy *)): New.
15930 (extern debug (ira_allocno &)): New.
15931 (extern debug (ira_allocno *)): New.
15933 (debug (live_range &)): New.
15934 (debug (live_range *)): New.
15936 (debug (lra_live_range &)): New.
15937 (debug (lra_live_range *)): New.
15939 (debug (lra_live_range &)): New.
15940 (debug (lra_live_range *)): New.
15942 (debug (omega_pb_d &)): New.
15943 (debug (omega_pb_d *)): New.
15945 (extern debug (omega_pb_d &)): New.
15946 (extern debug (omega_pb_d *)): New.
15948 (debug (const rtx_def &)): New.
15949 (debug (const rtx_def *)): New.
15951 (debug_tree (tree): Move within file.
15952 (debug_raw (const tree_node &)): New.
15953 (debug_raw (const tree_node *)): New.
15954 (dump_tree_via_hooks (const tree_node *, int)): New.
15955 (debug (const tree_node &)): New.
15956 (debug (const tree_node *)): New.
15957 (debug_verbose (const tree_node &)): New.
15958 (debug_verbose (const tree_node *)): New.
15959 (debug_head (const tree_node &)): New.
15960 (debug_head (const tree_node *)): New.
15961 (debug_body (const tree_node &)): New.
15962 (debug_body (const tree_node *)): New.
15963 (debug_vec_tree (tree): Move and reimplement in terms of dump.
15964 (debug (vec<tree, va_gc> &)): New.
15965 (debug (vec<tree, va_gc> *)): New.
15967 (extern debug (const rtx_def &)): New.
15968 (extern debug (const rtx_def *)): New.
15970 (debug_raw (simple_bitmap_def &)): New.
15971 (debug_raw (simple_bitmap_def *)): New.
15972 (debug (simple_bitmap_def &)): New.
15973 (debug (simple_bitmap_def *)): New.
15975 (extern debug (simple_bitmap_def &)): New.
15976 (extern debug (simple_bitmap_def *)): New.
15977 (extern debug_raw (simple_bitmap_def &)): New.
15978 (extern debug_raw (simple_bitmap_def *)): New.
15980 (debug (vinsn_def &)): New.
15981 (debug (vinsn_def *)): New.
15982 (debug_verbose (vinsn_def &)): New.
15983 (debug_verbose (vinsn_def *)): New.
15984 (debug (expr_def &)): New.
15985 (debug (expr_def *)): New.
15986 (debug_verbose (expr_def &)): New.
15987 (debug_verbose (expr_def *)): New.
15988 (debug (vec<rtx> &)): New.
15989 (debug (vec<rtx> *)): New.
15991 (extern debug (vinsn_def &)): New.
15992 (extern debug (vinsn_def *)): New.
15993 (extern debug_verbose (vinsn_def &)): New.
15994 (extern debug_verbose (vinsn_def *)): New.
15995 (extern debug (expr_def &)): New.
15996 (extern debug (expr_def *)): New.
15997 (extern debug_verbose (expr_def &)): New.
15998 (extern debug_verbose (expr_def *)): New.
15999 (extern debug (vec<rtx> &)): New.
16000 (extern debug (vec<rtx> *)): New.
16002 (_list_iter_cond_expr): Make inline instead of static.
16004 (debug (sreal &)): New.
16005 (debug (sreal *)): New.
16007 (extern debug (sreal &)): New.
16008 (extern debug (sreal *)): New.
16010 (extern debug_raw (const tree_node &)): New.
16011 (extern debug_raw (const tree_node *)): New.
16012 (extern debug (const tree_node &)): New.
16013 (extern debug (const tree_node *)): New.
16014 (extern debug_verbose (const tree_node &)): New.
16015 (extern debug_verbose (const tree_node *)): New.
16016 (extern debug_head (const tree_node &)): New.
16017 (extern debug_head (const tree_node *)): New.
16018 (extern debug_body (const tree_node &)): New.
16019 (extern debug_body (const tree_node *)): New.
16020 (extern debug (vec<tree, va_gc> &)): New.
16021 (extern debug (vec<tree, va_gc> *)): New.
16023 (debug (struct loop &)): New.
16024 (debug (struct loop *)): New.
16025 (debug_verbose (struct loop &)): New.
16026 (debug_verbose (struct loop *)): New.
16027 * tree-dump.c: Add header dependence.
16029 (extern debug (struct loop &)): New.
16030 (extern debug (struct loop *)): New.
16031 (extern debug_verbose (struct loop &)): New.
16032 (extern debug_verbose (struct loop *)): New.
16034 (debug (data_reference &)): New.
16035 (debug (data_reference *)): New.
16036 (debug (vec<data_reference_p> &)): New.
16037 (debug (vec<data_reference_p> *)): New.
16038 (debug (vec<ddr_p> &)): New.
16039 (debug (vec<ddr_p> *)): New.
16041 (extern debug (data_reference &)): New.
16042 (extern debug (data_reference *)): New.
16043 (extern debug (vec<data_reference_p> &)): New.
16044 (extern debug (vec<data_reference_p> *)): New.
16045 (extern debug (vec<ddr_p> &)): New.
16046 (extern debug (vec<ddr_p> *)): New.
16048 (debug (pt_solution &)): New.
16049 (debug (pt_solution *)): New.
16051 (extern debug (pt_solution &)): New.
16052 (extern debug (pt_solution *)): New.
16054 (debug (_var_map &)): New.
16055 (debug (_var_map *)): New.
16056 (debug (tree_live_info_d &)): New.
16057 (debug (tree_live_info_d *)): New.
16059 (extern debug (_var_map &)): New.
16060 (extern debug (_var_map *)): New.
16061 (extern debug (tree_live_info_d &)): New.
16062 (extern debug (tree_live_info_d *)): New.
16064 2013-03-28 Jan Hubicka <jh@suse.cz>
16066 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
16068 2013-03-28 Ian Bolton <ian.bolton@arm.com>
16070 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
16071 record only when desired or required.
16073 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
16075 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
16076 *vec_extractv2di_1_rex64. Use x64 isa attribute.
16078 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16080 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
16081 (*andsi3_compare0_uxtw): New pattern.
16082 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
16083 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
16085 2013-03-28 Jan Hubicka <jh@suse.cz>
16087 * data-streamer-in.c (streamer_read_gcov_count): New function.
16088 * gimple-streamer-out.c: Include value-prof.h.
16089 (output_gimple_stmt): Output histogram.
16090 (output_bb): Use streamer_write_gcov_count.
16091 * value-prof.c: Include data-streamer.h
16092 (dump_histogram_value): Add HIST_TYPE_MAX.
16093 (stream_out_histogram_value): New function.
16094 (stream_in_histogram_value): New function.
16095 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
16096 (stream_out_histogram_value, stream_in_histogram_value): Declare.
16097 * data-streamer-out.c (streamer_write_gcov_count): New function.
16098 (streamer_write_gcov_count_stream): New function.
16099 * lto-cgraph.c (lto_output_edge): Update counter streaming.
16100 (lto_output_node): Likewise.
16101 (input_node, input_edge): Likewise.
16102 * lto-streamer-out.c (output_cfg): Update streaming.
16103 * lto-streamer-in.c (input_cfg): Likewise.
16104 * data-streamer.h (streamer_write_gcov_count,
16105 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
16106 * gimple-streamer-in.c: Include value-prof.h
16107 (input_gimple_stmt): Input histograms.
16108 (input_bb): Update profile streaming.
16110 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16112 * genmodes.c (emit_max_int): New function.
16113 (emit_insn_modes_h): Added call to emit_max_function.
16114 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
16116 * machmode.def: Fixed comment.
16118 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16120 * combine.c (try_combine): Removed useless assert.
16121 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
16123 2013-03-28 Marek Polacek <polacek@redhat.com>
16124 Richard Biener <rguenther@suse.de>
16126 PR tree-optimization/56695
16127 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
16128 build signed result of a vector comparison.
16129 * tree-cfg.c (verify_gimple_comparison): Check that a result
16130 of a vector comparison has signed type.
16132 2013-03-28 Richard Biener <rguenther@suse.de>
16134 PR tree-optimization/37021
16135 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
16136 do not restrict gaps between groups.
16137 * tree-vect-stmts.c (vectorizable_load): Properly account for
16138 a gap between groups.
16140 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
16142 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
16143 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
16146 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
16148 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
16149 * doc/extend.texi (Named Address Spaces): Ditto.
16150 (Variable Attributes): Ditto.
16152 2013-03-27 Kai Tietz <ktietz@redhat.com>
16154 * config.build: Add support for cygwin x64 target.
16155 * config.gcc: Likewise.
16156 * config.host: Likewise.
16157 * configure.ac: Likewise
16158 * configure: Regenerated.
16160 2013-03-27 Kai Tietz <ktietz@redhat.com>
16162 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
16163 * config/i386/t-cygwin-w64: New file.
16164 * config/i386/cygwin-w64.h: New file.
16165 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
16166 and add support for x64-cygwin target.
16167 (CPP_SPEC): Likewise.
16168 (CXX_WRAP_SPEC_LIST): Undefine before define.
16169 (LIBGCJ_SONAME): Use 15 as version.
16171 2013-03-27 Richard Biener <rguenther@suse.de>
16173 PR tree-optimization/56716
16174 * tree-ssa-structalias.c (perform_var_substitution): Adjust
16175 dumping for ref nodes.
16177 2013-03-27 Martin Jambor <mjambor@suse.cz>
16179 PR tree-optimization/55334
16180 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
16181 restricted pointers to arrays.
16183 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
16185 * Makefile.in (.SUFFIXES): Add .cc.
16186 (.c.o): Apply same recipe for implicit rule .cc.o.
16188 2013-03-27 Richard Biener <rguenther@suse.de>
16190 PR tree-optimization/37021
16191 * tree-vect-data-refs.c (vect_check_strided_load): Allow
16192 REALPART/IMAGPART_EXPRs around the supported refs.
16193 * tree-ssa-structalias.c (find_func_aliases): Assume that
16194 floating-point values are not used to transfer pointers.
16196 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
16198 * target.def (TARGET_HAS_IFUNC_P): New target hook.
16199 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
16200 * doc/tm.texi: Regenerate.
16201 * targhooks.h (default_has_ifunc_p): New.
16202 * targhooks.c (default_has_ifunc_p): Ditto.
16203 * config/linux-protos.h: New file.
16204 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
16205 hook for linux which disables support of indirect functions in android.
16206 * config/linux-android.c: New file.
16207 * config/t-linux-android.c: Ditto.
16208 * config.gcc: Added new object file linux-android.o.
16209 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
16210 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
16211 * varasm.c (do_assemble_alias): Likewise.
16212 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
16213 doesn't support indirect functions.
16214 * configure: Regenerate.
16216 2013-03-27 Bin Cheng <bin.cheng@arm.com>
16219 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
16220 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
16223 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16225 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
16227 2013-03-27 Terry Guo <terry.guo@arm.com>
16229 * config/arm/arm-cores.def: Added core cortex-r7.
16230 * config/arm/arm-tune.md: Regenerated.
16231 * config/arm/arm-tables.opt: Regenerated.
16232 * doc/invoke.texi: Added entry for core cortex-r7.
16234 2013-03-27 Walter Lee <walt@tilera.com>
16236 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
16237 double-decrement of next_scratch_regno.
16239 2013-03-27 Walter Lee <walt@tilera.com>
16241 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
16243 (insn_v1mulus): Ditto.
16244 (insn_v2muls): Ditto.
16246 2013-03-27 Walter Lee <walt@tilera.com>
16248 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
16249 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
16251 2013-03-27 Walter Lee <walt@tilera.com>
16253 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
16254 (*sibcall_value): Ditto.
16256 2013-03-27 Walter Lee <walt@tilera.com>
16258 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
16259 (insn_mnz_v8qi): ... this ...
16260 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
16262 (insn_v<n>mnz): Replaced by ...
16263 (insn_v1mnz): ... this ...
16264 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
16266 (insn_mz_<mode>): Replaced by ...
16267 (insn_mz_v8qi): ... this ...
16268 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
16270 (insn_v<n>mz): Replaced by ...
16271 (insn_v1mz): ... this ...
16272 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
16275 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
16277 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
16279 2013-03-26 Roland McGrath <mcgrathr@google.com>
16281 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
16282 than fprintf with a non-constant, non-format string.
16284 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16286 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
16287 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
16288 operand 0 predicate.
16289 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
16290 attribute. Use general_x64nomem_operand as operand 1 predicate.
16291 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
16292 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16293 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
16294 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16295 (mov<mode>_insv_1): Remove expander. Merge insn with
16296 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
16297 Use general_x64nomem_operand as operand 1 predicate.
16298 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
16299 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
16300 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16301 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
16302 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
16303 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
16304 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16305 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
16306 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
16307 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
16308 (general_x64nomem_operand): Ditto.
16310 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
16312 * config/rtems.opt: Add -pthread option.
16314 2013-03-26 Richard Biener <rguenther@suse.de>
16316 * alias.c (find_base_term): Avoid redundant and not used recursion.
16317 (base_alias_check): Get the initial base term from the caller.
16318 (true_dependence_1): Compute and pass base terms to base_alias_check.
16319 (write_dependence_p): Likewise.
16320 (may_alias_p): Likewise.
16322 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
16324 * config/aarch64/aarch64.c (aarch64_classify_address): Support
16325 PC-relative load in SI modes and above only.
16327 2013-03-26 Xinyu Qi <xyqi@marvell.com>
16329 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
16330 * config/arm/iwmmxt.md (WCGR0): Update.
16331 (WCGR1, WCGR2, WCGR3): Likewise.
16333 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16335 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
16336 Use x64 and nox64 isa attributes.
16338 2013-03-26 Richard Biener <rguenther@suse.de>
16340 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
16341 alignment computations and rely on get_object_alignment_1
16342 for the !TYPE_P case.
16343 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
16345 2013-03-26 Walter Lee <walt@tilera.com>
16347 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
16348 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
16350 2013-03-25 Jeff Law <law@redhat.com>
16352 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
16353 check for INTEGRAL_TYPE_P that was missing due to checking in
16354 wrong version of prior patch.
16356 2013-03-25 Walter Lee <walt@tilera.com>
16358 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
16359 TILEGX_INSN_SHUFFLEBYTES1.
16360 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
16362 (tilegx_builtins): Ditto.
16363 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
16365 2013-03-25 Walter Lee <walt@tilera.com>
16367 * config/tilegx/tilegx.md (floatsisf2): New pattern.
16368 (floatunssisf2): New pattern.
16369 (floatsidf2): New pattern.
16370 (floatunssidf2): New pattern.
16372 2013-03-25 Walter Lee <walt@tilera.com>
16374 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
16375 tests for constraint J, K, N, P.
16377 2013-03-25 Walter Lee <walt@tilera.com>
16379 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
16380 Use indirect/pcrel encoding.
16381 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
16384 2013-03-25 Steve Ellcey <sellcey@mips.com>
16386 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
16387 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
16388 * config/mips/mips.c (mips_option_override): Set IMADD default.
16389 * config/mips/mips.h (PTF_AVOID_IMADD): New.
16390 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
16391 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
16392 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
16393 * doc/invoke.texi (-mimadd/-mno-imadd): New.
16395 2013-03-25 Jeff Law <law@redhat.com>
16397 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
16398 slightly to avoid creating and folding useless trees. Simplify
16399 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
16401 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
16403 * config/i386/i386.md (*zero_extendsidi2): Merge with
16404 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
16405 * config/i386/predicates.md (x86_64_zext_operand): Rename from
16406 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
16407 targets. Clarify comment.
16409 2013-03-25 Martin Jambor <mjambor@suse.cz>
16411 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
16412 pass-through jump functions differently.
16413 (ipa_read_jump_function): Likewise. Also use setter functions to set
16416 2013-03-25 Martin Jambor <mjambor@suse.cz>
16418 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
16419 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
16421 (ipa_get_indirect_edge_target): New function.
16422 (devirtualization_time_bonus): New parameter known_aggs, pass it to
16423 ipa_get_indirect_edge_target. Update all callers.
16424 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
16425 ipa_get_indirect_edge_target_1 instead of calling
16426 ipa_get_indirect_edge_target.
16427 (create_specialized_node): Pass aggvlas to
16428 ipcp_discover_new_direct_edges.
16430 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16432 * config/arm/arm.md (f_sels, f_seld): New types.
16433 (*cmov<mode>): New pattern.
16434 * config/arm/predicates.md (arm_vsel_comparison_operator): New
16437 2013-03-25 Kai Tietz <ktietz@redhat.com>
16439 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16440 POSIX-printf for mingw-hosted builds.
16442 2013-03-25 Richard Biener <rguenther@suse.de>
16444 PR middle-end/56694
16445 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16446 must-not-throw stmt location.
16448 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16450 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16451 Emit load-acquire versions when acq is true.
16452 (arm_emit_store_exclusive): Add rel parameter.
16453 Emit store-release versions when rel is true.
16454 (arm_split_compare_and_swap): Use acquire-release instructions
16456 of barriers when appropriate.
16457 (arm_split_atomic_op): Likewise.
16458 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16459 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16460 (VUNSPEC_SLX): Likewise.
16461 (VUNSPEC_LDA): Likewise.
16462 (VUNSPEC_STL): Likewise.
16463 * config/arm/sync.md (atomic_load<mode>): New pattern.
16464 (atomic_store<mode>): Likewise.
16465 (arm_load_acquire_exclusive<mode>): Likewise.
16466 (arm_load_acquire_exclusivesi): Likewise.
16467 (arm_load_acquire_exclusivedi): Likewise.
16468 (arm_store_release_exclusive<mode>): Likewise.
16470 2013-03-25 Catherine Moore <clm@codesourcery.com>
16472 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16473 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16474 * config/mip/predicates.md (lwsp_swsp_operand,
16475 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16476 sb16_operand, db4_operand, db7_operand, ib3_operand,
16477 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16478 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16479 andi16_operand): New predicates.
16480 * config/mips/mips.md (compression): New attribute.
16481 (enabled): New attribute.
16482 (length): Consider compression in computing length.
16483 (shift_compression): New code attribute.
16484 (*add<mode>3): New operands. Record compression.
16485 (sub<mode>3): Likewise.
16486 (one_cmpl<mode>2): Likewise.
16487 (*and<mode>3): Likewise.
16488 (*ior<mode>3): Likewise.
16489 (unnamed pattern for xor): Likewise.
16490 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16491 (*<optab><mode>3): Likewise.
16492 (*mov<mode>_internal: Likewise.
16493 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16494 (mips_unsigned_immediate_p): New.
16495 (umips_lwsp_swsp_address_p): New.
16496 (m16_based_address_p): New.
16497 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16498 (mips_unsigned_immediate_p): New prototype.
16499 (lwsp_swsp_address_p): New prototype.
16500 (m16_based_address_p): New prototype.
16501 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16502 (mips_signed_immediate_p): New function.
16503 (m16_based_address_p): New function.
16504 (lwsp_swsp_address_p): New function.
16505 (mips_print_operand_punctuation): Recognize short delay slot insns
16506 for microMIPS.add<mode>3"
16508 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16511 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16512 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16514 2013-03-25 Richard Biener <rguenther@suse.de>
16516 PR tree-optimization/56689
16517 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16520 2013-03-25 Richard Biener <rguenther@suse.de>
16522 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16524 (memory_references): Likewise.
16525 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16526 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16527 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16528 (gather_mem_refs_in_loops): Fold into ...
16529 (analyze_memory_references): ... this. Move initialization
16530 to tree_ssa_lim_initialize.
16531 (fill_always_executed_in): Rename to ...
16532 (fill_always_executed_in_1): ... this.
16533 (fill_always_executed_in): Move contains_call computation to
16534 this new function from ...
16535 (tree_ssa_lim_initialize): ... here.
16536 (tree_ssa_lim): Call fill_always_executed_in.
16538 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16540 * postreload.c (reload_combine): Fix code detecting returns.
16542 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16544 * function.c (emit_use_return_register_into_block): On cc0 targets,
16545 do not emit the sequence between cc0 setter and user.
16547 2013-03-25 Kai Tietz <ktietz@redhat.com>
16549 * config/i386/predicates.md (local_symbolic_operand): Interpret
16550 dll-imported symbols as none-local.
16552 2013-03-25 Richard Biener <rguenther@suse.de>
16554 * tree-ssa-loop-im.c (struct depend): Remove.
16555 (struct lim_aux_data): Make depends a vec of gimples.
16556 (free_lim_aux_data): Adjust.
16557 (add_dependency): Likewise.
16558 (set_level): Likewise.
16560 2013-03-25 Richard Biener <rguenther@suse.de>
16562 PR middle-end/56434
16563 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
16564 the pointer returned by calls with ECF_MALLOC set.
16566 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16568 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
16570 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16572 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
16573 using MMXMODE mode iterator.
16574 (*move<mode>_internal): Merge with *movv2sf_internal and
16575 *movv2sf_internal_rex64 using MMXMODE mode iterator.
16577 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16579 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
16580 (record_last_mem_set_info): Likewise.
16582 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
16583 of XNEWVEC followed by memset.
16584 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
16586 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16588 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
16589 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
16590 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
16591 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
16592 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
16593 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
16594 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16595 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16596 BARRIER_P instead of GET_CODE.
16598 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16600 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16601 inaccuracy in the probing code.
16603 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16604 (ctrapdi4): Likewise.
16606 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16608 * calls.c (expand_call): Add missing guard to code handling return
16609 of non-BLKmode structures in MSB.
16610 * function.c (expand_function_end): Likewise.
16612 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16614 * combine.c (try_combine): Adjust comment. Do not add the set of
16615 insn #0 if the destination indirectly is set or dies in insn #2.
16616 Tidy up code to distribute a new note.
16618 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16620 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16621 also for alternatives 16 and 17.
16623 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16625 * config/i386/sse.md (*mov<mode>_internal): Merge with
16626 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16627 Emit insn template depending on type attribute. Use
16628 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16629 movd instead of movq mnemonic for interunit moves. Rewrite mode
16630 attribute calculation. Remove unit attribute calculation.
16631 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16632 Set prefix_data16 attribute for DImode ssemov types.
16633 Use Ym instead of y for SSE-MMX conversion alternatives.
16634 Reorder operand constraints.
16636 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16638 * df.h (df_insn_delete): Adjust prototype.
16639 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16640 and let it decide whether mark the basic block dirty.
16641 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16642 * df-scan.c (df_insn_info_delete): New helper function, split
16643 off from df_insn_delete.
16644 (df_scan_free_bb_info): Use it.
16645 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16647 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16648 that the insn is actually an insn and it has a non-NULL basic block.
16649 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16651 2013-03-22 Richard Biener <rguenther@suse.de>
16653 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16655 (mem_ref_alloc): Do not allocate them.
16656 (refs_independent_p): Do not query or maintain a cache.
16658 2013-03-22 Richard Biener <rguenther@suse.de>
16660 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16661 (gather_mem_refs_in_loops): Do not compute it.
16662 (analyze_memory_references): Do not allocate it.
16663 (tree_ssa_lim_finalize): Do not free it.
16664 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16666 2013-03-22 Richard Biener <rguenther@suse.de>
16668 * is-a.h (as_a): Use gcc_checking_assert.
16670 2013-03-22 Ian Bolton <ian.bolton@arm.com>
16672 * config/aarch64/aarch64.c (aarch64_print_operand): New
16673 format specifier for printing a constant in hex.
16674 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16675 format specifier for printing second operand.
16677 2013-03-22 Richard Biener <rguenther@suse.de>
16679 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16681 (gather_mem_refs_in_loops): Perform store accumulation here.
16682 (create_vop_ref_mapping_loop): Remove.
16683 (create_vop_ref_mapping): Likewise.
16684 (analyze_memory_references): Initialize refs_stored_in_loop.
16685 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
16686 (record_indep_loop): Remove.
16687 (record_dep_loop): New function.
16688 (ref_indep_loop_p_1): Adjust to only walk over references
16689 in the loop, not its subloops.
16690 (ref_indep_loop_p): Rename to ...
16691 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16692 maintaining a more fine-grained cache.
16693 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16694 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16696 2013-03-22 Richard Biener <rguenther@suse.de>
16698 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16699 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16700 aggregate mem_ref_loc.
16701 (free_mem_ref_locs): Inline into ...
16702 (memref_free): ... this and adjust.
16703 (mem_ref_alloc): Adjust.
16704 (mem_ref_locs_alloc): Remove.
16705 (record_mem_ref_loc): Adjust.
16706 (get_all_locs_in_loop): Rewrite into ...
16707 (for_all_locs_in_loop): ... this iterator.
16708 (rewrite_mem_ref_loc): New functor.
16709 (rewrite_mem_refs): Use for_all_locs_in_loop.
16710 (sm_set_flag_if_changed): New functor.
16711 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16712 (ref_always_accessed): New functor.
16713 (ref_always_accessed_p): Use for_all_locs_in_loop.
16715 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16717 * tree-pass.h (PROP_gimple_lvec): New.
16718 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16719 (init_optimization_passes): Move pass_lower_vector.
16720 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16722 (pass_lower_vector): Provide PROP_gimple_lvec.
16723 (pass_lower_vector_ssa): Likewise.
16724 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16726 2013-03-21 Mark Wielaard <mjw@redhat.com>
16728 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16730 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16732 * config/i386/i386.md (*movdi_internal): Disparage slightly
16733 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16734 conversion alternatives.
16736 2013-03-21 Jakub Jelinek <jakub@redhat.com>
16738 PR middle-end/48087
16739 * diagnostic.def (DK_WERROR): New kind.
16740 * diagnostic.h (werrorcount): Define.
16741 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16742 promoted to DK_ERROR, increment DK_WERROR counter instead of
16744 * toplev.c (toplev_main): Call print_ignored_options even if
16745 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16746 even if just werrorcount is non-zero.
16749 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16751 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16752 (string_cst_pool_decl): New function.
16753 (optimize_one_addr_into_implicit_ptr): New function.
16754 (resolve_addr_in_expr): Optimize DWARF location expression
16755 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16756 which doesn't live in memory, but has DW_AT_location or
16757 DW_AT_const_value, or refers to a string literal, into
16758 DW_OP_GNU_implicit_pointer.
16759 (optimize_location_into_implicit_ptr): New function.
16760 (resolve_addr): If removing DW_AT_location of a variable because
16761 it was DW_OP_addr of address of the variable, but the variable doesn't
16762 live in memory, try to emit const value attribute for the initializer.
16764 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16766 * tree.h (VECTOR_TYPE_P): New macro.
16767 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16768 TYPE_MODE): Use it.
16769 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16770 VEC_COND_EXPR cannot be lvalues.
16771 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16773 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16775 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16776 Restrict the transformation to equal modes.
16778 2013-03-21 Richard Biener <rguenther@suse.de>
16780 PR tree-optimization/39326
16781 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16782 (MEM_ANALYZABLE): Adjust.
16783 (record_mem_ref_loc): Move bitmap ops ...
16784 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16785 unanalyzable refs, do not record locations for it.
16786 (analyze_memory_references): Allocate ref zero as shared
16788 (refs_independent_p): Do not test for unanalyzed mems here.
16789 (ref_indep_loop_p_1): Special-case disambiguation against
16790 the unanalyzed ref.
16791 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
16793 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
16795 * config/arm/arm-protos.h (tune_params): Add
16796 prefer_neon_for_64bits field.
16797 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
16798 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
16799 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
16800 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
16801 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
16802 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
16803 (arm_option_override): Handle -mneon-for-64bits new option.
16804 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
16805 (prefer_neon_for_64bits): Declare new variable.
16806 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
16807 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
16808 (arch_enabled): Handle new arch types. Remove support for onlya8
16810 (one_cmpldi2): Use new arch names.
16811 (zero_extend<mode>di2, extend<mode>di2): Ditto.
16812 * config/arm/arm.opt (mneon-for-64bits): Add option.
16813 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
16814 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
16815 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
16817 * doc/invoke.texi (-mneon-for-64bits): Document.
16819 2013-03-21 Richard Biener <rguenther@suse.de>
16821 PR tree-optimization/39326
16822 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
16823 (sort_bbs_in_loop_postorder_cmp): New function.
16824 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
16826 2013-03-21 Richard Biener <rguenther@suse.de>
16828 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
16829 (vect_insert_into_interleaving_chain): Likewise.
16830 (vect_drs_dependent_in_basic_block): Inline ...
16831 (vect_slp_analyze_data_ref_dependence): ... here. New function,
16833 (vect_analyze_data_ref_dependence): ... here. Simplify.
16834 (vect_check_interleaving): Simplify.
16835 (vect_analyze_data_ref_dependences): Likewise. Split out ...
16836 (vect_slp_analyze_data_ref_dependences): ... this new function.
16837 (dr_group_sort_cmp): New function.
16838 (vect_analyze_data_ref_accesses): Compute data-reference groups
16839 here instead of in vect_analyze_data_ref_dependence. Use
16840 a more efficient algorithm.
16841 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
16842 vect_slp_analyze_data_ref_dependences. Call
16843 vect_analyze_data_ref_accesses earlier.
16844 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
16845 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
16846 (vect_slp_analyze_data_ref_dependences): New prototype.
16848 2013-03-21 Richard Biener <rguenther@suse.de>
16850 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
16851 ref is stored in the loop.
16852 (find_refs_for_sm): Walk only over all stores.
16853 (store_motion_loop): Allocate from lim_bitmap_obstack.
16854 (store_motion): Likewise.
16856 2013-03-21 Richard Biener <rguenther@suse.de>
16858 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
16859 Update virtual SSA form.
16861 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16863 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
16864 * configure: Regenerate.
16865 * config.in: Regenerate.
16866 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
16867 if !HAVE_LD_EH_FRAME_CIEV3.
16869 2013-03-21 Richard Biener <rguenther@suse.de>
16871 * tree-cfg.c (verify_expr_no_block): New function.
16872 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
16873 nor DECL_VALUE_EXPR have locations with associated blocks.
16874 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
16875 (clear_unused_block_pointer): Remove code dealing with
16876 blocks in DECL_DEBUG_EXPR locations.
16878 2013-03-21 Richard Biener <rguenther@suse.de>
16880 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
16881 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
16882 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
16883 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
16884 instead of DECL_DEBUG_EXPR_IS_FROM.
16885 * gimplify.c (gimplify_modify_expr): Likewise.
16886 * tree-cfg.c (verify_expr_location_1): Likewise.
16887 * tree-complex.c (create_one_component_var): Likewise.
16888 * tree-sra.c (create_access_replacement): Likewise.
16889 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
16890 (clear_unused_block_pointer): Likewise.
16891 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
16892 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
16893 * var-tracking.c (var_debug_decl): Likewise.
16894 (track_expr_p): Likewise.
16895 * tree-inline.c (add_local_variables): Likewise. Set
16896 DECL_HAS_DEBUG_EXPR_P after copying it.
16897 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
16898 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
16900 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16903 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
16904 * configure: Regenerate.
16905 * config.in: Regenerate.
16906 * config/i386/i386.md (*movdf_internal): Use
16907 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16908 movd instead of movq mnemonic for interunit moves.
16909 (*movdi_internal): Ditto.
16911 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16913 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
16914 (abd<mode>_3): New pattern.
16915 (aba<mode>_3): New pattern.
16916 (fabd<mode>_3): New pattern.
16918 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16920 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
16921 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
16922 occurrence of REGISTER_PREFIX as its empty string.
16924 2013-03-20 Jeff Law <law@redhat.com>
16926 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
16927 addititional equivalences for equality comparisons between an SSA_NAME
16928 and a constant where the SSA_NAME was set from a widening conversion.
16930 2013-03-20 Walter Lee <walt@tilera.com>
16932 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
16934 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16936 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
16937 depending on type attribute.
16938 (*movti_internal): Ditto.
16939 (*movtf_internal): Ditto.
16940 (*movxf_internal): Ditto.
16941 (*movdf_internal): Ditto.
16942 (*movsf_internal): Ditto.
16944 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16946 * config/i386/i386.md (*movti_internal): Set prefix attribute to
16947 maybe_vex for sselog1 and ssemov types.
16948 (*movdi_internal): Reorder operand constraints.
16949 (*movsi_internal): Ditto. Set prefix attribute to
16950 maybe_vex for sselog1 and ssemov types.
16951 (*movtf_internal): Set prefix attribute to maybe_vex
16952 for sselog1 and ssemov types.
16953 (*movdf_internal): Ditto. Set prefix_data16 attribute for
16954 DImode ssemov types. Reorder operand constraints.
16955 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
16956 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
16957 attribute for SImode ssemov types. Reorder operand constraints.
16959 2013-03-20 Martin Jambor <mjambor@suse.cz>
16961 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
16962 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
16964 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
16966 * config/rs6000/predicates.md (indexed_address, update_address_mem
16967 update_indexed_address_mem): New predicates.
16968 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
16969 attribute for load/store instructions.
16970 * config/rs6000/dfp.md (movsd_store): Likewise.
16971 (movsd_load): Likewise.
16972 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
16973 (unnamed HI->DI extend define_insn): Likewise.
16974 (unnamed SI->DI extend define_insn): Likewise.
16975 (unnamed QI->SI extend define_insn): Likewise.
16976 (unnamed QI->HI extend define_insn): Likewise.
16977 (unnamed HI->SI extend define_insn): Likewise.
16978 (unnamed HI->SI extend define_insn): Likewise.
16979 (extendsfdf2_fpr): Likewise.
16980 (movsi_internal1): Likewise.
16981 (movsi_internal1_single): Likewise.
16982 (movhi_internal): Likewise.
16983 (movqi_internal): Likewise.
16984 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
16985 attribute for load/store instructions.
16986 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
16988 (mov<mode>_softfloat): Likewise.
16989 (mov<mode>_hardfloat32): Likewise.
16990 (mov<mode>_hardfloat64): Likewise.
16991 (mov<mode>_softfloat64): Likewise.
16992 (movdi_internal32): Likewise.
16993 (movdi_internal64): Likewise.
16994 (probe_stack_<mode>): Likewise.
16996 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
16998 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
16999 floating point, and decimal floating point to reload iterator.
17001 * config/rs6000/constraints.md (wl constraint): New constraints to
17002 return FLOAT_REGS if certain options are used to reduce the number
17003 of separate patterns that exist in the file.
17004 (wx constraint): Likewise.
17005 (wz constraint): Likewise.
17007 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17008 -mdebug=reg, print wg, wl, wx, and wz constraints.
17009 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17010 Initialize the reload functions for 64-bit binary/decimal floating
17012 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17013 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17014 create the buffer on the stack to overcome not having a 32-bit
17016 (rs6000_emit_move): Likewise.
17017 (rs6000_secondary_memory_needed_rtx): Likewise.
17018 (rs6000_alloc_sdmode_stack_slot): Likewise.
17019 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17020 via xxlxor, just like DFmode 0.0.
17022 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
17023 define as 1 if we are running on a power7 or newer.
17024 (enum r6000_reg_class_enum): Add new constraints.
17026 * config/rs6000/dfp.md (movsd): Delete, combine with binary
17027 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17028 with other moves by using conditional constraits (wg). Use LFIWZX
17029 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
17030 (movsd splitter): Likewise.
17031 (movsd_hardfloat): Likewise.
17032 (movsd_softfloat): Likewise.
17034 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
17035 binary and decimal floating point moves.
17036 (fmove_ok): New attributes to combine binary and decimal floating
17037 point moves, and to combine power6x (mfpgpr) moves along normal
17039 (real_value_to_target): Likewise.
17040 (f32_lr): Likewise.
17041 (f32_lm): Likewise.
17042 (f32_li): Likewise.
17043 (f32_sr): Likewise.
17044 (f32_sm): Likewise.
17045 (f32_si): Likewise.
17046 (movsf): Combine binary and decimal floating point moves. Combine
17047 power6x (mfpgpr) moves with other moves by using conditional
17048 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
17049 (mov<mode> for SFmode/SDmode); Likewise.
17050 (SFmode/SDmode splitters): Likewise.
17051 (movsf_hardfloat): Likewise.
17052 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
17053 (movsf_softfloat): Likewise.
17054 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
17056 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
17057 wx and wz constraints.
17059 * config/rs6000/constraints.md (wg constraint): New constraint to
17060 return FLOAT_REGS if -mmfpgpr (power6x) was used.
17062 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
17065 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17066 -mdebug=reg, print wg, wl, wx, and wz constraints.
17067 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17068 Initialize the reload functions for 64-bit binary/decimal floating
17070 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17071 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17072 create the buffer on the stack to overcome not having a 32-bit
17074 (rs6000_emit_move): Likewise.
17075 (rs6000_secondary_memory_needed_rtx): Likewise.
17076 (rs6000_alloc_sdmode_stack_slot): Likewise.
17077 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17078 via xxlxor, just like DFmode 0.0.
17080 * config/rs6000/dfp.md (movdd): Delete, combine with binary
17081 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17082 with other moves by using conditional constraits (wg). Use LFIWZX
17083 and STFIWX for loading SDmode on power7.
17084 (movdd splitters): Likewise.
17085 (movdd_hardfloat32): Likewise.
17086 (movdd_softfloat32): Likewise.
17087 (movdd_hardfloat64_mfpgpr): Likewise.
17088 (movdd_hardfloat64): Likewise.
17089 (movdd_softfloat64): Likewise.
17091 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
17092 64-bit binary and decimal floating point moves.
17093 (FMOVE64X): Likewise.
17094 (movdf): Combine 64-bit binary and decimal floating point moves.
17095 Combine power6x (mfpgpr) moves with other moves by using
17096 conditional constraits (wg).
17097 (mov<mode> for DFmode/DDmode): Likewise.
17098 (DFmode/DDmode splitters): Likewise.
17099 (movdf_hardfloat32): Likewise.
17100 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
17101 (movdf_softfloat32): Likewise.
17102 (movdf_hardfloat64_mfpgpr): Likewise.
17103 (movdf_hardfloat64): Likewise.
17104 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
17105 (movdf_softfloat64): Likewise.
17106 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
17107 (reload_<mode>_load): Move to later in the file so they aren't in
17108 the middle of the floating point move insns.
17109 (reload_<mode>_store): Likewise.
17111 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
17114 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
17115 constraint if -mdebug=reg.
17116 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
17117 Enable using dd reload support if needed.
17119 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
17120 binary and decimal floating point moves in rs6000.md.
17121 (movtd_internal): Likewise.
17123 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
17124 decimal floating point moves.
17126 (movtf_internal): Likewise.
17127 (mov<mode>_internal, TDmode/TFmode): Likewise.
17128 (movtf_softfloat): Likewise.
17129 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
17131 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
17132 movdi_internal64, using wg constraint for move direct operations.
17133 (movdi_internal64): Likewise.
17135 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
17136 MODES_TIEABLE_P for selected modes. Print the numerical value of
17137 the various virtual registers. Use GPR/FPR first/last values,
17138 instead of hard coding the register numbers. Print which modes
17139 have reload functions registered.
17140 (rs6000_option_override_internal): If -mdebug=reg, trace the options
17141 settings before/after setting cpu, target and subtarget settings.
17142 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
17143 and for secondary reload failures in rs6000_secondary_reload_inner.
17144 (rs6000_secondary_reload_fail): Likewise.
17145 (rs6000_secondary_reload_inner): Likewise.
17147 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
17148 macros for first/last GPR and FPR registers.
17149 (LAST_GPR_REGNO): Likewise.
17150 (FIRST_FPR_REGNO): Likewise.
17151 (LAST_FPR_REGNO): Likewise.
17153 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
17154 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
17155 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
17156 (vcond<mode><mode>): Likewise.
17157 (vcondu<mode><mode>): Likewise.
17158 (vector_gtu<mode>): Likewise.
17159 (vector_gte<mode>): Likewise.
17160 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
17161 to prevent the compiler from converting DImode operations to TImode.
17162 (ior<mode>3): Likewise.
17163 (and<mode>3): Likewise.
17164 (one_cmpl<mode>2): Likewise.
17165 (nor<mode>3): Likewise.
17166 (andc<mode>3): Likewise.
17168 * config/rs6000/constraints.md (wt constraint): New constraint
17169 that returns VSX_REGS if TImode is allowed in VSX registers.
17171 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
17172 constant under VSX.
17174 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
17175 similar to TImode, but it is restricted to being in the GPRs.
17177 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
17178 TImode to occupy a single VSX register.
17180 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
17181 -mvsx-timode for power7/power8.
17182 (power7 cpu): Likewise.
17183 (power8 cpu): Likewise.
17185 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
17186 sure that TFmode/TDmode take up two registers if they are ever
17187 allowed in the upper VSX registers.
17188 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
17190 (rs6000_init_hard_regno_mode_ok): Likewise.
17191 (rs6000_debug_reg_global): Add debugging for PTImode and wt
17192 constraint. Print if LRA is turned on.
17193 (rs6000_option_override_internal): Give an error if -mvsx-timode
17194 and VSX is not enabled.
17195 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
17196 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
17197 to reg+offset addressing. Use PTImode when checking offset
17198 addresses for validity.
17199 (reg_offset_addressing_ok_p): Likewise.
17200 (rs6000_legitimate_offset_address_p): Likewise.
17201 (rs6000_legitimize_address): Likewise.
17202 (rs6000_legitimize_reload_address): Likewise.
17203 (rs6000_legitimate_address_p): Likewise.
17204 (rs6000_eliminate_indexed_memrefs): Likewise.
17205 (rs6000_emit_move): Likewise.
17206 (rs6000_secondary_reload): Likewise.
17207 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
17208 reloads to fpr registers to continue to use reg+offset addressing,
17209 but 64-bit reloads to altivec registers need reg+reg addressing.
17210 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
17211 it. Treat LO_SUM like a PLUS operation.
17212 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
17213 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
17214 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
17215 registers to share a register with a smaller sized type, since VSX
17216 puts scalars in the upper 64-bits.
17217 (print_operand): Add support for PTImode.
17218 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
17219 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
17220 registers, but don't have arithmetic support.
17221 (rs6000_memory_move_cost): Add test for VSX.
17222 (rs6000_opt_masks): Add -mvsx-timode.
17224 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
17227 (VSr): Use wt constraint for TImode.
17228 (VSv): Drop TImode support.
17229 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
17230 (vsx_movti_64bit): Likewise.
17231 (vsx_movti_32bit): Likewise.
17232 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
17233 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
17234 one '?' on the appropriate output constraint. Do not allow TImode
17235 logical operations on 32-bit systems.
17236 (vsx_ior<mode>3): Likewise.
17237 (vsx_xor<mode>3): Likewise.
17238 (vsx_one_cmpl<mode>2): Likewise.
17239 (vsx_nor<mode>3): Likewise.
17240 (vsx_andc<mode>3): Likewise.
17241 (vsx_concat_<mode>): Likewise.
17242 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
17244 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
17245 OPTION_MASK_VSX_TIMODE.
17246 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
17247 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
17249 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
17250 (TI2 iterator): New iterator for TImode, PTImode.
17251 (wd mode attribute): Add values for vector types.
17252 (movti_string): Replace TI move operations with operations for TImode
17253 and PTImode. Add support for TImode being allowed in VSX registers.
17254 (mov<mode>_string, TImode/PTImode): Likewise.
17255 (movti_ppc64): Likewise.
17256 (mov<mode>_ppc64, TImode/PTImode): Likewise.
17257 (TI mode splitters): Likewise.
17259 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
17262 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
17264 PR tree-optimization/56355
17265 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
17266 Also handle integers with undefined overflow.
17268 2013-03-20 Catherine Moore <clm@codesourcery.com>
17269 Maciej W. Rozycki <macro@codesourcery.com>
17270 Tom de Vries <tom@codesourcery.com>
17271 Nathan Sidwell <nathan@codesourcery.com>
17272 Iain Sandoe <iain@codesourcery.com>
17273 Nathan Froyd <froydnj@codesourcery.com>
17274 Chao-ying Fu <fu@mips.com>
17276 * doc/extend.texi: (micromips, nomicromips, nocompression):
17277 Document new function attributes.
17278 * doc/invoke.texi (minterlink-compressed, mmicromips,
17279 m14k, m14ke, m14kec): Document new options.
17280 (minterlink-mips16): Update documentation.
17281 * doc/md.texi (ZC, ZD): Document new constraints.
17282 * configure.ac (gcc_cv_as_micromips): Check if linker
17283 supports the .set micromips directive.
17284 * configure: Regenerate.
17285 * config.in: Regenerate.
17286 * config/mips/mips-tables.opt: Regenerate.
17287 * config/mips/micromips.md: New file.
17288 * constraints.md (ZC, ZD): New constraints.
17289 * config/mips/predicates.md (movep_src_register): New predicate.
17290 (movep_src_operand): New predicate.
17291 (non_volatile_mem_operand): New predicate.
17292 * config/mips/mips.md (multimem): New type.
17293 (length): Differentiate between 17-bit and 18-bit branch offsets.
17294 (MOVEP1, MOVEP2): New mode iterator.
17295 (mov_<load>l): Use ZC constraint.
17296 (mov_<load>r): Likewise.
17297 (mov_<store>l): Likewise.
17298 (mov_<store>r): Likewise.
17299 (*branch_equality<mode>_inverted): Add microMIPS support.
17300 (*branch_equality<mode>): Likewise.
17301 (*jump_absolute): Likewise.
17302 (indirect_jump_<mode>): Likewise.
17303 (tablejump_<mode>): Likewise.
17304 (<optab>_internal): Likewise.
17305 (sibcall_internal): Likewise.
17306 (sibcall_value_internal): Likewise.
17307 (prefetch): Use constraint ZD.
17308 * config/mips/mips.opt (minterlink-compressed): New option.
17309 (minterlink-mips16): Now an alias for minterlink-compressed.
17310 (mmicromips): New option.
17311 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17312 (compare_and_swap_12): Likewise.
17313 (sync_add<mode>): Likewise.
17314 (sync_<optab>_12): Likewise.
17315 (sync_old_<optab>_12): Likewise.
17316 (sync_new_<optab>_12): Likewise.
17317 (sync_nand_12): Likewise.
17318 (sync_old_nand_12): Likewise.
17319 (sync_new_nand_12): Likewise.
17320 (sync_sub<mode>): Likewise.
17321 (sync_old_add<mode>): Likewise.
17322 (sync_old_sub<mode>): Likewise.
17323 (sync_new_add<mode>): Likewise.
17324 (sync_new_sub<mode>): Likewise.
17325 (sync_<optab><mode>): Likewise.
17326 (sync_old_<optab><mode>): Likewise.
17327 (sync_new_<optab><mode>): Likewise.
17328 (sync_nand<mode>): Likewise.
17329 (sync_old_nand<mode>): Likewise.
17330 (sync_new_nand<mode>): Likewise.
17331 (sync_lock_test_and_set<mode>): Likewise.
17332 (test_and_set_12): Likewise.
17333 (atomic_compare_and_swap<mode>): Likewise.
17334 (atomic_exchange<mode>_llsc): Likewise.
17335 (atomic_fetch_add<mode>_llsc): Likewise.
17336 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17337 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17338 (umips_save_restore_pattern_p): Likewise.
17339 (umips_load_store_pair_p): Likewise.
17340 (umips_output_load_store_pair): Likewise.
17341 (umips_movep_target_p): Likewise.
17342 (umips_12bit_offset_address_p): Likewise.
17343 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17344 (mips_base_mips16): Rename this...
17345 (mips_base_compression_flags): ...to this. Update all uses.
17346 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17347 (mips_mips16_decl_p): Delete.
17348 (mips_nomips16_decl_p): Delete.
17349 (mips_get_compress_on_flags): New function.
17350 (mips_get_compress_off_flags): New function.
17351 (mips_get_compress_mode): New function.
17352 (mips_get_compress_on_name): New function.
17353 (mips_get_compress_off_name): New function.
17354 (mips_insert_attributes): Support multiple compression types.
17355 (mips_merge_decl_attributes): Likewise.
17356 (umips_12bit_offset_address_p): New function.
17357 (mips_start_function_definition): Emit .set micromips directive.
17358 (mips_call_may_need_jalx_p): New function.
17359 (mips_function_ok_for_sibcall): Add microMIPS support.
17360 (mips_print_operand_punctuation): Support short delay slots and
17362 (umips_swm_mask, umips_swm_encoding): New.
17363 (umips_build_save_restore): New function.
17364 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17365 (was_mips16_p): Remove.
17366 (old_compression_mode): New.
17367 (mips_set_compression_mode): New function.
17368 (mips_set_current_function): Add microMIPS support.
17369 (mips_option_override): Likewise.
17370 (umips_save_restore_pattern_p): New function.
17371 (umips_output_save_restore): New function.
17372 (umips_load_store_pair_p_1): New function.
17373 (umips_load_store_pair_p): New function.
17374 (umips_output_load_store_pair_1): New function.
17375 (umips_output_load_store_pair): New function.
17376 (umips_movep_target_p) New function.
17377 (mips_prepare_pch_save): Add microMIPS support.
17378 * config/mips/mips.h (TARGET_COMPRESSION): New.
17379 (TARGET_CPU_CPP_BUILTINS): Update macro
17380 to use new compression flags and to support microMIPS.
17381 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17382 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17383 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17384 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17385 (ASM_SPEC): Support mmicromips and mno-micromips.
17386 (M16STORE_REG_P): New macro.
17387 (MIPS_CALL): Support TARGET_MICROMIPS.
17388 (MICROMIPS_J): New macro.
17389 (mips_base_mips16): Rename this...
17390 (mips_base_compression_flags): ...to this.
17391 (UMIPS_12BIT_OFFSET_P): New macro.
17392 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17393 (MULTILIB_DIRNAMES): Likewise.
17394 2013-03-20 Richard Biener <rguenther@suse.de>
17396 PR tree-optimization/56661
17397 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
17398 the result does not have to be distinct.
17400 2013-03-20 Richard Biener <rguenther@suse.de>
17402 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
17405 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17406 Steven Bosscher <steven@gcc.gnu.org>
17408 PR rtl-optimization/56605
17409 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
17411 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17414 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
17415 that require movd instead of movq.
17417 2013-03-20 Richard Biener <rguenther@suse.de>
17419 * tree-ssa-structalias.c (struct variable_info): Add pointer
17420 to the first field of an aggregate with sub-vars. Make
17421 this and the pointer to the next subfield its ID.
17422 (vi_next): New function.
17423 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
17424 storedanything_id, integer_id): Increment by one.
17425 (new_var_info, get_call_vi, lookup_call_clobber_vi,
17426 get_call_clobber_vi): Adjust.
17427 (solution_set_expand): Simplify and speedup.
17428 (solution_set_add): Inline into ...
17429 (set_union_with_increment): ... this. Adjust accordingly.
17430 (do_sd_constraint): Likewise.
17431 (do_ds_constraint): Likewise.
17432 (do_complex_constraint): Simplify.
17433 (build_pred_graph): Adjust.
17434 (solve_graph): Likewise. Simplify and speedup.
17435 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
17436 get_constraint_for_component_ref, get_constraint_for_1,
17437 first_vi_for_offset, first_or_preceding_vi_for_offset,
17438 create_function_info_for, create_variable_info_for_1,
17439 create_variable_info_for, intra_create_variable_infos): Adjust.
17440 (init_base_vars): Push NULL for ID zero.
17441 (compute_points_to_sets): Adjust.
17443 2013-03-20 Richard Biener <rguenther@suse.de>
17445 * cfgloop.c (verify_loop_structure): Streamline and avoid
17446 ICEing on corrupt loop tree.
17447 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17450 2013-03-20 Richard Biener <rguenther@suse.de>
17452 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17453 check whether an SSA update is needed.
17455 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17457 * config/mips/constraints.md (T): Rename to...
17461 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17462 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17464 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17466 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17467 (*subsi3_carryin_uxtw): Likewise.
17469 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17471 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17472 (*rorsi3_insn_uxtw): Likewise.
17474 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17476 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17477 (*extrsi5_insn_uxtw): Likewise.
17479 2013-03-19 Richard Biener <rguenther@suse.de>
17481 PR tree-optimization/56273
17482 * passes.c (init_optimization_passes): Move second VRP after DOM.
17484 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
17486 * config/i386/i386.md (*movti_internal): Merge from
17487 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17488 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17489 nox64 isa attributes.
17491 2013-03-18 Richard Biener <rguenther@suse.de>
17493 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17495 (merge_node_constraints): Likewise.
17496 (build_succ_graph): Likewise.
17497 (valid_graph_edge): Inline into single caller.
17498 (unify_nodes): Likewise. Use bitmap_set_bit return value
17500 (scc_visit): Fix formatting and variable use.
17501 (do_sd_constraint): Use gcc_checking_assert.
17502 (do_ds_constraint): Likewise.
17503 (do_complex_constraint): Likewise.
17504 (condense_visit): Likewise. Cleanup.
17505 (dump_pred_graph): New function.
17506 (perform_var_substitution): Dump the pred-graph before
17507 variable substitution.
17508 (find_equivalent_node): Use gcc_checking_assert.
17509 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17511 2013-03-18 Richard Biener <rguenther@suse.de>
17513 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17514 Remove cond_expr_stmt_list argument and do not gimplify the
17516 (vect_loop_versioning): Adjust.
17517 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17518 Cleanup to use less temporaries.
17519 (vect_create_data_ref_ptr): Cleanup.
17521 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17523 PR tree-optimization/56635
17524 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17525 require types_compatible_p types.
17527 2013-03-18 Nick Clifton <nickc@redhat.com>
17529 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
17530 spurious backslash.
17532 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
17533 Add missing line to comment describing function.
17535 2013-03-18 Richard Biener <rguenther@suse.de>
17537 PR tree-optimization/56210
17538 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17539 Handle string / character search functions.
17540 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
17542 2013-03-18 Richard Biener <rguenther@suse.de>
17544 PR middle-end/56483
17545 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
17546 and implement properly.
17547 * gimple.h (gimple_cond_single_var_p): Remove.
17549 2013-03-18 Richard Biener <rguenther@suse.de>
17551 * tree-data-ref.h (find_data_references_in_loop): Declare.
17552 * tree-data-ref.c (get_references_in_stmt): Use a stack
17553 vector pre-allocated in the callers.
17554 (find_data_references_in_stmt): Adjust.
17555 (graphite_find_data_references_in_stmt): Likewise.
17556 (create_rdg_vertices): Likewise.
17557 (find_data_references_in_loop): Export.
17558 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
17559 Compute dependences here...
17560 (vect_analyze_data_refs): ...not here. When we encounter
17561 a non-vectorizable data reference in basic-block vectorization
17562 truncate the data reference vector. Do not bother to
17563 fixup data-dependence information for gather loads.
17564 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
17565 of data references, as reported.
17567 2013-03-18 Richard Biener <rguenther@suse.de>
17569 PR tree-optimization/3713
17570 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
17571 has_constants and expr.
17572 (stmt_has_constants): Properly valueize SSA names when deciding
17573 whether the stmt has constants.
17575 2013-03-18 Richard Biener <rguenther@suse.de>
17577 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
17578 whole function when there is nothing to do.
17579 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
17580 * tree-vectorizer.c (vectorize_loops): Update virtual and
17581 loop-closed SSA once.
17582 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
17584 2013-03-18 Richard Biener <rguenther@suse.de>
17586 PR middle-end/56113
17587 * domwalk.c (bb_postorder): New global static.
17588 (cmp_bb_postorder): New function.
17589 (walk_dominator_tree): Replace scheme imposing an order for
17590 visiting dominator sons by one sorting them at the time they
17591 are pushed on the stack.
17593 2013-03-18 Richard Biener <rguenther@suse.de>
17595 PR tree-optimization/39326
17596 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17597 (struct mem_ref): Replace mem member with ao_ref typed member.
17598 (MEM_ANALYZABLE): Adjust.
17599 (memref_eq): Likewise.
17600 (mem_ref_alloc): Likewise.
17601 (gather_mem_refs_stmt): Likewise.
17602 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17603 (execute_sm_if_changed_flag_set): Adjust.
17604 (execute_sm): Likewise.
17605 (ref_always_accessed_p): Likewise.
17606 (refs_independent_p): Likewise.
17607 (can_sm_ref_p): Likewise.
17609 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17612 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17613 return 1 even for !unsignedp.
17615 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17617 * config/i386/i386.md (isa): Add x64 and nox64.
17618 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17619 (*pushtf): Enable *roF alternative for x64 isa only.
17620 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17621 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17622 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17623 (*movtf_internal): Merge from *movtf_internal_rex64 and
17624 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17625 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17626 nox64 isa attributes.
17627 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17628 nox64 isa attributes.
17629 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17631 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17633 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17635 2013-03-17 Jakub Jelinek <jakub@redhat.com>
17638 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17639 class names. Remove trailing comma after "ALL_REGS".
17641 2013-03-16 Jan Hubicka <jh@suse.cz>
17643 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17644 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17645 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17646 of cgraph_get_create_node.
17647 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17649 2013-03-16 Jason Merrill <jason@redhat.com>
17652 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17653 with DW_AT_default_value.
17655 2013-03-16 Jakub Jelinek <jakub@redhat.com>
17657 * BASE-VER: Set to 4.9.0.
17659 2013-03-14 Andi Kleen <ak@linux.intel.com>
17662 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17663 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17664 Document _x* TSX intrinsics.
17666 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17667 David Holsgrove <david.holsgrove@xilinx.com>
17669 * configure.ac: Add MicroBlaze TLS support detection.
17670 * configure: Regenerate.
17671 * config/microblaze/microblaze-protos.h
17672 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17673 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17674 * config/microblaze/microblaze.c (microblaze_address_type): Add
17675 ADDRESS_TLS and tls_reloc address types.
17676 (microblaze_address_info): Add tls_reloc.
17677 (TARGET_HAVE_TLS): Define.
17678 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17679 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
17680 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17681 load_tls_operand, microblaze_call_tls_get_addr,
17682 microblaze_legitimize_tls_address): New functions.
17683 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17684 (get_base_reg): Use microblaze_tls_symbol_p.
17685 (microblaze_classify_address): Handle TLS.
17686 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17687 label_mentioned_p and microblaze_tls_referenced_p.
17688 (microblaze_legitimize_address): Handle TLS.
17689 (microblaze_address_insns): Handle ADDRESS_TLS.
17690 (pic_address_needs_scratch): Handle TLS.
17691 (print_operand_address): Handle TLS.
17692 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17693 (microblaze_expand_move): Handle TLS.
17694 (microblaze_legitimate_constant_p): Check
17695 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
17696 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17697 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17698 (PIC_OFFSET_TABLE_REGNUM): Set.
17699 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17700 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17701 (addsi3, movsi_internal2, movdf_internal): Update constraints
17702 * config/microblaze/predicates.md (arith_plus_operand): Define
17703 (move_operand): Redefine as move_src_operand,
17704 check microblaze_tls_referenced_p.
17706 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17708 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17709 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17711 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17713 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17716 2013-03-14 Jakub Jelinek <jakub@redhat.com>
17718 PR tree-optimization/53265
17719 * common.opt (Waggressive-loop-optimizations): New option.
17720 * tree-ssa-loop-niter.c: Include tree-pass.h.
17721 (do_warn_aggressive_loop_optimizations): New function.
17722 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17723 if number_of_latch_executions returned constant.
17724 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17725 early. If number_of_latch_executions returned constant, set
17726 nb_iterations_upper_bound back to it.
17727 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17729 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17730 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17732 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17733 (MULTILIB_OSDIRNAMES): Set.
17734 * genmultilib: If defaultosdirname doesn't start with :: , set
17735 defaultosdirname2 instead, clear it and emit two . multilib_raw
17736 entries instead of just one.
17738 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17740 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17741 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17742 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17743 (SUBTARGET_OVERRIDE_OPTIONS): New.
17745 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17748 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17749 (musermode): Convert to Var(TARGET_USERMODE).
17750 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17751 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17752 * config/sh/sh.c (sh_option_override): Use
17753 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17754 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17756 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17758 (udivsi3_i4_single, divsi3_i4_single): Use
17759 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17761 2013-03-13 Dave Korn <dave.korn.cygwin@....>
17763 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17766 2013-03-13 Richard Biener <rguenther@suse.de>
17768 PR tree-optimization/56608
17769 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17770 calls when vectorizing basic-blocks.
17772 2013-03-13 Jakub Jelinek <jakub@redhat.com>
17775 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17778 2013-03-12 Jakub Jelinek <jakub@redhat.com>
17780 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17782 2013-03-11 Jan Hubicka <jh@suse.cz>
17785 * lto-streamer-out.c (output_symbol_p): Skip references from
17786 constructors of external variables.
17788 2013-03-11 Jan Hubicka <jh@suse.cz>
17790 PR middle-end/56571
17791 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
17793 * emit-rtl.c (verify_rtx_sharing): Likewise.
17794 (copy_insn_1): Likewise.
17795 * rtl.c (copy_rtx): Likewise.
17797 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
17800 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
17801 output_operand_lossage message.
17803 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
17806 * arm.c (shift_op): Validate RTL pattern on the fly.
17807 (arm_print_operand, case 'S'): Don't use shift_operator to validate
17810 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17813 * config/pa/pa.md (call_value): Check for calls to powf and direct to
17814 new call patterns that clobber %fr12.
17815 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
17816 split and postreload patterns.
17817 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
17818 registers %fr12 and %fr12R as call used.
17820 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
17822 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
17823 (canon_address, record_store, replace_read, check_mem_read_rtx,
17824 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
17825 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
17826 rest_of_handle_dse): Likewise.
17828 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
17830 PR middle-end/56524
17831 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
17833 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
17834 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
17835 (save_optabs_if_changed): Replace with...
17836 (init_tree_optimization_optabs): ...this.
17837 * optabs.c (save_optabs_if_changed): Rename to...
17838 (init_tree_optimization_optabs): ...this. Take the optimization node
17839 as argument. Do nothing if the base optabs are already correct.
17840 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
17841 to recompute optabs.
17842 * function.h (function): Remove optabs field.
17843 * function.c (invoke_set_current_function_hook): Call
17844 init_tree_optimization_optabs. Use the result to initialize
17847 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
17849 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
17850 if GTMA_HAS_NO_INSTRUMENTATION.
17851 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
17852 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
17853 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
17854 * gimple-pretty-print.c (dump_gimple_transaction): Handle
17855 GTMA_HAS_NO_INSTRUMENTATION.
17857 2013-03-08 Jakub Jelinek <jakub@redhat.com>
17859 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
17862 2013-03-08 Marek Polacek <polacek@redhat.com>
17863 Jakub Jelinek <jakub@redhat.com>
17865 PR tree-optimization/56478
17866 * predict.c (is_comparison_with_loop_invariant_p): Change the
17867 type of loop_step to tree.
17868 (predict_loops): Adjust.
17869 (predict_iv_comparison): Perform the computations on double_ints.
17871 2013-03-08 Richard Biener <rguenther@suse.de>
17873 PR tree-optimization/56570
17874 * tree-cfg.c (verify_expr_location_1): Verify locations for
17876 * tree-sra.c (create_access_replacement): Strip locations
17877 from DECL_DEBUG_EXPRs.
17879 2013-03-08 Richard Biener <rguenther@suse.de>
17881 * tree-inline.c (expand_call_inline): Do not associate
17882 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
17883 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
17885 2013-03-08 Richard Biener <rguenther@suse.de>
17887 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
17888 or block changes with -Og. Fix for location / block encoding
17889 changes and PHI arguments with locations.
17891 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
17893 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
17895 (struct output_info): Likewise.
17896 (register_overhead): Remove bad gcc_assert.
17897 (bitmap_find_bit): If there is only a single bitmap element, do not
17898 count a miss as a search.
17899 (print_statistics): Update for counter type changes.
17900 (dump_bitmap_statistics): Likewise. Print headers such that they
17901 are properly lined up with the printed counters.
17903 2013-03-07 Jakub Jelinek <jakub@redhat.com>
17905 PR tree-optimization/56559
17906 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
17907 check that it has only a single use.
17909 2013-03-07 Richard Biener <rguenther@suse.de>
17911 * doc/invoke.texi (fwhole-program): Discourage use in combination
17914 2013-03-06 Jakub Jelinek <jakub@redhat.com>
17916 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
17918 PR tree-optimization/56539
17919 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
17920 instead of GSI_CONTINUE_LINKING as last argument to
17921 force_gimple_operand_gsi. Adjust function comment.
17923 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
17926 PR middle-end/56548
17927 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
17928 promoted mode, convert the result back to the original mode.
17930 2013-03-06 Richard Biener <rguenther@suse.de>
17932 PR middle-end/56294
17933 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
17934 (insert_updated_phi_nodes_compare_uids): New function.
17935 (update_ssa): Sort symbols_to_rename after UID before
17936 traversing it to insert PHI nodes.
17938 2013-03-06 Richard Biener <rguenther@suse.de>
17940 PR middle-end/50494
17941 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
17942 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
17945 2013-02-13 Richard Biener <rguenther@suse.de>
17948 * varasm.c (output_constant_def_1): Get the decl representing
17949 the constant as argument.
17950 (output_constant_def): Wrap output_constant_def_1.
17951 (make_decl_rtl): Use output_constant_def_1 with the decl
17952 representing the constant.
17953 (build_constant_desc): Optionally re-use a decl already
17954 representing the constant.
17955 (tree_output_constant_def): Adjust.
17957 2013-03-06 Joey Ye <joey.ye@arm.com>
17960 * gcc.c (convert_white_space): New function.
17961 (main): Handles white space in function name.
17963 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
17966 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
17967 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
17968 to SH_DIV_CALL_TABLE for TARGET_SH2.
17969 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
17971 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
17972 call-table options.
17974 2013-03-05 Sterling Augustine <saugustine@google.com>
17975 Cary Coutant <ccoutant@google.com>
17978 * dwarf2out.c (resolve_addr): Don't call
17979 remove_loc_list_addr_table_entries a second time for the same
17982 2013-03-05 Jakub Jelinek <jakub@redhat.com>
17985 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
17986 (avoid_complex_debug_insns): New function.
17987 (expand_debug_locations): Call it.
17989 PR rtl-optimization/56484
17990 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
17991 lifetimes of hard registers on small register class machines.
17993 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
17995 * config/microblaze/microblaze-protos.h: Rename
17996 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
17997 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
17999 (microblaze_fast_interrupt_function_p): New function.
18000 (microblaze_is_interrupt_handler): Rename to
18001 microblaze_is_interrupt_variant and add fast_interrupt check.
18002 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
18003 (save_restore_insns): Likewise.
18004 (compute_frame_size): Likewise.
18005 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
18006 (microblaze_globalize_label): Likewise.
18007 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
18008 * config/microblaze/microblaze.md: Use wrapper
18009 microblaze_is_interrupt_variant.
18011 2013-03-05 Kai Tietz <ktietz@redhat.com>
18013 * sdbout.c (sdbout_one_type): Switch to current function's section
18014 supporting cold/hot.
18016 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18018 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
18021 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18023 PR middle-end/56461
18024 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
18025 if VALGRIND_GET_VBITS is defined, temporarily make object
18026 memory all defined, and restore previous valgrind addressability
18027 and definability afterwards. Free this_object at the end.
18029 PR middle-end/56461
18030 * lra.c (lra): Call lra_clear_live_ranges if live_p,
18031 right before calling lra_create_live_ranges, also call it
18032 when clearing live_p. Only call lra_clear_live_ranges
18033 at the end if live_p.
18035 PR middle-end/56461
18036 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
18038 2013-03-05 Richard Biener <rguenther@suse.de>
18040 PR tree-optimization/56521
18041 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
18044 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
18047 * except.h (remove_unreachable_eh_regions): New prototype.
18048 * except.c (remove_eh_handler_splicer): New function, split out
18049 of remove_eh_handler.
18050 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
18051 warning about running it on many EH regions one at a time.
18052 (remove_unreachable_eh_regions_worker): New function, walk the
18053 EH tree in depth-first order and remove non-marked regions.
18054 (remove_unreachable_eh_regions): New function.
18055 * tree-eh.c (mark_reachable_handlers): New function, split out
18056 from remove_unreachable_handlers.
18057 (remove_unreachable_handlers): Use mark_reachable_handlers and
18058 remove_unreachable_eh_regions.
18059 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
18060 and remove_unreachable_eh_regions.
18062 2013-03-05 Richard Biener <rguenther@suse.de>
18064 PR middle-end/56525
18065 * loop-init.c (fix_loop_structure): Remove loops in two stages,
18066 not freeing them until the end.
18068 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18070 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
18072 2013-03-05 Richard Biener <rguenther@suse.de>
18074 PR tree-optimization/56270
18075 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
18076 of loads after scheduling an SLP instance.
18078 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18080 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
18082 (check_gcc_parallelize): Run guality.exp as a separate job from
18083 vect.exp with unsorted.exp and $(dg_target_exps) separately from
18084 struct-layout-1.exp with stackalign.exp.
18086 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
18088 PR middle-end/56461
18089 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
18090 load_index sbitmap even if some bit in it isn't set.
18092 PR middle-end/56461
18093 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
18094 (discover_iteration_bound_by_body_walk): Change queues to
18095 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
18096 spelling in comment. Call safe_push on queues[bound_index] directly.
18097 Release queues[queue_index] in every iteration unconditionally.
18098 Release bounds vector.
18100 PR middle-end/56461
18101 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
18102 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
18103 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
18106 2013-03-05 Richard Biener <rguenther@suse.de>
18109 * tree-inline.c (remap_blocks_to_null): New function.
18110 (expand_call_inline): When expanding a call stmt without
18111 an associated block inline remap all callee blocks to NULL.
18113 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18115 PR rtl-optimization/56494
18116 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
18117 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
18118 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
18120 PR middle-end/56461
18121 * sel-sched-ir.c (free_sched_pools): Release
18122 succs_info_pool.stack[succs_info_pool.max_top] vectors too
18123 if succs_info_pool.max_top isn't -1.
18126 * opts.c (opts_obstack, opts_concat): Moved to...
18127 * opts-common.c (opts_obstack, opts_concat): ... here.
18129 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18131 PR middle-end/56461
18132 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
18134 2013-03-04 Martin Jambor <mjambor@suse.cz>
18136 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
18137 all appropriate places.
18139 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
18141 PR tree-optimization/56424
18142 * ipa-split.c (split_function): Do not set the RSO flag if result is
18143 not by reference and its type is a register type.
18145 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
18147 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
18148 (microblaze_legitimate_pic_operand): Likewise
18149 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
18150 new function microblaze_legitimate_pic_operand
18151 * config/microblaze/microblaze-protos.h
18152 (microblaze_legitimate_pic_operand): Declare.
18154 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18156 * config/microblaze/predicates.md (call_insn_simple_operand):
18157 New predicate for supported rtx code types.
18158 * config/microblaze/microblaze.md (call_internal1): Use
18159 call_insn_simple_operand predicate.
18161 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18163 PR middle-end/56461
18164 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
18165 partitions.ordered_remove.
18167 PR middle-end/56461
18168 * tree-vect-stmts.c (vectorizable_conversion): Don't call
18169 vec_oprnds0.create (1) for modifier == NONE.
18171 PR middle-end/56461
18172 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
18173 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
18174 vec_oprnds1 right before pushing anything to it for
18177 PR middle-end/56461
18178 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
18179 set nbbs to 0 instead of having separate code path.
18180 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
18181 instead of false as last argument if returning NULL.
18183 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
18185 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
18186 the attribute is now called "target" instead of "option".
18187 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
18188 * doc/tm.texi.in (Target Attributes): Likewise document the correct
18189 attribute/pragma name for TARGET_OPTION_VALID_P and
18190 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
18191 * doc/tm.texi: Regenerated.
18193 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
18195 * config/microblaze/microblaze.c:
18196 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
18197 * config/microblaze/microblaze.h: Add -mxl-reorder to
18199 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
18200 instructions emitted if TARGET_REORDER.
18201 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
18202 or 0 for -m/-mno case, but initialises as 2 to detect default use case
18205 2013-03-01 Xinliang David Li <davidxl@google.com>
18207 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
18210 2013-03-01 Jakub Jelinek <jakub@redhat.com>
18212 PR middle-end/56461
18213 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
18214 vector even when returning true. Fix up function comment formatting.
18216 PR middle-end/56461
18217 * ira-build.c (ira_loop_nodes_count): New variable.
18218 (create_loop_tree_nodes): Initialize it.
18219 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
18221 PR middle-end/56461
18222 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
18223 method on dr_chain and result_chain.
18224 * tree-vect-stmts.c (vectorizable_store): Only call
18225 result_chain.create if j == 0.
18227 PR middle-end/56461
18228 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
18229 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
18230 before overwriting it.
18232 2013-03-01 Tobias Burnus <burnus@net-b.de>
18234 * doc/extended.texi (C Extensions): Change order in @menu
18236 (Other MIPS Built-in Functions): Move last MIPS entry before
18237 "picoChip Built-in Functions".
18238 (SH Built-in Functions): Move after RX Built-in Functions.
18239 * doc/gcc.texi (Introduction): Change order in @menu
18241 * doc/md.texi (Constraints): Ditto.
18242 * gty.texi (Type Information): Ditto.
18243 (User-provided marking routines for template types): Make
18245 * doc/invoke.texi (AArch64 Options): Move before
18246 "Adapteva Epiphany Options".
18248 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
18249 Jakub Jelinek <jakub@redhat.com>
18252 * asan.c (gate_asan): Lookup no_sanitize_address instead of
18253 no_address_safety_analysis attribute.
18254 * doc/extend.texi (no_address_safety_attribute): Rename to
18255 no_sanitize_address attribute, mention no_address_safety_analysis
18256 attribute as deprecated alias.
18258 2013-02-28 Jakub Jelinek <jakub@redhat.com>
18260 PR middle-end/56461
18261 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
18262 type to vec<vec<tree> > *.
18263 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
18264 to be vec<tree> instead of vec<tree> *, set vec_defs
18265 to vNULL and call vec_defs.create (number_of_vects), adjust other
18267 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
18268 vectorizable_condition): Adjust vect_get_slp_defs callers.
18270 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18272 * config/aarch64/aarch64.c
18273 (aarch64_float_const_representable): Remove unused variable.
18275 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18277 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
18279 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18281 * config/aarch64/aarch64-builtins.c
18282 (aarch64_init_simd_builtins): Make static.
18284 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18286 * config/aarch64/aarch64.c
18287 (aarch64_simd_make_constant): Make static.
18289 2013-02-28 Martin Jambor <mjambor@suse.cz>
18291 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
18292 with no initialization to the RHS of debug statements.
18294 2013-02-28 Martin Jambor <mjambor@suse.cz>
18296 PR tree-optimization/56294
18297 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
18299 (get_access_replacement): Do not call create_access_replacement.
18300 Assert a replacement exists.
18301 (get_repl_default_def_ssa_name): Create the replacement declaration
18304 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18306 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
18307 final_end_function.
18309 2013-02-28 Marek Polacek <polacek@redhat.com>
18311 PR rtl-optimization/56466
18312 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
18313 if we're changing a loop.
18314 (peel_loops_completely): Likewise.
18316 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
18319 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
18321 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18324 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
18325 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
18326 INTX_FTYPE_FX, FX_FTYPE_INTX.
18327 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
18329 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18331 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
18332 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18333 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
18334 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
18335 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
18336 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
18337 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
18338 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
18339 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
18340 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
18341 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
18342 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
18343 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
18344 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
18345 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
18346 (avrxmega6): Increase max flash segments from 5 to 6.
18347 * config/avr/t-multilib: Regenerate.
18348 * config/avr/avr-tables.opt: Regenerate.
18349 * doc/avr-mmcu.texi: Regenerate.
18351 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18353 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
18354 (avr_device_to_arch): Rename to avr_device_to_ld.
18355 (avr_device_to_as): New prototype.
18356 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
18357 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
18358 * config/avr/driver-avr.c (avr_device_to_as): New.
18359 (avr_device_to_arch): Rename to avr_device_to_ld.
18361 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18363 PR middle-end/56461
18364 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
18365 method on dr_chain and result_chain.
18367 PR middle-end/56461
18368 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
18369 pointer_set_destroy on not_executed_last_iteration.
18371 PR middle-end/56461
18372 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
18374 PR middle-end/56461
18375 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
18376 FOR_EACH_DEFINED_FUNCTION when freeing state.
18378 PR middle-end/56461
18379 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
18381 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
18384 PR middle-end/56461
18385 * ipa-cp.c (decide_whether_version_node): Call vec_free on
18386 known_aggs[i].items and release known_aggs vector.
18388 PR middle-end/56461
18389 * ipa-reference.c (propagate): Free node_info even for alias nodes.
18391 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18393 * config/microblaze/microblaze.c (microblaze_emit_compare):
18394 Use xor for EQ/NE comparisions.
18395 * config/microblaze/microblaze.md (cstoresf4): Add constraints
18396 (cbranchsf4): Adjust operator to comparison_operator.
18398 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18400 PR middle-end/56461
18401 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
18403 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
18404 vec_safe_push, always update *slot.
18405 (redirect_edge_var_map_clear): Use vec_free.
18406 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
18407 (free_var_map_entry): Use vec_free.
18408 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
18409 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
18411 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
18413 PR middle-end/45472
18414 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
18415 when the may_trap_p bit of the exprs being merged differs.
18416 Reorder tests for speculativeness in the logical and operator.
18418 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18420 * incpath.c (add_standard_paths): Use reconcat instead of concat
18421 where appropriate and avoid leaking memory.
18423 * opts.h: Include obstack.h.
18424 (opts_concat): New prototype.
18425 (opts_obstack): New declaration.
18426 * opts.c (opts_concat): New function.
18427 (opts_obstack): New variable.
18428 (init_options_struct): Call gcc_init_obstack on opts_obstack.
18429 (finish_options): Use opts_concat instead of concat
18430 and XOBNEWVEC instead of XNEWVEC.
18431 * opts-common.c (generate_canonical_option, decode_cmdline_option,
18432 generate_option): Likewise.
18433 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
18434 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
18437 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18438 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18440 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18442 PR middle-end/56461
18443 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18445 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18447 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18448 (arm_block_move_unaligned_straight): Likewise.
18449 (arm_adjust_block_mem): Likewise.
18451 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18454 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18455 temp, cond and label.
18456 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18459 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18460 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18463 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18464 of prologue/epilogue functions.
18467 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18468 Remove unused variable cfa_offset.
18469 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18472 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18473 type promotion to unsigned.
18476 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18477 for HOST_WIDE_INT of 32 bit / same size as int.
18478 (arm_block_move_unaligned_straight): Likewise.
18479 (arm_adjust_block_mem): Likewise.
18482 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18485 2013-02-26 Marek Polacek <polacek@redhat.com>
18487 PR tree-optimization/56426
18488 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
18490 2013-02-26 Richard Biener <rguenther@suse.de>
18493 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18494 unused variable loops.
18496 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18498 PR tree-optimization/56448
18499 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18500 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18501 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18502 later operands of the references, or even first operand for
18503 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18505 PR tree-optimization/56443
18506 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18507 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18508 to type_for_mode langhook.
18510 2013-02-25 Matt Turner <mattst88@gmail.com>
18512 * doc/invoke.texi: Document r4700.
18514 2013-02-25 Richard Biener <rguenther@suse.de>
18516 PR tree-optimization/56175
18517 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18519 (simplify_bitwise_binary): ... here. Also guard the conversion
18520 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18522 2013-02-25 Catherine Moore <clm@codesourcery.com>
18525 2013-02-24 Catherine Moore <clm@codesourcery.com>
18526 Maciej W. Rozycki <macro@codesourcery.com>
18527 Tom de Vries <tom@codesourcery.com>
18528 Nathan Sidwell <nathan@codesourcery.com>
18529 Iain Sandoe <iain@codesourcery.com>
18530 Nathan Froyd <froydnj@codesourcery.com>
18531 Chao-ying Fu <fu@mips.com>
18533 * doc/extend.texi: (micromips, nomicromips, nocompression):
18534 Document new function attributes.
18535 * doc/invoke.texi (minterlink-compressed, mmicromips,
18536 m14k, m14ke, m14kec): Document new options.
18537 (minterlink-mips16): Update documentation.
18538 * doc/md.texi (ZC, ZD): Document new constraints.
18539 * configure.ac (gcc_cv_as_micromips): Check if linker
18540 supports the .set micromips directive.
18541 * configure: Regenerate.
18542 * config.in: Regenerate.
18543 * config/mips/mips-tables.opt: Regenerate.
18544 * config/mips/micromips.md: New file.
18545 * constraints.md (ZC, AD): New constraints.
18546 * config/mips/predicates.md (movep_src_register): New predicate.
18547 (movep_src_operand): New predicate.
18548 (non_volatile_mem_operand): New predicate.
18549 * config/mips/mips.md (multimem): New type.
18550 (length): Differentiate between 17-bit and 18-bit branch offsets.
18551 (MOVEP1, MOVEP2): New mode iterator.
18552 (mov_<load>l): Use ZC constraint.
18553 (mov_<load>r): Likewise.
18554 (mov_<store>l): Likewise.
18555 (mov_<store>r): Likewise.
18556 (*branch_equality<mode>_inverted): Add microMIPS support.
18557 (*branch_equality<mode>): Likewise.
18558 (*jump_absolute): Likewise.
18559 (indirect_jump_<mode>): Likewise.
18560 (tablejump_<mode>): Likewise.
18561 (<optab>_internal): Likewise.
18562 (sibcall_internal): Likewise.
18563 (sibcall_value_internal): Likewise.
18564 (prefetch): Use constraint ZD.
18565 * config/mips/mips.opt (minterlink-compressed): New option.
18566 (minterlink-mips16): Now an alias for minterlink-compressed.
18567 (mmicromips): New option.
18568 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18569 (compare_and_swap_12): Likewise.
18570 (sync_add<mode>): Likewise.
18571 (sync_<optab>_12): Likewise.
18572 (sync_old_<optab>_12): Likewise.
18573 (sync_new_<optab>_12): Likewise.
18574 (sync_nand_12): Likewise.
18575 (sync_old_nand_12): Likewise.
18576 (sync_new_nand_12): Likewise.
18577 (sync_sub<mode>): Likewise.
18578 (sync_old_add<mode>): Likewise.
18579 (sync_old_sub<mode>): Likewise.
18580 (sync_new_add<mode>): Likewise.
18581 (sync_new_sub<mode>): Likewise.
18582 (sync_<optab><mode>): Likewise.
18583 (sync_old_<optab><mode>): Likewise.
18584 (sync_new_<optab><mode>): Likewise.
18585 (sync_nand<mode>): Likewise.
18586 (sync_old_nand<mode>): Likewise.
18587 (sync_new_nand<mode>): Likewise.
18588 (sync_lock_test_and_set<mode>): Likewise.
18589 (test_and_set_12): Likewise.
18590 (atomic_compare_and_swap<mode>): Likewise.
18591 (atomic_exchange<mode>_llsc): Likewise.
18592 (atomic_fetch_add<mode>_llsc): Likewise.
18593 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18594 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18595 (umips_save_restore_pattern_p): Likewise.
18596 (umips_load_store_pair_p): Likewise.
18597 (umips_output_load_store_pair): Likewise.
18598 (umips_movep_target_p): Likewise.
18599 (umips_12bit_offset_address_p): Likewise.
18600 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18601 (mips_base_mips16): Rename this...
18602 (mips_base_compression_flags): ...to this. Update all uses.
18603 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18604 (mips_mips16_decl_p): Delete.
18605 (mips_nomips16_decl_p): Delete.
18606 (mips_get_compress_on_flags): New function.
18607 (mips_get_compress_off_flags): New function.
18608 (mips_get_compress_mode): New function.
18609 (mips_get_compress_on_name): New function.
18610 (mips_get_compress_off_name): New function.
18611 (mips_insert_attributes): Support multiple compression types.
18612 (mips_merge_decl_attributes): Likewise.
18613 (umips_12bit_offset_address_p): New function.
18614 (mips_start_function_definition): Emit .set micromips directive.
18615 (mips_call_may_need_jalx_p): New function.
18616 (mips_function_ok_for_sibcall): Add microMIPS support.
18617 (mips_print_operand_punctuation): Support short delay slots and
18619 (umips_swm_mask, umips_swm_encoding): New.
18620 (umips_build_save_restore): New function.
18621 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18622 (was_mips16_p): Remove.
18623 (old_compression_mode): New.
18624 (mips_set_compression_mode): New function.
18625 (mips_set_current_function): Add microMIPS support.
18626 (mips_option_override): Likewise.
18627 (umips_save_restore_pattern_p): New function.
18628 (umips_output_save_restore): New function.
18629 (umips_load_store_pair_p_1): New function.
18630 (umips_load_store_pair_p): New function.
18631 (umips_output_load_store_pair_1): New function.
18632 (umips_output_load_store_pair): New function.
18633 (umips_movep_target_p) New function.
18634 (mips_prepare_pch_save): Add microMIPS support.
18635 * config/mips/mips.h (TARGET_COMPRESSION): New.
18636 (TARGET_CPU_CPP_BUILTINS): Update macro
18637 to use new compression flags and to support microMIPS.
18638 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18639 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18640 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18641 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18642 (ASM_SPEC): Support mmicromips and mno-micromips.
18643 (M16STORE_REG_P): New macro.
18644 (MIPS_CALL): Support TARGET_MICROMIPS.
18645 (MICROMIPS_J): New macro.
18646 (mips_base_mips16): Rename this...
18647 (mips_base_compression_flags): ...to this.
18648 (UMIPS_12BIT_OFFSET_P): New macro.
18649 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18650 (MULTILIB_DIRNAMES): Likewise.
18652 2013-02-25 Tom de Vries <tom@codesourcery.com>
18654 PR rtl-optimization/56131
18655 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18656 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18657 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18659 2013-02-25 Tobias Burnus <burnus@net-b.de>
18661 * doc/invoke.texi (-fsanitize=): Move from optimization
18662 to debugging options.
18664 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18666 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18668 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18669 Alexander Monakov <amonakov@ispras.ru>
18671 PR middle-end/56077
18672 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18673 flush pending lists also on non-jumps. Adjust comment.
18675 2013-02-24 Catherine Moore <clm@codesourcery.com>
18676 Maciej W. Rozycki <macro@codesourcery.com>
18677 Tom de Vries <tom@codesourcery.com>
18678 Nathan Sidwell <nathan@codesourcery.com>
18679 Iain Sandoe <iain@codesourcery.com>
18680 Nathan Froyd <froydnj@codesourcery.com>
18681 Chao-ying Fu <fu@mips.com>
18683 * doc/extend.texi: (micromips, nomicromips, nocompression):
18684 Document new function attributes.
18685 * doc/invoke.texi (minterlink-compressed, mmicromips,
18686 m14k, m14ke, m14kec): Document new options.
18687 (minterlink-mips16): Update documentation.
18688 * doc/md.texi (ZC, ZD): Document new constraints.
18689 * configure.ac (gcc_cv_as_micromips): Check if linker
18690 supports the .set micromips directive.
18691 * configure: Regenerate.
18692 * config.in: Regenerate.
18693 * config/mips/mips-tables.opt: Regenerate.
18694 * config/mips/micromips.md: New file.
18695 * constraints.md (ZC, AD): New constraints.
18696 * config/mips/predicates.md (movep_src_register): New predicate.
18697 (movep_src_operand): New predicate.
18698 (non_volatile_mem_operand): New predicate.
18699 * config/mips/mips.md (multimem): New type.
18700 (length): Differentiate between 17-bit and 18-bit branch offsets.
18701 (MOVEP1, MOVEP2): New mode iterator.
18702 (mov_<load>l): Use ZC constraint.
18703 (mov_<load>r): Likewise.
18704 (mov_<store>l): Likewise.
18705 (mov_<store>r): Likewise.
18706 (*branch_equality<mode>_inverted): Add microMIPS support.
18707 (*branch_equality<mode>): Likewise.
18708 (*jump_absolute): Likewise.
18709 (indirect_jump_<mode>): Likewise.
18710 (tablejump_<mode>): Likewise.
18711 (<optab>_internal): Likewise.
18712 (sibcall_internal): Likewise.
18713 (sibcall_value_internal): Likewise.
18714 (prefetch): Use constraint ZD.
18715 * config/mips/mips.opt (minterlink-compressed): New option.
18716 (minterlink-mips16): Now an alias for minterlink-compressed.
18717 (mmicromips): New option.
18718 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18719 (compare_and_swap_12): Likewise.
18720 (sync_add<mode>): Likewise.
18721 (sync_<optab>_12): Likewise.
18722 (sync_old_<optab>_12): Likewise.
18723 (sync_new_<optab>_12): Likewise.
18724 (sync_nand_12): Likewise.
18725 (sync_old_nand_12): Likewise.
18726 (sync_new_nand_12): Likewise.
18727 (sync_sub<mode>): Likewise.
18728 (sync_old_add<mode>): Likewise.
18729 (sync_old_sub<mode>): Likewise.
18730 (sync_new_add<mode>): Likewise.
18731 (sync_new_sub<mode>): Likewise.
18732 (sync_<optab><mode>): Likewise.
18733 (sync_old_<optab><mode>): Likewise.
18734 (sync_new_<optab><mode>): Likewise.
18735 (sync_nand<mode>): Likewise.
18736 (sync_old_nand<mode>): Likewise.
18737 (sync_new_nand<mode>): Likewise.
18738 (sync_lock_test_and_set<mode>): Likewise.
18739 (test_and_set_12): Likewise.
18740 (atomic_compare_and_swap<mode>): Likewise.
18741 (atomic_exchange<mode>_llsc): Likewise.
18742 (atomic_fetch_add<mode>_llsc): Likewise.
18743 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18744 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18745 (umips_save_restore_pattern_p): Likewise.
18746 (umips_load_store_pair_p): Likewise.
18747 (umips_output_load_store_pair): Likewise.
18748 (umips_movep_target_p): Likewise.
18749 (umips_12bit_offset_address_p): Likewise.
18750 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18751 (mips_base_mips16): Rename this...
18752 (mips_base_compression_flags): ...to this. Update all uses.
18753 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18754 (mips_mips16_decl_p): Delete.
18755 (mips_nomips16_decl_p): Delete.
18756 (mips_get_compress_on_flags): New function.
18757 (mips_get_compress_off_flags): New function.
18758 (mips_get_compress_mode): New function.
18759 (mips_get_compress_on_name): New function.
18760 (mips_get_compress_off_name): New function.
18761 (mips_insert_attributes): Support multiple compression types.
18762 (mips_merge_decl_attributes): Likewise.
18763 (umips_12bit_offset_address_p): New function.
18764 (mips_start_function_definition): Emit .set micromips directive.
18765 (mips_call_may_need_jalx_p): New function.
18766 (mips_function_ok_for_sibcall): Add microMIPS support.
18767 (mips_print_operand_punctuation): Support short delay slots and
18769 (umips_swm_mask, umips_swm_encoding): New.
18770 (umips_build_save_restore): New function.
18771 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18772 (was_mips16_p): Remove.
18773 (old_compression_mode): New.
18774 (mips_set_compression_mode): New function.
18775 (mips_set_current_function): Add microMIPS support.
18776 (mips_option_override): Likewise.
18777 (umips_save_restore_pattern_p): New function.
18778 (umips_output_save_restore): New function.
18779 (umips_load_store_pair_p_1): New function.
18780 (umips_load_store_pair_p): New function.
18781 (umips_output_load_store_pair_1): New function.
18782 (umips_output_load_store_pair): New function.
18783 (umips_movep_target_p) New function.
18784 (mips_prepare_pch_save): Add microMIPS support.
18785 * config/mips/mips.h (TARGET_COMPRESSION): New.
18786 (TARGET_CPU_CPP_BUILTINS): Update macro
18787 to use new compression flags and to support microMIPS.
18788 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18789 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18790 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18791 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18792 (ASM_SPEC): Support mmicromips and mno-micromips.
18793 (M16STORE_REG_P): New macro.
18794 (MIPS_CALL): Support TARGET_MICROMIPS.
18795 (MICROMIPS_J): New macro.
18796 (mips_base_mips16): Rename this...
18797 (mips_base_compression_flags): ...to this.
18798 (UMIPS_12BIT_OFFSET_P): New macro.
18799 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18800 (MULTILIB_DIRNAMES): Likewise.
18802 2013-02-24 Jakub Jelinek <jakub@redhat.com>
18805 * target-globals.c (save_target_globals): For init_reg_sets and
18806 target_reinit remporarily set this_fn_optabs to this_target_optabs.
18808 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
18810 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
18811 * config/aarch64/t-aarch64
18812 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
18814 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
18816 PR inline-asm/56148
18817 * lra-constraints.c (process_alt_operands): Reload operand
18818 conflicting with earlier clobber only if no more other conflicting
18821 2013-02-22 Jakub Jelinek <jakub@redhat.com>
18824 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
18825 if not linking a shared library.
18827 2013-02-22 Seth LaForge <sethml@google.com>
18829 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
18831 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
18833 * config/arm/arm.md (split for extendsidi): Update condition.
18834 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
18835 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
18836 (qhs_zextenddi_cstr): Likewise.
18838 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18840 PR middle-end/56420
18841 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
18842 avoid signed wrapping.
18843 (expand_mult): Handle properly multiplication by
18844 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
18845 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
18846 in the compiler if coeff is HOST_WIDE_INT_MIN.
18847 (expand_divmod): Don't make ext_op1 static, change it's type to
18848 uhwi. Avoid undefined behavior in -INTVAL (op1).
18850 PR rtl-optimization/50339
18851 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
18853 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
18854 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
18855 into splitting_ashiftrt field.
18856 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
18858 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
18861 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
18863 PR middle-end/56108
18864 * trans-mem.c (execute_tm_mark): Do not expand transactions that
18865 are sure to go irrevocable.
18867 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
18869 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
18870 scalars are valid operands.
18872 2013-02-21 Martin Jambor <mjambor@suse.cz>
18874 PR tree-optimization/56310
18875 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
18876 only matching indices and non-negative final offsets.
18877 (intersect_aggregates_with_edge): Pass src_idx to
18878 agg_replacements_to_vector. Pass src_idx insstead of index to
18879 intersect_with_agg_replacements.
18881 2013-02-21 Martin Jambor <mjambor@suse.cz>
18883 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
18884 instead of hard-wired defaults.
18886 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
18888 * doc/invoke.texi (MIPS Options): Update documentation of the
18889 floating-point multiply-accumulate instruction restrictions.
18891 2013-02-21 Kostya Serebryany <kcc@google.com>
18893 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
18894 asan_shadow_offset on x86_64 linux.
18896 2013-02-21 Richard Biener <rguenther@suse.de>
18898 PR tree-optimization/56415
18900 2013-02-11 Richard Biener <rguenther@suse.de>
18902 PR tree-optimization/56273
18903 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18906 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18909 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
18912 PR inline-asm/56405
18913 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
18914 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
18916 2013-02-20 Jan Hubicka <jh@suse.cz>
18918 PR tree-optimization/56265
18919 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
18920 when target is referenced for first time.
18922 2013-02-20 Richard Biener <rguenther@suse.de>
18924 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
18925 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
18926 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
18927 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
18928 not return anything.
18929 (rename_ssa_copies): Do not remove unused locals.
18930 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
18931 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
18932 * passes.c (execute_function_todo): Do not schedule unused locals
18933 removal if cleanup_tree_cfg did something.
18934 * tree-ssa-live.c (remove_unused_locals): Dump statistics
18935 about the number of removed locals.
18937 2013-02-20 Richard Biener <rguenther@suse.de>
18939 PR tree-optimization/56398
18940 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
18942 2013-02-20 Martin Jambor <mjambor@suse.cz>
18944 PR tree-optimization/55334
18945 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
18946 restricted pointers to arrays.
18948 2013-02-20 Richard Biener <rguenther@suse.de>
18949 Jakub Jelinek <jakub@redhat.com>
18951 PR tree-optimization/56396
18952 * tree-ssa-ccp.c (n_const_val): New static variable.
18953 (get_value): Return NULL for SSA names we don't have a lattice
18955 (ccp_initialize): Initialize n_const_val.
18956 * tree-ssa-copy.c (n_copy_of): New static variable.
18957 (init_copy_prop): Initialize n_copy_of.
18958 (get_value): Return NULL_TREE for SSA names we don't have a
18961 2013-02-20 Martin Jambor <mjambor@suse.cz>
18963 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
18965 2013-02-20 Richard Biener <rguenther@suse.de>
18967 * genpreds.c (write_lookup_constraint): Do not compare first
18968 letter of the constraint again.
18970 2013-02-20 Richard Biener <rguenther@suse.de>
18972 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
18974 (get_use_iv_cost): Terminate hashtable walk when coming across
18977 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
18979 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
18980 reassociation for avx2 targets.
18982 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18984 * config/microblaze/microblaze.c: microblaze_has_clz = 0
18985 Add version check for v8.10.a to enable microblaze_has_clz
18986 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
18987 version and TARGET_PATTERN_COMPARE check
18988 * config/microblaze/microblaze.md: New clzsi2 instruction
18990 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18992 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
18993 function before branching.
18995 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
18997 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
18999 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
19001 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19003 PR middle-end/55889
19004 * sel-sched.c: Include ira.h.
19005 (implicit_clobber_conflict_p): New function.
19006 (moveup_expr): Use it.
19007 * Makefile.in (sel-sched.o): Depend on ira.h.
19009 2013-02-19 Richard Biener <rguenther@suse.de>
19011 PR tree-optimization/56384
19012 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
19013 (vn_hash_type): Split out from ...
19014 (vn_hash_constant_with_type): ... here.
19015 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
19016 (vn_phi_eq): Compare types from vn_phi_s structure.
19017 (vn_phi_lookup): Populate vn_phi_s type.
19018 (vn_phi_insert): Likewise.
19020 2013-02-19 Jakub Jelinek <jakub@redhat.com>
19022 PR tree-optimization/56350
19023 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
19024 if haven't found reduction or nested cycle operand, rather than
19025 asserting we must find it.
19027 PR tree-optimization/56381
19028 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
19031 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
19032 Jakub Jelinek <jakub@redhat.com>
19035 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
19036 (swap_optab_enable): Same.
19037 (init_all_optabs): Use argument instead of global.
19038 * tree.h (struct tree_optimization_option): New field target_optabs.
19039 * expr.h (init_all_optabs): Add argument to prototype.
19040 (TREE_OPTIMIZATION_OPTABS): New.
19041 (save_optabs_if_changed): Protoize.
19042 * optabs.h: Declare this_fn_optabs.
19043 * optabs.c (save_optabs_if_changed): New.
19044 Declare this_fn_optabs.
19045 (init_optabs): Add argument to init_all_optabs() call.
19046 * function.c (invoke_set_current_function_hook): Handle per
19048 * function.h (struct function): New field optabs.
19049 * config/mips/mips.c (mips_set_mips16_mode): Handle when
19050 optimization_current_node has changed.
19051 * target-globals.h (save_target_globals_default_opts): Protoize.
19052 * target-globals.c (save_target_globals_default_opts): New.
19054 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19057 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
19058 registers %fr12 and %fr12R as call used.
19061 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
19062 and HImode, require all displacements to be an integer multiple of
19064 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
19065 only allow QImode and HImode when reload is in progress and strict is
19066 true. Likewise for symbolic addresses. Use base14_operand to check
19067 displacements in REG+BASE addresses.
19069 2013-02-18 Richard Biener <rguenther@suse.de>
19071 PR tree-optimization/56366
19072 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19073 handle sign-conversion of outer-loop initial induction value.
19075 2013-02-18 Richard Biener <rguenther@suse.de>
19077 PR middle-end/56349
19078 * cfghooks.c (merge_blocks): If we merge a latch into another
19079 block adjust references to it.
19080 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
19081 (verify_loop_structure): Verify that a recorded latch is in fact
19084 2013-02-18 Richard Biener <rguenther@suse.de>
19086 PR tree-optimization/56321
19087 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
19088 order SSA name release and virtual operand unlinking.
19090 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19092 * config/microblaze/microblaze.md (save_stack_block): Define.
19093 (restore_stack_block): Likewise.
19095 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19097 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
19098 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
19099 * config/microblaze/microblaze.c (microblaze_option_override):
19100 Bail out early for PIC modes when target does not support PIC.
19102 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19104 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
19105 Replace with a microblaze version.
19106 (microblaze_trampoline_init): Adapt for microblaze.
19107 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
19110 2013-02-16 Jakub Jelinek <jakub@redhat.com>
19111 Dodji Seketeli <dodji@redhat.com>
19114 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
19115 (instrument_mem_region_access): Do not forget to always put
19116 instrumentation of the of 'base' and 'base + len' in a "if (len !=
19117 0) statement, even for cases where either 'base' or 'base + len'
19118 are not instrumented -- because they have been previously
19119 instrumented. Simplify the logic by putting all the statements
19120 instrument 'base + len' inside a sequence, and then insert that
19121 sequence right before the current insertion point. Then, to
19122 instrument 'base + len', just get an iterator on that statement.
19123 And do not forget to update the pointer to iterator the function
19124 received as argument.
19126 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
19128 PR rtl-optimization/56348
19129 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
19131 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
19133 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
19134 (clean_graph_dump_file): Pass base to start_graph_dump.
19136 2013-02-14 Richard Henderson <rth@redhat.com>
19139 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
19141 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
19143 * collect2-aix.h: Define F_LOADONLY.
19145 2013-02-14 Richard Biener <rguenther@suse.de>
19148 * varasm.c (output_constant_def_1): Get the decl representing
19149 the constant as argument.
19150 (output_constant_def): Wrap output_constant_def_1.
19151 (make_decl_rtl): Use output_constant_def_1 with the decl
19152 representing the constant.
19153 (build_constant_desc): Optionally re-use a decl already
19154 representing the constant.
19155 (tree_output_constant_def): Adjust.
19157 2013-02-14 Dodji Seketeli <dodji@redhat.com>
19160 * asan.c (instrument_builtin_call): Really put the length of the
19161 second source argument into src1_len.
19163 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19165 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
19166 argument. If it is false, don't create edge from then_bb to
19168 (insert_if_then_before_iter): Pass true to it.
19169 (build_check_stmt): Pass false to it.
19170 (transform_statements): Flush hash table only on extended basic
19171 block boundaries, rather than at the beginning of every bb.
19172 Don't flush hash table on nonfreeing_call_p calls.
19173 * tree-flow.h (nonfreeing_call_p): New prototype.
19174 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
19176 2013-02-13 David S. Miller <davem@davemloft.net>
19178 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
19180 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
19183 * ira.c (max_regno_before_ira): Move from ...
19185 (fix_reg_equiv_init): Use max_regno_before_ira instead of
19188 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19190 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
19192 2013-02-13 Richard Biener <rguenther@suse.de>
19195 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
19196 globals in MEM_REFs.
19198 2013-02-13 Richard Biener <rguenther@suse.de>
19200 * loop-init.c (loop_optimizer_init): Clear loop state when
19201 re-initializing preserved loops.
19202 * loop-unswitch.c (unswitch_single_loop): Return whether
19203 we unswitched the loop. Do not verify loop state here.
19204 (unswitch_loops): When we unswitched a loop discover new loops.
19206 2013-02-13 Kostya Serebryany <kcc@google.com>
19208 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
19210 * sanitizer.def: Rename __asan_init to __asan_init_v1.
19212 2013-02-12 Dodji Seketeli <dodji@redhat.com>
19214 Avoid instrumenting duplicated memory access in the same basic block
19215 * Makefile.in (asan.o): Add new dependency on hash-table.h
19216 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
19217 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
19218 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
19219 (free_mem_ref_resources, has_mem_ref_been_instrumented)
19220 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
19221 (get_mem_ref_of_assignment): New functions.
19222 (get_mem_refs_of_builtin_call): Extract from
19223 instrument_builtin_call and tweak a little bit to make it fit with
19225 (instrument_builtin_call): Use the new
19226 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
19227 of is_gimple_builtin_call.
19228 (instrument_derefs, instrument_mem_region_access): Insert the
19229 instrumented memory reference into the hash table.
19230 (maybe_instrument_assignment): Renamed instrument_assignment into
19231 this, and change it to advance the iterator when instrumentation
19232 actually happened and return true in that case. This makes it
19233 homogeneous with maybe_instrument_assignment, and thus give a
19234 chance to callers to be more 'regular'.
19235 (transform_statements): Clear the memory reference hash table
19236 whenever we enter a new BB, when we cross a function call, or when
19237 we are done transforming statements. Use
19238 maybe_instrument_assignment instead of instrumentation. No more
19239 need to special case maybe_instrument_assignment and advance the
19240 iterator after calling it; it's now handled just like
19241 maybe_instrument_call. Update comment.
19243 2013-02-13 Richard Biener <rguenther@suse.de>
19245 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
19246 Fix loop discovery code.
19248 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
19250 PR inline-asm/56148
19251 * lra-constraints.c (process_alt_operands): Match early clobber
19252 operand with itself. Check conflicts with earlyclobber only if
19253 the operand is not reloaded. Prefer to reload conflicting operand
19254 if earlyclobber and matching operands are the same.
19256 2013-02-12 Richard Biener <rguenther@suse.de>
19259 * lto-streamer-out.c (write_symbol): Do not output symbols
19260 for hard register variables.
19262 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
19265 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
19266 (umulsidi3_insn, mulsidi3_insn): New insns.
19268 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
19270 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
19271 (struct tune_params): Add vec_costs field.
19272 * config/arm/arm.c (arm_builtin_vectorization_cost)
19273 (arm_add_stmt_cost): New functions.
19274 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
19275 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
19276 (arm_default_vec_cost): New struct of type cpu_vec_costs.
19277 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
19278 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
19279 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
19280 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
19282 2013-02-12 Richard Biener <rguenther@suse.de>
19285 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
19286 decls again if possible.
19288 2013-02-12 Richard Biener <rguenther@suse.de>
19290 PR middle-end/56288
19291 * tree-ssa.c (verify_ssa_name): Fix check, move
19292 SSA_NAME_IN_FREE_LIST check up.
19294 2013-02-12 Jakub Jelinek <jakub@redhat.com>
19295 Steven Bosscher <steven@gcc.gnu.org>
19297 PR rtl-optimization/56151
19298 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
19299 equal to op0 or op1, and last_insn pattern is CODE operation
19300 with MEM dest and one of the operands matches that MEM.
19302 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
19304 * doc/extend.texi: Document Function Multiversioning and "default"
19305 parameter string to target attribute.
19306 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
19307 target attribute parameter is "default".
19308 (ix86_compare_version_priority): Remove checks for target attribute.
19309 (ix86_mangle_function_version_assembler_name): Change error to sorry.
19310 Remove check for target attribute equal to NULL. Add assert.
19311 (ix86_generate_version_dispatcher_body): Change error to sorry.
19313 2013-02-11 Iain Sandoe <iain@codesourcery.com>
19314 Jack Howarth <howarth@bromo.med.uc.edu>
19315 Patrick Marlier <patrick.marlier@gmail.com>
19318 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
19319 define ENDFILE_SPEC as TM_DESTRUCTOR.
19320 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
19322 2013-02-11 Alexander Potapenko <glider@google.com>
19323 Jack Howarth <howarth@bromo.med.uc.edu>
19324 Jakub Jelinek <jakub@redhat.com>
19327 * config/darwin.c (cdtor_record): Rename ctor_record.
19328 (sort_cdtor_records): Rename sort_ctor_records.
19329 (finalize_dtors): New routine to sort destructors by
19330 priority before use in assemble_integer.
19331 (machopic_asm_out_destructor): Use finalize_dtors if needed.
19333 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
19335 PR rtl-optimization/56275
19336 * simplify-rtx.c (avoid_constant_pool_reference): Check that
19337 offset is non-negative and less than cmode size before
19338 calling simplify_subreg.
19340 2013-02-11 Richard Biener <rguenther@suse.de>
19342 PR tree-optimization/56264
19343 * cfgloop.h (fix_loop_structure): Adjust prototype.
19344 * loop-init.c (fix_loop_structure): Return the number of
19345 newly discovered loops.
19346 * tree-cfgcleanup.c (repair_loop_structures): When new loops
19347 are discovered, do a full loop-closed SSA rewrite.
19349 2013-02-11 Richard Biener <rguenther@suse.de>
19351 PR tree-optimization/56273
19352 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19354 (check_array_ref): Fix missing newline in dumps.
19355 (search_for_addr_array): Likewise.
19357 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
19359 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
19361 2013-02-09 Jakub Jelinek <jakub@redhat.com>
19364 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
19366 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
19368 PR rtl-optimization/56246
19369 * lra-constraints.c (simplify_operand_subreg): Try to reuse
19371 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
19372 constraints are satisfied.
19374 2013-02-08 Jeff Law <law@redhat.com>
19377 * emit-rtl.c (reg_is_parm_p): New function.
19378 * regs.h (reg_is_parm_p): New prototype.
19379 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
19380 callee-clobbered registers.
19382 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
19385 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
19386 If there is no implicit builtin declaration, just return NULL.
19388 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19390 * config/i386/sse.md (FMAMODEM): New mode iterator.
19391 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
19392 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
19394 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19396 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
19397 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
19398 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
19400 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19402 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
19403 (microblaze*-*-elf): Likewise.
19404 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
19406 * config/microblaze/microblaze-c.c: Add builtin defines for
19407 _LITTLE_ENDIAN and _BIG_ENDIAN.
19408 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
19409 add to TARGET_DEFAULT flags.
19410 Expand ASM_SPEC and LINK_SPEC.
19411 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
19412 * config/microblaze/microblaze.md: Update extendsidi2 and
19413 movdi_internal instructions to use low-order / high-order reg
19415 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
19416 options and inversemask / mask of LITTLE_ENDIAN.
19417 * config/microblaze/t-microblaze: Expand multilib options to
19418 include mlittle-endian (le) and update exceptions patterns.
19420 2013-02-08 Jakub Jelinek <jakub@redhat.com>
19422 PR rtl-optimization/56195
19423 * lra-constraints.c (get_reload_reg): Don't reuse regs
19424 if they have smaller mode than requested, if they have
19425 wider mode than requested, try to return a SUBREG.
19427 PR tree-optimization/56250
19428 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
19429 if type is unsigned and code isn't MULT_EXPR.
19431 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19433 PR tree-optimization/56064
19434 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
19435 bits according to mode.
19436 * fixed-value.h (fixed_from_double_int)
19437 (const_fixed_from_double_int): Adjust comments.
19439 2013-02-08 Richard Biener <rguenther@suse.de>
19442 * lto-streamer.h (struct data_in): Remove current_file, current_line
19443 and current_col members.
19444 * lto-streamer-out.c (lto_output_location): Stream changed bits
19445 en-block for efficiency.
19446 * lto-streamer-in.c (clear_line_info): Remove.
19447 (lto_input_location): Cache current file, line and column
19448 globally via local statics. Read changed bits en-block.
19449 (input_function): Do not call clear_line_info.
19450 (lto_read_body): Likewise.
19451 (lto_input_toplevel_asms): Likewise.
19453 2013-02-08 Michael Matz <matz@suse.de>
19455 PR tree-optimization/52448
19456 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19457 (nt_call_phase): New static.
19458 (add_or_mark_expr): Only mark accesses with newer phase than any
19460 (nonfreeing_call_p): New.
19461 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19462 (nt_fini_block): Keep blocks marked as visited.
19463 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19465 2013-02-08 Richard Biener <rguenther@suse.de>
19467 * ira.c (ira): Free broken dominator information.
19469 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19471 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19473 2013-02-08 Marek Polacek <polacek@redhat.com>
19475 * cfgloop.c (verify_loop_structure): Add more checking of headers.
19477 2013-02-08 Richard Biener <rguenther@suse.de>
19479 PR middle-end/56181
19480 * cfgloop.h (flow_loops_find): Adjust.
19481 (bb_loop_header_p): Declare.
19482 * cfgloop.c (bb_loop_header_p): New function split out from ...
19483 (flow_loops_find): ... here. Adjust function signature,
19484 support incremental loop structure update.
19485 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19486 * cfgloopmanip.c (fix_loop_structure): Move ...
19487 * loop-init.c (fix_loop_structure): ... here.
19488 (apply_loop_flags): Split out from ...
19489 (loop_optimizer_init): ... here.
19490 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19491 in incremental mode, only remove dead loops here.
19493 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19496 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19497 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19498 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19499 (*round<mode>3.libgcc): New insns for fixed-modes.
19500 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19501 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19502 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19503 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19504 implementations. Define to __builtin_avr_absFX,
19505 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19506 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19507 __builtin_avr_countlsFX, respectively.
19508 * config/avr/avr-c.c (target.h): Include it.
19509 (enum avr_builtin_id): New enum.
19510 (avr_resolve_overloaded_builtin): New static function.
19511 (avr_register_target_pragmas): Use it to set
19512 targetm.resolve_overloaded_builtin.
19513 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19514 tree nodes used by DEF_BUILTIN.
19515 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19516 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19517 <AVR_BUILTIN_xxBITS>: Same.
19519 2013-02-08 Richard Biener <rguenther@suse.de>
19521 * cfgloop.c (verify_loop_structure): Properly handle
19522 a loop exiting to another loop header.
19523 * ira-int.h (ira_loops): Remove.
19524 * ira.c (ira_loops): Remove.
19525 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19526 (do_reload): Use loop_optimizer_finalize.
19527 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19528 number_of_loops to access the loop tree.
19529 (more_one_region_p): Likewise.
19530 (finish_loop_tree_nodes): Likewise.
19531 (rebuild_regno_allocno_maps): Likewise.
19532 (mark_loops_for_removal): Likewise.
19533 (mark_all_loops_for_removal): Likewise.
19534 (remove_unnecessary_regions): Likewise.
19535 (ira_build): Likewise.
19536 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19538 2013-02-08 Richard Biener <rguenther@suse.de>
19540 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
19541 * ipa-pure-const.c (analyze_function): Avoid calling
19542 mark_irreducible_loops twice.
19543 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
19545 2013-02-07 David S. Miller <davem@davemloft.net>
19547 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
19549 * var-tracking.c (vt_add_function_parameter): Test the presence of
19550 HAVE_window_save properly and do not remap argument registers when
19551 we have a leaf function.
19553 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
19556 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
19558 * config/i386/i386.c (ix86_print_operand): Ditto.
19560 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19562 * lra-constraints.c (process_alt_operands): Fix recently added comment.
19564 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19566 PR rtl-optimization/56225
19567 * lra-constraints.c (process_alt_operands): Check that reload hard
19568 reg can hold value for strict_low_part.
19570 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19573 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
19574 dwarf2out_end_function.
19575 (in_first_function_p, maybe_at_text_label_p,
19576 first_loclabel_num_not_at_text_label): New variables.
19577 (dwarf2out_var_location): In the first function find out
19578 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
19579 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
19582 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
19584 PR rtl-optimization/56178
19585 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
19586 SUBREG of a register. Tidy up related block of code.
19587 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
19588 note if the source is a register or a SUBREG of a register.
19590 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19593 * config/rs6000/rs6000.md (ptrm): New mode attr.
19594 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19595 call_value_indirect_aix<pttrsize>,
19596 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19599 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19601 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19602 if -bnortl. Convert to strcmp and strncmp.
19604 2013-02-07 Alan Modra <amodra@gmail.com>
19607 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19608 addresses won't wrap when offsetting.
19609 (rs6000_secondary_reload): Provide secondary reloads needed for
19610 wrapping LO_SUM addresses.
19612 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19614 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19615 MACH, just __MACH__.
19617 2013-02-06 Richard Biener <rguenther@suse.de>
19619 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19620 instead of calling fix_loop_structure.
19622 2013-02-06 Jakub Jelinek <jakub@redhat.com>
19624 PR middle-end/56217
19625 * omp-low.c (use_pointer_for_field): Return false if
19626 lower_send_shared_vars doesn't generate any copy-out code.
19628 2013-02-06 Tom de Vries <tom@codesourcery.com>
19630 PR rtl-optimization/56131
19631 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19632 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19633 of the label is NULL. Add comment.
19635 2013-02-05 Jakub Jelinek <jakub@redhat.com>
19637 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19640 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19641 (STATIC_LIBTSAN_LIBS): Likewise.
19642 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19643 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19644 is defined, don't add anything else beyond that.
19645 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19646 (LINK_COMMAND_SPEC): Use them.
19648 PR tree-optimization/56205
19649 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19650 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19651 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19653 2013-02-05 Richard Biener <rguenther@suse.de>
19655 PR tree-optimization/53342
19656 PR tree-optimization/53185
19657 * tree-vectorizer.h (vect_check_strided_load): Remove.
19658 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19659 not disallow peeling for vectorized strided loads.
19660 (vect_check_strided_load): Make static and simplify.
19661 (vect_analyze_data_refs): Adjust.
19662 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19663 correctly when vectorizing strided loads.
19665 2013-02-05 Richard Biener <rguenther@suse.de>
19667 * doc/install.texi: Refer to ISL, not PPL.
19669 2013-02-05 Jan Hubicka <jh@suse.cz>
19671 PR tree-optimization/55789
19672 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19674 2013-02-05 Jan Hubicka <jh@suse.cz>
19676 PR tree-optimization/55789
19677 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19678 the dead call anyway.
19680 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19683 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19685 2013-02-04 Alexander Potapenko <glider@google.com>
19686 Jack Howarth <howarth@bromo.med.uc.edu>
19687 Jakub Jelinek <jakub@redhat.com>
19690 * config/darwin.c (sort_ctor_records): Stabilized qsort
19691 on constructor priority by using original position.
19692 (finalize_ctors): New routine to sort constructors by
19693 priority before use in assemble_integer.
19694 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19696 2013-02-04 Jakub Jelinek <jakub@redhat.com>
19699 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19700 about visibility on artificial decls.
19701 * config/sol2.c (solaris_assemble_visibility): Likewise.
19703 2013-02-04 Kai Tietz <ktietz@redhat.com>
19706 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19707 argument and improve checking of return-argument types for 16-byte
19709 (ix86_function_value_1): Add additional valtype argument on call
19710 of function_value_64.
19711 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19712 handling infunction_value_64 function.
19714 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19716 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
19718 2013-02-04 Richard Biener <rguenther@suse.de>
19720 PR tree-optimization/56188
19721 * tree-ssa-structalias.c (label_visit): Consider case with
19722 initially non-empty points-to set.
19723 (perform_var_substitution): Dump node mapping and clean up.
19725 2013-02-04 Richard Guenther <rguenther@suse.de>
19728 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19729 node prevail as last resort.
19730 (lto_symtab_merge_decls): Remove guard on LTRANS here.
19731 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
19733 2013-02-04 Richard Biener <rguenther@suse.de>
19735 PR tree-optimization/56113
19736 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19738 (equiv_class_lookup_or_add): ... this.
19739 (label_visit): Adjust and fix error in previous patch.
19740 (perform_var_substitution): Adjust.
19742 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19744 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19745 * config/sh/sh4-300.md: Likewise.
19746 * config/sh/sh4a.md: Likewise.
19747 * config/sh/constraints.md: Likewise.
19748 * config/sh/sh.md: Likewise.
19749 * config/sh/netbsd-elf.h: Likewise.
19750 * config/sh/predicates.md: Likewise.
19751 * config/sh/sh-protos.h: Likewise.
19752 * config/sh/ushmedia.h: Likewise.
19753 * config/sh/linux.h: Likewise.
19754 * config/sh/sh.c: Likewise.
19755 * config/sh/superh.h: Likewise.
19756 * config/sh/elf.h: Likewise.
19757 * config/sh/sh4.md: Likewise.
19758 * config/sh/sh.h: Likewise.
19760 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19762 * config/pa/constraints.md: Adjust unused letters. Change "T"
19763 constraint to match_test floating_point_store_memory_operand().
19764 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19765 (base14_operand): New.
19766 (floating_point_store_memory_operand): New.
19767 (integer_store_memory_operand): Revise to use base14_operand and
19768 reg_plus_base_memory_operand.
19769 (move_dest_operand): Allow symbolic_memory_operands.
19770 (symbolic_memory_operand): Check for LO_SOM.
19771 (symbolic_operand): Change default case to break.
19772 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19773 CONST_DOUBLE values to be reloaded by putting them into memory when
19774 the destination is a floating point register.
19775 (movdf): Remove code to handle CONST_DOUBLE.
19777 (reload_indf_r1): New.
19778 (reload_insf_r1): New.
19779 Consistently use "Q" and "T" constraints with integer and floating
19780 point move instructions, respectively.
19781 (movdi): Remove FAIL.
19782 Change predicate for source operand unamed DImode move from
19783 general_operand to move_src_operand.
19784 (umulsidi3): Change predicate for destination operand to
19786 Likewise for similar unamed patterns.
19787 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19788 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19789 (hppa_legitimize_address): Simplify mask calculation.
19790 (pa_emit_move_sequence): Revised handling of secondary reloads from
19791 REG+D addresses for floating point loads and stores. Directly handle
19792 loading CONST0_RTX (mode) to a floating point register.
19793 (pa_secondary_reload): Handle reloading DF and SFmode constant values
19794 to floating point registers. Don't restrict secondary reloads to
19795 floating point registers to integer modes. Revise some comments and
19797 (TARGET_LEGITIMATE_ADDRESS_P): Define.
19798 (pa_legitimate_address_p): New.
19799 (pa_legitimize_reload_address): New.
19800 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
19801 (STRICT_REG_OK_FOR_BASE_P): New.
19802 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
19803 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
19805 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
19806 Andrew Dixie <andrewd@gentrack.com>
19808 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
19811 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
19813 * expmed.c (extract_bit_field_1): Pass the full width of the
19814 structure to get_best_reg_extraction_insn.
19816 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19819 * configure.ac (use_cxa_atexit): Add AIX.
19820 * configure: Regenerate.
19822 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
19824 2013-02-01 Jakub Jelinek <jakub@redhat.com>
19827 * final.c (need_profile_function): New variable.
19828 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
19829 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
19830 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
19831 notes, targetm.asm_out.function_prologue doesn't emit anything,
19832 HAVE_prologue and profiler should be emitted before prologue,
19833 set need_profile_function instead of emitting it.
19834 (final_scan_insn): If need_profile_function, emit
19835 profile_function on the first NOTE_INSN_BASIC_BLOCK or
19836 NOTE_INSN_FUNCTION_BEG note.
19838 2013-02-01 Richard Henderson <rth@redhat.com>
19840 * config/rs6000/rs6000.md (smulditi3): New.
19843 * config/alpha/alpha.md (umulditi3): New.
19845 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19847 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
19848 (ASM_OUTPUT_ALIGNED_LOCAL): New.
19850 2013-02-01 Richard Biener <rguenther@suse.de>
19852 PR tree-optimization/56113
19853 * tree-ssa-structalias.c (label_visit): Reduce work for
19854 single-predecessor nodes.
19856 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
19858 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
19859 range isn't testing for zero.
19861 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
19863 PR middle-end/56113
19864 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
19866 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
19867 Nick Clifton <nickc@redhat.com>
19869 * config/v850/constraints.md (Q): Define as a memory constraint.
19870 * config/v850/predicates.md (label_ref_operand): New predicate.
19871 (e3v5_shift_operand): New predicate.
19872 (ior_operator): New predicate.
19873 * config/v850/t-v850: Add e3v5 multilib.
19874 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
19875 (v850_gen_movdi): Prototype.
19876 * config/v850/v850.c: Add support for e3v5 architecture.
19877 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
19879 (construct_save_jarl): Add e3v5 long JARL support.
19880 (v850_adjust_insn_length): New function. Adjust length of call
19881 insns when using e3v5 instructions.
19882 (v850_gen_movdi): New function: Generate instructions to move a
19884 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
19885 (CPP_SPEC): Define __v850e3v5__ as appropriate.
19886 (TARGET_USE_FPU): Enable for e3v5.
19887 (CONST_OK_FOR_W): New macro.
19888 (ADJUST_INSN_LENGTH): Define.
19889 * config/v850/v850.md (UNSPEC_LOOP): Define.
19890 (attr cpu): Add v850e3v5.
19891 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
19892 (movdi): New pattern.
19893 (movdi_internal): New pattern.
19894 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
19895 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
19896 (cstoresf4): Likewise.
19897 (cstoredf4): Likewise.
19898 (insv): New pattern.
19899 (rotlso3_a): New pattern.
19900 (rotlsi3_b): New pattern
19901 (rotlsi3_v850e3v5): New pattern.
19902 (doloop_begin): New pattern.
19903 (fix_loop_counter): New pattern.
19904 (doloop_end): New pattern.
19905 (branch_normal): Add e3v5 long branch support.
19906 (branch_invert): Likewise.
19907 (branch_z_normal): Likewise.
19908 (branch_z_invert): Likewise.
19909 (branch_nz_normal): Likewise.
19910 (branch_nz_invert): Likewise.
19911 (call_internal_short): Add e3v5 register-indirect JARL support.
19912 (call_internal_long): Likewise.
19913 (call_value_internal_short): Likewise.
19914 (call_value_internal_long): Likewise.
19915 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
19916 (mloop): New option.
19917 * config.gcc: Add support for configuring v840e3v5 target.
19918 * doc/invoke.texi: Document new v850 specific command line options.
19920 2013-01-31 Paul Koning <ni1d@arrl.net>
19924 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
19925 children if parent is a class.
19926 (prune_unused_types_prune): Don't add DW_AT_declaration.
19928 2013-01-31 Richard Biener <rguenther@suse.de>
19930 PR tree-optimization/56157
19931 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
19932 match up operand with SLP child.
19934 2013-01-31 Jason Merrill <jason@redhat.com>
19937 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
19938 parameters the first time.
19939 (gen_scheduled_generic_parms_dies): Check completeness here.
19941 2013-01-31 Richard Biener <rguenther@suse.de>
19943 PR middle-end/53073
19944 * common.opt (faggressive-loop-optimizations): New flag,
19945 enabled by default.
19946 * doc/invoke.texi (faggressive-loop-optimizations): Document.
19947 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
19948 infer_loop_bounds_from_undefined by it.
19950 2013-01-31 Richard Biener <rguenther@suse.de>
19952 PR tree-optimization/56150
19953 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
19954 visit virtual operands.
19955 (find_uses_to_rename_bb): Likewise.
19957 2013-01-31 Richard Biener <rguenther@suse.de>
19959 PR tree-optimization/56150
19960 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
19961 mixed store non-store stmts.
19963 2013-01-30 Jakub Jelinek <jakub@redhat.com>
19966 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
19967 LIBASAN_EARLY_SPEC is defined.
19968 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
19969 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
19971 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
19974 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
19975 invalid args instead of ICEing on it.
19976 (ix86_valid_target_attribute_tree): Return error_mark_node if
19977 ix86_valid_target_attribute_inner_p failed.
19978 (ix86_valid_target_attribute_p): Return false only if
19979 ix86_valid_target_attribute_tree returned error_mark_node. Allow
19980 target("default") attribute.
19981 (sorted_attr_string): Change argument from const char * to tree,
19982 merge in all target attribute arguments rather than just one.
19983 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
19984 instead of free. Avoid using strcat.
19985 (ix86_mangle_function_version_assembler_name): Mangle
19986 target("default") as if no target attribute is present. Adjust
19987 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
19988 instead of xmalloc and XDELETEVEC instead of free.
19989 (ix86_function_versions): Don't return true if one of the decls
19990 doesn't have target attribute. If they don't and one of the decls
19991 is DECL_FUNCTION_VERSIONED, report an error. Adjust
19992 sorted_attr_string caller. Use XDELETEVEC instead of free.
19993 (ix86_supports_function_versions): Remove.
19994 (make_name): Fix up formatting.
19995 (make_dispatcher_decl): Remove resolver_name and its initialization.
19996 Avoid leaking memory.
19997 (is_function_default_version): Return true if there is
19998 target("default") attribute rather than no target attribute at all.
19999 (make_resolver_func): Avoid leaking memory.
20000 (ix86_generate_version_dispatcher_body): Likewise.
20001 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
20002 * target.def (supports_function_versions): Remove.
20003 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
20004 * doc/tm.texi: Regenerated.
20006 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
20008 PR rtl-optimization/56144
20009 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
20010 for values with side effects.
20012 2013-01-30 Richard Biener <rguenther@suse.de>
20014 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
20015 (sparseset_pop): Likewise.
20016 * cfganal.c (compute_idf): Likewise. Increase work-stack size
20017 to be able to use quick_push in the worker loop.
20019 2013-01-30 Marek Polacek <polacek@redhat.com>
20021 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
20023 2013-01-30 Richard Biener <rguenther@suse.de>
20026 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
20028 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
20030 PR tree-optimization/56064
20031 * fixed-value.c (fixed_from_double_int): New function.
20032 * fixed-value.h (fixed_from_double_int): New prototype.
20033 (const_fixed_from_double_int): New static inline function.
20034 * fold-const.c (native_interpret_fixed): New static function.
20035 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
20036 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
20037 (native_encode_fixed): New static function.
20038 (native_encode_expr) <FIXED_CST>: Use it.
20039 (native_interpret_int): Move double_int worker code to...
20040 * double-int.c (double_int::from_buffer): ...this new static method.
20041 * double-int.h (double_int::from_buffer): Prototype it.
20043 2013-01-30 Richard Biener <rguenther@suse.de>
20045 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
20046 New pointer-map and obstack.
20047 (init_alias_vars): Allocate pointer-map and obstack.
20048 (delete_points_to_sets): Free them.
20049 (find_what_var_points_to): Cache result.
20050 (find_what_p_points_to): Adjust for changed interface of
20051 find_what_var_points_to.
20052 (compute_points_to_sets): Likewise.
20053 (ipa_pta_execute): Likewise.
20055 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20057 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
20058 * configure: Regenerate.
20059 * config.in: Regenerate.
20060 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
20061 #nobits/#progbits if supported.
20063 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
20066 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
20067 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
20068 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
20070 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20072 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
20073 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
20075 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20077 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
20078 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
20080 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20082 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
20084 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
20085 * config/arm/cortex-a7.md: New bypasses using
20086 arm_mac_accumulator_is_result.
20088 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20090 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
20091 (cortex_a7_neon_mla): Likewise.
20092 (cortex_a7_fpfmad): New reservation.
20093 (cortex_a7_fpmacs): Use ffmas and update required units.
20094 (cortex_a7_fpmuld): Update required units and latency.
20095 (cortex_a7_fpmacd): Likewise.
20096 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
20097 (cortex_a7_neon). Likewise.
20098 (bypass) Update participating units.
20100 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20102 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
20103 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
20105 * config/arm/vfp11.md (vfp_farith): Use ffmas.
20106 (vfp_fmul): Use ffmad.
20107 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
20108 (cortex_r4_fmacd): Use ffmad.
20109 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
20110 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
20111 (cortex_a9_fmacd): Use ffmad.
20112 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
20113 (cortex_a8_vfp_macd): Use ffmad.
20114 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
20115 (cortex_a5_fpmacd): Use ffmad.
20116 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
20117 (cortex_a15_vfp_macd): Use ffmad.
20118 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
20120 2013-01-29 Jason Merrill <jason@redhat.com>
20123 * varasm.c (default_assemble_visibility): Don't warn about
20124 visibility on artificial decls.
20126 2013-01-29 Richard Biener <rguenther@suse.de>
20128 PR tree-optimization/56113
20129 * tree-ssa-structalias.c (equiv_class_lookup): Also return
20131 (label_visit): Free duplicate bitmaps and record the leader instead.
20132 (perform_var_substitution): Adjust.
20134 2013-01-29 Richard Biener <rguenther@suse.de>
20136 PR tree-optimization/55270
20137 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
20138 the CFG, schedule loops for fixup.
20140 2013-01-29 Nick Clifton <nickc@redhat.com>
20142 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
20145 2013-01-28 Leif Ekblad <leif@rdos.net>
20147 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
20148 * config/i386/i386.h (TARGET_RDOS): New macro.
20149 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
20150 * config/i386/i386.c (ix86_option_override_internal): For 64bit
20151 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
20152 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
20153 DEFAULT_LARGE_SECTION_THRESHOLD.
20154 * config/i386/i386.md (R14_REG, R15_REG): New constants.
20155 * config/i386/rdos.h: New file.
20156 * config/i386/rdos64.h: New file.
20158 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
20161 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
20164 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20166 PR rtl-optimization/56117
20167 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
20168 call cselib_lookup_from_insn on the MEM before calling
20169 add_insn_mem_dependence.
20171 2013-01-28 Richard Biener <rguenther@suse.de>
20173 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
20174 to a stmt that didn't have one.
20175 (copy_phis_for_bb): Likewise for PHI arguments.
20176 (copy_debug_stmt): Likewise for debug stmts.
20178 2013-01-28 Richard Biener <rguenther@suse.de>
20180 PR tree-optimization/56034
20181 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
20182 (partition_builtin_p): Adjust.
20183 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
20184 it is the last partition.
20185 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
20186 up the vertex for the definition.
20187 (classify_partition): Classify whether a partition is a
20188 PKIND_REDUCTION, thus has uses outside of the loop.
20189 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
20190 Merge all PKIND_REDUCTION partitions into the last partition.
20191 (tree_loop_distribution): Seed partitions from reductions as well.
20193 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20195 PR tree-optimization/56125
20196 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
20197 pow(x,c) into sqrt(x) * powi(x, n/2) or
20198 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
20199 optimizing for size.
20200 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
20201 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
20204 PR tree-optimization/56094
20205 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
20206 to UNKNOWN_LOCATION while gimplifying expr.
20208 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
20211 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
20212 operand 0 in movabs insn template for -masm=intel asm alternative.
20213 (*movabs<mode>_2): Ditto for operand 1.
20215 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
20218 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
20221 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20223 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
20226 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20228 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
20229 Undef to avoid warning.
20231 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20233 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
20234 * configure: Regenerate.
20236 2013-01-25 Jakub Jelinek <jakub@redhat.com>
20238 PR tree-optimization/56098
20239 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
20240 for stmts with volatile ops.
20241 (cond_store_replacement): Don't optimize if assign has volatile ops.
20242 (cond_if_else_store_replacement_1): Don't optimize if either
20243 then_assign or else_assign have volatile ops.
20244 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
20247 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20249 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
20251 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20253 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
20254 missing ':' in asm example.
20256 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
20258 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
20259 entries into lane and laneq entries.
20260 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
20261 Remove AdvSIMD scalar modes.
20262 (aarch64_sq<r>dmulh_laneq<mode>): New.
20263 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
20265 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
20266 builtin implementations to relfect changes in RTL in aarch64-simd.md.
20267 * config/aarch64/iterators.md (VCOND): New.
20270 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20273 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
20274 Add NULL LIBNAME argument to existing definitions.
20275 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
20276 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
20277 * config/avr/avr.c (DEF_BUILTIN): Same.
20278 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
20279 (avr_expand_builtin): Expand to a vanilla call if a libgcc
20280 implementation is available (DECL_ASSEMBLER_NAME is set).
20281 (avr_fold_absfx): New static function.
20282 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
20283 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
20284 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
20285 AVR_BUILTIN_ABSLLK.
20286 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
20287 (abshk, absk, abslk, absllk): Provide as static inline functions.
20289 2013-01-25 Marek Polacek <polacek@redhat.com>
20291 PR tree-optimization/56035
20292 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
20294 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
20296 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
20297 (*movtf_internal_rex64): Add (!o,C) alternative
20298 (*movxf_internal_rex64): Ditto.
20299 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
20301 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
20303 * doc/invoke.texi: fix typo.
20304 * doc/objc.texi: fix typo.
20306 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
20308 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
20309 for the first two alternatives.
20311 2013-01-24 Diego Novillo <dnovillo@google.com>
20313 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
20314 (ggc-zone.o): Remove.
20315 * configure.ac: Remove option --with-gc.
20316 * configure: Re-generate.
20317 * doc/install.texi: Remove documentation for --with-gc.
20318 * gengtype.c (write_enum_defn): Remove. Update all users.
20319 (write_Types_process_field): Remove generation of gt_e_* argument.
20320 (output_type_enum): Remove. Update all users.
20321 (write_enum_defn): Remove. Update all users.
20322 (enum alloc_zone): Remove. Update all users.
20323 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
20324 * ggc-common.c (ggc_splay_alloc): Remove first argument.
20325 Update all callers.
20326 (struct ptr_data): Remove field TYPE. Update all users.
20327 (gt_pch_note_object): Remove argument TYPE. Update all users.
20328 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
20330 * ggc-none.c (ggc_alloc_typed_stat): Remove.
20331 (struct alloc_zone): Remove.
20332 (ggc_internal_alloc_zone_stat): Remove.
20333 (ggc_internal_cleared_alloc_zone_stat): Remove.
20334 * ggc-page.c (ggc_alloc_typed_stat): Remove.
20335 (ggc_pch_count_object): Remove last argument. Update all users.
20336 (ggc_pch_alloc_object): Remove last argument. Update all users.
20337 (struct alloc_zone): Remove.
20338 * ggc-zone.c: Remove.
20339 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
20340 (struct alloc_zone): Remove.
20341 (ggc_alloc_typed_stat): Remove.
20342 (ggc_alloc_typed): Remove.
20343 (ggc_splay_alloc): Remove first argument.
20344 (rtl_zone): Remove. Update all users.
20345 (tree_zone): Remove. Update all users.
20346 (tree_id_zone): Remove. Update all users.
20347 (ggc_internal_zone_alloc_stat): Remove. Update all users.
20348 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
20349 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
20350 * tree-ssanames.c: Remove references to zone allocator in comments.
20352 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
20354 * config/avr/avr.c (avr_out_fract): Make register numbers that
20355 might be outside of source operand signed.
20357 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
20359 * config/i386/constraints.md (Yf): New constraint.
20360 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
20361 of f constraint to conditionaly disable x87 register preferences.
20362 (*movdf_internal): Ditto.
20363 (*movsf_internal): Ditto.
20365 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
20367 PR inline-asm/55934
20368 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
20369 that have operands with impossible constraints.
20370 Add a FIXME for a speed-up opportunity.
20371 * lra-constraints.c (process_alt_operands): Verify that a class
20372 selected from constraints on asms is valid for the operand mode.
20373 (curr_insn_transform): Remove incorrect comment.
20375 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
20377 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
20378 TOC operand is a valid symbol ref in the constant pool.
20380 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20382 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
20384 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
20387 * config/avr/stdfix.h: New file.
20388 * t-avr (stdfix-gcc.h): New rule to build it.
20389 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
20391 2013-01-23 Kostya Serebryany <kcc@google.com>
20393 * config/darwin.h: remove dependency on
20394 CoreFoundation (asan on Mac OS).
20396 2013-01-23 Jakub Jelinek <jakub@redhat.com>
20399 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
20400 instead of cmpdi_operand for first comparison operand.
20401 Don't assert that comparison operands aren't both constants.
20403 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20405 * doc/install.texi (Downloading the Source): Update references to
20406 downloading separate components.
20408 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20410 * doc/extend.texi (__int128): Improve grammar.
20412 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
20415 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
20416 alternative to (o,r).
20417 (*movdi_internal_rex64): Remove (!o,n) alternative.
20418 (DImode immediate->memory splitter): Remove.
20419 (DImode immediate->memory peephole2): Remove.
20420 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
20421 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
20422 alternative to (!o,*r).
20423 (*movtf_internal_sse): New pattern.
20424 (*movxf_internal_rex64): New pattern.
20425 (*movxf_internal): Disable for TARGET_64BIT.
20426 (*movdf_internal_rex64): Remove (!o,F) alternative.
20428 2013-01-22 Jakub Jelinek <jakub@redhat.com>
20430 PR middle-end/56074
20431 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
20432 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
20433 * tree-vect-loop-manip.c (find_loop_location): Also ignore
20434 stmt locations where LOCATION_LOCUS of the stmt location is
20435 UNKNOWN_LOCATION or BUILTINS_LOCATION.
20438 * config/i386/i386.md (UNSPEC_STOS): New.
20439 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20440 *strsetqi_1): Add UNSPEC_STOS.
20442 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20445 * doc/invoke.texi: Remove left over -Wsynth example.
20447 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20449 PR tree-optimization/56051
20450 * fold-const.c (fold_binary_loc): Don't fold
20451 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20452 a narrowing conversion, or widening conversion from signed
20455 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20457 PR rtl-optimization/56023
20458 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20459 dependent on debug instruction.
20461 2013-01-21 Martin Jambor <mjambor@suse.cz>
20463 PR middle-end/56022
20464 * function.c (allocate_struct_function): Call
20465 invoke_set_current_function_hook earlier.
20467 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20469 * reload1.c (init_reload): Only initialize reload_obstack
20470 during the first call.
20472 2013-01-21 Marek Polacek <polacek@redhat.com>
20474 * cfgloop.c (verify_loop_structure): Fix up grammar.
20476 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20478 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20479 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20481 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20484 * config/arm/marvell-pj4.md: Update copyright year.
20485 Fix up use of alu to alu_reg and simple_alu_imm.
20487 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20489 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20491 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20494 * lra-constraints.c (curr_insn_transform): Don't reuse original
20495 insn for secondary memory move when memory mode should be different.
20497 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20499 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20500 atomic_storedi_1): New patterns.
20502 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20504 btver2 pipeline descriptions.
20505 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20507 * config/i386/i386.md (btver2_decode): New type attributes.
20508 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
20510 * config/i386/btver2.md: New file describing btver2 pipelines.
20512 2013-01-19 Andrew Pinski <apinski@cavium.com>
20514 PR tree-optimization/52631
20515 * tree-ssa-sccvn (visit_use): Before looking up the original
20516 statement, try looking up the simplified expression.
20518 2013-01-19 Anthony Green <green@moxielogic.com>
20520 * config/moxie/moxie.c (moxie_expand_prologue): Set
20521 current_function_static_stack_size.
20523 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20525 PR tree-optimization/56029
20526 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20527 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20529 2013-01-18 Sharad Singhai <singhai@google.com>
20531 PR tree-optimization/55995
20532 * dumpfile.c (dump_loc): Print location only if available.
20533 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
20535 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
20538 * lra-constraints.c (curr_insn_transform): Reuse original insn for
20539 secondary memory move.
20540 (inherit_reload_reg): Use rclass instead of cl for
20541 check_secondary_memory_needed_p.
20543 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20545 PR middle-end/56015
20546 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
20547 the case where writing real complex part of target modifies op1.
20549 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
20551 * config/aarch64/aarch64-simd.md
20552 (aarch64_vcond_internal<mode>): Handle unordered cases.
20553 * config/aarch64/iterators.md (v_cmp_result): New.
20555 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
20556 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20558 * config/arm/marvell-pj4.md: New file.
20559 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
20560 * config/arm/arm.md (generic_sched): Add marvell_pj4.
20561 (generic_vfp): Likewise.
20562 * config/arm/arm-cores.def: Add marvell-pj4.
20563 * config/arm/arm-tune.md: Regenerate.
20564 * config/arm/arm-tables.opt: Regenerate.
20565 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
20566 * doc/invoke.texi: Document marvell-pj4.
20568 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
20570 * config/aarch64/arm_neon.h: Map scalar types to standard types.
20572 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20577 * var-tracking.c (negative_power_of_two_p): New.
20578 (global_get_addr_cache, local_get_addr_cache): New.
20579 (get_addr_from_global_cache, get_addr_from_local_cache): New.
20580 (vt_canonicalize_addr): Rewrite using the above. Adjust the
20582 (vt_stack_offset_p): Remove.
20583 (vt_canon_true_dep): Always canonicalize loc's address.
20584 (clobber_overlapping_mems): Make sure we have a MEM.
20585 (local_get_addr_clear_given_value): New.
20586 (val_reset): Clear local cached entries.
20587 (compute_bb_dataflow): Create and release the local cache.
20588 Disable duplicate MEMs clobbering.
20589 (emit_notes_in_bb): Clobber MEMs likewise.
20590 (vt_emit_notes): Create and release the local cache.
20591 (vt_initialize, vt_finalize): Create and release the global
20592 cache, respectively.
20593 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
20595 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20597 PR libmudflap/53359
20598 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20599 not found in the symtab.
20601 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20604 PR rtl-optimization/55547
20605 PR rtl-optimization/53827
20608 * alias.c (offset_overlap_p): New, factored out of...
20609 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20610 the conservative special case for symbolic constants. Don't
20611 adjust zero sizes on alignment.
20613 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20615 PR rtl-optimization/52573
20616 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20617 REG_UNUSED for the same register.
20619 2013-01-17 Richard Biener <rguenther@suse.de>
20620 Marek Polacek <polacek@redhat.com>
20622 PR rtl-optimization/55833
20623 * loop-unswitch.c (unswitch_loops): Move loop verification...
20624 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20625 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20626 Set it to true when we're removing a loop from hierarchy tree in
20627 an irreducible region.
20628 (fix_bb_placements): Adjust caller.
20629 (fix_loop_placements): Likewise.
20631 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
20633 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20634 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20635 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20636 Remove ID. Adjust comments.
20637 * config/avr/avr-c.c (avr_builtin_name): Remove.
20638 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20639 * config/avr/avr.c (avr_tolower): New static function.
20640 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20641 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20642 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20645 2013-01-17 Jan Hubicka <jh@suse.cz>
20647 PR tree-optimization/55273
20648 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20650 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
20653 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20654 store through atomic_store<mode>_1.
20655 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20657 2013-01-17 Martin Jambor <mjambor@suse.cz>
20659 PR tree-optimizations/55264
20660 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20661 for virtual methods.
20662 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20663 virtual methods before inlining is over.
20664 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20666 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20669 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20671 PR rtl-optimization/56005
20672 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20673 pending reads for prefetch.
20675 2013-01-16 Ian Bolton <ian.bolton@arm.com>
20677 * config/aarch64/aarch64.md
20678 (*cstoresi_neg_uxtw): New pattern.
20679 (*cmovsi_insn_uxtw): New pattern.
20680 (*<optab>si3_uxtw): New pattern.
20681 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20682 (*<optab>si3_insn_uxtw): New pattern.
20683 (*bswapsi2_uxtw): New pattern.
20685 2013-01-16 Richard Biener <rguenther@suse.de>
20687 * tree-inline.c (tree_function_versioning): Remove set but
20688 never used variable.
20690 2013-01-16 Richard Biener <rguenther@suse.de>
20692 PR tree-optimization/55964
20693 * tree-flow.h (rename_variables_in_loop): Remove.
20694 (rename_variables_in_bb): Likewise.
20695 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20696 (copy_loop_before): Adjust and delete update-ssa status.
20697 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
20698 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
20699 (rename_variables_in_loop): Remove.
20700 (slpeel_update_phis_for_duplicate_loop): Likewise.
20701 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20702 use available cfg machinery instead of duplicating it.
20703 Update PHI nodes and perform poor-mans SSA update here.
20704 (slpeel_tree_peel_loop_to_edge): Adjust.
20706 2013-01-16 Richard Biener <rguenther@suse.de>
20708 PR tree-optimization/54767
20709 PR tree-optimization/53465
20710 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20711 (vrp_visit_phi_node): For PHI arguments coming via backedges
20712 drop all symbolical range information.
20713 (execute_vrp): Compute backedges.
20715 2013-01-16 Richard Biener <rguenther@suse.de>
20717 * doc/install.texi: Update CLooG and ISL requirements to
20720 2013-01-16 Christian Bruel <christian.bruel@st.com>
20723 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20724 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20725 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20727 2013-01-16 DJ Delorie <dj@redhat.com>
20729 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20730 (UNSPECV_SP_SWITCH_E): New.
20731 (sp_switch_1): Change to an unspec.
20732 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20735 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
20737 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20738 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20739 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20740 (expand_mem_thread_fence): Ditto.
20741 (expand_mem_signal_fence): Ditto.
20742 (expand_atomic_load): Ditto.
20743 (expand_atomic_store): Ditto.
20745 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
20747 PR rtl-optimization/55547
20748 PR rtl-optimization/53827
20751 * alias.c (memrefs_conflict_p): Set sizes to negative after
20754 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20757 * function.c (thread_prologue_and_epilogue_insns): Always
20758 add crtl->drap_reg to set_up_by_prologue.set, even if
20759 stack_realign_drap is false.
20761 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20763 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20764 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20765 *call): Fix indention.
20767 2013-01-15 Tom de Vries <tom@codesourcery.com>
20770 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20773 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20775 PR rtl-optimization/55153
20776 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20778 2013-01-15 Martin Jambor <mjambor@suse.cz>
20780 PR tree-optimization/55920
20781 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20782 accesses as grp_to_be_debug_replaced.
20784 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20786 PR tree-optimization/55920
20787 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20788 there is non-useless type conversion needed from debug rhs to lhs,
20789 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20791 2013-01-15 Joseph Myers <joseph@codesourcery.com>
20792 Mikael Pettersson <mikpe@it.uu.se>
20795 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
20797 (ASM_OUTPUT_CASE_LABEL): Remove.
20798 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
20799 * final.c (shorten_branches): Update alignment of labels before
20800 jump tables if CASE_VECTOR_SHORTEN_MODE.
20802 2013-01-15 Richard Biener <rguenther@suse.de>
20805 * system.h: Do not include gmp.h for building host tools.
20807 2013-01-15 Richard Biener <rguenther@suse.de>
20809 PR middle-end/55882
20810 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
20811 account for bitpos when computing alignment.
20813 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
20815 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
20816 (ix86_target_macros_internal): Likewise.
20818 * config/i386/i386.c (m_CORE2I7): Removed.
20819 (m_CORE_HASWELL): New macro.
20820 (m_CORE_ALL): Likewise.
20821 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
20822 (initial_ix86_arch_features): Likewise.
20823 (processor_target_table): Initializations for Core avx2.
20824 (cpu_names): New names "core-avx2".
20825 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
20826 PROCESSOR_CORE_HASWELL.
20827 (ix86_issue_rate): New case.
20828 (ia32_multipass_dfa_lookahead): Likewise.
20829 (ix86_sched_init_global): Likewise.
20831 * config/i386/i386.h (TARGET_HASWELL): New macro.
20832 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
20833 (processor_type): New PROCESSOR_HASWELL.
20835 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20837 PR tree-optimization/55955
20838 * tree-vect-loop.c (vectorizable_reduction): Give up early on
20839 *SHIFT_EXPR and *ROTATE_EXPR codes.
20841 PR tree-optimization/48766
20842 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
20843 -ftrapv disable -fwrapv.
20845 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20848 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
20849 etc. to 1 and not to __flash.
20850 Use LL suffix for __INT24_MAX__ with -mint8.
20851 Use ULL suffix for __UINT24_MAX__ with -mint8.
20853 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20855 * config/avr/avr-arch.h
20856 (struct base_arch_s): Use typedef avr_arch_t instead.
20857 (struct arch_info_s): Use typedef avr_arch_info_t instead.
20858 (struct mcu_type_s): Use typedef avr_mcu_t instead.
20859 * config/avr/avr.c: Same.
20860 * config/avr/avr-devices.c: Same.
20861 * config/avr/driver-avr.c: Same.
20862 * config/avr/gen-avr-mmcu-texi.c: Same.
20863 * config/avr/avr-mcus.def: Adjust comment.
20865 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
20867 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
20868 * config/aarch64/iterators.md (VALLDI): New.
20870 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
20871 Andi Kleen <ak@linux.intel.com>
20874 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
20875 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
20878 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20880 * config/avr/avr-stdint.h: Remove trailing blanks.
20881 * config/avr/avr-log.h: Same.
20882 * config/avr/avr-arch.h: Same.
20883 * config/avr/avr-devices.c: Same.
20884 * config/avr/avr-dimode.md: Same.
20885 * config/avr/predicates.md: Same.
20886 * config/avr/avr-c.c: Same. And fix typo.
20888 * config/avr/avr-protos.h: Same. And:
20889 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
20890 (init_cumulative_args): Rename to avr_init_cumulative_args.
20891 (expand_prologue): Rename to avr_expand_prologue.
20892 (expand_epilogue): Rename to avr_expand_epilogue.
20893 (adjust_insn_length): Rename to avr_adjust_insn_length.
20894 (notice_update_cc): Rename to avr_notice_update_cc.
20895 (final_prescan_insn): Rename to avr_final_prescan_insn.
20896 * config/avr/avr.c: Same.
20897 * config/avr/avr.h: Same.
20898 * config/avr/avr.md: Remove trailing blanks.
20899 (prologue): Use avr_expand_prologue.
20900 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
20902 2013-01-14 Richard Biener <rguenther@suse.de>
20904 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
20905 verify_location, collect_subblocks): New functions.
20906 (verify_gimple_in_cfg): Verify that locations only reference
20907 BLOCKs in the functions BLOCK tree.
20909 2013-01-14 Richard Biener <rguenther@suse.de>
20911 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
20913 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
20915 (insert_out_of_ssa_copy_on_edge): Likewise.
20916 (rewrite_close_phi_out_of_ssa): Likewise.
20917 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
20919 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
20920 propagated constants.
20921 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
20924 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20926 * config/avr/avr-modes.def: Add GPL copyright notice.
20928 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
20930 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
20931 MEMMODEL_MASK to determine memory model.
20932 (atomic_store<mode>): Ditto from operands[2].
20933 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
20935 2013-01-13 Jakub Jelinek <jakub@redhat.com>
20938 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
20939 (fold_gimple_assign): Don't call unshare_expr here.
20940 (fold_ctor_reference): Call unshare_expr.
20942 2013-01-13 Terry Guo <terry.guo@arm.com>
20944 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
20945 * doc/fragments.texi: Document MULTILIB_REUSE.
20946 * gcc.c (multilib_reuse): New internal spec.
20947 (set_multilib_dir): Also search multilib from multilib_reuse.
20948 * genmultilib (tmpmultilib3): Refactor code.
20949 (tmpmultilib4): Ditto.
20950 (multilib_reuse): New multilib argument.
20952 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
20954 * Makefile.in: Update copyright.
20956 2013-01-12 Tom de Vries <tom@codesourcery.com>
20958 PR middle-end/55890
20959 * calls.c (expand_call): Check if arg_nr is valid.
20961 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
20963 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
20964 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
20965 documentation. Add missing '__' in front of
20966 __builtin_ia32_packssdw256.
20968 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20971 * config/s390/s390.c (s390_preferred_reload_class): Do not return
20972 NO_REGS for larl operands.
20973 (s390_reload_larl_operand): Use s390_load_address instead of
20976 2013-01-11 Richard Biener <rguenther@suse.de>
20978 * tree-cfg.c (verify_node_sharing_1): Split out from ...
20979 (verify_node_sharing): ... here.
20980 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
20982 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
20984 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
20985 Substitute TREECHECKING.
20986 * configure: Regenerate.
20987 * Makefile.in (TREECHECKING): New.
20989 2013-01-11 Richard Guenther <rguenther@suse.de>
20991 PR tree-optimization/44061
20992 * tree-vrp.c (extract_range_basic): Compute zero as
20993 value-range for __builtin_constant_p of function parameters.
20995 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
20997 Update copyright years.
20999 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
21001 PR rtl-optimization/55672
21002 * lra-eliminations.c (mark_not_eliminable): Permit addition with
21003 const to be eliminable.
21005 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
21007 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
21008 * configure: Regenerate.
21010 2013-01-10 Richard Biener <rguenther@suse.de>
21012 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
21014 2013-01-10 Richard Biener <rguenther@suse.de>
21017 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
21018 locations for virtual PHI arguments.
21019 (rewrite_update_phi_arguments): Likewise.
21021 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
21023 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
21026 2013-01-10 Jakub Jelinek <jakub@redhat.com>
21028 PR tree-optimization/55921
21029 * tree-complex.c (expand_complex_asm): New function.
21030 (expand_complex_operations_1): Call it for GIMPLE_ASM.
21032 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21035 * config/s390/s390.c (s390_symref_operand_p)
21036 (s390_loadrelative_operand_p): Merge the two functions.
21037 (s390_check_qrst_address, print_operand_address): Add parameters
21038 to s390_loadrelative_operand_p invokation.
21039 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
21040 (s390_reload_larl_operand, s390_secondary_reload): Use
21041 s390_loadrelative_operand_p instead of s390_symref_operand_p.
21042 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
21044 2013-01-09 Mike Stump <mikestump@comcast.net>
21046 * dse.c (record_store): Remove unnecessary assert.
21048 2013-01-09 Jan Hubicka <jh@suse.cz>
21050 PR tree-optimization/55569
21051 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
21052 * cfgloop.h (scale_loop_profile): Likewise.
21054 2013-01-09 Jan Hubicka <jh@suse.cz>
21057 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
21059 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
21061 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
21063 PR middle-end/55114
21064 * expr.h (maybe_emit_group_store): Declare.
21065 * expr.c (maybe_emit_group_store): New function.
21066 * builtins.c (expand_builtin_int_roundingfn): Call it.
21067 (expand_builtin_int_roundingfn_2): Likewise.
21069 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
21071 PR rtl-optimization/55829
21072 * lra-constraints.c (match_reload): Add code for absent output.
21073 (curr_insn_transform): Add code for reloads of matched inputs
21076 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
21078 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
21079 attribute of movddup insn to DF.
21080 (*vec_interleave_lowv2df): Ditto.
21081 (vec_dupv2df): Ditto.
21083 2013-01-09 Jan Hubicka <jh@suse.cz>
21085 PR tree-optimiation/55875
21086 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
21087 EVERY_ITERATION parameter.
21088 (number_of_iterations_exit): Check if exit is executed every iteration.
21089 (idx_infer_loop_bounds): Similarly here.
21090 (n_of_executions_at_most): Simplify
21091 to only test for cases where statement is dominated by the
21092 particular bound; handle correctly the "postdominance" test.
21093 (scev_probably_wraps_p): Use max loop iterations info
21094 as a global bound first.
21096 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
21097 Nick Clifton <nickc@redhat.com>
21099 * config/v850/v850.md (cbranchsf4): New pattern.
21100 (cstoresf4): New pattern.
21101 (cbranchdf4): New pattern.
21102 (cstoredf4): New pattern.
21103 (movsicc): Disallow floating point comparisons.
21104 (cmpsf_le_insn): Fix order of operators.
21105 (cmpsf_lt_insn): Likewise.
21106 (cmpsf_eq_insn): Likewise.
21107 (cmpdf_le_insn): Likewise.
21108 (cmpdf_lt_insn): Likewise.
21109 (cmpdf_eq_insn): Likewise.
21110 (cmpsf_ge_insn): Use LE comparison.
21111 (cmpdf_ge_insn): Likewise.
21112 (cmpsf_gt_insn): Use LT comparison.
21113 (cmpdf_gt_insn): Likewise.
21114 (cmpsf_ne_insn): Delete pattern.
21115 (cmpdf_ne_insn): Delete pattern.
21116 * config/v850/v850.c (v850_gen_float_compare): Use
21117 gen_cmpdf_eq_insn for NE comparison.
21118 (v850_float_z_comparison_operator)
21119 (v850_float_nz_comparison_operator): Move from here ...
21120 * config/v850/predicates.md: ... to here. Move GT and GE
21121 comparisons into v850_float_z_comparison_operator.
21122 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
21124 (v850_float_nz_comparison_operator): Likewise.
21126 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21128 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
21129 with calls to gen_insvsi/gen_insvdi.
21131 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21133 * config/i386/i386.c (initial_ix86_tune_features): Set up
21134 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
21136 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
21137 Jakub Jelinek <jakub@redhat.com>
21139 PR tree-optimization/48189
21140 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
21141 If nitercst is 0, don't predict the exit edge.
21143 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21145 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
21146 in asm_fprintf with reg_names.
21147 (aarch64_print_operand_address): Likewise.
21148 (aarch64_return_addr): Likewise.
21149 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
21151 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21153 * config/pa/pa.h (VAL_U6_BITS_P): Define.
21154 (INT_U6_BITS): Likewise.
21155 * config/pa/predicates.md (uint6_operand): New predicate.
21156 (shift5_operand, shift6_operand): Likewise.
21157 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
21159 (lshrdi3): Use shift6_operand.
21160 (shrpsi4, shrpdi4): New insn patterns.
21161 (extzv): Delete expander.
21162 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
21163 predicates in unamed zero extract patterns. Tighten common constraint.
21164 (extv): Delete expander.
21165 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
21166 predicates in unamed sign extract patterns. Tighten common constraint.
21167 (insv): Delete expander.
21168 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
21169 predicates in unamed insert patterns. Tighten common constraint.
21170 Change uint32_operand predicate to uint6_operand predicate in unamed
21171 DImode pattern to insert constant values of type 1...1xxxx.
21173 2013-01-04 Jan Hubicka <jh@suse.cz>
21175 PR tree-optimization/55823
21176 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
21179 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21180 Uros Bizjak <ubizjak@gmail.com>
21182 PR rtl-optimization/55845
21183 * df-problems.c (can_move_insns_across): Stop scanning at
21184 volatile_insn_p source instruction or give up if
21185 across_from .. across_to range contains any volatile_insn_p
21188 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21190 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
21191 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
21193 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
21194 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
21196 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21199 * asan.c (asan_clear_shadow): New function.
21200 (asan_emit_stack_protection): Use it.
21202 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21204 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
21205 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
21206 with tab instead of space.
21208 2013-01-08 Nick Clifton <nickc@redhat.com>
21210 * config/rl78/rl78.c (rl78_expand_prologue): Always select
21211 register bank 0 at the start of an interrupt handler.
21212 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
21215 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21217 * config/aarch64/aarch64-simd.md
21218 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
21219 (aarch64_simd_bsl): Likewise.
21220 (aarch64_vcond_internal<mode>): Likewise.
21221 (vcond<mode><mode>): Likewise.
21222 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
21223 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
21225 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21227 * config/aarch64/aarch64-builtins.c
21228 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
21230 2013-01-08 Martin Jambor <mjambor@suse.cz>
21233 * tree-sra.c (analyze_access_subtree): Return true also after
21234 potentially creating a debug-only replacement.
21236 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21238 PR middle-end/55890
21239 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
21241 PR tree-optimization/54120
21242 * tree-vrp.c (range_fits_type_p): Don't allow
21243 src_precision < precision from signed vr to unsigned_p
21244 if vr->min or vr->max is negative.
21245 (simplify_float_conversion_using_ranges): Test can_float_p
21246 against CODE_FOR_nothing.
21248 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21249 Richard Biener <rguenther@suse.de>
21251 PR middle-end/55851
21252 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
21253 types instead of just INTEGER_TYPE types.
21255 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
21257 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
21260 2013-01-07 Steve Ellcey <sellcey@mips.com>
21263 * config/mips/mips.opt: Change mad to mmad to match documentation.
21265 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21268 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
21269 .progmemx.data now.
21271 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21274 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
21275 (avr_addrspace_t): Add .section_name field.
21276 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
21278 (avr_addrspace): Same. Initialize .section_name. Remove last
21279 NULL entry. Put __memx into .progmemx.data.
21280 (progmem_section_prefix): Remove.
21281 (avr_asm_init_sections): No need to initialize progmem_section.
21282 (avr_asm_named_section): Use avr_addrspace[].section_name to get
21283 section name prefix.
21284 (avr_asm_select_section): Ditto. And use get_unnamed_section to
21285 retrieve the progmem section.
21286 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
21287 boundary to run over avr_addrspace[].
21288 (avr_register_target_pragmas): Ditto.
21290 2013-01-06 Jakub Jelinek <jakub@redhat.com>
21292 * varasm.c (output_constant_def_contents): For asan_protect_global
21293 protected strings, adjust DECL_ALIGN if needed, before testing for
21295 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
21296 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
21298 (output_object_block): For asan protected decls, emit asan padding
21299 after their contents.
21300 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
21301 (asan_finish_file): Test it here instead.
21303 2013-01-07 Nick Clifton <nickc@redhat.com>
21304 Matthias Klose <doko@debian.org>
21305 Doug Kwan <dougkwan@google.com>
21306 H.J. Lu <hongjiu.lu@intel.com>
21309 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
21311 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
21313 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
21315 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
21317 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
21319 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21322 * doc/install.texi (Cross-Compiler-Specific Options): Document
21325 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
21327 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
21328 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
21329 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
21330 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
21331 vqmovun_high_s64): Fix source operand number and update copyright.
21333 2013-01-07 Richard Biener <rguenther@suse.de>
21335 PR middle-end/55890
21336 * gimple.h (gimple_call_builtin_p): New overload.
21337 * gimple.c (validate_call): New function.
21338 (gimple_call_builtin_p): Likewise.
21339 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21340 Use gimple_call_builtin_p.
21341 (find_func_clobbers): Likewise.
21342 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
21343 (strlen_optimize_stmt): Likewise.
21345 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21347 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
21348 (vld1q_dup_*): Likewise.
21349 (vld1_*): Likewise.
21350 (vld1q_*): Likewise.
21351 (vld1_lane_*): Likewise.
21352 (vld1q_lane_*): Likewise.
21354 2013-01-07 Richard Biener <rguenther@suse.de>
21356 * lto-streamer.h (LTO_minor_version): Bump to 2.
21358 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21360 * config/aarch64/aarch64-protos.h
21361 (aarch64_const_double_zero_rtx_p): Rename to...
21362 (aarch64_float_const_zero_rtx_p): ...this.
21363 (aarch64_float_const_representable_p): New.
21364 (aarch64_output_simd_mov_immediate): Likewise.
21365 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
21366 move immediate case.
21367 * config/aarch64/aarch64.c
21368 (aarch64_const_double_zero_rtx_p): Rename to...
21369 (aarch64_float_const_zero_rtx_p): ...this.
21370 (aarch64_print_operand): Allow printing of new constants.
21371 (aarch64_valid_floating_const): New.
21372 (aarch64_legitimate_constant_p): Check for valid floating-point
21374 (aarch64_simd_valid_immediate): Likewise.
21375 (aarch64_vect_float_const_representable_p): New.
21376 (aarch64_float_const_representable_p): Likewise.
21377 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
21378 (aarch64_output_simd_mov_immediate): New.
21379 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
21380 (*movdf_aarch64): Likewise.
21381 * config/aarch64/constraints.md (Ufc): New.
21382 (Y): call aarch64_float_const_zero_rtx.
21383 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
21385 2013-01-07 Richard Biener <rguenther@suse.de>
21387 PR tree-optimization/55888
21388 PR tree-optimization/55862
21389 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
21390 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
21391 not if it is contained therein.
21393 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21395 * config/avr/t-avr: Typo.
21397 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21400 * config/avr/t-avr: Don't automatically rebuild
21401 $(srcdir)/config/avr/t-multilib
21402 $(srcdir)/config/avr/avr-tables.opt
21403 $(srcdir)/doc/avr-mmcu.texi
21404 (avr-mcus): New phony target to build them on request.
21405 (s-avr-mlib, s-avr-mmcu-texi): Remove.
21406 * avr/avr-mcus.def: Adjust comments.
21408 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
21410 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
21412 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21414 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
21416 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21418 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
21420 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
21422 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
21423 to generate profiling.
21424 * config/rs6000/aix64.h (LIB_SPEC): Same.
21426 2013-01-04 Andrew Pinski <apinski@cavium.com>
21428 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
21430 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
21432 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
21434 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
21436 (ix86_expand_move): Ditto.
21437 (ix86_zero_extend_to_Pmode): Ditto.
21438 (ix86_expand_call): Ditto.
21439 (ix86_expand_special_args_builtin): Ditto.
21440 (ix86_expand_builtin): Ditto.
21442 2013-01-04 Richard Biener <rguenther@suse.de>
21444 PR tree-optimization/55862
21445 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21446 translating them through PHI nodes.
21448 2013-01-04 Martin Jambor <mjambor@suse.cz>
21450 PR tree-optimization/55755
21451 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21452 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21454 2013-01-04 Marek Polacek <polacek@redhat.com>
21456 PR middle-end/55859
21457 * opts.c (default_options_optimization): Clarify error message.
21459 2013-01-04 Richard Biener <rguenther@suse.de>
21461 PR middle-end/55863
21462 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21465 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21468 * config/pa/pa.md (movsi): Revert previous change.
21469 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21472 2013-01-03 Richard Henderson <rth@redhat.com>
21474 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21475 after eliminating TLS symbols.
21477 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
21480 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21481 * graphite-poly.c (debug_gmp_value): Likewise.
21483 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
21486 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21487 selected code model, define __code_mode_small__, __code_model_medium__,
21488 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21489 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21490 xchg temporary register with %k. Declare temporary register as
21492 [__x86_64__]: For medium and large code models, preserve %rbx register.
21494 2013-01-03 Richard Biener <rguenther@suse.de>
21496 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
21497 (dump_subscript): Adjust.
21498 (finalize_ddr_dependent): Do not dump redundant info.
21499 (analyze_siv_subscript): Adjust.
21500 (subscript_dependence_tester): Likewise.
21501 (compute_affine_dependence): Likewise.
21503 2013-01-03 Richard Biener <rguenther@suse.de>
21506 2013-01-03 Richard Biener <rguenther@suse.de>
21508 PR tree-optimization/55857
21509 * tree-vect-stmts.c (vectorizable_load): Do not setup
21510 re-alignment for invariant loads.
21512 2013-01-02 Richard Biener <rguenther@suse.de>
21514 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21515 invariant load do not generate a vector load from the scalar location.
21517 2013-01-03 Richard Biener <rguenther@suse.de>
21519 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21520 for not vectorizing.
21521 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21522 not build INDIRECT_REFs, call get_name once only.
21523 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21524 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21526 2013-01-03 Richard Biener <rguenther@suse.de>
21528 PR tree-optimization/55857
21529 * tree-vect-stmts.c (vectorizable_load): Do not setup
21530 re-alignment for invariant loads.
21532 2013-01-03 Richard Biener <rguenther@suse.de>
21535 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
21536 prefer a built-in decl.
21538 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21540 * gcc.c (process_command): Update copyright notice dates.
21541 * gcov.c (print_version): Likewise.
21542 * gcov-dump.c (print_version): Likewise.
21544 PR rtl-optimization/55838
21545 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
21546 iv0.step, iv1.step and step.
21548 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21549 Marc Glisse <marc.glisse@inria.fr>
21551 PR tree-optimization/55832
21552 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
21553 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
21554 integer_{one,zero}_node.
21556 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21559 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
21560 * var-tracking.c (reverse_op): Don't add reverse ops to
21561 VALUEs that have already
21562 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
21564 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
21566 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
21568 2013-01-02 Teresa Johnson <tejohnson@google.com>
21570 * dumpfile.c (dump_loc): Print filename with location.
21571 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
21572 new location_t parameter to emit complete unroll message with
21573 new dump framework.
21574 (canonicalize_loop_induction_variables): Compute loops location
21575 and pass to try_unroll_loop_completely.
21576 * loop-unroll.c (report_unroll_peel): New function.
21577 (peel_loops_completely): Use new dump format with location
21578 for main dumpfile message, and invoke report_unroll_peel on success.
21579 (decide_unrolling_and_peeling): Ditto.
21580 (decide_peel_once_rolling): Remove old dumpfile message subsumed
21581 by report_unroll_peel.
21582 (decide_peel_completely): Ditto.
21583 (decide_unroll_constant_iterations): Ditto.
21584 (decide_unroll_runtime_iterations): Ditto.
21585 (decide_peel_simple): Ditto.
21586 (decide_unroll_stupid): Ditto.
21587 * cfgloop.c (get_loop_location): New function.
21588 * cfgloop.h (get_loop_location): Declare.
21590 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21592 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21595 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21597 PR middle-end/55198
21598 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21599 BLKmode objects when EXPAND_MEMORY is specified.
21601 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21603 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21605 (fold_builtin_cpu): Do not share cpu model decls across statements.
21607 2013-01-02 Jason Merrill <jason@redhat.com>
21610 * tree.c (build_array_type_1): Revert earlier change.
21612 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21614 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21616 * config/aarch64/aarch64-tune.md: Re-generate.
21618 2013-01-02 Richard Biener <rguenther@suse.de>
21620 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21621 invariant load do not generate a vector load from the scalar location.
21623 2013-01-02 Richard Biener <rguenther@suse.de>
21626 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21627 * configure: Regenerate.
21629 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21631 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21632 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21633 (expand_builtin_int_roundingfn_2): Keep the original target around
21634 for the fallback case.
21636 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21638 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21639 to be clear for sign changes.
21641 2013-01-01 Jan Hubicka <jh@suse.cz>
21643 * ipa-inline-analysis.c: Fix formatting.
21645 2013-01-01 Jakub Jelinek <jakub@redhat.com>
21647 PR tree-optimization/55831
21648 * tree-vect-loop.c (get_initial_def_for_induction): Use
21649 gsi_after_labels instead of gsi_start_bb.
21651 Copyright (C) 2013 Free Software Foundation, Inc.
21653 Copying and distribution of this file, with or without modification,
21654 are permitted in any medium without royalty provided the copyright
21655 notice and this notice are preserved.