1 2013-09-12 DJ Delorie <dj@redhat.com>
3 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
6 2013-09-12 Martin Jambor <mjambor@suse.cz>
9 * ipa-prop.c (remove_described_reference): Give up if the edge in the
10 reference descriptor is NULL.
11 (ipa_edge_removal_hook): If owning a reference descriptor, set its
14 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
16 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
17 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
18 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
19 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
20 make_temp_ssa_name): move to tree-ssanames.h
21 * tree-ssa-alias.h: Move prototype.
22 * tree-ssa.h: Include tree-ssanames.h.
23 * tree-ssanames.c (FREE_SSANAMES): Move to here.
24 * tree-ssanames.h: New. Move items from tree-flow*.h
25 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
27 2013-09-12 Richard Biener <rguenther@suse.de>
29 PR tree-optimization/58404
30 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
31 propagate non-invariant addresses into dereferences wrapped
32 in component references.
34 2013-09-12 Richard Biener <rguenther@suse.de>
36 PR tree-optimization/58402
37 * passes.def: Move pass_late_warn_uninitialized later.
39 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
41 * tree-ssa.h: New. Move content from tree-flow.h and
43 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
44 Move prototypes belonging to tree-ssa.c.
45 * tree-flow-inline.h (redirect_edge_var_map_def,
46 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
48 * gimple.h: Adjust prototypes.
49 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move to...
50 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
51 * tree.h: Move prototype to tree-ssa.h.
52 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
53 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
54 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
55 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
56 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
57 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
58 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
59 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
60 graphite-blocking.c, graphite-clast-to-gimple.c, graphite-dependences.c,
61 graphite-interchange.c, graphite-optimize-isl.c, graphite-poly.c,
62 graphite-scop-detection.c, graphite-sese-to-poly.c, graphite.c,
63 ipa-cp.c, ipa-inline-analysis.c, ipa-inline-transform.c, ipa-inline.c,
64 ipa-prop.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa-utils.c,
65 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
66 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
67 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
68 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
69 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
70 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
71 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
72 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
73 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
74 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
75 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
76 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
77 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
78 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
79 value-prof.c, var-tracking.c,
80 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
82 2013-09-12 Richard Biener <rguenther@suse.de>
84 PR tree-optimization/58396
85 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
86 (build_rdg): Take a loop-nest parameter, fix memleaks.
87 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
89 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
91 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
94 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
96 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
97 The FMA instruction names should have a 'v' prefix.
99 2013-09-12 Richard Biener <rguenther@suse.de>
101 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
102 (dot_rdg): Use popen instead of system in optional code.
103 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
104 (already_processed_vertex_p): Adjust.
105 (has_anti_or_output_dependence, predecessor_has_mem_write,
106 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
107 rdg_flag_uses): Remove.
108 (rdg_flag_vertex): Simplify.
109 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
111 (build_rdg_partition_for_component): Process the first vertex
113 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
115 2013-09-12 Alan Modra <amodra@gmail.com>
117 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
119 2013-09-11 DJ Delorie <dj@redhat.com>
120 Nick Clifton <nickc@redhat.com>
122 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
123 (rl78_stack_based_mem): New.
124 * config/rl78/constraints.md (Iv08): New.
133 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
135 (movsi): Change from expand to insn-and-split.
136 (ashrsi3): Clobber AX.
140 * config/rl78/rl78.md (CC_REG): Fix.
141 (addsi3): Allow memory and immediate operands.
142 (addsi3_internal): Split into...
143 (addsi3_internal_virt): ...new, and ...
144 (addsi3_internal_real): ...new.
146 (subsi3_internal_virt): New.
147 (subsi3_internal_real): New.
148 (mulsi3): Add memory operand.
149 (mulsi3_rl78): Likewise.
150 (mulsi3_g13): Likewise.
151 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
152 (cbranchqi4_real): Add more constraint options.
153 (cbranchhi4_real): Expand pattern.
154 (cbranchhi4_real_signed): New.
155 (cbranchhi4_real_inverted): New.
156 (cbranchsi4_real_lt): New.
157 (cbranchsi4_real_ge): New.
158 (cbranchsi4_real_signed): New.
159 (cbranchsi4_real): New.
161 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
163 (lshrsi3_virt): Likewise.
164 (ashlsi3_virt): Likewise.
165 (cbranchqi4_virt_signed): New.
166 (cbranchhi4_virt_signed): New.
167 (cbranchsi4_virt): New.
168 * config/rl78/rl78.c: Whitespace fixes throughout.
169 (move_elim_pass): New.
170 (pass_data_rl78_move_elim): New.
171 (pass_rl78_move_elim): New.
172 (make_pass_rl78_move_elim): New.
173 (rl78_devirt_info): Run devirt earlier.
174 (rl78_move_elim_info): New.
175 (rl78_asm_file_start): Register it.
176 (rl78_split_movsi): New.
177 (rl78_as_legitimate_address): Allow virtual base registers when
179 (rl78_addr_space_convert): Remove spurious debug stuff.
180 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
181 (rl78_print_operand): More cases for not printing '#'.
182 (rl78_expand_compare): Remove most of the logic.
183 (content_memory): New.
184 (clear_content_memory): New.
185 (get_content_index): New.
186 (get_content_name): New.
187 (display_content_memory): New.
188 (update_content): New.
189 (record_content): New.
190 (already_contains): New.
191 (insn_ok_now): Re-recog insns with virtual registers.
192 (add_postponed_content_update): New.
193 (process_postponed_content_update): New.
194 (gen_and_emit_move): New.
195 (transcode_memory_rtx): Record new location content.
196 Use gen_and_emit_move.
197 (force_into_acc): New.
198 (move_to_acc): Use gen_and_emit_move.
199 (move_from_acc): Likewise.
200 (move_acc_to_reg): Likewise.
201 (move_to_x): Likewise.
202 (move_to_hl): Likewise.
203 (move_to_de): Likewise.
204 (rl78_alloc_physical_registers_op1): Record location content.
205 (has_constraint): New.
206 (rl78_alloc_physical_registers_op2): Record location content.
208 (rl78_alloc_physical_registers_ro1): Likewise.
209 (rl78_alloc_physical_registers_cmp): Likewise.
210 (rl78_alloc_physical_registers_umul): Likewise.
211 (rl78_alloc_address_registers_macax): New.
212 (rl78_alloc_physical_registers): Initialize and set location
213 content memory as needed.
214 (rl78_reorg): Make sure split2 is called.
215 (rl78_rtx_costs): New.
217 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
219 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
220 for (not (neg ...)) and (neg (not ...)) cases.
222 2013-09-11 Richard Biener <rguenther@suse.de>
225 * passes.def: Split critical edges before late uninit warning passes.
226 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
228 2013-09-11 Jakub Jelinek <jakub@redhat.com>
230 PR tree-optimization/58385
231 * fold-const.c (build_range_check): If both low and high are NULL,
232 use omit_one_operand_loc to preserve exp side-effects.
234 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
236 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
238 2013-09-11 Richard Biener <rguenther@suse.de>
240 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
241 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
242 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
243 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
244 stmts_from_loop, known_dependences_p, build_empty_rdg,
245 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
246 * tree-loop-distribution.c: ... here.
247 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
248 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
249 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
250 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
251 * tree-loop-distribution.c: ... here.
252 * tree-loop-distribution.c: Include gimple-pretty-print.h.
253 (struct partition_s): Add loops member.
254 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
255 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
256 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
258 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
259 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
260 Sergey Lega <sergey.s.lega@intel.com>
261 Anna Tikhonova <anna.tikhonova@intel.com>
262 Ilya Tocar <ilya.tocar@intel.com>
263 Andrey Turetskiy <andrey.turetskiy@intel.com>
264 Ilya Verbin <ilya.verbin@intel.com>
265 Kirill Yukhin <kirill.yukhin@intel.com>
266 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
268 * config/i386/constraints.md (k): New.
270 * config/i386/i386.c (const regclass_map): Add new mask registers.
271 (dbx_register_map): Ditto.
272 (dbx64_register_map): Ditto.
273 (svr4_dbx_register_map): Ditto.
274 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
276 (ix86_preferred_reload_class): Disable constants for mask registers.
277 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
278 (ix86_hard_regno_mode_ok): Support new mask registers.
279 (x86_order_regs_for_local_alloc): Ditto.
280 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
281 (FIXED_REGISTERS): Add new mask registers.
282 (CALL_USED_REGISTERS): Ditto.
283 (REG_ALLOC_ORDER): Ditto.
284 (VALID_MASK_REG_MODE): New.
285 (FIRST_MASK_REG): Ditto.
286 (LAST_MASK_REG): Ditto.
287 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
288 (MAYBE_MASK_CLASS_P): New.
289 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
290 (REG_CLASS_CONTENTS): Ditto.
292 (ANY_MASK_REG_P): Ditto.
293 (HI_REGISTER_NAMES): Add new mask registers.
294 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
295 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
297 (attribute "type"): Add mskmov, msklog.
298 (attribute "length_immediate"): Support them.
299 (attribute "memory"): Ditto.
300 (attribute "prefix_0f"): Ditto.
301 (*movhi_internal): Support new mask registers.
302 (*movqi_internal): Ditto.
303 (define_split): Split out clobber pattern is a logic
304 insn on mask registers.
305 (*k<logic><mode>): New.
306 (*andhi_1): Extend to support mask regs.
307 (*andqi_1): Extend to support mask regs.
309 (define_split): Split and-not to and and not if operands
311 (*<code><mode>_1): Separate HI mode to new pattern...
313 (*<code>qi_1): Extend to support mask regs.
318 (*one_cmpl<mode>2_1): Remove HImode and handle it...
319 (*one_cmplhi2_1): ...Here, now with mask registers support.
320 (*one_cmplqi2_1): Support new mask registers.
321 (HI/QImode arithmetics splitter): Don't split if mask registers
323 (HI/QImode not splitter): Ditto.
324 * config/i386/predicated.md (mask_reg_operand): New.
325 (general_reg_operand): Ditto.
327 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
329 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
330 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
332 2013-09-10 Jeff Law <law@redhat.com>
334 PR tree-optimization/58380
335 * tree-ssa-threadupdate.c (thread_block): Recognize another case
336 of threading through a buried loop header.
338 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
339 return value for single successor case.
341 2013-09-10 Jan Hubicka <jh@suse.cz>
343 * ipa-devirt.c (ipa_devirt): Enable with LTO.
345 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
348 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
349 support conditional execution.
350 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
352 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
354 * lra.c (lra): Clear lra_optional_reload_pseudos before every
356 * lra-constraints.c (curr_insn_transform): Switch on optional
357 reloads. Check destination too to check move insn.
358 (undo_optional_reloads): Add check that the original peudo did not
359 changed its allocation and the optional reload was inherited on
360 last inheritance pass. Break loop after deciding to keep optional
362 (lra_undo_inheritance): Add check that inherited pseudo still in
365 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
367 * config/aarch64/aarch64.md (generic_sched): New.
368 * config/aarch64/aarch64-generic.md (load): Make conditional
369 on generic_sched attribute.
372 2013-09-10 Jan Hubicka <jh@suse.cz>
374 * lto-cgraph.c: Include ipa-utils.h.
375 (compute_ltrans_boundary): Also add possible targets into the boundary.
377 2013-09-10 Jan Hubicka <jh@suse.cz>
379 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
380 VAR_DECL of vtable rather than full expression.
382 2013-09-10 Jan Hubicka <jh@suse.cz>
383 Paolo Carlini <paolo.carlini@oracle.com>
385 * cgraphunit.c (analyze_functions): Save input_location, set it
386 to UNKNOWN_LOCATION and restore it at the end.
388 2013-09-10 Martin Jambor <mjambor@suse.cz>
390 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
391 represented by a thunk.
393 2013-09-10 Jeff Law <law@redhat.com>
395 PR tree-optimization/58343
396 * tree-ssa-threadupdate.c (thread_block): Identify and disable
397 jump threading requests through loop headers buried in the middle
398 of a jump threading path.
400 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
401 in return value/type.
403 2013-09-10 Jakub Jelinek <jakub@redhat.com>
405 PR rtl-optimization/58365
406 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
407 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
410 2013-09-10 Richard Biener <rguenther@suse.de>
412 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
413 * tree-data-ref.c (create_rdg_vertices): Collect all data
414 references, signal failure to the caller, use data-ref API.
415 (build_rdg): Compute data references only once. Maintain lifetime
416 of data references and data dependences from within RDG.
417 (free_rdg): Free dependence relations.
418 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
419 inventing extra dependences.
420 (distribute_loop): Update for RDG API changes.
422 2013-09-10 Kai Tietz <ktietz@redhat.com>
424 * doc/invoke.texi (fms-extensions): Document changed
425 behavior for ms-abi targets.
426 * config/i386/i386.c (ix86_option_override_internal):
427 Set default value of option -fms-extension for ms-abi targets.
429 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
431 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
433 2013-09-10 Alan Modra <amodra@gmail.com>
436 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
438 2013-09-10 Alan Modra <amodra@gmail.com>
440 * config/rs6000/predicates.md (add_cint_operand): New.
441 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
442 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
443 using add_cint_operand.
444 (largetoc_high_plus_aix): Likewise.
446 2013-09-09 Jakub Jelinek <jakub@redhat.com>
448 PR tree-optimization/58364
449 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
450 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
451 the current range can't be an unconditional true or false.
453 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
455 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
457 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
459 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
461 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
464 * doc/invoke.texi (-Wdelete-incomplete): Document it.
466 2013-09-09 Ian Bolton <ian.bolton@arm.com>
468 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
469 NO_REGS for immediate that can't be moved directly into FP_REGS.
471 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
473 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
474 comparison with negated operand.
475 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
478 2013-09-09 Richard Biener <rguenther@suse.de>
481 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
482 of a subloop record all its block as affecting loop-closed SSA form.
484 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
486 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
487 of an rtx/bitpos pair.
488 (store_fixed_bit_field): Update accordingly.
490 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
492 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
494 * builtins.c (expand_errno_check): Likewise.
495 * dwarf2cfi.c (init_return_column_size): Likewise.
496 * except.c (sjlj_mark_call_sites): Likewise.
497 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
498 * lra-constraints.c (emit_inc): Likewise.
499 * ree.c (combine_set_extension): Likewise.
500 * regmove.c (fixup_match_2): Likewise.
501 * reload1.c (inc_for_reload): Likewise.
503 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
505 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
506 (force_to_mode, simplify_shift_const_1, simplify_comparison):
507 Use gen_int_mode with the mode of the associated simplify_* call.
508 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
509 * expmed.c (expand_shift_1): Likewise.
510 * function.c (instantiate_virtual_regs_in_insn): Likewise.
511 * loop-iv.c (iv_number_of_iterations): Likewise.
512 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
513 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
515 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
517 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
518 of the associated expand_* call.
519 (asan_emit_stack_protection): Likewise.
520 * builtins.c (round_trampoline_addr): Likewise.
521 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
522 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
523 (emit_store_flag): Likewise.
524 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
525 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
527 * function.c (instantiate_virtual_regs_in_insn): Likewise.
528 * ifcvt.c (noce_try_store_flag_constants): Likewise.
529 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
530 * modulo-sched.c (generate_prolog_epilog): Likewise.
531 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
532 (expand_ctz, expand_ffs, expand_unop): Likewise.
534 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
536 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
537 of the associated gen_rtx_* call.
538 * caller-save.c (init_caller_save): Likewise.
539 * combine.c (find_split_point, make_extraction): Likewise.
540 (make_compound_operation): Likewise.
541 * dwarf2out.c (mem_loc_descriptor): Likewise.
542 * explow.c (plus_constant, probe_stack_range): Likewise.
543 * expmed.c (expand_mult_const): Likewise.
544 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
545 * reload1.c (init_reload): Likewise.
546 * valtrack.c (cleanup_auto_inc_dec): Likewise.
547 * var-tracking.c (adjust_mems): Likewise.
548 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
549 rather than gen_rtx_fmt_ee.
551 2013-09-09 Jan Hubicka <jh@suse.cz>
554 * value-prof.c (gimple_ic): Copy also abnormal edges.
556 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
558 * asan.c (asan_shadow_cst): Use gen_int_mode.
560 2013-09-08 Jan Hubicka <jh@suse.cz>
562 * ipa-profile.c: Add toplevel comment.
563 (ipa_propagate_frequency_1): Be more conservative when profile is read.
564 (contains_hot_call_p): New function.
565 (ipa_propagate_frequency): Set frequencies based on counts when
567 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
568 profile; do not tamper with profile after inlining if it is read.
570 2013-09-08 Jan Hubicka <jh@suse.cz>
572 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
575 2013-09-08 Jan Hubicka <jh@suse.cz>
577 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
580 2013-09-08 Jeff Law <law@redhat.com>
583 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
586 2013-09-08 Andi Kleen <ak@linux.intel.com>
588 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
590 2013-09-08 Jan Hubicka <jh@suse.cz>
592 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
594 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
596 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
598 * lra.c (new_insn_reg): Take the containing insn as a parameter.
599 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
600 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
603 2013-09-08 Jan Hubicka <jh@suse.cz>
605 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
606 targets and devirtualize to BUILT_IN_UNREACHABLE.
607 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
608 * ipa.c (walk_polymorphic_call_targets): New function.
609 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
610 functions; use the new timevar.
611 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
612 was removed from the program.
613 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
614 not consider it in the walk when its vtable is dead.
615 (possible_polymorphic_call_targets_1): Pass anonymous flag to
617 (devirt_variable_node_removal_hook): New function.
618 (possible_polymorphic_call_targets): Also register
619 devirt_variable_node_removal_hook.
620 (ipa_devirt): Do not do non-speculative devirtualization.
621 (gate_ipa_devirt): One execute if devirtualizing speculatively.
623 2013-09-08 Jan Hubicka <jh@suse.cz>
625 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
626 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
627 varpool_remove_variable_insertion_hook): Declare.
628 * varpool.c (varpool_node_hook_list): New structure.
629 (first_varpool_node_removal_hook,
630 first_varpool_variable_insertion_hook): New variables.
631 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
632 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
633 varpool_remove_variable_insertion_hook,
634 varpool_call_variable_insertion_hooks): New functions.
635 (varpool_remove_node): Use it.
637 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
640 * diagnostic.c (diagnostic_build_prefix): When s.file is
641 "<built-in>" don't output line and column numbers.
643 2013-09-06 Jan Hubicka <jh@suse.cz>
645 * cgraphunit.c (expand_thunk): Get body before touching arguments.
646 * lto-streamer-out.c: Stream thunks, too.
647 * lto-streamer-in.c (input_function): Pop cfun here
648 (lto_read_body): Instead of here.
650 2013-09-06 Caroline Tice <cmtice@google.com>
652 * doc/install.texi: Add documentation for the --enable-vtable-verify
653 and the --disable-libvtv configure options.
655 2013-09-06 Jeff Law <law@redhat.com>
657 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
658 edge implied equivalences into successor phis.
660 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
662 * resource.c (mark_referenced_resources): Handle COND_EXEC.
664 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
666 * resource.c (mark_target_live_regs): Compute resources taking
667 into account if a call is predicated or not.
669 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
671 * toplev.c (output_stack_usage): Be prepared for suffixes created by
672 the compiler in the function names.
674 2013-09-06 Jan Hubicka <jh@suse.cz>
677 * ipa-inline.c (has_caller_p): New function.
678 (want_inline_function_to_all_callers_p): Use it.
679 (sum_callers, inline_to_all_callers): Break out from ...
680 (ipa_inline): ... here.
682 2013-09-06 Jan Hubicka <jh@suse.cz>
684 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
685 only when AVX is enabled.
687 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
689 * config/aarch64/aarch64.md
690 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
691 is fpsimd_<load/store>2.
692 (load_pair<mode>): Likewise.
693 (store_pair<mode>): Likewise.
695 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
697 * config/arm/types.md (type): Add "mrs" type.
698 * config/aarch64/aarch64.md
699 (aarch64_load_tp_hard): Make type "mrs".
701 (load_tp_hard): Make type "mrs".
702 * config/arm/cortex-a15.md: Update with new attributes.
703 * config/arm/cortex-a5.md: Update with new attributes.
704 * config/arm/cortex-a53.md: Update with new attributes.
705 * config/arm/cortex-a7.md: Update with new attributes.
706 * config/arm/cortex-a8.md: Update with new attributes.
707 * config/arm/cortex-a9.md: Update with new attributes.
708 * config/arm/cortex-m4.md: Update with new attributes.
709 * config/arm/cortex-r4.md: Update with new attributes.
710 * config/arm/fa526.md: Update with new attributes.
711 * config/arm/fa606te.md: Update with new attributes.
712 * config/arm/fa626te.md: Update with new attributes.
713 * config/arm/fa726te.md: Update with new attributes.
715 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
717 * config/aarch64/aarch64.md
718 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
719 (*movtf_aarch64): Likewise.
721 (thumb1_movdi_insn): Use "multiple" for type where more than one
722 instruction is used for a move.
723 (*arm32_movhf): Likewise.
724 (*thumb_movdf_insn): Likewise.
726 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
728 * config/arm/types.md (type): Rename fcpys to fmov.
730 (*arm_movsi_vfp): Rename type fcpys as fmov.
731 (*thumb2_movsi_vfp): Likewise
732 (*movhf_vfp_neon): Likewise
733 (*movhf_vfp): Likewise
734 (*movsf_vfp): Likewise
735 (*thumb2_movsf_vfp): Likewise
736 (*movsfcc_vfp): Likewise
737 (*thumb2_movsfcc_vfp): Likewise
738 * config/aarch64/aarch64-simd.md
739 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
740 * config/aarch64/aarch64.md
741 (*movsi_aarch64): Replace type mov_reg with fmovs.
742 (*movdi_aarch64): Likewise
743 (*movsf_aarch64): Likewise
744 (*movdf_aarch64): Likewise
746 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
747 * config/arm/iwmmxt.md
748 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
749 * config/arm/arm1020e.md: Update with new attributes.
750 * config/arm/cortex-a15-neon.md: Update with new attributes.
751 * config/arm/cortex-a5.md: Update with new attributes.
752 * config/arm/cortex-a53.md: Update with new attributes.
753 * config/arm/cortex-a7.md: Update with new attributes.
754 * config/arm/cortex-a8-neon.md: Update with new attributes.
755 * config/arm/cortex-a9.md: Update with new attributes.
756 * config/arm/cortex-m4-fpu.md: Update with new attributes.
757 * config/arm/cortex-r4f.md: Update with new attributes.
758 * config/arm/marvell-pj4.md: Update with new attributes.
759 * config/arm/vfp11.md: Update with new attributes.
761 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
763 * config/aarch64/aarch64.md
764 (*madd<mode>): Fix type attribute.
765 (*maddsi_uxtw): Likewise.
766 (*msub<mode>): Likewise.
767 (*msubsi_uxtw): Likewise.
768 (<su_optab>maddsidi4): Likewise.
769 (<su_optab>msubsidi4): Likewise.
771 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
773 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
774 * config/arm/arm.md (core_cycles): Remove fdiv.
776 (*sqrtsf2_vfp): Update for attribute changes.
777 (*sqrtdf2_vfp): Likewise.
778 * config/aarch64/aarch64.md:
779 (sqrt<mode>2): Update for attribute changes.
780 * config/arm/arm1020e.md: Update with new attributes.
781 * config/arm/cortex-a15-neon.md: Update with new attributes.
782 * config/arm/cortex-a5.md: Update with new attributes.
783 * config/arm/cortex-a53.md: Update with new attributes.
784 * config/arm/cortex-a7.md: Update with new attributes.
785 * config/arm/cortex-a8-neon.md: Update with new attributes.
786 * config/arm/cortex-a9.md: Update with new attributes.
787 * config/arm/cortex-m4-fpu.md: Update with new attributes.
788 * config/arm/cortex-r4f.md: Update with new attributes.
789 * config/arm/marvell-pj4.md: Update with new attributes.
790 * config/arm/vfp11.md: Update with new attributes.
792 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
794 * config/arm/types.md
795 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
796 * config/aarch64/aarch64.md
797 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
799 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
800 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
801 (float<GPI:mode><GPF:mode>2): Likewise.
803 (*truncsisf2_vfp): Update with new attributes.
804 (*truncsidf2_vfp): Likewise.
805 (fixuns_truncsfsi2): Likewise.
806 (fixuns_truncdfsi2): Likewise.
807 (*floatsisf2_vfp): Likewise.
808 (*floatsidf2_vfp): Likewise.
809 (floatunssisf2): Likewise.
810 (floatunssidf2): Likewise.
811 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
812 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
813 * config/arm/arm1020e.md: Update with new attributes.
814 * config/arm/cortex-a15-neon.md: Update with new attributes.
815 * config/arm/cortex-a5.md: Update with new attributes.
816 * config/arm/cortex-a53.md: Update with new attributes.
817 * config/arm/cortex-a7.md: Update with new attributes.
818 * config/arm/cortex-a8-neon.md: Update with new attributes.
819 * config/arm/cortex-a9.md: Update with new attributes.
820 * config/arm/cortex-m4-fpu.md: Update with new attributes.
821 * config/arm/cortex-r4f.md: Update with new attributes.
822 * config/arm/marvell-pj4.md: Update with new attributes.
823 * config/arm/vfp11.md: Update with new attributes.
825 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
827 * config/aarch64/arm_neon.h
828 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
829 (vqtbx<1,2,3,4><q>_s8): Likewise.
831 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
833 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
834 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
835 (add<mode>3): Add type attribute.
836 (add<mode>3): Likewise.
837 (usadd<mode>3): Likewise.
838 (ssadd<mode>3): Likewise.
839 (sub<mode>3): Likewise.
840 (sub<mode>3): Likewise.
841 (ussub<mode>3): Likewise.
842 (sssub<mode>3): Likewise.
843 (ssmulsa3): Likewise.
844 (usmulusa3): Likewise.
845 (arm_usatsihi): Likewise.
847 (*movdi_vfp): Add types for all instructions.
848 (*movdi_vfp_cortexa8): Likewise.
849 (*movhf_vfp_neon): Likewise.
850 (*movhf_vfp): Likewise.
851 (*movdf_vfp): Likewise.
852 (*thumb2_movdf_vfp): Likewise.
853 (*thumb2_movdfcc_vfp): Likewise.
854 * config/arm/arm.md: Add type attribute to all insn patterns.
855 (*thumb1_adddi3): Add type attribute.
856 (*arm_adddi3): Likewise.
857 (*adddi_sesidi_di): Likewise.
858 (*adddi_zesidi_di): Likewise.
859 (*thumb1_addsi3): Likewise.
860 (addsi3_compare0): Likewise.
861 (*addsi3_compare0_scratch): Likewise.
862 (*compare_negsi_si): Likewise.
863 (cmpsi2_addneg): Likewise.
864 (*addsi3_carryin_<optab>): Likewise.
865 (*addsi3_carryin_alt2_<optab>): Likewise.
866 (*addsi3_carryin_clobercc_<optab>): Likewise.
867 (*subsi3_carryin): Likewise.
868 (*subsi3_carryin_const): Likewise.
869 (*subsi3_carryin_compare): Likewise.
870 (*subsi3_carryin_compare_const): Likewise.
871 (*arm_subdi3): Likewise.
872 (*thumb_subdi3): Likewise.
873 (*subdi_di_zesidi): Likewise.
874 (*subdi_di_sesidi): Likewise.
875 (*subdi_zesidi_di): Likewise.
876 (*subdi_sesidi_di): Likewise.
877 (*subdi_zesidi_ze): Likewise.
878 (thumb1_subsi3_insn): Likewise.
879 (*arm_subsi3_insn): Likewise.
880 (*anddi3_insn): Likewise.
881 (*anddi_zesidi_di): Likewise.
882 (*anddi_sesdi_di): Likewise.
883 (*ne_zeroextracts): Likewise.
884 (*ne_zeroextracts): Likewise.
885 (*ite_ne_zeroextr): Likewise.
886 (*ite_ne_zeroextr): Likewise.
887 (*anddi_notdi_di): Likewise.
888 (*anddi_notzesidi): Likewise.
889 (*anddi_notsesidi): Likewise.
890 (andsi_notsi_si): Likewise.
891 (thumb1_bicsi3): Likewise.
892 (*iordi3_insn): Likewise.
893 (*iordi_zesidi_di): Likewise.
894 (*iordi_sesidi_di): Likewise.
895 (*thumb1_iorsi3_insn): Likewise.
896 (*xordi3_insn): Likewise.
897 (*xordi_zesidi_di): Likewise.
898 (*xordi_sesidi_di): Likewise.
899 (*arm_xorsi3): Likewise.
900 (*andsi_iorsi3_no): Likewise.
902 (*smax_m1): Likewise.
903 (*arm_smax_insn): Likewise.
905 (*arm_smin_insn): Likewise.
906 (*arm_umaxsi3): Likewise.
907 (*arm_uminsi3): Likewise.
908 (*minmax_arithsi): Likewise.
909 (*minmax_arithsi_): Likewise.
910 (*satsi_<SAT:code>): Likewise.
911 (arm_ashldi3_1bit): Likewise.
912 (arm_ashrdi3_1bit): Likewise.
913 (arm_lshrdi3_1bit): Likewise.
914 (*arm_negdi2): Likewise.
915 (*thumb1_negdi2): Likewise.
916 (*arm_negsi2): Likewise.
917 (*thumb1_negsi2): Likewise.
918 (*negdi_extendsid): Likewise.
919 (*negdi_zero_extend): Likewise.
920 (*arm_abssi2): Likewise.
921 (*thumb1_abssi2): Likewise.
922 (*arm_neg_abssi2): Likewise.
923 (*thumb1_neg_abss): Likewise.
924 (one_cmpldi2): Likewise.
925 (extend<mode>di2): Likewise.
926 (*compareqi_eq0): Likewise.
927 (*arm_extendhisi2addsi): Likewise.
928 (*arm_movdi): Likewise.
929 (*thumb1_movdi_insn): Likewise.
930 (*arm_movt): Likewise.
931 (*thumb1_movsi_insn): Likewise.
932 (pic_add_dot_plus_four): Likewise.
933 (pic_add_dot_plus_eight): Likewise.
934 (tls_load_dot_plus_eight): Likewise.
935 (*thumb1_movhi_insn): Likewise.
936 (*thumb1_movsf_insn): Likewise.
937 (*movdf_soft_insn): Likewise.
938 (*thumb_movdf_insn): Likewise.
939 (cbranchsi4_insn): Likewise.
940 (cbranchsi4_scratch): Likewise.
941 (*negated_cbranchsi4): Likewise.
942 (*tbit_cbranch): Likewise.
943 (*tlobits_cbranch): Likewise.
944 (*tstsi3_cbranch): Likewise.
945 (*cbranchne_decr1): Likewise.
946 (*addsi3_cbranch): Likewise.
947 (*addsi3_cbranch_scratch): Likewise.
948 (*arm_cmpdi_insn): Likewise.
949 (*arm_cmpdi_unsig): Likewise.
950 (*arm_cmpdi_zero): Likewise.
951 (*thumb_cmpdi_zero): Likewise.
952 (*deleted_compare): Likewise.
953 (*mov_scc): Likewise.
954 (*mov_negscc): Likewise.
955 (*mov_notscc): Likewise.
956 (*cstoresi_eq0_thumb1_insn): Likewise.
957 (cstoresi_nltu_thumb1): Likewise.
958 (cstoresi_ltu_thu): Likewise.
959 (thumb1_addsi3_addgeu): Likewise.
960 (*arm_jump): Likewise.
961 (*thumb_jump): Likewise.
962 (*check_arch2): Likewise.
963 (arm_casesi_internal): Likewise.
964 (thumb1_casesi_dispatch): Likewise.
965 (*arm_indirect_jump): Likewise.
966 (*thumb1_indirect_jump): Likewise.
968 (*and_scc): Likewise.
969 (*ior_scc): Likewise.
970 (*compare_scc): Likewise.
971 (*cond_move): Likewise.
972 (*cond_arith): Likewise.
973 (*cond_sub): Likewise.
974 (*cmp_ite0): Likewise.
975 (*cmp_ite1): Likewise.
976 (*cmp_and): Likewise.
977 (*cmp_ior): Likewise.
978 (*ior_scc_scc): Likewise.
979 (*ior_scc_scc_cmp): Likewise.
980 (*and_scc_scc): Likewise.
981 (*and_scc_scc_cmp): Likewise.
982 (*and_scc_scc_nod): Likewise.
984 (movcond_addsi): Likewise.
986 (*ifcompare_plus_move): Likewise.
987 (*if_plus_move): Likewise.
988 (*ifcompare_move_plus): Likewise.
989 (*if_move_plus): Likewise.
990 (*ifcompare_arith_arith): Likewise.
991 (*if_arith_arith): Likewise.
992 (*ifcompare_arith_move): Likewise.
993 (*if_arith_move): Likewise.
994 (*ifcompare_move_arith): Likewise.
995 (*if_move_arith): Likewise.
996 (*ifcompare_move_not): Likewise.
997 (*if_move_not): Likewise.
998 (*ifcompare_not_move): Likewise.
999 (*if_not_move): Likewise.
1000 (*ifcompare_shift_move): Likewise.
1001 (*if_shift_move): Likewise.
1002 (*ifcompare_move_shift): Likewise.
1003 (*if_move_shift): Likewise.
1004 (*ifcompare_shift_shift): Likewise.
1005 (*ifcompare_not_arith): Likewise.
1006 (*ifcompare_arith_not): Likewise.
1007 (*if_arith_not): Likewise.
1008 (*ifcompare_neg_move): Likewise.
1009 (*if_neg_move): Likewise.
1010 (*ifcompare_move_neg): Likewise.
1011 (*if_move_neg): Likewise.
1012 (prologue_thumb1_interwork): Likewise.
1013 (*cond_move_not): Likewise.
1014 (*sign_extract_onebit): Likewise.
1015 (*not_signextract_onebit): Likewise.
1016 (stack_tie): Likewise.
1017 (align_4): Likewise.
1018 (align_8): Likewise.
1019 (consttable_end): Likewise.
1020 (consttable_1): Likewise.
1021 (consttable_2): Likewise.
1022 (consttable_4): Likewise.
1023 (consttable_8): Likewise.
1024 (consttable_16): Likewise.
1025 (*thumb1_tablejump): Likewise.
1026 (prefetch): Likewise.
1027 (force_register_use): Likewise.
1028 (thumb_eh_return): Likewise.
1029 (load_tp_hard): Likewise.
1030 (load_tp_soft): Likewise.
1031 (tlscall): Likewise.
1032 (*arm_movtas_ze): Likewise.
1033 (*arm_rev): Likewise.
1034 (*arm_revsh): Likewise.
1035 (*arm_rev16): Likewise.
1036 * config/arm/thumb2.md
1037 (*thumb2_smaxsi3): Likewise.
1038 (*thumb2_sminsi3): Likewise.
1039 (*thumb32_umaxsi3): Likewise.
1040 (*thumb2_uminsi3): Likewise.
1041 (*thumb2_negdi2): Likewise.
1042 (*thumb2_abssi2): Likewise.
1043 (*thumb2_neg_abss): Likewise.
1044 (*thumb2_movsi_insn): Likewise.
1045 (tls_load_dot_plus_four): Likewise.
1046 (*thumb2_movhi_insn): Likewise.
1047 (*thumb2_mov_scc): Likewise.
1048 (*thumb2_mov_negs): Likewise.
1049 (*thumb2_mov_negs): Likewise.
1050 (*thumb2_mov_nots): Likewise.
1051 (*thumb2_mov_nots): Likewise.
1052 (*thumb2_movsicc_): Likewise.
1053 (*thumb2_movsfcc_soft_insn): Likewise.
1054 (*thumb2_indirect_jump): Likewise.
1055 (*thumb2_and_scc): Likewise.
1056 (*thumb2_ior_scc): Likewise.
1057 (*thumb2_ior_scc_strict_it): Likewise.
1058 (*thumb2_cond_move): Likewise.
1059 (*thumb2_cond_arith): Likewise.
1060 (*thumb2_cond_ari): Likewise.
1061 (*thumb2_cond_sub): Likewise.
1062 (*thumb2_negscc): Likewise.
1063 (*thumb2_movcond): Likewise.
1064 (thumb2_casesi_internal): Likewise.
1065 (thumb2_casesi_internal_pic): Likewise.
1066 (*thumb2_alusi3_short): Likewise.
1067 (*thumb2_mov<mode>_shortim): Likewise.
1068 (*thumb2_addsi_short): Likewise.
1069 (*thumb2_subsi_short): Likewise.
1070 (thumb2_addsi3_compare0): Likewise.
1071 (*thumb2_cbz): Likewise.
1072 (*thumb2_cbnz): Likewise.
1073 (*thumb2_one_cmplsi2_short): Likewise.
1074 (*thumb2_negsi2_short): Likewise.
1075 (*orsi_notsi_si): Likewise.
1076 * config/arm/arm1020e.md: Update with new attributes.
1077 * config/arm/arm1026ejs.md: Update with new attributes.
1078 * config/arm/arm1136jfs.md: Update with new attributes.
1079 * config/arm/arm926ejs.md: Update with new attributes.
1080 * config/arm/cortex-a15.md: Update with new attributes.
1081 * config/arm/cortex-a5.md: Update with new attributes.
1082 * config/arm/cortex-a53.md: Update with new attributes.
1083 * config/arm/cortex-a7.md: Update with new attributes.
1084 * config/arm/cortex-a8.md: Update with new attributes.
1085 * config/arm/cortex-a9.md: Update with new attributes.
1086 * config/arm/cortex-m4.md: Update with new attributes.
1087 * config/arm/cortex-r4.md: Update with new attributes.
1088 * config/arm/fa526.md: Update with new attributes.
1089 * config/arm/fa606te.md: Update with new attributes.
1090 * config/arm/fa626te.md: Update with new attributes.
1091 * config/arm/fa726te.md: Update with new attributes.
1093 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1095 * config/aarch64/aarch64-simd.md
1096 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
1097 <vwx> iterator to ensure correct register choice.
1098 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1099 (aarch64_sqdmull_n<mode>): Likewise.
1100 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1101 * config/aarch64/arm_neon.h
1102 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
1103 (vml<as><q>_n_<su>16): Likewise.
1104 (vml<as>l_high_lane<q>_<su>16): Likewise.
1105 (vml<as>l_high_n_<su>16): Likewise.
1106 (vml<as>l_lane<q>_<su>16): Likewise.
1107 (vml<as>l_n_<su>16): Likewise.
1108 (vmul<q>_lane<q>_<su>16): Likewise.
1109 (vmul<q>_n_<su>16): Likewise.
1110 (vmull_lane<q>_<su>16): Likewise.
1111 (vmull_n_<su>16): Likewise.
1112 (vmull_high_lane<q>_<su>16): Likewise.
1113 (vmull_high_n_<su>16): Likewise.
1114 (vqrdmulh<q>_n_s16): Likewise.
1116 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
1118 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
1119 have the correct lane parameter.
1121 2013-09-06 Richard Biener <rguenther@suse.de>
1123 * cfganal.c (control_dependences::~control_dependences):
1124 Properly free all of the vector.
1126 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
1129 * config/i386/i386.c (ix86_conditional_register_usage):
1130 Proper initialize extended SSE registers.
1132 2013-09-06 Jan Hubicka <jh@suse.cz>
1134 PR tree-optimization/58311
1135 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
1137 2013-09-06 Jan Hubicka <jh@suse.cz>
1139 * Makefile.in (tree-sra.o): Update dependencies.
1140 * tree-sra.c: Include ipa-utils.h
1141 (scan_function): Use recursive_call_p.
1142 (has_caller_p): New function.
1143 (cgraph_for_node_and_aliases): Count also callers of aliases.
1145 2013-09-06 Jan Hubicka <jh@suse.cz>
1148 * cgraph.h (symtab_semantically_equivalent_p): Declare.
1149 * tree-tailcall.c: Include ipa-utils.h.
1150 (find_tail_calls): Use it.
1151 * ipa-pure-const.c (check_call): Likewise.
1152 * ipa-utils.c (recursive_call_p): New function.
1153 * ipa-utils.h (recursive_call_p): Dclare.
1154 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
1155 (symtab_semantically_equivalent_p): New function.
1156 * Makefile.in (tree-tailcall.o): Update dependencies.
1158 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1160 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
1163 2013-09-06 Richard Biener <rguenther@suse.de>
1165 * lto-streamer.h (lto_global_var_decls): Remove.
1166 * Makefile.in (OBJS): Remove lto-symtab.o.
1167 (lto-symtab.o): Remove.
1168 (GTFILES): Remove lto-symtab.c
1169 * lto-symtab.c: Move to lto/
1171 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1173 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
1174 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
1175 (UNSPEC_FPINT_RINT): New constant definitions.
1176 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
1177 definition with 2 attributes.
1178 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
1179 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
1182 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1184 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
1185 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
1186 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
1187 attribute to "z196_alone".
1188 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
1191 2013-09-06 Richard Biener <rguenther@suse.de>
1193 * basic-block.h (class control_dependences): New.
1194 * tree-ssa-dce.c (control_dependence_map): Remove.
1196 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
1197 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
1198 find_pdom, find_control_dependence, find_all_control_dependences):
1200 (mark_control_dependent_edges_necessary,
1201 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
1202 tree_dce_done, perform_tree_ssa_dce): Adjust.
1203 * cfganal.c (set_control_dependence_map_bit,
1204 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
1205 find_all_control_dependences): Move from tree-ssa-dce.c and
1206 implement as methods of control_dependences class.
1207 (control_dependences::control_dependences): New.
1208 (control_dependences::~control_dependences): Likewise.
1209 (control_dependences::get_edges_dependent_on): Likewise.
1210 (control_dependences::get_edge): Likewise.
1212 2013-09-04 Jan Hubicka <jh@suse.cz>
1214 * tree.c (types_same_for_odr): Drop overactive check.
1215 * ipa-devirt.c (hash_type_name): Likewise.
1217 2013-09-04 Jan Hubicka <jh@suse.cz>
1219 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
1220 (analyze_functions): ... here.
1222 2013-09-04 Jan Hubicka <jh@suse.cz>
1225 * cgraphunit.c (analyze_functions): Clear AUX fields
1226 after processing; initialize assembler name has.
1228 2013-09-05 Jeff Law <law@redhat.com>
1230 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
1231 from thread_around_empty_block. Record threading path into PATH.
1232 Recurse if threading through the initial block is successful.
1233 (thread_across_edge): Corresponding changes to slightly simplify.
1235 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1237 * config/aarch64/aarch64.md
1238 (type): Remove frecpe, frecps, frecpx.
1239 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
1240 fix to be a TARGET_SIMD instruction.
1241 (aarch64_frecps): Remove.
1242 * config/aarch64/aarch64-simd.md
1243 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
1244 (aarch64_frecps<mode>): Handle all float/vector of float modes.
1246 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1247 Sofiane Naci <sofiane.naci@arm.com>
1249 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
1250 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
1251 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
1252 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
1253 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
1254 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
1255 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
1256 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
1257 into "clz, "rbit". Rename "shift" to "shift_imm".
1258 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
1259 changes. Update for attribute changes all occurrences of arlo_* and
1261 * config/arm/arm-fixed.md: Update for attribute changes
1262 all occurrences of arlo_* types.
1263 * config/arm/thumb2.md: Update for attribute changes all occurrences
1265 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
1266 (cortexa7_older_only): Likewise.
1267 (cortexa7_younger): Likewise.
1268 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
1269 (1020alu_shift_op): Likewise.
1270 (1020alu_shift_reg_op): Likewise.
1271 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
1272 (alu_shift_op): Likewise.
1273 (alu_shift_reg_op): Likewise.
1274 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
1275 (11_alu_shift_op): Likewise.
1276 (11_alu_shift_reg_op): Likewise.
1277 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
1278 (9_alu_shift_reg_op): Likewise.
1279 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
1281 (cortex_a15_alu_shift): Likewise.
1282 (cortex_a15_alu_shift_reg): Likewise.
1283 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
1285 (cortex_a5_alu_shift): Likewise.
1286 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
1288 (cortex_a53_alu_shift): Likewise.
1289 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
1291 (cortex_a7_alu_reg): Likewise.
1292 (cortex_a7_alu_shift): Likewise.
1293 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
1295 (cortex_a8_alu_shift): Likewise.
1296 (cortex_a8_alu_shift_reg): Likewise.
1297 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
1298 (cortex_a9_dp_shift): Likewise.
1299 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
1301 * config/arm/cortex-r4.md
1302 (cortex_r4_alu): Update for attribute changes.
1303 (cortex_r4_mov): Likewise.
1304 (cortex_r4_alu_shift_reg): Likewise.
1305 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
1306 (526_alu_shift_op): Likewise.
1307 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
1308 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
1309 (626te_alu_shift_op): Likewise.
1310 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
1311 (726te_alu_shift_op): Likewise.
1312 (726te_alu_shift_reg_op): Likewise.
1313 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
1314 (mp626_alu_shift_op): Likewise.
1315 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
1316 (pj4_alu_conds): Likewise.
1317 (pj4_shift): Likewise.
1318 (pj4_shift_conds): Likewise.
1319 (pj4_alu_shift): Likewise.
1320 (pj4_alu_shift_conds): Likewise.
1321 * config/aarch64/aarch64.md: Update for attribute change
1322 all occurrences of arlo_* and shift* types.
1324 2013-09-05 Mike Stump <mikestump@comcast.net>
1326 * tree.h: Move documentation for tree_function_decl to tree-core.h
1327 with the declaration.
1329 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
1332 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
1333 looking for widest mode.
1335 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
1337 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
1339 2013-09-05 Richard Biener <rguenther@suse.de>
1341 PR tree-optimization/58137
1342 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
1343 Do not create vectors of pointers.
1344 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
1345 types for the components of the vector initializer.
1346 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
1347 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
1349 2013-09-05 Martin Jambor <mjambor@suse.cz>
1351 * ipa-prop.c (remove_described_reference): Accept missing references,
1352 return false if that hppens, otherwise return true.
1353 (cgraph_node_for_jfunc): New function.
1354 (try_decrement_rdesc_refcount): Likewise.
1355 (try_make_edge_direct_simple_call): Use them.
1356 (ipa_edge_removal_hook): Remove references from rdescs.
1357 (ipa_edge_duplication_hook): Clone rdescs and their references
1358 when the new edge has the same caller as the old one.
1359 * cgraph.c (cgraph_resolve_speculation): Remove speculative
1360 reference before removing any edges.
1362 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
1364 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
1366 * thumb2.md (thumb2_storewb_parisi): New pattern.
1368 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
1370 * config/aarch64/aarch64-option-extensions.def: Add
1371 AARCH64_OPT_EXTENSION of 'crc'.
1372 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
1373 (AARCH64_ISA_CRC): Ditto.
1374 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
1375 description of the CRC extension.
1377 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
1379 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
1380 * config/rs6000/linux.h: Ditto.
1381 * alpha/linux.h: Ditto.
1382 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
1383 no_c99_libc_has_function.
1384 * config/c6x/uclinux-elf.h: Ditto.
1385 * config/lm32/uclinux-elf.h: Ditto.
1386 * config/m68k/uclinux.h: Ditto.
1387 * config/moxie/uclinux.h: Ditto.
1388 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
1389 (crisv32-*-linux*, cris-*-linux*): Ditto.
1390 * config/bfin/bfin.c: Include "tm_p.h".
1392 2013-09-05 Richard Biener <rguenther@suse.de>
1394 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
1395 check for a definition without a basic-block.
1397 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1398 Sofiane Naci <sofiane.naci@arm.com>
1400 * config/aarch64/aarch64.md
1401 (*movti_aarch64): Rename r_2_f and f_2_r.
1402 (*movsf_aarch64): Likewise.
1403 (*movdf_aarch64): Likewise.
1404 (*movtf_aarch64): Likewise.
1405 (aarch64_movdi_<mode>low): Likewise.
1406 (aarch64_movdi_<mode>high): Likewise.
1407 (aarch64_mov<mode>high_di): Likewise.
1408 (aarch64_mov<mode>low_di): Likewise.
1409 (aarch64_movtilow_tilow): Likewise.
1410 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
1411 values to config/arm/types.md
1412 (attribute "conds"): Update for attribute change.
1413 (anddi3_insn): Likewise.
1414 (iordi3_insn): Likewise.
1415 (xordi3_insn): Likewise.
1416 (one_cmpldi2): Likewise.
1417 * config/arm/types.md (type): Add Neon types.
1418 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
1419 use "type" attribute.
1420 (movmisalign<mode>_neon_store): Likewise.
1421 (movmisalign<mode>_neon_load): Likewise.
1422 (vec_set<mode>_internal): Likewise.
1423 (vec_setv2di_internal): Likewise.
1424 (vec_extract<mode>): Likewise.
1425 (vec_extractv2di): Likewise.
1426 (add<mode>3_neon): Likewise.
1427 (adddi3_neon): Likewise.
1428 (sub<mode>3_neon): Likewise.
1429 (subdi3_neon): Likewise.
1430 (mul<mode>3_neon): Likewise.
1431 (mul<mode>3add<mode>_neon): Likewise.
1432 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
1433 (fma<VCVTF:mode>4)): Likewise.
1434 (fma<VCVTF:mode>4_intrinsic): Likewise.
1435 (fmsub<VCVTF:mode>4)): Likewise.
1436 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1437 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1438 (ior<mode>3): Likewise.
1439 (and<mode>3): Likewise.
1440 (anddi3_neon): Likewise.
1441 (orn<mode>3_neon): Likewise.
1442 (orndi3_neon): Likewise.
1443 (bic<mode>3_neon): Likewise.
1444 (bicdi3_neon): Likewise.
1445 (xor<mode>3): Likewise.
1446 (one_cmpl<mode>2): Likewise.
1447 (abs<mode>2): Likewise.
1448 (neg<mode>2): Likewise.
1449 (umin<mode>3_neon): Likewise.
1450 (umax<mode>3_neon): Likewise.
1451 (smin<mode>3_neon): Likewise.
1452 (smax<mode>3_neon): Likewise.
1453 (vashl<mode>3): Likewise.
1454 (vashr<mode>3_imm): Likewise.
1455 (vlshr<mode>3_imm): Likewise.
1456 (ashl<mode>3_signed): Likewise.
1457 (ashl<mode>3_unsigned): Likewise.
1458 (neon_load_count): Likewise.
1459 (ashldi3_neon_noclobber): Likewise.
1460 (signed_shift_di3_neon): Likewise.
1461 (unsigned_shift_di3_neon): Likewise.
1462 (ashrdi3_neon_imm_noclobber): Likewise.
1463 (lshrdi3_neon_imm_noclobber): Likewise.
1464 (widen_ssum<mode>3): Likewise.
1465 (widen_usum<mode>3): Likewise.
1466 (quad_halves_<code>v4si): Likewise.
1467 (quad_halves_<code>v4sf): Likewise.
1468 (quad_halves_<code>v8hi): Likewise.
1469 (quad_halves_<code>v16qi): Likewise.
1470 (reduc_splus_v2di): Likewise.
1471 (neon_vpadd_internal<mode>): Likewise.
1472 (neon_vpsmin<mode>): Likewise.
1473 (neon_vpsmax<mode>): Likewise.
1474 (neon_vpumin<mode>): Likewise.
1475 (neon_vpumax<mode>): Likewise.
1476 (ss_add<mode>_neon): Likewise.
1477 (us_add<mode>_neon): Likewise.
1478 (ss_sub<mode>_neon): Likewise.
1479 (us_sub<mode>_neon): Likewise.
1480 (neon_vadd<mode>_unspec): Likewise.
1481 (neon_vaddl<mode>): Likewise.
1482 (neon_vaddw<mode>): Likewise.
1483 (neon_vhadd<mode>): Likewise.
1484 (neon_vqadd<mode>): Likewise.
1485 (neon_vaddhn<mode>): Likewise.
1486 (neon_vmul<mode>): Likewise.
1487 (neon_vmla<mode>): Likewise.
1488 (neon_vmlal<mode>): Likewise.
1489 (neon_vmls<mode>): Likewise.
1490 (neon_vmlsl<mode>): Likewise.
1491 (neon_vqdmulh<mode>): Likewise.
1492 (neon_vqdmlal<mode>): Likewise.
1493 (neon_vqdmlsl<mode>): Likewise.
1494 (neon_vmull<mode>): Likewise.
1495 (neon_vqdmull<mode>): Likewise.
1496 (neon_vsub<mode>_unspec): Likewise.
1497 (neon_vsubl<mode>): Likewise.
1498 (neon_vsubw<mode>): Likewise.
1499 (neon_vqsub<mode>): Likewise.
1500 (neon_vhsub<mode>): Likewise.
1501 (neon_vsubhn<mode>): Likewise.
1502 (neon_vceq<mode>): Likewise.
1503 (neon_vcge<mode>): Likewise.
1504 (neon_vcgeu<mode>): Likewise.
1505 (neon_vcgt<mode>): Likewise.
1506 (neon_vcgtu<mode>): Likewise.
1507 (neon_vcle<mode>): Likewise.
1508 (neon_vclt<mode>): Likewise.
1509 (neon_vcage<mode>): Likewise.
1510 (neon_vcagt<mode>): Likewise.
1511 (neon_vtst<mode>): Likewise.
1512 (neon_vabd<mode>): Likewise.
1513 (neon_vabdl<mode>): Likewise.
1514 (neon_vaba<mode>): Likewise.
1515 (neon_vabal<mode>): Likewise.
1516 (neon_vmax<mode>): Likewise.
1517 (neon_vmin<mode>): Likewise.
1518 (neon_vpaddl<mode>): Likewise.
1519 (neon_vpadal<mode>): Likewise.
1520 (neon_vpmax<mode>): Likewise.
1521 (neon_vpmin<mode>): Likewise.
1522 (neon_vrecps<mode>): Likewise.
1523 (neon_vrsqrts<mode>): Likewise.
1524 (neon_vqabs<mode>): Likewise.
1525 (neon_vqneg<mode>): Likewise.
1526 (neon_vcls<mode>): Likewise.
1527 (clz<mode>2): Likewise.
1528 (popcount<mode>2): Likewise.
1529 (neon_vrecpe): Likewise.
1530 (neon_vrsqrte): Likewise.
1531 (neon_vget_lane<mode>_sext_internal): Likewise.
1532 (neon_vget_lane<mode>_zext_internal): Likewise.
1533 (neon_vdup_n<mode>): Likewise.
1534 (neon_vdup_nv2di): Likewise.
1535 (neon_vdpu_lane<mode>_internal): Likewise.
1536 (neon_vswp<mode>): Likewise.
1537 (float<mode><V_cvtto>2): Likewise.
1538 (floatuns<mode><V_cvtto>2): Likewise.
1539 (fix_trunc<mode><V_cvtto>)2): Likewise
1540 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1541 (neon_vcvt<mode>): Likewise.
1542 (neon_vcvtv4sfv4hf): Likewise.
1543 (neon_vcvtv4hfv4sf): Likewise.
1544 (neon_vcvt_n<mode>): Likewise.
1545 (neon_vmovn<mode>): Likewise.
1546 (neon_vqmovn<mode>): Likewise.
1547 (neon_vqmovun<mode>): Likewise.
1548 (neon_vmovl<mode>): Likewise.
1549 (neon_vmul_lane<mode>): Likewise.
1550 (neon_vmull_lane<mode>): Likewise.
1551 (neon_vqdmull_lane<mode>): Likewise.
1552 (neon_vqdmulh_lane<mode>): Likewise.
1553 (neon_vmla_lane<mode>): Likewise.
1554 (neon_vmlal_lane<mode>): Likewise.
1555 (neon_vqdmlal_lane<mode>): Likewise.
1556 (neon_vmls_lane<mode>): Likewise.
1557 (neon_vmlsl_lane<mode>): Likewise.
1558 (neon_vqdmlsl_lane<mode>): Likewise.
1559 (neon_vext<mode>): Likewise.
1560 (neon_vrev64<mode>): Likewise.
1561 (neon_vrev32<mode>): Likewise.
1562 (neon_vrev16<mode>): Likewise.
1563 (neon_vbsl<mode>_internal): Likewise.
1564 (neon_vshl<mode>): Likewise.
1565 (neon_vqshl<mode>): Likewise.
1566 (neon_vshr_n<mode>): Likewise.
1567 (neon_vshrn_n<mode>): Likewise.
1568 (neon_vqshrn_n<mode>): Likewise.
1569 (neon_vqshrun_n<mode>): Likewise.
1570 (neon_vshl_n<mode>): Likewise.
1571 (neon_vqshl_n<mode>): Likewise.
1572 (neon_vqshlu_n<mode>): Likewise.
1573 (neon_vshll_n<mode>): Likewise.
1574 (neon_vsra_n<mode>): Likewise.
1575 (neon_vsri_n<mode>): Likewise.
1576 (neon_vsli_n<mode>): Likewise.
1577 (neon_vtbl1v8qi): Likewise.
1578 (neon_vtbl2v8qi): Likewise.
1579 (neon_vtbl3v8qi): Likewise.
1580 (neon_vtbl4v8qi): Likewise.
1581 (neon_vtbx1v8qi): Likewise.
1582 (neon_vtbx2v8qi): Likewise.
1583 (neon_vtbx3v8qi): Likewise.
1584 (neon_vtbx4v8qi): Likewise.
1585 (neon_vtrn<mode>_internal): Likewise.
1586 (neon_vzip<mode>_internal): Likewise.
1587 (neon_vuzp<mode>_internal): Likewise.
1588 (neon_vld1<mode>): Likewise.
1589 (neon_vld1_lane<mode>): Likewise.
1590 (neon_vld1_dup<mode>): Likewise.
1591 (neon_vld1_dupv2di): Likewise.
1592 (neon_vst1<mode>): Likewise.
1593 (neon_vst1_lane<mode>): Likewise.
1594 (neon_vld2<mode>): Likewise.
1595 (neon_vld2_lane<mode>): Likewise.
1596 (neon_vld2_dup<mode>): Likewise.
1597 (neon_vst2<mode>): Likewise.
1598 (neon_vst2_lane<mode>): Likewise.
1599 (neon_vld3<mode>): Likewise.
1600 (neon_vld3qa<mode>): Likewise.
1601 (neon_vld3qb<mode>): Likewise.
1602 (neon_vld3_lane<mode>): Likewise.
1603 (neon_vld3_dup<mode>): Likewise.
1604 (neon_vst3<mode>): Likewise.
1605 (neon_vst3qa<mode>): Likewise.
1606 (neon_vst3qb<mode>): Likewise.
1607 (neon_vst3_lane<mode>): Likewise.
1608 (neon_vld4<mode>): Likewise.
1609 (neon_vld4qa<mode>): Likewise.
1610 (neon_vld4qb<mode>): Likewise.
1611 (neon_vld4_lane<mode>): Likewise.
1612 (neon_vld4_dup<mode>): Likewise.
1613 (neon_vst4<mode>): Likewise.
1614 (neon_vst4qa<mode>): Likewise.
1615 (neon_vst4qb<mode>): Likewise.
1616 (neon_vst4_lane<mode>): Likewise.
1617 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1618 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1619 (neon_vec_<US>mult_lo_<mode>): Likewise.
1620 (neon_vec_<US>mult_hi_<mode>): Likewise.
1621 (neon_vec_<US>shiftl_<mode>): Likewise.
1622 (neon_unpack<US>_<mode>): Likewise.
1623 (neon_vec_<US>mult_<mode>): Likewise.
1624 (vec_pack_trunc_<mode>): Likewise.
1625 (neon_vec_pack_trunk_<mode>): Likewise.
1626 (neon_vabd<mode>_2): Likewise.
1627 (neon_vabd<mode>_3): Likewise.
1628 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1629 (thumb2_movsi_vfp): Likewise.
1630 (movdi_vfp): Likewise.
1631 (movdi_vfp_cortexa8): Likewise.
1632 (movhf_vfp_neon): Likewise.
1633 (movhf_vfp): Likewiwse.
1634 (movsf_vfp): Likewiwse.
1635 (thumb2_movsf_vfp): Likewiwse.
1636 (movdf_vfp): Likewise.
1637 (thumb2_movdf_vfp): Likewise.
1638 (movsfcc_vfp): Likewise.
1639 (thumb2_movsfcc_vfp): Likewise.
1640 (movdfcc_vfp): Likewise.
1641 (thumb2_movdfcc_vfp): Likewise.
1642 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1643 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1644 (v10_v2c): Likewise.
1645 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1647 (cortex_a15_neon_int_2): Likewise.
1648 (cortex_a15_neon_int_3): Likewise.
1649 (cortex_a15_neon_int_4): Likewise.
1650 (cortex_a15_neon_int_5): Likewise.
1651 (cortex_a15_neon_vqneg_vqabs): Likewise.
1652 (cortex_a15_neon_vmov): Likewise.
1653 (cortex_a15_neon_vaba): Likewise.
1654 (cortex_a15_neon_vaba_qqq): Likewise.
1655 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1656 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1657 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1658 scalar_64_32_long_scalar): Likewise.
1659 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1660 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1661 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1662 lotype_qdd_64_32_long): Likewise.
1663 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1664 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1665 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1666 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1667 (cortex_a15_neon_shift_1): Likewise.
1668 (cortex_a15_neon_shift_2): Likewise.
1669 (cortex_a15_neon_shift_3): Likewise.
1670 (cortex_a15_neon_vshl_ddd): Likewise.
1671 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1672 (cortex_a15_neon_vsra_vrsra): Likewise.
1673 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1674 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1675 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1676 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1677 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1678 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1679 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1680 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1681 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1682 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1683 (cortex_a15_neon_bp_simple): Likewise.
1684 (cortex_a15_neon_bp_2cycle): Likewise.
1685 (cortex_a15_neon_bp_3cycle): Likewise.
1686 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1687 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1688 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1689 (cortex_a15_neon_vld2_4_regs): Likewise.
1690 (cortex_a15_neon_vld3_vld4): Likewise.
1691 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1692 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1693 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1694 (cortex_a15_neon_vst3_vst4): Likewise.
1695 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1696 (cortex_a15_neon_vld3_vld4_lane" 10
1697 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1698 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1699 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1700 (cortex_a15_neon_ldm_2): Likewise.0
1701 (cortex_a15_neon_stm_2): Likewise.
1702 (cortex_a15_neon_mcr): Likewise.
1703 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1704 (cortex_a15_neon_mrc): Likewise.
1705 (cortex_a15_neon_mrrc): Likewise.
1706 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1708 (cortex_a15_alu_shift): Likewise.
1709 (cortex_a15_alu_shift_reg): Likewise.
1710 (cortex_a15_mult32): Likewise.
1711 (cortex_a15_mult64): Likewise.
1712 (cortex_a15_block): Likewise.
1713 (cortex_a15_branch): Likewise.
1714 (cortex_a15_load1): Likewise.
1715 (cortex_a15_load3): Likewise.
1716 (cortex_a15_store1): Likewise.
1717 (cortex_a15_store3): Likewise.
1718 (cortex_a15_call): Likewise.
1719 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
1720 (cortex_a5_f2r): Likewise.
1721 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1723 (cortex_a53_f2r): Likewise.
1724 * config/arm/cortex-a7.md
1725 (cortex_a7_branch): Update for attribute change.
1726 (cortex_a7_call): Likewise.
1727 (cortex_a7_alu_imm): Likewise.
1728 (cortex_a7_alu_reg): Likewise.
1729 (cortex_a7_alu_shift): Likewise.
1730 (cortex_a7_mul): Likewise.
1731 (cortex_a7_load1): Likewise.
1732 (cortex_a7_store1): Likewise.
1733 (cortex_a7_load2): Likewise.
1734 (cortex_a7_store2): Likewise.
1735 (cortex_a7_load3): Likewise.
1736 (cortex_a7_store3): Likewise.
1737 (cortex_a7_load4): Likewise.
1738 (cortex_a7_store4): Likewise.
1739 (cortex_a7_fpalu): Likewise.
1740 (cortex_a7_fconst): Likewise.
1741 (cortex_a7_fpmuls): Likewise.
1742 (cortex_a7_neon_mul): Likewise.
1743 (cortex_a7_fpmacs): Likewise.
1744 (cortex_a7_neon_mla: Likewise.
1745 (cortex_a7_fpmuld: Likewise.
1746 (cortex_a7_fpmacd: Likewise.
1747 (cortex_a7_fpfmad: Likewise.
1748 (cortex_a7_fdivs: Likewise.
1749 (cortex_a7_fdivd: Likewise.
1750 (cortex_a7_r2f: Likewise.
1751 (cortex_a7_f2r: Likewise.
1752 (cortex_a7_f_flags: Likewise.
1753 (cortex_a7_f_loads: Likewise.
1754 (cortex_a7_f_loadd: Likewise.
1755 (cortex_a7_f_stores: Likewise.
1756 (cortex_a7_f_stored: Likewise.
1757 (cortex_a7_neon): Likewise.
1758 * config/arm/cortex-a8-neon.md
1759 (cortex_a8_neon_mrc): Update for attribute change.
1760 (cortex_a8_neon_mrrc): Likewise.
1761 (cortex_a8_neon_int_1): Likewise.
1762 (cortex_a8_neon_int_2): Likewise.
1763 (cortex_a8_neon_int_3): Likewise.
1764 (cortex_a8_neon_int_4): Likewise.
1765 (cortex_a8_neon_int_5): Likewise.
1766 (cortex_a8_neon_vqneg_vqabs): Likewise.
1767 (cortex_a8_neon_vmov): Likewise.
1768 (cortex_a8_neon_vaba): Likewise.
1769 (cortex_a8_neon_vaba_qqq): Likewise.
1770 (cortex_a8_neon_vsma): Likewise.
1771 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1772 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1773 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1774 long_scalar): Likewise.
1775 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1776 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1777 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1778 long_scalar_qdd_64_32_long): Likewise.
1779 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1780 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1781 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1782 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1783 (cortex_a8_neon_shift_1): Likewise.
1784 (cortex_a8_neon_shift_2): Likewise.
1785 (cortex_a8_neon_shift_3): Likewise.
1786 (cortex_a8_neon_vshl_ddd): Likewise.
1787 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1788 (cortex_a8_neon_vsra_vrsra): Likewise.
1789 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1790 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1791 (cortex_a8_neon_fp_vsum): Likewise.
1792 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1793 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1794 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1795 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1796 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1797 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1798 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1799 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1800 (cortex_a8_neon_bp_simple): Likewise.
1801 (cortex_a8_neon_bp_2cycle): Likewise.
1802 (cortex_a8_neon_bp_3cycle): Likewise.
1803 (cortex_a8_neon_ldr): Likewise.
1804 (cortex_a8_neon_str): Likewise.
1805 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1806 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1807 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1808 (cortex_a8_neon_vld2_4_regs): Likewise.
1809 (cortex_a8_neon_vld3_vld4): Likewise.
1810 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1811 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1812 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1813 (cortex_a8_neon_vst3_vst4): Likewise.
1814 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1815 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1816 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1817 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1818 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1819 (cortex_a8_neon_mcr): Likewise.
1820 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1821 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
1822 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
1824 (ca9_neon_mrrc): Likewise.
1825 (cortex_a9_neon_int_1): Likewise.
1826 (cortex_a9_neon_int_2): Likewise.
1827 (cortex_a9_neon_int_3): Likewise.
1828 (cortex_a9_neon_int_4): Likewise.
1829 (cortex_a9_neon_int_5): Likewise.
1830 (cortex_a9_neon_vqneg_vqabs): Likewise.
1831 (cortex_a9_neon_vmov): Likewise.
1832 (cortex_a9_neon_vaba): Likewise.
1833 (cortex_a9_neon_vaba_qqq): Likewise.
1834 (cortex_a9_neon_vsma): Likewise.
1835 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1836 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1837 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1838 long_scalar): Likewise.
1839 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1840 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1841 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1842 long_scalar_qdd_64_32_long): Likewise.
1843 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1844 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1845 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1846 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1847 (cortex_a9_neon_shift_1): Likewise.
1848 (cortex_a9_neon_shift_2): Likewise.
1849 (cortex_a9_neon_shift_3): Likewise.
1850 (cortex_a9_neon_vshl_ddd): Likewise.
1851 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1852 (cortex_a9_neon_vsra_vrsra): Likewise.
1853 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1854 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1855 (cortex_a9_neon_fp_vsum): Likewise.
1856 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1857 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1858 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1859 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1860 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1861 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1862 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1863 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1864 (cortex_a9_neon_bp_simple): Likewise.
1865 (cortex_a9_neon_bp_2cycle): Likewise.
1866 (cortex_a9_neon_bp_3cycle): Likewise.
1867 (cortex_a9_neon_ldr): Likewise.
1868 (cortex_a9_neon_str): Likewise.
1869 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1870 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1871 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1872 (cortex_a9_neon_vld2_4_regs): Likewise.
1873 (cortex_a9_neon_vld3_vld4): Likewise.
1874 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1875 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1876 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1877 (cortex_a9_neon_vst3_vst4): Likewise.
1878 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1879 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1880 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1881 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1882 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1883 (cortex_a9_neon_mcr): Likewise.
1884 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1885 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
1886 (cortex_a9_fps): Likewise.
1887 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
1889 (cortex_m4_fmuls): Likewise.
1890 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
1892 (cortex_r4_mrc): Likewise.
1893 * config/arm/iterators.md: Update comment referring to neon_type.
1894 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
1895 (iwmmxt_movsi_insn): Likewise.
1896 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
1898 (pj4_core_to_vfp): Likewise.
1899 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
1901 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
1902 (vfp_fstore): Likewise.
1903 * doc/md.texi: Change references to neon_type to refer to type.
1905 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1907 * tree.h (DECL_BUILT_IN): Fix typo in comment.
1909 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
1911 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
1914 2013-09-04 Easwaran Raman <eraman@google.com>
1917 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
1918 (build_and_add_sum): Use it.
1919 (appears_later_in_bb): Simplify code.
1921 2013-09-04 Teresa Johnson <tejohnson@google.com>
1923 * dumpfile.c (dump_finish): Don't close stderr/stdout.
1925 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
1927 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
1929 2013-09-04 Jan Hubicka <jh@suse.cz>
1931 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
1932 * ipa-devirt.c: Include diganostic.h
1933 (odr_type_d): Add types and types_set.
1934 (hash_type_name): Work for types with vtables during LTO.
1935 (odr_hasher::remove): Fix comment; destroy types_set.
1936 (add_type_duplicate): New function,
1937 (get_odr_type): Use it.
1938 (dump_type_inheritance_graph): Dump type duplicates.
1939 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
1941 * tree.c (types_same_for_odr): Give exact answers on types with
1944 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1946 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
1947 explaining their differences.
1949 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
1951 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
1953 2013-09-03 Jeff Law <law@redhat.com>
1955 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
1956 when not threading through a joiner block. Pass joiner/no joiner
1957 state to register_jump_thread.
1958 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
1959 state from argument rather than implying on path length.
1960 Dump the entire jump thread path into debugging dump.
1961 * tree-flow.h (register_jump_thread): Update prototype.
1963 2013-08-29 Xinliang David Li <davidxl@google.com>
1965 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1966 Remove a trivial gcc_assert.
1968 2013-08-29 Xinliang David Li <davidxl@google.com>
1970 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
1971 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
1972 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1973 Delay base decl alignment adjustment.
1974 * tree-vectorizer.c (vect_destroy_datarefs): New function.
1975 * tree-vectorizer.h: New data structure.
1976 (set_dr_misalignment): New function.
1977 (dr_misalignment): Ditto.
1978 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
1979 (vectorizable_load): Ditto.
1980 (ensure_base_align): New function.
1981 (vectorize_loops): Add dbg_cnt support.
1982 (execute_vect_slp): Ditto.
1983 * dbgcnt.def: New debug counter.
1984 * Makefile: New dependency.
1986 2013-09-03 Meador Inge <meadori@codesourcery.com>
1990 2013-08-30 Meador Inge <meadori@codesourcery.com>
1992 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1994 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
1996 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
1997 function descriptor.
1999 2013-09-03 Richard Biener <rguenther@suse.de>
2001 * tree-affine.c (add_elt_to_tree): Fix association issue,
2002 avoid useless converts and make sure to always return a
2003 properly typed result.
2005 2013-09-03 Richard Biener <rguenther@suse.de>
2008 * fold-const.c (negate_expr_p): Fix division case.
2009 (negate_expr): Likewise.
2011 2013-09-03 Richard Biener <rguenther@suse.de>
2014 * tree-streamer-out.c: Include tm.h.
2015 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
2017 2013-09-03 Jan Hubicka <jh@suse.cz>
2019 * tree-profile.c (tree_profiling): Cleanup CFG when done.
2021 2013-09-03 Alan Modra <amodra@gmail.com>
2023 * config.gcc (powerpc*-*-linux*): Add support for little-endian
2024 multilibs to big-endian target and vice versa.
2025 * config/rs6000/t-linux64: Use := assignment on all vars.
2026 (MULTILIB_EXTRA_OPTS): Remove fPIC.
2027 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
2028 * config/rs6000/t-linux64le: New file.
2029 * config/rs6000/t-linux64bele: New file.
2030 * config/rs6000/t-linux64lebe: New file.
2032 2013-09-02 Jan Hubicka <jh@suse.cz>
2034 * ipa-inline-transform.c (inline_transform): Do not
2035 optimize_inline_calls when not optimizing.
2037 2013-09-02 Jan Hubicka <jh@suse.cz>
2039 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
2040 duplicated nodes for assembler names.
2041 * symtab.c (symtab_unregister_node): Do not attempt to unlink
2042 hard registers from assembler name hash.
2044 2013-09-02 Jan Hubicka <jh@suse.cz>
2046 * ipa-split.c (execute_split_functions): Split externally visible
2047 functions called once.
2049 2013-09-02 Martin Jambor <mjambor@suse.cz>
2052 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
2053 linked list. When finding the correct duplicate, also consider also
2054 the caller in additon to its inlined_to node.
2056 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
2058 * config/aarch64/aarch64-simd-builtins.def
2059 (dup_lane_scalar): Remove.
2060 * config/aarch64/aarch64-simd.md
2061 (aarch64_simd_dup): Add 'w->w' alternative.
2062 (aarch64_dup_lane<mode>): Allow for VALL.
2063 (aarch64_dup_lane_scalar<mode>): Remove.
2064 (aarch64_dup_lane_<vswap_width_name><mode>): New.
2065 (aarch64_get_lane_signed<mode>): Add w->w altenative.
2066 (aarch64_get_lane_unsigned<mode>): Likewise.
2067 (aarch64_get_lane<mode>): Likewise.
2068 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
2069 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
2070 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
2071 (VCON): Change container of V2SF.
2072 (vswap_width_name): Likewise.
2073 * config/aarch64/arm_neon.h
2074 (__aarch64_vdup_lane_any): New.
2075 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2076 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
2077 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2079 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
2082 * expr.c (emit_move_complex): Do not move complex FP values as parts if
2083 the source or the destination is a single hard register.
2085 2013-09-02 Richard Biener <rguenther@suse.de>
2088 * tree-scalar-evolution.c (instantiate_scev_name): Allow
2091 2013-09-02 Richard Biener <rguenther@suse.de>
2093 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
2094 arithmetic to sizetype.
2096 2013-09-02 Bin Cheng <bin.cheng@arm.com>
2098 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
2099 Find auto-increment use both before and after candidate.
2101 2013-09-02 Marek Polacek <polacek@redhat.com>
2103 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
2105 2013-09-01 Jan Hubicka <jh@suse.cz>
2107 * Makefile.in: Add ipa-profile.o
2108 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
2109 * cgraph.c (struct cgraph_propagate_frequency_data,
2110 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
2111 ipa-profile.c; replace cgraph_ by ipa_ prefix.
2112 * cgraph.h (cgraph_propagate_frequency): Remove.
2113 * ipa-inline-analysis.c: Include ipa-utils.h;
2114 drop duplicated cfgloop.h.
2115 (inline_update_callee_summaries): Update.
2116 * ipa-profile.c: New file.
2117 * ipa-utils.h (ipa_propagate_frequency): Declare.
2118 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
2119 data-streamer.h, value-prof.h.
2120 (symtab_remove_unreachable_nodes): Update profile.
2121 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
2122 account_time_size, cmp_counts, dump_histogram,
2123 ipa_profile_generate_summary, ipa_profile_write_summary,
2124 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
2125 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
2126 Move to ipa-profile.c.
2128 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
2130 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
2132 2013-09-01 Jan Hubicka <jh@suse.cz>
2134 * common.opt (fdevirtualize-speculatively): New function.
2135 * invoke.texi (fdevirtualize-speculatively): Document.
2136 * ipa-devirt.c: Include ipa-inline.h
2137 (likely_target_p): New function.
2138 (ipa_devirt): New function.
2139 (gate_ipa_devirt): New function.
2140 (pass_data_ipa_devirt): New static var.
2141 (pass_ipa_devirt): Likewise.
2142 (make_pass_ipa_devirt): New function.
2143 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
2144 (common_handle_option): Disable devirtualization when
2145 value range profiling is available.
2146 * passes.def (pass_ipa_devirt): Add.
2147 * timever.def (TV_IPA_DEVIRT): New timevar.
2148 * tree-pass.h (make_pass_ipa_devirt):
2150 2013-09-01 Iain Sandoe <iain@codesourcery.com>
2152 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
2153 include sanitize(undefined).
2155 2013-08-31 Diego Novillo <dnovillo@google.com>
2157 * Makefile.in (TREE_CORE_H): Define.
2159 (GTFILES): Add tree-core.h.
2160 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
2162 * tree-core.h: New file.
2163 Move all data structures, enum, typedefs, global
2164 declarations and constants from ...
2167 2013-08-31 Jan Hubicka <jh@suse.cz>
2169 * bulitins.c (expand_builtin): Do not early exit for gcov
2170 instrumented functions.
2172 2013-08-31 Marek Polacek <polacek@redhat.com>
2174 * ubsan.c: Include tm_p.h.
2176 2013-08-31 Jan Hubicka <jh@suse.cz>
2178 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
2181 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
2182 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
2183 * tree-cfg.c (verify_gimple_label): ... here.
2184 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
2185 (ipa_merge_profiles): New function.
2186 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
2187 (lto_input_function_body): Likewise.
2188 * ipa-utils.h (ipa_merge_profiles): Declare.
2189 * lto-streamer.h (lto_input_function_body): Update prototype.
2190 (emit_label_in_global_context_p): Remove.
2191 * lto-symtab.c: Include ipa-utils.h
2192 (lto_cgraph_replace_node): Use ipa_merge_profiles.
2194 2013-08-31 Jan Hubicka <jh@suse.cz>
2196 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
2198 2013-08-31 Jan Hubicka <jh@suse.cz>
2200 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
2202 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
2204 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
2205 "cmp" RTX before signed_comparison_operator check to account
2208 2013-08-30 Jan Hubicka <jh@suse.cz>
2210 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
2211 (detect_type_change_1): Rename to ...
2212 (detect_type_change): ... this one; early return on non-polymorphic
2214 (detect_type_change_ssa): Add comp_type parameter; update
2215 use of detect_type_change.
2216 (compute_complex_assign_jump_func): Add param_type parameter;
2217 update use of detect_type_change_ssa.
2218 (compute_complex_ancestor_jump_func): Likewise.
2219 (ipa_get_callee_param_type): New function.
2220 (ipa_compute_jump_functions_for_edge): Compute parameter type;
2221 update calls to the jump function computation functions.
2223 2013-08-30 Teresa Johnson <tejohnson@google.com>
2224 Steven Bosscher <steven@gcc.gnu.org>
2226 * cfgrtl.c (fixup_new_cold_bb): New routine.
2227 (commit_edge_insertions): Invoke fixup_partitions.
2228 (find_partition_fixes): New routine.
2229 (fixup_partitions): Ditto.
2230 (verify_hot_cold_block_grouping): Update comments.
2231 (rtl_verify_edges): Invoke find_partition_fixes.
2232 (rtl_verify_bb_pointers): Update comments.
2233 (rtl_verify_bb_layout): Ditto.
2234 * basic-block.h (probably_never_executed_edge_p): Declare.
2235 (fixup_partitions): Ditto.
2236 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
2237 * bb-reorder.c (sanitize_hot_paths): New function.
2238 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
2240 * predict.c (probably_never_executed_edge_p): New routine.
2241 * cfg.c (check_bb_profile): Add partition insanity warnings.
2243 2013-08-30 Meador Inge <meadori@codesourcery.com>
2245 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2247 2013-08-30 Marek Polacek <polacek@redhat.com>
2249 * Makefile.in (ubsan.o): Add.
2250 (c-family/c-ubsan.o): Add.
2251 (builtins.o): Add ubsan.h dependency.
2252 * ubsan.h: New file.
2253 * ubsan.c: New file.
2254 * common.opt: Add -fsanitize=undefined option.
2255 (flag_sanitize): Add variable.
2256 (fsanitize=): Add option. Add Driver.
2257 (fsanitize=thread): Remove option.
2258 (fsanitize=address): Likewise.
2259 (static-libubsan): New option.
2260 * doc/invoke.texi: Document the new flag and -static-libubsan.
2261 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
2262 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
2263 * builtin-attrs.def (ATTR_COLD): Define.
2264 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2265 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
2266 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
2267 * flag-types.h (sanitize_code): New enum.
2268 * opts.c (common_handle_option): Parse command line arguments
2269 of -fsanitize=. Add -fsanitize=unreachable option.
2270 * varasm.c (get_variable_section): Adjust.
2271 (assemble_noswitch_variable): Likewise.
2272 (assemble_variable): Likewise.
2273 (output_constant_def_contents): Likewise.
2274 (categorize_decl_for_section): Likewise.
2275 (place_block_symbol): Likewise.
2276 (output_object_block): Likewise.
2277 * builtins.def: Likewise.
2278 * toplev.c (compile_file): Likewise.
2279 (process_options): Likewise.
2280 * cppbuiltin.c: Likewise.
2281 * tsan.c (tsan_pass): Likewise.
2282 (tsan_gate): Likewise.
2283 (tsan_gate_O0): Likewise.
2284 * cfgexpand.c (partition_stack_vars): Likewise.
2285 (expand_stack_vars): Likewise.
2286 (defer_stack_allocation): Likewise.
2287 (expand_used_vars): Likewise.
2288 * cfgcleanup.c (old_insns_match_p): Likewise.
2289 * asan.c (asan_finish_file): Likewise.
2290 (asan_instrument): Likewise.
2291 (gate_asan): Likewise.
2292 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
2293 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2294 (asan_global_struct): Use pointer_sized_int_node instead
2295 calling build_nonstandard_integer_type.
2296 (initialize_sanitizer_builtins): Likewise.
2297 (asan_finish_file): Likewise.
2298 * gcc.c: Document %{%:function(args):X}.
2299 (static_spec_functions): Add sanitize.
2300 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
2301 store funcval != NULL there.
2302 (do_spec_1): Adjust handle_spec_function caller.
2303 (handle_braces): Allow %:function(args) as condition.
2304 (sanitize_spec_function): New function.
2305 (ADD_STATIC_LIBUBSAN_LIBS): Define.
2306 (LIBUBSAN_SPEC): Likewise.
2307 (LIBUBSAN_EARLY_SPEC): Likewise.
2308 (SANITIZER_SPEC): Handle libubsan.
2309 (SANITIZER_EARLY_SPEC): Likewise.
2310 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
2311 instead of fsanitize=address.
2312 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
2313 instead of fsanitize=address*.
2314 * builtins.c: Include ubsan.h.
2315 (fold_builtin_0): Instrument __builtin_unreachable.
2316 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
2317 instead of flag_asan.
2318 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
2319 (pointer_sized_int_node): Define.
2320 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
2322 2013-08-30 Mike Stump <mikestump@comcast.net>
2324 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
2327 2013-08-29 Jan Hubicka <jh@suse.cz>
2329 * cgraph.c (cgraph_function_body_availability): Handle weakref
2331 * passes.def: Remove pass_fixup_cfg.
2332 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
2333 track when we need to remove functions.
2334 (gate_ipa_inline): Execute inlining always; add comment why.
2335 (pass_data_ipa_inline): Remove TODO_remove_functions.
2336 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
2337 do not produce summaries.
2338 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
2339 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
2340 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
2341 constant pool and vtable.
2343 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
2345 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
2346 New arm_neon.h's internal macros to specify 64-bit constants.
2347 Avoid using stdint.h's macros.
2349 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
2351 * recog.c (verify_changes): Verify that changes[i].old is non-zero
2352 before applying REG_P.
2354 2013-08-30 Jakub Jelinek <jakub@redhat.com>
2356 PR tree-optimization/58277
2357 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
2358 after seeing too many stmts with vdef in between dombb and current
2359 bb, invalidate everything.
2361 2013-08-30 Richard Biener <rguenther@suse.de>
2363 * fold-const.c (fold_single_bit_test): Fix overflow test.
2365 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
2367 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
2368 and which can live in a register, always retrieve the value on entry.
2369 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
2370 passed by invisible reference specially.
2371 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
2372 (vt_add_function_parameter): Correctly deal with a parameter passed by
2373 invisible reference.
2375 2013-08-30 Jan Hubicka <jh@suse.cz>
2377 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
2379 2013-08-30 Richard Biener <rguenther@suse.de>
2381 PR tree-optimization/58228
2382 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
2383 allow invariant loads in nested loop vectorization.
2385 2013-08-30 Richard Biener <rguenther@suse.de>
2387 PR tree-optimization/58223
2388 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
2389 (has_anti_or_output_dependence): ... this and adjust to also
2390 look for output dependences.
2391 (mark_nodes_having_upstream_mem_writes): Adjust.
2392 (rdg_flag_uses): Likewise.
2394 2013-08-30 Richard Biener <rguenther@suse.de>
2396 PR tree-optimization/58010
2397 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
2398 assert that we have a loop-closed PHI.
2400 2013-08-29 Jan Hubicka <jh@suse.cz>
2402 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
2403 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
2404 * lto-section-in.c (lto_free_function_in_decl_state): New function.
2405 (lto_free_function_in_decl_state_for_node): New function.
2407 2013-08-29 Xinliang David Li <davidxl@google.com>
2409 * loop-unroll.c (report_unroll_peel): Minor message change.
2410 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
2411 Emit alignment peeling message with default -fopt-info.
2412 (vect_loop_versioning): Emit loop version info message.
2413 * tree-vectorizer.c (vectorize_loops): Minor message change.
2414 (execute_vect_slp): Ditto.
2416 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
2418 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
2419 of the clone from the DECL_NAME of the original function.
2421 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
2423 * passes.c (register_pass): Add overload.
2424 * tree-pass.h (register_pass): Forward declare it. Add comment.
2426 2013-08-29 Jan Hubicka <jh@suse.cz>
2428 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
2429 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
2430 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
2431 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2432 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
2433 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
2434 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2435 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
2437 2013-08-29 Teresa Johnson <tejohnson@google.com>
2439 * dumpfile.c (dump_loc): Output column number.
2440 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
2441 * doc/invoke.texi: Document optall -fopt-info flag.
2442 * profile.c (read_profile_edge_counts): Use new dump framework.
2443 (compute_branch_probabilities): Ditto.
2444 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
2445 when pass not in any opt group.
2446 * pass_manager.h (pass_manager::get_pass_profile): New method.
2447 * value-prof.c (check_counter): Use new dump framework.
2448 (check_ic_target): Ditto.
2449 * coverage.c (get_coverage_counts): Ditto.
2450 (coverage_init): Setup new dump framework.
2452 2013-08-29 Richard Biener <rguenther@suse.de>
2454 PR tree-optimization/58246
2455 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
2456 handle the dominance check inside a basic-block.
2458 2013-08-29 Richard Biener <rguenther@suse.de>
2461 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
2462 of default defs that appear in abnormal PHI nodes.
2464 2013-08-29 Richard Biener <rguenther@suse.de>
2466 PR tree-optimization/57685
2467 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2468 single-use operands to avoid exponential complexity.
2470 2013-08-28 Dehao Chen <dehao@google.com>
2472 * ipa-inline.c (edge_badness): Fix integer underflow.
2474 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
2476 * gtm-builtins.def (_ITM_free): Declare leaf.
2478 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2481 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2482 (*tls_local_dynamic_base_64_largepic): Likewise.
2483 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2484 Remove predicate from call operand.
2485 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2486 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2488 2013-08-28 Jeff Law <law@redhat.com>
2490 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2491 checks for the number of predecessors and successors allowed.
2492 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2493 which require copying a joiner block if there is a request which
2494 is a subpath that requires no joiner block copying.
2496 2013-08-28 Jan Hubicka <jh@suse.cz>
2498 * lto-streamer-out.c (DFS_write_tree_body): Drop
2499 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2500 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2501 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2502 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2503 Do not read DECL_ERROR_ISSUED.
2504 (unpack_ts_decl_with_vis_value_fields): Do not read
2506 (lto_input_ts_binfo_tree_pointers): Do not read
2507 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
2508 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2509 write DECL_ERROR_ISSUED..
2510 (pack_ts_decl_with_vis_value_fields): Do not write
2512 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
2513 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
2514 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2515 * tree.h (tree_decl_common): Update comment.
2516 (DECL_ERROR_ISSUED): Remove.
2518 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2521 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2523 2013-08-28 Jan Hubicka <jh@suse.cz>
2525 * builtins.def (free): Declare leaf.
2527 2013-08-27 David Malcolm <dmalcolm@redhat.com>
2530 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2531 * configure: Regenerate.
2533 2013-08-27 Martin Jambor <mjambor@suse.cz>
2535 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2536 (ipa_ancestor_jf_data): Likewise.
2537 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2538 (ipa_get_jf_pass_through_type_preserved): New function.
2539 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2540 (ipa_get_jf_ancestor_type_preserved): New function.
2541 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
2542 (ipa_get_jf_ancestor_result): Likewise.
2543 (propagate_vals_accross_pass_through): Use
2544 ipa_get_jf_pass_through_result to do all the value mappings.
2545 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2546 type_preserved flag.
2547 (ipa_set_jf_cst_copy): New function.
2548 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2549 (ipa_set_jf_arith_pass_through): Likewise.
2550 (ipa_set_ancestor_jf): Likewise.
2551 (compute_complex_assign_jump_func): Set type_preserved instead of
2553 (ipa_compute_jump_functions_for_edge): Likewise.
2554 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2555 (update_jump_functions_after_inlining): Update type_preserved.
2556 Explicitely create jump functions when combining one with pass_through.
2557 (ipa_write_jump_function): Stream the type_preserved flags.
2558 (ipa_read_jump_function): Likewise.
2560 2013-08-27 Jakub Jelinek <jakub@redhat.com>
2561 Aldy Hernandez <aldyh@redhat.com>
2563 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
2564 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
2565 * function.h (struct function): Add has_force_vect_loops and
2567 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
2568 * gimple.c (gimple_build_omp_critical): Add KIND argument and
2570 * gimple.def: Update CLAUSES comments.
2571 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
2572 (gimple_build_omp_for): Add argument to prototype.
2573 (gimple_omp_for_kind): New.
2574 (gimple_omp_for_set_kind): New.
2575 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
2576 GOVD_DATA_SHARE_CLASS.
2577 (enum omp_region_type): Add ORT_SIMD.
2578 (gimple_add_tmp_var): Handle ORT_SIMD.
2579 (gimplify_var_or_parm_decl): Same.
2580 (is_gimple_stmt): Same.
2581 (omp_firstprivatize_variable): Same.
2582 (omp_add_variable): Only use splay_tree_insert if lookup failed.
2583 (omp_notice_variable): Handle ORT_SIMD.
2584 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
2585 (omp_check_private): Handle ORT_SIMD.
2586 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2588 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
2589 Handle OMP_CLAUSE_LASTPRIVATE.
2590 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2592 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
2593 (gimplify_expr): Handle OMP_SIMD.
2594 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
2595 (expand_GOMP_SIMD_VF): New.
2596 (expand_GOMP_SIMD_LAST_LANE): New.
2597 * internal-fn.def (GOMP_SIMD_LANE): New.
2598 (GOMP_SIMD_VF): New.
2599 (GOMP_SIMD_LAST_LANE): New.
2600 * omp-low.c: Include target.h.
2601 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
2603 (check_omp_nesting_restrictions): Same.
2605 (lower_rec_simd_input_clauses): New.
2606 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
2608 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
2609 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
2610 (expand_omp_build_assign): New.
2611 (expand_omp_for_init_counts): New.
2612 (expand_omp_for_init_vars): New.
2613 (extract_omp_for_update_vars): New.
2614 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
2615 and rewrite accordingly.
2616 (expand_omp_simd): New.
2617 (expand_omp_for): Use expand_omp_simd.
2618 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2619 (lower_omp_for): Do not lower the body.
2620 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2622 * tree-flow.h (find_omp_clause): Remove prototype.
2623 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2624 forcing vectorization of the loop, or if flag_tree_vectorize.
2625 (gate_tree_if_conversion): Similarly.
2626 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2627 gimple_build_omp_for.
2628 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2629 * tree-parloops (create_parallel_loop): Pass kind argument to
2630 gimple_build_omp_for.
2631 * tree-pretty-print.c (dump_omp_clause): Add cases for
2632 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2633 OMP_CLAUSE__SIMDUID_.
2634 (dump_generic_node): Handle OMP_SIMD.
2635 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2636 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2637 unroll OMP_SIMD loops here.
2638 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
2639 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
2641 (vect_analyze_data_refs): Handle simd loops.
2642 * tree-vect-loop.c (vectorizable_live_operation): Handle
2644 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
2645 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2646 (vectorizable_load): Same.
2647 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
2648 (struct simduid_to_vf): New.
2649 (simduid_to_vf::hash): New.
2650 (simduid_to-vf::equal): New.
2651 (struct simd_array_to_simduid): New.
2652 (simd_array_to_simduid::hash): New.
2653 (simd_array_to_simduid::equal): New.
2654 (adjust_simduid_builtins): New.
2655 (struct note_simd_array_uses_struct): New.
2656 (note_simd_array_uses_cb): New.
2657 (note_simd_array_uses): New.
2658 (vectorize_loops): Handle simd hints and adjust simd builtins
2660 * tree-vectorizer.h (struct _stmt_vec_info): Add
2661 simd_lane_access_p field.
2662 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2663 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2664 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2665 (omp_clause_code_name): Same.
2666 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2667 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2668 * tree.def (OMP_SIMD): New entry.
2669 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2670 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
2671 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2672 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2673 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2674 (OMP_CLAUSE_LINEAR_STEP): New.
2675 (OMP_CLAUSE_SAFELEN_EXPR): New.
2676 (OMP_CLAUSE__SIMDUID__DECL): New.
2677 (find_omp_clause): New prototype.
2679 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2681 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2682 Haswell processor detection.
2684 2013-08-27 Christian Widmer <shadow@umbrox.de>
2687 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2688 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2689 AVX2 capable processors.
2691 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2693 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2694 of vget_low_* with implementations in terms of other intrinsics.
2696 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
2699 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2700 values to -1, 0 and 1.
2702 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2704 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2705 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2706 (<optab><mode>3_insn): Remove.
2707 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2708 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2709 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2710 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2711 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2712 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2713 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2714 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2715 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2716 (ror<mode>3_insn): Likewise.
2717 * config/aarch64/predicates.md (aarch64_simd_register): New.
2719 2013-08-27 Richard Biener <rguenther@suse.de>
2721 PR tree-optimization/57521
2722 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2723 one edge is non-critical.
2724 (find_phi_replacement_condition): Make sure to use a non-critical
2725 edge. Cleanup and remove old bug workarounds.
2726 (bb_postdominates_preds): Remove.
2727 (if_convertible_loop_p_1): Do not compute post-dominators.
2728 (combine_blocks): Do not free post-dominators.
2729 (main_tree_if_conversion): Likewise.
2730 (pass_data_if_conversion): Add TODO_verify_ssa.
2732 2013-08-27 DJ Delorie <dj@redhat.com>
2734 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2736 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2738 * function.c (assign_parm_find_data_types): Set passed_mode and
2739 nominal_mode to the TYPE_MODE of nominal_type for the built
2740 pointer type in case of the struct-pass-by-reference.
2742 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2744 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2745 (UINT16_TYPE): Change default to "unsigned int".
2747 * config/avr/avr.opt (mfract-convert-truncate): New option.
2748 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2749 is set, round negative fractional integers according to n1169
2750 when converting to integer types.
2752 2013-08-26 Jan Hubicka <jh@suse.cz>
2754 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2755 methods can not be called indirectly when their address is not taken.
2757 2013-08-26 Jan Hubicka <jh@suse.cz>
2759 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2762 2013-08-26 Jan Hubicka <jh@suse.cz>
2764 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2767 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2769 * reload.c (find_valid_class): Allow classes that do not include
2770 FIRST_PSEUDO_REGISTER - 1.
2772 2013-08-26 Jan Hubicka <jh@suse.cz>
2774 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2775 fix edge count/frequency when speculation failed; fix type check
2776 for the direct call.
2778 2013-08-26 Jan Hubicka <jh@suse.cz>
2780 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2782 2013-08-26 Jan Hubicka <jh@suse.cz>
2784 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2785 to be changed by edge redirection.
2787 2013-08-26 Jan Hubicka <jh@suse.cz>
2789 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2790 formating; add sanity check.
2791 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2792 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2793 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2794 (resolve_noninline_speculation): Update callee keys, too.
2796 2013-08-26 Jan Hubicka <jh@suse.cz>
2798 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2799 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2801 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2803 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2806 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
2808 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2809 (x86_64_elf_select_section): Ditto.
2810 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2812 (ix86_pass_by_reference): Ditto.
2813 (output_set_got): Ditto.
2814 (ix86_unary_operator_ok): Ditto.
2815 (ix86_expand_builtin): Ditto.
2817 2013-08-23 Jan Hubicka <jh@suse.cz>
2819 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
2821 2013-08-23 Jan Hubicka <jh@suse.cz>
2823 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
2824 (tree_decl_with_vis): Add FINAL field.
2826 2013-08-23 Jeff Law <law@redhat.com>
2828 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
2829 the debugging dump when the expression is fully redundant.
2831 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2833 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
2834 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
2835 * pretty-print.c (pp_formatted_text_data): Likewise.
2836 (pp_write_text_to_stream): Likewise.
2837 (pp_write_text_as_dot_label_to_stream): Likewise.
2838 (pp_append_r): Likewise.
2839 (pp_format): Likewise.
2840 (pp_flush): Likewise.
2841 (pp_clear_output_area): Likewise.
2842 (pp_append_text): Likewise.
2843 (pp_formatted_text): Likewise.
2844 (pp_remaining_character_count_for_line): Likewise.
2845 (pp_newline): Likewise.
2846 (pp_character): Likewise.
2847 (output_buffer::~output_buffer): Define.
2848 (pretty_printer::~pretty_printer): Destruct output buffer.
2849 * pretty-print.h (output_buffer::~output_buffer): Declare.
2850 (pretty_printer::~pretty_printer): Declare virtual.
2852 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
2855 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
2856 HOST_WIDE_INT_M1U): New macros.
2857 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
2858 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
2859 unsigned -1 for lshift.
2860 * cse.c (cse_insn): Likewise.
2861 * double-int.c (rshift_double, lshift_double): Likewise.
2862 * builtins.c (fold_builtin_bitop): Likewise.
2863 * combine.c (force_to_mode): Likewise.
2864 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
2865 * simplify-rtx.c (simplify_const_unary_operation,
2866 simplify_const_binary_operation): Likewise.
2867 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
2868 check_va_list_escapes): Likewise.
2869 * rtlanal.c (nonzero_bits1): Likewise.
2870 * expmed.c (expand_smod_pow2): Likewise.
2871 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
2873 2013-08-23 Jan Hubicka <jh@suse.cz>
2875 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
2876 as having address taken.
2878 2013-08-23 Jan Hubicka <jh@suse.cz>
2880 * ipa-utils.h (method_class_type): Declare.
2881 * ipa-devirt.c (method_class_type): Export.
2883 * cgraphunit.c (analyze_functions): Do basic devirtualization;
2884 do not walk base classes of anonymous types.
2886 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
2888 PR rtl-optimization/58220
2890 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
2891 handle SEQUENCE insns properly.
2893 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2895 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
2897 (pp_newline_and_indent): Likewise.
2898 (pp_separate_with): Likewise.
2899 * pretty-print.c (pp_newline_and_flush): Define.
2900 (pp_newline_and_indent): Likewise.
2901 (pp_separate_with): Likewise.
2903 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2906 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
2907 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
2909 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
2911 * config/i386/predicates.md (ext_sse_reg_operand): New.
2912 * config/i386/i386.md (*movti_internal): Use
2913 predicate to determine if EVEX is needed.
2914 (*movsi_internal): Ditto.
2915 (*movdf_internal): Ditto.
2916 (*movsf_internal): Ditto.
2917 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2919 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2921 PR tree-optimization/58209
2922 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
2923 (find_tail_calls): Give up for pointer result types if m is non-NULL.
2924 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
2925 emit POINTER_PLUS_EXPR.
2926 (create_tailcall_accumulator): For pointer result type accumulate in
2929 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
2931 * configure.ac: Add backslashes missing from the last change.
2932 * configure: Regenerate.
2934 2013-08-22 Jan Hubicka <jh@susue.cz>
2936 * ipa.c (function_and_variable_visibility): First remember function
2937 was global and then make it local.
2939 2013-08-22 Julian Brown <julian@codesourcery.com>
2941 * configure.ac: Add aarch64 to list of arches which use "nop" in
2943 * configure: Regenerate.
2945 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2947 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
2948 gnu_libc_has_function.
2949 * config/s390/tpf.h: Likewise.
2951 2013-08-22 Jan Hubicka <jh@susue.cz>
2953 * timevar.c (validate_phases): Add cast.
2955 2013-08-22 Jan Hubicka <jh@susue.cz>
2957 * timevar.c (validate_phases): Use size_t for memory.
2958 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
2960 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
2962 * pretty-print.h (output_buffer::output_buffer): Declare.
2963 (pretty_printer::pretty_printer): Likewise.
2964 (pp_construct): Remove.
2965 * pretty-print.c (output_buffer::output_buffer): Define.
2966 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
2967 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
2968 (print_gimple_expr): Likewise.
2969 (print_gimple_seq): Likewise.
2970 (gimple_dump_bb): Likewise.
2971 * sched-vis.c (dump_value_slim): Likewise.
2972 (dump_insn_slim): Likewise.
2973 (dump_rtl_slim): Likewise.
2974 (str_pattern_slim): Likewise.
2975 * tree-mudflap.c (mf_varname_tree): Likewise.
2976 * graph.c (print_graph_cfg): Likewise.
2977 (start_graph_dump): Likewise.
2978 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
2980 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
2981 pretty printer initialization.
2982 * coretypes.h (diagnostic_context): Remove superflous type alias
2984 (pretty_printer): Likewise. Declare directly as a class.
2985 (pretty_print_info): Remove declaration as class.
2986 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
2987 and pp_clear_output_area.
2988 (asan_add_global): Likewise.
2990 2013-08-22 Jan Hubicka <jh@suse.cz>
2992 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
2993 * ipa-utils.h (update_type_inheritance_graph): Declare.
2994 (possible_polymorphic_call_target_p): Declare.
2995 (possible_polymorphic_call_target_p): New.
2996 * ipa-devirt.c: Update toplevel comments.
2997 (cached_polymorphic_call_targets): Move up.
2998 (odr_type_d): Move ID down.
2999 (polymorphic_type_binfo_p): Update comment.
3000 (odr_hasher::remove): Likewise;
3001 (get_odr_type): Set anonymous_namespace.
3002 (dump_odr_type): Dump it.
3003 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
3004 (maybe_record_node): Record node in cached_polymorphic_call_targets.
3005 (record_binfo): Add comment.
3006 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
3008 (devirt_node_removal_hook): Do not iCE when cache is freed.
3009 (possible_polymorphic_call_target_p): New predicate.
3010 (update_type_inheritance_graph): New function.
3012 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
3013 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3014 Sergey Lega <sergey.s.lega@intel.com>
3015 Anna Tikhonova <anna.tikhonova@intel.com>
3016 Ilya Tocar <ilya.tocar@intel.com>
3017 Andrey Turetskiy <andrey.turetskiy@intel.com>
3018 Ilya Verbin <ilya.verbin@intel.com>
3019 Kirill Yukhin <kirill.yukhin@intel.com>
3020 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3022 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
3023 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
3024 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
3025 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
3026 (OPTION_MASK_ISA_AVX2_UNSET): Update.
3027 (OPTION_MASK_ISA_AVX512F_UNSET): New.
3028 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
3029 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
3030 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
3031 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
3032 OPT_mavx512pf, OPT_mavx512er cases.
3033 * config/i386/constraints.md (v): New constraint.
3034 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
3035 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
3036 (bit_AVX512CD): New.
3037 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3038 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
3039 * config/i386/i386-c.c (ix86_target_macros_internal):
3040 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
3042 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
3043 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
3044 * config/i386/i386.c (regclass_map, dbx_register_map)
3045 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
3046 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
3047 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
3049 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
3050 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
3051 -mavx512pf options. Fix formatting.
3052 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
3053 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
3054 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
3055 -mavx512cd, -mavx512pf options.
3056 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
3057 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
3058 (ix86_preferred_output_reload_class): Replace SSE_REGS with
3060 (ix86_hard_regno_mode_ok): Support 512-bit registers.
3061 (ix86_set_reg_reg_cost): Ditto.
3062 (x86_order_regs_for_local_alloc): Ditto.
3063 (MAX_VECT_LEN): Extend to 64-byte.
3064 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
3065 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
3066 (TARGET_AVX512ER, TARGET_AVX512CD): New.
3067 (BIGGEST_ALIGNMENT): Extend to 512-bits.
3068 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
3069 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
3070 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
3071 (SSE_REG_MODE_P): Support new modes.
3072 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
3073 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
3074 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
3075 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
3076 (REG_CLASS_CONTENTS): Add new registers.
3077 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
3078 (EXT_REX_SSE_REGNO_P): New.
3079 (HI_REGISTER_NAMES): Add new registers.
3080 * config/i386/i386.md: Define constants for new registers.
3081 (mode): Add new 512-bit modes.
3082 (prefix): Support evex prefix.
3083 (isa): Support avx512f, noavx512f, fma_avx512f.
3084 (ssemodesuffix): Add new 512-bit modes.
3086 (*movxi_internal_avx512f): Ditto.
3087 (*movdi_internal): Replace constraint "x" with the new constraint "v".
3089 (*movsi_internal): Likewise.
3090 (*movdf_internal): Likewise.
3091 (*movsf_internal): Likewise.
3092 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
3093 (<code><mode>3): Likewise.
3094 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
3096 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
3097 with the new constraint "v".
3098 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
3100 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
3101 with the new constraint "v".
3102 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
3103 (<sse2>_storedqu<avxsizesuffix>): Likewise.
3104 (*<plusminus_insn><mode>3): Likewise.
3105 (<sse>_vm<plusminus_insn><mode>3): Likewise.
3106 (*mul<mode>3): Likewise.
3107 (<sse>_vmmul<mode>3): Likewise.
3108 (<sse>_div<mode>3): Likewise.
3109 (<sse>_vmdiv<mode>3): Likewise.
3110 (<sse>_sqrt<mode>2): Likewise.
3111 (<sse>_vmsqrt<mode>2): Likewise.
3112 (*<code><mode>3_finite): Likewise.
3113 (*<code><mode>3) <smaxmin>: Likewise.
3114 (<sse>_vm<code><mode>3): Likewise.
3115 (*<code><mode>3) <any_logic>: Likewise.
3116 (*fma_fmadd_<mode>): Likewise.
3117 (*fma_fmsub_<mode>): Likewise.
3118 (*fma_fnmadd_<mode>): Likewise.
3119 (*fma_fnmsub_<mode>): Likewise.
3120 (*fma_fmaddsub_<mode>): Likewise.
3121 (*fma_fmsubadd_<mode>): Likewise.
3122 (*fmai_fmadd_<mode>): Likewise.
3123 (*fmai_fmsub_<mode>): Likewise.
3124 (*fmai_fnmadd_<mode>): Likewise.
3125 (*fmai_fnmsub_<mode>): Likewise.
3126 (sse_cvtsi2ss): Likewise.
3127 (sse_cvtsi2ssq): Likewise.
3128 (sse_cvtss2si): Likewise.
3129 (sse_cvtss2si_2): Likewise.
3130 (sse_cvtss2siq): Likewise.
3131 (sse_cvtss2siq_2): Likewise.
3132 (sse_cvttss2si): Likewise.
3133 (sse_cvtss2siq_2): Likewise.
3134 (float<sseintvecmodelower><mode>2): Likewise.
3135 (sse2_cvtsd2si_2): Likewise.
3136 (sse2_cvtsd2siq_2): Likewise.
3137 (*<plusminus_insn><mode>3): Likewise.
3138 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
3139 (*<sse4_1_avx2>_mul<mode>3): Likewise.
3140 (ashr<mode>3): Likewise.
3141 (<shift_insn><mode>3): Likewise.
3142 (avx2_<code><mode>3): Likewise.
3143 (*avx2_<code><mode>3): Likewise.
3144 (*andnot<mode>3): Likewise.
3145 (*<code><mode>3) <any_logic>: Likewise.
3146 (abs<mode>2): Likewise.
3147 (avx2_permvar<mode>): Likewise.
3148 (avx2_perm<mode>_1): Likewise.
3149 (*avx_vpermilp<mode>): Likewise.
3150 (avx_vpermilvar<mode>3): Likewise.
3151 (avx2_ashrv<mode>): Likewise.
3152 (avx2_<shift_insn>v<mode>): Likewise.
3153 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
3155 * doc/rtl.texi: Document XImode.
3157 2013-08-21 Jeff Law <law@redhat.com>
3159 * tree-flow.h (register_jump_thread): Pass vector of edges
3160 instead of each important edge.
3161 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
3162 thread path into a vector and pass that to register_jump_thread.
3163 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
3164 passed in edge vector to the current 3-edge form.
3167 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3169 * dce.c (fini_dce): Call df_analyze again just in case
3170 delete_unmarked_insns removed anything.
3172 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
3174 * reload.h (struct reg_equivs): Rename to ..
3175 (struct reg_equivs_s): .. this.
3177 2013-08-20 Martin Liska <marxin.liska@gmail.com>
3179 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
3181 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3183 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
3185 2013-08-21 Jeff Law <law@redhat.com>
3187 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
3188 simplify assignments too. If the RHS collapses to a singleton
3189 range, then return the value for the range.
3191 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
3193 * config/i386/sse.md (V16): Rename to...
3195 (mov<mode>): Update iterator name.
3196 (*mov<mode>_internal): Ditto.
3197 (push<mode>1): Ditto.
3198 (movmisalign<mode>): Ditto.
3200 2013-08-20 Jan Hubicka <jh@suse.cz>
3203 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
3204 entry for direct edges.
3205 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
3207 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3209 Revert my last two changes, r201865 and r201864:
3212 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3214 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3215 instances can own GC refs.
3217 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3218 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3219 (gcc::context::gt_pch_nx): Likewise.
3220 (gcc::context::gt_pch_nx): Likewise.
3221 * ggc.h (gt_ggc_mx <T>): New.
3222 (gt_pch_nx_with_op <T>): New.
3223 (gt_pch_nx <T>): New.
3224 * passes.c (opt_pass::gt_ggc_mx): New.
3225 (opt_pass::gt_pch_nx): New.
3226 (opt_pass::gt_pch_nx_with_op): New.
3227 (pass_manager::gt_ggc_mx): New.
3228 (pass_manager::gt_pch_nx): New.
3229 (pass_manager::gt_pch_nx_with_op): New.
3230 (pass_manager::operator new): Use
3231 ggc_internal_cleared_alloc_stat rather than xcalloc.
3232 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3233 (pass_manager::gt_ggc_mx): New.
3234 (pass_manager::gt_pch_nx): New.
3235 (pass_manager::gt_pch_nx_with_op): New.
3236 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3237 (opt_pass::operator new): New.
3238 (opt_pass::gt_ggc_mx): New.
3239 (opt_pass::gt_pch_nx): New.
3240 (opt_pass::gt_pch_nx_with_op): New.
3243 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3245 * Makefile.in (GTFILES): Add context.h.
3246 * context.c (gcc::context::operator new): New.
3247 (gcc::context::gt_ggc_mx): New.
3248 (gcc::context::gt_pch_nx): New.
3249 (gcc::context::gt_pch_nx): New.
3250 * context.h (gcc::context): Add GTY((user)) marking.
3251 (gcc::context::operator new): New.
3252 (gcc::context::gt_ggc_mx): New.
3253 (gcc::context::gt_pch_nx): New.
3254 (gcc::context::gt_pch_nx): New.
3255 (g): Add GTY marking.
3256 (gt_ggc_mx (gcc::context *)): New.
3257 (gt_pch_nx (gcc::context *)): New.
3258 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3259 void *cookie)): New.
3260 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3262 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3264 * dce.c (fini_dce): Call df_analyze again just in case
3265 delete_unmarked_insns removed anything.
3267 2013-08-20 Teresa Johnson <tejohnson@google.com>
3269 PR rtl-optimizations/57451
3270 * final.c (reemit_insn_block_notes): Prevent lexical blocks
3271 from crossing split section boundaries.
3273 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
3275 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
3276 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
3277 with MULTLIB_DEFAULTS.
3279 2013-08-20 Nick Clifton <nickc@redhat.com>
3281 * target.def (narrow_volatile_bitfield): Note that the default
3282 value is false, not !TARGET_STRICT_ALIGN.
3283 * doc/tm.texi: Regenerate.
3285 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
3287 Fix LIB_SPEC for systems without libpthread.
3289 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
3290 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
3292 * config/i386/linux-common.h: Likewise.
3293 * config/mips/linux-common.h: Likewise.
3295 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
3297 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
3300 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3302 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3303 instances can own GC refs.
3305 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3306 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3307 (gcc::context::gt_pch_nx): Likewise.
3308 (gcc::context::gt_pch_nx): Likewise.
3309 * ggc.h (gt_ggc_mx <T>): New.
3310 (gt_pch_nx_with_op <T>): New.
3311 (gt_pch_nx <T>): New.
3312 * passes.c (opt_pass::gt_ggc_mx): New.
3313 (opt_pass::gt_pch_nx): New.
3314 (opt_pass::gt_pch_nx_with_op): New.
3315 (pass_manager::gt_ggc_mx): New.
3316 (pass_manager::gt_pch_nx): New.
3317 (pass_manager::gt_pch_nx_with_op): New.
3318 (pass_manager::operator new): Use
3319 ggc_internal_cleared_alloc_stat rather than xcalloc.
3320 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3321 (pass_manager::gt_ggc_mx): New.
3322 (pass_manager::gt_pch_nx): New.
3323 (pass_manager::gt_pch_nx_with_op): New.
3324 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3325 (opt_pass::operator new): New.
3326 (opt_pass::gt_ggc_mx): New.
3327 (opt_pass::gt_pch_nx): New.
3328 (opt_pass::gt_pch_nx_with_op): New.
3330 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3332 * Makefile.in (GTFILES): Add context.h.
3333 * context.c (gcc::context::operator new): New.
3334 (gcc::context::gt_ggc_mx): New.
3335 (gcc::context::gt_pch_nx): New.
3336 (gcc::context::gt_pch_nx): New.
3337 * context.h (gcc::context): Add GTY((user)) marking.
3338 (gcc::context::operator new): New.
3339 (gcc::context::gt_ggc_mx): New.
3340 (gcc::context::gt_pch_nx): New.
3341 (gcc::context::gt_pch_nx): New.
3342 (g): Add GTY marking.
3343 (gt_ggc_mx (gcc::context *)): New.
3344 (gt_pch_nx (gcc::context *)): New.
3345 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3346 void *cookie)): New.
3347 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3349 2013-08-20 Alan Modra <amodra@gmail.com>
3352 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
3353 (rs6000_emit_epilogue): Likewise.
3355 2013-08-19 Dehao Chen <dehao@google.com>
3357 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
3359 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
3360 Jakub Jelinek <jakub@redhat.com>
3362 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
3363 (BUILT_IN_FABSD64): Likewise.
3364 (BUILT_IN_FABSD128): Likewise.
3365 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
3366 (fold_builtin_1): Likewise.
3367 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
3368 destination and source operands.
3369 (*abstd2_fpr): Likewise.
3370 (*nabstd2_fpr): Likewise.
3372 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
3374 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
3377 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
3379 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
3382 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
3384 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
3385 * builtins.c (default_libc_has_function): New.
3386 (gnu_libc_has_function): Ditto.
3387 (no_c99_libc_has_function): Ditto.
3388 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
3389 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
3390 (fold_builtin_sincos): Likewise.
3391 (fold_builtin_cexp): Likewise.
3392 * builtins.def (DEF_C94_BUILTIN): Likewise.
3393 (DEF_C99_BUILTIN): Likewise.
3394 (DEF_C99_C90RES_BUILTIN): Likewise.
3395 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
3396 definitions to using this define.
3397 * config/darwin-protos.h (darwin_libc_has_function): New.
3398 * config/darwin.c (darwin_libc_has_function): Ditto.
3399 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
3400 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
3401 * config/darwin.h: Ditto.
3402 * config/elfos.h: Ditto.
3403 * config/freebsd.h: Ditto.
3404 * config/i386/cygming.h: Ditto.
3405 * config/i386/djgpp.h: Ditto.
3406 * config/i386/i386-interix.h: Ditto.
3407 * config/microblaze/microblaze.h: Ditto.
3408 * config/mmix/mmix.h: Ditto.
3409 * config/gnu-user.h: Ditto.
3410 * config/ia64/hpux.h: Ditto.
3411 * config/pa/pa-hpux.h: Ditto.
3412 * config/pdp11/pdp11.h: Ditto.
3413 * config/picochip/picochip.h: Ditto.
3414 * config/linux.h: Ditto.
3415 * config/netbsd.h: Ditto.
3416 * config/openbsd.h: Ditto.
3417 * config/rs6000/aix43.h: Ditto.
3418 * config/rs6000/aix51.h: Ditto.
3419 * config/rs6000/aix52.h: Ditto.
3420 * config/rs6000/aix53.h: Ditto.
3421 * config/rs6000/aix61.h: Ditto.
3422 * config/rs6000/darwin.h: Ditto.
3423 * config/rs6000/linux.h: Ditto.
3424 * config/rs6000/linux64.h: Ditto.
3425 * config/s390/tpf.h: Ditto.
3426 * config/sol2-10.h: Ditto.
3427 * config/sol2.h: Ditto.
3428 * config/vms/vms.h: Ditto.
3429 * config/vxworks.h: Ditto.
3430 * config/linux-android.c (linux_android_libc_has_function):
3431 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
3432 * config/linux-protos.h (linux_android_libc_has_function):
3434 * config/i386/i386.c (ix86_libc_has_function): New.
3435 * config/i386/i386-protos.h
3436 (ix86_libc_has_function): New declaration.
3437 * config/i386/i386.md
3438 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
3439 ("isinf<mode>2): Likewise.
3440 * convert.c (convert_to_integer): Using new target hook
3441 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
3442 TARGET_C99_FUNCTIONS.
3443 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
3444 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
3445 * coretypes.h (function_class): New enum for different
3446 classes of functions.
3447 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
3448 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
3449 (TARGET_HAS_SINCOS): Likewise.
3450 (TARGET_LIBC_HAS_FUNCTION): New.
3451 * doc/tm.texi: Regenerated.
3452 * targhooks.h (default_libc_has_function): New declaration.
3453 (no_c99_libc_has_function): Ditto.
3454 (gnu_libc_has_function): Ditto.
3455 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
3456 and TARGET_HAS_SINCOS.
3458 2013-08-18 Jan Hubicka <jh@suse.cz>
3460 * Makeifle-in (ipa-devirt.o): New.
3461 (GTFILES): Add ipa-utils.h and ipa-devirt.c
3462 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
3463 (analyze_functions): Look into possible targets of polymorphic call.
3464 * dumpfile.c (dump_files): Add type-inheritance dump.
3465 * dumpfile.h (TDI_inheritance): New.
3466 * ipa-devirt.c: New file.
3467 * ipa-utils.h (odr_type_d): Forward declare.
3468 (odr_type): New type.
3469 (build_type_inheritance_graph): Declare.
3470 (possible_polymorphic_call_targets): Declare and introduce inline
3471 variant when only edge is pased.
3472 (dump_possible_polymorphic_call_targets): Likewise.
3473 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3474 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3475 (types_same_for_odr): ... here.
3476 * tree.h (type_in_anonymous_namespace_p): Declare.
3478 2013-08-18 Jakub Jelinek <jakub@redhat.com>
3480 PR tree-optimization/58006
3481 * tree-parloops.c (take_address_of): Don't ICE if get_name
3483 (eliminate_local_variables_stmt): Remove clobber stmts.
3485 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3487 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3488 error message is issued for an alias to undefined symbol.
3490 2013-08-18 Jan Hubicka <jh@suse.cz>
3492 * cgraph.c (cgraph_create_indirect_edge): Discover
3493 polymorphic calls and record basic info into indirect_info.
3494 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3495 devirtualization, ignore objc function calls.
3496 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3497 call with no parm index info.
3498 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3499 * tree.c (virtual_method_call_p): New function.
3500 * tree.h (virtual_method_call_p): Declare.
3502 2013-08-16 Jan Hubicka <jh@suse.cz>
3505 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3507 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
3509 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3512 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3514 * gengtype.c (type_for_name): Add special-case support for
3515 locating types within the "gcc::" namespace.
3516 (open_base_files): Emit a "using namespace gcc" directive.
3518 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3521 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3522 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3524 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3525 array instead of each individual operand as a separate argument.
3526 (emit_fusion_gpr_load): Likewise.
3527 (expand_fusion_gpr_load): Add new function declaration.
3529 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3530 signature to have the operands passed as an array, instead of as
3531 separate arguments. Allow ZERO_EXTEND to be in the memory
3532 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3533 depend on the register live/dead flags when peepholes are run.
3534 (expand_fusion_gpr_load): New function to be called from the
3535 peephole2 pass, to change the register that addis sets to be the
3537 (emit_fusion_gpr_load): Change the calling signature to have the
3538 operands passed as an array, instead of as separate arguments.
3539 Allow ZERO_EXTEND to be in the memory address, and also
3540 SIGN_EXTEND if -mpower8-fusion-sign.
3542 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3544 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3545 adjust the register addis loads up in the peephole2 pass. Do not
3546 depend on the register live/dead state when the peephole pass is done.
3548 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3550 * gengtype.c (create_user_defined_type): Ensure that the kind
3551 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3552 declaration is seen before the GTY((user)) marking.
3554 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
3557 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
3559 2013-08-16 Jan Hubicka <jh@suse.cz>
3561 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
3562 arugment expected_type.
3563 (gimple_fold_call): Use it.
3564 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
3565 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
3566 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
3567 (try_make_edge_direct_virtual_call): Likewise.
3568 * tree.c (obj_type_ref_class): New.
3569 * tree.h (obj_type_ref_class): Use it.
3571 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
3573 * sched-vis.c (rtl_slim_pp_initialized): Remove.
3574 (rtl_slim_pp): Likewise.
3575 (init_rtl_slim_pretty_print): Likewise.
3576 (dump_value_slim): Don't call it. Use local pretty printer.
3577 (dump_insn_slim): Likewise.
3578 (dump_rtl_slim): Likewise.
3579 (str_pattern_slim): Likewise.
3580 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
3583 2013-08-16 Jakub Jelinek <jakub@redhat.com>
3585 PR tree-optimization/58164
3586 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
3587 walk gimple_goto_dest of GIMPLE_GOTO.
3589 PR tree-optimization/58165
3590 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
3591 bi_call must be the last stmt in a bb, don't split_block, instead
3592 use fallthru edge from it and give up if there is none.
3593 Release conds vector when returning early.
3595 2013-08-14 Xinliang David Li <davidxl@google.com>
3597 * config/i386/i386.c (ix86_option_override_internal):
3598 Remove unused variable and field.
3600 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3603 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
3604 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
3605 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
3606 and Linux, correct BLKmode alignment when 128-bit alignment is
3607 required and compatibility flag is not set.
3608 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
3609 for zero-size arguments when compatibility flag is not set.
3611 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3613 PR tree-optimization/58145
3614 * tree-sra.c (build_ref_for_offset): If prev_base has
3615 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
3617 2013-08-14 Xinliang David Li <davidxl@google.com>
3619 * config/i386/i386.c (ix86_option_override_internal):
3620 Fix uninitialized variable error.
3622 2013-08-14 Xinliang David Li <davidxl@google.com>
3624 * config/i386/i386.opt: Define two new options.
3625 * config/i386/x86-tune.def: Add arch selector field in macros.
3626 * config/i386/i386.h: Adjust macro definition.
3627 * config/i386/i386.c (ix86_option_override_internal):
3629 (parse_mtune_ctrl_str): New function.
3630 (set_ix86_tune_features): New function.
3631 (ix86_function_specific_restore): Call the new helper function.
3633 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
3635 PR rtl-optimization/57662
3636 * sel-sched.c (code_motion_process_successors): When the current insn
3637 is removed after the recursive traversal, break from the loop.
3638 Add comments and debug printouts.
3640 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3641 Alexandre Oliva <aoliva@redhat.com>
3644 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3645 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3646 there also UNSPEC_PLTOFF.
3648 2013-08-14 Marek Polacek <polacek@redhat.com>
3650 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3651 NUM_CONDITIONS bit positions.
3653 2013-08-13 Cary Coutant <ccoutant@google.com>
3655 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3656 (attr_checksum): Hash vector contents instead of pointer.
3657 (attr_checksum_ordered): Likewise.
3659 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
3661 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3662 when Pmode != word_mode. Add length_address attribute.
3663 (sse3_monitor_<mode>): Merge from sse3_monitor and
3664 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3665 Pmode != word_mode. Update insn length attribute.
3666 * config/i386/i386.c (ix86_option_override_internal): Update
3667 ix86_gen_monitor selection for merged sse3_monitor insn.
3669 2013-08-13 Julian Brown <julian@codesourcery.com>
3671 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3672 perform invalid legitimization on greater-than-word-size modes for
3675 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3677 * ira.c (setup_class_translate_array): Use aclass instead of cl
3678 for classes not fully covered by allocno classes.
3680 2013-08-13 Jakub Jelinek <jakub@redhat.com>
3682 PR tree-optimization/57661
3683 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3684 * tree-inline.c (tree_function_versioning): Initialize it.
3685 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3686 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3687 that is not being copied.
3690 * asan.c (instrument_strlen_call): Fix typo in comment.
3691 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3693 2013-08-13 Steve Ellcey <sellcey@mips.com>
3695 * config/mips/mips.md (prefetch): Use lw instead of ld on
3696 loongson in 32bit mode.
3698 2013-08-13 Nick Clifton <nickc@redhat.com>
3700 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3702 2013-08-13 Jan Hubicka <jh@suse.cz>
3704 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3705 introduced edge; fix typo in sanity check.
3706 (cgraph_resolve_speculation): Export; improve diagnostic.
3707 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3708 speculation at type mismatch.
3709 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3710 (cgraph_resolve_speculation): Declare.
3711 (symtab_can_be_discarded): New function.
3712 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3713 * ipa-inline-transform.c (speculation_removed): New global var.
3714 (clone_inlined_nodes): See if speculation can be removed.
3715 (inline_call): If speculations was removed, we growths may not match.
3716 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3717 (speculation_useful_p): New function.
3718 (resolve_noninline_speculation): New function.
3719 (inline_small_functions): Resolve useless speculations.
3720 * ipa-inline.h (speculation_useful_p): Declare
3721 * ipa.c (can_replace_by_local_alias): Simplify.
3722 (ipa_profile): Produce speculative calls in non-lto, too;
3723 add simple cost model; produce local aliases.
3725 2013-08-13 David Malcolm <dmalcolm@redhat.com>
3727 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
3730 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3732 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3734 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
3736 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3737 iterator instead of X87MODEF.
3739 2013-08-12 Perez Read <netfirewall@gmail.com>
3742 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3743 operand 0 for intel asm alternative.
3744 (*movabs<mode>_2): Ditto for operand 1.
3746 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3748 * config/aarch64/arm_none.h
3749 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3751 2013-08-12 Nick Clifton <nickc@redhat.com>
3753 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3755 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3757 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3758 expand for QI/HImode operand to produce more effictive code for
3759 unsigned char(short) --> float(double) conversion.
3761 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
3763 * doc/invoke.texi: Mention that -ftls-model does not force the final
3766 2013-08-12 Marek Polacek <polacek@redhat.com>
3767 Marc Glisse <marc.glisse@inria.fr>
3769 PR tree-optimization/57980
3770 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3771 when creating -1 constant.
3773 2013-08-10 Jan Hubicka <jh@suse.cz>
3775 Workaround binutils PR14342.
3776 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3777 (gimple_init_edge_profiler): Likewise.
3778 (gimple_gen_ic_func_profiler): Likewise.
3780 2013-08-09 Jan Hubicka <jh@suse.cz>
3782 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3784 2013-08-09 Xinliang David Li <davidxl@google.com>
3786 * config/i386/stringop.def: New file.
3787 * config/i386/stringop.opt: New file.
3788 * config/i386/i386-opts.h: Include stringopt.def.
3789 * config/i386/i386.opt: Include stringopt.opt.
3790 * config/i386/i386.c (ix86_option_override_internal):
3791 Override default size based stringop inline strategies with options.
3792 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3795 2013-08-09 Jan Hubicka <jh@suse.cz>
3797 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
3799 2013-08-09 Jan Hubicka <jh@suse.cz>
3801 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3803 (dump_cgraph_node): Dump profile-id.
3804 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3805 and common_target_probability.
3806 * lto-cgraph.c (lto_output_edge): Stream common targets.
3807 (lto_output_node): Stream profile ids.
3808 (input_node): Stream profile ids.
3809 (input_edge): Stream common targets.
3810 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3811 * ipa.c: Include value-prof.h
3812 (ipa_profile_generate_summary): Turn indirect call statement histograms
3813 into common targets.
3814 (ipa_profile): Turn common targets into speculative edges.
3816 2013-08-09 Jan Hubicka <jh@suse.cz>
3818 * cgraph.h (cgraph_node): Add profile_id.
3819 * value-prof.c (cgraph_node_map): Turn into pointer_map.
3820 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
3821 (del_node_map): Update.
3822 (find_func_by_funcdef_no): Replace by ...
3823 (find_func_by_profile_id): ... this one.
3824 (gimple_ic_transform): Do not remove useful histograms when
3825 speculation is not done; dump info when indirect call removal
3827 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
3828 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
3829 (__gcov_indirect_call_profiler_v2): .. this one.
3830 * profile.h (init_node_map): Update.
3831 * coverage.c (coverage_compute_profile_id): New function.
3832 * coverage.h (coverage_compute_profile_id): Declare.
3833 * tree-profile.c (init_ic_make_global_vars): Make
3834 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
3835 (gimple_init_edge_profiler): Update prototype of
3836 __gcov_indirect_call_profiler.
3837 (gimple_gen_ic_func_profiler): Simplify.
3838 (tree_profiling): Use init_node_map
3840 2013-08-09 Jan Hubicka <jh@suse.cz>
3842 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
3843 non-speculative refs.
3844 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
3845 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
3846 (cgraph_set_call_stmt): Likewise.
3847 (cgraph_create_edge_1): Fix release checking compilatoin;
3849 (cgraph_free_edge): Free indirect info.
3850 (cgraph_turn_edge_to_speculative): New function.
3851 (cgraph_speculative_call_info): New function.
3852 (cgraph_make_edge_direct): Return direct edge; handle speculation.
3853 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
3854 (dump_cgraph_node): Dump speculation.
3855 (verify_edge_count_and_frequency): Accept speculative edges.
3856 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
3857 (verify_cgraph_node): Handle speculation.
3858 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
3859 (cgraph_set_call_stmt): Update prototype.
3860 (cgraph_make_edge_direct): Update prototype.
3861 (cgraph_speculative_call_info): Declare.
3862 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
3863 to change; update call of ipa_find_references.
3864 * ipa-ref.c (ipa_record_reference): Fix return value; clear
3865 lto_stmt_uid and speculative flags.
3866 (ipa_dump_references): Dump speculation.
3867 (ipa_clone_references): Clone speculative flag.
3868 (ipa_clone_referring): Likewise.
3869 (ipa_clone_ref): New function.
3870 (ipa_find_reference): Look into lto_stmt_uids
3871 (ipa_clear_stmts_in_references): Do not clear speculative calls.
3872 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
3873 (ipa_find_reference): Update declaration.
3874 (ipa_clone_ref): Declare.
3875 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
3876 stream speculative flag.
3877 (lto_output_ref): Stream statements uids and speculation.
3878 (input_ref): Likewise.
3879 (input_edge): Stream speuclation.
3880 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
3881 (cgraph_set_call_stmt_including_clones): Handle speculation.
3882 * ipa-inline.c (heap_edge_removal_hook): New function.
3883 (inline_small_functions): Register it.
3884 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
3885 also initialize refs.
3886 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
3888 (try_make_edge_direct_simple_call): Likewise.
3889 (try_make_edge_direct_simple_call): Likewise.
3890 (update_indirect_edges_after_inlining): Likewise.
3891 (remove_described_reference): Look proper lto_stmt_uid.
3892 (propagate_controlled_uses): Likewise.
3893 (propagate_controlled_uses): Liekwise.
3894 * tree-inline.c (copy_bb): Copy speculative edges.
3895 (redirect_all_calls): New function.
3896 (copy_cfg_body): Do redirection after loop info is updated.
3897 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
3899 2013-08-09 Jan Hubicka <jh@suse.cz>
3901 * lto-streamer-out.c (output_function): Renumber PHIs.
3902 * lto-streamer-in.c (input_function): Likewise.
3904 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
3906 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
3907 (get_lane_unsigned): Likewise.
3908 (dup_lane_scalar): Likewise.
3909 (get_lane): enable for VALL.
3910 * config/aarch64/aarch64-simd.md
3911 (aarch64_dup_lane_scalar<mode>): Remove.
3912 (aarch64_get_lane_signed<mode>): Likewise.
3913 (aarch64_get_lane_unsigned<mode>): Likewise.
3914 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
3915 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
3916 (aarch64_get_lane<mode>): Enable for all vector modes.
3917 (aarch64_get_lanedi): Remove misleading constraints.
3918 * config/aarch64/arm_neon.h
3919 (__aarch64_vget_lane_any): Define.
3920 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
3921 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
3922 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
3923 * config/aarch64/iterators.md (VDQQH): New.
3927 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
3929 * configure.ac: Add GAS check for LEON instructions on SPARC.
3930 * configure: Regenerate.
3931 * config.in: Likewise.
3932 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
3934 * config/sparc/sparc.opt (LEON, LEON3): New masks.
3935 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
3937 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
3938 (AS_LEON_FLAG): New macro.
3939 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
3940 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
3941 Deal with LEON and LEON3 for the memory model.
3942 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
3943 (atomic_compare_and_swap<mode>_1): Likewise.
3944 (*atomic_compare_and_swap<mode>_1): Likewise.
3946 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3948 * config/arm/neon.md (vcond): Fix floating-point vector
3949 comparisons against 0.
3951 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
3953 * lra-constraints.c (emit_spill_move): Remove assert.
3954 (process_alt_operands): Add more debugging
3955 output. Increase reject for spilling into memory. Decrease
3956 reject for reloading scratch.
3957 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
3959 2013-08-08 Steve Ellcey <sellcey@mips.com>
3961 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
3962 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
3963 micromips incompatible. Add nan2008.
3964 (MULTILIB_DIRNAMES): Add nan2008.
3965 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3966 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
3967 and micromips incompatible. Add nan2008.
3968 (MULTILIB_DIRNAMES): Add nan2008.
3969 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3971 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
3973 PR rtl-optimization/58079
3974 * combine.c (combine_simplify_rtx): Avoid using SUBST if
3975 simplify_comparison has widened a comparison with an integer.
3977 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3979 * config/arm/neon.md (movmisalign<mode>): Disable when we
3980 don't allow unaligned accesses.
3981 (*movmisalign<mode>_neon_store): Likewise.
3982 (*movmisalign<mode>_neon_load): Likewise.
3983 (*movmisalign<mode>_neon_store): Likewise.
3984 (*movmisalign<mode>_neon_load): Likewise.
3986 2013-08-08 Jan Hubicka <jh@suse.cz>
3988 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
3989 (make_pass_rebuild_cgraph_edges): Also clear references.
3990 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
3991 * ipa-inline-transform.c (inline_transform): Remove all references
3993 * cgraphunit.c (expand_function): Remove all references after
3995 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
3996 (ipa_find_reference): Rewrite to iterator.
3997 (remove_stmt_references): Likewise.
3998 (ipa_clear_stmts_in_references): New function.
3999 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
4000 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
4002 * ipa-split.c (split_function): Remove references in split function.
4004 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4007 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
4008 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
4010 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4013 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
4014 suggested mode for the assignment isn't compatible with the
4017 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
4020 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
4022 2013-08-07 Xinliang David Li <davidxl@google.com>
4024 * config/i386/i386.opt: New option -mtune-ctrl=.
4025 * config/i386/x86-tune.def: New file.
4026 * config/i386/i386.h: include x86-tune.def.
4027 * config/i386/i386.c (ix86_option_override_internal):
4028 Parsing -mtune-ctrl= option and set tune features.
4030 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
4033 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
4034 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
4037 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4039 * rtl.h (update_alignments): Declare.
4040 * final.c (grow_label_align): New function extracted from...
4041 (shorten_branches): ...here. Call it.
4042 (update_alignments): New function.
4043 * reorg.c (sibling_labels): New variable.
4044 (get_label_before): Add SIBLING parameter. If it is non-zero, push
4045 the new label along with it onto the sibling_labels vector.
4046 (fill_simple_delay_slots): Adjust call to get_label_before.
4047 (fill_slots_from_thread): Likewise.
4048 (relax_delay_slots): Likewise.
4049 (make_return_insns): Likewise.
4050 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
4052 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4054 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
4055 document its semantics.
4056 (diagnostic_report_diagnostic): Adjust accordingly.
4058 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4060 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
4061 (sparc_option_override): ...and port to new C++ pass API.
4062 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
4064 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
4066 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
4068 2013-08-06 Caroline Tice <cmtice@google.com>
4070 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4071 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
4072 * tree-pass.h: Add pass_vtable_verify.
4073 * varasm.c (assemble_variable): Add code to properly set the comdat
4074 section and name for the .vtable_map_vars section.
4075 (assemble_vtyv_preinit_initializer): New function.
4076 (default_sectin_type_flags): Make sure .vtable_map_vars section has
4078 * output.h: Add function decl for assemble_vtv_preinit_initializer.
4079 * vtable-verify.c: New file.
4080 * vtable-verify.h: New file.
4081 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
4082 initialiation levels.
4083 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
4084 * passes.def: Insert pass_vtable_verify.
4085 * aclocal.m4: Reorder includes.
4086 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
4087 -fvtv-counts options.
4088 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
4089 as appropriate, if -fvtable-verify=... is used.
4090 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
4091 -fvtable-verify=... is used.
4092 * Makefile.in (OBJS): Add vtable-verify.o to list.
4093 (vtable-verify.o): Add new build rule.
4094 (GTFILES): Add vtable-verify.c to list.
4095 * common.opt (fvtable-verify=): New flag.
4096 (vtv_priority): Values for fvtable-verify= flag.
4097 (fvtv-counts): New flag.
4098 (fvtv-debug): New flag.
4099 * tree.h (save_vtable_map_decl): New extern function decl.
4101 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4103 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
4104 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
4105 (pass_data_rl78_devirt): ...new pass_data instance and...
4106 (make_pass_rl78_devirt): ...new function.
4107 (rl78_asm_file_start): Port pass registration to new C++ API.
4109 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4111 * coretypes.h (rtl_opt_pass): Add.
4112 (gcc::context): Add.
4113 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
4114 (epiphany_init): Port to new C++ pass API.
4115 (epiphany_optimize_mode_switching): Likewise.
4116 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
4117 (pass_manager::get_pass_mode_switching): New.
4118 (pass_manager::get_pass_peephole2): New.
4119 * mode-switching.c (pass_mode_switching): Add clone method.
4120 * recog.c (pass_peephole2): Add clone method.
4121 (pass_split_all_insns): Add clone method.
4123 2013-08-06 David Malcolm <dmalcolm@redhat.com>
4125 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
4126 (mips_option_override): ...here, porting to new C++ API for passes.
4128 2013-08-06 Jan Hubicka <jh@suse.cz>
4130 * cgraph.c (cgraph_get_body): New function based on lto.c
4132 * cgraph.h (cgraph_get_body): Declare.
4133 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
4135 * cgraphunit.c (expand_function): Get body prior expanding.
4136 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
4137 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
4139 * passes.c (do_per_function_toporder): Get body.
4140 * tree-inline.c (expand_call_inline): Get body prior inlining it.
4141 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
4143 2013-08-06 Martin Jambor <mjambor@suse.cz>
4146 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
4147 is not re-finalized. Rename second parameter to no_collect.
4149 2013-08-06 Martin Jambor <mjambor@suse.cz>
4152 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
4153 MEM_REF has proper alignment information.
4155 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
4158 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
4160 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
4161 rtx (*) (rtx, ...) with insn_gen_fn.
4162 * genoutput.c (output_insn_data): Cast gen_? function pointers to
4163 insn_gen_fn::stored_funcptr. Add initializer braces.
4165 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4167 Rewrite how instances of passes are cloned to remove assumptions
4168 about their sizes (thus allowing pass subclasses to have
4169 additional data fields, albeit non-GC-managed ones at this point).
4171 * passes.c (make_pass_instance): Now that passes have clone
4172 methods, rewrite this function to eliminate XNEW and memcpy
4173 calls that used hardcoded sizes. Since this function no longer
4174 creates pass instances, rename it to...
4175 (add_pass_instance): ...this. Document the old way that passes were
4176 numbered and flagged, and rework this function to continue using it.
4177 (next_pass_1): Add an initial_pass argument for use by
4179 (position_pass): When adding multiple instances of a pass, use
4180 the pass's clone method, rather than relying on the XNEW/memcpy
4181 within the former make_pass_instance (now add_pass_instance).
4182 (pass_manager::pass_manager): When invoking next_pass_1, also supply
4183 the initial instance of the current pass within the pass manager.
4185 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4187 This is the automated part of the conversion of passes from C
4188 structs to C++ classes.
4190 Patch autogenerated by refactor_passes.py from
4191 https://github.com/davidmalcolm/gcc-refactoring-scripts
4192 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
4194 * asan.c (pass_asan): Convert from a global struct to a subclass of
4195 gimple_opt_pass along with...
4196 (pass_data_asan): ...new pass_data instance and...
4197 (make_pass_asan): ...new function.
4198 (pass_asan_O0): Convert from a global struct to a subclass of
4199 gimple_opt_pass along with...
4200 (pass_data_asan_O0): ...new pass_data instance and...
4201 (make_pass_asan_O0): ...new function.
4202 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
4203 subclass of rtl_opt_pass along with...
4204 (pass_data_inc_dec): ...new pass_data instance and...
4205 (make_pass_inc_dec): ...new function.
4206 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
4207 a subclass of rtl_opt_pass along with...
4208 (pass_data_reorder_blocks): ...new pass_data instance and...
4209 (make_pass_reorder_blocks): ...new function.
4210 (pass_duplicate_computed_gotos): Convert from a global struct to a
4211 subclass of rtl_opt_pass along with...
4212 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
4213 (make_pass_duplicate_computed_gotos): ...new function.
4214 (pass_partition_blocks): Convert from a global struct to a subclass of
4215 rtl_opt_pass along with...
4216 (pass_data_partition_blocks): ...new pass_data instance and...
4217 (make_pass_partition_blocks): ...new function.
4218 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
4219 struct to a subclass of rtl_opt_pass along with...
4220 (pass_data_branch_target_load_optimize1): ...new pass_data instance
4222 (make_pass_branch_target_load_optimize1): ...new function.
4223 (pass_branch_target_load_optimize2): Convert from a global struct to a
4224 subclass of rtl_opt_pass along with...
4225 (pass_data_branch_target_load_optimize2): ...new pass_data instance
4227 (make_pass_branch_target_load_optimize2): ...new function.
4228 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
4229 of rtl_opt_pass along with...
4230 (pass_data_jump): ...new pass_data instance and...
4231 (make_pass_jump): ...new function.
4232 (pass_jump2): Convert from a global struct to a subclass of
4233 rtl_opt_pass along with...
4234 (pass_data_jump2): ...new pass_data instance and...
4235 (make_pass_jump2): ...new function.
4236 * cfgexpand.c (pass_expand): Convert from a global struct to a
4237 subclass of rtl_opt_pass along with...
4238 (pass_data_expand): ...new pass_data instance and...
4239 (make_pass_expand): ...new function.
4240 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
4241 of rtl_opt_pass along with...
4242 (pass_data_free_cfg): ...new pass_data instance and...
4243 (make_pass_free_cfg): ...new function.
4244 (pass_into_cfg_layout_mode): Convert from a global struct to a
4245 subclass of rtl_opt_pass along with...
4246 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
4247 (make_pass_into_cfg_layout_mode): ...new function.
4248 (pass_outof_cfg_layout_mode): Convert from a global struct to a
4249 subclass of rtl_opt_pass along with...
4250 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
4251 (make_pass_outof_cfg_layout_mode): ...new function.
4252 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
4253 struct to a subclass of gimple_opt_pass along with...
4254 (pass_data_build_cgraph_edges): ...new pass_data instance and...
4255 (make_pass_build_cgraph_edges): ...new function.
4256 (pass_rebuild_cgraph_edges): Convert from a global struct to a
4257 subclass of gimple_opt_pass along with...
4258 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
4259 (make_pass_rebuild_cgraph_edges): ...new function.
4260 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
4261 subclass of gimple_opt_pass along with...
4262 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
4264 (make_pass_remove_cgraph_callee_edges): ...new function.
4265 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
4266 struct to a subclass of rtl_opt_pass along with...
4267 (pass_data_stack_adjustments): ...new pass_data instance and...
4268 (make_pass_stack_adjustments): ...new function.
4269 * combine.c (pass_combine): Convert from a global struct to a subclass
4270 of rtl_opt_pass along with...
4271 (pass_data_combine): ...new pass_data instance and...
4272 (make_pass_combine): ...new function.
4273 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
4274 global struct to a subclass of rtl_opt_pass along with...
4275 (pass_data_compare_elim_after_reload): ...new pass_data instance
4277 (make_pass_compare_elim_after_reload): ...new function.
4278 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
4279 of rtl_opt_pass along with...
4280 (pass_data_rtl_cprop): ...new pass_data instance and...
4281 (make_pass_rtl_cprop): ...new function.
4282 * cse.c (pass_cse): Convert from a global struct to a subclass of
4283 rtl_opt_pass along with...
4284 (pass_data_cse): ...new pass_data instance and...
4285 (make_pass_cse): ...new function.
4286 (pass_cse2): Convert from a global struct to a subclass of
4287 rtl_opt_pass along with...
4288 (pass_data_cse2): ...new pass_data instance and...
4289 (make_pass_cse2): ...new function.
4290 (pass_cse_after_global_opts): Convert from a global struct to a
4291 subclass of rtl_opt_pass along with...
4292 (pass_data_cse_after_global_opts): ...new pass_data instance and...
4293 (make_pass_cse_after_global_opts): ...new function.
4294 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
4295 of rtl_opt_pass along with...
4296 (pass_data_ud_rtl_dce): ...new pass_data instance and...
4297 (make_pass_ud_rtl_dce): ...new function.
4298 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
4299 rtl_opt_pass along with...
4300 (pass_data_fast_rtl_dce): ...new pass_data instance and...
4301 (make_pass_fast_rtl_dce): ...new function.
4302 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
4303 a subclass of rtl_opt_pass along with...
4304 (pass_data_df_initialize_opt): ...new pass_data instance and...
4305 (make_pass_df_initialize_opt): ...new function.
4306 (pass_df_initialize_no_opt): Convert from a global struct to a
4307 subclass of rtl_opt_pass along with...
4308 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
4309 (make_pass_df_initialize_no_opt): ...new function.
4310 (pass_df_finish): Convert from a global struct to a subclass of
4311 rtl_opt_pass along with...
4312 (pass_data_df_finish): ...new pass_data instance and...
4313 (make_pass_df_finish): ...new function.
4314 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
4315 rtl_opt_pass along with...
4316 (pass_data_rtl_dse1): ...new pass_data instance and...
4317 (make_pass_rtl_dse1): ...new function.
4318 (pass_rtl_dse2): Convert from a global struct to a subclass of
4319 rtl_opt_pass along with...
4320 (pass_data_rtl_dse2): ...new pass_data instance and...
4321 (make_pass_rtl_dse2): ...new function.
4322 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
4323 subclass of rtl_opt_pass along with...
4324 (pass_data_dwarf2_frame): ...new pass_data instance and...
4325 (make_pass_dwarf2_frame): ...new function.
4326 * except.c (pass_set_nothrow_function_flags): Convert from a global
4327 struct to a subclass of rtl_opt_pass along with...
4328 (pass_data_set_nothrow_function_flags): ...new pass_data instance
4330 (make_pass_set_nothrow_function_flags): ...new function.
4331 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
4332 subclass of rtl_opt_pass along with...
4333 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
4335 (make_pass_convert_to_eh_region_ranges): ...new function.
4336 * final.c (pass_compute_alignments): Convert from a global struct to a
4337 subclass of rtl_opt_pass along with...
4338 (pass_data_compute_alignments): ...new pass_data instance and...
4339 (make_pass_compute_alignments): ...new function.
4340 (pass_final): Convert from a global struct to a subclass of
4341 rtl_opt_pass along with...
4342 (pass_data_final): ...new pass_data instance and...
4343 (make_pass_final): ...new function.
4344 (pass_shorten_branches): Convert from a global struct to a subclass of
4345 rtl_opt_pass along with...
4346 (pass_data_shorten_branches): ...new pass_data instance and...
4347 (make_pass_shorten_branches): ...new function.
4348 (pass_clean_state): Convert from a global struct to a subclass of
4349 rtl_opt_pass along with...
4350 (pass_data_clean_state): ...new pass_data instance and...
4351 (make_pass_clean_state): ...new function.
4352 * function.c (pass_instantiate_virtual_regs): Convert from a global
4353 struct to a subclass of rtl_opt_pass along with...
4354 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
4355 (make_pass_instantiate_virtual_regs): ...new function.
4356 (pass_leaf_regs): Convert from a global struct to a subclass of
4357 rtl_opt_pass along with...
4358 (pass_data_leaf_regs): ...new pass_data instance and...
4359 (make_pass_leaf_regs): ...new function.
4360 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
4361 subclass of rtl_opt_pass along with...
4362 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
4364 (make_pass_thread_prologue_and_epilogue): ...new function.
4365 (pass_match_asm_constraints): Convert from a global struct to a
4366 subclass of rtl_opt_pass along with...
4367 (pass_data_match_asm_constraints): ...new pass_data instance and...
4368 (make_pass_match_asm_constraints): ...new function.
4369 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
4370 subclass of rtl_opt_pass along with...
4371 (pass_data_rtl_fwprop): ...new pass_data instance and...
4372 (make_pass_rtl_fwprop): ...new function.
4373 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
4374 rtl_opt_pass along with...
4375 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
4376 (make_pass_rtl_fwprop_addr): ...new function.
4377 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
4378 rtl_opt_pass along with...
4379 (pass_data_rtl_pre): ...new pass_data instance and...
4380 (make_pass_rtl_pre): ...new function.
4381 (pass_rtl_hoist): Convert from a global struct to a subclass of
4382 rtl_opt_pass along with...
4383 (pass_data_rtl_hoist): ...new pass_data instance and...
4384 (make_pass_rtl_hoist): ...new function.
4385 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
4386 subclass of gimple_opt_pass along with...
4387 (pass_data_lower_cf): ...new pass_data instance and...
4388 (make_pass_lower_cf): ...new function.
4389 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
4390 from a global struct to a subclass of gimple_opt_pass along with...
4391 (pass_data_strength_reduction): ...new pass_data instance and...
4392 (make_pass_strength_reduction): ...new function.
4393 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
4394 of rtl_opt_pass along with...
4395 (pass_data_rtl_ifcvt): ...new pass_data instance and...
4396 (make_pass_rtl_ifcvt): ...new function.
4397 (pass_if_after_combine): Convert from a global struct to a subclass of
4398 rtl_opt_pass along with...
4399 (pass_data_if_after_combine): ...new pass_data instance and...
4400 (make_pass_if_after_combine): ...new function.
4401 (pass_if_after_reload): Convert from a global struct to a subclass of
4402 rtl_opt_pass along with...
4403 (pass_data_if_after_reload): ...new pass_data instance and...
4404 (make_pass_if_after_reload): ...new function.
4405 * init-regs.c (pass_initialize_regs): Convert from a global struct to
4406 a subclass of rtl_opt_pass along with...
4407 (pass_data_initialize_regs): ...new pass_data instance and...
4408 (make_pass_initialize_regs): ...new function.
4409 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
4410 of ipa_opt_pass_d along with...
4411 (pass_data_ipa_cp): ...new pass_data instance and...
4412 (make_pass_ipa_cp): ...new function.
4413 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
4414 global struct to a subclass of gimple_opt_pass along with...
4415 (pass_data_inline_parameters): ...new pass_data instance and...
4416 (make_pass_inline_parameters): ...new function.
4417 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
4418 subclass of gimple_opt_pass along with...
4419 (pass_data_early_inline): ...new pass_data instance and...
4420 (make_pass_early_inline): ...new function.
4421 (pass_ipa_inline): Convert from a global struct to a subclass of
4422 ipa_opt_pass_d along with...
4423 (pass_data_ipa_inline): ...new pass_data instance and...
4424 (make_pass_ipa_inline): ...new function.
4425 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
4426 struct to a subclass of gimple_opt_pass along with...
4427 (pass_data_local_pure_const): ...new pass_data instance and...
4428 (make_pass_local_pure_const): ...new function.
4429 (pass_ipa_pure_const): Convert from a global struct to a subclass of
4430 ipa_opt_pass_d along with...
4431 (pass_data_ipa_pure_const): ...new pass_data instance and...
4432 (make_pass_ipa_pure_const): ...new function.
4433 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
4434 to a subclass of ipa_opt_pass_d along with...
4435 (pass_data_ipa_reference): ...new pass_data instance and...
4436 (make_pass_ipa_reference): ...new function.
4437 * ipa-split.c (pass_split_functions): Convert from a global struct to
4438 a subclass of gimple_opt_pass along with...
4439 (pass_data_split_functions): ...new pass_data instance and...
4440 (make_pass_split_functions): ...new function.
4441 (pass_feedback_split_functions): Convert from a global struct to a
4442 subclass of gimple_opt_pass along with...
4443 (pass_data_feedback_split_functions): ...new pass_data instance and...
4444 (make_pass_feedback_split_functions): ...new function.
4445 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
4446 global struct to a subclass of simple_ipa_opt_pass along with...
4447 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
4449 (make_pass_ipa_function_and_variable_visibility): ...new function.
4450 (pass_ipa_free_inline_summary): Convert from a global struct to a
4451 subclass of simple_ipa_opt_pass along with...
4452 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
4453 (make_pass_ipa_free_inline_summary): ...new function.
4454 (pass_ipa_whole_program_visibility): Convert from a global struct to a
4455 subclass of ipa_opt_pass_d along with...
4456 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
4458 (make_pass_ipa_whole_program_visibility): ...new function.
4459 (pass_ipa_profile): Convert from a global struct to a subclass of
4460 ipa_opt_pass_d along with...
4461 (pass_data_ipa_profile): ...new pass_data instance and...
4462 (make_pass_ipa_profile): ...new function.
4463 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
4464 ipa_opt_pass_d along with...
4465 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4466 (make_pass_ipa_cdtor_merge): ...new function.
4467 * ira.c (pass_ira): Convert from a global struct to a subclass of
4468 rtl_opt_pass along with...
4469 (pass_data_ira): ...new pass_data instance and...
4470 (make_pass_ira): ...new function.
4471 (pass_reload): Convert from a global struct to a subclass of
4472 rtl_opt_pass along with...
4473 (pass_data_reload): ...new pass_data instance and...
4474 (make_pass_reload): ...new function.
4475 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4476 subclass of rtl_opt_pass along with...
4477 (pass_data_cleanup_barriers): ...new pass_data instance and...
4478 (make_pass_cleanup_barriers): ...new function.
4479 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4480 of rtl_opt_pass along with...
4481 (pass_data_loop2): ...new pass_data instance and...
4482 (make_pass_loop2): ...new function.
4483 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4484 rtl_opt_pass along with...
4485 (pass_data_rtl_loop_init): ...new pass_data instance and...
4486 (make_pass_rtl_loop_init): ...new function.
4487 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4488 rtl_opt_pass along with...
4489 (pass_data_rtl_loop_done): ...new pass_data instance and...
4490 (make_pass_rtl_loop_done): ...new function.
4491 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4492 subclass of rtl_opt_pass along with...
4493 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4494 (make_pass_rtl_move_loop_invariants): ...new function.
4495 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4496 rtl_opt_pass along with...
4497 (pass_data_rtl_unswitch): ...new pass_data instance and...
4498 (make_pass_rtl_unswitch): ...new function.
4499 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4500 subclass of rtl_opt_pass along with...
4501 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4503 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4504 (pass_rtl_doloop): Convert from a global struct to a subclass of
4505 rtl_opt_pass along with...
4506 (pass_data_rtl_doloop): ...new pass_data instance and...
4507 (make_pass_rtl_doloop): ...new function.
4508 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4509 a subclass of rtl_opt_pass along with...
4510 (pass_data_lower_subreg): ...new pass_data instance and...
4511 (make_pass_lower_subreg): ...new function.
4512 (pass_lower_subreg2): Convert from a global struct to a subclass of
4513 rtl_opt_pass along with...
4514 (pass_data_lower_subreg2): ...new pass_data instance and...
4515 (make_pass_lower_subreg2): ...new function.
4516 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4517 struct to a subclass of ipa_opt_pass_d along with...
4518 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4519 (make_pass_ipa_lto_gimple_out): ...new function.
4520 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4521 of ipa_opt_pass_d along with...
4522 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4523 (make_pass_ipa_lto_finish_out): ...new function.
4524 * mode-switching.c (pass_mode_switching): Convert from a global struct
4525 to a subclass of rtl_opt_pass along with...
4526 (pass_data_mode_switching): ...new pass_data instance and...
4527 (make_pass_mode_switching): ...new function.
4528 * modulo-sched.c (pass_sms): Convert from a global struct to a
4529 subclass of rtl_opt_pass along with...
4530 (pass_data_sms): ...new pass_data instance and...
4531 (make_pass_sms): ...new function.
4532 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4533 subclass of gimple_opt_pass along with...
4534 (pass_data_expand_omp): ...new pass_data instance and...
4535 (make_pass_expand_omp): ...new function.
4536 (pass_lower_omp): Convert from a global struct to a subclass of
4537 gimple_opt_pass along with...
4538 (pass_data_lower_omp): ...new pass_data instance and...
4539 (make_pass_lower_omp): ...new function.
4540 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4541 of gimple_opt_pass along with...
4542 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4543 (make_pass_diagnose_omp_blocks): ...new function.
4544 * passes.c (pass_early_local_passes): Convert from a global struct to
4545 a subclass of simple_ipa_opt_pass along with...
4546 (pass_data_early_local_passes): ...new pass_data instance and...
4547 (make_pass_early_local_passes): ...new function.
4548 (pass_all_early_optimizations): Convert from a global struct to a
4549 subclass of gimple_opt_pass along with...
4550 (pass_data_all_early_optimizations): ...new pass_data instance and...
4551 (make_pass_all_early_optimizations): ...new function.
4552 (pass_all_optimizations): Convert from a global struct to a subclass
4553 of gimple_opt_pass along with...
4554 (pass_data_all_optimizations): ...new pass_data instance and...
4555 (make_pass_all_optimizations): ...new function.
4556 (pass_all_optimizations_g): Convert from a global struct to a subclass
4557 of gimple_opt_pass along with...
4558 (pass_data_all_optimizations_g): ...new pass_data instance and...
4559 (make_pass_all_optimizations_g): ...new function.
4560 (pass_rest_of_compilation): Convert from a global struct to a subclass
4561 of rtl_opt_pass along with...
4562 (pass_data_rest_of_compilation): ...new pass_data instance and...
4563 (make_pass_rest_of_compilation): ...new function.
4564 (pass_postreload): Convert from a global struct to a subclass of
4565 rtl_opt_pass along with...
4566 (pass_data_postreload): ...new pass_data instance and...
4567 (make_pass_postreload): ...new function.
4568 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
4569 subclass of rtl_opt_pass along with...
4570 (pass_data_gcse2): ...new pass_data instance and...
4571 (make_pass_gcse2): ...new function.
4572 * postreload.c (pass_postreload_cse): Convert from a global struct to
4573 a subclass of rtl_opt_pass along with...
4574 (pass_data_postreload_cse): ...new pass_data instance and...
4575 (make_pass_postreload_cse): ...new function.
4576 * predict.c (pass_profile): Convert from a global struct to a subclass
4577 of gimple_opt_pass along with...
4578 (pass_data_profile): ...new pass_data instance and...
4579 (make_pass_profile): ...new function.
4580 (pass_strip_predict_hints): Convert from a global struct to a subclass
4581 of gimple_opt_pass along with...
4582 (pass_data_strip_predict_hints): ...new pass_data instance and...
4583 (make_pass_strip_predict_hints): ...new function.
4584 * recog.c (pass_peephole2): Convert from a global struct to a subclass
4585 of rtl_opt_pass along with...
4586 (pass_data_peephole2): ...new pass_data instance and...
4587 (make_pass_peephole2): ...new function.
4588 (pass_split_all_insns): Convert from a global struct to a subclass of
4589 rtl_opt_pass along with...
4590 (pass_data_split_all_insns): ...new pass_data instance and...
4591 (make_pass_split_all_insns): ...new function.
4592 (pass_split_after_reload): Convert from a global struct to a subclass
4593 of rtl_opt_pass along with...
4594 (pass_data_split_after_reload): ...new pass_data instance and...
4595 (make_pass_split_after_reload): ...new function.
4596 (pass_split_before_regstack): Convert from a global struct to a
4597 subclass of rtl_opt_pass along with...
4598 (pass_data_split_before_regstack): ...new pass_data instance and...
4599 (make_pass_split_before_regstack): ...new function.
4600 (pass_split_before_sched2): Convert from a global struct to a subclass
4601 of rtl_opt_pass along with...
4602 (pass_data_split_before_sched2): ...new pass_data instance and...
4603 (make_pass_split_before_sched2): ...new function.
4604 (pass_split_for_shorten_branches): Convert from a global struct to a
4605 subclass of rtl_opt_pass along with...
4606 (pass_data_split_for_shorten_branches): ...new pass_data instance
4608 (make_pass_split_for_shorten_branches): ...new function.
4609 * ree.c (pass_ree): Convert from a global struct to a subclass of
4610 rtl_opt_pass along with...
4611 (pass_data_ree): ...new pass_data instance and...
4612 (make_pass_ree): ...new function.
4613 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
4614 subclass of rtl_opt_pass along with...
4615 (pass_data_stack_regs): ...new pass_data instance and...
4616 (make_pass_stack_regs): ...new function.
4617 (pass_stack_regs_run): Convert from a global struct to a subclass of
4618 rtl_opt_pass along with...
4619 (pass_data_stack_regs_run): ...new pass_data instance and...
4620 (make_pass_stack_regs_run): ...new function.
4621 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4622 subclass of rtl_opt_pass along with...
4623 (pass_data_cprop_hardreg): ...new pass_data instance and...
4624 (make_pass_cprop_hardreg): ...new function.
4625 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4626 subclass of rtl_opt_pass along with...
4627 (pass_data_reginfo_init): ...new pass_data instance and...
4628 (make_pass_reginfo_init): ...new function.
4629 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4630 of rtl_opt_pass along with...
4631 (pass_data_regmove): ...new pass_data instance and...
4632 (make_pass_regmove): ...new function.
4633 * regrename.c (pass_regrename): Convert from a global struct to a
4634 subclass of rtl_opt_pass along with...
4635 (pass_data_regrename): ...new pass_data instance and...
4636 (make_pass_regrename): ...new function.
4637 * reorg.c (pass_delay_slots): Convert from a global struct to a
4638 subclass of rtl_opt_pass along with...
4639 (pass_data_delay_slots): ...new pass_data instance and...
4640 (make_pass_delay_slots): ...new function.
4641 (pass_machine_reorg): Convert from a global struct to a subclass of
4642 rtl_opt_pass along with...
4643 (pass_data_machine_reorg): ...new pass_data instance and...
4644 (make_pass_machine_reorg): ...new function.
4645 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4646 of rtl_opt_pass along with...
4647 (pass_data_sched): ...new pass_data instance and...
4648 (make_pass_sched): ...new function.
4649 (pass_sched2): Convert from a global struct to a subclass of
4650 rtl_opt_pass along with...
4651 (pass_data_sched2): ...new pass_data instance and...
4652 (make_pass_sched2): ...new function.
4653 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4654 to a subclass of rtl_opt_pass along with...
4655 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4656 (make_pass_stack_ptr_mod): ...new function.
4657 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4658 to a subclass of rtl_opt_pass along with...
4659 (pass_data_rtl_store_motion): ...new pass_data instance and...
4660 (make_pass_rtl_store_motion): ...new function.
4661 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4662 of gimple_opt_pass along with...
4663 (pass_data_tracer): ...new pass_data instance and...
4664 (make_pass_tracer): ...new function.
4665 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4666 to a subclass of gimple_opt_pass along with...
4667 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4668 (make_pass_diagnose_tm_blocks): ...new function.
4669 (pass_lower_tm): Convert from a global struct to a subclass of
4670 gimple_opt_pass along with...
4671 (pass_data_lower_tm): ...new pass_data instance and...
4672 (make_pass_lower_tm): ...new function.
4673 (pass_tm_init): Convert from a global struct to a subclass of
4674 gimple_opt_pass along with...
4675 (pass_data_tm_init): ...new pass_data instance and...
4676 (make_pass_tm_init): ...new function.
4677 (pass_tm_mark): Convert from a global struct to a subclass of
4678 gimple_opt_pass along with...
4679 (pass_data_tm_mark): ...new pass_data instance and...
4680 (make_pass_tm_mark): ...new function.
4681 (pass_tm_edges): Convert from a global struct to a subclass of
4682 gimple_opt_pass along with...
4683 (pass_data_tm_edges): ...new pass_data instance and...
4684 (make_pass_tm_edges): ...new function.
4685 (pass_tm_memopt): Convert from a global struct to a subclass of
4686 gimple_opt_pass along with...
4687 (pass_data_tm_memopt): ...new pass_data instance and...
4688 (make_pass_tm_memopt): ...new function.
4689 (pass_ipa_tm): Convert from a global struct to a subclass of
4690 simple_ipa_opt_pass along with...
4691 (pass_data_ipa_tm): ...new pass_data instance and...
4692 (make_pass_ipa_tm): ...new function.
4693 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4694 subclass of gimple_opt_pass along with...
4695 (pass_data_call_cdce): ...new pass_data instance and...
4696 (make_pass_call_cdce): ...new function.
4697 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4698 subclass of gimple_opt_pass along with...
4699 (pass_data_build_cfg): ...new pass_data instance and...
4700 (make_pass_build_cfg): ...new function.
4701 (pass_split_crit_edges): Convert from a global struct to a subclass of
4702 gimple_opt_pass along with...
4703 (pass_data_split_crit_edges): ...new pass_data instance and...
4704 (make_pass_split_crit_edges): ...new function.
4705 (pass_warn_function_return): Convert from a global struct to a
4706 subclass of gimple_opt_pass along with...
4707 (pass_data_warn_function_return): ...new pass_data instance and...
4708 (make_pass_warn_function_return): ...new function.
4709 (pass_warn_function_noreturn): Convert from a global struct to a
4710 subclass of gimple_opt_pass along with...
4711 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4712 (make_pass_warn_function_noreturn): ...new function.
4713 (pass_warn_unused_result): Convert from a global struct to a subclass
4714 of gimple_opt_pass along with...
4715 (pass_data_warn_unused_result): ...new pass_data instance and...
4716 (make_pass_warn_unused_result): ...new function.
4717 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4718 a subclass of gimple_opt_pass along with...
4719 (pass_data_merge_phi): ...new pass_data instance and...
4720 (make_pass_merge_phi): ...new function.
4721 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4722 a subclass of gimple_opt_pass along with...
4723 (pass_data_lower_complex): ...new pass_data instance and...
4724 (make_pass_lower_complex): ...new function.
4725 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4726 gimple_opt_pass along with...
4727 (pass_data_lower_complex_O0): ...new pass_data instance and...
4728 (make_pass_lower_complex_O0): ...new function.
4729 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4730 subclass of gimple_opt_pass along with...
4731 (pass_data_lower_eh): ...new pass_data instance and...
4732 (make_pass_lower_eh): ...new function.
4733 (pass_refactor_eh): Convert from a global struct to a subclass of
4734 gimple_opt_pass along with...
4735 (pass_data_refactor_eh): ...new pass_data instance and...
4736 (make_pass_refactor_eh): ...new function.
4737 (pass_lower_resx): Convert from a global struct to a subclass of
4738 gimple_opt_pass along with...
4739 (pass_data_lower_resx): ...new pass_data instance and...
4740 (make_pass_lower_resx): ...new function.
4741 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4742 of gimple_opt_pass along with...
4743 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4744 (make_pass_lower_eh_dispatch): ...new function.
4745 (pass_cleanup_eh): Convert from a global struct to a subclass of
4746 gimple_opt_pass along with...
4747 (pass_data_cleanup_eh): ...new pass_data instance and...
4748 (make_pass_cleanup_eh): ...new function.
4749 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4750 to a subclass of simple_ipa_opt_pass along with...
4751 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4752 (make_pass_ipa_lower_emutls): ...new function.
4753 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4754 a subclass of gimple_opt_pass along with...
4755 (pass_data_if_conversion): ...new pass_data instance and...
4756 (make_pass_if_conversion): ...new function.
4757 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4758 subclass of gimple_opt_pass along with...
4759 (pass_data_build_ssa): ...new pass_data instance and...
4760 (make_pass_build_ssa): ...new function.
4761 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4762 global struct to a subclass of gimple_opt_pass along with...
4763 (pass_data_loop_distribution): ...new pass_data instance and...
4764 (make_pass_loop_distribution): ...new function.
4765 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4766 subclass of gimple_opt_pass along with...
4767 (pass_data_mudflap_1): ...new pass_data instance and...
4768 (make_pass_mudflap_1): ...new function.
4769 (pass_mudflap_2): Convert from a global struct to a subclass of
4770 gimple_opt_pass along with...
4771 (pass_data_mudflap_2): ...new pass_data instance and...
4772 (make_pass_mudflap_2): ...new function.
4773 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4774 subclass of gimple_opt_pass along with...
4775 (pass_data_mudflap_1): ...new pass_data instance and...
4776 (make_pass_mudflap_1): ...new function.
4777 (pass_mudflap_2): Convert from a global struct to a subclass of
4778 gimple_opt_pass along with...
4779 (pass_data_mudflap_2): ...new pass_data instance and...
4780 (make_pass_mudflap_2): ...new function.
4781 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4782 gimple_opt_pass along with...
4783 (pass_data_nrv): ...new pass_data instance and...
4784 (make_pass_nrv): ...new function.
4785 (pass_return_slot): Convert from a global struct to a subclass of
4786 gimple_opt_pass along with...
4787 (pass_data_return_slot): ...new pass_data instance and...
4788 (make_pass_return_slot): ...new function.
4789 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4790 to a subclass of gimple_opt_pass along with...
4791 (pass_data_object_sizes): ...new pass_data instance and...
4792 (make_pass_object_sizes): ...new function.
4793 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4794 global struct to a subclass of gimple_opt_pass along with...
4795 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4797 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4798 (pass_fixup_cfg): Convert from a global struct to a subclass of
4799 gimple_opt_pass along with...
4800 (pass_data_fixup_cfg): ...new pass_data instance and...
4801 (make_pass_fixup_cfg): ...new function.
4802 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4803 (make_pass_mudflap_1): ...new function.
4804 (pass_mudflap_2): Replace declaration with that of...
4805 (make_pass_mudflap_2): ...new function.
4806 (pass_asan): Replace declaration with that of...
4807 (make_pass_asan): ...new function.
4808 (pass_asan_O0): Replace declaration with that of...
4809 (make_pass_asan_O0): ...new function.
4810 (pass_tsan): Replace declaration with that of...
4811 (make_pass_tsan): ...new function.
4812 (pass_tsan_O0): Replace declaration with that of...
4813 (make_pass_tsan_O0): ...new function.
4814 (pass_lower_cf): Replace declaration with that of...
4815 (make_pass_lower_cf): ...new function.
4816 (pass_refactor_eh): Replace declaration with that of...
4817 (make_pass_refactor_eh): ...new function.
4818 (pass_lower_eh): Replace declaration with that of...
4819 (make_pass_lower_eh): ...new function.
4820 (pass_lower_eh_dispatch): Replace declaration with that of...
4821 (make_pass_lower_eh_dispatch): ...new function.
4822 (pass_lower_resx): Replace declaration with that of...
4823 (make_pass_lower_resx): ...new function.
4824 (pass_build_cfg): Replace declaration with that of...
4825 (make_pass_build_cfg): ...new function.
4826 (pass_early_tree_profile): Replace declaration with that of...
4827 (make_pass_early_tree_profile): ...new function.
4828 (pass_cleanup_eh): Replace declaration with that of...
4829 (make_pass_cleanup_eh): ...new function.
4830 (pass_sra): Replace declaration with that of...
4831 (make_pass_sra): ...new function.
4832 (pass_sra_early): Replace declaration with that of...
4833 (make_pass_sra_early): ...new function.
4834 (pass_early_ipa_sra): Replace declaration with that of...
4835 (make_pass_early_ipa_sra): ...new function.
4836 (pass_tail_recursion): Replace declaration with that of...
4837 (make_pass_tail_recursion): ...new function.
4838 (pass_tail_calls): Replace declaration with that of...
4839 (make_pass_tail_calls): ...new function.
4840 (pass_tree_loop): Replace declaration with that of...
4841 (make_pass_tree_loop): ...new function.
4842 (pass_tree_loop_init): Replace declaration with that of...
4843 (make_pass_tree_loop_init): ...new function.
4844 (pass_lim): Replace declaration with that of...
4845 (make_pass_lim): ...new function.
4846 (pass_tree_unswitch): Replace declaration with that of...
4847 (make_pass_tree_unswitch): ...new function.
4848 (pass_predcom): Replace declaration with that of...
4849 (make_pass_predcom): ...new function.
4850 (pass_iv_canon): Replace declaration with that of...
4851 (make_pass_iv_canon): ...new function.
4852 (pass_scev_cprop): Replace declaration with that of...
4853 (make_pass_scev_cprop): ...new function.
4854 (pass_empty_loop): Replace declaration with that of...
4855 (make_pass_empty_loop): ...new function.
4856 (pass_record_bounds): Replace declaration with that of...
4857 (make_pass_record_bounds): ...new function.
4858 (pass_graphite): Replace declaration with that of...
4859 (make_pass_graphite): ...new function.
4860 (pass_graphite_transforms): Replace declaration with that of...
4861 (make_pass_graphite_transforms): ...new function.
4862 (pass_if_conversion): Replace declaration with that of...
4863 (make_pass_if_conversion): ...new function.
4864 (pass_loop_distribution): Replace declaration with that of...
4865 (make_pass_loop_distribution): ...new function.
4866 (pass_vectorize): Replace declaration with that of...
4867 (make_pass_vectorize): ...new function.
4868 (pass_slp_vectorize): Replace declaration with that of...
4869 (make_pass_slp_vectorize): ...new function.
4870 (pass_complete_unroll): Replace declaration with that of...
4871 (make_pass_complete_unroll): ...new function.
4872 (pass_complete_unrolli): Replace declaration with that of...
4873 (make_pass_complete_unrolli): ...new function.
4874 (pass_parallelize_loops): Replace declaration with that of...
4875 (make_pass_parallelize_loops): ...new function.
4876 (pass_loop_prefetch): Replace declaration with that of...
4877 (make_pass_loop_prefetch): ...new function.
4878 (pass_iv_optimize): Replace declaration with that of...
4879 (make_pass_iv_optimize): ...new function.
4880 (pass_tree_loop_done): Replace declaration with that of...
4881 (make_pass_tree_loop_done): ...new function.
4882 (pass_ch): Replace declaration with that of...
4883 (make_pass_ch): ...new function.
4884 (pass_ccp): Replace declaration with that of...
4885 (make_pass_ccp): ...new function.
4886 (pass_phi_only_cprop): Replace declaration with that of...
4887 (make_pass_phi_only_cprop): ...new function.
4888 (pass_build_ssa): Replace declaration with that of...
4889 (make_pass_build_ssa): ...new function.
4890 (pass_build_alias): Replace declaration with that of...
4891 (make_pass_build_alias): ...new function.
4892 (pass_build_ealias): Replace declaration with that of...
4893 (make_pass_build_ealias): ...new function.
4894 (pass_dominator): Replace declaration with that of...
4895 (make_pass_dominator): ...new function.
4896 (pass_dce): Replace declaration with that of...
4897 (make_pass_dce): ...new function.
4898 (pass_dce_loop): Replace declaration with that of...
4899 (make_pass_dce_loop): ...new function.
4900 (pass_cd_dce): Replace declaration with that of...
4901 (make_pass_cd_dce): ...new function.
4902 (pass_call_cdce): Replace declaration with that of...
4903 (make_pass_call_cdce): ...new function.
4904 (pass_merge_phi): Replace declaration with that of...
4905 (make_pass_merge_phi): ...new function.
4906 (pass_split_crit_edges): Replace declaration with that of...
4907 (make_pass_split_crit_edges): ...new function.
4908 (pass_pre): Replace declaration with that of...
4909 (make_pass_pre): ...new function.
4910 (pass_profile): Replace declaration with that of...
4911 (make_pass_profile): ...new function.
4912 (pass_strip_predict_hints): Replace declaration with that of...
4913 (make_pass_strip_predict_hints): ...new function.
4914 (pass_lower_complex_O0): Replace declaration with that of...
4915 (make_pass_lower_complex_O0): ...new function.
4916 (pass_lower_complex): Replace declaration with that of...
4917 (make_pass_lower_complex): ...new function.
4918 (pass_lower_vector): Replace declaration with that of...
4919 (make_pass_lower_vector): ...new function.
4920 (pass_lower_vector_ssa): Replace declaration with that of...
4921 (make_pass_lower_vector_ssa): ...new function.
4922 (pass_lower_omp): Replace declaration with that of...
4923 (make_pass_lower_omp): ...new function.
4924 (pass_diagnose_omp_blocks): Replace declaration with that of...
4925 (make_pass_diagnose_omp_blocks): ...new function.
4926 (pass_expand_omp): Replace declaration with that of...
4927 (make_pass_expand_omp): ...new function.
4928 (pass_expand_omp_ssa): Replace declaration with that of...
4929 (make_pass_expand_omp_ssa): ...new function.
4930 (pass_object_sizes): Replace declaration with that of...
4931 (make_pass_object_sizes): ...new function.
4932 (pass_strlen): Replace declaration with that of...
4933 (make_pass_strlen): ...new function.
4934 (pass_fold_builtins): Replace declaration with that of...
4935 (make_pass_fold_builtins): ...new function.
4936 (pass_stdarg): Replace declaration with that of...
4937 (make_pass_stdarg): ...new function.
4938 (pass_early_warn_uninitialized): Replace declaration with that of...
4939 (make_pass_early_warn_uninitialized): ...new function.
4940 (pass_late_warn_uninitialized): Replace declaration with that of...
4941 (make_pass_late_warn_uninitialized): ...new function.
4942 (pass_cse_reciprocals): Replace declaration with that of...
4943 (make_pass_cse_reciprocals): ...new function.
4944 (pass_cse_sincos): Replace declaration with that of...
4945 (make_pass_cse_sincos): ...new function.
4946 (pass_optimize_bswap): Replace declaration with that of...
4947 (make_pass_optimize_bswap): ...new function.
4948 (pass_optimize_widening_mul): Replace declaration with that of...
4949 (make_pass_optimize_widening_mul): ...new function.
4950 (pass_warn_function_return): Replace declaration with that of...
4951 (make_pass_warn_function_return): ...new function.
4952 (pass_warn_function_noreturn): Replace declaration with that of...
4953 (make_pass_warn_function_noreturn): ...new function.
4954 (pass_cselim): Replace declaration with that of...
4955 (make_pass_cselim): ...new function.
4956 (pass_phiopt): Replace declaration with that of...
4957 (make_pass_phiopt): ...new function.
4958 (pass_forwprop): Replace declaration with that of...
4959 (make_pass_forwprop): ...new function.
4960 (pass_phiprop): Replace declaration with that of...
4961 (make_pass_phiprop): ...new function.
4962 (pass_tree_ifcombine): Replace declaration with that of...
4963 (make_pass_tree_ifcombine): ...new function.
4964 (pass_dse): Replace declaration with that of...
4965 (make_pass_dse): ...new function.
4966 (pass_nrv): Replace declaration with that of...
4967 (make_pass_nrv): ...new function.
4968 (pass_rename_ssa_copies): Replace declaration with that of...
4969 (make_pass_rename_ssa_copies): ...new function.
4970 (pass_sink_code): Replace declaration with that of...
4971 (make_pass_sink_code): ...new function.
4972 (pass_fre): Replace declaration with that of...
4973 (make_pass_fre): ...new function.
4974 (pass_check_data_deps): Replace declaration with that of...
4975 (make_pass_check_data_deps): ...new function.
4976 (pass_copy_prop): Replace declaration with that of...
4977 (make_pass_copy_prop): ...new function.
4978 (pass_vrp): Replace declaration with that of...
4979 (make_pass_vrp): ...new function.
4980 (pass_uncprop): Replace declaration with that of...
4981 (make_pass_uncprop): ...new function.
4982 (pass_return_slot): Replace declaration with that of...
4983 (make_pass_return_slot): ...new function.
4984 (pass_reassoc): Replace declaration with that of...
4985 (make_pass_reassoc): ...new function.
4986 (pass_rebuild_cgraph_edges): Replace declaration with that of...
4987 (make_pass_rebuild_cgraph_edges): ...new function.
4988 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
4989 (make_pass_remove_cgraph_callee_edges): ...new function.
4990 (pass_build_cgraph_edges): Replace declaration with that of...
4991 (make_pass_build_cgraph_edges): ...new function.
4992 (pass_local_pure_const): Replace declaration with that of...
4993 (make_pass_local_pure_const): ...new function.
4994 (pass_tracer): Replace declaration with that of...
4995 (make_pass_tracer): ...new function.
4996 (pass_warn_unused_result): Replace declaration with that of...
4997 (make_pass_warn_unused_result): ...new function.
4998 (pass_diagnose_tm_blocks): Replace declaration with that of...
4999 (make_pass_diagnose_tm_blocks): ...new function.
5000 (pass_lower_tm): Replace declaration with that of...
5001 (make_pass_lower_tm): ...new function.
5002 (pass_tm_init): Replace declaration with that of...
5003 (make_pass_tm_init): ...new function.
5004 (pass_tm_mark): Replace declaration with that of...
5005 (make_pass_tm_mark): ...new function.
5006 (pass_tm_memopt): Replace declaration with that of...
5007 (make_pass_tm_memopt): ...new function.
5008 (pass_tm_edges): Replace declaration with that of...
5009 (make_pass_tm_edges): ...new function.
5010 (pass_split_functions): Replace declaration with that of...
5011 (make_pass_split_functions): ...new function.
5012 (pass_feedback_split_functions): Replace declaration with that of...
5013 (make_pass_feedback_split_functions): ...new function.
5014 (pass_strength_reduction): Replace declaration with that of...
5015 (make_pass_strength_reduction): ...new function.
5016 (pass_ipa_lower_emutls): Replace declaration with that of...
5017 (make_pass_ipa_lower_emutls): ...new function.
5018 (pass_ipa_function_and_variable_visibility): Replace declaration with
5020 (make_pass_ipa_function_and_variable_visibility): ...new function.
5021 (pass_ipa_tree_profile): Replace declaration with that of...
5022 (make_pass_ipa_tree_profile): ...new function.
5023 (pass_early_local_passes): Replace declaration with that of...
5024 (make_pass_early_local_passes): ...new function.
5025 (pass_ipa_whole_program_visibility): Replace declaration with that
5027 (make_pass_ipa_whole_program_visibility): ...new function.
5028 (pass_ipa_lto_gimple_out): Replace declaration with that of...
5029 (make_pass_ipa_lto_gimple_out): ...new function.
5030 (pass_ipa_increase_alignment): Replace declaration with that of...
5031 (make_pass_ipa_increase_alignment): ...new function.
5032 (pass_ipa_inline): Replace declaration with that of...
5033 (make_pass_ipa_inline): ...new function.
5034 (pass_ipa_free_lang_data): Replace declaration with that of...
5035 (make_pass_ipa_free_lang_data): ...new function.
5036 (pass_ipa_free_inline_summary): Replace declaration with that of...
5037 (make_pass_ipa_free_inline_summary): ...new function.
5038 (pass_ipa_cp): Replace declaration with that of...
5039 (make_pass_ipa_cp): ...new function.
5040 (pass_ipa_reference): Replace declaration with that of...
5041 (make_pass_ipa_reference): ...new function.
5042 (pass_ipa_pure_const): Replace declaration with that of...
5043 (make_pass_ipa_pure_const): ...new function.
5044 (pass_ipa_pta): Replace declaration with that of...
5045 (make_pass_ipa_pta): ...new function.
5046 (pass_ipa_lto_finish_out): Replace declaration with that of...
5047 (make_pass_ipa_lto_finish_out): ...new function.
5048 (pass_ipa_tm): Replace declaration with that of...
5049 (make_pass_ipa_tm): ...new function.
5050 (pass_ipa_profile): Replace declaration with that of...
5051 (make_pass_ipa_profile): ...new function.
5052 (pass_ipa_cdtor_merge): Replace declaration with that of...
5053 (make_pass_ipa_cdtor_merge): ...new function.
5054 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
5056 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5057 (pass_init_datastructures): Replace declaration with that of...
5058 (make_pass_init_datastructures): ...new function.
5059 (pass_fixup_cfg): Replace declaration with that of...
5060 (make_pass_fixup_cfg): ...new function.
5061 (pass_expand): Replace declaration with that of...
5062 (make_pass_expand): ...new function.
5063 (pass_instantiate_virtual_regs): Replace declaration with that of...
5064 (make_pass_instantiate_virtual_regs): ...new function.
5065 (pass_rtl_fwprop): Replace declaration with that of...
5066 (make_pass_rtl_fwprop): ...new function.
5067 (pass_rtl_fwprop_addr): Replace declaration with that of...
5068 (make_pass_rtl_fwprop_addr): ...new function.
5069 (pass_jump): Replace declaration with that of...
5070 (make_pass_jump): ...new function.
5071 (pass_jump2): Replace declaration with that of...
5072 (make_pass_jump2): ...new function.
5073 (pass_lower_subreg): Replace declaration with that of...
5074 (make_pass_lower_subreg): ...new function.
5075 (pass_cse): Replace declaration with that of...
5076 (make_pass_cse): ...new function.
5077 (pass_fast_rtl_dce): Replace declaration with that of...
5078 (make_pass_fast_rtl_dce): ...new function.
5079 (pass_ud_rtl_dce): Replace declaration with that of...
5080 (make_pass_ud_rtl_dce): ...new function.
5081 (pass_rtl_dce): Replace declaration with that of...
5082 (make_pass_rtl_dce): ...new function.
5083 (pass_rtl_dse1): Replace declaration with that of...
5084 (make_pass_rtl_dse1): ...new function.
5085 (pass_rtl_dse2): Replace declaration with that of...
5086 (make_pass_rtl_dse2): ...new function.
5087 (pass_rtl_dse3): Replace declaration with that of...
5088 (make_pass_rtl_dse3): ...new function.
5089 (pass_rtl_cprop): Replace declaration with that of...
5090 (make_pass_rtl_cprop): ...new function.
5091 (pass_rtl_pre): Replace declaration with that of...
5092 (make_pass_rtl_pre): ...new function.
5093 (pass_rtl_hoist): Replace declaration with that of...
5094 (make_pass_rtl_hoist): ...new function.
5095 (pass_rtl_store_motion): Replace declaration with that of...
5096 (make_pass_rtl_store_motion): ...new function.
5097 (pass_cse_after_global_opts): Replace declaration with that of...
5098 (make_pass_cse_after_global_opts): ...new function.
5099 (pass_rtl_ifcvt): Replace declaration with that of...
5100 (make_pass_rtl_ifcvt): ...new function.
5101 (pass_into_cfg_layout_mode): Replace declaration with that of...
5102 (make_pass_into_cfg_layout_mode): ...new function.
5103 (pass_outof_cfg_layout_mode): Replace declaration with that of...
5104 (make_pass_outof_cfg_layout_mode): ...new function.
5105 (pass_loop2): Replace declaration with that of...
5106 (make_pass_loop2): ...new function.
5107 (pass_rtl_loop_init): Replace declaration with that of...
5108 (make_pass_rtl_loop_init): ...new function.
5109 (pass_rtl_move_loop_invariants): Replace declaration with that of...
5110 (make_pass_rtl_move_loop_invariants): ...new function.
5111 (pass_rtl_unswitch): Replace declaration with that of...
5112 (make_pass_rtl_unswitch): ...new function.
5113 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
5114 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5115 (pass_rtl_doloop): Replace declaration with that of...
5116 (make_pass_rtl_doloop): ...new function.
5117 (pass_rtl_loop_done): Replace declaration with that of...
5118 (make_pass_rtl_loop_done): ...new function.
5119 (pass_web): Replace declaration with that of...
5120 (make_pass_web): ...new function.
5121 (pass_cse2): Replace declaration with that of...
5122 (make_pass_cse2): ...new function.
5123 (pass_df_initialize_opt): Replace declaration with that of...
5124 (make_pass_df_initialize_opt): ...new function.
5125 (pass_df_initialize_no_opt): Replace declaration with that of...
5126 (make_pass_df_initialize_no_opt): ...new function.
5127 (pass_reginfo_init): Replace declaration with that of...
5128 (make_pass_reginfo_init): ...new function.
5129 (pass_inc_dec): Replace declaration with that of...
5130 (make_pass_inc_dec): ...new function.
5131 (pass_stack_ptr_mod): Replace declaration with that of...
5132 (make_pass_stack_ptr_mod): ...new function.
5133 (pass_initialize_regs): Replace declaration with that of...
5134 (make_pass_initialize_regs): ...new function.
5135 (pass_combine): Replace declaration with that of...
5136 (make_pass_combine): ...new function.
5137 (pass_if_after_combine): Replace declaration with that of...
5138 (make_pass_if_after_combine): ...new function.
5139 (pass_ree): Replace declaration with that of...
5140 (make_pass_ree): ...new function.
5141 (pass_partition_blocks): Replace declaration with that of...
5142 (make_pass_partition_blocks): ...new function.
5143 (pass_match_asm_constraints): Replace declaration with that of...
5144 (make_pass_match_asm_constraints): ...new function.
5145 (pass_regmove): Replace declaration with that of...
5146 (make_pass_regmove): ...new function.
5147 (pass_split_all_insns): Replace declaration with that of...
5148 (make_pass_split_all_insns): ...new function.
5149 (pass_fast_rtl_byte_dce): Replace declaration with that of...
5150 (make_pass_fast_rtl_byte_dce): ...new function.
5151 (pass_lower_subreg2): Replace declaration with that of...
5152 (make_pass_lower_subreg2): ...new function.
5153 (pass_mode_switching): Replace declaration with that of...
5154 (make_pass_mode_switching): ...new function.
5155 (pass_sms): Replace declaration with that of...
5156 (make_pass_sms): ...new function.
5157 (pass_sched): Replace declaration with that of...
5158 (make_pass_sched): ...new function.
5159 (pass_ira): Replace declaration with that of...
5160 (make_pass_ira): ...new function.
5161 (pass_reload): Replace declaration with that of...
5162 (make_pass_reload): ...new function.
5163 (pass_clean_state): Replace declaration with that of...
5164 (make_pass_clean_state): ...new function.
5165 (pass_branch_prob): Replace declaration with that of...
5166 (make_pass_branch_prob): ...new function.
5167 (pass_value_profile_transformations): Replace declaration with that
5169 (make_pass_value_profile_transformations): ...new function.
5170 (pass_postreload_cse): Replace declaration with that of...
5171 (make_pass_postreload_cse): ...new function.
5172 (pass_gcse2): Replace declaration with that of...
5173 (make_pass_gcse2): ...new function.
5174 (pass_split_after_reload): Replace declaration with that of...
5175 (make_pass_split_after_reload): ...new function.
5176 (pass_branch_target_load_optimize1): Replace declaration with that
5178 (make_pass_branch_target_load_optimize1): ...new function.
5179 (pass_thread_prologue_and_epilogue): Replace declaration with that
5181 (make_pass_thread_prologue_and_epilogue): ...new function.
5182 (pass_stack_adjustments): Replace declaration with that of...
5183 (make_pass_stack_adjustments): ...new function.
5184 (pass_peephole2): Replace declaration with that of...
5185 (make_pass_peephole2): ...new function.
5186 (pass_if_after_reload): Replace declaration with that of...
5187 (make_pass_if_after_reload): ...new function.
5188 (pass_regrename): Replace declaration with that of...
5189 (make_pass_regrename): ...new function.
5190 (pass_cprop_hardreg): Replace declaration with that of...
5191 (make_pass_cprop_hardreg): ...new function.
5192 (pass_reorder_blocks): Replace declaration with that of...
5193 (make_pass_reorder_blocks): ...new function.
5194 (pass_branch_target_load_optimize2): Replace declaration with that
5196 (make_pass_branch_target_load_optimize2): ...new function.
5197 (pass_leaf_regs): Replace declaration with that of...
5198 (make_pass_leaf_regs): ...new function.
5199 (pass_split_before_sched2): Replace declaration with that of...
5200 (make_pass_split_before_sched2): ...new function.
5201 (pass_compare_elim_after_reload): Replace declaration with that of...
5202 (make_pass_compare_elim_after_reload): ...new function.
5203 (pass_sched2): Replace declaration with that of...
5204 (make_pass_sched2): ...new function.
5205 (pass_stack_regs): Replace declaration with that of...
5206 (make_pass_stack_regs): ...new function.
5207 (pass_stack_regs_run): Replace declaration with that of...
5208 (make_pass_stack_regs_run): ...new function.
5209 (pass_df_finish): Replace declaration with that of...
5210 (make_pass_df_finish): ...new function.
5211 (pass_compute_alignments): Replace declaration with that of...
5212 (make_pass_compute_alignments): ...new function.
5213 (pass_duplicate_computed_gotos): Replace declaration with that of...
5214 (make_pass_duplicate_computed_gotos): ...new function.
5215 (pass_variable_tracking): Replace declaration with that of...
5216 (make_pass_variable_tracking): ...new function.
5217 (pass_free_cfg): Replace declaration with that of...
5218 (make_pass_free_cfg): ...new function.
5219 (pass_machine_reorg): Replace declaration with that of...
5220 (make_pass_machine_reorg): ...new function.
5221 (pass_cleanup_barriers): Replace declaration with that of...
5222 (make_pass_cleanup_barriers): ...new function.
5223 (pass_delay_slots): Replace declaration with that of...
5224 (make_pass_delay_slots): ...new function.
5225 (pass_split_for_shorten_branches): Replace declaration with that of...
5226 (make_pass_split_for_shorten_branches): ...new function.
5227 (pass_split_before_regstack): Replace declaration with that of...
5228 (make_pass_split_before_regstack): ...new function.
5229 (pass_convert_to_eh_region_ranges): Replace declaration with that
5231 (make_pass_convert_to_eh_region_ranges): ...new function.
5232 (pass_shorten_branches): Replace declaration with that of...
5233 (make_pass_shorten_branches): ...new function.
5234 (pass_set_nothrow_function_flags): Replace declaration with that of...
5235 (make_pass_set_nothrow_function_flags): ...new function.
5236 (pass_dwarf2_frame): Replace declaration with that of...
5237 (make_pass_dwarf2_frame): ...new function.
5238 (pass_final): Replace declaration with that of...
5239 (make_pass_final): ...new function.
5240 (pass_rtl_seqabstr): Replace declaration with that of...
5241 (make_pass_rtl_seqabstr): ...new function.
5242 (pass_release_ssa_names): Replace declaration with that of...
5243 (make_pass_release_ssa_names): ...new function.
5244 (pass_early_inline): Replace declaration with that of...
5245 (make_pass_early_inline): ...new function.
5246 (pass_inline_parameters): Replace declaration with that of...
5247 (make_pass_inline_parameters): ...new function.
5248 (pass_update_address_taken): Replace declaration with that of...
5249 (make_pass_update_address_taken): ...new function.
5250 (pass_convert_switch): Replace declaration with that of...
5251 (make_pass_convert_switch): ...new function.
5252 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
5253 to a subclass of simple_ipa_opt_pass along with...
5254 (pass_data_ipa_tree_profile): ...new pass_data instance and...
5255 (make_pass_ipa_tree_profile): ...new function.
5256 * tree-sra.c (pass_sra_early): Convert from a global struct to a
5257 subclass of gimple_opt_pass along with...
5258 (pass_data_sra_early): ...new pass_data instance and...
5259 (make_pass_sra_early): ...new function.
5260 (pass_sra): Convert from a global struct to a subclass of
5261 gimple_opt_pass along with...
5262 (pass_data_sra): ...new pass_data instance and...
5263 (make_pass_sra): ...new function.
5264 (pass_early_ipa_sra): Convert from a global struct to a subclass of
5265 gimple_opt_pass along with...
5266 (pass_data_early_ipa_sra): ...new pass_data instance and...
5267 (make_pass_early_ipa_sra): ...new function.
5268 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
5269 subclass of gimple_opt_pass along with...
5270 (pass_data_ccp): ...new pass_data instance and...
5271 (make_pass_ccp): ...new function.
5272 (pass_fold_builtins): Convert from a global struct to a subclass of
5273 gimple_opt_pass along with...
5274 (pass_data_fold_builtins): ...new pass_data instance and...
5275 (make_pass_fold_builtins): ...new function.
5276 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
5277 subclass of gimple_opt_pass along with...
5278 (pass_data_copy_prop): ...new pass_data instance and...
5279 (make_pass_copy_prop): ...new function.
5280 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
5281 global struct to a subclass of gimple_opt_pass along with...
5282 (pass_data_rename_ssa_copies): ...new pass_data instance and...
5283 (make_pass_rename_ssa_copies): ...new function.
5284 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
5285 subclass of gimple_opt_pass along with...
5286 (pass_data_dce): ...new pass_data instance and...
5287 (make_pass_dce): ...new function.
5288 (pass_dce_loop): Convert from a global struct to a subclass of
5289 gimple_opt_pass along with...
5290 (pass_data_dce_loop): ...new pass_data instance and...
5291 (make_pass_dce_loop): ...new function.
5292 (pass_cd_dce): Convert from a global struct to a subclass of
5293 gimple_opt_pass along with...
5294 (pass_data_cd_dce): ...new pass_data instance and...
5295 (make_pass_cd_dce): ...new function.
5296 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
5297 subclass of gimple_opt_pass along with...
5298 (pass_data_dominator): ...new pass_data instance and...
5299 (make_pass_dominator): ...new function.
5300 (pass_phi_only_cprop): Convert from a global struct to a subclass of
5301 gimple_opt_pass along with...
5302 (pass_data_phi_only_cprop): ...new pass_data instance and...
5303 (make_pass_phi_only_cprop): ...new function.
5304 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
5305 subclass of gimple_opt_pass along with...
5306 (pass_data_dse): ...new pass_data instance and...
5307 (make_pass_dse): ...new function.
5308 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
5309 a subclass of gimple_opt_pass along with...
5310 (pass_data_forwprop): ...new pass_data instance and...
5311 (make_pass_forwprop): ...new function.
5312 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
5313 struct to a subclass of gimple_opt_pass along with...
5314 (pass_data_tree_ifcombine): ...new pass_data instance and...
5315 (make_pass_tree_ifcombine): ...new function.
5316 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
5317 subclass of gimple_opt_pass along with...
5318 (pass_data_ch): ...new pass_data instance and...
5319 (make_pass_ch): ...new function.
5320 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
5321 subclass of gimple_opt_pass along with...
5322 (pass_data_tree_loop): ...new pass_data instance and...
5323 (make_pass_tree_loop): ...new function.
5324 (pass_tree_loop_init): Convert from a global struct to a subclass of
5325 gimple_opt_pass along with...
5326 (pass_data_tree_loop_init): ...new pass_data instance and...
5327 (make_pass_tree_loop_init): ...new function.
5328 (pass_lim): Convert from a global struct to a subclass of
5329 gimple_opt_pass along with...
5330 (pass_data_lim): ...new pass_data instance and...
5331 (make_pass_lim): ...new function.
5332 (pass_tree_unswitch): Convert from a global struct to a subclass of
5333 gimple_opt_pass along with...
5334 (pass_data_tree_unswitch): ...new pass_data instance and...
5335 (make_pass_tree_unswitch): ...new function.
5336 (pass_predcom): Convert from a global struct to a subclass of
5337 gimple_opt_pass along with...
5338 (pass_data_predcom): ...new pass_data instance and...
5339 (make_pass_predcom): ...new function.
5340 (pass_vectorize): Convert from a global struct to a subclass of
5341 gimple_opt_pass along with...
5342 (pass_data_vectorize): ...new pass_data instance and...
5343 (make_pass_vectorize): ...new function.
5344 (pass_graphite): Convert from a global struct to a subclass of
5345 gimple_opt_pass along with...
5346 (pass_data_graphite): ...new pass_data instance and...
5347 (make_pass_graphite): ...new function.
5348 (pass_graphite_transforms): Convert from a global struct to a subclass
5349 of gimple_opt_pass along with...
5350 (pass_data_graphite_transforms): ...new pass_data instance and...
5351 (make_pass_graphite_transforms): ...new function.
5352 (pass_check_data_deps): Convert from a global struct to a subclass of
5353 gimple_opt_pass along with...
5354 (pass_data_check_data_deps): ...new pass_data instance and...
5355 (make_pass_check_data_deps): ...new function.
5356 (pass_iv_canon): Convert from a global struct to a subclass of
5357 gimple_opt_pass along with...
5358 (pass_data_iv_canon): ...new pass_data instance and...
5359 (make_pass_iv_canon): ...new function.
5360 (pass_scev_cprop): Convert from a global struct to a subclass of
5361 gimple_opt_pass along with...
5362 (pass_data_scev_cprop): ...new pass_data instance and...
5363 (make_pass_scev_cprop): ...new function.
5364 (pass_record_bounds): Convert from a global struct to a subclass of
5365 gimple_opt_pass along with...
5366 (pass_data_record_bounds): ...new pass_data instance and...
5367 (make_pass_record_bounds): ...new function.
5368 (pass_complete_unroll): Convert from a global struct to a subclass of
5369 gimple_opt_pass along with...
5370 (pass_data_complete_unroll): ...new pass_data instance and...
5371 (make_pass_complete_unroll): ...new function.
5372 (pass_complete_unrolli): Convert from a global struct to a subclass of
5373 gimple_opt_pass along with...
5374 (pass_data_complete_unrolli): ...new pass_data instance and...
5375 (make_pass_complete_unrolli): ...new function.
5376 (pass_parallelize_loops): Convert from a global struct to a subclass
5377 of gimple_opt_pass along with...
5378 (pass_data_parallelize_loops): ...new pass_data instance and...
5379 (make_pass_parallelize_loops): ...new function.
5380 (pass_loop_prefetch): Convert from a global struct to a subclass of
5381 gimple_opt_pass along with...
5382 (pass_data_loop_prefetch): ...new pass_data instance and...
5383 (make_pass_loop_prefetch): ...new function.
5384 (pass_iv_optimize): Convert from a global struct to a subclass of
5385 gimple_opt_pass along with...
5386 (pass_data_iv_optimize): ...new pass_data instance and...
5387 (make_pass_iv_optimize): ...new function.
5388 (pass_tree_loop_done): Convert from a global struct to a subclass of
5389 gimple_opt_pass along with...
5390 (pass_data_tree_loop_done): ...new pass_data instance and...
5391 (make_pass_tree_loop_done): ...new function.
5392 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
5393 struct to a subclass of gimple_opt_pass along with...
5394 (pass_data_cse_reciprocals): ...new pass_data instance and...
5395 (make_pass_cse_reciprocals): ...new function.
5396 (pass_cse_sincos): Convert from a global struct to a subclass of
5397 gimple_opt_pass along with...
5398 (pass_data_cse_sincos): ...new pass_data instance and...
5399 (make_pass_cse_sincos): ...new function.
5400 (pass_optimize_bswap): Convert from a global struct to a subclass of
5401 gimple_opt_pass along with...
5402 (pass_data_optimize_bswap): ...new pass_data instance and...
5403 (make_pass_optimize_bswap): ...new function.
5404 (pass_optimize_widening_mul): Convert from a global struct to a
5405 subclass of gimple_opt_pass along with...
5406 (pass_data_optimize_widening_mul): ...new pass_data instance and...
5407 (make_pass_optimize_widening_mul): ...new function.
5408 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
5409 subclass of gimple_opt_pass along with...
5410 (pass_data_phiopt): ...new pass_data instance and...
5411 (make_pass_phiopt): ...new function.
5412 (pass_cselim): Convert from a global struct to a subclass of
5413 gimple_opt_pass along with...
5414 (pass_data_cselim): ...new pass_data instance and...
5415 (make_pass_cselim): ...new function.
5416 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
5417 subclass of gimple_opt_pass along with...
5418 (pass_data_phiprop): ...new pass_data instance and...
5419 (make_pass_phiprop): ...new function.
5420 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
5421 subclass of gimple_opt_pass along with...
5422 (pass_data_pre): ...new pass_data instance and...
5423 (make_pass_pre): ...new function.
5424 (pass_fre): Convert from a global struct to a subclass of
5425 gimple_opt_pass along with...
5426 (pass_data_fre): ...new pass_data instance and...
5427 (make_pass_fre): ...new function.
5428 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
5429 subclass of gimple_opt_pass along with...
5430 (pass_data_reassoc): ...new pass_data instance and...
5431 (make_pass_reassoc): ...new function.
5432 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
5433 subclass of gimple_opt_pass along with...
5434 (pass_data_sink_code): ...new pass_data instance and...
5435 (make_pass_sink_code): ...new function.
5436 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
5437 subclass of gimple_opt_pass along with...
5438 (pass_data_strlen): ...new pass_data instance and...
5439 (make_pass_strlen): ...new function.
5440 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
5441 struct to a subclass of gimple_opt_pass along with...
5442 (pass_data_build_alias): ...new pass_data instance and...
5443 (make_pass_build_alias): ...new function.
5444 (pass_build_ealias): Convert from a global struct to a subclass of
5445 gimple_opt_pass along with...
5446 (pass_data_build_ealias): ...new pass_data instance and...
5447 (make_pass_build_ealias): ...new function.
5448 (pass_ipa_pta): Convert from a global struct to a subclass of
5449 simple_ipa_opt_pass along with...
5450 (pass_data_ipa_pta): ...new pass_data instance and...
5451 (make_pass_ipa_pta): ...new function.
5452 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
5453 subclass of gimple_opt_pass along with...
5454 (pass_data_uncprop): ...new pass_data instance and...
5455 (make_pass_uncprop): ...new function.
5456 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
5457 global struct to a subclass of gimple_opt_pass along with...
5458 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
5459 (make_pass_late_warn_uninitialized): ...new function.
5460 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
5461 to a subclass of gimple_opt_pass along with...
5462 (pass_data_init_datastructures): ...new pass_data instance and...
5463 (make_pass_init_datastructures): ...new function.
5464 (pass_early_warn_uninitialized): Convert from a global struct to a
5465 subclass of gimple_opt_pass along with...
5466 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5467 (make_pass_early_warn_uninitialized): ...new function.
5468 (pass_update_address_taken): Convert from a global struct to a
5469 subclass of gimple_opt_pass along with...
5470 (pass_data_update_address_taken): ...new pass_data instance and...
5471 (make_pass_update_address_taken): ...new function.
5472 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5473 struct to a subclass of gimple_opt_pass along with...
5474 (pass_data_release_ssa_names): ...new pass_data instance and...
5475 (make_pass_release_ssa_names): ...new function.
5476 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5477 subclass of gimple_opt_pass along with...
5478 (pass_data_stdarg): ...new pass_data instance and...
5479 (make_pass_stdarg): ...new function.
5480 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5481 global struct to a subclass of gimple_opt_pass along with...
5482 (pass_data_convert_switch): ...new pass_data instance and...
5483 (make_pass_convert_switch): ...new function.
5484 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5485 to a subclass of gimple_opt_pass along with...
5486 (pass_data_tail_recursion): ...new pass_data instance and...
5487 (make_pass_tail_recursion): ...new function.
5488 (pass_tail_calls): Convert from a global struct to a subclass of
5489 gimple_opt_pass along with...
5490 (pass_data_tail_calls): ...new pass_data instance and...
5491 (make_pass_tail_calls): ...new function.
5492 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5493 struct to a subclass of gimple_opt_pass along with...
5494 (pass_data_lower_vector): ...new pass_data instance and...
5495 (make_pass_lower_vector): ...new function.
5496 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5497 gimple_opt_pass along with...
5498 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5499 (make_pass_lower_vector_ssa): ...new function.
5500 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5501 to a subclass of gimple_opt_pass along with...
5502 (pass_data_slp_vectorize): ...new pass_data instance and...
5503 (make_pass_slp_vectorize): ...new function.
5504 (pass_ipa_increase_alignment): Convert from a global struct to a
5505 subclass of simple_ipa_opt_pass along with...
5506 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5507 (make_pass_ipa_increase_alignment): ...new function.
5508 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5509 gimple_opt_pass along with...
5510 (pass_data_vrp): ...new pass_data instance and...
5511 (make_pass_vrp): ...new function.
5512 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5513 subclass of simple_ipa_opt_pass along with...
5514 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5515 (make_pass_ipa_free_lang_data): ...new function.
5516 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5517 gimple_opt_pass along with...
5518 (pass_data_tsan): ...new pass_data instance and...
5519 (make_pass_tsan): ...new function.
5520 (pass_tsan_O0): Convert from a global struct to a subclass of
5521 gimple_opt_pass along with...
5522 (pass_data_tsan_O0): ...new pass_data instance and...
5523 (make_pass_tsan_O0): ...new function.
5524 * var-tracking.c (pass_variable_tracking): Convert from a global
5525 struct to a subclass of rtl_opt_pass along with...
5526 (pass_data_variable_tracking): ...new pass_data instance and...
5527 (make_pass_variable_tracking): ...new function.
5528 * web.c (pass_web): Convert from a global struct to a subclass of
5529 rtl_opt_pass along with...
5530 (pass_data_web): ...new pass_data instance and...
5531 (make_pass_web): ...new function.
5532 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5533 declaration with that of...
5534 (make_pass_mode_switch_use): ...new function.
5535 (pass_resolve_sw_modes): Replace declaration with that of...
5536 (make_pass_resolve_sw_modes): ...new function.
5537 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5538 from a global struct to a subclass of rtl_opt_pass along with...
5539 (pass_data_mode_switch_use): ...new pass_data instance and...
5540 (make_pass_mode_switch_use): ...new function.
5541 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5542 from a global struct to a subclass of rtl_opt_pass along with...
5543 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5544 (make_pass_resolve_sw_modes): ...new function.
5545 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5546 struct to a subclass of rtl_opt_pass along with...
5547 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5548 (make_pass_insert_vzeroupper): ...new function.
5549 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5550 global struct to a subclass of rtl_opt_pass along with...
5551 (pass_data_work_around_errata): ...new pass_data instance and...
5552 (make_pass_work_around_errata): ...new function.
5553 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5554 struct to a subclass of rtl_opt_pass along with...
5555 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
5556 (make_pass_mips_machine_reorg2): ...new function.
5558 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5560 * passes.c (pass_manager::operator new): New.
5562 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5564 Handwritten part of conversion of passes to C++ classes.
5566 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
5567 (toplev.o): Add dep on PASS_MANAGER_H.
5568 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
5569 of early local pases to reflect this moving from a global to a
5570 member of gcc::pass_manager.
5571 (cgraph_add_new_function): Likewise.
5572 * lto-cgraph.c (lto_output_node): Update for conversion of
5573 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
5574 * passes.c (opt_pass::clone): New.
5575 (opt_pass::gate): New.
5576 (opt_pass::execute): New.
5577 (opt_pass::opt_pass): New.
5578 (pass_manager::execute_early_local_passes): New.
5579 (pass_manager::execute_pass_mode_switching): new.
5580 (finish_optimization_passes): Convert to...
5581 (pass_manager::finish_optimization_passes): ...this.
5582 (finish_optimization_passes): Update for conversion of passes to
5584 (register_dump_files_1): Use has_gate since we cannot portably
5585 check a vtable entry against NULL.
5586 (dump_one_pass): Likewise.
5587 (ipa_write_summaries_2): Likewise.
5588 (ipa_write_optimization_summaries_1): Likewise.
5589 (ipa_read_summaries_1): Likewise.
5590 (ipa_read_optimization_summaries_1): Likewise.
5591 (execute_ipa_stmt_fixups): Likewise.
5592 (pass_manager::pass_manager): Rewrite pass-creation, invoking
5593 pass-creation functions rather than wiring up globals, and
5594 storing the results in fields of pass_manager generated using
5596 (pass_manager::dump_profile_report): Update for conversion of
5597 passes to C++ classes.
5598 (pass_manager::execute_ipa_summary_passes): Likewise.
5599 (execute_one_ipa_transform_pass): Likewise.
5600 (execute_one_pass): Use has_gate and has_execute since we cannot
5601 portably check a vtable entry against NULL.
5602 * pass_manager.h (pass_manager::finish_optimization_passes): New.
5603 (pass_manager): Use pass-instances.def to add fields for the
5604 various pass instances.
5605 * toplev.c (finalize): Update for move of
5606 finish_optimization_passes to a method of gcc::pass_manager.
5607 * toplev.h (finish_optimization_passes): Move to method of class
5609 * tree-pass.h (struct pass_data): New.
5610 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
5611 (opt_pass::gate): Convert to virtual function.
5612 (opt_pass::~opt_pass): New.
5613 (opt_pass::clone): New.
5614 (opt_pass::execute): Convert to virtual function.
5615 (opt_pass::opt_pass): New.
5616 (opt_pass::ctxt_): new.
5617 (gimple_opt_pass): Convert to subclass of opt_pass.
5618 (gimple_opt_pass::gimple_opt_pass): New.
5619 (rtl_opt_pass): Convert to subclass of opt_pass.
5620 (rtl_opt_pass::rtl_opt_pass): New.
5621 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5622 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5623 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5624 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5625 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5626 invocation of pass_mode_switching to reflect this moving from a
5627 global to a member of gcc::pass_manager.
5628 (ix86_option_override): Rework how pass_insert_vzeroupper is
5629 added to the pass_manager to reflect autogenerated changes.
5630 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
5632 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
5634 PR rtl-optimization/57708
5635 * recog.c (peep2_find_free_register): Validate all regs in a
5638 2013-08-05 Jan Hubicka <jh@suse.cz>
5641 * cgraph.c (verify_cgraph_node): Accept local flags from other
5643 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5644 (function_and_variable_visibility): Likewise.
5645 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5647 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5649 * graph.c (init_graph_slim_pretty_print): Remove.
5650 (print_graph_cfg): Do not call it. Use local pretty printer.
5651 (start_graph_dump): Likewise.
5653 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5655 * gimple-pretty-print.c (buffer): Remove.
5656 (initialized): Likewise.
5657 (maybe_init_pretty_print): Likewise.
5658 (print_gimple_stmt): Do not call it. Use non-static local
5659 pretty_printer variable.
5660 (print_gimple_expr): Likewise.
5661 (print_gimple_seq): Likewise.
5662 (gimple_dump_bb): Likewise.
5664 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5666 * asan.c (asan_pp): Remove.
5667 (asan_pp_initialized): Likewise.
5668 (asan_pp_initialize): Likewise.
5669 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5670 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5671 (asan_add_global): Likewise.
5673 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5675 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5676 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5677 * pretty-print.c (pp_formatted_text_data): Likewise.
5678 (pp_indent): Rename from pp_base_indent.
5679 (pp_format): Rename from pp_base_format.
5680 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5681 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5682 (pp_flush): Rename from pp_base_flush.
5683 (pp_set_line_maximum_length): Rename from
5684 pp_base_set_line_maximum_length.
5685 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5686 (pp_set_prefix): Rename from pp_base_set_prefix.
5687 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5688 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5689 (pp_append_text): Rename from pp_base_append_text.
5690 (pp_formatted_text): Rename from pp_base_formatted_text.
5691 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5692 (pp_remaining_character_count_for_line): Rename from
5693 pp_base_remaining_character_count_for_line.
5694 (pp_newline): Rename from pp_base_newline.
5695 (pp_character): Rename from pp_base_character.
5696 (pp_string): Rename from pp_base_string.
5697 (pp_maybe_space): Rename from pp_base_maybe_space.
5698 * asan.c (asan_pp_string): Adjust.
5699 (asan_emit_stack_protection): Likewise.
5700 (asan_add_global): Likewise.
5701 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5702 * tree-mudflap.c (mf_varname_tree): Likewise.
5703 * tree-pretty-print.c (pp_tree_identifier): Rename from
5704 pp_base_tree_identifier.
5705 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5706 Declare as function.
5708 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5710 * pretty-print.h (pp_bar_bar): New.
5711 (pp_ampersand_ampersand): Likewise.
5712 (pp_less_equal): Likewise.
5713 (pp_greater_equal): Likewise.
5714 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5715 printer functions instead of pp_string or operators and punctuators.
5716 (dump_gimple_call): Likewise.
5717 (dump_gimple_omp_for): Likewise.
5718 (dump_gimple_transaction): Likewise.
5719 (dump_gimple_phi): Likewise.
5720 (pp_gimple_stmt_1): Likewise.
5721 * sched-vis.c (print_insn): Likewise.
5722 * tree-mudflap.c (mf_varname_tree): Likewise.
5723 * tree-pretty-print.c (dump_block_node): Likewise.
5724 (dump_generic_node): Likewise.
5726 2013-08-02 Jan Hubicka <jh@suse.cz>
5728 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5730 * lto-streamer-out.c (tree_is_indexable): Results decls and
5731 parm decls are not indexable.
5732 (DFS_write_tree_body): Do not follow args and results.
5733 (hash_tree): Likewise.
5734 (output_functions): Rearrange so struct function is needed
5735 only when real body is output; be able to also ouptut abstract
5736 functions; output DECL_ARGUMENTS and DECL_RESULT.
5737 (lto_output): When not in WPA, ale store abstract functions.
5738 (write_symbol): Do not care about RESULT_DECL.
5739 (output_symbol_p): Handle correctly sbtract decls.
5740 * lto-streamer-in.c (input_function): Rearrange so struct
5741 function can be NULL at entry; allow streaming of
5742 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5743 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5744 sanity check during LTO.
5745 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5746 RESULT_DECl and DECL_ARGUMENTS.
5747 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5750 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5752 * pretty-print.h (pp_underscore): New.
5754 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5755 printer functions instead of pp_character.
5756 (dump_binary_rhs): Likewise.
5757 (dump_ternary_rhs): Likewise.
5758 (dump_gimple_call_args): Likewise.
5759 (pp_points_to_solution): Likewise.
5760 (dump_gimple_call): Likewise.
5761 (dump_gimple_switch): Likewise.
5762 (dump_gimple_cond): Likewise.
5763 (dump_gimple_bind): Likewise.
5764 (dump_gimple_try): Likewise.
5765 (dump_gimple_omp_for): Likewise.
5766 (dump_gimple_omp_continue): Likewise.
5767 (dump_gimple_omp_single): Likewise.
5768 (dump_gimple_omp_sections): Likewise.
5769 (dump_gimple_omp_block): Likewise.
5770 (dump_gimple_omp_critical): Likewise.
5771 (dump_gimple_transaction): Likewise.
5772 (dump_gimple_asm): Likewise.
5773 (dump_gimple_phi): Likewise.
5774 (dump_gimple_omp_parallel): Likewise.
5775 (dump_gimple_omp_task): Likewise.
5776 (dump_gimple_omp_atomic_load): Likewise.
5777 (dump_gimple_omp_atomic_store): Likewise.
5778 (dump_gimple_mem_ops): Likewise.
5779 (pp_gimple_stmt_1): Likewise.
5780 (pp_cfg_jump): Likewise.
5781 (dump_implicit_edges): Likewise.
5782 (gimple_dump_bb_for_graph): Likewise.
5783 * graph.c (draw_cfg_node): Likewise.
5784 * langhooks.c (lhd_print_error_function): Likewise.
5785 * sched-vis.c (print_exp): Likewise.
5786 (print_value): Likewise.
5787 (print_pattern): Likewise.
5788 (print_insn): Likewise.
5789 (rtl_dump_bb_for_graph): Likewise.
5790 * tree-pretty-print.c (dump_function_declaration): Likewise.
5791 (dump_array_domain): Likewise.
5792 (dump_omp_clause): Likewise.
5793 (dump_location): Likewise.
5794 (dump_generic_node): Likewise.
5795 (print_struct_decl): Likewise.
5796 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5798 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5800 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5801 candidate table when replacing a candidate statement.
5802 (replace_rhs_if_not_dup): Likewise.
5803 (replace_one_candidate): Likewise.
5805 2013-08-02 Jan Hubicka <jh@suse.cz>
5806 Martin Liska <marxin.liska@gmail.com>
5808 * cgraphunit.c (add_new_function): Fix logic when adding from
5810 (assemble_thunk): Rename to ...
5811 (expand_thunk); .. this one; export; get it working with
5812 general functions; make produced gimple valid.
5813 * cgraph.h (expand_thunk): Declare.
5815 2013-08-02 Jan Hubicka <jh@suse.cz>
5817 * ipa-cp.c (gather_context_independent_values): Use
5818 ipa_get_param_move_cost.
5819 (get_replacement_map): Remove PARAM; move parameter folding
5821 (create_specialized_node): Update.
5822 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
5823 assert that we have gimple body; update move_cost.
5824 (count_formal_params): Assert that we have gimple body.
5825 (ipa_dump_param): New function.
5826 (ipa_alloc_node_params): Break out from ...
5827 (ipa_initialize_node_params): ... here.
5828 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
5829 (ipa_write_node_info): Stream move costs.
5830 (ipa_read_node_info): Read move costs.
5831 (ipa_update_after_lto_read): Do not recompute node params.
5832 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
5833 (ipa_get_param): Check we are not in WPA.
5834 (ipa_get_param_move_cost): New.
5835 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
5836 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
5837 parm numbers to be present.
5839 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5841 PR rtl-optimization/58048
5842 * lra-constraints.c (process_alt_operands): Don't check asm
5843 operand on register.
5845 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
5847 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
5848 the implied StoreLoad barrier for atomic operations if before.
5850 2013-08-02 Jan Hubicka <jh@suse.cz>
5851 Martin Liska <marxin.liska@gmail.com>
5853 * cgraph.c (cgraph_function_body_availability): Do not check
5855 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
5856 symtab_node_availability): Declare.
5857 * ipa.c (can_replace_by_local_alias): New.
5858 (function_and_variable_visibility): Use it.
5859 * symtab.c (symtab_for_node_and_aliases,
5860 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
5862 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5864 PR rtl-optimization/57963
5865 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
5866 (lra_constraints): Use them.
5868 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
5870 * config/arm/types.md (define_attr "type"): Add "load_acq"
5872 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
5874 (cortex_a53_store1): Likewise.
5876 2013-08-01 Jan Hubicka <jh@suse.cz>
5878 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
5879 partitions are not needed.
5881 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
5883 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
5884 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
5885 MAYBE_NON_Q_CLASS_P where appropriate.
5887 2013-08-01 Jan Hubicka <jh@suse.cz>
5889 * cgraph.h (release_function_body): Declare.
5890 * tree.c (free_lang_data_in_decl): Free, parameters and return values
5891 of unused delcarations.
5893 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5895 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
5896 RTL form when subtracting a constant.
5898 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5900 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
5901 Generate canonical plus rtx with negated immediate instead of minus
5903 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
5905 2013-08-01 Jan Hubicka <jh@suse.cz>
5907 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
5908 (cgraph_release_function_body): Likewise.
5909 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5910 * cgraph.h (cgrpah_node): Rename abstract_and_needed
5911 to used_as_abstract_origin.
5912 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
5913 symbols used as abstract origins.
5914 * cgraphunit.c (analyze_functions): Update.
5915 * ipa.c (symtab_remove_unreachable_nodes): Recompute
5916 used_as_abstract_origin.
5917 * tree-inline.c (tree_function_versioning): Update
5918 used_as_abstract_origin; be ready for DECL_RESULT and
5919 DECL_ARGUMENTS to be NULL.
5921 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
5922 for abstract functions.
5923 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
5926 2013-08-01 Jan Hubicka <jh@suse.cz>
5928 * profile.c (compute_value_histograms): Fix thinko.
5930 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5932 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
5933 aarch-common-protos.h to extra_headers.
5934 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5935 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
5936 * config/aarch64/t-aarch64 (aarch-common.o): Define.
5938 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5940 * config/aarch64/aarch64.md (define_attr "type"): Delete.
5941 Include "../arm/types.md". Define "type" attribute for all patterns.
5942 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
5945 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
5947 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
5948 to support power8 load fusion.
5949 (fusion_gpr_mem_load): Likewise.
5951 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
5953 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
5954 declarations for power8 load fusion.
5955 (emit_fusion_gpr_load): Likewise.
5957 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
5958 tuning for power8, turn on fusion mode by default. Turn on sign
5959 extending fusion mode if normal fusion mode is on, and we are at
5961 (fusion_gpr_load_p): New function, return true if we can fuse an
5962 addis instruction with a dependent load to a GPR.
5963 (emit_fusion_gpr_load): Emit the instructions for power8 load
5966 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
5967 (VSX load fusion peepholes): New peepholes to fuse together an
5968 addi instruction with a VSX load instruction.
5970 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
5971 peepholes to fuse an addis instruction with a load to a GPR base
5972 register. If we are supporting sign extending fusions, convert
5973 sign extending loads to zero extending loads and add an explicit
5976 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5978 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
5979 aarch-common-protos.h to extra_headers.
5980 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5981 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
5982 (arm_early_store_addr_dep): Likewise.
5983 (arm_no_early_alu_shift_dep): Likewise.
5984 (arm_no_early_alu_shift_value_dep): Likewise.
5985 (arm_no_early_mul_dep): Likewise.
5986 (arm_no_early_store_addr_dep): Likewise.
5987 (arm_mac_accumulator_is_mul_result): Likewise.
5988 (arm_mac_accumulator_is_result): Likewise.
5989 * config/arm/aarch-common.c: ... here. New file.
5990 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
5992 (arm_early_store_addr_dep): Likewise.
5993 (arm_no_early_alu_shift_dep): Likewise.
5994 (arm_no_early_alu_shift_value_dep): Likewise.
5995 (arm_no_early_mul_dep): Likewise.
5996 (arm_no_early_store_addr_dep): Likewise.
5997 (arm_mac_accumulator_is_mul_result): Likewise.
5998 (arm_mac_accumulator_is_result): Likewise.
5999 * config/arm/aarch-common-protos.h: ... here. New file.
6000 * config/arm/t-arm (aarch-common.o): Define.
6002 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6004 * config/arm/arm.md: Include new file "types.md".
6005 (define_attr "type"): Move from here to ...
6006 (define_attr "mul32"): Likewise.
6007 (define_attr "mul64"): Likewise.
6008 * config/arm/types.md: ... here. New file.
6010 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
6012 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
6013 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
6015 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6017 * gen-pass-instances.awk: Fix offset of substr().
6019 2013-07-31 David Malcolm <dmalcolm@redhat.com>
6021 * Makefile.in (pass-instances.def): New.
6022 (passes.o): Replace dependency on passes.def with one on
6025 * gen-pass-instances.awk: New.
6027 * passes.c (pass_manager::pass_manager): Use pass-instances.def
6028 rather than passes.def, updating local definition of NEXT_PASS
6029 macro to add an extra NUM parameter (currently unused).
6031 2013-07-30 David Malcolm <dmalcolm@redhat.com>
6033 * Makefile.in (PASS_MANAGER_H): New.
6034 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
6035 (passes.o): Likewise.
6036 (statistics.o): Likewise.
6037 (cgraphunit.o): Likewise.
6038 (context.o): Depend on PASS_MANAGER_H.
6040 * pass_manager.h: New.
6042 * cgraphunit.c (cgraph_add_new_function): Update for moves
6043 of globals to fields of pass_manager.
6044 (analyze_function): Likewise.
6045 (expand_function): Likewise.
6046 (ipa_passes): Likewise.
6047 (compile): Likewise.
6049 * context.c (context::context): New.
6050 * context.h (context::context): New.
6051 (context::get_passes): New.
6052 (context::passes_): New.
6054 * lto-cgraph.c (input_node): Update for moves of globals to
6055 fields of pass_manager.
6057 * passes.c (all_passes): Remove, in favor of a field of the
6058 same name within the new class pass_manager.
6059 (all_small_ipa_passes): Likewise.
6060 (all_lowering_passes): Likewise.
6061 (all_regular_ipa_passes): Likewise.
6062 (all_late_ipa_passes): Likewise.
6063 (all_lto_gen_passes): Likewise.
6064 (passes_by_id): Likewise.
6065 (passes_by_id_size): Likewise.
6066 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6067 the new class pass_manager.
6068 (set_pass_for_id): Convert to...
6069 (pass_manager::set_pass_for_id): ...method.
6070 (get_pass_for_id): Convert to...
6071 (pass_manager::get_pass_for_id): ...method.
6072 (register_one_dump_file): Move body of implementation into...
6073 (pass_manager::register_one_dump_file): ...here.
6074 (register_dump_files_1): Convert to...
6075 (pass_manager::register_dump_files_1): ...method.
6076 (register_dump_files): Convert to...
6077 (pass_manager::register_dump_files): ...method.
6078 (create_pass_tab): Update for moves of globals to fields of
6080 (dump_passes): Move body of implementation into...
6081 (pass_manager::dump_passes): ...here.
6082 (register_pass): Move body of implementation into...
6083 (pass_manager::register_pass): ...here.
6084 (init_optimization_passes): Convert into...
6085 (pass_manager::pass_manager): ...constructor for new
6086 pass_manager class, and initialize the pass_lists array.
6087 (check_profile_consistency): Update for moves of globals to
6088 fields of pass_manager.
6089 (dump_profile_report): Move body of implementation into...
6090 (pass_manager::dump_profile_report): ...here.
6091 (ipa_write_summaries_1): Update for moves of pass lists from
6092 being globals to fields of pass_manager.
6093 (ipa_write_optimization_summaries): Likewise.
6094 (ipa_read_summaries): Likewise.
6095 (ipa_read_optimization_summaries): Likewise.
6096 (execute_all_ipa_stmt_fixups): Likewise.
6098 * statistics.c (statistics_fini): Update for moves of globals to
6099 fields of pass_manager.
6101 * toplev.c (general_init): Replace call to
6102 init_optimization_passes with construction of the pass_manager
6105 * tree-pass.h (all_passes): Remove, in favor of a field of the
6106 same name within the new class pass_manager.
6107 (all_small_ipa_passes): Likewise.
6108 (all_lowering_passes): Likewise.
6109 (all_regular_ipa_passes): Likewise.
6110 (all_lto_gen_passes): Likewise.
6111 (all_late_ipa_passes): Likewise.
6112 (passes_by_id): Likewise.
6113 (passes_by_id_size): Likewise.
6114 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6115 the new class pass_manager.
6116 (get_pass_for_id): Remove.
6118 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6120 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
6123 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6125 * arm.md (mulhi3): New expand pattern.
6127 2013-07-30 Jan Hubicka <jh@suse.cz>
6128 Martin Liska <marxin.liska@gmail.com>
6130 * profile.c (compute_value_histograms): Do not ICE when
6131 there is mismatch only on some counters.
6133 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6135 PR rtl-optimization/57637
6136 * function.c (move_insn_for_shrink_wrap): Also check the
6137 GEN set of the LIVE problem for the liveness analysis
6138 if it exists, otherwise give up.
6140 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6142 PR tree-optimization/57993
6143 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
6144 replaced statement in the candidate table.
6145 (phi_add_costs): Return infinite cost when the hidden basis does
6146 not dominate all phis on which the candidate is dependent.
6147 (replace_one_candidate): Record replaced statement in the
6150 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
6152 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
6153 (ashlv2si3): New expander.
6154 (*ashlv2si3_i): New define_insn_and_split.
6155 * predicates.md (float_operation): Allow patterns with three
6158 PR rtl-optimization/58021
6159 * mode-switching.c (create_pre_exit): Always split off preceding
6160 insns if we are not at the basic block head.
6162 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
6164 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
6165 (UCLIBC_DYNAMIC_LINKER): New macro.
6166 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
6168 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
6169 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
6171 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
6172 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
6173 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
6174 for SF and DF modes. Use ieee_quad_format for TF mode.
6175 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
6176 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
6177 (mips_option_override): Handle `-mnan=legacy'.
6178 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
6179 `-mabs=2008' and `-mnan=2008'.
6180 (OPTION_DEFAULT_SPECS): Add "nan" default.
6181 (ASM_SPEC): Handle `-mnan='.
6182 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
6183 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
6184 comment accordingly.
6185 (neg<mode>2): Likewise.
6186 * config/mips/mips.opt (mabs, mnan): New options.
6187 * doc/install.texi (Configuration): Document `--with-nan=' option.
6188 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
6190 (MIPS Options): Document them.
6191 * config.gcc <mips*-*-*>: Handle `--with-nan='.
6192 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
6193 * configure: Regenerate.
6194 * config.in: Regenerate.
6196 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6198 * config/i386/i386.md (float post-reload splitters): Do not check
6199 for subregs of SSE registers.
6201 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6202 H.J. Lu <hongjiu.lu@intel.com>
6206 * config/i386/i386.md (post-reload splitter
6207 to avoid partial SSE reg dependency stalls): New pattern.
6209 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
6211 * config/s390/s390.md ("movcc"): Swap load and store instructions.
6213 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
6215 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
6216 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
6218 2013-07-26 Cary Coutant <ccoutant@google.com>
6220 * dwarf2out.c (die_checksum_ordered): Don't include template
6221 instantiations in signature.
6222 (is_template_parameter): New function.
6223 (is_template_instantiation): New function.
6224 (generate_skeleton_bottom_up): Don't include template instantiations
6226 (generate_skeleton): Likewise.
6227 (break_out_comdat_types): Move recursive call to break out nested
6229 (prune_unused_types_mark_generic_parms_dies): Call
6230 is_template_parameter.
6232 2013-07-26 Ian Bolton <ian.bolton@arm.com>
6234 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
6235 uses vector registers.
6236 * config/aarch64/iterators.md: Add attributes rtn and vas.
6238 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6239 Richard Earnshaw <richard.earnshaw@arm.com>
6241 * combine.c (simplify_comparison): Re-canonicalize operands
6243 * config/arm/arm.md (movcond_addsi): New splitter.
6245 2013-07-25 Sterling Augustine <saugustine@google.com>
6247 * dwarf2out.c (size_of_pubnames): Move code to...
6248 (include_pubname_in_output): ...here. New.
6249 (want_pubnames): Rearrange.
6250 (output_pubnames): Call include_pubname_in_output. Move assertion.
6252 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6254 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
6256 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6259 * doc/extend.texi: Remove obsolete builtins. Fix
6260 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
6262 2013-07-25 Jan Hubicka <jh@suse.cz>
6264 * cgraph.c (release_function_body): Break out from ...
6265 (cgraph_release_function_body): ... this one; also release DECL_RESULT
6267 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
6268 old_tree in the map.
6269 (create_specialized_node): Update.
6270 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
6272 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
6273 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
6274 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
6275 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
6278 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6280 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
6281 addsi3_carryin_alt2_<optab>): Correct output template.
6283 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6285 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
6286 Adjust for arm_restrict_it.
6287 Remove trailing whitespace.
6289 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
6291 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
6292 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
6294 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
6296 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
6298 PR rtl-optimization/57960
6299 * lra-constraints.c (process_alt_operands): Use the right mode
6300 when checking strict_low.
6302 2013-07-25 Jan Hubicka <jh@suse.cz>
6304 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
6305 * cgraph.c (cgraph_remove_node): Do not release function body
6306 when in cgraph streaming.
6307 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
6308 in other partitions are not considered reachable; fix handling of
6311 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6313 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
6315 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6320 * config/arm/arm.md ("*sibcall_insn): Replace use of
6321 Ss with US. Adjust output for v5 and v4t.
6322 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
6324 * config/arm/constraints.md ("Ss"): Rename to US.
6326 2013-07-25 Terry Guo <terry.guo@arm.com>
6328 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
6329 shift_add/shift_sub0/shift_sub1 RTXs.
6331 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
6332 Anton Blanchard <anton@au1.ibm.com>
6334 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
6335 (altivec_vpks<VI_char>ss): Likewise.
6336 (altivec_vpks<VI_char>us): Likewise.
6337 (altivec_vpku<VI_char>us): Likewise.
6338 (altivec_vpku<VI_char>um): Likewise.
6340 2013-07-24 David Malcolm <dmalcolm@redhat.com>
6342 Introduce context class.
6344 * Makefile.in (CONTEXT_H): New.
6345 (OBJS): Add context.o.
6346 (toplev.o): Add CONTEXT_H to dependencies.
6349 * toplev.c (general_init): Create the singleton gcc::context instance.
6355 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
6357 PR rtl-optimization/57968
6358 * mode-switching.c (create_pre_exit): Allow instructions that
6359 don't set a return register to need a non-exit mode.
6361 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6362 Anton Blanchard <anton@au1.ibm.com>
6364 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
6365 operands to vperm for little endian.
6366 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
6367 of lvsl to create the control mask for a vperm for little endian.
6369 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6370 Anton Blanchard <anton@au1.ibm.com>
6372 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6373 two operands for little-endian.
6375 2013-07-23 Steve Ellcey <sellcey@mips.com>
6377 * config/mips/mips.c (mips_case_values_threshold): New.
6378 (TARGET_CASE_VALUES_THRESHOLD): Define.
6380 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6381 Anton Blanchard <anton@au1.ibm.com>
6383 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
6384 selection of field for vector splat in little endian mode.
6386 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6388 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
6389 expanders to rs6000.md.
6390 (ior<mode>3): Likewise.
6391 (and<mode>3): Likewise.
6392 (one_cmpl<mode>2): Likewise.
6393 (nor<mode>3): Likewise.
6394 (andc<mode>3): Likewise.
6395 (eqv<mode>3): Likewise.
6396 (nand<mode>3): Likewise.
6397 (orc<mode>3): Likewise.
6399 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
6402 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
6403 to split multi-word logical operations.
6404 (rs6000_split_logical_di): Likewise.
6405 (rs6000_split_logical): Likewise.
6407 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
6408 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
6409 and allow TImode operations in 32-bit.
6410 (vsx_and<mode>3_64bit): Likewise.
6411 (vsx_ior<mode>3_32bit): Likewise.
6412 (vsx_ior<mode>3_64bit): Likewise.
6413 (vsx_xor<mode>3_32bit): Likewise.
6414 (vsx_xor<mode>3_64bit): Likewise.
6415 (vsx_one_cmpl<mode>2_32bit): Likewise.
6416 (vsx_one_cmpl<mode>2_64bit): Likewise.
6417 (vsx_nor<mode>3_32bit): Likewise.
6418 (vsx_nor<mode>3_64bit): Likewise.
6419 (vsx_andc<mode>3_32bit): Likewise.
6420 (vsx_andc<mode>3_64bit): Likewise.
6421 (vsx_eqv<mode>3_32bit): Likewise.
6422 (vsx_eqv<mode>3_64bit): Likewise.
6423 (vsx_nand<mode>3_32bit): Likewise.
6424 (vsx_nand<mode>3_64bit): Likewise.
6425 (vsx_orc<mode>3_32bit): Likewise.
6426 (vsx_orc<mode>3_64bit): Likewise.
6428 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
6429 logical types in GPRs.
6431 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
6432 logical insns to rs6000.md, and allow TImode operations in
6434 (altivec_ior<mode>3): Likewise.
6435 (altivec_xor<mode>3): Likewise.
6436 (altivec_one_cmpl<mode>2): Likewise.
6437 (altivec_nor<mode>3): Likewise.
6438 (altivec_andc<mode>3): Likewise.
6440 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
6441 attributes for moving the 128-bit logical operations into
6443 (BOOL_REGS_OUTPUT): Likewise.
6444 (BOOL_REGS_OP1): Likewise.
6445 (BOOL_REGS_OP2): Likewise.
6446 (BOOL_REGS_UNARY): Likewise.
6447 (BOOL_REGS_AND_CR0): Likewise.
6448 (one_cmpl<mode>2): Add support for DI logical operations on
6449 32-bit, splitting the operations to 32-bit.
6453 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
6454 changes to combine the 32/64-bit code, allow logical operations on
6455 TI mode in 32-bit, and to use similar match_operator patterns like
6456 scalar mode uses. Combine the Altivec and VSX code for logical
6457 operations, and move it here.
6458 (ior<mode>3, 128-bit types): Likewise.
6459 (xor<mode>3, 128-bit types): Likewise.
6460 (one_cmpl<mode>3, 128-bit types): Likewise.
6461 (nor<mode>3, 128-bit types): Likewise.
6462 (andc<mode>3, 128-bit types): Likewise.
6463 (eqv<mode>3, 128-bit types): Likewise.
6464 (nand<mode>3, 128-bit types): Likewise.
6465 (orc<mode>3, 128-bit types): Likewise.
6466 (and<mode>3_internal): Likewise.
6467 (bool<mode>3_internal): Likewise.
6468 (boolc<mode>3_internal1): Likewise.
6469 (boolc<mode>3_internal2): Likewise.
6470 (boolcc<mode>3_internal1): Likewise.
6471 (boolcc<mode>3_internal2): Likewise.
6472 (eqv<mode>3_internal1): Likewise.
6473 (eqv<mode>3_internal2): Likewise.
6474 (one_cmpl1<mode>3_internal): Likewise.
6476 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6478 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6479 Rename flag_stack_usage to flag_stack_usage_info.
6481 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6483 * config/microblaze/sync.md: New file.
6484 * config/microblaze/microblaze.md: Include sync.md
6485 * config/microblaze/microblaze.c: Add print_operand 'y'.
6486 * config/microblaze/constraints.md: Add memory_contraint
6487 'Q' which is a single register.
6489 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6491 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6493 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6495 * reload.c (find_reloads): Exit loop once we find this operand
6496 cannot be reloaded somehow for this alternative.
6498 * reload.c (find_reloads): Exit loop once we find a hard register.
6500 * rtlanal.c (computed_jump_p): Exit loop once we find label
6503 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6505 * cfgloopmanip.c (remove_path): Exit loop after setting
6508 * gensupport.c (subst_dup): Avoid loop if code is not
6509 MATCH_DUP nor MATCH_OP_DUP.
6511 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6513 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6515 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6517 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6518 true for SP_REGNUM if mode == ptr_mode.
6519 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6520 with value R0_REGNUM + 31.
6522 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6524 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6525 pad pointer-typed argument downward.
6527 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6529 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6530 and __ILP32__ when the ILP32 model is in use.
6532 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6534 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6535 (aarch64_load_symref_appropriately): In the case of
6536 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6537 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6538 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6539 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6540 if 'mode' doesn't equal to 'ptr_mode'.
6541 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6542 'vcall_offset'; change to call aarch64_emit_move differently depending
6543 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6544 to calculate the upper bound of 'vcall_offset'.
6545 (aarch64_cannot_force_const_mem): Change to also return true if
6547 (aarch64_legitimize_reload_address): In the case of large
6548 displacements, add new local variable 'xmode' and an assertion
6549 based on it; change to use 'xmode' to generate the new rtx and
6551 (aarch64_asm_trampoline_template): Change to generate the template
6552 differently depending on TARGET_ILP32 or not; change to use
6553 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6554 (aarch64_trampoline_size): Removed.
6555 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
6556 and replace immediate literals with it. Change to use 'ptr_mode'
6557 instead of 'DImode' and call convert_memory_address if the mode
6558 of 'fnaddr' doesn't equal to 'ptr_mode'.
6559 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
6561 (aarch64_elf_asm_destructor): Likewise.
6562 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
6563 on TARGET_ILP32 instead of aarch64_trampoline_size.
6564 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
6565 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
6566 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
6567 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
6568 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
6569 (storewb_pair<GPI:mode>_<P:mode>): ... this.
6570 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
6571 depending on the value of 'mode'.
6572 (add_losym_<mode>): New.
6573 (ldr_got_small_<mode>): New, based on ldr_got_small.
6574 (ldr_got_small): Remove.
6575 (ldr_got_small_sidi): New.
6576 * config/aarch64/iterators.md (P): New.
6577 (PTR): Change to 'ptr_mode' in the condition.
6579 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6581 * config.gcc (aarch64*-*-*): Support --with-abi.
6582 (aarch64*-*-elf): Support --with-multilib-list.
6583 (aarch64*-*-linux*): Likewise.
6584 (supported_defaults): Add abi to aarch64*-*-*.
6585 * configure.ac: Mention AArch64 for --with-multilib-list.
6586 * configure: Re-generated.
6587 * config/aarch64/biarchilp32.h: New file.
6588 * config/aarch64/biarchlp64.h: New file.
6589 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
6591 (MULTILIB_DEFAULTS): Ditto.
6592 (DRIVER_SELF_SPECS): Ditto.
6593 (ASM_SPEC): Update to also substitute -mabi.
6594 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
6595 file whose name depends on -mabi= and -mbig-endian.
6596 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
6598 (POINTER_SIZE): New define.
6599 (POINTERS_EXTEND_UNSIGNED): Ditto.
6600 (enum aarch64_abi_type): New enumeration tag.
6601 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
6602 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
6603 (TARGET_ILP32): New define.
6604 * config/aarch64/aarch64.opt (mabi): New.
6606 (ilp32, lp64): New values for -mabi.
6607 * config/aarch64/t-aarch64 (comma): New define.
6608 (MULTILIB_OPTIONS): Ditto.
6609 (MULTILIB_DIRNAMES): Ditto.
6610 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
6611 * doc/invoke.texi: Document -mabi for AArch64.
6613 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
6615 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
6617 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6618 Anton Blanchard <anton@au1.ibm.com>
6620 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6621 endianness when selecting field to splat.
6623 2013-07-22 Eric Christopher <echristo@gmail.com>
6625 * dwarf2out.c (die_odr_checksum): New function to use
6626 CHECKSUM_ macros and ULEB128 for DIE tag.
6627 (generate_type_signature): Use.
6629 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6631 * config.gcc (sparc*-*-*): Accept leon3 processor.
6632 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6633 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6634 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6635 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6636 (mfix-ut699): Adjust comment.
6637 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6638 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6639 (CPP_CPU_SPEC): Likewise.
6640 (ASM_CPU_SPEC): Likewise.
6641 * config/sparc/sparc.c (leon3_cost): New constant.
6642 (sparc_option_override): Add leon3 support.
6643 (mem_ref): New function.
6644 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6645 (sparc_do_work_around_errata): Look into the instruction in the delay
6646 slot and adjust accordingly. Add fix for the data cache nullify issues
6647 of the UT699. Change insertion position for the NOP.
6648 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6649 (leon3_load): New reservation.
6650 (leon_store): Bump latency to 2.
6651 (grfpu): New automaton.
6652 (grfpu_alu): New unit.
6653 (grfpu_ds): Likewise.
6654 (leon_fp_alu): Adjust.
6655 (leon_fp_mult): Delete.
6656 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6657 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6658 * config/sparc/sparc.md (cpu): Add leon3.
6659 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6661 (atomic_test_and_set): Likewise.
6664 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6666 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6667 default for R5900 targets.
6668 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6669 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6670 * config/mips/mips.c (mips_option_override): Report an error for
6671 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6672 for -march=r5900 -mhard-float.
6674 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6676 * df-problems.c (can_move_insns_across): Exit loop once we
6677 find a non-fixed, non-global register.
6679 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6682 * omega.c (omega_eliminate_red): Break after setting red_found.
6683 (omega_problem_has_red_equations): Similarly after setting found.
6684 (omega_query_variable): Similarly after setting coupled.
6686 2013-07-22 Marek Polacek <polacek@redhat.com>
6688 * gimplify.c: Don't include gimple.h twice.
6690 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6692 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6693 instead of TARGET_THUMB1.
6694 (Pz): New constraint.
6695 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6697 (compare_negsi_si): Likewise.
6698 (compare_addsi2_op0): Likewise.
6699 (compare_addsi2_op1): Likewise.
6700 (addsi3_carryin_<optab>): Likewise.
6701 (addsi3_carryin_alt2_<optab>): Likewise.
6702 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6703 for arm_restrict_it.
6704 (subsi3_carryin): Likewise.
6705 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6706 (minmax_arithsi): Disable for arm_restrict_it.
6707 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6708 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6709 (satsi_<SAT:code>_shift): Likewise.
6710 (arm_shiftsi3): Add alternative for 16-bit encoding.
6711 (arm32_movhf): Disable for arm_restrict_it.
6712 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6713 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6715 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6717 * config/arm/arm.md (attribute "insn"): Delete.
6718 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6719 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6720 (not_shiftsi): Update for attribute change.
6721 (not_shiftsi_compare0): Likewise.
6722 (not_shiftsi_compare0_scratch): Likewise.
6723 (arm_one_cmplsi2): Likewise.
6724 (thumb1_one_cmplsi2): Likewise.
6725 (notsi_compare0): Likewise.
6726 (notsi_compare0_scratch): Likewise.
6727 (thumb1_movdi_insn): Likewise.
6728 (arm_movsi_insn): Likewise.
6729 (movhi_insn_arch4): Likewise.
6730 (movhi_bytes): Likewise.
6731 (arm_movqi_insn): Likewise.
6732 (thumb1_movqi_insn): Likewise.
6733 (arm32_movhf): Likewise.
6734 (thumb1_movhf): Likewise.
6735 (arm_movsf_soft_insn): Likewise.
6736 (thumb1_movsf_insn): Likewise.
6737 (thumb_movdf_insn): Likewise.
6738 (movsicc_insn): Likewise.
6739 (movsfcc_soft_insn): Likewise.
6740 (and_scc): Likewise.
6741 (cond_move): Likewise.
6742 (if_move_not): Likewise.
6743 (if_not_move): Likewise.
6744 (if_shift_move): Likewise.
6745 (if_move_shift): Likewise.
6746 (if_shift_shift): Likewise.
6747 (if_not_arith): Likewise.
6748 (if_arith_not): Likewise.
6749 (cond_move_not): Likewise.
6750 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6751 (neon_mov<mode>): Likewise.
6752 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6753 (thumb2_movsi_vfp): Likewise.
6754 (movsf_vfp): Likewise.
6755 (thumb2_movsf_vfp): Likewise.
6756 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6758 (cortexa7_older_only): Likewise.
6759 (cortexa7_younger): Likewise.
6760 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6761 (1020alu_shift_op): Likewise.
6762 (1020alu_shift_reg_op): Likewise.
6763 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6764 (alu_shift_op): Likewise.
6765 (alu_shift_reg_op): Likewise.
6766 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6767 (11_alu_shift_op): Likewise.
6768 (11_alu_shift_reg_op): Likewise.
6769 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6770 (9_alu_shift_reg_op): Likewise.
6771 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6773 (cortex_a15_alu_shift): Likewise.
6774 (cortex_a15_alu_shift_reg): Likewise.
6775 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6776 (cortex_a5_alu_shift): Likewise.
6777 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6779 (cortex_a53_alu_shift): Likewise.
6780 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6782 (cortex_a7_alu_reg): Likewise.
6783 (cortex_a7_alu_shift): Likewise.
6784 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6785 (cortex_a8_alu_shift): Likewise.
6786 (cortex_a8_alu_shift_reg): Likewise.
6787 (cortex_a8_mov): Likewise.
6788 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6789 (cortex_a9_dp_shift): Likewise.
6790 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6791 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6792 (cortex_r4_mov): Likewise.
6793 (cortex_r4_alu_shift): Likewise.
6794 (cortex_r4_alu_shift_reg): Likewise.
6795 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6796 (526_alu_shift_op): Likewise.
6797 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6798 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6799 (626te_alu_shift_op): Likewise.
6800 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6801 (726te_alu_op): Likewise.
6802 (726te_alu_shift_op): Likewise.
6803 (726te_alu_shift_reg_op): Likewise.
6804 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6805 (mp626_alu_shift_op): Likewise.
6806 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6807 (pj4_alu_e1_conds): Likewise.
6808 (pj4_alu): Likewise.
6809 (pj4_alu_conds): Likewise.
6810 (pj4_shift): Likewise.
6811 (pj4_shift_conds): Likewise.
6812 (pj4_alu_shift): Likewise.
6813 (pj4_alu_shift_conds): Likewise.
6815 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6817 * config/arm/predicates.md (shiftable_operator_strict_it):
6819 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
6820 Disable cond_exec version for arm_restrict_it.
6821 (thumb2_smaxsi3): Convert to generate cond_exec.
6822 (thumb2_sminsi3): Likewise.
6823 (thumb32_umaxsi3): Likewise.
6824 (thumb2_uminsi3): Likewise.
6825 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
6826 (thumb2_neg_abssi2): Likewise.
6827 (thumb2_mov_scc): Add alternative for 16-bit encoding.
6828 (thumb2_movsicc_insn): Adjust alternatives.
6829 (thumb2_mov_negscc): Disable for arm_restrict_it.
6830 (thumb2_mov_negscc_strict_it): New pattern.
6831 (thumb2_mov_notscc_strict_it): New pattern.
6832 (thumb2_mov_notscc): Disable for arm_restrict_it.
6833 (thumb2_ior_scc): Likewise.
6834 (thumb2_ior_scc_strict_it): New pattern.
6835 (thumb2_cond_move): Adjust for arm_restrict_it.
6836 (thumb2_cond_arith): Disable for arm_restrict_it.
6837 (thumb2_cond_arith_strict_it): New pattern.
6838 (thumb2_cond_sub): Adjust for arm_restrict_it.
6839 (thumb2_movcond): Likewise.
6840 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
6841 (thumb2_zero_extendhisi2_v6): Likewise.
6842 (thumb2_zero_extendqisi2_v6): Likewise.
6843 (orsi_notsi_si): Likewise.
6844 (orsi_not_shiftsi_si): Likewise.
6846 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
6848 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
6849 instruction sequence is 1 byte shorter.
6851 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
6853 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
6854 it is not needed after split.
6856 2013-07-20 Iain Sandoe <iain@codesourcery.com>
6859 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
6860 second label for nonlocal goto receivers. Don't output pic base labels
6861 unless we're producing PIC; mark that action unreachable().
6862 (ix86_save_reg): If the function contains a nonlocal label, save the
6864 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
6865 * config/darwin.c (emitted_pic_label_num): New GTY.
6866 (update_pic_label_number_if_needed): New.
6867 (machopic_output_function_base_name): Adjust for nonlocal receiver
6869 (machopic_should_output_picbase_label): New.
6870 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
6871 (nonlocal_goto_receiver): New insn and split.
6873 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
6875 * config/aarch64/aarch64-builtins.c
6876 (aarch64_fold_builtin): Fold abs in all modes.
6877 * config/aarch64/aarch64-simd-builtins.def
6878 (abs): Enable for all modes.
6879 * config/aarch64/arm_neon.h
6880 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
6881 (vabs_f64): Add missing intrinsic.
6883 2013-07-19 Ian Bolton <ian.bolton@arm.com>
6885 * config/aarch64/arm_neon.h (vabs_s64): New function
6887 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
6890 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
6891 * config/avr/avr.md (adjust_len): Add `round'.
6892 * config/avr/avr-protos.h (avr_out_round): New prototype.
6893 (avr_out_plus): Add `out_label' argument.
6894 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
6895 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
6896 Handle the case where `insn' is just a pattern.
6897 (avr_out_bitop): Handle the case where `insn' is just a pattern.
6898 (avr_out_round): New function.
6899 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
6901 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
6903 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6904 Add check for flag_stack_usage to handle -fstack-usage support
6906 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
6908 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
6909 interaction for new Power8 flags and VSX.
6911 2013-07-18 Sriraman Tallam <tmsriram@google.com>
6914 * tree-inline.c (expand_call_inline): Emit errors during
6915 early_inlining only if optimization is not turned on.
6917 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6921 * passes.c (init_optimization_passes): Move the construction of
6922 the pass hierarchy into a new passes.def file.
6924 * Makefile.in (passes.o): Add dependency on passes.def.
6926 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6928 * passes.c (init_optimization_passes): Introduce macros for
6929 constructing the tree of passes (INSERT_PASSES_AFTER,
6930 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
6931 TERMINATE_PASS_LIST).
6933 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
6934 Wei Mi <wmi@google.com>
6936 PR rtl-optimization/57878
6937 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
6939 (reload_pseudo_compare_func): Check nregs first for reload
6942 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6944 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
6946 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
6948 * read-rtl.c (validate_const_int): Once an invalid character is
6949 seen, quit the loop.
6951 * gengtype.c (write_roots): Similarly once we find the "deletable"
6952 or "if_marked" option.
6954 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6956 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
6957 "xtab" and "sat". Move value "clz" from here to ...
6958 (attriubte "type"): ... here.
6959 (satsi_<SAT:code>): Delete "insn" attribute.
6960 (satsi_<SAT:code>_shift): Likewise.
6961 (arm_zero_extendqisi2addsi): Likewise.
6962 (arm_extendqisi2addsi): Likewise.
6963 (clzsi2): Update for attribute changes.
6964 (rbitsi2): Likewise.
6965 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
6967 (arm_usatsihi): Likewise.
6968 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6970 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6972 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
6973 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
6974 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
6975 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
6976 in alphabetical order.
6977 (attribute "core_cycles"): Update for attribute changes.
6978 (arm_addsi3): Likewise.
6979 (addsi3_compare0): Likewise.
6980 (addsi3_compare0_scratch): Likewise.
6981 (addsi3_compare_op1): Likewise.
6982 (addsi3_compare_op2): Likewise.
6983 (compare_addsi2_op0): Likewise.
6984 (compare_addsi2_op1): Likewise.
6985 (addsi3_carryin_shift_<optab>): Likewise.
6986 (subsi3_carryin_shift): Likewise.
6987 (rsbsi3_carryin_shift): Likewise.
6988 (arm_subsi3_insn): Likewise.
6989 (subsi3_compare0): Likewise.
6990 (subsi3_compare): Likewise.
6991 (arm_andsi3_insn): Likewise.
6992 (thumb1_andsi3_insn): Likewise.
6993 (andsi3_compare0): Likewise.
6994 (andsi3_compare0_scratch): Likewise.
6995 (zeroextractsi_compare0_scratch
6996 (andsi_not_shiftsi_si): Likewise.
6997 (iorsi3_insn): Likewise.
6998 (iorsi3_compare0): Likewise.
6999 (iorsi3_compare0_scratch): Likewise.
7000 (arm_xorsi3): Likewise.
7001 (thumb1_xorsi3_insn): Likewise.
7002 (xorsi3_compare0): Likewise.
7003 (xorsi3_compare0_scratch): Likewise.
7004 (satsi_<SAT:code>_shift): Likewise.
7006 (arm_shiftsi3): Likewise.
7007 (shiftsi3_compare0): Likewise.
7008 (not_shiftsi): Likewise.
7009 (not_shiftsi_compare0): Likewise.
7010 (not_shiftsi_compare0_scratch): Likewise.
7011 (arm_one_cmplsi2): Likewise.
7012 (thumb_one_complsi2): Likewise.
7013 (notsi_compare0): Likewise.
7014 (notsi_compare0_scratch): Likewise.
7015 (thumb1_zero_extendhisi2): Likewise.
7016 (arm_zero_extendhisi2): Likewise.
7017 (arm_zero_extendhisi2_v6): Likewise.
7018 (arm_zero_extendhisi2addsi): Likewise.
7019 (thumb1_zero_extendqisi2): Likewise.
7020 (thumb1_zero_extendqisi2_v6): Likewise.
7021 (arm_zero_extendqisi2): Likewise.
7022 (arm_zero_extendqisi2_v6): Likewise.
7023 (arm_zero_extendqisi2addsi): Likewise.
7024 (thumb1_extendhisi2): Likewise.
7025 (arm_extendhisi2): Likewise.
7026 (arm_extendhisi2_v6): Likewise.
7027 (arm_extendqisi): Likewise.
7028 (arm_extendqisi_v6): Likewise.
7029 (arm_extendqisi2addsi): Likewise.
7030 (thumb1_extendqisi2): Likewise.
7031 (thumb1_movdi_insn): Likewise.
7032 (arm_movsi_insn): Likewise.
7033 (movsi_compare0): Likewise.
7034 (movhi_insn_arch4): Likewise.
7035 (movhi_bytes): Likewise.
7036 (arm_movqi_insn): Likewise.
7037 (thumb1_movqi_insn): Likewise.
7038 (arm32_movhf): Likewise.
7039 (thumb1_movhf): Likewise.
7040 (arm_movsf_soft_insn): Likewise.
7041 (thumb1_movsf_insn): Likewise.
7042 (movdf_soft_insn): Likewise.
7043 (thumb_movdf_insn): Likewise.
7044 (arm_cmpsi_insn): Likewise.
7045 (cmpsi_shiftsi): Likewise.
7046 (cmpsi_shiftsi_swp): Likewise.
7047 (arm_cmpsi_negshiftsi_si): Likewise.
7048 (movsicc_insn): Likewise.
7049 (movsfcc_soft_insn): Likewise.
7050 (arith_shiftsi): Likewise.
7051 (arith_shiftsi_compare0
7052 (arith_shiftsi_compare0_scratch
7053 (sub_shiftsi): Likewise.
7054 (sub_shiftsi_compare0
7055 (sub_shiftsi_compare0_scratch
7056 (and_scc): Likewise.
7057 (cond_move): Likewise.
7058 (if_plus_move): Likewise.
7059 (if_move_plus): Likewise.
7060 (if_move_not): Likewise.
7061 (if_not_move): Likewise.
7062 (if_shift_move): Likewise.
7063 (if_move_shift): Likewise.
7064 (if_shift_shift): Likewise.
7065 (if_not_arith): Likewise.
7066 (if_arith_not): Likewise.
7067 (cond_move_not): Likewise.
7068 (thumb1_ashlsi3): Set type attribute.
7069 (thumb1_ashrsi3): Likewise.
7070 (thumb1_lshrsi3): Likewise.
7071 (thumb1_rotrsi3): Likewise.
7072 (shiftsi3_compare0_scratch): Likewise.
7073 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
7074 (neon_mov<mode>): Likewise.
7075 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
7077 (thumb2_movsi_insn): Likewise.
7078 (thumb2_cmpsi_neg_shiftsi): Likewise.
7079 (thumb2_extendqisi_v6): Likewise.
7080 (thumb2_zero_extendhisi2_v6): Likewise.
7081 (thumb2_zero_extendqisi2_v6): Likewise.
7082 (thumb2_shiftsi3_short): Likewise.
7083 (thumb2_addsi3_compare0_scratch): Likewise.
7084 (orsi_not_shiftsi_si): Likewise.
7085 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7086 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
7088 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7089 (1020alu_shift_op): Likewise.
7090 (1020alu_shift_reg_op): Likewise.
7091 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7092 (alu_shift_op): Likewise.
7093 (alu_shift_reg_op): Likewise.
7094 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7095 (11_alu_shift_op): Likewise.
7096 (11_alu_shift_reg_op): Likewise.
7097 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7098 (9_alu_shift_reg_op): Likewise.
7099 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7101 (cortex_a15_alu_shift): Likewise.
7102 (cortex_a15_alu_shift_reg): Likewise.
7103 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
7105 (cortex_a5_alu_shift): Likewise.
7106 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
7108 (cortex_a53_alu_shift): Likewise.
7109 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7111 (cortex_a7_alu_reg): Likewise.
7112 (cortex_a7_alu_shift): Likewise.
7113 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
7115 (cortex_a8_alu_shift): Likewise.
7116 (cortex_a8_alu_shift_reg): Likewise.
7117 (cortex_a8_mov): Likewise.
7118 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7119 (cortex_a9_dp_shift): Likewise.
7120 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
7122 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
7124 (cortex_r4_mov): Likewise.
7125 (cortex_r4_alu_shift): Likewise.
7126 (cortex_r4_alu_shift_reg): Likewise.
7127 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7128 (526_alu_shift_op): Likewise.
7129 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7130 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7131 (626te_alu_shift_op): Likewise.
7132 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
7133 (726te_alu_op): Likewise.
7134 (726te_alu_shift_op): Likewise.
7135 (726te_alu_shift_reg_op): Likewise.
7136 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7137 (mp626_alu_shift_op): Likewise.
7138 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
7139 (pj4_alu_e1_conds): Likewise.
7140 (pj4_alu): Likewise.
7141 (pj4_alu_conds): Likewise.
7142 (pj4_shift): Likewise.
7143 (pj4_shift_conds): Likewise.
7144 (pj4_alu_shift): Likewise.
7145 (pj4_alu_shift_conds): Likewise.
7146 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7148 (cortexa7_older_only): Likewise.
7149 (cortexa7_younger): Likewise.
7151 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7153 * ipa-pure-const.c (generate_summary): Rename to...
7154 (pure_const_generate_summary): ... this.
7156 2013-07-17 Iain Sandoe <iain@codesourcery.com>
7158 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
7160 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
7163 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
7166 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7168 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
7169 enabled without -march=zEC12.
7170 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
7173 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7175 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
7176 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
7177 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
7178 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
7179 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
7180 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
7181 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
7183 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
7184 and ISA_HAS_NMADD3_NMSUB3.
7185 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
7186 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
7187 (nmsub4<mode>, nmsub3<mode>): Likewise.
7188 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
7190 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7192 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
7193 TARGET_MIPS5400 checking.
7195 2013-07-16 Jakub Jelinek <jakub@redhat.com>
7196 Peter Bergner <bergner@vnet.ibm.com>
7198 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
7199 registers in the comment.
7200 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
7201 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
7202 rather than FIRST_PSEUDO_REGISTERS.
7204 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
7206 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
7207 enable extra ISA flags with TARGET_HTM.
7209 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7211 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
7214 2013-07-15 Cong Hou <congh@google.com>
7216 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
7217 in compare function for sorting.
7219 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
7221 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
7222 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
7223 * config/rs6000/rs6000.opt: Add -mhtm option.
7224 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
7225 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
7226 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7227 __HTM__ if the HTM instructions are available.
7228 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
7229 htm_spr_reg_operand): New define_predicates.
7230 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
7231 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
7233 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
7234 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
7235 HTM builtin functions.
7236 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
7237 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
7238 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
7239 (rs6000_builtin_mask_calculate): Likewise.
7240 (rs6000_option_override_internal): Likewise.
7241 (bdesc_htm): Add new HTM builtin support.
7242 (htm_spr_num): New function.
7243 (htm_spr_regno): Likewise.
7244 (rs6000_htm_spr_icode): Likewise.
7245 (htm_expand_builtin): Likewise.
7246 (htm_init_builtins): Likewise.
7247 (rs6000_expand_builtin): Add support for HTM builtin functions.
7248 (rs6000_init_builtins): Likewise.
7249 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
7251 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
7252 (TARGET_HTM, MASK_HTM): Define macros.
7253 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
7254 (FIXED_REGISTERS): Likewise.
7255 (CALL_USED_REGISTERS): Likewise.
7256 (CALL_REALLY_USED_REGISTERS): Likewise.
7257 (REG_ALLOC_ORDER): Likewise.
7258 (enum reg_class): Likewise.
7259 (REG_CLASS_NAMES): Likewise.
7260 (REG_CLASS_CONTENTS): Likewise.
7261 (REGISTER_NAMES): Likewise.
7262 (ADDITIONAL_REGISTER_NAMES): Likewise.
7263 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
7264 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
7265 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
7266 * config/rs6000/htm.md: New file.
7267 * config/rs6000/htmintrin.h: New file.
7268 * config/rs6000/htmxlintrin.h: New file.
7270 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
7272 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7273 Define SYMBOL_TINY_GOT, update comment.
7274 * config/aarch64/aarch64.c
7275 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
7276 (aarch64_expand_mov_immediate): Likewise.
7277 (aarch64_print_operand): Likewise.
7278 (aarch64_classify_symbol): Likewise.
7279 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
7280 (ldr_got_tiny): Define.
7282 2013-07-13 Tobias Grosser <tobias@grosser.es>
7284 PR tree-optimization/54094
7285 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
7286 scheduling dimension for the parallelism check from the polyhedral
7287 information in the AST.
7288 * graphite-dependences.c (carries_deps): Do not assume the schedule is
7291 2013-07-13 Jason Merrill <jason@redhat.com>
7293 * print-tree.c (debug_vec_tree): Use debug_raw.
7294 (debug_raw (vec<tree, va_gc> &)): New.
7295 (debug_raw (vec<tree, va_gc> *)): New.
7296 * tree.h: Declare them.
7298 2013-07-13 Bin Cheng <bin.cheng@arm.com>
7300 * ifcvt.c (ifcvt_after_combine): New static variable.
7301 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
7303 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
7304 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
7305 rest_of_handle_if_after_reload): Pass new argument for if_convert.
7307 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
7309 * config/mips/mips.c (mips_expand_call): Remove empty statement.
7311 2013-07-12 Michael Matz <matz@suse.de>
7314 * convert.c (convert_to_real): Reject non-float inner types.
7316 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
7318 * config/aarch64/aarch64-protos.h
7319 (aarch64_simd_immediate_valid_for_move): Remove.
7320 * config/aarch64/aarch64.c (simd_immediate_info): New member.
7321 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
7323 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
7325 2013-07-11 Steve Ellcey <sellcey@mips.com>
7327 * config/mips/mips.c (mips_conditional_register_usage): Do not
7328 use t[0-7] registers in MIPS16 mode when optimizing for size.
7330 2013-07-11 Sriraman Tallam <tmsriram@google.com>
7332 * config/i386/i386.c (dispatch_function_versions): Fix array
7333 indexing of function_version_info to match actual_versions.
7335 2013-07-11 Teresa Johnson <tejohnson@google.com>
7337 * vec.h (struct va_gc): Move release out-of-line.
7338 (va_gc::release): Call ggc_free on released vec.
7340 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7342 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
7343 Require GOT register as additional operand in UNSPEC.
7344 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
7345 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7346 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7347 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
7348 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
7349 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
7350 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
7352 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
7355 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
7356 name seen by assembler/linker rather if available.
7358 2013-07-11 Andreas Schwab <schwab@suse.de>
7360 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
7362 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
7364 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
7366 2013-07-10 Joseph Myers <joseph@codesourcery.com>
7368 * doc/tm.texi.in: Move hook documentation to ....
7369 * target.def: ... here.
7371 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
7373 * doc/tm.texi: Regenerate.
7375 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
7378 * doc/invoke.texi: Document -Wconditionally-supported.
7380 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7383 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
7386 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7389 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
7390 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
7391 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
7392 Remove duplicate devices.
7393 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
7394 * config/avr/t-multilib: Regenerate.
7395 * config/avr/avr-tables.opt: Regenerate.
7396 * doc/avr-mmcu.texi: Regenerate.
7398 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7401 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
7403 2013-07-10 Graham Stott <graham.stott@btinternet.com>
7405 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
7406 the cost of MULT when optimizing for size.
7408 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7410 * config/cr16/cr16-protos.h: Don't include target.h.
7412 2013-07-09 Joseph Myers <joseph@codesourcery.com>
7414 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
7415 adjust register size for TDmode and TFmode for VSX registers.
7417 2013-07-08 Kai Tietz <ktietz@redhat.com>
7420 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
7421 hook_bool_const_tree_true.
7423 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7425 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
7426 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
7427 * config/s390/s390.md: Define FPR*_REGNUM constants.
7428 Fix FPR2_REGNUM constant (18 -> 17).
7429 ("*trunc<BFP:mode><DFP_ALL:mode>2")
7430 ("*trunc<DFP_ALL:mode><BFP:mode>2")
7431 ("trunc<BFP:mode><DFP_ALL:mode>2")
7432 ("trunc<DFP_ALL:mode><BFP:mode>2")
7433 ("*extend<BFP:mode><DFP_ALL:mode>2")
7434 ("*extend<DFP_ALL:mode><BFP:mode>2")
7435 ("extend<BFP:mode><DFP_ALL:mode>2")
7436 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
7439 2013-07-08 Graham Stott <graham.stott@btinternet.com>
7441 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
7443 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7445 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
7446 and cfun_fpr_bit_p to cfun_fpr_save_p.
7447 (s390_frame_area, s390_register_info, s390_frame_info)
7448 (s390_emit_prologue, s390_emit_epilogue)
7449 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
7451 * config/s390/s390.h: Define *_REGNUM macros for floating point
7454 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
7456 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
7458 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
7460 PR rtl-optimization/57786
7461 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
7462 and break out of the loop when it is set to false.
7464 2013-07-08 Jakub Jelinek <jakub@redhat.com>
7467 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7468 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7469 (const_int 63)) 0)).
7470 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7471 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7472 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7474 PR rtl-optimization/57829
7475 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7476 mask bits outside of mode are just sign-extension from mode to HWI.
7478 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7480 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7481 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7482 adjust_address instead of change_address to keep info about alignment.
7483 (emit_strmov): Remove.
7484 (emit_memmov): New function.
7485 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7486 (expand_movmem_epilogue): Likewise and return updated rtx for
7488 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7489 destination and source.
7490 (decide_alignment): Refactor, handle vector_loop.
7491 (ix86_expand_movmem): Likewise.
7492 (ix86_expand_setmem): Likewise.
7493 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7495 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
7497 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7498 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7500 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7502 * config/i386/sse.md (sse_movlhps): Change alternative 3
7503 of operand 2 to "m".
7505 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7508 * config/i386/sse.md (iptr): New mode attribute.
7509 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7510 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7511 (<sse>_vmmul<mode>3): Ditto.
7512 (<sse>_vmdiv<mode>3): Ditto.
7513 (sse_vmrcpv4sf2): Ditto.
7514 (<sse>_vmsqrt<mode>2): Ditto.
7515 (sse_vmrsqrtv4sf2): Ditto.
7516 (<sse>_vm<code><mode>3): Ditto.
7517 (avx_vmcmp<mode>3): Ditto.
7518 (<sse>_vmmaskcmp<mode>3): Ditto.
7519 (<sse>_comi): Ditto.
7520 (<sse>_ucomi): Ditto.
7521 (*xop_vmfrcz_<mode>): Ditto.
7522 (*fmai_fmadd_<mode>): Ditto.
7523 (*fmai_fmsub_<mode>): Ditto.
7524 (*fmai_fnmadd_<mode>): Ditto.
7525 (*fmai_fnmsub_<mode>): Ditto.
7526 (*fma4i_vmfmadd_<mode>): Ditto.
7527 (*fma4i_vmfmsub_<mode>): Ditto.
7528 (*fma4i_vmfnmadd_<mode>): Ditto.
7529 (*fma4i_vmfnmsub_<mode>): Ditto.
7530 (*xop_vmfrcz_<mode>): Ditto.
7531 (sse_cvtps2pi): Ditto.
7532 (sse_cvttps2pi): Ditto.
7533 (sse_cvtss2si): Ditto.
7534 (sse_cvtss2si_2): Ditto.
7535 (sse_cvtss2siq_2): Ditto.
7536 (sse_cvttss2si): Ditto.
7537 (sse_cvttss2siq): Ditto.
7538 (sse_cvtsd2si): Ditto.
7539 (sse_cvtsd2si_2): Ditto.
7540 (sse_cvtsd2siq_2): Ditto.
7541 (sse_cvttsd2si): Ditto.
7542 (sse_cvttsd2siq): Ditto.
7543 (sse_cvtsd2ss): Ditto.
7544 (sse_cvtss2sd): Ditto.
7545 (avx2_pbroadcast<mode>): Ditto.
7546 (avx2_pbroadcast<mode>_1): Ditto.
7547 (*avx_vperm_broadcast_v4sf): Ditto.
7549 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7550 (sse_movlhps): Ditto.
7551 (sse_storehps): Ditto.
7552 (sse_loadhps): Ditto.
7553 (sse_storelps): Ditto.
7554 (sse_loadlps): Ditto.
7555 (*vec_concatv4sf): Ditto.
7556 (*vec_interleave_highv2df): Ditto.
7557 (*vec_interleave_lowv2df): Ditto.
7558 (*vec_extractv2df_1_sse): Ditto.
7559 (*vec_extractv2df_0_sse): Ditto.
7560 (sse2_storelpd): Ditto.
7561 (sse2_loadlpd): Ditto.
7562 (sse2_movsd): Ditto.
7563 (*vec_concatv4si): Ditto.
7564 (vec_concatv2di): Ditto.
7566 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
7567 for Intel asm dialect.
7568 (mmx_punpcklwd): Ditto.
7569 (mmx_punpckldq): Ditto.
7571 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
7572 for intel assembler dialect.
7574 2013-07-06 Jakub Jelinek <jakub@redhat.com>
7577 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
7578 for BUILT_IN_C{LZ,LRSB}*.
7579 * tree.h (CASE_INT_FN): Add FN##IMAX case.
7580 * tree-vrp.c (extract_range_basic): Handle
7581 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
7582 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
7583 fall thru to code calling set_value*.
7584 * builtins.c (expand_builtin): Remove *IMAX cases.
7585 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
7586 if width is bigger than 2*HWI.
7588 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
7590 PR rtl-optimization/55342
7591 * lra-int.h (lra_subreg_reload_pseudos): New.
7592 * lra.c: Add undoing optional reloads to the block diagram.
7593 (lra_subreg_reload_pseudos): New.
7594 (lra_optional_reload_pseudos): Change comments.
7595 (lra): Init and clear lra_subreg_reload_pseudos. Clear
7596 lra_optional_reload_pseudos after undo transformations.
7597 * lra-assigns.c (pseudo_prefix_title): New.
7598 (lra_setup_reg_renumber): Use it.
7599 (spill_for): Ditto. Check subreg reload pseudos too.
7600 (assign_by_spills): Consider subreg reload pseudos too.
7601 * lra-constraints.c (simplify_operand_subreg): Use
7602 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
7603 (curr_insn_transform): Recognize and do optional reloads.
7604 (undo_optional_reloads): New.
7605 (lra_undo_inheritance): Call undo_optional_reloads.
7607 2013-07-05 Thomas Quinot <quinot@adacore.com>
7609 * tree-complex.c (expand_complex_operations_1): Fix typo.
7611 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
7613 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
7614 (tune_params): New member 'const vec_costs'.
7615 * config/aarch64/aarch64.c (generic_vector_cost): New.
7616 (generic_tunings): New member 'generic_vector_cost'.
7617 (aarch64_builtin_vectorization_cost): New.
7618 (aarch64_add_stmt_cost): New.
7619 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7620 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7622 2013-07-03 Jakub Jelinek <jakub@redhat.com>
7625 * config/i386/predicates.md (vsib_address_operand): Disallow
7626 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7628 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7631 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7632 expand_builtin_setjmp_receiver.
7633 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7634 with NULL for the label parameter.
7635 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7636 the frame-pointer. Adjust comments.
7637 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7638 only if LABEL is non-NULL.
7640 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7642 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7643 (ARM_ABI_AAPCS64): Ditto.
7645 (ARM_DEFAULT_ABI): Ditto.
7647 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7649 * config/aarch64/aarch64-builtins.c
7650 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7651 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7653 * config/aarch64/aarch64-simd.md
7654 (aarch64_ld1<VALL:mode>): New.
7655 (aarch64_st1<VALL:mode>): Likewise.
7656 * config/aarch64/arm_neon.h
7657 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7659 2013-07-02 Sriraman Tallam <tmsriram@google.com>
7661 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7663 (ix86_option_override_internal):Turn on all -mavx target flags by
7664 default as they are dependent on AVX anyway.
7666 2013-07-02 Cary Coutant <ccoutant@google.com>
7668 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7670 (loc_checksum_ordered): Likewise.
7671 (hash_loc_operands): Remove inline keyword.
7673 2013-07-02 Jakub Jelinek <jakub@redhat.com>
7675 PR tree-optimization/57741
7676 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7677 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7678 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7679 Allow REAL_CST step_exprs if flag_associative_math.
7680 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7682 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7684 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
7686 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7688 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7690 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7692 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7694 (iorsi3_insn): Likewise.
7695 (arm_xorsi3): Likewise.
7697 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7699 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7701 (attribute "type"): ... here, and prefix with "wmmx_".
7702 (attribute "core_cycles"): Update for attribute changes.
7703 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7704 (tbcstv4hi): Likewise.
7705 (tbcstv2si): Likewise.
7706 (iwmmxt_iordi3): Likewise.
7707 (iwmmxt_xordi3): Likewise.
7708 (iwmmxt_anddi3): Likewise.
7709 (iwmmxt_nanddi3): Likewise.
7710 (iwmmxt_arm_movdi): Likewise.
7711 (iwmmxt_movsi_insn): Likewise.
7712 (mov<mode>_internal): Likewise.
7713 (and<mode>3_iwmmxt): Likewise.
7714 (ior<mode>3_iwmmxt): Likewise.
7715 (xor<mode>3_iwmmxt): Likewise.
7716 (add<mode>3_iwmmxt): Likewise.
7717 (ssaddv8qi3): Likewise.
7718 (ssaddv4hi3): Likewise.
7719 (ssaddv2si3): Likewise.
7720 (usaddv8qi3): Likewise.
7721 (usaddv4hi3): Likewise.
7722 (usaddv2si3): Likewise.
7723 (sub<mode>3_iwmmxt): Likewise.
7724 (sssubv8qi3): Likewise.
7725 (sssubv4hi3): Likewise.
7726 (sssubv2si3): Likewise.
7727 (ussubv8qi3): Likewise.
7728 (ussubv4hi3): Likewise.
7729 (ussubv2si3): Likewise.
7730 (mulv4hi3_iwmmxt): Likewise.
7731 (smulv4hi3_highpart): Likewise.
7732 (umulv4hi3_highpart): Likewise.
7733 (iwmmxt_wmacs): Likewise.
7734 (iwmmxt_wmacsz): Likewise.
7735 (iwmmxt_wmacu): Likewise.
7736 (iwmmxt_wmacuz): Likewise.
7737 (iwmmxt_clrdi): Likewise.
7738 (iwmmxt_clrv8qi): Likewise.
7739 (iwmmxt_clr4hi): Likewise.
7740 (iwmmxt_clr2si): Likewise.
7741 (iwmmxt_uavgrndv8qi3): Likewise.
7742 (iwmmxt_uavgrndv4hi3): Likewise.
7743 (iwmmxt_uavgv8qi3): Likewise.
7744 (iwmmxt_uavgv4hi3): Likewise.
7745 (iwmmxt_tinsrb): Likewise.
7746 (iwmmxt_tinsrh): Likewise.
7747 (iwmmxt_tinsrw): Likewise.
7748 (iwmmxt_textrmub): Likewise.
7749 (iwmmxt_textrmsb): Likewise.
7750 (iwmmxt_textrmuh): Likewise.
7751 (iwmmxt_textrmsh): Likewise.
7752 (iwmmxt_textrmw): Likewise.
7753 (iwmxxt_wshufh): Likewise.
7754 (eqv8qi3): Likewise.
7755 (eqv4hi3): Likewise.
7756 (eqv2si3): Likewise.
7757 (gtuv8qi3): Likewise.
7758 (gtuv4hi3): Likewise.
7759 (gtuv2si3): Likewise.
7760 (gtv8qi3): Likewise.
7761 (gtv4hi3): Likewise.
7762 (gtv2si3): Likewise.
7763 (smax<mode>3_iwmmxt): Likewise.
7764 (umax<mode>3_iwmmxt): Likewise.
7765 (smin<mode>3_iwmmxt): Likewise.
7766 (umin<mode>3_iwmmxt): Likewise.
7767 (iwmmxt_wpackhss): Likewise.
7768 (iwmmxt_wpackwss): Likewise.
7769 (iwmmxt_wpackdss): Likewise.
7770 (iwmmxt_wpackhus): Likewise.
7771 (iwmmxt_wpackwus): Likewise.
7772 (iwmmxt_wpackdus): Likewise.
7773 (iwmmxt_wunpckihb): Likewise.
7774 (iwmmxt_wunpckihh): Likewise.
7775 (iwmmxt_wunpckihw): Likewise.
7776 (iwmmxt_wunpckilb): Likewise.
7777 (iwmmxt_wunpckilh): Likewise.
7778 (iwmmxt_wunpckilw): Likewise.
7779 (iwmmxt_wunpckehub): Likewise.
7780 (iwmmxt_wunpckehuh): Likewise.
7781 (iwmmxt_wunpckehuw): Likewise.
7782 (iwmmxt_wunpckehsb): Likewise.
7783 (iwmmxt_wunpckehsh): Likewise.
7784 (iwmmxt_wunpckehsw): Likewise.
7785 (iwmmxt_wunpckelub): Likewise.
7786 (iwmmxt_wunpckeluh): Likewise.
7787 (iwmmxt_wunpckeluw): Likewise.
7788 (iwmmxt_wunpckelsb): Likewise.
7789 (iwmmxt_wunpckelsh): Likewise.
7790 (iwmmxt_wunpckelsw): Likewise.
7791 (ror<mode>3): Likewise.
7792 (ashr<mode>3_iwmmxt): Likewise.
7793 (lshr<mode>3_iwmmxt): Likewise.
7794 (ashl<mode>3_iwmmxt): Likewise.
7795 (ror<mode>3_di): Likewise.
7796 (ashr<mode>3_di): Likewise.
7797 (lshr<mode>3_di): Likewise.
7798 (ashl<mode>3_di): Likewise.
7799 (iwmmxt_wmadds): Likewise.
7800 (iwmmxt_wmaddu): Likewise.
7801 (iwmmxt_tmia): Likewise.
7802 (iwmmxt_tmiaph): Likewise.
7803 (iwmmxt_tmiabb): Likewise.
7804 (iwmmxt_tmiatb): Likewise.
7805 (iwmmxt_tmiabt): Likewise.
7806 (iwmmxt_tmiatt): Likewise.
7807 (iwmmxt_tmovmskb): Likewise.
7808 (iwmmxt_tmovmskh): Likewise.
7809 (iwmmxt_tmovmskw): Likewise.
7810 (iwmmxt_waccb): Likewise.
7811 (iwmmxt_wacch): Likewise.
7812 (iwmmxt_waccw): Likewise.
7813 (iwmmxt_waligni): Likewise.
7814 (iwmmxt_walignr): Likewise.
7815 (iwmmxt_walignr0): Likewise.
7816 (iwmmxt_walignr1): Likewise.
7817 (iwmmxt_walignr2): Likewise.
7818 (iwmmxt_walignr3): Likewise.
7819 (iwmmxt_wsadb): Likewise.
7820 (iwmmxt_wsadh): Likewise.
7821 (iwmmxt_wsadbz): Likewise.
7822 (iwmmxt_wsadhz): Likewise.
7823 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
7824 (iwmmxt_wabsdiffb): Likewise.
7825 (iwmmxt_wabsdiffh): Likewise.
7826 (iwmmxt_wabsdiffw): Likewise.
7827 (iwmmxt_waddsubhx): Likewise
7828 (iwmmxt_wsubaddhx): Likewise.
7829 (addc<mode>3): Likewise.
7830 (iwmmxt_avg4): Likewise.
7831 (iwmmxt_avg4r): Likewise.
7832 (iwmmxt_wmaddsx): Likewise.
7833 (iwmmxt_wmaddux): Likewise.
7834 (iwmmxt_wmaddsn): Likewise.
7835 (iwmmxt_wmaddun): Likewise.
7836 (iwmmxt_wmulwsm): Likewise.
7837 (iwmmxt_wmulwum): Likewise.
7838 (iwmmxt_wmulsmr): Likewise.
7839 (iwmmxt_wmulumr): Likewise.
7840 (iwmmxt_wmulwsmr): Likewise.
7841 (iwmmxt_wmulwumr): Likewise.
7842 (iwmmxt_wmulwl): Likewise.
7843 (iwmmxt_wqmulm): Likewise.
7844 (iwmmxt_wqmulwm): Likewise.
7845 (iwmmxt_wqmulmr): Likewise.
7846 (iwmmxt_wqmulwmr): Likewise.
7847 (iwmmxt_waddbhusm): Likewise.
7848 (iwmmxt_waddbhusl): Likewise.
7849 (iwmmxt_wqmiabb): Likewise.
7850 (iwmmxt_wqmiabt): Likewise.
7851 (iwmmxt_wqmiatb): Likewise.
7852 (iwmmxt_wqmiatt): Likewise.
7853 (iwmmxt_wqmiabbn): Likewise.
7854 (iwmmxt_wqmiabtn): Likewise.
7855 (iwmmxt_wqmiatbn): Likewise.
7856 (iwmmxt_wqmiattn): Likewise.
7857 (iwmmxt_wmiabb): Likewise.
7858 (iwmmxt_wmiabt): Likewise.
7859 (iwmmxt_wmiatb): Likewise.
7860 (iwmmxt_wmiatt): Likewise.
7861 (iwmmxt_wmiabbn): Likewise.
7862 (iwmmxt_wmiabtn): Likewise.
7863 (iwmmxt_wmiatbn): Likewise.
7864 (iwmmxt_wmiattn): Likewise.
7865 (iwmmxt_wmiawbb): Likewise.
7866 (iwmmxt_wmiawbt): Likewise.
7867 (iwmmxt_wmiawtb): Likewise.
7868 (iwmmxt_wmiawtt): Likewise.
7869 (iwmmxt_wmiawbbn): Likewise.
7870 (iwmmxt_wmiawbtn): Likewise.
7871 (iwmmxt_wmiawtbn): Likewise.
7872 (iwmmxt_wmiawttn): Likewise.
7873 (iwmmxt_wmerge): Likewise.
7874 (iwmmxt_tandc<mode>3): Likewise.
7875 (iwmmxt_torc<mode>3): Likewise.
7876 (iwmmxt_torvsc<mode>3): Likewise.
7877 (iwmmxt_textrc<mode>3): Likewise.
7878 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
7879 (wmmxt_pack): Likewise.
7880 (wmmxt_mult_c1): Likewise.
7881 (wmmxt_mult_c2): Likewise.
7882 (wmmxt_alu_c1): Likewise.
7883 (wmmxt_alu_c2): Likewise.
7884 (wmmxt_alu_c3): Likewise.
7885 (wmmxt_transfer_c1): Likewise.
7886 (wmmxt_transfer_c2): Likewise.
7887 (wmmxt_transfer_c3): Likewise.
7888 (marvell_f_iwmmxt_wstr): Likewise.
7889 (marvell_f_iwmmxt_wldr): Likewise.
7891 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
7893 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
7895 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7898 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7899 * lra-constraints.c (need_for_split_p): Check call used hard regs
7900 living through calls.
7902 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
7903 call used regs for call insn.
7905 2013-06-28 Jakub Jelinek <jakub@redhat.com>
7908 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
7909 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
7911 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7913 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
7915 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7917 * lra-constraints.c (need_for_split_p): Check call used hard regs
7918 living through calls.
7920 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
7923 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
7924 to tie with any other modes. Eliminate Altivec vector mode tests,
7925 since these are a subset of ALTIVEC or VSX vector modes. Simplify
7926 code, to return 0 if testing MODE2 for a condition, if we've
7927 already tested MODE1 for the same condition.
7929 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7931 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
7934 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7936 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7937 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
7939 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7941 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
7943 (aarch64_symbolic_constant_p): Remove.
7944 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
7945 static. Fix line length and white space.
7946 (aarch64_symbolic_constant_p): Remove.
7947 * config/aarch64/predicates.md (aarch64_valid_symref):
7948 Use aarch64_classify_symbol_expression.
7950 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7952 * config/arm/constraints.md (Ts): New constraint.
7953 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
7955 (compare_scc): Use "Ts" constraint for operand 0.
7956 (ior_scc_scc): Likewise.
7957 (and_scc_scc): Likewise.
7958 (and_scc_scc_nodom): Likewise.
7959 (ior_scc_scc_cmp): Likewise for operand 7.
7960 (and_scc_scc_cmp): Likewise.
7961 * config/arm/thumb2.md (thumb2_movsi_insn):
7962 Add alternatives for 16-bit encodings.
7963 (thumb2_movhi_insn): Likewise.
7964 (thumb2_movsicc_insn): Likewise.
7965 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
7966 (thumb2_negscc): Use "Ts" constraint.
7967 Move mvn instruction outside cond_exec block.
7968 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
7969 for 16-bit encodings.
7971 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7973 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
7975 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
7976 (mulsi3subsi): Likewise.
7977 (mulsidi3adddi): Likewise.
7978 (mulsidi3_v6): Likewise.
7979 (umulsidi3_v6): Likewise.
7980 (umulsidi3adddi_v6): Likewise.
7981 (smulsi3_highpart_v6): Likewise.
7982 (umulsi3_highpart_v6): Likewise.
7983 (mulhisi3tb): Likewise.
7984 (mulhisi3bt): Likewise.
7985 (mulhisi3tt): Likewise.
7986 (maddhisi4): Likewise.
7987 (maddhisi4tb): Likewise.
7988 (maddhisi4tt): Likewise.
7989 (maddhidi4): Likewise.
7990 (maddhidi4tb): Likewise.
7991 (maddhidi4tt): Likewise.
7992 (zeroextractsi_compare0_scratch): Likewise.
7993 (insv_zero): Likewise.
7994 (insv_t2): Likewise.
7995 (anddi_notzesidi_di): Likewise.
7996 (anddi_notsesidi_di): Likewise.
7997 (andsi_notsi_si): Likewise.
7998 (iordi_zesidi_di): Likewise.
7999 (xordi_zesidi_di): Likewise.
8000 (andsi_iorsi3_notsi): Likewise.
8002 (smax_m1): Likewise.
8004 (not_shiftsi): Likewise.
8005 (unaligned_loadsi): Likewise.
8006 (unaligned_loadhis): Likewise.
8007 (unaligned_loadhiu): Likewise.
8008 (unaligned_storesi): Likewise.
8009 (unaligned_storehi): Likewise.
8010 (extv_reg): Likewise.
8011 (extzv_t2): Likewise.
8013 (udivsi3): Likewise.
8014 (arm_zero_extendhisi2addsi): Likewise.
8015 (arm_zero_extendqisi2addsi): Likewise.
8016 (compareqi_eq0): Likewise.
8017 (arm_extendhisi2_v6): Likewise.
8018 (arm_extendqisi2addsi): Likewise.
8019 (arm_movt): Likewise.
8020 (thumb2_ldrd): Likewise.
8021 (thumb2_ldrd_base): Likewise.
8022 (thumb2_ldrd_base_neg): Likewise.
8023 (thumb2_strd): Likewise.
8024 (thumb2_strd_base): Likewise.
8025 (thumb2_strd_base_neg): Likewise.
8026 (arm_negsi2): Add alternative for 16-bit encoding.
8027 (arm_one_cmplsi2): Likewise.
8029 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8031 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
8032 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
8033 (movdfcc): Likewise.
8034 * config/arm/vfp.md (*thumb2_movsf_vfp):
8035 Disable predication for arm_restrict_it.
8036 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
8037 (*thumb2_movdfcc_vfp): Likewise.
8038 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
8039 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
8040 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
8041 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
8042 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
8043 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
8044 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
8045 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
8046 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
8047 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
8048 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
8049 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
8050 Disable predication for arm_restrict_it.
8052 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
8054 * config/i386/bmiintrin.h (_bextr_u32): New.
8055 (_bextr_u64): Ditto.
8057 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
8059 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
8060 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
8061 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
8062 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
8063 * config/mips/n32-elf.h: ...this new file.
8065 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
8068 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
8069 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
8071 2013-06-27 Catherine Moore <clm@codesourcery.com>
8073 * config/mips/mips-tables.opt: Regenerate.
8074 * config/mips/mips-cpus.def: Add m14ke and m14kec.
8075 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8076 * doc/invoke.texi: Add -m14kc.
8078 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8081 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
8082 constraints of operand 1 and 2.
8085 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
8086 to match RTL canonicalization. Swap predicates and
8087 constraints of operand 1 and 2.
8089 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
8091 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
8092 Process OP_INOUT regs for splitting too.
8094 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8096 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
8097 decl before the loop, initialize to NULL.
8098 (vectorizable_load): Initialize ptr_incr to NULL.
8100 2013-06-27 Martin Jambor <mjambor@suse.cz>
8103 * ipa-ref.h (ipa_maybe_record_reference): Declare.
8104 * ipa-ref.c (ipa_maybe_record_reference): New function.
8105 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
8106 * ipa-cp.c (create_specialized_node): Record potential references from
8108 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
8110 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8112 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
8113 parameter 'mode' of type 'enum machine_mode mode'; change to pass
8114 'mode' to force_reg.
8115 (aarch64_add_offset): Update calls to aarch64_force_temporary.
8116 (aarch64_expand_mov_immediate): Likewise.
8118 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8120 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
8121 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
8123 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8125 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
8126 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8127 (struct machine_function): Add tbegin_p.
8128 (s390_canonicalize_comparison): Fold CC mode compares to
8129 conditional jump if possible.
8130 (s390_emit_jump): Return the emitted jump.
8131 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
8132 Handle CCRAWmode compares.
8133 (s390_option_override): Default to -mhtm if available.
8134 (s390_reg_clobbered_rtx): Handle floating point regs as well.
8135 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
8136 FPRs instead of df_regs_ever_live_p.
8137 (s390_optimize_nonescaping_tx): New function.
8138 (s390_init_frame_layout): Extend clobbered_regs array to cover
8140 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
8141 (s390_expand_tbegin): New function.
8142 (enum s390_builtin): New enum definition.
8143 (code_for_builtin): New array definition.
8144 (s390_init_builtins): New function.
8145 (s390_expand_builtin): New function.
8146 (TARGET_INIT_BUILTINS): Define.
8147 (TARGET_EXPAND_BUILTIN): Define.
8148 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
8149 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
8150 (s390_alc_comparison): Likewise.
8151 * config/s390/s390-modes.def: Add CCRAWmode.
8152 * config/s390/s390.h (processor_flags): Add PF_TX.
8153 (TARGET_CPU_HTM): Define macro.
8154 (TARGET_HTM): Define macro.
8155 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
8156 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
8157 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8158 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
8159 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
8160 (TBEGIN_MASK, TBEGINC_MASK): New constants.
8161 ("*cc_to_int"): Move up.
8162 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
8163 constants other than 0.
8164 ("*ccraw_to_int"): New insn and splitter definition.
8165 ("tbegin", "tbegin_nofloat", "tbegin_retry")
8166 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
8167 ("tx_assist"): New expander.
8168 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
8169 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
8170 * config/s390/s390.opt: Add -mhtm option.
8171 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
8172 * config/s390/htmxlintrin.h: New file.
8173 * config/s390/htmintrin.h: New file.
8174 * config/s390/s390intrin.h: New file.
8175 * doc/extend.texi: Document htm builtins.
8176 * config.gcc: Add the new header files to extra_headers.
8178 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
8180 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
8181 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
8183 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
8184 Pat Haugen <pthaugen@us.ibm.com>
8185 Peter Bergner <bergner@vnet.ibm.com>
8187 * config/rs6000/power8.md: New.
8188 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
8189 setting for power8 entry.
8190 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
8191 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
8192 test for Power4/Power5 only.
8193 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
8195 (force_new_group): Adjust comment.
8196 * config/rs6000/rs6000.md: Include power8.md.
8198 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
8200 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
8201 * config/arm/arm-protos.h (arm_max_conditional_execute): New
8203 (tune_params): Update comment.
8204 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
8205 (arm_max_conditional_execute): New function.
8206 (thumb2_final_prescan_insn): Use max_insn_skipped and
8207 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
8209 2013-06-25 Jakub Jelinek <jakub@redhat.com>
8211 PR tree-optimization/57705
8212 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
8213 SSA_NAME step, provided that it is not defined inside the loop.
8214 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
8215 (get_initial_def_for_induction): Handle SSA_NAME IV step.
8217 2013-06-25 Martin Jambor <mjambor@suse.cz>
8220 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
8221 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
8223 (ipa_note_param_call): Initialize member_ptr flag.
8224 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
8225 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
8226 (ipa_write_indirect_edge_info): Stream member_ptr flag.
8227 (ipa_read_indirect_edge_info): Likewise.
8229 2013-06-25 Richard Biener <rguenther@suse.de>
8232 * passes.c (init_optimization_passes): Move pass_fold_builtins
8233 and pass_dce earlier with -Og.
8235 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
8237 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
8238 <BIT_FIELD_REF>: Remove trailing TAB.
8239 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
8242 2013-06-24 Martin Jambor <mjambor@suse.cz>
8244 PR tree-optimization/57358
8245 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
8246 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
8247 (ipa_analyze_params_uses): Generate pessimistic info when true.
8249 2013-06-24 Martin Jambor <mjambor@suse.cz>
8251 PR tree-optimization/57539
8252 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
8253 global.inlined_to of the new node to it. All callers changed.
8254 * ipa-inline-transform.c (clone_inlined_nodes): New variable
8255 inlining_into, pass it to cgraph_clone_node.
8256 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
8257 ipa_free_edge_args_substructures.
8258 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
8259 rdesc linked list. Do not assert rdesc edges have inlined caller.
8260 Assert we have found an rdesc in the rdesc list.
8262 2013-06-24 Richard Biener <rguenther@suse.de>
8264 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
8265 (pointer_set_lookup): Declare.
8266 (class pointer_map): New template class implementing a
8267 generic pointer to T map.
8268 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
8269 pointer_map<T>::contains, pointer_map<T>::insert,
8270 pointer_map<T>::traverse): New functions.
8271 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
8272 (pointer_set_lookup): New function.
8273 (pointer_set_contains): Use pointer_set_lookup.
8274 (pointer_set_insert): Likewise.
8275 (insert_aux): Remove.
8276 (struct pointer_map_t): Embed a pointer_set_t.
8277 (pointer_map_create): Adjust.
8278 (pointer_map_destroy): Likewise.
8279 (pointer_map_contains): Likewise.
8280 (pointer_map_insert): Likewise.
8281 (pointer_map_traverse): Likewise.
8282 * tree-streamer.h (struct streamer_tree_cache_d): Use a
8283 pointer_map<unsigned> instead of a pointer_map_t.
8284 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
8285 (streamer_tree_cache_lookup): Likewise.
8286 (streamer_tree_cache_create): Likewise.
8287 (streamer_tree_cache_delete): Likewise.
8288 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
8289 pointer_map<unsigned> instead of a pointer_map_t.
8290 (lto_init_tree_ref_encoder): Adjust.
8291 (lto_destroy_tree_ref_encoder): Likewise.
8292 * lto-section-out.c (lto_output_decl_index): Likewise.
8293 (lto_record_function_out_decl_state): Likewise.
8294 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
8296 2013-06-24 Richard Biener <rguenther@suse.de>
8298 PR tree-optimization/57488
8299 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
8301 2013-06-24 Alan Modra <amodra@gmail.com>
8303 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
8304 (gen_easy_altivec_constant): Likewise.
8305 * config/rs6000/predicates.md (easy_vector_constant_add_self,
8306 easy_vector_constant_msb): Likewise.
8308 2013-06-23 Jakub Jelinek <jakub@redhat.com>
8311 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
8312 add missing return true.
8314 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
8317 * config/sh/predicates.md (general_extend_operand): Invoke
8318 general_movsrc_operand for memory operands.
8319 (general_movsrc_operand): Allow reg+reg addressing, do not use
8320 general_operand for memory operands.
8322 2013-06-23 Sriraman Tallam <tmsriram@google.com>
8324 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
8325 when current target options does not apply.
8326 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
8327 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
8328 * config/i386/bmiintrin.h: Pass appropriate target
8329 attributes to header.
8330 * config/i386/mmintrin.h: Ditto.
8331 * config/i386/nmmintrin.h: Ditto.
8332 * config/i386/avx2intrin.h: Ditto.
8333 * config/i386/fxsrintrin.h: Ditto.
8334 * config/i386/tbmintrin.h: Ditto.
8335 * config/i386/xsaveintrin.h: Ditto.
8336 * config/i386/f16cintrin.h: Ditto.
8337 * config/i386/xtestintrin.h: Ditto.
8338 * config/i386/xsaveoptintrin.h: Ditto.
8339 * config/i386/bmi2intrin.h: Ditto.
8340 * config/i386/lzcntintrin.h: Ditto.
8341 * config/i386/smmintrin.h: Ditto.
8342 * config/i386/wmmintrin.h: Ditto.
8343 * config/i386/x86intrin.h: Remove all header include guards.
8344 * config/i386/prfchwintrin.h: Ditto.
8345 * config/i386/pmmintrin.h: Ditto.
8346 * config/i386/tmmintrin.h: Ditto.
8347 * config/i386/xmmintrin.h: Ditto.
8348 * config/i386/popcntintrin.h: Ditto.
8349 * config/i386/rdseedintrin.h: Ditto.
8350 * config/i386/ammintrin.h: Ditto.
8351 * config/i386/emmintrin.h: Ditto.
8352 * config/i386/immintrin.h: Remove all header include guards.
8353 * config/i386/fma4intrin.h: Ditto.
8354 * config/i386/lwpintrin.h: Ditto.
8355 * config/i386/xopintrin.h: Ditto.
8356 * config/i386/ia32intrin.h: Ditto.
8357 * config/i386/avxintrin.h: Ditto.
8358 * config/i386/rtmintrin.h: Ditto.
8359 * config/i386/fmaintrin.h: Ditto.
8360 * config/i386/mm3dnow.h: Ditto.
8362 2013-06-22 Sriraman Tallam <tmsriram@google.com>
8364 * common/config/i386/i386-common.c: Handle LZCNT.
8366 2013-06-22 Andi Kleen <ak@linux.intel.com>
8368 * doc/extend.texi: Use __atomic_store_n instead of
8369 __atomic_store in HLE example.
8371 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
8373 * config/sh/sh.c: Remove <cstdlib> workaround.
8375 2013-06-21 Andi Kleen <ak@linux.intel.com>
8377 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
8379 2013-06-21 Andi Kleen <ak@linux.intel.com>
8381 * doc/extend.texi: Document that __atomic_clear and
8382 __atomic_test_and_set should only be used with bool.
8384 2013-06-20 Jan Hubicka <jh@suse.cz>
8386 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
8388 * tree.c (decls_same_for_odr): New function.
8389 (same_for_edr): New function.
8390 (types_same_for_odr): New function.
8391 (get_binfo_at_offset): Use it.
8392 * tree.h (types_same_for_odr): Declare.
8394 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
8395 Jason Merrill <jason@redhat.com>
8397 * system.h: Include <cstdlib> as well as <stdlib.h>.
8399 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
8402 * config/i386/i386.c (construct_container): Report error if
8403 long double is used with disabled x87 float returns.
8405 2013-06-20 Jan Hubicka <jh@suse.cz>
8407 * lto-cgraph.c (input_symtab): Do not set cgraph state.
8409 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
8411 PR rtl-optimization/57425
8412 PR rtl-optimization/57569
8413 * alias.c (write_dependence_p): Remove parameters mem_mode and
8414 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
8415 Changed all callers.
8416 (canon_anti_dependence): Get comments and semantics in sync.
8417 Add parameter mem_canonicalized. Changed all callers.
8418 * rtl.h (canon_anti_dependence): Update prototype.
8420 2013-06-20 Richard Biener <rguenther@suse.de>
8422 * data-streamer-in.c (streamer_read_uhwi): Optimize single
8423 byte case, inline streamer_read_uchar and defer section
8426 2013-06-20 Richard Biener <rguenther@suse.de>
8428 PR tree-optimization/57584
8429 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
8430 SSA names into the expanded expression that take part in
8431 abnormal coalescing.
8433 2013-06-19 Sharad Singhai <singhai@google.com>
8435 * gcov.c (print_usage): Handle new option.
8436 (process_args): Ditto.
8437 (get_gcov_intermediate_filename): New function.
8438 (output_intermediate_file): New function.
8439 (output_gcov_file): New function
8440 (generate_results): Handle new option.
8441 (release_function): Relase demangled name.
8442 (read_graph_file): Handle demangled name.
8443 (output_lines): Ditto.
8444 * doc/gcov.texi: Document gcov intermediate format.
8446 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
8449 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
8450 (lra_emit_add): Use the functions. Add comment about Y as an
8453 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
8456 * collect2.c (collect_atexit): New.
8457 (collect_exit): Delete.
8458 (main): Register collect_atexit with atexit.
8459 (collect_wait): Change collect_exit to exit.
8461 * collect2.h (collect_exit): Delete.
8462 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
8464 2013-06-19 Wei Mi <wmi@google.com>
8466 PR rtl-optimization/57518
8467 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8468 if regno is used in paradoxical subreg.
8469 (update_equiv_regs): Check pdx_subregs[regno] before
8470 set a reg to be equivalent with a mem.
8472 2013-06-19 Matthias Klose <doko@ubuntu.com>
8475 * file-find.h (find_a_file): Add a mode parameter.
8476 * file-find.c (find_a_file): Likewise.
8477 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8478 with X_OK for the executables.
8479 * collect2.c (main): Call find_a_file with X_OK.
8481 2013-06-19 Steve Ellcey <sellcey@mips.com>
8484 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8485 Use NEXT_INSN instead of next_real_insn.
8487 2013-06-19 Jan Hubicka <jh@suse.cz>
8489 * cgraph.h (const_value_known_p): Replace by ...
8490 (ctor_for_folding): .. this one.
8491 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8492 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8493 * expr.c (expand_expr_real_1): Likewise.
8494 (string_constant): Likewise.
8495 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8496 * ipa.c (process_references): Likewise.
8497 (symtab_remove_unreachable_nodes): Likewise.
8498 * ipa-inline-analysis.c (param_change_prob): Likewise.
8499 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8500 (get_base_constructor): Likwise.
8501 * varpool.c (varpool_remove_node): Likewise.
8502 (varpool_remove_initializer): LIkewise.
8503 (dump_varpool_node): LIkwise.
8504 (const_value_known_p): Rewrite to ...
8505 (ctor_for_folding): ... this one.
8507 2013-06-19 Jakub Jelinek <jakub@redhat.com>
8510 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8511 PERSONALITY in $PATH derived prefixes.
8513 2013-06-19 Jeff Law <law@redhat.com>
8515 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8518 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8519 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8522 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8524 * config/arm/vfp.md: Move VFP instruction classification documentation
8526 * config/arm/arm.md: ... here. Update instruction classification
8529 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
8531 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8532 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8533 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8535 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8538 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8539 with NEXT_INSN. Use tablejump_p to check for jump table data
8542 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8545 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8546 that now in C++ the value is correct per the C++ standards.
8548 2013-06-19 Richard Biener <rguenther@suse.de>
8550 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8553 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8556 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8559 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8560 with next_active_insn.
8562 2013-06-18 Sriraman Tallam <tmsriram@google.com>
8564 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
8565 functions are inlined during failures to flag an error.
8566 * tree-inline.c (expand_call_inline): Allow the error to be flagged
8567 in early inline pass.
8569 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
8571 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
8574 2013-06-18 Julian Brown <julian@codesourcery.com>
8576 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
8577 Permit virtual register pre-reload if !strict.
8578 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
8580 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
8582 * config/arm/neon.md (movmisalign<mode>): Use
8583 neon_perm_struct_or_reg_operand instead of
8584 neon_struct_or_register_operand.
8585 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
8586 neon_permissive_struct_operand instead of neon_struct_operand.
8587 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
8588 neon_vector_mem_operand.
8589 * config/arm/predicates.md (neon_struct_operand): Adjust call to
8590 neon_vector_mem_operand.
8591 (neon_permissive_struct_operand): New.
8592 (neon_struct_or_register_operand): Rename to...
8593 (neon_perm_struct_or_reg_operand): This. Adjust call to
8594 neon_vector_mem_operand.
8596 2013-06-18 Richard Biener <rguenther@suse.de>
8598 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
8599 * lto-streamer.h: Include pointer-set.h.
8600 (struct lto_decl_slot): Remove.
8601 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
8602 Remove next_index entry.
8603 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8604 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8605 (lto_init_tree_ref_encoder): Adjust.
8606 (lto_destroy_tree_ref_encoder): Likewise.
8607 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8608 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8609 (lto_output_decl_index): Adjust.
8610 (lto_new_out_decl_state): Likewise.
8611 (lto_record_function_out_decl_state): Likewise.
8612 * lto-streamer-out.c (copy_function): Likewise.
8614 2013-06-18 Richard Biener <rguenther@suse.de>
8616 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
8617 * cgraphunit.c: Include cfgloop.h.
8618 (init_lowered_empty_function): Initialize the loop tree.
8619 (assemble_thunk): Insert new BBs into loops.
8621 2013-06-18 Richard Biener <rguenther@suse.de>
8623 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8624 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8625 the map from cache entry to cache index optional.
8626 (streamer_tree_cache_replace_tree): Adjust accordingly.
8627 (streamer_tree_cache_append): Likewise.
8628 (streamer_tree_cache_delete): Likewise.
8629 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8630 streamer cache map from cache entry to cache index.
8631 * lto-streamer-out.c (create_output_block): Adjust.
8633 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8635 * config/arm/arm.md (attribute "insn"): Move multiplication and
8636 division attributes to...
8637 (attribute "type"): ... here. Remove mult.
8638 (attribute "mul32"): New attribute.
8639 (attribute "mul64"): Add umaal.
8640 (*arm_mulsi3): Update attributes.
8641 (*arm_mulsi3_v6): Likewise.
8642 (*thumb_mulsi3): Likewise.
8643 (*thumb_mulsi3_v6): Likewise.
8644 (*mulsi3_compare0): Likewise.
8645 (*mulsi3_compare0_v6): Likewise.
8646 (*mulsi_compare0_scratch): Likewise.
8647 (*mulsi_compare0_scratch_v6): Likewise.
8648 (*mulsi3addsi): Likewise.
8649 (*mulsi3addsi_v6): Likewise.
8650 (*mulsi3addsi_compare0): Likewise.
8651 (*mulsi3addsi_compare0_v6): Likewise.
8652 (*mulsi3addsi_compare0_scratch): Likewise.
8653 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8654 (*mulsi3subsi): Likewise.
8655 (*mulsidi3adddi): Likewise.
8656 (*mulsi3addsi_v6): Likewise.
8657 (*mulsidi3adddi_v6): Likewise.
8658 (*mulsidi3_nov6): Likewise.
8659 (*mulsidi3_v6): Likewise.
8660 (*umulsidi3_nov6): Likewise.
8661 (*umulsidi3_v6): Likewise.
8662 (*umulsidi3adddi): Likewise.
8663 (*umulsidi3adddi_v6): Likewise.
8664 (*smulsi3_highpart_nov6): Likewise.
8665 (*smulsi3_highpart_v6): Likewise.
8666 (*umulsi3_highpart_nov6): Likewise.
8667 (*umulsi3_highpart_v6): Likewise.
8668 (mulhisi3): Likewise.
8669 (*mulhisi3tb): Likewise.
8670 (*mulhisi3bt): Likewise.
8671 (*mulhisi3tt): Likewise.
8672 (maddhisi4): Likewise.
8673 (*maddhisi4tb): Likewise.
8674 (*maddhisi4tt): Likewise.
8675 (maddhidi4): Likewise.
8676 (*maddhidi4tb): Likewise.
8677 (*maddhidi4tt): Likewise.
8679 (udivsi3): Likewise.
8680 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8681 (thumb2_mulsi_short_compare0): Likewise.
8682 (thumb2_mulsi_short_compare0_scratch): Likewise.
8683 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8684 (1020mult2): Likewise.
8685 (1020mult3): Likewise.
8686 (1020mult4): Likewise.
8687 (1020mult5): Likewise.
8688 (1020mult6): Likewise.
8689 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8691 (cortex_a15_mult64): Likewise.
8692 (cortex_a15_sdiv): Likewise.
8693 (cortex_a15_udiv): Likewise.
8694 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8700 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8701 (pj4_ir_div): Likewise.
8702 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8703 (11_mult2): Likewise.
8704 (11_mult3): Likewise.
8705 (11_mult4): Likewise.
8706 (11_mult5): Likewise.
8707 (11_mult6): Likewise.
8708 (11_mult7): Likewise.
8709 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8710 (cortex_a8_mla): Likewise.
8711 (cortex_a8_mull): Likewise.
8712 (cortex_a8_smulwy): Likewise.
8713 (cortex_a8_smlald): Likewise.
8714 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8715 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8716 (cortex_r4_mul_3): Likewise.
8717 (cortex_r4_mla_4): Likewise.
8718 (cortex_r4_mla_3): Likewise.
8719 (cortex_r4_smlald): Likewise.
8720 (cortex_r4_mull): Likewise.
8721 (cortex_r4_sdiv): Likewise.
8722 (cortex_r4_udiv): Likewise.
8723 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8724 (cortex_a7_idiv): Likewise.
8725 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8726 (9_mult2): Likewise.
8727 (9_mult3): Likewise.
8728 (9_mult4): Likewise.
8729 (9_mult5): Likewise.
8730 (9_mult6): Likewise.
8731 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8732 (cortex_a53_sdiv): Likewise.
8733 (cortex_a53_udiv): Likewise.
8734 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8735 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8736 (mp626_mult2): Likewise.
8737 (mp626_mult3): Likewise.
8738 (mp626_mult4): Likewise.
8739 * config/arm/fa526.md (526_mult1): Update attribute change.
8740 (526_mult2): Likewise.
8741 * config/arm/arm-generic.md (mult): Update attribute change.
8742 (mult_ldsched_strongarm): Likewise.
8743 (mult_ldsched): Likewise.
8744 (multi_cycle): Likewise.
8745 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8746 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8747 (606te_mult2): Likewise.
8748 (606te_mult3): Likewise.
8749 (606te_mult4): Likewise.
8750 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8751 (cortex_a9_mac16): Likewise.
8752 (cortex_a9_multiply): Likewise.
8753 (cortex_a9_mac): Likewise.
8754 (cortex_a9_multiply_long): Likewise.
8755 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8756 (626te_mult2): Likewise.
8757 (626te_mult3): Likewise.
8758 (626te_mult4): Likewise.
8760 2013-06-18 Richard Biener <rguenther@suse.de>
8763 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8765 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8768 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8769 with next_active_insn.
8771 2013-06-18 Alan Modra <amodra@gmail.com>
8773 * config/rs6000/rs6000.h (enum data_align): New.
8774 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8775 (DATA_ABI_ALIGNMENT): Define.
8776 (CONSTANT_ALIGNMENT): Correct comment.
8777 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8778 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8780 2013-06-17 David Malcolm <dmalcolm@redhat.com>
8782 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8783 ATTRIBUTE_UNUSED marking.
8785 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8787 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8788 alternative and update.
8789 (aarch64_dup_lanedi): Delete.
8790 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8791 * config/aarch64/aarch64-simd-builtins.def: Update.
8793 2013-06-17 Richard Biener <rguenther@suse.de>
8795 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8796 (lto_input_scc): Declare.
8797 (lto_input_tree_1): Likewise.
8798 (struct lto_stats_d): Add num_tree_bodies_output and
8799 num_pickle_refs_output.
8800 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8801 (lto_read_tree_1): Split out from ...
8802 (lto_read_tree): ... this.
8803 (lto_input_scc): New function.
8804 (lto_input_tree_1): Split out from ...
8805 (lto_input_tree): ... this. Handle LTO_tree_scc.
8806 (lto_data_in_create): Create the streamer cache without hashes.
8807 * lto-streamer-out.c (create_output_block): Create the streamer
8808 cache with hashes when not doing WPA.
8809 (lto_write_tree_1): Split out from ...
8810 (lto_write_tree): ... this.
8811 (get_symbol_initial_value): New function.
8812 (lto_output_tree_1): Split out from ...
8813 (lto_output_tree): ... this. Write trees as series of SCCs
8814 using a DFS walk via DFS_write_tree.
8815 (struct sccs, struct scc_entry): New types.
8816 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
8817 (DFS_write_tree_body): New function.
8818 (DFS_write_tree): Likewise.
8819 (hash_tree): Likewise.
8820 (scc_entry_compare): Likewise.
8821 (hash_scc): Likewise.
8822 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
8823 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
8824 TREE_CHAIN as regular reference.
8825 (streamer_read_integer_cst): Remove.
8826 (streamer_get_pickled_tree): Adjust.
8827 * tree-streamer-out.c (streamer_write_chain): Disable streaming
8828 of DECL_EXTERNALs in BLOCK_VARS for now.
8829 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
8831 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
8832 Add hash value argument and record that if hashes are recorded
8834 (streamer_tree_cache_insert_1): Adjust.
8835 (streamer_tree_cache_insert): Likewise.
8836 (streamer_tree_cache_insert_at): Rename to ...
8837 (streamer_tree_cache_replace_tree): ... this and adjust.
8838 (streamer_tree_cache_append): Adjust.
8839 (record_common_node): Likewise.
8840 (streamer_tree_cache_create): Add argument whether to
8841 record hash values together with trees.
8842 (streamer_tree_cache_delete): Adjust.
8843 * tree-streamer.h (struct streamer_tree_cache_d): Add
8845 (streamer_read_integer_cst): Remove.
8846 (streamer_tree_cache_insert): Adjust.
8847 (streamer_tree_cache_append): Likewise.
8848 (streamer_tree_cache_insert_at): Rename to ...
8849 (streamer_tree_cache_replace_tree): ... this and adjust.
8850 (streamer_tree_cache_create): Add argument whether to record hashes.
8851 (streamer_tree_cache_get): Rename to ...
8852 (streamer_tree_cache_get_tree): ... this.
8853 (streamer_tree_cache_get_hash): New function.
8854 * tree.c (cache_integer_cst): New function.
8855 * tree.h (cache_integer_cst): Declare.
8856 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
8857 * lto-symtab.c (lto_varpool_replace_node): Only release
8858 DECL_INITIAL of non-prevailing decls.
8859 * varpool.c (varpool_remove_initializer): Do not release
8860 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
8862 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8864 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
8865 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
8866 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
8867 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
8868 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
8869 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
8870 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
8871 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
8872 instead of TARGET_64BIT.
8873 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
8874 Require ISA_HAS_<D>DIV.
8876 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
8878 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
8879 (mips*-*-linux*): Move default with_llsc setting to where other
8881 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
8883 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
8884 Likewise. Remove default with_tune setting. Move default float
8885 setting to its own block. Handle with_llsc in the same block as above.
8887 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
8889 PR rtl-optimization/57425
8890 PR rtl-optimization/57569
8891 * alias.c (write_dependence_p): Add new parameters mem_mode,
8892 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
8893 Changed all callers.
8894 (canon_anti_dependence): New function.
8895 * cse.c (check_dependence): Use canon_anti_dependence.
8896 * cselib.c (cselib_invalidate_mem): Likewise.
8897 * rtl.h (canon_anti_dependence): Declare.
8899 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8901 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
8902 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
8903 ".set mips3" for 64-bit targets.
8905 2013-06-15 Dehao Chen <dehao@google.com>
8907 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
8908 * gimple-low.c (gimple_check_call_matching_types): Likewise.
8909 (gimple_check_call_args): Likewise.
8910 * value-prof.c (check_ic_target): Likewise.
8911 * ipa-inline.c (early_inliner): Likewise.
8912 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
8913 * cgraph.c (cgraph_create_edge_1): Likewise.
8914 (cgraph_make_edge_direct): Likewise.
8916 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
8919 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
8920 rs6000_output_move_128bit to handle emitting quad memory
8921 operations. Set attribute length to 8 bytes.
8923 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
8925 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
8927 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
8928 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
8929 (aarch64_<su>mlsl<mode>): Likewise.
8931 2013-06-14 Mike Stump <mikestump@comcast.net>
8933 * Makefile.in (TARGET_H): Add insn-codes.h.
8935 2013-06-14 Alan Modra <amodra@gmail.com>
8939 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
8940 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
8941 bitfield expansion when EXPAND_MEMORY.
8942 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
8944 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
8946 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
8947 test for clearing quad memory on 32-bit later.
8949 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8951 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
8952 (fold_negate_expr): Likewise.
8953 (fold_real_zero_addition_p): Handle vectors.
8954 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
8956 2013-06-14 Alan Modra <amodra@gmail.com>
8958 * varasm.c (force_const_mem): Revert 2013-06-07 change.
8960 2013-06-13 Jan Hubicka <jh@suse.cz>
8962 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
8963 Local comdats are not externally visible.
8964 * symtab.c (dump_symtab_base): Dump externally visible.
8965 (verify_symtab_base): Verify back links in the symtab hash.
8967 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8969 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
8970 CONVERT_EXPR as equal nodes.
8972 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8974 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8976 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8978 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
8979 Generalize to complex and vector.
8980 * tree.c (build_all_ones_cst): New function.
8981 * tree.h (build_all_ones_cst): Declare it.
8983 2013-06-13 Alan Modra <amodra@gmail.com>
8985 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
8986 * config/rs6000/rs6000.md (signbittf2): New insn.
8987 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
8988 (abstf2_internal, cmptf_internal2): Likewise.
8989 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
8991 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8992 Pat Haugen <pthaugen@us.ibm.com>
8993 Peter Bergner <bergner@vnet.ibm.com>
8995 * config/rs6000/rs6000.c (emit_load_locked): Add support for
8996 power8 byte, half-word, and quad-word atomic instructions.
8997 (emit_store_conditional): Likewise.
8998 (rs6000_expand_atomic_compare_and_swap): Likewise.
8999 (rs6000_expand_atomic_op): Likewise.
9001 * config/rs6000/sync.md (larx): Add new modes for power8.
9003 (AINT): New mode iterator to include TImode as well as normal
9004 integer modes on power8.
9005 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
9006 that VSX registers are not considered. Use AINT mode iterator
9007 instead of INT1 to allow inclusion of quad word atomic operations
9009 (load_locked<mode>): Likewise.
9010 (store_conditional<mode>): Likewise.
9011 (atomic_compare_and_swap<mode>): Likewise.
9012 (atomic_exchange<mode>): Likewise.
9013 (atomic_nand<mode>): Likewise.
9014 (atomic_fetch_<fetchop_name><mode>): Likewise.
9015 (atomic_nand_fetch<mode>): Likewise.
9016 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
9018 (ATOMIC): On power8, add QImode, HImode modes.
9019 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
9020 modes that promote to SImode.
9021 (load_lockedti): Convert TImode arguments to PTImode, so that we
9022 get a guaranteed even/odd register pair.
9023 (load_lockedpti): Likewise.
9024 (store_conditionalti): Likewise.
9025 (store_conditionalpti): Likewise.
9027 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
9028 atomic load/store instructions.
9031 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
9033 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
9035 (insn_count): New attribute, with most cases extracted from...
9036 (length): ...here. Redefine most cases in terms of insn_count.
9037 (single_insn): Delete.
9038 (can_delay): Use insn_count to check for single instructions.
9039 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
9040 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
9041 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
9042 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
9043 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
9044 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
9045 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
9046 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
9047 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
9048 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
9049 rather than "length".
9050 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
9051 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
9052 Use "insn_count" rather than "length".
9053 * config/mips/mips-dsp.md
9054 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
9055 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
9058 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
9060 PR tree-optimization/57361
9061 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
9063 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
9065 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
9067 (aarch64_simd_combine<mode>): New instruction expansion.
9068 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
9070 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
9071 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
9073 2013-06-12 Jan Hubicka <jh@suse.cz>
9075 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
9076 decl has when in streaming stage.
9077 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
9078 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
9080 2013-06-12 Roland Stigge <stigge@antcom.de>
9083 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
9085 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9087 PR tree-optimization/57537
9088 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
9089 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
9091 2013-06-12 Richard Biener <rguenther@suse.de>
9093 * data-streamer.h (streamer_write_char_stream): CSE
9094 obs->current_pointer.
9095 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
9096 streamer_write_char_stream manually and optimize the resulting loop.
9097 (streamer_write_hwi_stream): Likewise.
9099 2013-06-12 Jan Hubicka <jh@suse.cz>
9101 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
9102 * cgraph.h (varpool_create_empty_node): Declare.
9103 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
9105 * symtab.c (symtab_unregister_node): Be lax about missin entries
9107 (symtab_get_node): Update comment.
9108 * varpool.c (varpool_create_empty_node): Break out from ...
9109 (varpool_node_for_decl): ... here.
9110 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
9112 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
9114 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
9115 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
9116 part. Use straight-line flow at the end.
9117 <COMPONENT_REF>: Remove superfluous else.
9118 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
9120 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9123 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
9124 target hook even for !TREE_PUBLIC decls. If no resolution info
9125 is available, return false for common and external decls.
9127 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9129 * config/rl78/constraints.md (U): New constraint.
9130 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
9133 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9136 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
9137 to allow returning address to AT_PLATFORM name.
9139 2013-06-11 Jan Hubicka <jh@suse.cz>
9141 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
9142 * cgraph.h (symtab_node_base): Add weakref flag.
9143 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
9144 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
9145 (output_weakrefs): Use weakref flag.
9146 * fold-const.c (simple_operand_p): Handle WEAK.
9147 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
9148 * ipa.c (varpool_externally_visible_p): Drop weakref.
9149 (function_and_variable_visibility): Update comment; fix weakref
9150 sanity checks; do not clear DECL_WEAK on them.
9151 * lto-cgraph.c (lto_output_node): update.
9152 (lto_output_varpool_node): Update.
9153 (input_overwrite_node): Update.
9154 (input_node): Update.
9155 (input_varpool_node): Update.
9156 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
9157 (lto_symtab_merge_symbols): Add sanity check.
9158 (lto_symtab_prevailing_decl): Do not special case weakrefs.
9159 * passes.c (rest_of_decl_compilation): Set static flag, too.
9160 * symtab.c (dump_symtab_base): Dump weakref.
9161 (verify_symtab_base): Sanity check weakrefs.
9162 (symtab_make_decl_local): Remove duplicated code.
9163 (symtab_alias_ultimate_target): Simplify.
9164 * varpool.c (varpool_create_variable_alias): Set weakref flag.
9166 2013-06-11 Tom de Vries <tom@codesourcery.com>
9168 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
9169 sequence_vect == NULL.
9171 2013-06-11 DJ Delorie <dj@redhat.com>
9173 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
9174 (rl78_unwind_word_mode): New.
9176 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9178 * final.c (debug_prefix_maps): Make static.
9180 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9182 * function.c (initial_trampoline): Remove stray copy.
9184 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
9186 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
9188 2013-06-11 Martin Jambor <mjambor@suse.cz>
9190 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
9191 within bounds at the beginning of the function.
9193 2013-06-11 Alan Modra <amodra@gmail.com>
9195 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
9197 (get_named_section): Don't NULL !DECL_P decl.
9199 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
9201 * doc/invoke.texi (core-avx2): Document.
9203 (atom): Updated with MOVBE.
9205 2013-06-11 Richard Biener <rguenther@suse.de>
9207 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
9209 2013-06-11 Anton Blanchard <anton@samba.org>
9211 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
9212 correct shift value in little-endian mode.
9214 2013-06-11 Jakub Jelinek <jakub@redhat.com>
9217 * varasm.c (get_variable_align): Move #endif to the right place.
9219 2013-06-10 Cary Coutant <ccoutant@google.com>
9221 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
9222 for hash so that hash table traversal order is deterministic.
9224 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
9225 Pat Haugen <pthaugen@us.ibm.com>
9226 Peter Bergner <bergner@vnet.ibm.com>
9228 * config/rs6000/vector.md (GPR move splitter): Do not split moves
9229 of vectors in GPRS if they are direct moves or quad word load or
9232 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
9234 (direct_move_p): Likewise.
9235 (quad_load_store_p): Likewise.
9237 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
9238 classes into bins based on the physical register type.
9239 (reg_class_to_reg_type): Likewise.
9240 (IS_STD_REG_TYPE): Likewise.
9241 (IS_FP_VECT_REG_TYPE): Likewise.
9242 (reload_fpr_gpr): Arrays to determine what insn to use if we can
9243 use direct move instructions.
9244 (reload_gpr_vsx): Likewise.
9245 (reload_vsx_gpr): Likewise.
9246 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
9247 information that is a simplification of register classes. Also
9248 precalculate direct move reload helpers.
9249 (direct_move_p): New function to return true if the operation can
9250 be done as a direct move instruciton.
9251 (quad_load_store_p): New function to return true if the operation
9252 is a quad memory operation.
9253 (rs6000_legitimize_address): If quad memory, only allow register
9254 indirect for TImode addresses.
9255 (rs6000_legitimate_address_p): Likewise.
9256 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
9257 (rs6000_reload_register_type): Likewise.
9258 (register_to_reg_type): Return register type.
9259 (rs6000_secondary_reload_simple_move): New helper function for
9260 secondary reload and secondary memory needed to identify anything
9261 that is a simple move, and does not need reloading.
9262 (rs6000_secondary_reload_direct_move): New helper function for
9263 secondary reload to identify cases that can be done with several
9264 instructions via the direct move instructions.
9265 (rs6000_secondary_reload_move): New helper function for secondary
9266 reload to identify moves between register types that can be done.
9267 (rs6000_secondary_reload): Add support for quad memory operations
9268 and for direct move.
9269 (rs6000_secondary_memory_needed): Likewise.
9270 (rs6000_debug_secondary_memory_needed): Change argument names.
9271 (rs6000_output_move_128bit): New function to return the move to
9272 use for 128-bit moves, including knowing about the various
9273 limitations of quad memory operations.
9275 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
9276 memory operations. call rs6000_output_move_128bit for the actual
9277 instruciton(s) to generate.
9278 (vsx_movti_64bit): Likewise.
9280 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
9281 (UNSPEC_P8V_MTVSRWZ): Likewise.
9282 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
9283 (UNSPEC_P8V_MTVSRD): Likewise.
9284 (UNSPEC_P8V_XXPERMDI): Likewise.
9285 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
9286 (UNSPEC_FUSION_GPR): Likewise.
9287 (FMOVE128_GPR): New iterator for direct move.
9288 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
9291 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
9292 loads and direct move instructions.
9293 (zero_extendsidi2_lfiwzx): Likewise.
9294 (extendsidi2_lfiwax): Likewise.
9295 (extendsidi2_nocell): Likewise.
9296 (floatsi<mode>2_lfiwax): Likewise.
9298 (floatunssi<mode>2_lfiwzx): Likewise.
9300 (fix_trunc<mode>_stfiwx): Likewise.
9301 (fixuns_trunc<mode>_stfiwx): Likewise.
9302 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
9303 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
9304 (parity<mode>2_cmpb): Set length/type attr.
9305 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
9306 for 'mr.' to fast_compare.
9307 (bpermd_<mode>): Change type attr to popcnt.
9308 (p8_fmrgow_<mode>): New insns for power8 direct move support.
9309 (p8_mtvsrwz_1): Likewise.
9310 (p8_mtvsrwz_2): Likewise.
9311 (reload_fpr_from_gpr<mode>): Likewise.
9312 (p8_mtvsrd_1): Likewise.
9313 (p8_mtvsrd_2): Likewise.
9314 (p8_xxpermdi_<mode>): Likewise.
9315 (reload_vsx_from_gpr<mode>): Likewise.
9316 (reload_vsx_from_gprsf): Likewise.
9317 (p8_mfvsrd_3_<mode>): LIkewise.
9318 (reload_gpr_from_vsx<mode>): Likewise.
9319 (reload_gpr_from_vsxsf): Likewise.
9320 (p8_mfvsrd_4_disf): Likewise.
9321 (multi-word GPR splits): Do not split direct moves or quad memory
9324 2013-06-10 David Malcolm <dmalcolm@redhat.com>
9326 * tree-into-ssa.c (interesting_blocks): Make static.
9328 2013-06-10 Jakub Jelinek <jakub@redhat.com>
9331 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
9332 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
9333 Use DATA_ABI_ALIGNMENT for that case instead if defined.
9334 (get_variable_align): New function.
9335 (get_variable_section, emit_bss, emit_common,
9336 assemble_variable_contents, place_block_symbol): Use
9337 get_variable_align instead of DECL_ALIGN.
9338 (assemble_noswitch_variable): Add align argument, use it
9339 instead of DECL_ALIGN.
9340 (assemble_variable): Adjust caller. Use get_variable_align
9341 instead of DECL_ALIGN.
9342 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
9344 (DATA_ABI_ALIGNMENT): Define.
9345 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
9346 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
9347 opt is false, only return the psABI mandated alignment increase.
9348 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
9349 (DATA_ABI_ALIGNMENT): ... this.
9350 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
9351 (DATA_ABI_ALIGNMENT): ... this.
9352 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
9353 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
9354 (DATA_ABI_ALIGNMENT): ... this.
9355 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
9356 * doc/tm.texi: Regenerated.
9358 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
9360 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
9361 cmp_code to construct REG_EQUAL note.
9363 2013-06-09 Jakub Jelinek <jakub@redhat.com>
9366 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
9367 that operands[2] doesn't overlap with operands[0].
9369 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
9370 Jan Hubicka <jh@suse.cz>
9372 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
9373 hack to mark symbols as used.
9375 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
9377 PR rtl-optimization/57559
9378 * lra-constraints.c (process_alt_operands): Don't discourage
9379 memory with known offset for offsetable memory constraint.
9380 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
9382 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
9384 * varasm.c (struct oc_local_state): Reorder fields.
9385 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
9386 and adjust accordingly.
9387 (output_constructor): Reorder initialization code and adjust call to
9388 output_constructor_bitfield.
9390 2013-06-07 Jan Hubicka <jh@suse.cz>
9392 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
9394 2013-06-07 David Malcolm <dmalcolm@redhat.com>
9396 * tree-object-size.c (unknown): Make const.
9398 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9400 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
9401 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
9402 for last alternative in the cpu_facility attribute.
9404 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9407 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
9408 (xordi3): Change operand 2 constraint to arm_xordi_operand.
9409 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
9410 * config/arm/constraints.md (Dg): New constraint.
9411 * config/arm/neon.md (xordi3_neon): Remove.
9412 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
9413 * config/arm/predicates.md (arm_xordi_operand): New predicate.
9415 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9417 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
9418 Clean up alternatives.
9420 2013-06-07 Alan Modra <amodra@gmail.com>
9422 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
9425 2013-06-07 Alan Modra <amodra@gmail.com>
9427 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
9429 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9431 * config/arm/constraints.md (Df): New constraint.
9432 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
9433 Correct length attribute for last two alternatives.
9435 2013-06-07 Alan Modra <amodra@gmail.com>
9437 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9438 override user -mfp-in-toc.
9439 (offsettable_ok_by_alignment): Consider just the current access
9440 rather than the whole object, unless BLKmode. Handle
9441 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
9442 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
9443 for -mcmodel=medium.
9444 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
9445 override user -mfp-in-toc or -msum-in-toc. Default to
9446 -mno-fp-in-toc for -mcmodel=medium.
9448 2013-06-06 DJ Delorie <dj@redhat.com>
9450 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
9451 TARGET_VALID_POINTER_MODE.
9453 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
9454 Pat Haugen <pthaugen@us.ibm.com>
9455 Peter Bergner <bergner@vnet.ibm.com>
9457 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9458 Document new power8 builtins.
9460 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
9461 condition code register, to allow 128-bit logical operations to be
9462 done in the VSX or GPR registers.
9463 (nor<mode>3): Use the canonical form for nor.
9464 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9465 vclz*, and vpopcnt* vector instructions.
9466 (nand<mode>3): Likewise.
9467 (orc<mode>3): Likewise.
9468 (clz<mode>2): LIkewise.
9469 (popcount<mode>2): Likewise.
9471 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9472 that only the GPRs are recognized.
9474 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9475 support for new power8 builtins.
9477 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9479 (xscvdpspn): Likewise.
9485 (vpopcnt): Likewise.
9486 (vpopcntb): Likewise.
9487 (vpopcnth): Likewise.
9488 (vpopcntw): Likewise.
9489 (vpopcntd): Likewise.
9494 (eqv_v16qi3): Likewise.
9495 (eqv_v8hi3): Likewise.
9496 (eqv_v4si3): Likewise.
9497 (eqv_v2di3): Likewise.
9498 (eqv_v4sf3): Likewise.
9499 (eqv_v2df3): Likewise.
9501 (nand_v16qi3): Likewise.
9502 (nand_v8hi3): Likewise.
9503 (nand_v4si3): Likewise.
9504 (nand_v2di3): Likewise.
9505 (nand_v4sf3): Likewise.
9506 (nand_v2df3): Likewise.
9508 (orc_v16qi3): Likewise.
9509 (orc_v8hi3): Likewise.
9510 (orc_v4si3): Likewise.
9511 (orc_v2di3): Likewise.
9512 (orc_v4sf3): Likewise.
9513 (orc_v2df3): Likewise.
9515 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9516 allow power8 quad mode in 64-bit.
9517 (rs6000_builtin_vectorized_function): Add support to vectorize
9518 ISA 2.07 count leading zeros, population count builtins.
9519 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9520 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9521 (builtin_function_type): Add vgbbd builtin function which takes an
9523 (altivec_expand_vec_perm_const): Add support for new power8 merge
9526 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9527 that does not include TImdoe for use with 32-bit.
9528 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9530 (UNSPEC_VSX_CVDPSPN): Likewise.
9531 (vsx_xscvdpspn): Likewise.
9532 (vsx_xscvspdpn): Likewise.
9533 (vsx_xscvdpspn_scalar): Likewise.
9534 (vsx_xscvspdpn_directmove): Likewise.
9535 (vsx_and<mode>3): Split logical operations into 32-bit and
9536 64-bit. Add support to do logical operations on TImode as well as
9537 VSX vector types. Allow logical operations to be done in either
9538 VSX registers or in general purpose registers in 64-bit mode. Add
9539 splitters if GPRs were used. For AND, add clobber of CCmode to
9540 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9542 (vsx_and<mode>3_32bit): Likewise.
9543 (vsx_and<mode>3_64bit): Likewise.
9544 (vsx_ior<mode>3): Likewise.
9545 (vsx_ior<mode>3_32bit): Likewise.
9546 (vsx_ior<mode>3_64bit): Likewise.
9547 (vsx_xor<mode>3): Likewise.
9548 (vsx_xor<mode>3_32bit): Likewise.
9549 (vsx_xor<mode>3_64bit): Likewise.
9550 (vsx_one_cmpl<mode>2): Likewise.
9551 (vsx_one_cmpl<mode>2_32bit): Likewise.
9552 (vsx_one_cmpl<mode>2_64bit): Likewise.
9553 (vsx_nor<mode>3): Likewise.
9554 (vsx_nor<mode>3_32bit): Likewise.
9555 (vsx_nor<mode>3_64bit): Likewise.
9556 (vsx_andc<mode>3): Likewise.
9557 (vsx_andc<mode>3_32bit): Likewise.
9558 (vsx_andc<mode>3_64bit): Likewise.
9559 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
9560 and xxlorc instructions.
9561 (vsx_eqv<mode>3_64bit): Likewise.
9562 (vsx_nand<mode>3_32bit): Likewise.
9563 (vsx_nand<mode>3_64bit): Likewise.
9564 (vsx_orc<mode>3_32bit): Likewise.
9565 (vsx_orc<mode>3_64bit): Likewise.
9567 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
9569 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
9571 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
9572 (p8_vmrgow): Likewise.
9573 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
9574 GPRs to be split under VSX.
9575 (p8v_clz<mode>2): Add power8 count leading zero support.
9576 (p8v_popcount<mode>2): Add power8 population count support.
9577 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
9580 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
9583 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
9585 (vec_nand): Likewise.
9586 (vec_vclz): Likewise.
9587 (vec_vclzb): Likewise.
9588 (vec_vclzd): Likewise.
9589 (vec_vclzh): Likewise.
9590 (vec_vclzw): Likewise.
9591 (vec_vgbbd): Likewise.
9592 (vec_vmrgew): Likewise.
9593 (vec_vmrgow): Likewise.
9594 (vec_vpopcnt): Likewise.
9595 (vec_vpopcntb): Likewise.
9596 (vec_vpopcntd): Likewise.
9597 (vec_vpopcnth): Likewise.
9598 (vec_vpopcntw): Likewise.
9600 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9602 PR rtl-optimization/57468
9603 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
9606 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9608 PR rtl-optimization/57459
9609 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
9610 type when setting live regs.
9612 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9614 * config/s390/s390.opt (mlra): New option.
9615 * config/s390/s390.c (s390_decompose_address): Check displacement
9616 for all registers for LRA.
9617 (s390_secondary_reload): Don't used secondary reloads for LRA.
9618 (s390_lra_p): New function.
9619 (TARGET_LRA_P): Define.
9620 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
9621 of attribute cpu_facility to zarch for the last alternative.
9622 (*cmpmem_short): Ditto.
9624 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9626 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9627 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9628 (arm_expand_prologue): Likewise.
9630 2013-06-06 Teresa Johnson <tejohnson@google.com>
9633 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9634 as this is now done by redirect_edge_and_branch_force.
9635 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9636 barriers, and fix interaction with splitting.
9637 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9638 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9639 reflect changes made in the routine.
9640 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9641 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9642 since this is called in cfglayout mode, and replace partition fixup
9643 with assert as that is now done by force_nonfallthru_and_redirect.
9644 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9645 already be marked with region crossing note.
9646 (insert_section_boundary_note): Make non-static, gate on flag
9647 has_bb_partition, rewrite to also check for multiple partitions.
9648 (rest_of_handle_reorder_blocks): Remove call to
9649 insert_section_boundary_note, now done later during free_cfg.
9650 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9651 * bb-reorder.h (insert_section_boundary_note): Declare.
9652 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9653 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9654 invoke insert_section_boundary_note.
9655 (try_redirect_by_replacing_jump): Remove unnecessary
9656 check for region crossing note.
9657 (fixup_partition_crossing): New function.
9658 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9659 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9660 in non-cfglayout mode.
9661 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9662 remove old code that tried to do this. Emit barrier correctly
9663 when we are in cfglayout mode.
9664 (last_bb_in_partition): New function.
9665 (rtl_split_edge): Correctly fixup partition boundaries.
9666 (commit_one_edge_insertion): Remove old code that tried to
9667 fixup region crossing edge since this is now handled in
9668 split_block, and set up insertion point correctly since
9669 block may now end in a jump.
9670 (verify_hot_cold_block_grouping): Guard against checking when not in
9671 linearized RTL mode.
9672 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9674 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9675 rtl_verify_flow_info, so not called in cfglayout mode.
9676 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9677 (fixup_reorder_chain): Remove old code that attempted to fixup region
9678 crossing note as this is now handled in force_nonfallthru_and_redirect.
9679 (duplicate_insn_chain): Don't duplicate switch section notes.
9680 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9682 * basic-block.h (emit_barrier_after_bb): Declare.
9684 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9686 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9687 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9688 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9690 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9692 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9694 * config/arm/ldmstm.md: Regenerate.
9696 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9698 * config/arm/sync.md (atomic_loaddi_1):
9699 Disable predication for arm_restrict_it.
9700 (arm_load_exclusive<mode>): Likewise.
9701 (arm_load_exclusivesi): Likewise.
9702 (arm_load_exclusivedi): Likewise.
9703 (arm_load_acquire_exclusive<mode>): Likewise.
9704 (arm_load_acquire_exclusivesi): Likewise.
9705 (arm_load_acquire_exclusivedi): Likewise.
9706 (arm_store_exclusive<mode>): Likewise.
9707 (arm_store_exclusive<mode>): Likewise.
9708 (arm_store_release_exclusivedi): Likewise.
9709 (arm_store_release_exclusive<mode>): Likewise.
9711 2013-06-06 Richard Biener <rguenther@suse.de>
9713 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9715 (lto_tag_is_tree_code_p): Adjust.
9716 (lto_tag_is_gimple_code_p): Likewise.
9717 (lto_gimple_code_to_tag): Likewise.
9718 (lto_tag_to_gimple_code): Likewise.
9719 (lto_tree_code_to_tag): Likewise.
9720 (lto_tag_to_tree_code): Likewise.
9721 * data-streamer.h (streamer_write_hwi_in_range): Use
9722 uhwi streaming to stream the normalized range.
9723 (streamer_read_hwi_in_range): Likewise.
9725 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9727 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9728 (predicable_short_it): Likewise.
9729 (predicated): Likewise.
9730 (enabled): Handle above.
9731 (define_cond_exec): Set predicated attribute to yes.
9733 2013-06-05 Mike Stump <mikestump@comcast.net>
9735 * gdbinit.in (__FUNCTION__): Add.
9737 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
9739 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9740 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9742 2013-06-05 Jan Hubicka <jh@suse.cz>
9744 * varasm.c (mark_decl_referenced): Revert the removal until targets
9747 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
9749 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9750 instead of mark_decl_referenced.
9752 2013-06-05 Jan Hubicka <jh@suse.cz>
9754 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9755 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9756 and symtab_used_from_object_file_p.
9757 (cgraph_make_node_local_1): Clear forced_by_abi.
9758 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9759 * cgraph.h (symtab_node_base): Add forced_by_abi.
9760 (decide_is_variable_needed): Remove.
9761 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9762 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9763 (decide_is_symbol_needed): ... this one; handle symbols in general;
9764 always analyze virtuals; honnor forced_by_abi.
9765 (cgraph_finalize_function): Update.
9766 (varpool_finalize_decl): Update.
9767 (symbol_defined_and_needed): Remove.
9768 (analyze_functions): Update.
9769 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9770 output_refs, input_overwrite_node): Handle forced_by_abi.
9771 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9772 (address_taken_from_non_vtable_p): ... this one.
9773 (comdat_can_be_unshared_p_1): New function.
9774 (cgraph_comdat_can_be_unshared_p): Rename to ...
9775 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9776 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9777 (function_and_variable_visibility): Clear forced_by_abi as needed.
9778 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9779 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
9780 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9781 * varpool.c (decide_is_variable_needed): Remove.
9783 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9785 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9786 (arm_option_override): Override arm_restrict_it where appropriate.
9787 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9788 * config/arm/arm.opt (mrestrict-it): New command-line option.
9789 * doc/invoke.texi: Document -mrestrict-it.
9791 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9793 * tsan.c (tsan_atomic_table): Make const.
9795 2013-06-05 Richard Biener <rguenther@suse.de>
9797 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9798 index associated with the tree we are supposed to replace.
9799 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9800 TREE_ASM_WRITTEN as zero for everything but SSA names.
9802 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9804 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9806 2013-06-04 Jan Hubicka <jh@suse.cz>
9808 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9809 (input_node, input_varpool_node): Handle correctly external same
9811 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9812 nodes at ltrans stage.
9814 2013-06-04 Jan Hubicka <jh@suse.cz>
9816 * ipa-inline.c (update_caller_keys): Fix availability test.
9817 (update_callee_keys): Likewise.
9818 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
9819 to follow ELF standard.
9821 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
9823 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
9824 (mips64r5900el-*-elf*): New configurations.
9825 * config/mips/mips-cpus.def (r5900): New processor.
9826 * config/mips/mips-tables.opt: Regenerate.
9827 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
9828 (mips_issue_rate): Handle PROCESSOR_R5900.
9829 (mips_reorg_process_insns): Force reorder mode for the R5900.
9830 * config/mips/mips.h (TARGET_MIPS5900): Define.
9831 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
9833 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
9835 * config/mips/mips.md (processor): Add r5900.
9836 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
9838 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9840 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
9841 into function to generate MOVI instruction.
9842 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
9843 (aarch64_preferred_simd_mode): Turn into wrapper.
9844 (aarch64_output_scalar_simd_mov_immediate): New function.
9845 * config/aarch64/aarch64-protos.h: Add prototype for above.
9847 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9849 * config/aarch64/aarch64.c (simd_immediate_info): Remove
9850 element_char member.
9851 (sizetochar): Return signed char.
9852 (aarch64_simd_valid_immediate): Remove elchar and other
9853 unnecessary variables.
9854 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
9855 Calculate element_char as required.
9856 * config/aarch64/aarch64-protos.h: Update and move prototype
9857 for aarch64_output_simd_mov_immediate.
9858 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
9861 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9863 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
9864 information completed by aarch64_simd_valid_immediate.
9865 (aarch64_legitimate_constant_p): Update arguments.
9866 (aarch64_simd_valid_immediate): Work with struct rather than many
9868 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
9869 (aarch64_simd_make_constant): Update arguments.
9870 (aarch64_output_simd_mov_immediate): Work with struct rather than
9871 many pointers. Output immediate directly rather than as operand.
9872 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9874 * config/aarch64/constraints.md (Dn): Update arguments.
9876 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9878 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
9880 (aarch64_simd_immediate_valid_for_move): Remove.
9881 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
9882 (aarch64_simd_make_constant): Update call.
9883 (aarch64_output_simd_mov_immediate): Update call.
9884 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9886 * config/aarch64/constraints.md (Dn): Update call.
9888 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9890 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
9891 return type to bool for prototype.
9892 (aarch64_legitimate_constant_p): Check for true instead of not -1.
9893 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
9894 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
9896 2013-06-04 Catherine Moore <clm@codesourcery.com>
9898 * config/mips/mips.opt (meva): New.
9899 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
9900 (ASM_SPEC): Handle -meva.
9901 * doc/invoke.texi (meva): Document.
9903 2013-06-04 Alan Modra <amodra@gmail.com>
9905 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
9908 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9910 * rtl.def: Add extra fourth optional field to define_cond_exec.
9911 * gensupport.c (process_one_cond_exec): Process attributes from
9913 * doc/md.texi: Document fourth field in define_cond_exec.
9915 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
9917 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
9918 out the processing order as in store_bit_field_1.
9920 2013-06-04 Jan Hubicka <jh@suse.cz>
9923 * cgraphunit.c (cgraph_process_same_body_aliases): Create
9924 non-VAR_DECL node if it does not exist yet.
9926 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
9928 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
9929 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
9930 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
9931 target_cpu_default setting.
9933 2013-06-03 Teresa Johnson <tejohnson@google.com>
9935 * dumpfile.c (opt_info_switch_p): Change -fopt-info
9936 default to -fopt-info=optimized instead of all.
9937 * doc/invoke.texi: Ditto.
9938 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
9939 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
9940 (execute_vect_slp): Emit BB vectorization success under
9941 MSG_OPTIMIZED_LOCATIONS.
9942 * tree-vect-slp.c (vect_slp_transform_bb): Change
9943 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
9944 * tree-vect-loop.c (vect_transform_loop): Ditto.
9946 2013-06-03 Jason Merrill <jason@redhat.com>
9949 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9950 Use TARGET_EXPR for C++.
9952 2013-06-03 Jakub Jelinek <jakub@redhat.com>
9954 PR rtl-optimization/57268
9955 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
9956 if DEBUG_INSN_P (insn).
9959 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9961 PR rtl-optimization/57268
9962 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9963 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9965 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9967 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
9968 (ix86_avoid_lea_for_addr): Likewise.
9969 (exact_dependency_1): Likewise.
9970 (ix86_adjust_cost): Likewise.
9971 (swap_top_of_ready_list): Fix formatting and !reload_completed check
9973 (do_reorder_for_imul): Fix typo, formatting and
9974 !reload_completed check removed.
9975 (ix86_sched_reorder): Fix typo and formatting.
9976 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
9979 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
9981 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
9983 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
9985 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
9986 <STRING_CST>: Likewise.
9987 <VECTOR_CST>: Likewise.
9989 2013-06-01 Janus Weil <janus@gcc.gnu.org>
9990 Mikael Morin <mikael@gcc.gnu.org>
9992 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
9993 * config.in: Regenerated.
9994 * configure: Regenerated.
9996 2013-06-01 Jan Hubicka <jh@suse.cz>
9999 * cgraphunit.c (compile): When weakref is not supported,
10000 set up transparent aliases before final output pass.
10001 * varasm.c (assemble_alias): Do not try to do it here.
10003 2013-06-01 Jan Hubicka <jh@suse.cz>
10005 PR middle-end/57467
10006 * passes.c (for_per_function): Skip unanalyzed functions.
10008 2013-06-01 Jan Hubicka <jh@suse.cz>
10010 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
10011 (lto_symtab_merge_symbols_1): ... this one.
10012 (lto_symtab_merge_cgraph_nodes): Rename to ...
10013 (lto_symtab_merge_symbols): ... this one; simplify.
10014 * cgraph.c (same_body_aliases_done): Rename to ...
10015 (cpp_implicit_aliases_done): ... this one.
10016 (cgraph_create_function_alias): Update.
10017 (cgraph_same_body_alias): Update.
10018 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
10019 (verify_edge_corresponds_to_fndecl): Simplify.
10020 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
10021 (cgraph_node): Remove same_body_alias.
10022 (varpool_node): Remove alias_of and extra_name_alias.
10023 (same_body_aliases_done): Rename to ..
10024 (cpp_implicit_aliases_done): ... this one.
10025 (symtab_alias_ultimate_target): Add default parameter.
10026 (symtab_resolve_alias): New function.
10027 (fixup_same_cpp_alias_visibility): Declare.
10028 (cgraph_function_node): Add default parameter.
10029 (cgraph_node_asm_name): Likewise.
10030 (cgraph_function_or_thunk_node): Add default parameter; do
10031 not ICE when it is NULL.
10032 (varpool_variable_node): Likewise.
10033 * tree-emutls.c (create_emultls_var): Update.
10034 (ipa_lower_emutls): Update.
10035 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
10036 (cgraph_reset_node): Reset alias info.
10037 (cgraph_finalize_function): Update.
10038 (fixup_same_cpp_alias_visibility): Move to symtab.c.
10039 (analyze_function): Simplify.
10040 (cgraph_process_same_body_aliases): Simplify.
10041 (analyze_functions): Fixup same body aliases.
10042 (handle_alias_pairs): Simplify.
10043 (assemble_thunk): Update.
10044 (assemble_thunks_and_aliases): Update.
10045 (output_weakrefs): Rewrite.
10046 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
10047 (lto_output_varpool_node): Likewise.
10048 (compute_ltrans_boundary): Remve assert.
10049 (get_alias_symbol): New functoin.
10050 (input_node): Rewrite alias handling.
10051 (input_varpool_node): Likewise.
10052 * ipa-pure-const.c (propagate_pure_const): Fix formating.
10053 * ipa.c (process_references): Handle weakrefs correctly.
10054 (symtab_remove_unreachable_nodes): Likewise.
10055 * trans-mem.c (get_cg_data): Update.
10056 (ipa_tm_create_version_alias): Update.
10057 (ipa_tm_execute): Update.
10058 * symtab.c (dump_symtab_base): Dump aliases.
10059 (verify_symtab_base): Verify aliases.
10060 (symtab_node_availability): New function.
10061 (symtab_alias_ultimate_target): Simplify.
10062 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
10063 handle all the fixup cases.
10064 (symtab_resolve_alias): New function.
10065 * passes.c (ipa_write_summaries): Handle weakrefs.
10066 * varpool.c (varpool_analyze_node): Simplify.
10067 (assemble_aliases): Update.
10068 (varpool_create_variable_alias): Simplify.
10069 (varpool_extra_name_alias): Simplify.
10070 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
10071 (lto_symtab_merge_symbols): ... this one.
10073 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
10076 PR rtl-optimization/57268
10077 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10078 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10080 2013-06-01 Tobias Burnus <burnus@net-b.de>
10082 Partially reverted:
10083 2013-05-31 Tobias Burnus <burnus@net-b.de>
10085 PR middle-end/57073
10086 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
10089 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10091 PR rtl-optimization/57268
10092 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10093 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10095 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10097 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
10098 unordered comparison operators when -fno-trapping-math is in effect
10100 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
10101 and implement unordered comparison operators properly on the e500.
10103 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10105 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
10106 for constant scalar integers.
10107 (simplify_relational_operation_1): Likewise.
10109 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
10111 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
10112 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
10115 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10116 Igor Zamyatin <igor.zamyatin@intel.com>
10118 Silvermont (SLM) architecture performance tuning.
10119 * config/i386/i386.h (enum ix86_tune_indices): Add
10120 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
10121 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
10123 * config/i386/i386.c (initial_ix86_tune_features)
10124 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
10125 (ix86_lea_outperforms): Handle Silvermont tuning.
10126 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
10128 (ix86_use_lea_for_mov): Likewise.
10129 (ix86_avoid_lea_for_addr): Likewise.
10130 (ix86_lea_for_add_ok): Likewise.
10131 (exact_dependency_1): New function.
10132 (exact_store_load_dependency): Likewise.
10133 (ix86_adjust_cost): Handle Silvermont tuning.
10134 (do_reoder_for_imul): Likewise.
10135 (swap_top_of_ready_list): New function.
10136 (ix86_sched_reorder): Changed to handle Silvermont tuning.
10138 * config/i386/i386.md (peepholes that split memory operand in fp
10141 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10143 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
10144 Remove un-necessary braces.
10146 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10148 * config/aarch64/aarch64.c (aarch64_classify_symbol):
10149 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
10151 2013-05-31 Tobias Burnus <burnus@net-b.de>
10153 PR middle-end/57073
10154 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
10156 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10159 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
10160 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
10161 * config/arm/neon.md (iordi3_neon): Remove.
10162 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
10163 * config/arm/predicates.md (imm_for_neon_logic_operand):
10164 Move to earlier in the file.
10165 (neon_logic_op2): Likewise.
10166 (arm_iordi_operand_neon): New predicate.
10168 2013-05-31 Richard Biener <rguenther@suse.de>
10170 PR tree-optimization/57478
10171 PR tree-optimization/57453
10172 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
10175 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10177 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
10178 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
10180 2013-05-30 Tobias Burnus <burnus@net-b.de>
10181 Thomas Koenig <tkoenig@gcc.gnu.org>
10183 PR middle-end/57073
10184 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
10185 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
10187 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
10189 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
10191 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
10193 * target.def (register_usage_leveling_p): New hook.
10194 * targhooks.c (default_register_usage_leveling_p): New.
10195 * targhooks.h (default_register_usage_leveling_p): New prototype.
10196 * lra-assigns.c (register_usage_leveling_p): Use the hook.
10197 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
10198 * doc/tm.texi: Update.
10199 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
10201 2013-05-30 Ian Bolton <ian.bolton@arm.com>
10203 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
10204 (*insv_reg<mode>): New define_insn.
10206 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
10208 PR rtl-optimization/57439
10209 * postreload.c (move2add_valid_value_p): Check that we have
10210 a zero subreg_regno_offset when accessing the register in
10211 the requested mode.
10213 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10214 Igor Zamyatin <igor.zamyatin@intel.com>
10216 Silvermont (SLM) architecture pipeline model, tuning and
10218 * config.gcc: Add slm config options and target.
10220 * config/i386/slm.md: New.
10222 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
10224 * config/i386/i386-c.c (ix86_target_macros_internal): New case
10226 (ix86_target_macros_internal): Likewise.
10228 * config/i386/i386.c (slm_cost): New cost.
10229 (m_SLM): New macro flag.
10230 (initial_ix86_tune_features): Set m_SLM.
10231 (x86_accumulate_outgoing_args): Likewise.
10232 (x86_arch_always_fancy_math_387): Likewise.
10233 (processor_target_table): Add slm cost.
10234 (cpu_names): Add slm cpu name.
10235 (x86_option_override_internal): Set SLM ISA.
10236 (ix86_issue_rate): New case PROCESSOR_SLM.
10237 (ia32_multipass_dfa_lookahead): Likewise.
10238 (fold_builtin_cpu): Add slm.
10240 * config/i386/i386.h (TARGET_SLM): New target macro.
10241 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
10242 (processor_type): Add PROCESSOR_SLM.
10244 * config/i386/i386.md (cpu): Add new value "slm".
10245 (slm.md): Include slm.md.
10247 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
10248 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10250 * config/arm/arm-protos.h: Add and update function protos.
10251 * config/arm/arm.c (use_simple_return_p): New added.
10252 (thumb2_expand_return): Check simple_return flag.
10253 * config/arm/arm.md: Add simple_return and conditional simple_return.
10254 * config/arm/iterators.md: Add iterator for return and simple_return.
10256 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10258 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
10259 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
10260 (arm_emit_vfp_multi_reg_pop): Likewise.
10261 (thumb2_emit_ldrd_pop): Likewise.
10262 (arm_expand_epilogue): Add misc REG_CFA notes.
10263 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
10265 2013-05-29 Lawrence Crowl <crowl@google.com>
10267 * config/arm/t-arm: Update for below.
10269 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
10270 Change type to hash_table. Update dependent calls and types.
10272 * config/i386/t-cygming: Update for below.
10274 * config/i386/t-interix: Update for below.
10276 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
10277 Change type to hash_table. Update dependent calls and types.
10278 (i386_find_on_wrapper_list::wrappers): Likewise.
10280 * config/ia64/t-ia64: Update for below.
10282 * config/ia64/ia64.c (bundle_state_table):
10283 Change type to hash_table. Update dependent calls and types.
10285 * config/mips/mips.c (mips_reorg_process_insns::htab):
10286 Change type to hash_table. Update dependent calls and types.
10288 * config/sol2.c (solaris_comdat_htab):
10289 Change type to hash_table. Update dependent calls and types.
10291 * config/t-sol2: Update for above.
10293 2013-05-29 Teresa Johnson <tejohnson@google.com>
10295 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
10296 functions are not yet marked as defined.
10298 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
10299 Pat Haugen <pthaugen@us.ibm.com>
10300 Peter Bergner <bergner@vnet.ibm.com>
10302 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
10306 (vrotl<mode>3): Likewise.
10307 (vashl<mode>3): Likewise.
10308 (vlshr<mode>3): Likewise.
10309 (vashr<mode>3): Likewise.
10311 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10312 support for power8 V2DI builtins.
10314 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
10315 power8 V2DI builtins.
10316 (vupkhsw): Likewise.
10317 (vupklsw): Likewise.
10318 (vaddudm): Likewise.
10319 (vminsd): Likewise.
10320 (vmaxsd): Likewise.
10321 (vminud): Likewise.
10322 (vmaxud): Likewise.
10323 (vpkudum): Likewise.
10324 (vpksdss): Likewise.
10325 (vpkudus): Likewise.
10326 (vpksdus): Likewise.
10331 (vsubudm): Likewise.
10332 (vcmpequd): Likewise.
10333 (vcmpgtsd): Likewise.
10334 (vcmpgtud): Likewise.
10335 (vcmpequd_p): Likewise.
10336 (vcmpgtsd_p): Likewise.
10337 (vcmpgtud_p): Likewise.
10338 (vupkhsw): Likewise.
10339 (vupklsw): Likewise.
10340 (vaddudm): Likewise.
10341 (vmaxsd): Likewise.
10342 (vmaxud): Likewise.
10343 (vminsd): Likewise.
10344 (vminud): Likewise.
10345 (vpksdss): Likewise.
10346 (vpksdus): Likewise.
10347 (vpkudum): Likewise.
10348 (vpkudus): Likewise.
10353 (vsubudm): Likewise.
10355 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
10356 support for power8 V2DI instructions.
10358 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
10359 power8 V2DI instructions. Combine pack and unpack insns to use an
10360 iterator for each mode. Check whether a particular mode supports
10361 Altivec instructions instead of just checking TARGET_ALTIVEC.
10362 (UNSPEC_VPKUWUM): Likewise.
10363 (UNSPEC_VPKSHSS): Likewise.
10364 (UNSPEC_VPKSWSS): Likewise.
10365 (UNSPEC_VPKUHUS): Likewise.
10366 (UNSPEC_VPKSHUS): Likewise.
10367 (UNSPEC_VPKUWUS): Likewise.
10368 (UNSPEC_VPKSWUS): Likewise.
10369 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
10370 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
10371 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
10372 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
10373 (UNSPEC_VUPKHSB): Likewise.
10374 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
10375 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
10376 (UNSPEC_VUPKHSH): Likewise.
10377 (UNSPEC_VUPKLSB): Likewise.
10378 (UNSPEC_VUPKLSH): Likewise.
10380 (VI_char): Likewise.
10381 (VI_scalar): Likewise.
10382 (VI_unit): Likewise.
10384 (VP_small): Likewise.
10385 (VP_small_lc): Likewise.
10386 (VU_char): Likewise.
10387 (add<mode>3): Likewise.
10388 (altivec_vaddcuw): Likewise.
10389 (altivec_vaddu<VI_char>s): Likewise.
10390 (altivec_vadds<VI_char>s): Likewise.
10391 (sub<mode>3): Likewise.
10392 (altivec_vsubcuw): Likewise.
10393 (altivec_vsubu<VI_char>s): Likewise.
10394 (altivec_vsubs<VI_char>s): Likewise.
10395 (altivec_vavgs<VI_char>): Likewise.
10396 (altivec_vcmpbfp): Likewise.
10397 (altivec_eq<mode>): Likewise.
10398 (altivec_gt<mode>): Likewise.
10399 (altivec_gtu<mode>): Likewise.
10400 (umax<mode>3): Likewise.
10401 (smax<mode>3): Likewise.
10402 (umin<mode>3): Likewise.
10403 (smin<mode>3): Likewise.
10404 (altivec_vpkuhum): Likewise.
10405 (altivec_vpkuwum): Likewise.
10406 (altivec_vpkshss): Likewise.
10407 (altivec_vpkswss): Likewise.
10408 (altivec_vpkuhus): Likewise.
10409 (altivec_vpkshus): Likewise.
10410 (altivec_vpkuwus): Likewise.
10411 (altivec_vpkswus): Likewise.
10412 (altivec_vpks<VI_char>ss): Likewise.
10413 (altivec_vpks<VI_char>us): Likewise.
10414 (altivec_vpku<VI_char>us): Likewise.
10415 (altivec_vpku<VI_char>um): Likewise.
10416 (altivec_vrl<VI_char>): Likewise.
10417 (altivec_vsl<VI_char>): Likewise.
10418 (altivec_vsr<VI_char>): Likewise.
10419 (altivec_vsra<VI_char>): Likewise.
10420 (altivec_vsldoi_<mode>): Likewise.
10421 (altivec_vupkhsb): Likewise.
10422 (altivec_vupkhs<VU_char>): Likewise.
10423 (altivec_vupkls<VU_char>): Likewise.
10424 (altivec_vupkhsh): Likewise.
10425 (altivec_vupklsb): Likewise.
10426 (altivec_vupklsh): Likewise.
10427 (altivec_vcmpequ<VI_char>_p): Likewise.
10428 (altivec_vcmpgts<VI_char>_p): Likewise.
10429 (altivec_vcmpgtu<VI_char>_p): Likewise.
10430 (abs<mode>2): Likewise.
10431 (vec_unpacks_hi_v16qi): Likewise.
10432 (vec_unpacks_hi_v8hi): Likewise.
10433 (vec_unpacks_lo_v16qi): Likewise.
10434 (vec_unpacks_hi_<VP_small_lc>): Likewise.
10435 (vec_unpacks_lo_v8hi): Likewise.
10436 (vec_unpacks_lo_<VP_small_lc>): Likewise.
10437 (vec_pack_trunc_v8h): Likewise.
10438 (vec_pack_trunc_v4si): Likewise.
10439 (vec_pack_trunc_<mode>): Likewise.
10441 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
10443 (vec_vmaxsd): Likewise.
10444 (vec_vmaxud): Likewise.
10445 (vec_vminsd): Likewise.
10446 (vec_vminud): Likewise.
10447 (vec_vpksdss): Likewise.
10448 (vec_vpksdus): Likewise.
10449 (vec_vpkudum): Likewise.
10450 (vec_vpkudus): Likewise.
10451 (vec_vrld): Likewise.
10452 (vec_vsld): Likewise.
10453 (vec_vsrad): Likewise.
10454 (vec_vsrd): Likewise.
10455 (vec_vsubudm): Likewise.
10456 (vec_vupkhsw): Likewise.
10457 (vec_vupklsw): Likewise.
10459 2013-05-29 Jan Hubicka <jh@suse.cz>
10461 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
10462 flags; reorder rest of fields in more consistent way.
10463 (varpool_node): Remove analyzed, finalized and alias.
10464 (cgraph_ndoe): Likewise.
10465 (symtab_alias_ultimate_target): New function.
10466 (cgraph_function_node): Move offline.
10467 (cgraph_reset_node): Declare.
10468 (cgraph_comdat_can_be_unshared_p): Remove.
10469 (varpool_remove_initializer): Declare.
10470 (varpool_first_defined_variable, varpool_next_defined_variable
10471 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10472 (cgraph_function_with_gimple_body_p): Update.
10473 (varpool_all_refs_explicit_p): Update.
10474 (symtab_alias_target): New function.
10475 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10476 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
10477 (cgraph_function_or_thunk_node): Simplify using
10478 symtab_alias_ultimate_target.
10479 (varpool_variable_node): Likewise.
10480 * cgraph.c (cgraph_create_function_alias): Update.
10481 (cgraph_add_thunk): Update.
10482 (cgraph_remove_node): Update.
10483 (dump_cgraph_node): Do not dump removed flags.
10484 (cgraph_function_body_availability): Update.
10485 (cgraph_propagate_frequency): Update.
10486 (verify_cgraph_node): Check sanity of local flag.
10487 (cgraph_function_node): Move here from cgraph.h; revamp for
10488 cgraph_function_or_thunk_node.
10489 * lto-symtab.c (lto_varpool_replace_node): Update.
10490 (lto_symtab_resolve_can_prevail_p): Update.
10491 (lto_symtab_merge_cgraph_nodes): Update.
10492 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10493 propagate_constants_accross_call, devirtualization_time_bonus,
10494 ipcp_propagate_stage): Update.
10495 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
10496 * ipa-inline-transform.c (clone_inlined_nodes,
10497 preserve_function_body_p): Update.
10498 * ipa-reference.c (propagate): Update.
10499 (write_node_summary_p): Update.
10500 * toplev.c (wrapup_global_declaration_2): Update.
10501 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10502 (analyze_function) ... this one.
10503 (cgraph_process_new_functions): Update.
10504 (cgraph_reset_node): Export.
10505 (cgraph_finalize_function): Update.
10506 (cgraph_add_new_function): Update.
10507 (process_function_and_variable_attributes): Update.
10508 (varpool_finalize_decl): Update.
10509 (symbol_finalized): Remove.
10510 (symbol_finalized_and_needed): Rename to ...
10511 (symbol_defined_and_needed): ... update.
10512 (cgraph_analyze_functions): Update.
10513 (handle_alias_pairs): Update.
10514 (mark_functions_to_output): Update.
10515 (assemble_thunk): Update.
10516 (output_in_order): Update.
10517 (output_weakrefs): Update.
10518 (finalize_compilation_unit): Update.
10519 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10520 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10521 input_node, input_varpool_node): Update.
10522 * dbxout.c (dbxout_expand_expr): Update.
10523 * cgraphclones.c (cgraph_clone_node): Update.
10524 (cgraph_copy_node_for_versioning): Update.
10525 (cgraph_materialize_clone): Update.
10526 (cgraph_materialize_all_clones): Update.
10527 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10528 propagate_pure_const, propagate_nothrow): Update.
10529 * lto-streamer-out.c (lto_output, write_symbol): Update.
10530 * ipa-utils.c (ipa_reverse_postorder): Update.
10531 * ipa-inline.c (can_inline_edge_p): Update.
10532 (update_caller_keys, ipa_inline): Update.
10533 * dwarf2out.c (reference_to_unused,
10534 premark_types_used_by_global_vars_helper): Update.
10535 * tree-eh.c (tree_could_trap_p): Update.
10536 * ipa-split.c (consider_split, execute_split_functions): Update.
10537 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
10538 has_addr_references_p): Update; move ahead in file for better
10540 (process_references): Simplify.
10541 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10542 bodies are removed.
10543 (cgraph_comdat_can_be_unshared_p): Make static.
10544 (cgraph_externally_visible_p): Update.
10545 (varpool_externally_visible_p): Update.
10546 (function_and_variable_visibility): Update.
10547 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10548 ipa_tm_mark_force_output_node): Update.
10549 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10550 estimate_edge_devirt_benefit, inline_generate_summary,
10551 inline_write_summary): Update.
10552 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10553 * ipa-prop.c (ipa_compute_jump_functions): Update.
10554 (ipa_print_node_params, ipa_prop_read_section,
10555 ipa_update_after_lto_read, read_replacements_section): Update.
10556 * varasm.c (mark_decl_referenced): Update.
10557 (assemble_alias, dump_tm_clone_pairs): Update.
10558 * tree-inline.c (copy_bb): Update.
10559 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
10561 * symtab.c (dump_symtab_base): Print new flags.
10562 (verify_symtab_base): Verify new flags.
10563 (symtab_alias_ultimate_target): New function.
10564 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
10565 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
10567 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
10569 * i386.c (ix86_get_function_versions_dispatcher,
10570 ix86_generate_version_dispatcher_body): Update.
10571 (fold_builtin_cpu): Use varpool_add_new_variable.
10572 * varpool.c (varpool_remove_initializer): Break out from ...
10573 (varpool_remove_node): ... this one.
10574 (dump_varpool_node, varpool_node_for_asm,
10575 cgraph_variable_initializer_availability, varpool_analyze_node,
10576 varpool_assemble_decl, varpool_remove_unreferenced_decls,
10577 varpool_finalize_named_section_flags, varpool_create_variable_alias):
10580 2013-05-29 Jan Hubicka <jh@suse.cz>
10582 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
10584 2013-05-29 Easwaran Raman <eraman@google.com>
10586 PR tree-optimization/57442
10587 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
10588 when control exits the main loop.
10590 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
10592 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
10594 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
10595 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
10596 * rx/t-rx: Add rx100 under multi library matches option for nofpu
10599 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10601 PR tree-optimization/57441
10602 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
10603 Don't limit size of incr_vec to number of candidates.
10605 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
10607 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
10608 and mips16 directories.
10609 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
10610 (MULTILIB_DIRNAMES): Ditto.
10611 (MULTILIB_EXCEPTIONS): Add new exceptions.
10612 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
10613 (MULTILIB_DIRNAMES): Ditto.
10614 (MULTILIB_EXCEPTIONS): Add new exceptions.
10616 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10617 Marcus Shawcroft <marcus.shawcroft@arm.com>
10619 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
10620 SYMBOL_TINY_ABSOLUTE.
10621 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
10622 SYMBOL_TINY_ABSOLUTE.
10623 (aarch64_expand_mov_immediate): Likewise.
10624 (aarch64_classify_symbol): Likewise.
10625 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10626 Permit SYMBOL_TINY_ABSOLUTE.
10627 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10629 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10630 Marcus Shawcroft <marcus.shawcroft@arm.com>
10632 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10633 Refactor if/switch. Replace gcc_assert with if.
10635 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10637 * config/i386/i386.c (initial_ix86_tune_features): Enable
10638 FP Reassociation for AMD bdver1 and bdver2.
10640 2013-05-29 Martin Jambor <mjambor@suse.cz>
10642 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10643 and IMAGPART_EXPR do not occur within other handled_components.
10645 2013-05-29 Richard Biener <rguenther@suse.de>
10647 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10648 access on whether the use is in the BB we currently try to
10650 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10651 vectorize to vect_bb_slp_scalar_cost.
10653 2013-05-29 Richard Biener <rguenther@suse.de>
10655 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10656 computing scalar cost offsetted by stmts that are kept live
10658 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10659 for computation of scalar cost.
10661 2013-05-28 Steve Ellcey <sellcey@mips.com>
10663 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10665 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10667 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10668 array notation built-in reduction functions.
10669 * doc/passes.texi (Passes): Added documentation about changes done
10671 * doc/invoke.texi (C Dialect Options): Added documentation about
10672 the -fcilkplus flag.
10673 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10674 (BUILTINS_DEF): Depend on cilkplus.def.
10675 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10676 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10677 * cilkplus.def: New file.
10679 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
10681 PR rtl-optimization/57439
10682 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10684 2013-05-28 Easwaran Raman <eraman@google.com>
10686 PR tree-optimization/57337
10687 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10688 (find_insert_point): Correctly identify the insertion point
10689 when two statements with the same UID is compared.
10691 2013-05-28 Richard Biener <rguenther@suse.de>
10693 PR tree-optimization/56787
10694 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10695 from the list of data references.
10696 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10698 (vect_analyze_loop_operations): Likewise.
10699 (vect_transform_loop): Remove clobbers.
10701 2013-05-28 Martin Jambor <mjambor@suse.cz>
10703 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10704 and REALPART_EXPRs have scalar type.
10706 2013-05-28 Richard Biener <rguenther@suse.de>
10708 PR tree-optimization/57411
10709 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10711 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10712 virtual operand propagation.
10714 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10716 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10717 destination register for bmasksi_vis.
10718 (vector_init_bshuffle): Likewise.
10719 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10721 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10723 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10724 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10725 mode if the instruction isn't available in the original mode.
10726 * config/sparc/sparc.opt (mfix-ut699): New option.
10727 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10728 (divdf3): Turn into expander.
10729 (divdf3_nofix): New insn.
10730 (divdf3_fix): Likewise.
10731 (divsf3): Disable if -mfix-ut699.
10732 (sqrtdf2): Turn into expander.
10733 (sqrtdf2_nofix): New insn.
10734 (sqrtdf2_fix): Likewise.
10735 (sqrtsf2): Disable if -mfix-ut699.
10737 2013-05-27 Richard Biener <rguenther@suse.de>
10739 PR middle-end/57412
10740 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10741 block for the new loop.
10743 2013-05-27 Richard Biener <rguenther@suse.de>
10745 PR tree-optimization/57343
10746 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10747 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10748 (number_of_iterations_cond): Do not build the folded tree.
10750 2013-05-27 Richard Biener <rguenther@suse.de>
10753 PR middle-end/57381
10754 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10755 OEP_CONSTANT_ADDRESS_OF retained.
10757 PR tree-optimization/57417
10758 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10759 for unchanged base.
10760 (set_ssa_val_to): Compare addresses using
10761 get_addr_base_and_unit_offset.
10763 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
10765 PR rtl-optimization/56833
10766 * postreload.c (move2add_record_mode): New function.
10767 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10768 (move2add_use_add2_insn): Use move2add_record_sym_value.
10769 (move2add_use_add3_insn): Likewise.
10770 (reload_cse_move2add): Use move2add_valid_value_p and
10771 move2add_record_mode. Invalidate call-clobbered and REG_INC
10772 affected regs by setting reg_mode to VOIDmode.
10773 (move2add_note_store): Don't pretend the inside of a SUBREG is
10774 the actual destination. Invalidate single/leading registers by
10775 setting reg_mode to VOIDmode.
10776 Use move2add_record_sym_value, move2add_valid_value_p and
10777 move2add_record_mode.
10779 2013-05-27 Richard Biener <rguenther@suse.de>
10781 PR tree-optimization/57396
10782 * tree-affine.c (double_int_constant_multiple_p): Properly
10783 return false for val == 0 and div != 0.
10785 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10787 * config/mips/mips.h: Use #elif in preprocessor conditions.
10789 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10792 * config/mips/constraints.md (kl): New constraint.
10793 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10794 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10795 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10796 from LO for MIPS16.
10797 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10799 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10802 * config/mips/mips.c (mips_can_inline_p): New function.
10803 (TARGET_CAN_INLINE_P): Define.
10805 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10807 * sched-int.h (ds_t, dw_t): Make unsigned int.
10808 Fix documentation that describes how all the ds_t bits are used.
10809 Reserve the last bit for delayed-branch scheduling.
10810 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10811 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10812 (gen_dep_weak_1): Remove prototype.
10813 * sched-deps.c (get_dep_weak_1): Make static.
10814 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10815 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10816 * doc/tm.texi: Regenerate.
10817 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
10819 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
10822 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
10824 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
10825 Sandra Loosemore <sandra@codesourcery.com>
10827 * config.gcc (powerpc-*): Allow native for with-cpu.
10829 2013-05-24 Jeff Law <law@redhat.com>
10831 PR tree-optimization/57124
10832 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
10833 conversion feeding a condition if the range has an overflow
10834 if -fstrict-overflow. Add warnings for when we do make the
10837 2013-05-24 Dehao Chen <dehao@google.com>
10839 * tree-cfg.c (locus_discrim_map): Fix the typo.
10840 (locus_discrim_hasher): Likewise.
10841 (locus_discrim_hasher::hash): Likewise.
10842 (locus_discrim_hasher::equal): Likewise.
10844 2013-05-24 Martin Jambor <mjambor@suse.cz>
10846 PR tree-optimization/57294
10847 * cgraph.h (ipa_record_stmt_references): Declare.
10848 * cgraphbuild.c (ipa_record_stmt_references): New function.
10849 (build_cgraph_edges): Use ipa_record_stmt_references.
10850 (rebuild_cgraph_edges): Likewise.
10851 (cgraph_rebuild_references): Likewise.
10852 * ipa-prop.c (ipa_modify_call_arguments): Discard references
10853 associated with the old statement and build references from the
10854 newly built statements.
10855 * ipa-ref.c (ipa_remove_stmt_references): New function.
10856 * ipa-ref.h (ipa_remove_stmt_references): Declare.
10858 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
10860 * lra-constraints.c (emit_spill_move): Use smaller mode for
10862 (check_and_process_move): Consider mem-reg moves for secondary
10864 (curr_insn_transform): Don't lose insns emitted before for
10865 secondary memory moves.
10866 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
10867 reg set up in the current insn.
10869 2013-05-24 Dehao Chen <dehao@google.com>
10871 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
10873 (locus_descrim_hasher::equal): Likewise.
10874 (build_gimple_cfg): New discriminator assignment algorithm.
10875 (make_edges): Likewise.
10876 (next_discriminator_for_locus): Likewise.
10877 (same_line_p): Likewise.
10878 (assign_discriminators): Likewise.
10879 (make_cond_expr_edges): Likewise.
10880 (make_gimple_switch_edges): Likewise.
10881 (make_goto_expr_edges): Likewise.
10882 (make_gimple_asm_edges): Likewise.
10884 2013-05-24 Ian Bolton <ian.bolton@arm.com>
10886 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
10887 X format specifier to only display bottom 16 bits.
10888 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
10889 immediate to match for operand 2, since it will be masked.
10891 2013-05-24 Richard Biener <rguenther@suse.de>
10893 PR tree-optimization/57287
10894 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
10895 all SSA names that occur in abnormal PHIs.
10897 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
10900 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
10901 that index is not negative.
10903 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
10905 PR rtl-optimization/55177
10906 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
10907 (simplify_byte_swapping_operation): New.
10908 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
10909 (simplify_relational_operation_1): Deal with BSWAP.
10911 2013-05-23 Richard Henderson <rth@redhat.com>
10914 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
10915 (ix86_reorg): Call it.
10917 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
10920 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
10921 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
10922 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
10924 2013-05-23 Christian Bruel <christian.bruel@st.com>
10927 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
10929 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10930 Marcus Shawcroft <marcus.shawcroft@arm.com>
10932 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
10933 * config/aarch64/constraints.md (Usa): Remove.
10934 * doc/md.texi (AArch64 Usa): Remove.
10936 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10937 Marcus Shawcroft <marcus.shawcroft@arm.com>
10939 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
10940 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
10941 * config/aarch64/predicates.md (aarch64_const_address): Remove.
10942 (aarch64_mov_operand): Use aarch64_mov_operand_p.
10944 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
10946 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
10947 instruction (AdvSIMD).
10948 * config/aarch64/aarch64-builtins.c
10949 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
10950 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
10952 2013-05-23 Martin Jambor <mjambor@suse.cz>
10954 PR middle-end/57347
10955 * tree.h (contains_bitfld_component_ref_p): Declare.
10956 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
10957 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
10959 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
10960 not access a bit-field. Assert all final offsets are byte-aligned.
10962 2013-05-23 Richard Biener <rguenther@suse.de>
10964 PR tree-optimization/57380
10965 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
10966 least one invariant or re-used load.
10967 * passes.c (init_optimization_passes): Move pass_phiprop before
10970 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
10972 * config/aarch64/aarch64-simd.md
10973 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
10975 2013-05-23 Richard Biener <rguenther@suse.de>
10977 PR middle-end/57381
10978 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10979 OEP_CONSTANT_ADDRESS_OF retained.
10981 2013-05-23 Jakub Jelinek <jakub@redhat.com>
10983 PR middle-end/57344
10984 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
10985 don't lower unit. Handle unit not being always BITS_PER_WORD.
10987 2013-05-23 Richard Biener <rguenther@suse.de>
10989 PR rtl-optimization/57341
10990 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
10991 instead of true_dependence.
10993 2013-05-22 David Malcolm <dmalcolm@redhat.com>
10995 * bb-reorder.c (branch_threshold): Make const.
10996 (exec_threshold): Ditto.
10998 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
10999 Pat Haugen <pthaugen@us.ibm.com>
11000 Peter Bergner <bergner@vnet.ibm.com>
11002 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
11003 documentation for the power8 crypto builtins.
11005 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
11007 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
11008 macros for defining power8 builtin functions.
11009 (BU_P8V_AV_2): Likewise.
11010 (BU_P8V_AV_P): Likewise.
11011 (BU_P8V_VSX_1): Likewise.
11012 (BU_P8V_OVERLOAD_1): Likewise.
11013 (BU_P8V_OVERLOAD_2): Likewise.
11014 (BU_CRYPTO_1): Likewise.
11015 (BU_CRYPTO_2): Likewise.
11016 (BU_CRYPTO_3): Likewise.
11017 (BU_CRYPTO_OVERLOAD_1): Likewise.
11018 (BU_CRYPTO_OVERLOAD_2): Likewise.
11019 (XSCVSPDP): Fix typo, point to the correct instruction.
11020 (VCIPHER): Add power8 crypto builtins.
11021 (VCIPHERLAST): Likewise.
11022 (VNCIPHER): Likewise.
11023 (VNCIPHERLAST): Likewise.
11024 (VPMSUMB): Likewise.
11025 (VPMSUMH): Likewise.
11026 (VPMSUMW): Likewise.
11027 (VPERMXOR_V2DI): Likewise.
11028 (VPERMXOR_V4SI: Likewise.
11029 (VPERMXOR_V8HI: Likewise.
11030 (VPERMXOR_V16QI: Likewise.
11031 (VSHASIGMAW): Likewise.
11032 (VSHASIGMAD): Likewise.
11033 (VPMSUM): Likewise.
11034 (VPERMXOR): Likewise.
11035 (VSHASIGMA): Likewise.
11037 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11038 __CRYPTO__ if the crypto instructions are available.
11039 (altivec_overloaded_builtins): Add support for overloaded power8
11042 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
11043 support for power8 crypto builtins.
11044 (builtin_function_type): Likewise.
11045 (altivec_init_builtins): Add support for builtins that take vector
11046 long long (V2DI) arguments.
11048 * config/rs6000/crypto.md: New file, define power8 crypto
11051 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11052 Pat Haugen <pthaugen@us.ibm.com>
11053 Peter Bergner <bergner@vnet.ibm.com>
11055 * doc/invoke.texi (Option Summary): Add power8 options.
11056 (RS/6000 and PowerPC Options): Likewise.
11058 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
11059 constraints.md instead of rs6000.h. Reorder w* constraints. Add
11060 wm, wn, wr documentation.
11062 * config/rs6000/constraints.md (wm): New constraint for VSX
11063 registers if direct move instructions are enabled.
11064 (wn): New constraint for no registers.
11065 (wq): New constraint for quad word even GPR registers.
11066 (wr): New constraint if 64-bit instructions are enabled.
11067 (wv): New constraint if power8 vector instructions are enabled.
11068 (wQ): New constraint for quad word memory locations.
11070 * config/rs6000/predicates.md (const_0_to_15_operand): New
11071 constraint for 0..15 for crypto instructions.
11072 (gpc_reg_operand): If VSX allow registers in VSX registers as well
11073 as GPR and floating point registers.
11074 (int_reg_operand): New predicate to match only GPR registers.
11075 (base_reg_operand): New predicate to match base registers.
11076 (quad_int_reg_operand): New predicate to match even GPR registers
11077 for quad memory operations.
11078 (vsx_reg_or_cint_operand): New predicate to allow vector logical
11079 operations in both GPR and VSX registers.
11080 (quad_memory_operand): New predicate for quad memory operations.
11081 (reg_or_indexed_operand): New predicate for direct move support.
11083 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
11084 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
11085 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
11086 (POWERPC_MASKS): Add power8 options.
11087 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
11090 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11091 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
11093 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
11094 (-mpower8-fusion): New power8 options.
11095 (-mpower8-fusion-sign): Likewise.
11096 (-mpower8-vector): Likewise.
11097 (-mcrypto): Likewise.
11098 (-mdirect-move): Likewise.
11099 (-mquad-memory): Likewise.
11101 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
11102 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
11104 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
11105 (rs6000_debug_vector_unit): Add p8_vector.
11106 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
11107 definitions. Also print fusion state.
11108 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
11109 (rs6000_builtin_mask_calculate): Add power8 builtin support.
11110 (rs6000_option_override_internal): Add support for power8.
11111 (rs6000_common_init_builtins): Add debugging for skipped builtins
11112 if -mdebug=builtin.
11113 (rs6000_adjust_cost): Add power8 support.
11114 (rs6000_issue_rate): Likewise.
11115 (insn_must_be_first_in_group): Likewise.
11116 (insn_must_be_last_in_group): Likewise.
11117 (force_new_group): Likewise.
11118 (rs6000_register_move_cost): Likewise.
11119 (rs6000_opt_masks): Likewise.
11121 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
11122 power8 capable assembler, default to power7 options.
11123 (TARGET_DIRECT_MOVE): Likewise.
11124 (TARGET_CRYPTO): Likewise.
11125 (TARGET_P8_VECTOR): Likewise.
11126 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
11127 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
11128 (VECTOR_MEM_P8_VECTOR_P): Likewise.
11129 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
11130 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
11131 (TARGET_XSCVDPSPN): Likewise.
11132 (TARGET_XSCVSPDPN): Likewsie.
11133 (TARGET_SYNC_HI_QI): Likewise.
11134 (TARGET_SYNC_TI): Likewise.
11135 (MASK_CRYPTO): Likewise.
11136 (MASK_DIRECT_MOVE): Likewise.
11137 (MASK_P8_FUSION): Likewise.
11138 (MASK_P8_VECTOR): Likewise.
11139 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
11140 temporary used by some of the direct move instructions to get two FP
11141 temporary registers does not force creation of a stack frame.
11142 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
11143 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
11144 that any VSX registers are tieable, even if they are also an
11145 Altivec vector mode.
11146 (r6000_reg_class_enum): Add wm, wr, wv constraints.
11147 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
11148 (RS6000_BTM_CRYPTO): Likewise.
11149 (RS6000_BTM_COMMON): Likewise.
11151 * config/rs6000/rs6000.md (cpu attribute): Add power8.
11152 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
11153 (enum rs6000_vector): Add power8 vector support.
11155 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11159 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
11160 (any_sibcall_could_use_r3): this and handle indirect calls.
11161 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
11163 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11165 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
11167 2013-05-22 Richard Biener <rguenther@suse.de>
11169 PR middle-end/57349
11170 * profile.c (branch_prob): Do not split blocks that are
11171 abnormally receiving from ECF_RETURNS_TWICE functions.
11173 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11175 * recog.c (offsettable_address_addr_space_p): Fix calculation of
11176 address mode. Move pointer mode initialization to the same place.
11178 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11180 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
11181 while it has any effect.
11183 2013-05-21 Easwaran Raman <eraman@google.com>
11185 PR tree-optimization/57322
11186 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
11187 UID of the statement added to the BB to be 1.
11189 2013-05-21 Jakub Jelinek <jakub@redhat.com>
11191 PR tree-optimization/57331
11192 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
11193 of conversion from pointer type to integral type with integer.
11195 2013-05-21 Martin Jambor <mjambor@suse.cz>
11198 * ipa-prop.c (ipa_read_node_info): Process param_used and
11199 controlled_uses in the same order as when writing.
11201 2013-05-21 Magnus Granberg <baldrick@free.fr>
11204 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
11206 2013-05-21 Richard Biener <rguenther@suse.de>
11208 PR tree-optimization/57318
11209 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
11210 estimate stmts with side-effects as likely eliminated.
11212 2013-05-21 Richard Biener <rguenther@suse.de>
11214 PR tree-optimization/57330
11215 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
11216 preserve the call stmts fntype.
11218 2013-05-21 Richard Biener <rguenther@suse.de>
11220 PR tree-optimization/57303
11221 * tree-ssa-sink.c (statement_sink_location): Improve killing
11222 stmt detection and properly handle self-assignments.
11224 2013-05-21 Christian Bruel <christian.bruel@st.com>
11226 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
11227 spanning registers. LEAF_REG_REMAP is supported only for contiguous
11228 registers. Set register size out of the PARALLEL loop.
11230 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
11233 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
11234 (*fmasf4, *fmasf4_media): New insns.
11236 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11238 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
11239 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
11240 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
11241 (mips_idiv_insns): Update the comments to say that the returned
11242 instruction counts are in units of BASE_INSN_LENGTH.
11243 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
11244 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
11245 using 2 rather than 4 as the length of indirect MIPS16 and
11246 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
11247 length of a NOP. Don't divide MIPS16 lengths by 2.
11248 (mips16_split_long_branches): Assume a branch is long if the
11249 length is greater than 4 rather than 8.
11250 * config/mips/mips.md (length): Give MIPS16 lengths directly,
11251 rather than multiplying them by 2. Multiply instruction counts
11252 by BASE_INSN_LENGTH rather than 4.
11253 (*jump_mips16, tls_get_tp_mips16_<mode>)
11254 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
11256 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11258 * config/mips/mips.md (extended_mips16): Remove branch case.
11259 (length): Remove duplicated extended_mips16 test.
11261 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11263 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
11265 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
11267 * recog.h (Recog_data): Rename to...
11268 (recog_data_d): ...this.
11269 (recog_data): Update accordingly.
11270 * recog.c (recog_data): Likewise.
11271 * reload.c (save_recog_data): Likewise.
11272 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
11273 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
11275 2013-05-17 Julian Brown <julian@codesourcery.com>
11277 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
11278 found in a REG_EQUAL note, invalidate it.
11280 2013-05-17 Easwaran Raman <eraman@google.com>
11282 * tree-ssa-reassoc.c (find_insert_point): New function.
11283 (insert_stmt_after): Likewise.
11284 (get_def_stmt): Likewise.
11285 (ensure_ops_are_available): Likewise.
11286 (not_dominated_by): Likewise.
11287 (rewrite_expr_tree): Do not move statements beyond what is
11288 necessary. Remove call to swap_ops_for_binary_stmt...
11289 (reassociate_bb): ... and move it here.
11290 (build_and_add_sum): Assign UIDs for new statements.
11291 (linearize_expr): Likewise.
11292 (do_reassoc): Renumber gimple statement UIDs.
11294 2013-05-17 Jan Hubicka <jh@suse.cz>
11296 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
11298 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
11299 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
11301 (output_weakrefs): Update.
11303 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
11304 Martin Jambor <mjambor@suse.cz>
11306 PR middle-end/57276
11307 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
11308 value that corresponds to the given aggval is found in values vector.
11310 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
11312 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
11313 sse, sse2, sse3, ssse3 and sse4a flags to options.
11315 2013-05-17 David Malcolm <dmalcolm@redhat.com>
11317 * gengtype-state.c: (s_expr_writer): New class, to handle
11318 prettifying of output layout of s-expressions.
11319 (state_writer): New class, to write out gtype.state.
11320 (state_written_type_count): Move this variable into member data of
11322 (s_expr_writer::s_expr_writer): New code: constructor for new class
11323 (state_writer::state_writer(): ditto
11324 (s_expr_writer::write_new_line): New function
11325 (s_expr_writer::write_any_indent): ditto
11326 (s_expr_writer::begin_s_expr): ditto
11327 (s_expr_writer::end_s_expr): ditto
11328 (write_state_fileloc): convert to method of state_writer...
11329 (state_writer:: write_state_fileloc): ...and use methods of
11330 s_expr_writer to write indentation into the gtype.state output file
11331 to visually represent the hierarchical structure of the list
11333 (write_state_fields): ditto, renaming to...
11334 (state_writer::write_state_fields)
11335 (write_state_a_string): ditto, renaming to...
11336 (state_writer::write_state_a_string)
11337 (write_state_string_option): ditto, renaming to...
11338 (state_writer::write_state_string_option)
11339 (write_state_type_option): ditto, renaming to...
11340 (state_writer::write_state_type_option)
11341 (write_state_nested_option): ditto, renaming to...
11342 (state_writer::write_state_nested_option)
11343 (write_state_option): ditto, renaming to...
11344 (state_writer::write_state_option)
11345 (write_state_options): ditto, renaming to...
11346 (state_writer::write_state_options)
11347 (write_state_lang_bitmap): ditto, renaming to...
11348 (state_writer::write_state_lang_bitmap)
11349 (write_state_version): ditto, renaming to...
11350 (state_writer::write_state_version)
11351 (write_state_scalar_type): ditto, renaming to...
11352 (state_writer::write_state_scalar_type)
11353 (write_state_string_type): ditto, renaming to...
11354 (state_writer::write_state_string_type)
11355 (write_state_undefined_type): ditto, renaming to...
11356 (state_writer::write_state_undefined_type)
11357 (write_state_struct_union_type): ditto, renaming to...
11358 (state_writer::write_state_struct_union_type)
11359 (write_state_struct_type): ditto, renaming to...
11360 (state_writer::write_state_struct_type)
11361 (write_state_user_struct_type): ditto, renaming to...
11362 (state_writer::write_state_user_struct_type)
11363 (write_state_lang_struct_type): ditto, renaming to...
11364 (state_writer::write_state_lang_struct_type)
11365 (write_state_param_struct_type): ditto, renaming to...
11366 (state_writer::write_state_param_struct_type)
11367 (write_state_pointer_type): ditto, renaming to...
11368 (state_writer::write_state_pointer_type)
11369 (write_state_array_type): ditto, renaming to...
11370 (state_writer::write_state_array_type)
11371 (write_state_gc_used): ditto, renaming to...
11372 (state_writer::write_state_gc_used)
11373 (write_state_common_type_content): ditto, renaming to...
11374 (state_writer::write_state_common_type_content)
11375 (write_state_type): ditto, renaming to...
11376 (state_writer::write_state_type)
11377 (write_state_pair_list): ditto, renaming to...
11378 (state_writer::write_state_pair_list)
11379 (write_state_pair): ditto, renaming to...
11380 (state_writer::write_state_pair)
11381 (write_state_typedefs): ditto, renaming to...
11382 (state_writer::write_state_typedefs)
11383 (write_state_structures): ditto, renaming to...
11384 (state_writer::write_state_structures)
11385 (write_state_param_structs): ditto, renaming to...
11386 (state_writer::write_state_param_structs)
11387 (write_state_variables): ditto, renaming to...
11388 (state_writer::write_state_variables)
11389 (write_state_srcdir): ditto, renaming to...
11390 (state_writer::write_state_srcdir)
11391 (write_state_files_list): ditto, renaming to...
11392 (state_writer::write_state_files_list)
11393 (write_state_languages): ditto, renaming to...
11394 (state_writer::write_state_languages)
11395 (write_state): create a state_writer instance and use it when
11396 writing out the state file
11398 2013-05-17 Mike Stump <mikestump@comcast.net>
11400 PR rtl-optimization/57304
11401 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
11402 accessing DF_REF_REAL_LOC.
11404 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11406 PR rtl-optimization/57281
11407 PR rtl-optimization/57300
11408 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
11409 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
11410 what the other splitter did if the registers are dead.
11412 2013-05-17 Richard Biener <rguenther@suse.de>
11414 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
11417 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11419 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
11422 2013-05-17 Marek Polacek <polacek@redhat.com>
11424 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
11425 length when doing non-zero store of storing '\0' to '\0'.
11427 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11429 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
11430 vect_external_def oprnd1 with loop_vinfo, try to emit
11431 optional cast, negation and and stmts on the loop preheader
11432 edge instead of into the pattern def seq.
11434 PR tree-optimization/57051
11435 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
11436 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
11438 2013-05-16 Nick Clifton <nickc@redhat.com>
11440 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
11441 (rl78_is_naked_func): New function.
11442 (rl78_expand_prologue): Skip prologue generation for naked functions.
11443 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
11444 * doc/extend.texi (naked): Add RL78 to the list of processors
11445 that supports this attribute.
11447 2013-05-16 Jeff Law <law@redhat.com>
11449 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
11451 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11453 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
11454 cache parameters using detect_caches_amd also for CYRIX,
11455 NSC and TM2 signatures.
11457 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11458 Dzianis Kahanovich <mahatma@eu.by>
11462 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11463 VIA/Centaur processors and determine their cache parameters
11464 using detect_caches_amd.
11466 2013-05-16 Teresa Johnson <tejohnson@google.com>
11468 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11469 (rtl_verify_edges): New function.
11470 (rtl_verify_bb_insns): Ditto.
11471 (rtl_verify_bb_pointers): Ditto.
11472 (rtl_verify_bb_insn_chain): Ditto.
11473 (rtl_verify_fallthru): Ditto.
11474 (rtl_verify_bb_layout): Ditto.
11475 (rtl_verify_flow_info_1): Outline checks into new functions.
11476 (rtl_verify_flow_info): Ditto.
11478 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
11480 * cfghooks.c (copy_bbs): Add update_dominance argument.
11481 * cfghooks.h (copy_bbs): Update prototype.
11482 * tree-cfg.c (gimple_duplicate_sese_region):
11483 Add update_dominance argument.
11484 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11485 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11486 gimple_duplicate_sese_region call.
11487 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11488 Update copy_bbs call.
11489 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11490 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11492 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11494 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11495 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11496 vect_recog_rotate_pattern.
11497 (vect_recog_rotate_pattern): New function.
11499 2013-05-16 Jason Merrill <jason@redhat.com>
11501 * Makefile.in (LLINKER): New variable.
11502 (mostlyclean): Remove link mutex.
11503 * configure.ac: Handle --enable-link-mutex.
11504 * lock-and-run.sh: New script.
11506 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11509 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11512 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11514 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11516 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11518 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11519 * config/arm/arm.c (next_consecutive_mem): New function.
11520 (gen_movmem_ldrd_strd): Likewise.
11521 * config/arm/arm.md (movmemqi): Update condition and code.
11522 (unaligned_loaddi, unaligned_storedi): New patterns.
11524 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11526 * config.gcc: Obsolete *-*-solaris2.9*.
11527 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11529 2013-05-16 Richard Biener <rguenther@suse.de>
11531 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11532 earlier, after GRAPHITE transforms and IV canonicalization.
11534 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11536 * omp-low.c (extract_omp_for_data): For collapsed loops,
11537 if at least one of the loops is known at compile time to
11538 iterate zero times, set count to 0.
11539 (expand_omp_regimplify_p): New function.
11540 (expand_omp_for_generic): For collapsed loops, if at least
11541 one of the loops isn't known to iterate at least once,
11542 add runtime check with setting count to 0.
11543 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11544 For unsigned types if it isn't known at compile time that
11545 the loop will iterate at least once, add runtime check to bypass
11546 the whole loop if initial condition isn't true.
11548 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11550 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11552 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
11554 PR middle-end/57286
11555 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
11556 transformations to avoid an infinite loop.
11558 2013-05-16 Marek Polacek <polacek@redhat.com>
11560 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
11562 2013-05-15 Leif Ekblad <leif@rdos.net>
11564 * config/i386/i386.c (ix86_decompose_address): Use
11565 DEFAULT_TLS_SEG_REG to access TLS segment register.
11566 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
11567 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
11568 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
11570 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
11573 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
11574 sibling calls to functions that would normally be lazily bound,
11575 unless $gp is call-clobbered.
11577 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
11579 * config/i386/i386.c (ix86_option_override_internal): Update
11580 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
11581 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
11582 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
11583 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
11585 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
11587 2013-05-15 Andreas Schwab <schwab@suse.de>
11589 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
11590 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
11593 2013-05-15 Teresa Johnson <tejohnson@google.com>
11595 * loop-unroll.c (report_unroll_peel): Check decision before
11596 emitting unroll/peel message.
11598 2013-05-15 Teresa Johnson <tejohnson@google.com>
11600 * function.h (has_bb_partition): New rtl_data flag.
11601 (bb_reorder_complete): Ditto.
11602 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
11603 instead of flag_reorder_blocks_and_partition.
11604 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
11605 with some enhancements.
11606 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
11607 * bb-reorder.c (connect_traces): Check for has_bb_partition
11608 instead of flag_reorder_blocks_and_partition.
11609 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
11610 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
11611 verify_hot_cold_block_grouping.
11612 (partition_hot_cold_basic_blocks): Set has_bb_partition.
11614 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11617 * config/arm/predicates.md (call_insn_operand): New predicate.
11618 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
11619 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
11620 if insn is not a tail call.
11621 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11623 * config/arm/arm.h (enum reg_class): New caller save register class.
11624 (REG_CLASS_NAMES): Likewise.
11625 (REG_CLASS_CONTENTS): Likewise.
11626 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11629 2013-05-15 Richard Biener <rguenther@suse.de>
11631 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11632 of MSG_OPTIMIZED_LOCATIONS.
11633 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11634 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11636 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11637 of MSG_OPTIMIZED_LOCATIONS.
11638 (execute_vect_slp): Likewise.
11639 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11640 (vect_create_cond_for_alias_checks): Likewise.
11641 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11642 (vect_recog_widen_mult_pattern): Likewise.
11643 (vect_recog_widen_sum_pattern): Likewise.
11644 (vect_recog_over_widening_pattern): Likewise.
11645 (vect_recog_widen_shift_pattern): Likewise.
11646 (vect_recog_vector_vector_shift_pattern): Likewise.
11647 (vect_recog_divmod_pattern): Likewise.
11648 (vect_recog_mixed_size_cond_pattern): Likewise.
11649 (vect_recog_bool_pattern): Likewise.
11650 (vect_pattern_recog_1): Likewise.
11652 2013-05-15 Martin Jambor <mjambor@suse.cz>
11654 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11655 non-functions to builtin_unreachable.
11656 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11657 correct when new direct edges were discovered.
11659 2013-05-15 Martin Jambor <mjambor@suse.cz>
11661 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11662 header, print symbol order instead of node uid, print more information
11663 about indirect edge targets.
11664 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11666 (ipa_make_edge_direct_to_target): Likewise.
11667 (remove_described_reference): Likewise.
11668 (propagate_controlled_uses): Likewise.
11669 (ipa_print_node_params): Also print symbol order.
11670 (ipcp_transform_function): Print symbol order instead of node uids.
11671 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11672 (cgraph_get_create_real_symbol_node): Likewise.
11673 * ipa-cp.c (print_lattice): Likewise.
11674 (print_all_lattices): Likewise.
11675 (determine_versionability): Likewise.
11676 (initialize_node_lattices): Likewise.
11677 (estimate_local_effects): Likewise.
11678 (update_profiling_info): Likewise.
11679 (create_specialized_node): Likewise.
11680 (perhaps_add_new_callers): Likewise.
11681 (decide_about_value): Likewise.
11682 (decide_whether_version_node): Likewise.
11683 (identify_dead_nodes): Likewise.
11684 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11685 (dump_inline_summary): Likewise.
11686 (estimate_node_size_and_time): Likewise.
11687 (inline_analyze_function): Likewise.
11688 * ipa-inline.c (report_inline_failed_reason): Likewise.
11689 (want_early_inline_function_p): Likewise.
11690 (edge_badness): Likewise.
11691 (update_edge_key): Likewise.
11692 (inline_small_functions): Likewise. Add dumping of order to two other
11694 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11695 instead of node uids.
11696 (propagate_pure_const): Likewise.
11697 (propagate_pure_const): Likewise.
11698 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11699 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11700 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11702 * tree-pretty-print.c (dump_function_header): Likewise.
11703 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11704 Print symbol order instead of node uids.
11706 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11708 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11709 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11711 2013-05-15 Richard Biener <rguenther@suse.de>
11713 PR tree-optimization/57275
11714 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11715 return value for fail to do runtime alias checks for gather loads.
11717 2013-05-15 Jan Hubicka <jh@suse.cz>
11721 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11722 weakrefs are not external.
11723 (lto_symtab_merge_decls): Fix thinko when dealing with
11724 non-lto_symtab decls.
11725 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11726 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11727 * varpool.c (dump_varpool_node): Dump more flags.
11729 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11731 * config/i386/i386.c (processor_alias_table): Add instruction
11732 FSGSBASE for AMD bdver3 architecture.
11734 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11736 * tree.c (warn_deprecated_use): Print file:line using locus color.
11737 * diagnostic.c (diagnostic_report_current_module): Print file:line
11738 and file:line:column using locus color.
11740 2013-05-14 Mike Stump <mikestump@comcast.net>
11742 * gdbinit.in: Add __null.
11744 2013-05-14 Mike Stump <mikestump@comcast.net>
11746 * recog.h: Rename struct recog_data to Recog_data.
11747 * recog.c: Likewise.
11748 * reload.c (can_reload_into): Likewise.
11749 * config/picochip/picochip.c: Likewise.
11751 2013-05-14 Mike Stump <mikestump@comcast.net>
11753 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11755 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11757 * resource.h (struct resources): Remove unch_memory member.
11758 (CLEAR_RESOURCE): Don't clear unch_memory.
11759 * resource.c (mark_referenced_resources): Don't set it.
11760 (mark_set_resources): Likewise.
11761 (mark_target_live_regs): Don't clear it.
11762 (init_resource_info): Likewise.
11763 * reorg.c (resource_conflicts_p): Don't compare it.
11764 (redundant_insn): Don't set it.
11766 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11768 * emit-rtl.c (next_label): Remove unused function.
11769 (skip_consecutive_labels, link_cc0_insns): Move to ...
11770 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11771 only place where these functions are used, and make them static.
11773 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11775 * fold-const.c (fold_negate_expr): Handle vectors.
11776 (fold_truth_not_expr): Make it static.
11777 (fold_invert_truthvalue): New static function.
11778 (invert_truthvalue_loc): Handle vectors. Do not call
11779 fold_truth_not_expr directly.
11780 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11781 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11782 (fold_comparison): Handle vector constants.
11783 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11784 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11785 * tree.h (fold_truth_not_expr): Remove declaration.
11787 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11789 * config/aarch64/aarch64-simd.md
11790 (aarch64_vcond_internal<mode>): Rename to...
11791 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11792 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11793 float modes. Clarify all iterator modes.
11794 (vcond<mode><mode>): Use new name for vcond expanders.
11795 (vcond<v_cmp_result><mode>): Likewise.
11796 (vcondu<mode><mode>: Likewise.
11797 * config/aarch64/iterators.md (VDQF_COND): New.
11799 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11802 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11803 variable for the shift amount. Check that we shift by non-negative
11806 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11809 * config/arm/arm.h (EPILOGUE_USES): Only return true
11810 for LR_REGNUM after epilogue_completed.
11812 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
11814 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
11815 is error_mark_node.
11817 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11820 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
11821 and Solaris 11+/x86 with gld.
11822 * configure: Regenerate.
11824 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11826 * expmed.c (expand_shift_1): Canonicalize rotates by
11827 constant bitsize / 2 to bitsize - 1.
11828 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
11829 case ROTATERT>: Likewise.
11832 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11834 * config/i386/i386.md (rotateinv): New code attr.
11835 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11836 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11837 roll $31, %eax, etc.
11839 2013-05-14 Richard Biener <rguenther@suse.de>
11841 PR middle-end/57235
11842 * tree-eh.c (sink_clobbers): Give up for successors with
11843 multiple predecessors and no virtual uses.
11845 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
11847 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
11848 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
11850 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11852 PR middle-end/57251
11853 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
11854 the case when both op0 and op1 have VOIDmode.
11856 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11858 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
11859 in multiply-accumulate mode.
11861 2013-05-13 Guozhi Wei <carrot@google.com>
11863 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
11865 2013-05-13 Kai Tietz <ktietz@redhat.com>
11868 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
11869 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
11870 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
11871 * config/i386/i386.c (ix86_option_override_internal): Likewise.
11872 (ix86_expand_prologue): Likewise.
11873 (ix86_expand_split_stack_prologue): Likewise.
11874 (legitimate_pic_address_disp_p): Likewise.
11875 (legitimize_pic_address): Likewise.
11876 (legitimize_tls_address): Likewise.
11877 (legitimize_pe_coff_symbol): Likewise.
11878 (output_pic_addr_const): Likewise.
11879 (construct_plt_address): Likewise.
11880 (ix86_expand_call): Likewise.
11881 (x86_output_mi_thunk): Likewise.
11882 (x86_function_profiler): Likewise.
11884 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
11886 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
11887 similar switch cases.
11888 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
11889 (aarch64_simd_mov_to_<mode>low): Delete.
11890 (aarch64_simd_mov_to_<mode>high): Delete.
11891 (move_lo_quad_<mode>): Add w<-r alternative.
11892 (aarch64_simd_move_hi_quad_<mode>): Likewise.
11893 (aarch64_simd_mov_from_*): Update type attribute.
11894 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
11897 2013-05-13 Jan Hubicka <jh@suse.cz>
11899 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
11900 * config/i386/i386.c (ix86_compute_frame_layout,
11901 ix86_expand_epilogue, emit_i387_cw_initialization,
11902 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
11903 ix86_local_alignment): Fix use of size/speed predicates.
11905 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11907 PR tree-optimization/45216
11908 PR tree-optimization/57157
11909 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
11910 the (-Y) & (B - 1) variant if OP is |.
11911 * expmed.c (expand_shift_1): For rotations by const0_rtx just
11912 return shifted. Use (-op1) & (prec - 1) as other_amount
11913 instead of prec - op1.
11915 2013-05-13 Martin Jambor <mjambor@suse.cz>
11917 PR middle-end/42371
11918 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
11919 (ipa_constant_data): New type.
11920 (ipa_jump_func): Use ipa_constant_data to hold information about
11921 constant jump functions.
11922 (ipa_get_jf_constant): Adjust to jump function type changes.
11923 (ipa_get_jf_constant_rdesc): New function.
11924 (ipa_param_descriptor): New field controlled_uses.
11925 (ipa_get_controlled_uses): New function.
11926 (ipa_set_controlled_uses): Likewise.
11927 * ipa-ref.h (ipa_find_reference): Declare.
11928 * ipa-prop.c (ipa_cst_ref_desc): New type.
11929 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
11931 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
11932 New parameter cs. Adjust all callers.
11933 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
11934 (remove_described_reference): New function.
11935 (jfunc_rdesc_usable): Likewise.
11936 (try_make_edge_direct_simple_call): Decrement controlled use count,
11937 attempt to remove reference if it hits zero.
11938 (combine_controlled_uses_counters): New function.
11939 (propagate_controlled_uses): Likewise.
11940 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
11941 (ipa_edge_duplication_hook): Duplicate reference descriptions.
11942 (ipa_print_node_params): Print described use counter.
11943 (ipa_write_jump_function): Adjust to jump function type changes.
11944 (ipa_read_jump_function): New parameter CS, pass it to
11945 ipa_set_jf_constant. Adjust caller.
11946 (ipa_write_node_info): Stream controlled use count
11947 (ipa_read_node_info): Likewise.
11948 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
11950 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
11951 count. Remove cloning-added reference if it reaches zero.
11952 * ipa-ref.c (ipa_find_reference): New function.
11954 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11956 * config/i386/i386.c (processor_target_table): Modified default
11957 alignment values for AMD BD and BT architectures.
11959 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
11961 * tree-vect-generic.c (uniform_vector_p): Move ...
11962 * tree.c (uniform_vector_p): ... here.
11963 * tree.h (uniform_vector_p): Declare it.
11964 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
11967 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11969 PR tree-optimization/57230
11970 * tree-ssa-strlen.c (handle_char_store): Record length for
11971 array store from STRING_CST.
11973 PR tree-optimization/57230
11974 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
11977 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
11979 * config/epiphany/epiphany.c (epiphany_init): Check size of
11980 NUM_MODES_FOR_MODE_SWITCHING.
11981 (epiphany_expand_prologue):
11982 Remove CONFIG_REGNUM initial value handling code.
11983 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
11984 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
11985 (emit_set_fp_mode, epiphany_mode_after): Likewise.
11986 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
11987 Don't return 1 for FP_MODE_NONE.
11988 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
11989 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
11990 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
11991 * config/epiphany/epiphany.md (save_config): New pattern.
11993 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
11995 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
11997 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
11999 * config/i386/i386.md (memory): Handle sseishft1.
12000 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
12001 (*vec_extractv2di_1): Ditto.
12003 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
12005 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
12008 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
12010 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
12011 Add mthumb/march=armv7-a multilib.
12012 Add mthumb/march=armv7-r multilib.
12013 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
12015 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
12017 * config/v850/t-rtems: Add more multilibs.
12019 2013-05-10 Richard Biener <rguenther@suse.de>
12021 PR tree-optimization/57214
12022 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
12023 not propagate from SSA names that occur in abnormal PHI nodes.
12025 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
12027 * stor-layout.c (element_precision): New function.
12028 * machmode.h (element_precision): Declare it.
12029 * tree.c (build_minus_one_cst): New function.
12030 (element_precision): Likewise.
12031 * tree.h (build_minus_one_cst): Declare new function.
12032 (element_precision): Likewise.
12033 * fold-const.c (operand_equal_p): Use element_precision.
12034 (fold_binary_loc): Handle vector types.
12035 * convert.c (convert_to_integer): Use element_precision.
12036 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
12039 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
12041 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
12042 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
12043 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
12044 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12045 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
12046 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
12047 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
12048 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12049 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
12050 (Uuw8): New constraints.
12051 (Usb4): Move into alphabetical order.
12052 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
12053 (sd8_operand, ub8_operand, uw8_operand): New predicates.
12054 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
12055 previously unnamed patterns.
12056 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
12057 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
12058 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
12059 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
12060 of set_attr_alternative/if_then_else. Use extended_mips16 instead
12061 of specific lengths.
12063 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12065 * config/i386/i386.md (rotateinv): New code attr.
12066 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12067 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12068 roll $31, %eax, etc.
12070 PR tree-optimization/45216
12071 PR tree-optimization/57157
12072 * tree-ssa-forwprop.c (simplify_rotate): New function.
12073 (ssa_forward_propagate_and_combine): Call it.
12075 2013-05-10 Richard Biener <rguenther@suse.de>
12077 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
12078 disable peeling when we version for aliasing.
12079 (vector_alignment_reachable_p): Honor explicit user alignment.
12080 (vect_supportable_dr_alignment): Likewise.
12081 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
12082 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
12083 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
12084 then peeling to arrange for the cost-model check to come first.
12086 2013-05-10 Alan Modra <amodra@gmail.com>
12088 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
12089 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
12090 * configure: Regenerate.
12092 2013-05-10 Alan Modra <amodra@gmail.com>
12095 * varasm.c (default_elf_select_section): Move !DECL_P check..
12096 (get_named_section): ..to here before calling get_section_name.
12098 (default_section_type_flags): Add DECL_P check.
12099 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
12100 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
12102 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
12104 * config/epiphany/epiphany.c (epiphany_expand_prologue):
12105 When using gen_stack_adjust_str with a register offset, add a
12106 REG_FRAME_RELATED_EXPR note.
12108 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
12110 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
12111 (*vec_extractv4si_zext_mem): Ditto.
12112 (*vec_extractv2di): Add 0->x and x->x alternatives.
12113 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
12114 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
12116 2013-05-09 Jason Merrill <jason@redhat.com>
12118 N3639 C++1y VLA support
12119 * gimplify.c (gimplify_vla_decl): Don't touch an existing
12122 * tree.c (build_constructor_va): New.
12123 * tree.h: Declare it.
12125 2013-05-09 Martin Jambor <mjambor@suse.cz>
12128 * gimple-fold.c (canonicalize_constructor_val): Call
12129 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
12131 2013-05-09 Jan Hubicka <jh@suse.cz>
12132 Richard Biener <rguenther@suse.de>
12135 * symtab.c (symtab_make_decl_local): Do not add private names.
12137 2013-05-09 Jan Hubicka <jh@suse.cz>
12140 * symtab.c (insert_to_assembler_name_hash): Handle clones.
12141 (unlink_from_assembler_name_hash): Likewise.
12142 (symtab_prevail_in_asm_name_hash, symtab_register_node,
12143 symtab_unregister_node, symtab_initialize_asm_name_hash,
12144 change_decl_assembler_name): Update.
12146 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
12148 * config/aarch64/aarch64.md: New movtf split.
12149 (*movtf_aarch64): Update.
12150 (aarch64_movdi_tilow): Handle TF modes and rename to
12151 aarch64_movdi_<mode>low.
12152 (aarch64_movdi_tihigh): Handle TF modes and rename to
12153 aarch64_movdi_<mode>high
12154 (aarch64_movtihigh_di): Handle TF modes and rename to
12155 aarch64_mov<mode>high_di
12156 (aarch64_movtilow_di): Handle TF modes and rename to
12157 aarch64_mov<mode>low_di
12158 (aarch64_movtilow_tilow): Remove spurious whitespace.
12159 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
12161 (aarch64_print_operand): Update.
12163 2013-05-09 Alan Modra <amodra@gmail.com>
12165 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
12167 * configure: Regenerate.
12169 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12171 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
12172 splitter preparation statements.
12173 * config/i386/sse.md (*vec_extract* splitters): Ditto.
12174 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
12177 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12179 * gimple-ssa-strength-reduction.c (count_candidates): Change
12180 return value to int.
12181 (analyze_candidates_and_replace): Change type of length to int.
12183 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12185 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
12186 (*vec_extract<mode>): Use VI12_128 mode iterator.
12187 (*vec_extract<mode>_mem): Ditto.
12188 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
12191 2013-05-08 Diego Novillo <dnovillo@google.com>
12196 2012-08-17 Diego Novillo <dnovillo@google.com>
12199 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
12200 * config.in: Regenerate.
12201 * configure: Regenerate.
12202 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
12204 2013-05-08 Jan Hubicka <jh@suse.cz>
12207 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
12208 * cgraph.h (symtab_node_base): Add unique_name.
12209 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12210 input_overwrite_node, input_varpool_node): Stream unique_name.
12211 * cgraphclones.c (cgraph_create_virtual_clone,
12212 cgraph_function_versioning): Set unique_name.
12213 * ipa.c (function_and_variable_visibility): Set unique_name.
12215 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12217 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
12218 (alloc_cand_and_find_basis): Restrict conditional candidate
12219 processing to CAND_MULTs.
12221 2013-05-08 Jan Hubicka <jh@suse.cz>
12224 lto-symtab.c (lto_symtab_symbol_p): New function.
12225 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
12226 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
12227 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
12228 Skip static symbols.
12230 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
12232 PR tree-optimization/57200
12233 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
12234 Only call inform if the preceding warning_at returns true.
12236 2013-05-07 Han Shen <shenhan@google.com>
12238 * cfgexpand.c (record_or_union_type_has_array_p): New function.
12239 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
12240 * common.opt (fstack-protector-strong): New option.
12241 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
12242 * doc/invoke.texi (Optimization Options): Document
12243 "-fstack-protector-strong".
12244 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
12246 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12248 * config/mips/mips.c (mips_machine_reorg2): Return 0.
12250 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
12252 * ira.c (update_equiv_regs): Add insn having equiv memory even if
12253 it is not lhs of the insn.
12254 (setup_reg_equiv): Remove insn having equiv memory which it is not
12256 * lra-constraints.c (process_address): Try to improve generation
12257 code for address base + disp.
12258 (lra_constraints): Make correct the code for checking insn setting
12259 up backward equivalence. Remove insn only if it is in the init
12261 * lra-eliminations.c (update_reg_eliminate): Change return value.
12262 (lra_eliminate): Use the result.
12264 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
12266 * config/i386/sse.md (ssescalarnummask): New mode attribute.
12267 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
12268 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
12269 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
12270 register target operands.
12271 (*vec_extractv8hi_sse2): New pattern.
12272 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
12273 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
12274 (*vec_extract<mode>_mem): New insn and split pattern.
12276 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
12278 * config/arm/arm.c (arm_asan_shadow_offset): New function.
12279 (TARGET_ASAN_SHADOW_OFFSET): Define.
12280 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
12281 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
12283 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12285 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
12286 (incr_vec_index): Return -1 if increment not found.
12287 (create_add_on_incoming_edge): Assert if increment not found.
12288 (record_increment): Limit number of increments recorded.
12289 (all_phi_incrs_profitable): Return false if an increment not found.
12290 (replace_profitable_candidates): Don't process increments that were
12292 (analyze_candidates_and_replace): Limit size of incr_vec.
12294 2013-05-07 Richard Biener <rguenther@suse.de>
12296 * calls.c (special_function_p): setjmp-like functions are leaf.
12297 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
12298 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
12300 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
12302 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
12303 (aarch64_simd_mov<mode>): New expander.
12304 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
12305 (aarch64_simd_mov_to_<mode>high): Likewise.
12306 (aarch64_simd_mov_from_<mode>low): Likewise.
12307 (aarch64_simd_mov_from_<mode>high): Likewise.
12308 (aarch64_dup_lane<mode>): Update.
12309 (aarch64_dup_lanedi): New instruction pattern.
12310 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
12311 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
12313 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12315 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
12316 (replace_mult_candidate): Remove unnecessary argument; remove
12317 unnecessary parameter from call to introduce_cast_before_cand.
12318 (replace_unconditional_candidate): Remove unnecessary parameter
12319 from call to replace_mult_candidate.
12320 (replace_conditional_candidate): Likewise.
12321 (insert_initializers): Use make_temp_ssa_name.
12322 (introduce_cast_before_cand): Remove unnecessary argument; use
12323 make_temp_ssa_name.
12324 (replace_one_candidate): Remove unnecessary argument; remove
12325 unnecessary parameter from calls to introduce_cast_before_cand.
12326 (replace_profitable_candidates): Remove unnecessary parameters
12327 from calls to replace_one_candidate.
12329 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12331 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
12332 phi def as possibly hiding a basis for a CAND_ADD whose operands
12333 have been commuted in the analysis.
12334 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
12336 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12338 * config/aarch64/aarch64.md
12339 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
12340 shift value between 0-4.
12342 2013-05-07 Richard Biener <rguenther@suse.de>
12344 * double-int.h (rshift): New overload.
12345 * double-int.c (rshift): New function.
12346 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
12347 (create_reference_ops_from_ref): Remove.
12348 (vn_reference_insert): Use shared ops for constructing the
12349 reference and copy it.
12351 2013-05-07 Richard Biener <rguenther@suse.de>
12353 PR middle-end/57190
12354 * tree-eh.c (sink_clobbers): Properly propagate
12355 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
12357 2013-05-07 Jakub Jelinek <jakub@redhat.com>
12359 PR tree-optimization/57149
12360 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
12361 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
12362 collect_phi_def_edges, execute_late_warn_uninitialized): Use
12363 uninit_undefined_value_p instead of ssa_undefined_value_p.
12366 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
12367 for modifier == EXPAND_INITIALIZER.
12369 2013-05-07 Anton Blanchard <anton@samba.org>
12371 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
12372 for powerpc64 little endian.
12373 * configure: Regenerate.
12375 2013-05-06 Graham Stott <grahams@btinternet.com>
12377 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
12378 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
12379 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
12380 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
12382 2013-05-06 Graham Stott <grahams@btinternet.com>
12384 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
12385 codes which allow non-lvalues.
12387 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
12389 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
12390 components are all 1s.
12391 (integer_minus_onep): New function.
12392 * tree.h (integer_minus_onep): Declare it.
12393 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
12394 integer_minus_onep instead of integer_all_onesp.
12396 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12399 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
12400 variations of these patterns.
12402 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12404 * config/i386/i386.md (isa): Add x64_sse4 member.
12405 (enabled): Handle x64_sse4.
12406 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
12407 instruction for 64bit SSE4_1 targets. Update insn attributes.
12408 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
12409 instruction for SSE4_1 targets. Update insn attributes.
12410 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12411 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
12412 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
12414 (*vec_extractv4si): Rename from *sse4_1_pextrd.
12415 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
12416 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
12418 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12421 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
12423 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
12425 * final.c (do_assembler_dialects): Don't handle curly braces and
12426 vertical bar escaped by % as dialect delimiters.
12427 (output_asm_insn): Print curly braces and vertical bar if escaped
12428 by % and ASSEMBLER_DIALECT defined.
12429 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
12430 * doc/tm.texi: Regenerated.
12432 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12434 * config/mips/mips.c: Include tree-pass.h.
12435 (mips_reorg): Split in pre- and post-dbr_schedule parts.
12436 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
12437 (pass_mips_machine_reorg2): New machine specific pass.
12438 (insert_pass_mips_machine_reorg2): New pass plugin definition.
12439 (mips_option_override): Register the new pass.
12440 * rtl.h (cleanup_barriers): Remove prototype.
12441 (dbr_schedule): Likewise.
12442 * jump.c (cleanup_barriers): Make static.
12443 * reorg.c (dbr_schedule): Likewise.
12445 2013-05-06 Richard Biener <rguenther@suse.de>
12447 PR tree-optimization/57185
12448 * tree-parloops.c (add_field_for_reduction): Handle anonymous
12449 SSA names properly.
12451 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12454 * config/i386/i386.c (add_parameter_dependencies): Add dependence
12455 between "first_arg" and "insn", not "last" and "insn".
12457 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12459 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
12460 (find_candidates_in_block): Re-enable slsr_process_phi.
12461 (create_phi_basis): Fix double counting of candidate adjustment.
12463 2013-05-06 Richard Biener <rguenther@suse.de>
12465 PR middle-end/57147
12466 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12467 the edge is also fallthru, preserve it and just clear the
12469 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12470 also complex, preserve that and just clear the fallthru flag.
12471 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12472 update virtual operands.
12474 2013-05-06 Alan Modra <amodra@gmail.com>
12476 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12477 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12478 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12479 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12480 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12481 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12483 2013-05-06 Alan Modra <amodra@gmail.com>
12485 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12486 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12487 (DEFAULT_ASM_ENDIAN): Define.
12488 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12489 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12490 Update -K PIC clause from sysv4.h.
12491 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12492 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12494 2013-05-06 Alan Modra <amodra@gmail.com>
12496 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12497 twice for little-endian.
12498 (ashrdi3_no_power, ashrdi3): Support little-endian.
12500 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12503 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12504 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12506 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12507 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12508 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12511 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12512 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12514 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12515 * config/pa/pa.opt: Make mbig-switch a no-op.
12516 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12517 (CASE_VECTOR_MODE): Always return SImode.
12518 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12519 for the !TARGET_BIG_SWITCH case.
12520 * config/pa/pa-linux.h: Likewise.
12521 * config/pa/pa-openbsd.h: Likewise.
12522 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12523 * config/pa/pa.md (short_jump): Remove define_insn.
12524 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12525 (casesi0): Remove define_insn.
12526 (type): Remove btable_branch.
12527 (pa_combine_type): Likewise.
12528 (in_nullified_branch_delay): Likewise.
12529 (in_call_delay): Likewise.
12530 (define_delay): Likewise.
12531 (define_insn_reservation "Z3"): Likewise.
12532 (define_insn_reservation "Z4"): Likewise.
12533 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12534 (pa_adjust_insn_length): Remove adjustment for btable branches.
12535 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12538 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
12540 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12541 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12542 Add m->r,x alternatives.
12543 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12544 splitters using SWI48x mode iterator.
12545 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12546 TARGET_64BIT. Add m->x alternative.
12547 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12548 Add o->x alternative. Enable for TARGET_SSE.
12549 (sse_storeq): Remove expander.
12550 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12551 with memory input operand.
12552 (*vec_extractv2di_1 splitter): New.
12553 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12554 * config/i386/i386.md (ssevecmodelower): New mode attribute.
12556 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12558 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
12559 (INT_LOWPART): Delete.
12560 (extract_MB): Adjust.
12561 (extract_ME): Adjust.
12562 (print_operand): Adjust.
12564 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12566 * config/rs6000/predicates.md (reg_or_add_cint_operand,
12567 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12568 (reg_or_logical_cint_operand, easy_fp_constant,
12569 logical_const_operand): Delete "CONST_DOUBLE" case.
12570 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
12571 "HOST_BITS_PER_WIDE_INT == 64" test.
12572 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
12573 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
12574 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
12575 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
12577 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
12578 CONST_DOUBLE DImode/VOIDmode case.
12579 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
12580 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12581 CONST_DOUBLE VOIDmode case.
12582 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12583 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
12584 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
12585 Delete CONST_DOUBLE case.
12586 (splitters for mov FMOVE64 const_double): Delete
12587 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12588 "HOST_BITS_PER_WIDE_INT >= 64" test.
12589 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
12591 (mov DI const_double): Delete.
12593 2013-05-04 Jakub Jelinek <jakub@redhat.com>
12595 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
12596 on op shows all bits zero in mode of a lowpart subreg, return zero.
12598 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
12601 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
12602 to save TFmode registers and DImode to save TImode registers for
12603 caller save operations.
12604 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
12605 mark being partially clobbered since they only use the first
12608 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
12609 and TDmode only use the upper 64-bits of each VSX register.
12611 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12613 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
12614 (find_candidates_in_block): Disable slsr_process_phi.
12616 2013-05-03 Guozhi Wei <carrot@google.com>
12618 * coverage.c (coverage_obj_init): Move the construction of gcov
12620 (build_init_ctor): ... here.
12622 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12624 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12625 (slsr_cand_d): Redefine def_phi.
12626 (stride_status, phi_adjust_status, count_phis_status): New enums.
12627 (find_phi_def): New.
12628 (find_basis_for_base_expr): New.
12629 (find_basis_for_candidate): Handle hidden bases.
12630 (alloc_cand_and_find_basis): Handle phi candidates.
12631 (slsr_process_phi): New.
12632 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12633 (create_mul_imm_cand): Likewise.
12634 (create_add_ssa_cand): Exclude phi base candidates.
12635 (create_add_imm_cand): Likewise.
12636 (slsr_process_cast): Likewise.
12637 (slsr_process_copy): Likewise.
12638 (find_candidates_in_block): Handle phi candidates.
12639 (dump_candidate): Likewise.
12640 (unconditional_cands): Delete.
12641 (unconditional_cands_with_known_stride_p): Delete.
12642 (phi_dependent_cand_p): New.
12643 (cand_increment): Handle phi-dependent candidates.
12644 (replace_dependent): Delete.
12645 (replace_mult_candidate): New.
12646 (replace_unconditional_candidate): New.
12647 (incr_vec_index): Move to avoid forward reference.
12648 (create_add_on_incoming_edge): New.
12649 (create_phi_basis): New.
12650 (replace_dependents): Delete.
12651 (replace_conditional_candidate): New.
12652 (phi_add_costs): New.
12653 (replace_uncond_cands_and_profitable_phis): New.
12654 (record_increment): Handle phi adjustments.
12655 (record_phi_increments): New.
12656 (record_increments): Handle phi adjustments.
12657 (phi_incr_cost): New.
12658 (lowest_cost_path): Handle phis.
12659 (total_savings): Likewise.
12660 (analyze_increments): Likewise.
12661 (ncd_with_phi): New.
12662 (ncd_of_cand_and_phis): New.
12663 (nearest_common_dominator_for_cands): Handle phi increments.
12664 (all_phi_incrs_profitable): New.
12665 (replace_profitable_candidates): Handle phi-dependent candidates.
12666 (analyze_candidates_and_replace): Likewise.
12668 2013-05-03 Teresa Johnson <tejohnson@google.com>
12671 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12672 do not exceed REG_BR_PROB_BASE.
12674 2013-05-03 Jeff Law <law@redhat.com>
12676 PR tree-optimization/57144
12677 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12678 operand of the condition will bit into the new type when eliminating
12679 a cast feeding a condition.
12681 2013-05-03 Jakub Jelinek <jakub@redhat.com>
12683 PR rtl-optimization/57130
12684 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12685 of COMPARE as in_code to the recursive call if needed.
12687 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
12689 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12690 (enabled): Handle new members.
12691 * config/i386/sse.md (*vec_concatv2si): Merge from
12692 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12693 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12695 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
12697 PR tree-optimization/57027
12698 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12699 for fnms opportunity, check we got the prerequisite kind
12700 of tree / gimple before using accessor functions.
12702 2013-05-03 Richard Biener <rguenther@suse.de>
12704 * double-int.h (lshift): New overload without precision
12705 and arith argument.
12706 (operator *=, operator +=, operator -=): Move ...
12707 * double-int.c (operator *=, operator +=, operator -=): ... here
12708 and implement more efficiently.
12709 (mul_double_with_sign): Remove.
12710 (lshift_double): Adjust to take unsinged shift argument, push
12711 dispatching code to callers.
12712 (mul_double_wide_with_sign): Add early out for callers that
12713 are not interested in high parts or overflow.
12714 (lshift): New function.
12715 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12716 dispatch code here.
12717 (lrotate, rrotate): Use logical shifts.
12718 * expr.c (get_inner_reference): Use lshift.
12719 * fixed-value.c (do_fixed_divide): Likewise.
12720 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12721 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12722 (indirect_refs_may_alias_p): Likewise.
12723 (stmt_kills_ref_p_1): Likewise.
12725 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12727 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12729 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12731 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12732 scalar form of FABD instruction.
12734 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12736 * lra-constraints.c (process_alt_operands): Add checking alt
12737 number to choose the best alternative.
12739 2013-05-02 Richard Biener <rguenther@suse.de>
12741 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12742 bitmap and its handling.
12743 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12745 2013-05-02 Richard Biener <rguenther@suse.de>
12747 PR middle-end/57140
12748 * tree-inline.c (copy_loops): Properly handle removed loops.
12749 (copy_cfg_body): Mark destination loops for fixup if source
12750 loops needed fixup.
12752 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12755 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12756 generating simple_return for naked functions.
12758 2013-05-02 Martin Jambor <mjambor@suse.cz>
12760 PR middle-end/56988
12761 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12762 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12764 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12765 ipa_agg_replacement_value structures.
12766 (known_aggs_to_agg_replacement_list): Likewise.
12767 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12768 (read_agg_replacement_chain): Likewise.
12769 (ipcp_transform_function): Also check that by_ref flags match.
12771 2013-05-02 Richard Biener <rguenther@suse.de>
12773 * graphds.h (struct graph): Add obstack member.
12774 * graphds.c (new_graph): Initialize obstack and allocate
12776 (add_edge): Allocate edge from the obstack.
12777 (free_graph): Free the obstack instead of all edges and vertices.
12779 2013-05-02 Teresa Johnson <tejohnson@google.com>
12781 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12783 * cfg.c (update_bb_profile_for_threading): Ditto.
12784 * tree-inline.c (copy_bb): Ditto.
12785 (copy_edges_for_bb): Ditto.
12786 (initialize_cfun): Ditto.
12787 (copy_cfg_body): Ditto.
12788 (expand_call_inline): Ditto.
12789 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12790 (estimate_node_size_and_time): Ditto.
12791 (inline_merge_summary): Ditto.
12792 * cgraphclones.c (cgraph_clone_edge): Ditto.
12793 (cgraph_clone_node): Ditto.
12794 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12795 (compute_trg_info): Ditto.
12797 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12799 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12800 S reg when fp attribute set.
12801 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12803 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12805 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12807 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12808 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12809 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12811 2013-05-02 Richard Biener <rguenther@suse.de>
12813 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12814 (struct instantiate_cache_entry): New type.
12815 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12816 (struct instantiate_cache_type): New type.
12817 (set_instantiated_value, get_instantiated_value): Remove.
12818 (get_instantiated_value_entry): New function.
12819 (instantiate_scev_name): Use the new cache and adjust.
12820 (instantiate_scev_poly): Adjust.
12821 (instantiate_scev_binary): Likewise.
12822 (instantiate_array_ref): Likewise.
12823 (instantiate_scev_convert): Likewise.
12824 (instantiate_scev_not): Likewise.
12825 (instantiate_scev_3): Likewise.
12826 (instantiate_scev_2): Likewise.
12827 (instantiate_scev_r): Likewise.
12828 (instantiate_scev): Likewise.
12829 (resolve_mixers): Likewise.
12831 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
12834 * lra-constraints.c (best_small_class_operands_num): Remove.
12835 (process_alt_operands): Remove small_class_operands_num. Take
12836 small classes operands into losers and only if the operand is not
12837 matched. Modify debugging output.
12838 (curr_insn_transform): Remove best_small_class_operands_num.
12841 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12843 * config/aarch64/aarch64-builtins.c
12844 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
12845 * config/aarch64/aarch64-simd-builtins.def
12846 (reduc_splus_): Add new modes.
12847 (reduc_uplus_): New.
12848 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
12849 (reduc_uplus_v4sf): Likewise.
12850 (reduc_splus_v4sf): Likewise.
12851 (aarch64_addv<mode>): Likewise.
12852 (reduc_uplus_<mode>): Likewise.
12853 (reduc_splus_<mode>): Likewise.
12854 (aarch64_addvv2di): Likewise.
12855 (reduc_uplus_v2di): Likewise.
12856 (reduc_splus_v2di): Likewise.
12857 (aarch64_addvv2si): Likewise.
12858 (reduc_uplus_v2si): Likewise.
12859 (reduc_splus_v2si): Likewise.
12860 (reduc_<sur>plus_<mode>): New.
12861 (reduc_<sur>plus_v2di): Likewise.
12862 (reduc_<sur>plus_v2si): Likewise.
12863 (reduc_<sur>plus_v4sf): Likewise.
12864 (aarch64_addpv4sf): Likewise.
12865 * config/aarch64/arm_neon.h
12866 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
12867 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
12868 add UNSPEC_SADDV, UNSPEC_UADDV.
12870 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
12872 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12874 * config/aarch64/arm_neon.h
12875 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
12877 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12879 * config/aarch64/aarch64-builtins
12880 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
12882 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12884 * config/aarch64/aarch64-simd-builtins.def
12885 (reduc_smax_): New.
12886 (reduc_smin_): Likewise.
12887 (reduc_umax_): Likewise.
12888 (reduc_umin_): Likewise.
12889 (reduc_smax_nan_): Likewise.
12890 (reduc_smin_nan_): Likewise.
12893 (smax): Update for V2SF, V4SF and V2DF modes.
12896 (smin_nan): Likewise.
12897 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
12898 (<su><maxmin><mode>3): ...This, refactor.
12899 (s<maxmin><mode>3): New.
12900 (<maxmin_uns><mode>3): Likewise.
12901 (reduc_<maxmin_uns>_<mode>): Refactor.
12902 (reduc_<maxmin_uns>_v4sf): Likewise.
12903 (reduc_<maxmin_uns>_v2si): Likewise.
12904 (aarch64_<fmaxmin><mode>: Remove.
12905 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
12907 (vmin<q>_f<32,64>): Likewise.
12908 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12910 (su): Add mappings for smax, smin, umax, umin.
12912 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12913 (FMAXMIN): Rename as...
12914 (FMAXMIN_UNS): ...This.
12916 (fmaxminv): Likewise.
12917 (fmaxmin): Likewise.
12919 (maxmin_uns_op): Likewise.
12921 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12923 * config/aarch64/arm_neon.h
12924 (vac<ge, gt><sd>_f<32, 64>): Rename to...
12925 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
12926 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
12928 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12930 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
12931 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
12933 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12935 * config/aarch64/aarch64-simd.md
12936 (vcond<mode>_internal): Handle special cases for constant masks.
12937 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
12938 (vcondu<mode><mode>): Likewise.
12939 (vcond<v_cmp_result><mode>): New.
12941 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12943 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
12944 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
12945 * config/aarch64/aarch64-simd-builtins.def
12946 (cmeq): Update to BUILTIN_VALLDI.
12951 * config/aarch64/arm_neon.h
12952 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
12953 to builtins or C as appropriate.
12955 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12957 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
12959 (cmhi): Rename to...
12961 * config/aarch64/aarch64-simd.md
12962 (simd_mode): Add SF.
12963 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
12964 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
12965 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
12966 (cstore<mode>_neg): ...This.
12967 * config/aarch64/iterators.md
12969 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
12970 (COMPARISONS): New.
12971 (UCOMPARISONS): Likewise.
12972 (optab): Add missing comparisons.
12978 (VCMP_S): Likewise.
12979 (VCMP_U): Likewise.
12980 (V_cmp_result): Add DF, SF modes.
12981 (v_cmp_result): Likewise.
12983 (vmtype): Likewise.
12984 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
12986 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
12988 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
12989 define_insn to define_insn_and_split.
12990 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
12991 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
12992 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
12993 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
12994 (thumb2_negscc): Likewise.
12996 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12998 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
13000 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13002 * config/arm/thumb2.md: Remove trailing whitespaces.
13004 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13006 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
13007 Use gen_int_mode rather than GEN_INT.
13009 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
13011 * value-prof.c (stream_in_histogram_value): Remove the strayed
13014 2013-04-30 Richard Biener <rguenther@suse.de>
13016 PR middle-end/57122
13017 * cfghooks.c (split_edge): Properly check for the loop latch edge.
13019 2013-04-30 Richard Biener <rguenther@suse.de>
13021 PR middle-end/57107
13022 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
13024 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
13026 PR rtl-optimization/56957
13027 PR rtl-optimization/57105
13028 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
13029 variable. Use just INSN_UID for determining whether an insn
13030 should be only disconnected from the insn stream.
13031 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
13033 2013-04-30 Jakub Jelinek <jakub@redhat.com>
13035 PR tree-optimization/57104
13036 * tsan.c (instrument_expr): Don't instrument accesses to
13037 DECL_HARD_REGISTER VAR_DECLs.
13039 2013-04-30 Richard Biener <rguenther@suse.de>
13041 * function.h (loops_for_fn): New inline function.
13042 (set_loops_for_fn): Likewise.
13043 * cfgloop.h (place_new_loop): Add struct function parameter.
13044 (get_loop): Likewise.
13045 (get_loops): Likewise.
13046 (number_of_loops): Likewise.
13047 (fel_next): Adjust.
13048 (fel_init): Likewise.
13049 * cfg.c (get_loop_copy): Adjust.
13050 * cfgloop.c (flow_loops_dump): Likewise.
13051 (record_loop_exits): Likewise.
13052 (verify_loop_structure): Likewise.
13053 * cfgloopanal.c (mark_irreducible_loops): Likewise.
13054 (estimate_reg_pressure_cost): Likewise.
13055 (mark_loop_exit_edges): Likewise.
13056 * cfgloopmanip.c (place_new_loop): Likewise.
13057 (add_loop): Likewise.
13058 (duplicate_loop): Likewise.
13059 * graph.c (draw_cfg_nodes): Likewise.
13060 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
13061 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
13062 (extract_affine_chrec): Likewise.
13063 (build_scop_iteration_domain): Likewise.
13064 * graphite.c (graphite_initialize): Likewise.
13065 * ira-build.c (create_loop_tree_nodes): Likewise.
13066 (more_one_region_p): Likewise.
13067 (rebuild_regno_allocno_maps): Likewise.
13068 (mark_loops_for_removal): Likewise.
13069 (mark_all_loops_for_removal): Likewise.
13070 (remove_unnecessary_regions): Likewise.
13071 (ira_build): Likewise.
13072 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
13073 * loop-init.c (fix_loop_structure): Likewise.
13074 (gate_rtl_move_loop_invariants): Likewise.
13075 (gate_rtl_unswitch): Likewise.
13076 (gate_rtl_unroll_and_peel_loops): Likewise.
13077 (rtl_doloop): Likewise.
13078 * lto-streamer-in.c (input_cfg): Likewise.
13079 * lto-streamer-out.c (output_cfg): Likewise.
13080 * modulo-sched.c (sms_schedule): Likewise.
13081 * predict.c (tree_estimate_probability): Likewise.
13082 (tree_estimate_probability_driver): Likewise.
13083 (estimate_loops): Likewise.
13084 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
13085 (move_sese_region_to_fn): Likewise.
13086 (debug_loop_num): Likewise.
13087 * tree-chrec.c (chrec_evaluate): Likewise.
13088 (hide_evolution_in_other_loops_than_loop): Likewise.
13089 (chrec_component_in_loop_num): Likewise.
13090 (reset_evolution_in_loop): Likewise.
13091 (evolution_function_is_invariant_rec_p): Likewise.
13092 * tree-if-conv.c (main_tree_if_conversion): Likewise.
13093 * tree-inline.c (copy_loops): Likewise.
13094 (copy_cfg_body): Likewise.
13095 (tree_function_versioning): Likewise.
13096 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
13097 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
13099 (add_to_evolution_1): Likewise.
13100 (scev_const_prop): Likewise.
13101 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
13102 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
13103 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
13104 (tree_ssa_lim_initialize): Likewise.
13105 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
13106 (verify_loop_closed_ssa): Likewise.
13107 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
13108 (tree_ssa_loop_im): Likewise.
13109 (tree_ssa_loop_unswitch): Likewise.
13110 (tree_vectorize): Likewise.
13111 (check_data_deps): Likewise.
13112 (tree_ssa_loop_ivcanon): Likewise.
13113 (tree_ssa_loop_bounds): Likewise.
13114 (tree_complete_unroll): Likewise.
13115 (tree_complete_unroll_inner): Likewise.
13116 (tree_parallelize_loops): Likewise.
13117 (tree_ssa_loop_prefetch): Likewise.
13118 (tree_ssa_loop_ivopts): Likewise.
13119 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
13120 * tree-vectorizer.c (vectorize_loops): Likewise.
13122 2013-04-29 Mike Frysinger <vapier@gentoo.org>
13124 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
13125 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
13126 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
13127 with EABI_LINK_SPEC.
13129 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13132 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
13135 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
13138 * lra-constraints.c (process_alt_operands): Discourage a bit more
13139 using memory for pseudos. Print cost dump for alternatives.
13140 Modify cost values for conflicts with early clobbers.
13141 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
13143 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13146 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
13148 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13150 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
13151 from/to S register.
13152 (movdi_aarch64): Support LDR/STR from/to D register.
13154 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13156 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
13157 or higher by default.
13159 2013-04-29 Richard Biener <rguenther@suse.de>
13161 PR middle-end/57075
13162 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
13163 even if not adding abnormal edges for calls that can make
13166 2013-04-29 Richard Biener <rguenther@suse.de>
13168 PR middle-end/57103
13169 * tree-cfg.c (move_stmt_op): Fix condition under which to update
13171 (move_stmt_r): Remove redundant checking.
13173 2013-04-29 Teresa Johnson <tejohnson@google.com>
13176 * basic-block.h (apply_scale): New function.
13177 (apply_probability): Use apply_scale.
13178 * gimple-streamer-in.c (input_bb): Ditto.
13179 * lto-streamer-in.c (input_cfg): Ditto.
13180 * lto-cgraph.c (merge_profile_summaries): Ditto.
13181 * tree-optimize.c (execute_fixup_cfg): Ditto.
13182 * tree-inline.c (copy_bb): Update comment to use apply_scale.
13183 (copy_edges_for_bb): Ditto.
13184 (copy_cfg_body): Ditto.
13186 2013-04-29 Tom de Vries <tom@codesourcery.com>
13188 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
13189 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
13190 (tail_merge_optimize): Handle current_loops == NULL.
13192 2013-04-26 Jeff Law <law@redhat.com>
13194 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
13195 (simplify_cond_using_ranges): Generalize code to simplify
13196 COND_EXPRs where one argument is a constant and the other
13197 is an SSA_NAME created by an integral type conversion.
13199 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13201 * config/arm/arm.md (store_minmaxsi): Use only when
13202 optimize_insn_for_size_p.
13204 2013-04-29 Christian Bruel <christian.bruel@st.com>
13207 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
13209 2013-04-29 Richard Biener <rguenther@suse.de>
13211 PR middle-end/57089
13212 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
13213 loop tree make sure to schedule a fixup for the child as well.
13214 (expand_omp_for_generic): Properly add loops.
13215 (expand_omp_for_static_nochunk): Likewise.
13216 (expand_omp_for_static_chunk): Likewise.
13217 (expand_omp_for): For the degenerate case fixup loops.
13218 (expand_omp_sections): Fix default bb placement in loops.
13219 (expand_omp_atomic_pipeline): Properly add loops.
13221 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13223 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
13225 2013-04-29 Tom de Vries <tom@codesourcery.com>
13227 * tree-ssa-tail-merge.c: Update header comment.
13229 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13231 * config/aarch64/arm_neon.h
13232 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
13233 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
13234 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
13235 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
13236 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
13237 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
13238 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
13239 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
13241 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13243 * config/aarch64/aarch64-simd.md
13244 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
13245 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
13246 fix_trunc, fixuns_trunc.
13247 (ftrunc<VDQF:mode>2): New.
13248 * config/aarch64/iterators.md (optab): Add fix, fixuns.
13249 (fix_trunc_optab): New.
13251 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13253 * config/aarch64/aarch64-builtins.c
13254 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
13255 iceilf, lround, iroundf.
13257 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13260 * config/i386/i386.h (enum ix86_tune_indices)
13261 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
13262 New, split from X86_TUNE_INTER_UNIT_MOVES.
13263 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
13264 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
13265 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
13266 (TARGET_INTER_UNIT_MOVES): Remove.
13267 * config/i386/i386.c (initial_ix86_tune_features): Update.
13268 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
13269 (ix86_expand_convert_uns_didf_sse): Use
13270 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13271 (ix86_expand_vector_init_one_nonzero): Ditto.
13272 (ix86_expand_vector_init_interleave): Ditto.
13273 (inline_secondary_memory_needed): Return true for moves from SSE class
13274 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
13275 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13276 * config/i386/constraints.md (Yi, Ym): Depend on
13277 TARGET_INTER_UNIT_MOVES_TO_VEC.
13278 (Yj, Yn): New constraints.
13279 * config/i386/i386.md (*movdi_internal): Change constraints of
13280 operand 1 from Yi to Yj and from Ym to Yn.
13281 (*movsi_internal): Ditto.
13282 (*movdf_internal): Ditto.
13283 (*movsf_internal): Ditto.
13284 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
13285 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13286 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
13287 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
13288 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
13289 * config/i386/sse.md (movdi_to_sse): Ditto.
13290 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
13291 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
13292 TARGET_INTER_UNIT_MOVES.
13293 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
13294 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
13295 instead of TARGET_INTER_UNIT_MOVES.
13296 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
13297 operand 1 from Yi to Yj and from Ym to Yn.
13299 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13301 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
13302 (float_truncate_hi_): Likewise.
13303 (float_extend_lo_): Likewise.
13304 (float_truncate_lo_): Likewise.
13305 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
13306 (aarch64_float_extend_lo_v2df): Likewise.
13307 (vec_unpacks_hi_v4sf): Likewise.
13308 (aarch64_float_truncate_lo_v2sf): Likewise.
13309 (aarch64_float_truncate_hi_v4sf): Likewise.
13310 (vec_pack_trunc_v2df): Likewise.
13311 (vec_pack_trunc_df): Likewise.
13313 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13315 * config/aarch64/aarch64-builtins.c
13316 (aarch64_fold_builtin): Fold float conversions.
13317 * config/aarch64/aarch64-simd-builtins.def
13318 (floatv2si, floatv4si, floatv2di): New.
13319 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
13320 * config/aarch64/aarch64-simd.md
13321 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
13322 * config/aarch64/iterators.md (FLOATUORS): New.
13323 (optab): Add float, floatuns.
13324 (su_optab): Likewise.
13326 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13328 * config/aarch64/aarch64-builtins.c
13329 (aarch64_builtin_vectorized_function): Use new names for
13331 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
13332 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
13333 (fcvtzu): Split as...
13334 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
13335 (fcvtas): Split as...
13336 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
13337 (fcvtau): Split as...
13338 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
13339 (fcvtps): Split as...
13340 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
13341 (fcvtpu): Split as...
13342 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
13343 (fcvtms): Split as...
13344 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
13345 (fcvtmu): Split as...
13346 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
13347 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
13348 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
13349 (lfrintnusf, lfrintnudf): Likewise.
13350 * config/aarch64/aarch64-simd.md
13351 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
13353 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
13354 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
13355 (fcvt_pattern): Likewise.
13357 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13359 * config/aarch64/aarch64-simd.md
13360 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
13361 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
13363 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13365 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
13366 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
13367 (vrnd<a,m,n,p>_f32): Implement using builtins.
13368 (vrnd<i,x><q>_f<32, 64>): New.
13370 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13372 * config/aarch64/aarch64-builtins.c
13373 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
13374 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
13375 (frintz): Rename to...
13377 (frintp): Rename to...
13379 (frintm): Rename to...
13381 (frinti): Rename to...
13382 (nearbyint): ...this.
13383 (frintx): Rename to...
13385 (frinta): Rename to...
13387 * config/aarch64/aarch64-simd.md
13388 (aarch64_frint<frint_suffix><mode>): Delete.
13389 (<frint_pattern><mode>2): Convert to insn.
13390 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
13391 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
13392 (frint_pattern): Likewise.
13393 (frint_suffix): Likewise.
13395 2013-04-29 Richard Biener <rguenther@suse.de>
13397 PR tree-optimization/57081
13398 * loop-init.c: Include tree-flow.h.
13399 (loop_optimizer_finalize): Free number of iteration estimates.
13400 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
13402 2013-04-29 Jakub Jelinek <jakub@redhat.com>
13404 PR tree-optimization/57083
13405 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
13406 non-singleton shift count range, zero extend low_bound for uns case.
13408 * config/i386/predicates.md (general_vector_operand): New predicate.
13409 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
13410 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
13411 if they aren't nonimmediate operands. If their original values
13412 satisfy const_vector_equal_evenodd_p, don't shift them.
13413 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
13414 predicates. For the SSE4.1 case force operands[{1,2}] into registers
13415 if not nonimmediate_operand.
13416 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
13417 instead of register_operand.
13418 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
13420 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
13422 * stor-layout.c (finalize_size_functions): Allocate a structure and
13423 reset cfun before dumping the functions.
13425 2013-04-27 Jakub Jelinek <jakub@redhat.com>
13427 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
13430 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
13431 use xop_pmacsdqh if uns_p.
13432 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
13433 the immediate rotate count.
13435 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
13437 * rtl.h (struct rtx_def): Add comment for field jump.
13438 (LRA_SUBREG_P): New macro.
13439 * recog.c (register_operand): Check LRA_SUBREG_P.
13440 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13442 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13443 (lra_final_code_change): Skip subreg change for operators.
13444 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13445 if there are no operand changes.
13446 * lra-constraints.c (curr_insn_set): New.
13447 (match_reload): Set LRA_SUBREG_P.
13448 (emit_spill_move): Ditto.
13449 (check_and_process_move): Use curr_insn_set. Process only single
13450 set insns. Don't initialize sec_mem_p and change_p.
13451 (simplify_operand_subreg): Use LRA_SUBREG_P.
13452 (reg_in_class_p): New function.
13453 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13454 of #ifdef. Add code to remove cycling.
13455 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13456 non-null disp. Reload inner instead of disp when base and index
13457 are null. Try to put lo_sum into register.
13458 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13459 (check_and_process_move): Move code for move cost check to
13460 simple_move_p. Remove equiv_substitution.
13461 (simple_move_p): New function.
13462 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13463 curr_insn_set. Call check_and_process_move only for single set
13464 insns. Use the new function. Move call of check_and_process_move
13465 after operand equiv substitution and address process.
13467 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13470 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13471 with nonlocal goto receivers or returns twice calls, ignore
13472 unininitialized values from abnormal edges to nl goto receiver
13473 or returns twice call.
13475 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13477 PR tree-optimization/57051
13478 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13479 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13482 2013-04-26 Richard Biener <rguenther@suse.de>
13484 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13485 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13486 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13487 (fixup_loop_arrays_after_move): New function.
13488 (move_sese_region_to_fn): Properly outline the loop tree parts
13489 of the SESE region.
13491 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
13493 * config/i386/i386.md (type, unit): Fix long lines.
13495 2013-04-26 Richard Biener <rguenther@suse.de>
13497 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13498 (lto-streamer-out.o): Likewise.
13499 * cfgloop.c (init_loops_structure): Export, add struct function
13500 argument and adjust.
13501 (flow_loops_find): Adjust.
13502 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13503 (init_loops_structure): Declare.
13504 * lto-streamer-in.c: Include cfgloop.h.
13505 (input_cfg): Input the loop tree.
13506 * lto-streamer-out.c: Include cfgloop.h.
13507 (output_cfg): Output the loop tree.
13508 (output_struct_function_base): Do not drop PROP_loops.
13510 2013-03-26 Richard Biener <rguenther@suse.de>
13512 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13513 (pass_build_cfg): Provide PROP_loops.
13514 (move_sese_region_to_fn): Remove loops that are outlined into fn
13516 * tree-inline.c: Include cfgloop.h.
13517 (initialize_cfun): Do not drop PROP_loops.
13518 (copy_loops): New function.
13519 (copy_cfg_body): Copy loop structure.
13520 (tree_function_versioning): Initialize destination loop tree.
13521 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13522 (pass_parallelize_loops): Do IL verification.
13523 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13524 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13525 the CFG make sure we fixup loops as well.
13526 * tree-ssa-tail-merge.c: Include cfgloop.h.
13527 (replace_block_by): When merging loop latches mark loops for fixup.
13528 * lto-streamer-out.c (output_struct_function_base): Drop
13529 PROP_loops for now.
13530 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13531 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13532 * ipa-split.c: Include cfgloop.h.
13533 (split_function): Add the new return block to the loop tree root.
13534 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13535 whether we have removed the forwarder block.
13536 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13537 * cfgloop.h (place_new_loop): Declare.
13538 * cfgloopmanip.c (place_new_loop): Export.
13539 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13540 (tree-switch-conversion.o): Likewise.
13541 (tree-complex.o): Likewise.
13542 (tree-inline.o): Likewise.
13543 (tree-ssa-tailmerge.o): Likewise.
13544 (ipa-split.o): Likewise.
13545 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13546 (tree-ssa-copy.o): Likewise.
13547 * tree-switch-conversion.c: Include cfgloop.h
13548 (process_switch): If we emit a bit-test cascade, schedule loops
13550 * tree-complex.c: Include cfgloop.h.
13551 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13552 * asan.c: Include cfgloop.h.
13553 (create_cond_insert_point): Properly add new basic-blocks to
13554 loops, schedule loop fixup.
13555 * cfgloop.c (verify_loop_structure): Check that looks are not
13557 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
13559 (expand_omp_for_generic): Likewise.
13560 (expand_omp_sections): Likewise.
13561 (expand_omp_atomic_pipeline): Schedule loops for fixup.
13562 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
13563 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
13564 is initialized, not when loops are present.
13565 * tree-parloops.c (parallelize_loops): Remove checking here.
13566 * passes.c (init_optimization_passes): Schedule a copy-propagation
13567 pass before complete unrolling of inner loops.
13569 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13571 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
13572 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
13573 (colorize_init): Add argument to _WIN32 version.
13574 * toplev.c: Include diagnostic-color.h.
13575 (process_options): Default to -fdiagnostics-color=auto if
13576 GCC_COLORS env var is in the environment.
13577 * common.opt (fdiagnostics-color=): Add Var and Init.
13578 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
13579 env var is in the environment, the default is auto rather than never.
13581 * diagnostic.h (file_name_as_prefix): Add context argument.
13582 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
13583 the string as locus.
13584 * langhooks.c (lhd_print_error_function): Adjust caller.
13586 2013-04-25 Lawrence Crowl <crowl@google.com>
13588 * var-tracking.c (shared_hash_def::htab):
13589 Change type to hash_table. Update dependent calls and types.
13591 2013-04-25 Lawrence Crowl <crowl@google.com>
13593 * Makefile.in: Update as needed below.
13595 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
13596 Move declaration to after the type's method definitons.
13598 * attribs.c (htab_t scoped_attributes::attribute_hash):
13599 Change type to hash_table. Update dependent calls and types.
13601 * bitmap.c (htab_t bitmap_desc_hash):
13602 Change type to hash_table. Update dependent calls and types.
13604 * cselib.c (htab_t cselib_hash_table):
13605 Change type to hash_table. Update dependent calls and types.
13607 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
13608 (hash_string_slot_node): Move implementation into lto-streamer.h
13609 struct string_slot_hasher.
13610 (eq_string_slot_node): Likewise.
13612 * data-streamer-out.c: Update output_block::string_hash_table
13613 dependent calls and types.
13615 * dwarf2cfi.c (htab_t trace_index):
13616 Change type to hash_table. Update dependent calls and types.
13618 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
13619 Change type to hash_table. Update dependent calls and types.
13620 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
13621 (htab_t optimize_external_refs::map): Likewise.
13622 (htab_t output_comp_unit::extern_map): Likewise.
13623 (htab_t output_comdat_type_unit::extern_map): Likewise.
13624 (htab_t output_macinfo::macinfo_htab): Likewise.
13625 (htab_t optimize_location_lists::htab): Likewise.
13626 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13628 * except.c (htab_t ehspec_hash_type):
13629 Change type to hash_table. Update dependent calls and types.
13630 (assign_filter_values::ttypes): Likewise.
13631 (assign_filter_values::ehspec): Likewise.
13632 (sjlj_assign_call_site_values::ar_hash): Likewise.
13633 (convert_to_eh_region_ranges::ar_hash): Likewise.
13635 * gcse.c (htab_t pre_ldst_table):
13636 Change type to hash_table. Update dependent calls and types.
13638 * ggc-common.c (htab_t saving_htab):
13639 Change type to hash_table. Update dependent calls and types.
13640 (htab_t loc_hash): Likewise.
13641 (htab_t ptr_hash): Likewise.
13642 (call_count): Rename ggc_call_count.
13643 (call_alloc): Rename ggc_call_alloc.
13644 (loc_descriptor): Rename make_loc_descriptor.
13645 (add_statistics): Rename ggc_add_statistics.
13647 * ggc-common.c (saving_htab):
13648 Change type to hash_table. Update dependent calls and types.
13650 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13651 (push_gimplify_context): Likewise.
13652 (pop_gimplify_context): Likewise.
13653 (struct gimple_temp_hash_elt): Added.
13654 (struct gimplify_hasher): Likewise.
13655 (struct gimplify_ctx.temp_htab):
13656 Change type to hash_table. Update dependent calls and types.
13658 * gimple-fold.c: Include gimplify-ctx.h.
13660 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13661 Change type to hash_table. Update dependent calls and types.
13662 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13663 avoid potential global name collision.
13665 * gimplify.c: Include gimplify-ctx.h.
13666 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13667 (htab_t gimplify_ctx::temp_htab):
13668 Update dependent calls and types for new type hash_table.
13669 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13670 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13672 * gimplify-ctx.h: New.
13673 (struct gimple_temp_hash_elt): Move from gimplify.c.
13674 (class gimplify_hasher): New.
13675 (struct gimplify_ctx): Move from gimple.h.
13676 (htab_t gimplify_ctx::temp_htab):
13677 Change type to hash_table. Update dependent calls and types.
13679 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13680 (htab_t ivs_params::newivs_index):
13681 Change type to hash_table. Update dependent calls and types.
13682 (htab_t ivs_params::params_index): Likewise.
13683 (htab_t print_generated_program::params_index): Likewise.
13684 (htab_t gloog::newivs_index): Likewise.
13685 (htab_t gloog::params_index): Likewise.
13687 * graphite.c: Include graphite-htab.h.
13688 4htab_t graphite_transform_loops::bb_pbb_mapping):
13689 Change type to hash_table. Update dependent calls and types.
13691 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13692 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13693 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13695 * graphite-dependences.c: Include graphite-htab.h.
13696 (loop_is_parallel_p): Change hash table type of parameter.
13698 * graphite-htab.h: New.
13699 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13700 (extern find_pbb_via_hash): Move from graphite-poly.h.
13701 (extern loop_is_parallel_p): Move from graphite-poly.h.
13702 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13704 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13705 (extern loop_is_parallel_p): Move to graphite-htab.h.
13706 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13708 * haifa-sched.c (htab_t delay_htab):
13709 Change type to hash_table. Update dependent calls and types.
13710 (htab_t delay_htab_i2): Likewise.
13712 * ira-color.c (htab_t allocno_hard_regs_htab):
13713 Change type to hash_table. Update dependent calls and types.
13715 * ira-costs.c (htab_t cost_classes_htab):
13716 Change type to hash_table. Update dependent calls and types.
13718 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13719 Change type to hash_table. Update dependent calls and types.
13721 * loop-iv.c (htab_t bivs):
13722 Change type to hash_table. Update dependent calls and types.
13724 * loop-unroll.c (htab_t opt_info::insns_to_split):
13725 Change type to hash_table. Update dependent calls and types.
13726 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13728 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13729 (struct string_slot_hasher): New.
13730 (htab_t output_block::string_hash_table):
13731 Change type to hash_table. Update dependent calls and types.
13733 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13734 (htab_t file_name_hash_table):
13735 Change type to hash_table. Update dependent calls and types.
13737 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13740 * lto-streamer.c (htab_t tree_htab):
13741 Change type to hash_table. Update dependent calls and types.
13743 * omp-low.c: Include gimplify-ctx.h.
13745 * passes.c (htab_t name_to_pass_map):
13746 Change type to hash_table. Update dependent calls and types.
13747 (pass_traverse): Rename to passes_pass_traverse.
13749 * plugin.c (htab_t event_tab):
13750 Change type to hash_table. Update dependent calls and types.
13752 * postreload-gcse.c (htab_t expr_table):
13753 Change type to hash_table. Update dependent calls and types.
13754 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13756 * sese.c (debug_rename_map_1): Make extern.
13757 (htab_t copy_bb_and_scalar_dependences::rename_map):
13758 Change type to hash_table. Update dependent calls and types.
13760 * sese.h (extern debug_rename_map): Move to .c file.
13762 * store-motion.c (htab_t store_motion_mems_table):
13763 Change type to hash_table. Update dependent calls and types.
13765 * trans-mem.c (htab_t tm_new_mem_hash):
13766 Change type to hash_table. Update dependent calls and types.
13768 * tree-browser.c (htab_t TB_up_ht):
13769 Change type to hash_table. Update dependent calls and types.
13771 * tree-cfg.c (htab_t discriminator_per_locus):
13772 Change type to hash_table. Update dependent calls and types.
13774 * tree-complex.c: Include tree-hasher.h
13775 (htab_t complex_variable_components):
13776 Change type to hash_table. Update dependent calls and types.
13778 * tree-eh.c (htab_t finally_tree):
13779 Change type to hash_table. Update dependent calls and types.
13781 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13782 struct int_tree_hasher.
13783 (extern int_tree_map_eq): Likewise.
13784 (uid_decl_map_hash): Removed.
13785 (extern decl_tree_map_eq): Likewise.
13787 * tree-hasher.h: New.
13788 (struct int_tree_hasher): New.
13789 (typedef int_tree_htab_type): New.
13791 * tree-inline.c: Include gimplify-ctx.h.
13793 * tree-mudflap.c: Include gimplify-ctx.h.
13795 * tree-parloops.c: Include tree-hasher.h.
13796 (htab_t eliminate_local_variables_stmt::decl_address):
13797 Change type to hash_table. Update dependent calls and types.
13798 (htab_t separate_decls_in_region::decl_copies): Likewise.
13800 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13801 Change type to hash_table. Update dependent calls and types.
13803 * tree-sra.c (candidates):
13804 Change type to hash_table. Update dependent calls and types.
13806 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13808 (int_tree_map_hash): Likewise.
13810 * tree-ssa-dom.c (htab_t avail_exprs):
13811 Change type to hash_table. Update dependent calls and types.
13813 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13814 Change type to hash_table. Update dependent calls and types.
13816 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
13817 Change type to hash_table. Update dependent calls and types.
13819 * tree-ssa-phiopt.c (seen_ssa_names):
13820 Change type to hash_table. Update dependent calls and types.
13822 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
13823 Change type to hash_table. Update dependent calls and types.
13825 * tree-ssa-uncprop.c (equiv):
13826 Change type to hash_table. Update dependent calls and types.
13828 2013-04-25 Jakub Jelinek <jakub@redhat.com>
13830 PR rtl-optimization/57003
13831 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
13832 call note_stores with kill_clobbered_value callback again after
13833 killing regs_invalidated_by_call.
13835 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13837 * config/aarch64/aarch64-simd.md
13838 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
13839 (aarch64_simd_bsl<mode>): Likewise.
13840 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
13842 2013-04-25 Marek Polacek <polacek@redhat.com>
13844 PR tree-optimization/57066
13845 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
13847 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
13849 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
13851 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13853 * config/aarch64/aarch64-builtins.c
13854 (aarch64_fold_builtin): New.
13855 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
13856 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
13857 * config/aarch64/aarch64-simd-builtins.def (abs): New.
13858 * config/aarch64/arm_neon.h
13859 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
13861 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13862 Tejas Belagod <tejas.belagod@arm.com>
13864 * config/aarch64/aarch64-builtins.c
13865 (aarch64_gimple_fold_builtin): New.
13866 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
13867 * config/aarch64/aarch64-simd-builtins.def (addv): New.
13868 * config/aarch64/aarch64-simd.md (addpv4sf): New.
13869 (addvv4sf): Update.
13870 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
13872 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13874 * config/aarch64/aarch64.md
13875 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
13877 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13879 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
13880 (*ngcsi_uxtw): New pattern.
13882 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13883 Julian Brown <julian@codesourcery.com>
13885 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
13886 (TB_DREG): Add T_V4HF.
13887 (v4hf_UP): New macro.
13888 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13889 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13890 Handle initialisation of V4HF. Adjust initialisation of reinterpret
13892 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13893 (arm_vector_mode_supported_p): Handle V4HF.
13894 (arm_mangle_map): Handle V4HFmode.
13895 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
13896 * config/arm/arm_neon_builtins.def: Add entries for
13897 vcvtv4hfv4sf, vcvtv4sfv4hf.
13898 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
13899 (neon_vcvtv4hfv4sf): Likewise.
13900 * config/arm/neon-gen.ml: Handle half-precision floating point
13902 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
13903 * config/arm/arm_neon.h: Regenerate.
13904 * config/arm/neon.ml (type elts): Add F16.
13905 (type vectype): Add T_float16x4, T_floatHF.
13906 (type vecmode): Add V4HF.
13907 (type features): Add Requires_FP_bit feature.
13908 (elt_width): Handle F16.
13909 (elt_class): Likewise.
13910 (elt_of_class_width): Likewise.
13911 (mode_of_elt): Refactor.
13912 (type_for_elt): Handle F16, fix error messages.
13913 (vectype_size): Handle T_float16x4.
13914 (vcvt_sh): New function.
13915 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
13916 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
13917 (string_of_mode): Handle V4HF.
13918 * doc/arm-neon-intrinsics.texi: Regenerate.
13920 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13922 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
13923 format specifier in 'X' case.
13925 2013-04-25 Alan Modra <amodra@gmail.com>
13928 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
13929 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
13930 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
13931 Repeat for many other rotate/shift and mask patterns using subregs.
13932 Name lshiftrt insns.
13933 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
13934 on WORDS_BIG_ENDIAN.
13936 2013-04-25 Alan Modra <amodra@gmail.com>
13938 * config.gcc: Support little-endian powerpc-linux targets.
13939 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
13940 (LINK_OS_LINUX_SPEC): Define.
13941 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
13942 Preserve MASK_LITTLE_ENDIAN.
13943 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
13944 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
13945 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
13946 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
13947 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
13948 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
13949 Correct fp word order for little-endian. Don't shift toc entries
13950 smaller than a word for little-endian.
13951 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
13952 (bswapdi2 splits): Correct low-part subreg for little-endian.
13953 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
13954 low/high where such is correct only for be.
13955 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
13956 little-endian for -mcall-aixdesc.
13958 2013-04-25 Alan Modra <amodra@gmail.com>
13960 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
13961 replace_equiv_address_nv.
13963 2013-04-25 Alan Modra <amodra@gmail.com>
13965 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
13967 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13970 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13971 * rtl.h (struct rtx_def): ...
13973 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13975 PR rtl-optimizations/57046
13976 * lra-constraints (split_reg): Set up lra_risky_transformations_p
13977 for multi-reg splits.
13979 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
13981 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
13983 2013-04-24 Sterling Augustine <saugustine@google.com>
13985 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
13986 (comp_dir_string, debug_str_dwo_section): New.
13987 (DEBUG_STR_DWO_SECTION): Rename to ...
13988 (DEBUG_DWO_STR_SECTION): ... this.
13989 (DEBUG_NORM_STR_SECTION): Delete.
13990 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
13991 (DEBUG_STR_DWO_SECTION_FLAGS): New.
13992 (find_AT_string): Move most logic to ...
13993 (find_AT_string_in_table): ... here. New.
13994 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
13995 add_skeleton_AT_string. Delete logic.
13996 (output_skeleton_debug_sections): Remove call to
13997 add_top_level_skeleton_die_attrs.
13998 (add_comp_dir_attribute): Move logic to comp_dir_string.
13999 (dwarf2out_init): Initialize debug_str_dwo_section.
14000 (output_indirect_string): Call find_string_form.
14001 (output_indirect_strings): Rewrite.
14002 (prune_unused_types): Empty skeleton_debug_str_hash.
14003 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
14004 (dwarf2out_finish): Call output_indirect_strings.
14006 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14008 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
14010 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14012 * rtl.h (struct rtx_def): Add comment for field jump.
14013 (LRA_SUBREG_P): New macro.
14014 * recog.c (register_operand): Check LRA_SUBREG_P.
14015 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14017 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14018 (lra_final_code_change): Skip subreg change for operators.
14019 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14020 if there are no operand changes.
14021 * lra-constraints.c (curr_insn_set): New.
14022 (match_reload): Set LRA_SUBREG_P.
14023 (emit_spill_move): Ditto.
14024 (check_and_process_move): Use curr_insn_set. Process only single
14025 set insns. Don't initialize sec_mem_p and change_p.
14026 (simplify_operand_subreg): Use LRA_SUBREG_P.
14027 (reg_in_class_p): New function.
14028 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14029 of #ifdef. Add code to remove cycling.
14030 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14031 non-null disp. Reload inner instead of disp when base and index
14032 are null. Try to put lo_sum into register.
14033 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14034 (check_and_process_move): Move code for move cost check to
14035 simple_move_p. Remove equiv_substitution.
14036 (simple_move_p): New function.
14037 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14038 curr_insn_set. Call check_and_process_move only for single set
14039 insns. Use the new function. Move call of check_and_process_move
14040 after operand equiv substitution and address process.
14042 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
14044 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
14045 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
14046 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
14048 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14050 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
14052 2013-04-24 Marek Polacek <polacek@redhat.com>
14054 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
14055 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
14056 (select_loops_exit_conditions): Likewise.
14057 (number_of_iterations_for_all_loops): Likewise.
14058 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
14059 (scev_analysis): Likewise.
14061 2013-04-02 Catherine Moore <clm@codesourcery.com>
14062 Chao-ying Fu <fu@mips.com>
14064 * config/mips/micromips.md (jraddiusp): New pattern.
14065 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
14066 instruction if possible.
14068 2013-04-24 Alan Modra <amodra@gmail.com>
14070 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
14072 2013-04-24 Julian Brown <julian@codesourcery.com>
14073 Chung-Lin Tang <cltang@codesourcery.com>
14075 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
14076 dependency behavior in enumeration type DIE generation. Add TODO note
14077 to comments about future DW_FORM_sdata/udata re-work of related code.
14079 2013-04-23 Lawrence Crowl <crowl@google.com>
14081 * Makefile.in: Update as needed below.
14083 * hash-table.h (class hash_table):
14084 Correct many methods with parameter types compare_type to the correct
14085 value_type. (Correct code was unlikely to notice the change.)
14086 (hash_table::elements_with_deleted) New.
14087 (class hashtable::iterator): New.
14088 (hashtable::begin()): New.
14089 (hashtable::end()): New.
14090 (FOR_EACH_HASH_TABLE_ELEMENT): New.
14092 * statistics.c (statistics_hashes):
14093 Change type to hash_table. Update dependent calls and types.
14095 * tree-into-ssa.c (var_infos):
14096 Change type to hash_table. Update dependent calls and types.
14098 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
14099 Change type to hash_table. Update dependent calls and types.
14101 * tree-ssa-loop-im.c (struct mem_ref.refs):
14102 Change type to hash_table. Update dependent calls and types.
14104 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
14105 Change type to hash_table. Update dependent calls and types.
14107 * tree-ssa-sccvn.c (vn_tables_s::nary):
14108 Change type to hash_table. Update dependent calls and types.
14109 (vn_tables_s::phis): Likewise.
14110 (vn_tables_s::references): Likewise.
14112 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
14113 (vn_reference_eq): Update parameter and return types.
14115 * tree-ssa-structalias.c (pointer_equiv_class_table):
14116 Change type to hash_table. Update dependent calls and types.
14117 (location_equiv_class_table): Likewise.
14119 * tree-vect-data-refs.c: Consequential changes for making
14120 peeling a hash_table.
14122 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
14123 (destroy_loop_vec_info): Dependent hash_table update.
14125 * tree-vectorizer.h (peeling_htab):
14126 Change type to hash_table. Update dependent calls and types.
14128 2013-04-23 Shiva Chen <shiva0217@gmail.com>
14130 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
14131 to check the register content is equal or not.
14132 * lra-constraints.c (match_reload): Use lra_assign_reg_val
14133 to assign register content record.
14134 * lra-eliminations.c (update_reg_eliminate): Use
14135 lra_update_reg_val_offset to update register content offset.
14136 * lra-int.h (struct lra_reg): Add offset member.
14137 (lra_reg_val_equal_p): New static inline function.
14138 (lra_update_reg_val_offset): New static inline function.
14139 (lra_assign_reg_val): New static inline function.
14140 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
14141 to assign register content record.
14142 (initialize_lra_reg_info_element): Initial offset to zero.
14144 2013-04-23 Catherine Moore <clm@codesourcery.com>
14146 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
14147 operands. Record compression.
14149 2013-04-23 Xinliang David Li <davidxl@google.com>
14151 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
14153 2013-04-23 Richard Biener <rguenther@suse.de>
14155 PR middle-end/57036
14156 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
14157 parameter, only add abnormal goto edges from the copied body
14158 if the call could perform abnormal gotos.
14159 (copy_cfg_body): Adjust.
14161 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
14163 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
14165 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
14167 * coretypes.h (gimple_stmt_iterator): Add struct to make
14170 2013-04-23 Richard Biener <rguenther@suse.de>
14172 PR tree-optimization/57026
14173 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
14174 from SSA names occuring in abnormal PHI nodes.
14176 2013-04-22 Andi Kleen <ak@linux.intel.com>
14178 * lto/lto.c (print_lto_report_1): Fix LTO report names.
14180 2013-04-22 Andi Kleen <ak@linux.intel.com>
14182 * lto/lto.c (print_lto_report_1): Declare early.
14183 (read_cgraph_and_symbols): Call print_lto_report_1 early.
14185 2013-04-22 Andi Kleen <ak@linux.intel.com>
14187 * common.opt (-flto-report-wpa): Add.
14188 * doc/invoke.texi (-flto-report-wpa): Add.
14189 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
14192 2013-04-22 Xinliang David Li <davidxl@google.com>
14194 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
14195 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
14196 * Makefile.in: New dependency
14198 David Daney <ddaney.cavm@gmail.com>
14200 * configure.ac (gcc_cv_as_micromips_support): Use the
14201 --fatal-warnings option.
14202 * configure: Regenerate.
14204 2013-04-22 Marek Polacek <polacek@redhat.com>
14207 * tsan.c (instrument_expr): Don't instrument expression
14208 in case its size is zero.
14210 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
14214 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14216 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14218 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14220 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
14221 (gimple_stmt_iterator): New typedef.
14222 * gimple.h (gimple_stmt_iterator): Rename to...
14223 (gimple_stmt_iterator_d): ... This.
14224 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
14225 trees be valid for GIMPLE and GENERIC.
14226 (TARGET_GIMPLE_FOLD_BUILTIN): New.
14227 * gimple-fold.c (gimple_fold_call): Call target hook
14228 gimple_fold_builtin.
14229 * hooks.c (hook_bool_gsiptr_false): New.
14230 * hooks.h (hook_bool_gsiptr_false): New.
14231 * target.def (fold_stmt): New.
14232 * doc/tm.texi: Regenerate.
14234 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
14237 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
14238 a set sp if no stack realignment.
14240 2013-04-22 Nick Clifton <nickc@redhat.com>
14242 * config.gcc (tilegx-linux): Extend extra_objs rather than
14244 (tilepro-linux): Likewise.
14246 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14248 * config/aarch64/aarch64-builtins.c
14250 (CF0, CF1, CF2, CF3, CF4, CF10): New.
14251 (VAR<1-12>): Add MAP parameter.
14252 (BUILTIN_*): Likewise.
14253 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
14254 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
14255 (aarch64_ushl_n<mode>): Likewise.
14256 (aarch64_sshr_n<mode>): Likewise.
14257 (aarch64_ushr_n<mode>): Likewise.
14258 (aarch64_<maxmin><mode>): Likewise.
14259 (aarch64_sqrt<mode>): Likewise.
14260 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
14261 (vshr<q>_n_*): Likewise.
14263 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14265 * config/aarch64/aarch64-builtins.c
14266 (aarch64_simd_builtin_type_mode): Handle SF types.
14268 (BUILTIN_GPF): Define.
14269 (aarch64_init_simd_builtins): Handle SF types.
14270 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
14271 (frecps): Likewise.
14272 (frecpx): Likewise.
14273 * config/aarch64/aarch64-simd.md
14274 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
14275 (aarch64_frecpe<mode>): New.
14276 (aarch64_frecps<mode>): Likewise.
14277 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
14278 (v8type): Add frecp<esx>.
14279 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
14280 (aarch64_frecps<mode>): Likewise.
14281 * config/aarch64/iterators.md (FRECP): New.
14282 (frecp_suffix): Likewise.
14283 * config/aarch64/arm_neon.h
14284 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
14286 2013-04-22 Christian Bruel <christian.bruel@st.com>
14289 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
14290 (REG_CLASS_NAMES): Idem.
14291 (REG_CLASS_CONTENTS): Idem.
14292 (REGCLASS_HAS_FP_REG): Idem.
14293 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
14294 (sh_conditional_register_usage): Idem.
14296 2013-04-21 Jeff Law <law@redhat.com>
14298 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
14299 (ssa_forward_propagate_and_combine): Use it.
14301 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14303 * lra.c: Update the flow chart diagram.
14305 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14307 PR rtl-optimization/56847
14308 * lra-constraints.c (process_alt_operands): Discourage alternative
14309 with non-matche doffsettable memory constraint fro memory with
14312 2013-04-19 Richard Biener <rguenther@suse.de>
14314 PR tree-optimization/56982
14315 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
14317 * gimplify.c (gimplify_call_expr): Notice special calls.
14318 (gimplify_modify_expr): Likewise.
14319 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
14320 abnormal control flow receivers.
14321 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
14322 in the same way as cfun->has_nonlocal_labels.
14323 (gimple_purge_dead_abnormal_call_edges): Likewise.
14324 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
14325 receivers start a basic-block.
14327 2013-04-19 Richard Biener <rguenther@suse.de>
14329 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
14331 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
14332 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
14333 (SLP_TREE_LOAD_PERMUTATION): Add.
14334 (vect_transform_slp_perm_load): Adjust prototype.
14335 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
14336 (vect_free_slp_instance): Likewise.
14337 (vect_create_new_slp_node): Likewise.
14338 (vect_supported_slp_permutation_p): Remove.
14339 (vect_slp_rearrange_stmts): Adjust.
14340 (vect_supported_load_permutation_p): Likewise. Inline
14341 vect_supported_slp_permutation_p here.
14342 (vect_analyze_slp_instance): Compute load permutations per
14343 slp node instead of per instance.
14344 (vect_get_slp_defs): Adjust.
14345 (vect_transform_slp_perm_load): Likewise.
14346 (vect_schedule_slp_instance): Remove redundant code.
14347 (vect_schedule_slp): Remove hack for PR56270, add it ...
14348 * tree-vect-stmts.c (vectorizable_load): ... here, do not
14349 CSE loads for SLP. Adjust.
14351 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14353 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
14354 spelling in two comments.
14356 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14359 * config/arm/arm.c (load_multiple_sequence): Require SP
14360 as base register for loads if SP is in the register list.
14362 2013-04-19 Martin Jambor <mjambor@suse.cz>
14364 PR tree-optimization/56718
14365 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
14366 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
14367 and made public. Adjusted all callers.
14368 (ipa_intraprocedural_devirtualization): New function.
14369 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
14370 (ipa_intraprocedural_devirtualization): Likewise.
14371 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
14373 2013-04-19 Richard Biener <rguenther@suse.de>
14375 PR tree-optimization/57000
14376 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
14378 2013-04-19 Terry Guo <terry.guo@arm.com>
14380 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
14382 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
14383 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
14384 (cortex_m4_fmacs): Use new reservations.
14385 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
14387 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
14389 PR rtl-optimization/56999
14390 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
14392 (lra_coalesce): Remove split_origin_bitmap and related code.
14393 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
14394 ranges if necessary.
14396 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
14398 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
14400 (ix86_expand_call): Remove clobbered_registers array and use
14401 x86_64_ms_sysv_extra_clobbered_registers instead.
14402 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
14404 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
14406 * config/i386/i386.md (*call_rex64_ms_sysv): Use
14407 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
14408 (*call_value_rex64_ms_sysv): Ditto.
14410 2013-04-18 Cary Coutant <ccoutant@google.com>
14412 * dwarf2out.c (output_pubnames): Check die_perennial_p of
14413 parent instead of die_mark.
14415 2013-04-18 Diego Novillo <dnovillo@google.com>
14417 * gimple.c (create_gimple_tmp): New.
14418 (get_expr_type): New.
14419 (build_assign): New.
14420 (build_type_cast): New.
14421 * gimple.h (enum ssa_mode): Define.
14422 (gimple_seq_set_location): New.
14423 * asan.c (build_check_stmt): Change some gimple_build_* calls
14424 to use build_assign and build_type_cast.
14426 2013-04-18 Richard Biener <rguenther@suse.de>
14428 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
14429 handle negative step. Remove redundant checks.
14430 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
14431 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
14432 for negative step and grouped loads fail to vectorize.
14434 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14436 * emit-rtl.c (reset_insn_used_flags): New function.
14437 (reset_all_used_flags): Use it.
14438 (verify_insn_sharing): New function.
14439 (verify_rtl_sharing): Fix verification for SEQUENCEs.
14441 2013-04-18 Jakub Jelinek <jakub@redhat.com>
14443 PR tree-optimization/56984
14444 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
14445 and (x >> M) >= N don't register any assertion if N << M is the
14448 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14450 * lower-subreg.c (resolve_simple_move): If called self-recursive,
14451 do not delete_insn insns that have not yet been emitted, only
14452 unlink them with remove_insn.
14453 * df-scan.c (df_insn_delete): Revert r197492.
14455 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
14457 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
14458 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
14460 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14462 * config/arm/arm.md (movsicc_insn): Convert define_insn into
14463 define_insn_and_split.
14464 (and_scc,ior_scc,negscc): Likewise.
14465 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14467 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14469 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14470 can benefit from using a sequence of LDRD instructions in epilogue
14471 instead of a single LDM instruction.
14473 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14476 * doc/extend.texi: Fix typo.
14478 2013-04-17 Richard Biener <rguenther@suse.de>
14480 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14481 (vect_build_slp_tree): ... here.
14482 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14483 match. Remove special-casing of mismatched complex loads.
14484 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14485 re-try the match with swapped commutative operands.
14486 (vect_supported_load_permutation_p): Remove special-casing of
14487 mismatched complex loads.
14488 (vect_analyze_slp_instance): Adjust.
14490 2013-04-17 Richard Biener <rguenther@suse.de>
14492 PR rtl-optimization/56921
14493 * cfgloop.h (struct loop): Add simple_loop_desc member.
14494 (struct niter_desc): Mark with GTY(()).
14495 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14496 * loop-iv.c (get_simple_loop_desc): Likewise.
14497 (free_simple_loop_desc): Likewise.
14500 2013-04-16 Richard Biener <rguenther@suse.de>
14502 PR rtl-optimization/56921
14503 * loop-init.c (pass_rtl_move_loop_invariants): Add
14504 TODO_do_not_ggc_collect to todo_flags_finish.
14505 (pass_rtl_unswitch): Same.
14506 (pass_rtl_unroll_and_peel_loops): Same.
14507 (pass_rtl_doloop): Same.
14509 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14511 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14512 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14513 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14515 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14516 * tree-streamer.c (record_common_node): Adjust reference in comment.
14518 2013-04-17 Terry Guo <terry.guo@arm.com>
14520 * config/arm/cortex-m4.md: Add a new bypass.
14522 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14524 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14526 (*subs_<optab><mode>_multp2): New pattern.
14527 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14528 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14530 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14532 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14533 (*subs_mul_imm_<mode>): New pattern.
14535 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
14538 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14539 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14540 (vsx_movti_32bit): Change j->wa to O->wa.
14542 2013-04-16 Richard Biener <rguenther@suse.de>
14544 PR rtl-optimization/56921
14545 * loop-init.c (pass_rtl_move_loop_invariants): Add
14546 TODO_do_not_ggc_collect to todo_flags_finish.
14547 (pass_rtl_unswitch): Same.
14548 (pass_rtl_unroll_and_peel_loops): Same.
14549 (pass_rtl_doloop): Same.
14551 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
14553 * config/arm/arm.c (emit_multi_reg_push): New declaration
14554 for an existing function.
14555 (arm_emit_strd_push): New function.
14556 (arm_expand_prologue): Used here.
14557 (arm_emit_ldrd_pop): New function.
14558 (arm_expand_epilogue): Used here.
14559 (arm_get_frame_offsets): Update condition.
14560 (arm_emit_multi_reg_pop): Add a special case for load of a single
14561 register with writeback.
14563 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
14565 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
14568 2013-04-16 Richard Biener <rguenther@suse.de>
14570 PR tree-optimization/56756
14571 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
14572 (first_mem_ref_loc): New.
14573 (execute_sm): Place the load temporarily before a previous
14574 access instead of in the latch edge to ensure its SSA dependencies
14575 are defined at points dominating the load.
14577 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
14579 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
14580 correct fix by moving header and footer insn to the footer of
14581 the merged basic block. Clear BB_END of the merged-away block.
14583 PR middle-end/43631
14584 * emit-rtl.c (make_note_raw): New function.
14585 (link_insn_into_chain): New static inline function.
14586 (add_insn): Use it.
14587 (add_insn_before, add_insn_after): Factor insn chain linking code...
14588 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
14589 using link_insn_into_chain.
14590 (note_outside_basic_block_p): New helper function for emit_note_after
14591 and emit_note_before.
14592 (emit_note_after): Use nobb variant of add_insn_after if the note
14593 should not be contained in a basic block.
14594 (emit_note_before): Use nobb variant of add_insn_before if the note
14595 should not be contained in a basic block.
14596 (emit_note_copy): Use make_note_raw.
14597 (emit_note): Likewise.
14598 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
14599 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
14600 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
14601 the moved barrier the tail of the basic block it follows.
14602 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
14604 2013-04-15 Jakub Jelinek <jakub@redhat.com>
14606 PR tree-optimization/56962
14607 * gimple-ssa-strength-reduction.c (record_increment): Only set
14608 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
14609 either rhs1 or rhs2 is equal to c->base_expr.
14611 2013-04-15 Richard Biener <rguenther@suse.de>
14613 PR tree-optimization/56933
14614 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
14616 (GROUP_READ_WRITE_DEPENDENCE): Remove.
14617 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
14618 * tree-vect-data-refs.c (vect_analyze_group_access): Move
14619 dependence check ...
14620 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14622 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14623 GROUP_READ_WRITE_DEPENDENCE.
14625 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14627 * emit-rtl.c (reset_all_used_flags): New function.
14628 (verify_rtl_sharing): Call reset_all_used_flags before and after
14629 performing the checks.
14631 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14633 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14634 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14635 * config/arm/constraints.md (De): New constraint.
14636 * config/arm/neon.md (anddi3_neon): Delete.
14637 (neon_vand<mode>): Expand to standard anddi3 pattern.
14638 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14639 Move earlier in the file.
14640 (neon_inv_logic_op2): Likewise.
14641 (arm_anddi_operand_neon): New predicate.
14643 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14645 * configure.ac (gcc_cv_ld_as_needed): Set
14646 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14647 Use -z ignore, -z record on *-*-solaris2*.
14648 (HAVE_LD_AS_NEEDED): Update comment.
14649 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14650 * configure: Regenerate.
14651 * config.in: Regenerate.
14652 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14653 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14654 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14655 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14656 equivalents. Fix markup.
14657 * doc/tm.texi: Regenerate.
14659 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14661 * config/i386/i386.opt: New option mstack-protector-guard=.
14662 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14663 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14664 TARGET_SSP_TLS_GUARD.
14665 * config/i386/i386.c (ix86_option_override_internal): Set
14666 ix86_stack_protector_guard.
14667 * config/i386/i386.md (stack_protect_set): Enable for
14668 TARGET_SSP_TLS_GUARD only.
14669 (stack_protect_set_<mode>): Ditto.
14670 (stack_protect_test): Ditto.
14671 (stack_protect_test_<mode>): Ditto.
14672 * doc/invoke.texi (i386 Option): Document.
14674 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14677 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14678 (S_MODES): Set H_MODE bit.
14679 (SF_MODES): Set only S_MODE and SF_MODE bits.
14680 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14681 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14682 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14683 <MODE_FLOAT>: Likewise.
14685 2013-04-15 Joey Ye <joey.ye@arm.com>
14687 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14689 2013-04-15 Joey Ye <joey.ye@arm.com>
14691 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14693 (thumb_far_jump_used_p): Count instruction size and set
14696 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14698 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14699 * resource.c (mark_target_live_regs): Reformat conditional block.
14701 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14703 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14704 notes, they are emitted only just before final.
14705 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14707 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14709 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14710 * cfgrtl.c (delete_insn): Call it here instead.
14711 * lra-spills.c (lra_final_code_change): Use delete_insn.
14712 * haifa-sched.c (sched_remove_insn): Likewise.
14713 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14714 returning to the nop pool.
14715 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14716 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14718 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14720 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14721 * doc/tm.texi: Regenerated.
14723 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
14725 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14728 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14730 * df-core.c (df_find_def): Compare register numbers.
14731 (df_find_use): Likewise.
14733 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14736 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14737 lra_in_progress for return.
14739 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14741 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14742 define_insn into define_insn_and_split and emit movsicc patterns.
14744 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14746 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14748 2013-04-12 Richard Biener <rguenther@suse.de>
14750 * tree-pass.h (TODO_do_not_ggc_collect): New.
14751 * passes.c (execute_one_ipa_transform_pass): Honor
14752 TODO_do_not_ggc_collect.
14753 (execute_one_pass): Likewise.
14756 2013-04-10 Richard Biener <rguenther@suse.de>
14758 * passes.c (init_optimization_passes): Remove reload pass.
14759 * ira.c (do_reload): Merge into ...
14761 (rest_of_handle_reload): Remove.
14762 (pass_reload): Likewise.
14763 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14764 of reload for vzeroupper pass placement.
14766 2013-04-12 Jakub Jelinek <jakub@redhat.com>
14768 PR tree-optimization/56918
14769 PR tree-optimization/56920
14770 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14771 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14772 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14773 use wide_mul_with_sign method.
14775 2013-04-12 Richard Biener <rguenther@suse.de>
14777 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14778 not be considered a gimple constant.
14780 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
14782 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14783 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14786 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14787 Jakub Jelinek <jakub@redhat.com>
14789 * opts.c: Include diagnostic-color.h.
14790 (common_handle_option): Handle OPT_fdiagnostics_color_.
14791 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14792 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14793 (diagnostic-color.o): New.
14794 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14795 (diagnostic_color_rule): New enum.
14796 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14797 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14798 the location string.
14799 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14800 either NULL, or color kind.
14801 * diagnostic-color.c: New file.
14802 * diagnostic-color.h: New file.
14803 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14805 * doc/invoke.texi (-fdiagnostics-color): Document.
14806 * pretty-print.h (pp_show_color): Define.
14807 (struct pretty_print_info): Add show_color field.
14808 * diagnostic.c: Include diagnostic-color.h.
14809 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14810 macros. Colorize error:, warning: etc. strings and also the location
14812 (diagnostic_show_locus): Colorize the caret line.
14813 * pretty-print.c: Include diagnostic-color.h.
14814 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14815 inside of %< %> quotes or quoted through q format modifier.
14817 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14819 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
14821 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14823 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
14824 code in CC_NZ mode.
14825 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
14828 2013-04-11 Marek Polacek <polacek@redhat.com>
14830 PR tree-optimization/48184
14831 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
14833 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
14835 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
14836 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
14837 (skip_simple_arithmetic): Tidy up.
14838 * tree.h (skip_simple_constant_arithmetic): Declare.
14840 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14842 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
14844 2013-04-11 Richard Biener <rguenther@suse.de>
14846 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14847 generate vector constants.
14849 2013-04-11 Richard Biener <rguenther@suse.de>
14851 PR tree-optimization/56878
14852 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
14853 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
14855 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
14856 Prefer to align the DR with the most invariant base address.
14858 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14860 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
14863 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
14865 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
14866 floating-point vector comparisons against 0.
14868 2013-04-11 Jakub Jelinek <jakub@redhat.com>
14870 PR tree-optimization/56899
14871 * fold-const.c (extract_muldiv_1): Apply distributive law
14872 only if TYPE_OVERFLOW_WRAPS (ctype).
14874 2013-04-11 Bin Cheng <bin.cheng@arm.com>
14877 * ira-costs.c (scan_one_insn): Check whether the source rtx of
14878 loading has side effect.
14880 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
14882 * config/sparc/sparc.c: Include tree-pass.h.
14883 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
14884 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
14885 head of file. Change return type. Split off gate function.
14886 (sparc_gate_work_around_errata): New function.
14887 (pass_work_around_errata): New pass definition.
14888 (insert_pass_work_around_errata) New pass insert definition to
14889 insert pass_work_around_errata just after delayed-branch scheduling.
14890 (sparc_option_override): Insert the pass.
14891 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
14893 2013-04-10 David S. Miller <davem@davemloft.net>
14895 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
14896 or -mcpu=hypersparc.
14898 * target.def (cstore_mode): New hook.
14899 * target.h: Include insn-codes.h
14900 * targhooks.c: Likewise.
14901 (default_cstore_mode): New function.
14902 * targhooks.h: Declare it.
14903 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
14904 * doc/tm.texi: Rebuild.
14905 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
14906 target hook, rather than inspecting the insn_data.
14907 * config/sparc/sparc.c (sparc_cstore_mode): New function.
14908 (TARGET_CSTORE_MODE): Redefine.
14909 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
14911 * config/sparc/predicates.md (cstore_result_operand): New special
14913 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
14914 Use it for operand 0.
14915 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
14916 (*snesi_special): Likewise.
14917 (*snesi_zero): Likewise.
14918 (*seqsi_zero): Likewise.
14919 (*sltu_insn): Likewise.
14920 (*sgeu_insn): Likewise.
14921 (*seqdi_special): Make operand 0 and comparison operation be of
14923 (*snedi_special): Likewise.
14924 (*snedi_special_vis3): Likewise.
14925 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
14926 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
14927 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
14928 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
14929 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
14930 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
14931 (*sltu_extend_sp64): Likewise.
14932 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
14933 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
14934 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
14935 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
14936 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
14938 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
14940 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
14941 (aarch64_start_file): Use the new function.
14943 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14944 Jason Merrill <jason@redhat.com>
14946 * common.opt: Add -gdwarf.
14947 * opts.c (common_handle_option): Handle it.
14948 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
14950 2013-04-10 Richard Biener <rguenther@suse.de>
14952 * passes.c (execute_todo): Do not call ggc_collect conditional here.
14953 (execute_one_ipa_transform_pass): But unconditionally here.
14954 (execute_one_pass): And here.
14955 (init_optimization_passes): Remove reload pass.
14956 * tree-pass.h (TODO_ggc_collect): Remove.
14957 (pass_reload): Likewise.
14958 * ira.c (do_reload): Merge into ...
14960 (rest_of_handle_reload): Remove.
14961 (pass_reload): Likewise.
14962 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14963 of reload for vzeroupper pass placement.
14964 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
14965 and todo_flags_finish of all passes.
14967 2013-04-10 Richard Biener <rguenther@suse.de>
14969 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
14970 first_const_oprnd field, rename first_def_type to first_op_type.
14971 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
14972 (vect_get_and_check_slp_defs): Always use the type of the
14973 operand. Allow mixed vect_external_def, vect_constant_def types.
14974 (vect_get_constant_vectors): Handle mixed vect_external_def,
14975 vect_constant_def types.
14977 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
14979 PR tree-optimization/55524
14980 * tree-ssa-math-opts.c
14981 (convert_mult_to_fma): Don't use an fms construct
14982 when we don't have an fms operation, but fnma, and it looks
14983 likely that we'll be able to use the latter.
14985 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
14987 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
14989 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
14990 inline fail caused by overwritable functions.
14992 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
14994 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
14995 unnecessary bits in the constant power of two case.
14997 2013-04-10 Richard Biener <rguenther@suse.de>
14999 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
15000 broken code swapping operands.
15001 (vect_build_slp_tree): Do not compute load permutations here.
15002 (vect_analyze_slp_instance): Compute load permutations here,
15003 after building the SLP tree.
15005 2013-04-09 Christian Bruel <christian.bruel@st.com>
15007 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
15008 of next/prev_real_insn.
15010 2013-04-09 Jan Hubicka <jh@suse.cz>
15012 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15013 Drop aliased parameter.
15014 (function_and_variable_visibility): Do not handle alias pairs.
15015 * cgraph.c (varpool_externally_visible_p): Update prototype.
15016 * varpool.c (varpool_add_new_variable): Update.
15018 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15020 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
15022 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
15024 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
15026 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
15028 2013-04-09 Marek Polacek <polacek@redhat.com>
15030 PR tree-optimization/48762
15031 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
15033 2013-04-09 Richard Biener <rguenther@suse.de>
15035 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
15037 (vect_build_slp_tree): Likewise.
15038 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
15039 calculating the cost of a SLP instance.
15040 (vect_analyze_slp_instance): Use it from here, after building
15043 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15045 PR middle-end/56883
15046 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15047 expand_omp_for_static_chunk): Use simple_p = true in
15048 force_gimple_operand_gsi calls when assigning to addressable decls.
15050 2013-04-09 Jeff Law <law@redhat.com>
15052 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
15053 when the boolean was created by converting a wider object which
15054 had a boolean range.
15056 2013-04-09 Richard Biener <rguenther@suse.de>
15058 * tree-vectorizer.h (slp_void_p): Remove.
15059 (slp_tree): Typedef before _slp_tree declaration.
15060 (struct _slp_tree): Use a vector of slp_tree as children.
15061 (vect_get_place_in_interleaving_chain): Remove.
15062 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
15064 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
15066 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
15067 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
15068 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
15069 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
15070 Use slp_node instead of slp_void_p and adjust.
15072 2013-04-09 Richard Biener <rguenther@suse.de>
15074 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
15075 work that is not necessary.
15077 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15079 PR tree-optimization/56854
15080 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
15081 forward into clobber stmts if it would change MEM_REF lhs into
15084 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
15086 * tree.c (type_hash_lookup, type_hash_add): Make static.
15087 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
15089 2013-04-09 Richard Biener <rguenther@suse.de>
15091 * tree.h (unsave_expr_now): Remove.
15092 * tree-inline.c (mark_local_for_remap_r): Remove.
15093 (unsave_expr_1): Likewise.
15094 (unsave_r): Likewise.
15095 (unsave_expr_now): Likewise.
15096 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
15097 (propagate_tree_value): Likewise.
15099 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
15101 * doc/rtl.texi (sequence): Rewrite documentation to match the
15102 current use of SEQUENCE rtl objects.
15103 * rtl.def (SEQUENCE): Likewise.
15105 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
15106 Update documentation.
15107 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
15108 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
15110 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
15112 2013-04-08 Teresa Johnson <tejohnson@google.com>
15114 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
15115 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
15117 (estimate_edge_size_and_time): Add comment to suggest using rounding
15119 (estimate_node_size_and_time): Ditto.
15120 (remap_edge_change_prob): Use helper rounding divide methods.
15121 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
15122 (gimple_mod_pow2_value_transform): Ditto.
15123 (gimple_mod_subtract_transform): Ditto.
15124 (gimple_ic_transform): Ditto.
15125 (gimple_stringops_transform): Ditto.
15126 * stmt.c (conditional_probability): Ditto.
15127 (emit_case_dispatch_table): Ditto.
15128 * lto-cgraph.c (merge_profile_summaries): Ditto.
15129 * tree-optimize.c (execute_fixup_cfg): Ditto.
15130 * cfgcleanup.c (try_forward_edges): Ditto.
15131 * cfgloopmanip.c (scale_loop_profile): Ditto.
15133 (duplicate_loop_to_header_edge): Ditto.
15134 (lv_adjust_loop_entry_edge): Ditto.
15135 * tree-vect-loop.c (vect_transform_loop): Ditto.
15136 * profile.c (compute_branch_probabilities): Ditto.
15137 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
15138 * lto-streamer-in.c (input_cfg): Ditto.
15139 * gimple-streamer-in.c (input_bb): Ditto.
15140 * ipa-cp.c (update_profiling_info): Ditto.
15141 (update_specialized_profile): Ditto.
15142 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
15143 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
15145 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15146 (compute_trg_info): Ditto.
15147 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
15148 (purge_dead_edges): Ditto.
15149 * loop-unswitch.c (unswitch_loop): Ditto.
15150 * cgraphclones.c (cgraph_clone_edge): Ditto.
15151 (cgraph_clone_node): Ditto.
15152 * tree-inline.c (copy_bb): Ditto.
15153 (copy_edges_for_bb): Ditto.
15154 (initialize_cfun): Ditto.
15155 (copy_cfg_body): Ditto.
15156 (expand_call_inline): Ditto.
15158 2013-04-08 Kai Tietz <ktietz@redhat.com>
15160 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
15161 TARGET_CYGWIN64 by TARGET_64BIT.
15163 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15165 * config/epiphany/epiphany.md (GPR_1): New constant.
15166 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
15167 * config/epiphany/epiphany.c (gen_compare_reg):
15168 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
15169 is already in place.
15170 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
15171 Don't require being called during rtl expansion; If y operlaps r0,
15173 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
15174 (epiphany_expand_epilogue): Likewise.
15176 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
15177 Don't use CC_FPmode for ORDERED / UNORDERED.
15178 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
15180 * config/epiphany/constraints.md (CnL): New constraint.
15181 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
15182 * config/epiphany/predicates.md (add_operand): Allow 1024.
15184 * config/epiphany/epiphany.md (logical_op): New code iterator.
15185 (op_mnc): New code attribute.
15186 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
15187 (mov_f+1, mov_f+2): New peephole2 patterns.
15189 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
15190 (cstoresi4): Also allow re-use of zero result when doing a NE
15191 comparison to a non-zero operand.
15192 Use (clobber (scratch)) for first insn if the gpr output is not needed.
15194 * config/epiphany/epiphany.md (<insn_opname>v2si3):
15195 Use gen_addsi3_i / gen_subsi3_i.
15197 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15201 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
15202 contain anything but clobbers, at most one __builtin_stack_restore,
15203 optionally debug stmts and final resx, and if it has at least one
15204 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
15205 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
15206 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
15207 which isn't defaut definition, remove them.
15208 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
15209 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
15210 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
15211 with MEM_REF LHS with SSA_NAME address.
15213 2013-04-08 Jeff Law <law@redhat.com>
15215 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
15217 2013-04-08 Richard Biener <rguenther@suse.de>
15219 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
15221 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
15222 determined vector type.
15223 (vect_analyze_data_refs): Likewise.
15224 (vect_get_new_vect_var): Adjust.
15225 (vect_create_destination_var): Preserve SSA name versions.
15226 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
15227 not dump anything here.
15229 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15231 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
15232 Add member lr_slot_known.
15233 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
15235 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15236 Remove code that sets lr_slot_offset according to what a previous
15237 version of epiphany_emit_save_restore used to do.
15238 (epiphany_emit_save_restore): When doing an lr save or restore,
15239 set/verify lr_slot_known and lr_slot_offset.
15241 2013-04-08 Xinyu Qi <xyqi@marvell.com>
15244 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
15247 2013-04-08 Richard Biener <rguenther@suse.de>
15249 * alias.c (find_base_term): Fix thinko in previous change.
15251 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15253 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
15254 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
15255 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
15256 if possible to compute val.
15257 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
15258 For QImode integers don't require anything about precision. Use
15259 const_with_all_bytes_same to find out if the constant doesn't have
15260 repeated bytes in it.
15262 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15264 * config/s390/s390.c (s390_expand_insv): Only accept insertions
15267 2013-04-08 Marek Polacek <polacek@redhat.com>
15269 PR rtl-optimization/48182
15270 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
15273 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15276 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
15277 nuses, make sure we have a label.
15279 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15282 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
15283 (rs6000_emit_swdiv_low_precision): Remove.
15284 (rs6000_emit_swdiv): Rewrite to handle between one and four
15285 iterations of Newton-Raphson generally; modify required number of
15286 iterations for some cases.
15287 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
15289 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
15291 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
15292 set-but-unused variable.
15294 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
15295 basic blocks of released function bodies garbage-collectable.
15297 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
15298 (struct rtl_opt_pass): Add TODO_df_finish.
15300 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
15302 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15304 * config/arm/constraints.md (q): New constraint.
15305 * config/arm/ldrdstrd.md: New file.
15306 * config/arm/arm.md (ldrdstrd.md) New include.
15307 (arm_movdi): Use "q" instead of "r" constraint
15308 for double-word memory access.
15309 (movdf_soft_insn): Likewise.
15310 * config/arm/vfp.md (movdi_vfp): Likewise.
15311 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
15312 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
15313 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
15314 (mem_ok_for_ldrd_strd): Likewise.
15315 (output_move_double): Update assertion.
15317 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15319 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
15321 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15323 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
15324 define_insn_and_split.
15325 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
15327 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15329 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
15330 define_insn_and_split.
15331 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
15332 (shiftsi3_compare): New pattern.
15333 (rrx): New pattern.
15334 * config/arm/unspecs.md (UNSPEC_RRX): New.
15336 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15338 * config/arm/arm.md (negdi_extendsidi): New pattern.
15339 (negdi_zero_extendsidi): Likewise.
15341 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15343 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
15344 define_insn_and_split.
15345 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
15346 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
15348 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15350 * config/arm/arm.md (arm_subdi3): Convert define_insn into
15351 define_insn_and_split.
15352 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
15353 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
15355 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15357 * config/arm/arm.md (subsi3_carryin): New pattern.
15358 (subsi3_carryin_const): Likewise.
15359 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
15360 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
15362 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15364 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
15366 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15368 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
15369 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
15371 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15373 * config/arm/arm.c (arm_expand_builtin): Change fcode
15374 type to unsigned int.
15376 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15378 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
15380 2013-04-04 Ian Lance Taylor <iant@google.com>
15382 * doc/standards.texi (Standards): The Go frontend supports the Go 1
15385 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
15387 PR middle-end/56729
15388 * df-scan.c (df_insn_delete): Disable failing assert.
15390 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15392 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
15393 New function prototype.
15394 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
15395 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
15396 (arm_builtin_vectorized_function): New function.
15398 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15400 * config/arm/arm_neon_builtins.def: New file.
15401 * config/arm/arm.c (neon_builtin_data): Move contents to
15402 arm_neon_builtins.def.
15403 (enum arm_builtins): Include neon builtin definitions.
15404 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
15405 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
15407 2013-04-04 Marek Polacek <polacek@redhat.com>
15409 PR tree-optimization/48186
15410 * predict.c (maybe_hot_frequency_p): Return false if
15411 HOT_BB_FREQUENCY_FRACTION is 0.
15412 (cgraph_maybe_hot_edge_p): Likewise.
15414 2013-04-04 Richard Biener <rguenther@suse.de>
15416 PR tree-optimization/56826
15417 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
15420 2013-04-04 Richard Biener <rguenther@suse.de>
15422 PR tree-optimization/56213
15423 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
15424 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
15426 2013-04-04 Richard Biener <rguenther@suse.de>
15428 PR tree-optimization/56837
15429 * tree-loop-distribution.c (classify_partition): For non-zero
15430 values require that the value has the same precision as its
15431 mode to be useful as memset value.
15433 2013-04-03 Nick Clifton <nickc@redhat.com>
15435 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
15436 (fmssf4): Use fmsf.s on E3V5 architectures.
15437 (fnmasf4): Use fnmaf.s on E3V5 architectures.
15438 (fnmssf4): Use fnmsf.s on E3V5 architectures.
15440 2013-04-03 Jeff Law <law@redhat.com>
15442 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
15443 (lra-eliminations.o): Likewise.
15445 2013-04-03 Teresa Johnson <tejohnson@google.com>
15447 * gcov-io.c (compute_working_sets): Moved most of body of old
15448 compute_working_sets here from profile.c.
15449 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
15450 (gcov_working_set_t): Moved typedef here from basic-block.h
15451 (compute_working_set): Declare.
15452 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
15453 (get_working_sets): Renamed from compute_working_set,
15454 replace most of body with call to new compute_working_sets.
15455 (get_exec_counts): Replace call to compute_working_sets
15456 to get_working_sets.
15457 * profile.h (get_working_sets): Renamed from compute_working_set.
15458 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
15459 to get_working_sets.
15460 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
15461 * gcov-dump.c (dump_working_sets): New function.
15463 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
15465 * hwint.c (sext_hwi, zext_hwi): New functions.
15466 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15467 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15468 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15469 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15470 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15471 (sext_hwi, zext_hwi): New functions.
15473 2013-04-03 Jeff Law <law@redhat.com>
15475 PR tree-optimization/56799
15476 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15477 back test for widening conversion erroneously dropped in prior change.
15479 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15482 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15483 instead of next_real_insn.
15485 2013-04-03 Marek Polacek <polacek@redhat.com>
15488 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
15490 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15493 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15495 (thumb1_output_casesi): Likewise.
15496 (thumb2_output_casesi): Likewise.
15498 2013-04-03 Richard Biener <rguenther@suse.de>
15500 PR tree-optimization/56817
15501 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15503 (tree_unroll_loops_completely_1): ... new function to manually
15504 walk the loop tree, properly defering outer loops of unrolled
15505 loops to later iterations.
15507 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15509 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15510 (vectorizable_load): Likewise.
15511 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15512 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15514 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15516 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15519 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15521 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15523 2013-04-03 Bin Cheng <bin.cheng@arm.com>
15525 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15527 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15529 PR tree-optimization/56790
15530 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15533 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15535 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15537 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15538 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15541 2013-04-03 Jakub Jelinek <jakub@redhat.com>
15544 * tree.h (force_folding_builtin_constant_p): New decl.
15545 * builtins.c (force_folding_builtin_constant_p): New variable.
15546 (fold_builtin_constant_p): Fold immediately also if
15547 force_folding_builtin_constant_p.
15549 2013-04-03 Richard Biener <rguenther@suse.de>
15551 PR tree-optimization/56812
15552 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15553 DRs of the same interleaving chain are independent.
15555 2013-04-02 Jason Merrill <jason@redhat.com>
15557 * gdbinit.in (pbb): Use debug fn.
15559 2013-04-02 Lawrence Crowl <crowl@google.com>
15561 * sese.h (struct ivtype_map_elt_s): Remove unused.
15562 (extern debug_ivtype_map): Remove unused.
15563 (extern eq_ivtype_map_elts): Remove unused.
15564 * sese.c (debug_ivtype_map): Removed unused.
15565 (debug_ivtype_map_1): Removed unused.
15566 (debug_ivtype_elt): Remove unused.
15567 (eq_ivtype_map_elts): Remove unused.
15570 2013-04-02 Kai Tietz <ktietz@redhat.com>
15573 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
15574 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
15575 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
15577 (legitimize_pe_coff_symbol): Likewise.
15578 (is_imported_p): New helper-function.
15579 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
15580 for Windows x64 targets.
15581 (ix86_expand_prologue): Optimize for pe-coff targets.
15582 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
15583 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
15584 medium/large code-model.
15585 (legitimize_pic_address): Likewise.
15586 (legitimize_tls_address): Likewise.
15587 (ix86_expand_call): Likewise.
15588 (x86_output_mi_thunk): Likewise.
15589 (get_dllimport_decl): Add new beimport argument.
15590 (construct_plt_address): Don't assert for x64 pe-coff targets.
15591 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
15593 (SYMBOL_FLAG_STUBVAR): New macro.
15594 (SYMBOL_REF_STUBVAR_P): Likewise.
15595 * config/i386/winnt.c (stub_list): New structure.
15596 (stub_head): New local variable.
15597 (i386_pe_record_stub): New function.
15598 (i386_pe_file_end): Emit refptr-stubs.
15600 2013-04-02 Jakub Jelinek <jakub@redhat.com>
15602 PR rtl-optimization/56745
15603 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
15604 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
15607 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
15608 and both of them are MEM_REFs, just compare first argument for
15609 equality and attempt to deal even with differing offsets.
15612 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
15613 of gimple_clobber_p to be MEM_REF.
15614 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
15615 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
15616 after gimplification.
15617 * asan.c (get_mem_ref_of_assignment): Don't instrument
15618 gimple_clobber_p stmts.
15619 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
15620 gimple_clobber_p stmt if they have MEM_REF lhs and
15621 are dead because of another gimple_clobber_p stmt.
15622 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15623 gimple_clobber_p stmts like debug stmts.
15624 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15625 that refer to unused VAR_DECLs or uninitialized values.
15626 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15627 gimple_clobber_p stmts if they refer to removed parameters.
15628 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15631 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
15633 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15634 using SWI48 mode attribute.
15636 2013-04-02 Wei Mi <wmi@google.com>
15638 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15639 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15640 *<rotate_insn><mode>3_mask in i386.md.
15642 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15644 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15646 2013-04-02 Richard Biener <rguenther@suse.de>
15648 PR tree-optimization/56778
15649 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15650 Runtime alias tests are not supported for gather loads.
15651 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15652 stmts referenced from SSA operands before updating SSA form.
15654 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15655 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15657 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
15658 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15659 * config/arm/cortex-a53.md: New file.
15660 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15661 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15662 * config/arm/arm.c (arm_issue_rate): Likewise.
15663 * config/arm/arm-tune.md: Regenerate
15664 * config/arm/arm-tables.opt: Regenerate.
15665 * config/arm/arm-cores.def: Add cortex-a53.
15667 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15669 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15672 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15674 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15675 scalar load/store operations using B/H registers.
15676 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15678 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15680 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15682 * config/aarch64/aarch64.c
15683 (aarch64_simd_scalar_immediate_valid_for_move): New.
15684 * config/aarch64/aarch64-protos.h
15685 (aarch64_simd_scalar_immediate_valid_for_move): New.
15686 * config/aarch64/constraints.md (Dh, Dq): New.
15687 * config/aarch64/iterators.md (hq): New.
15689 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15691 * reorg.c (get_branch_condition): Deal with conditional returns.
15692 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15694 2013-04-01 Wei Mi <wmi@google.com>
15696 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15697 Truncate operand 2 using %b asm operand modifier.
15698 (*<shift_insn><mode>3_mask): Ditto.
15699 (*<rotate_insn><mode>3_mask): Ditto.
15701 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15703 PR middle-end/56798
15704 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15706 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15708 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15710 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15712 2013-03-30 Lawrence Crowl <crowl@google.com>
15714 * dse.c (clear_alias_sets): Remove never set.
15715 (disqualified_clear_alias_sets): Remove never set.
15716 (clear_alias_mode_pool): Remove never set.
15717 (dse_step0): Remove condition that is never true.
15718 (canon_address): Remove condition that is never true.
15719 (dse_step7): Remove condition that is never true.
15720 (rest_of_handle_dse): Remove condition that is never true.
15721 (rest_of_handle_dse::did_global): Remove never read from above.
15722 (dse_step2_spill): Remove never called from above.
15723 (dse_step5_spill): Remove never called from above.
15725 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15727 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15728 JUMP_TABLE_DATA changes.
15729 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15730 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15731 (Insns) <jump_table_data>: New entry.
15732 * doc/tm.texi: Regenerate.
15734 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15736 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15737 for table jump at the end of a basic block using tablejump_p.
15738 * targhooks.c (default_invalid_within_doloop): Likewise.
15739 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15740 target hook implementation that is identical to the default hook.
15741 (rs6000_invalid_within_doloop): Remove.
15743 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15744 unused variable from tablejump_p call.
15746 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15747 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15748 (INSN_DELETED_P): Likewise.
15749 (emit_jump_table_data): New prototype.
15750 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15751 after 4th as unused.
15752 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15753 * sched-vis.c (print_insn): Likewise.
15754 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15755 insn for compatibility with back ends that use next_active_insn to
15756 identify jump table data.
15757 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15758 (remove_insn): Likewise.
15759 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15761 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15762 (emit_jump_table_data): New function.
15764 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15765 basic block, a JUMP_TABLE_DATA never is.
15766 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15767 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15768 off from code handling real insns.
15769 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15770 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15771 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15772 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15773 is not a NONDEBUG_INSN_P.
15774 * ira-costs.c (scan_one_insn): Likewise.
15775 * jump.c (mark_all_labels): Likewise.
15776 (mark_jump_label_1): Likewise.
15777 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15778 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15780 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15781 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15782 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15783 JUMP_TABLE_DATA_P insns.
15784 (calculate_elim_costs_all_insns): Likewise.
15785 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15786 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15787 (delete_output_reload): Code style fixups.
15788 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15789 insn flags on this non-insn.
15790 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15791 as scheduling barriers, for pre-change compatibility.
15792 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15793 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15795 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15796 redundant JUMP_TABLE_DATA_P test.
15797 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15798 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15799 (frv_for_each_packet): Likewise.
15800 * config/i386/i386.c (min_insn_size): Likewise.
15801 (ix86_avoid_jump_mispredicts): Likewise.
15802 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15803 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15804 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15805 (mips16_insn_length): Robustify.
15806 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15807 (mips16_split_long_branches): Likewise.
15808 * config/pa/pa.c (pa_combine_instructions): Likewise.
15809 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15810 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15811 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15812 as contributing to pool range lengths.
15813 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15814 Remove redundant JUMP_TABLE_DATA_P test.
15815 (sh_loop_align): Likewise.
15816 (split_branches): Likewise.
15817 (sh_insn_length_adjustment): Likewise.
15818 * config/spu/spu.c (get_branch_target): Likewise.
15820 2013-03-29 Jan Hubicka <jh@suse.cz>
15822 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
15823 gcov streaming; stream hot bb threshold to ltrans.
15824 * predict.c (get_hot_bb_threshold): Break out from ....
15825 (maybe_hot_count_p): ... here.
15826 (set_hot_bb_threshold): New function.
15827 * lto-section-in.c (lto_section_name): Add profile.
15828 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
15829 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
15830 and data-streamer.h
15831 (histogram_entry): New structure.
15832 (histogram, histogram_pool): New global vars.
15833 (histogram_hash): New structure.
15834 (histogram_hash::hash): New method.
15835 (histogram_hash::equal): Likewise.
15836 (account_time_size): New function.
15837 (cmp_counts): New function.
15838 (dump_histogram): New function.
15839 (ipa_profile_generate_summary): New function.
15840 (ipa_profile_write_summary): New function.
15841 (ipa_profile_read_summary): New function.
15842 (ipa_profile): Decide on threshold.
15843 (pass_ipa_profile): Add ipa_profile_write_summary and
15844 ipa_profile_read_summary.
15845 * Makefile.in (ipa.o): Update dependencies.
15846 * lto-streamer.h (LTO_section_ipa_profile): New section.
15848 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
15850 * tree.h (VAR_P): New.
15852 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
15855 * doc/invoke.texi ([-fwhole-program]): Fix typo.
15857 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
15859 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
15860 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15861 (control_flow_insn_p): Likewise.
15862 * cfgrtl.c (duplicate_insn_chain): Likewise.
15863 * final.c (get_attr_length_1): Likewise.
15864 (shorten_branches): Likewise.
15865 (final_scan_insn): Likewise.
15866 * function.c (instantiate_virtual_regs): Likewise.
15867 * gcse.c (insert_insn_end_basic_block): Likewise.
15868 * ira-costs.c (scan_one_insn): Likewise.
15869 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15870 * lra.c (check_rtl): Likewise.
15871 * reload1.c (elimination_costs_in_insn): Likewise.
15872 * reorg.c (follow_jumps): Likewise.
15874 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
15875 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15876 (thumb_far_jump_used_p): Likewise.
15877 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
15878 (workaround_speculation): Likewise.
15879 (add_sched_insns_for_speculation): Likewise.
15880 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
15881 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15882 (frv_for_each_packet): Likewise.
15883 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
15884 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
15885 (final_emit_insn_group_barriers): Likewise.
15886 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15887 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15888 (mips16_insn_length): Likewise.
15889 * config/pa/pa.c (pa_reorg): Likewise.
15890 (pa_combine_instructions): Likewise.
15891 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
15892 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
15893 (sh_reorg): Likewise.
15894 (split_branches): Likewise.
15895 * config/spu/spu.c (get_branch_target): Likewise.
15897 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
15900 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
15902 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
15903 Fix declaration name.
15905 2013-03-28 Lawrence Crowl <crowl@google.com>
15907 * graphds.h (struct graph.indicies): Remove unused.
15908 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
15909 (SCOP_ORIGINAL_PDDRS): Remove unused.
15910 * sese.h (extern insert_loop_close_phis): Removed unused.
15911 (extern insert_guard_phis): Removed unused.
15912 (extern ivtype_map_elt_info): Removed unused.
15913 (new_ivtype_map_elt): Removed unused.
15914 * sese.c (ivtype_map_elt_info): Removed unused.
15916 2013-03-28 Lawrence Crowl <crowl@google.com>
15918 * Makefile.in: Add several missing include dependences.
15920 (test-dump.o): New. This object is not added to any executable,
15921 but is present for ad-hoc testing.
15923 (debug (const bitmap_head_def &)): New.
15924 (debug (const bitmap_head_def *)): New.
15926 (extern debug (const bitmap_head_def &)): New.
15927 (extern debug (const bitmap_head_def *)): New.
15929 (debug (edge_def &)): New.
15930 (debug (edge_def *)): New.
15932 (debug (basic_block_def &)): New.
15933 (debug (basic_block_def *)): New.
15935 (dump_node (const_tree, int, FILE *)): Correct source file.
15937 (debug (die_struct &)): New.
15938 (debug (die_struct *)): New.
15940 (extern debug (die_struct &)): New.
15941 (extern debug (die_struct *)): New.
15942 * gimple-pretty-print.c
15943 (debug (gimple_statement_d &)): New.
15944 (debug (gimple_statement_d *)): New.
15945 * gimple-pretty-print.h
15946 (extern debug (gimple_statement_d &)): New.
15947 (extern debug (gimple_statement_d *)): New.
15949 (debug (ira_allocno_copy &)): New.
15950 (debug (ira_allocno_copy *)): New.
15951 (debug (ira_allocno &)): New.
15952 (debug (ira_allocno *)): New.
15954 (extern debug (ira_allocno_copy &)): New.
15955 (extern debug (ira_allocno_copy *)): New.
15956 (extern debug (ira_allocno &)): New.
15957 (extern debug (ira_allocno *)): New.
15959 (debug (live_range &)): New.
15960 (debug (live_range *)): New.
15962 (debug (lra_live_range &)): New.
15963 (debug (lra_live_range *)): New.
15965 (debug (lra_live_range &)): New.
15966 (debug (lra_live_range *)): New.
15968 (debug (omega_pb_d &)): New.
15969 (debug (omega_pb_d *)): New.
15971 (extern debug (omega_pb_d &)): New.
15972 (extern debug (omega_pb_d *)): New.
15974 (debug (const rtx_def &)): New.
15975 (debug (const rtx_def *)): New.
15977 (debug_tree (tree): Move within file.
15978 (debug_raw (const tree_node &)): New.
15979 (debug_raw (const tree_node *)): New.
15980 (dump_tree_via_hooks (const tree_node *, int)): New.
15981 (debug (const tree_node &)): New.
15982 (debug (const tree_node *)): New.
15983 (debug_verbose (const tree_node &)): New.
15984 (debug_verbose (const tree_node *)): New.
15985 (debug_head (const tree_node &)): New.
15986 (debug_head (const tree_node *)): New.
15987 (debug_body (const tree_node &)): New.
15988 (debug_body (const tree_node *)): New.
15989 (debug_vec_tree (tree): Move and reimplement in terms of dump.
15990 (debug (vec<tree, va_gc> &)): New.
15991 (debug (vec<tree, va_gc> *)): New.
15993 (extern debug (const rtx_def &)): New.
15994 (extern debug (const rtx_def *)): New.
15996 (debug_raw (simple_bitmap_def &)): New.
15997 (debug_raw (simple_bitmap_def *)): New.
15998 (debug (simple_bitmap_def &)): New.
15999 (debug (simple_bitmap_def *)): New.
16001 (extern debug (simple_bitmap_def &)): New.
16002 (extern debug (simple_bitmap_def *)): New.
16003 (extern debug_raw (simple_bitmap_def &)): New.
16004 (extern debug_raw (simple_bitmap_def *)): New.
16006 (debug (vinsn_def &)): New.
16007 (debug (vinsn_def *)): New.
16008 (debug_verbose (vinsn_def &)): New.
16009 (debug_verbose (vinsn_def *)): New.
16010 (debug (expr_def &)): New.
16011 (debug (expr_def *)): New.
16012 (debug_verbose (expr_def &)): New.
16013 (debug_verbose (expr_def *)): New.
16014 (debug (vec<rtx> &)): New.
16015 (debug (vec<rtx> *)): New.
16017 (extern debug (vinsn_def &)): New.
16018 (extern debug (vinsn_def *)): New.
16019 (extern debug_verbose (vinsn_def &)): New.
16020 (extern debug_verbose (vinsn_def *)): New.
16021 (extern debug (expr_def &)): New.
16022 (extern debug (expr_def *)): New.
16023 (extern debug_verbose (expr_def &)): New.
16024 (extern debug_verbose (expr_def *)): New.
16025 (extern debug (vec<rtx> &)): New.
16026 (extern debug (vec<rtx> *)): New.
16028 (_list_iter_cond_expr): Make inline instead of static.
16030 (debug (sreal &)): New.
16031 (debug (sreal *)): New.
16033 (extern debug (sreal &)): New.
16034 (extern debug (sreal *)): New.
16036 (extern debug_raw (const tree_node &)): New.
16037 (extern debug_raw (const tree_node *)): New.
16038 (extern debug (const tree_node &)): New.
16039 (extern debug (const tree_node *)): New.
16040 (extern debug_verbose (const tree_node &)): New.
16041 (extern debug_verbose (const tree_node *)): New.
16042 (extern debug_head (const tree_node &)): New.
16043 (extern debug_head (const tree_node *)): New.
16044 (extern debug_body (const tree_node &)): New.
16045 (extern debug_body (const tree_node *)): New.
16046 (extern debug (vec<tree, va_gc> &)): New.
16047 (extern debug (vec<tree, va_gc> *)): New.
16049 (debug (struct loop &)): New.
16050 (debug (struct loop *)): New.
16051 (debug_verbose (struct loop &)): New.
16052 (debug_verbose (struct loop *)): New.
16053 * tree-dump.c: Add header dependence.
16055 (extern debug (struct loop &)): New.
16056 (extern debug (struct loop *)): New.
16057 (extern debug_verbose (struct loop &)): New.
16058 (extern debug_verbose (struct loop *)): New.
16060 (debug (data_reference &)): New.
16061 (debug (data_reference *)): New.
16062 (debug (vec<data_reference_p> &)): New.
16063 (debug (vec<data_reference_p> *)): New.
16064 (debug (vec<ddr_p> &)): New.
16065 (debug (vec<ddr_p> *)): New.
16067 (extern debug (data_reference &)): New.
16068 (extern debug (data_reference *)): New.
16069 (extern debug (vec<data_reference_p> &)): New.
16070 (extern debug (vec<data_reference_p> *)): New.
16071 (extern debug (vec<ddr_p> &)): New.
16072 (extern debug (vec<ddr_p> *)): New.
16074 (debug (pt_solution &)): New.
16075 (debug (pt_solution *)): New.
16077 (extern debug (pt_solution &)): New.
16078 (extern debug (pt_solution *)): New.
16080 (debug (_var_map &)): New.
16081 (debug (_var_map *)): New.
16082 (debug (tree_live_info_d &)): New.
16083 (debug (tree_live_info_d *)): New.
16085 (extern debug (_var_map &)): New.
16086 (extern debug (_var_map *)): New.
16087 (extern debug (tree_live_info_d &)): New.
16088 (extern debug (tree_live_info_d *)): New.
16090 2013-03-28 Jan Hubicka <jh@suse.cz>
16092 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
16094 2013-03-28 Ian Bolton <ian.bolton@arm.com>
16096 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
16097 record only when desired or required.
16099 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
16101 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
16102 *vec_extractv2di_1_rex64. Use x64 isa attribute.
16104 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16106 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
16107 (*andsi3_compare0_uxtw): New pattern.
16108 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
16109 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
16111 2013-03-28 Jan Hubicka <jh@suse.cz>
16113 * data-streamer-in.c (streamer_read_gcov_count): New function.
16114 * gimple-streamer-out.c: Include value-prof.h.
16115 (output_gimple_stmt): Output histogram.
16116 (output_bb): Use streamer_write_gcov_count.
16117 * value-prof.c: Include data-streamer.h
16118 (dump_histogram_value): Add HIST_TYPE_MAX.
16119 (stream_out_histogram_value): New function.
16120 (stream_in_histogram_value): New function.
16121 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
16122 (stream_out_histogram_value, stream_in_histogram_value): Declare.
16123 * data-streamer-out.c (streamer_write_gcov_count): New function.
16124 (streamer_write_gcov_count_stream): New function.
16125 * lto-cgraph.c (lto_output_edge): Update counter streaming.
16126 (lto_output_node): Likewise.
16127 (input_node, input_edge): Likewise.
16128 * lto-streamer-out.c (output_cfg): Update streaming.
16129 * lto-streamer-in.c (input_cfg): Likewise.
16130 * data-streamer.h (streamer_write_gcov_count,
16131 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
16132 * gimple-streamer-in.c: Include value-prof.h
16133 (input_gimple_stmt): Input histograms.
16134 (input_bb): Update profile streaming.
16136 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16138 * genmodes.c (emit_max_int): New function.
16139 (emit_insn_modes_h): Added call to emit_max_function.
16140 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
16142 * machmode.def: Fixed comment.
16144 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16146 * combine.c (try_combine): Removed useless assert.
16147 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
16149 2013-03-28 Marek Polacek <polacek@redhat.com>
16150 Richard Biener <rguenther@suse.de>
16152 PR tree-optimization/56695
16153 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
16154 build signed result of a vector comparison.
16155 * tree-cfg.c (verify_gimple_comparison): Check that a result
16156 of a vector comparison has signed type.
16158 2013-03-28 Richard Biener <rguenther@suse.de>
16160 PR tree-optimization/37021
16161 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
16162 do not restrict gaps between groups.
16163 * tree-vect-stmts.c (vectorizable_load): Properly account for
16164 a gap between groups.
16166 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
16168 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
16169 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
16172 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
16174 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
16175 * doc/extend.texi (Named Address Spaces): Ditto.
16176 (Variable Attributes): Ditto.
16178 2013-03-27 Kai Tietz <ktietz@redhat.com>
16180 * config.build: Add support for cygwin x64 target.
16181 * config.gcc: Likewise.
16182 * config.host: Likewise.
16183 * configure.ac: Likewise
16184 * configure: Regenerated.
16186 2013-03-27 Kai Tietz <ktietz@redhat.com>
16188 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
16189 * config/i386/t-cygwin-w64: New file.
16190 * config/i386/cygwin-w64.h: New file.
16191 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
16192 and add support for x64-cygwin target.
16193 (CPP_SPEC): Likewise.
16194 (CXX_WRAP_SPEC_LIST): Undefine before define.
16195 (LIBGCJ_SONAME): Use 15 as version.
16197 2013-03-27 Richard Biener <rguenther@suse.de>
16199 PR tree-optimization/56716
16200 * tree-ssa-structalias.c (perform_var_substitution): Adjust
16201 dumping for ref nodes.
16203 2013-03-27 Martin Jambor <mjambor@suse.cz>
16205 PR tree-optimization/55334
16206 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
16207 restricted pointers to arrays.
16209 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
16211 * Makefile.in (.SUFFIXES): Add .cc.
16212 (.c.o): Apply same recipe for implicit rule .cc.o.
16214 2013-03-27 Richard Biener <rguenther@suse.de>
16216 PR tree-optimization/37021
16217 * tree-vect-data-refs.c (vect_check_strided_load): Allow
16218 REALPART/IMAGPART_EXPRs around the supported refs.
16219 * tree-ssa-structalias.c (find_func_aliases): Assume that
16220 floating-point values are not used to transfer pointers.
16222 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
16224 * target.def (TARGET_HAS_IFUNC_P): New target hook.
16225 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
16226 * doc/tm.texi: Regenerate.
16227 * targhooks.h (default_has_ifunc_p): New.
16228 * targhooks.c (default_has_ifunc_p): Ditto.
16229 * config/linux-protos.h: New file.
16230 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
16231 hook for linux which disables support of indirect functions in android.
16232 * config/linux-android.c: New file.
16233 * config/t-linux-android.c: Ditto.
16234 * config.gcc: Added new object file linux-android.o.
16235 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
16236 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
16237 * varasm.c (do_assemble_alias): Likewise.
16238 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
16239 doesn't support indirect functions.
16240 * configure: Regenerate.
16242 2013-03-27 Bin Cheng <bin.cheng@arm.com>
16245 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
16246 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
16249 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16251 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
16253 2013-03-27 Terry Guo <terry.guo@arm.com>
16255 * config/arm/arm-cores.def: Added core cortex-r7.
16256 * config/arm/arm-tune.md: Regenerated.
16257 * config/arm/arm-tables.opt: Regenerated.
16258 * doc/invoke.texi: Added entry for core cortex-r7.
16260 2013-03-27 Walter Lee <walt@tilera.com>
16262 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
16263 double-decrement of next_scratch_regno.
16265 2013-03-27 Walter Lee <walt@tilera.com>
16267 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
16269 (insn_v1mulus): Ditto.
16270 (insn_v2muls): Ditto.
16272 2013-03-27 Walter Lee <walt@tilera.com>
16274 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
16275 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
16277 2013-03-27 Walter Lee <walt@tilera.com>
16279 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
16280 (*sibcall_value): Ditto.
16282 2013-03-27 Walter Lee <walt@tilera.com>
16284 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
16285 (insn_mnz_v8qi): ... this ...
16286 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
16288 (insn_v<n>mnz): Replaced by ...
16289 (insn_v1mnz): ... this ...
16290 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
16292 (insn_mz_<mode>): Replaced by ...
16293 (insn_mz_v8qi): ... this ...
16294 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
16296 (insn_v<n>mz): Replaced by ...
16297 (insn_v1mz): ... this ...
16298 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
16301 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
16303 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
16305 2013-03-26 Roland McGrath <mcgrathr@google.com>
16307 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
16308 than fprintf with a non-constant, non-format string.
16310 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16312 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
16313 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
16314 operand 0 predicate.
16315 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
16316 attribute. Use general_x64nomem_operand as operand 1 predicate.
16317 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
16318 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16319 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
16320 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16321 (mov<mode>_insv_1): Remove expander. Merge insn with
16322 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
16323 Use general_x64nomem_operand as operand 1 predicate.
16324 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
16325 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
16326 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16327 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
16328 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
16329 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
16330 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16331 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
16332 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
16333 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
16334 (general_x64nomem_operand): Ditto.
16336 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
16338 * config/rtems.opt: Add -pthread option.
16340 2013-03-26 Richard Biener <rguenther@suse.de>
16342 * alias.c (find_base_term): Avoid redundant and not used recursion.
16343 (base_alias_check): Get the initial base term from the caller.
16344 (true_dependence_1): Compute and pass base terms to base_alias_check.
16345 (write_dependence_p): Likewise.
16346 (may_alias_p): Likewise.
16348 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
16350 * config/aarch64/aarch64.c (aarch64_classify_address): Support
16351 PC-relative load in SI modes and above only.
16353 2013-03-26 Xinyu Qi <xyqi@marvell.com>
16355 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
16356 * config/arm/iwmmxt.md (WCGR0): Update.
16357 (WCGR1, WCGR2, WCGR3): Likewise.
16359 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16361 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
16362 Use x64 and nox64 isa attributes.
16364 2013-03-26 Richard Biener <rguenther@suse.de>
16366 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
16367 alignment computations and rely on get_object_alignment_1
16368 for the !TYPE_P case.
16369 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
16371 2013-03-26 Walter Lee <walt@tilera.com>
16373 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
16374 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
16376 2013-03-25 Jeff Law <law@redhat.com>
16378 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
16379 check for INTEGRAL_TYPE_P that was missing due to checking in
16380 wrong version of prior patch.
16382 2013-03-25 Walter Lee <walt@tilera.com>
16384 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
16385 TILEGX_INSN_SHUFFLEBYTES1.
16386 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
16388 (tilegx_builtins): Ditto.
16389 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
16391 2013-03-25 Walter Lee <walt@tilera.com>
16393 * config/tilegx/tilegx.md (floatsisf2): New pattern.
16394 (floatunssisf2): New pattern.
16395 (floatsidf2): New pattern.
16396 (floatunssidf2): New pattern.
16398 2013-03-25 Walter Lee <walt@tilera.com>
16400 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
16401 tests for constraint J, K, N, P.
16403 2013-03-25 Walter Lee <walt@tilera.com>
16405 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
16406 Use indirect/pcrel encoding.
16407 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
16410 2013-03-25 Steve Ellcey <sellcey@mips.com>
16412 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
16413 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
16414 * config/mips/mips.c (mips_option_override): Set IMADD default.
16415 * config/mips/mips.h (PTF_AVOID_IMADD): New.
16416 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
16417 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
16418 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
16419 * doc/invoke.texi (-mimadd/-mno-imadd): New.
16421 2013-03-25 Jeff Law <law@redhat.com>
16423 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
16424 slightly to avoid creating and folding useless trees. Simplify
16425 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
16427 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
16429 * config/i386/i386.md (*zero_extendsidi2): Merge with
16430 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
16431 * config/i386/predicates.md (x86_64_zext_operand): Rename from
16432 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
16433 targets. Clarify comment.
16435 2013-03-25 Martin Jambor <mjambor@suse.cz>
16437 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
16438 pass-through jump functions differently.
16439 (ipa_read_jump_function): Likewise. Also use setter functions to set
16442 2013-03-25 Martin Jambor <mjambor@suse.cz>
16444 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
16445 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
16447 (ipa_get_indirect_edge_target): New function.
16448 (devirtualization_time_bonus): New parameter known_aggs, pass it to
16449 ipa_get_indirect_edge_target. Update all callers.
16450 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
16451 ipa_get_indirect_edge_target_1 instead of calling
16452 ipa_get_indirect_edge_target.
16453 (create_specialized_node): Pass aggvlas to
16454 ipcp_discover_new_direct_edges.
16456 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16458 * config/arm/arm.md (f_sels, f_seld): New types.
16459 (*cmov<mode>): New pattern.
16460 * config/arm/predicates.md (arm_vsel_comparison_operator): New
16463 2013-03-25 Kai Tietz <ktietz@redhat.com>
16465 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16466 POSIX-printf for mingw-hosted builds.
16468 2013-03-25 Richard Biener <rguenther@suse.de>
16470 PR middle-end/56694
16471 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16472 must-not-throw stmt location.
16474 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16476 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16477 Emit load-acquire versions when acq is true.
16478 (arm_emit_store_exclusive): Add rel parameter.
16479 Emit store-release versions when rel is true.
16480 (arm_split_compare_and_swap): Use acquire-release instructions
16482 of barriers when appropriate.
16483 (arm_split_atomic_op): Likewise.
16484 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16485 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16486 (VUNSPEC_SLX): Likewise.
16487 (VUNSPEC_LDA): Likewise.
16488 (VUNSPEC_STL): Likewise.
16489 * config/arm/sync.md (atomic_load<mode>): New pattern.
16490 (atomic_store<mode>): Likewise.
16491 (arm_load_acquire_exclusive<mode>): Likewise.
16492 (arm_load_acquire_exclusivesi): Likewise.
16493 (arm_load_acquire_exclusivedi): Likewise.
16494 (arm_store_release_exclusive<mode>): Likewise.
16496 2013-03-25 Catherine Moore <clm@codesourcery.com>
16498 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16499 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16500 * config/mip/predicates.md (lwsp_swsp_operand,
16501 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16502 sb16_operand, db4_operand, db7_operand, ib3_operand,
16503 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16504 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16505 andi16_operand): New predicates.
16506 * config/mips/mips.md (compression): New attribute.
16507 (enabled): New attribute.
16508 (length): Consider compression in computing length.
16509 (shift_compression): New code attribute.
16510 (*add<mode>3): New operands. Record compression.
16511 (sub<mode>3): Likewise.
16512 (one_cmpl<mode>2): Likewise.
16513 (*and<mode>3): Likewise.
16514 (*ior<mode>3): Likewise.
16515 (unnamed pattern for xor): Likewise.
16516 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16517 (*<optab><mode>3): Likewise.
16518 (*mov<mode>_internal: Likewise.
16519 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16520 (mips_unsigned_immediate_p): New.
16521 (umips_lwsp_swsp_address_p): New.
16522 (m16_based_address_p): New.
16523 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16524 (mips_unsigned_immediate_p): New prototype.
16525 (lwsp_swsp_address_p): New prototype.
16526 (m16_based_address_p): New prototype.
16527 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16528 (mips_signed_immediate_p): New function.
16529 (m16_based_address_p): New function.
16530 (lwsp_swsp_address_p): New function.
16531 (mips_print_operand_punctuation): Recognize short delay slot insns
16532 for microMIPS.add<mode>3"
16534 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16537 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16538 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16540 2013-03-25 Richard Biener <rguenther@suse.de>
16542 PR tree-optimization/56689
16543 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16546 2013-03-25 Richard Biener <rguenther@suse.de>
16548 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16550 (memory_references): Likewise.
16551 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16552 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16553 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16554 (gather_mem_refs_in_loops): Fold into ...
16555 (analyze_memory_references): ... this. Move initialization
16556 to tree_ssa_lim_initialize.
16557 (fill_always_executed_in): Rename to ...
16558 (fill_always_executed_in_1): ... this.
16559 (fill_always_executed_in): Move contains_call computation to
16560 this new function from ...
16561 (tree_ssa_lim_initialize): ... here.
16562 (tree_ssa_lim): Call fill_always_executed_in.
16564 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16566 * postreload.c (reload_combine): Fix code detecting returns.
16568 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16570 * function.c (emit_use_return_register_into_block): On cc0 targets,
16571 do not emit the sequence between cc0 setter and user.
16573 2013-03-25 Kai Tietz <ktietz@redhat.com>
16575 * config/i386/predicates.md (local_symbolic_operand): Interpret
16576 dll-imported symbols as none-local.
16578 2013-03-25 Richard Biener <rguenther@suse.de>
16580 * tree-ssa-loop-im.c (struct depend): Remove.
16581 (struct lim_aux_data): Make depends a vec of gimples.
16582 (free_lim_aux_data): Adjust.
16583 (add_dependency): Likewise.
16584 (set_level): Likewise.
16586 2013-03-25 Richard Biener <rguenther@suse.de>
16588 PR middle-end/56434
16589 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
16590 the pointer returned by calls with ECF_MALLOC set.
16592 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16594 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
16596 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16598 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
16599 using MMXMODE mode iterator.
16600 (*move<mode>_internal): Merge with *movv2sf_internal and
16601 *movv2sf_internal_rex64 using MMXMODE mode iterator.
16603 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16605 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
16606 (record_last_mem_set_info): Likewise.
16608 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
16609 of XNEWVEC followed by memset.
16610 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
16612 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16614 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
16615 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
16616 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
16617 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
16618 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
16619 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
16620 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16621 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16622 BARRIER_P instead of GET_CODE.
16624 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16626 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16627 inaccuracy in the probing code.
16629 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16630 (ctrapdi4): Likewise.
16632 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16634 * calls.c (expand_call): Add missing guard to code handling return
16635 of non-BLKmode structures in MSB.
16636 * function.c (expand_function_end): Likewise.
16638 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16640 * combine.c (try_combine): Adjust comment. Do not add the set of
16641 insn #0 if the destination indirectly is set or dies in insn #2.
16642 Tidy up code to distribute a new note.
16644 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16646 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16647 also for alternatives 16 and 17.
16649 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16651 * config/i386/sse.md (*mov<mode>_internal): Merge with
16652 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16653 Emit insn template depending on type attribute. Use
16654 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16655 movd instead of movq mnemonic for interunit moves. Rewrite mode
16656 attribute calculation. Remove unit attribute calculation.
16657 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16658 Set prefix_data16 attribute for DImode ssemov types.
16659 Use Ym instead of y for SSE-MMX conversion alternatives.
16660 Reorder operand constraints.
16662 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16664 * df.h (df_insn_delete): Adjust prototype.
16665 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16666 and let it decide whether mark the basic block dirty.
16667 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16668 * df-scan.c (df_insn_info_delete): New helper function, split
16669 off from df_insn_delete.
16670 (df_scan_free_bb_info): Use it.
16671 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16673 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16674 that the insn is actually an insn and it has a non-NULL basic block.
16675 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16677 2013-03-22 Richard Biener <rguenther@suse.de>
16679 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16681 (mem_ref_alloc): Do not allocate them.
16682 (refs_independent_p): Do not query or maintain a cache.
16684 2013-03-22 Richard Biener <rguenther@suse.de>
16686 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16687 (gather_mem_refs_in_loops): Do not compute it.
16688 (analyze_memory_references): Do not allocate it.
16689 (tree_ssa_lim_finalize): Do not free it.
16690 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16692 2013-03-22 Richard Biener <rguenther@suse.de>
16694 * is-a.h (as_a): Use gcc_checking_assert.
16696 2013-03-22 Ian Bolton <ian.bolton@arm.com>
16698 * config/aarch64/aarch64.c (aarch64_print_operand): New
16699 format specifier for printing a constant in hex.
16700 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16701 format specifier for printing second operand.
16703 2013-03-22 Richard Biener <rguenther@suse.de>
16705 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16707 (gather_mem_refs_in_loops): Perform store accumulation here.
16708 (create_vop_ref_mapping_loop): Remove.
16709 (create_vop_ref_mapping): Likewise.
16710 (analyze_memory_references): Initialize refs_stored_in_loop.
16711 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
16712 (record_indep_loop): Remove.
16713 (record_dep_loop): New function.
16714 (ref_indep_loop_p_1): Adjust to only walk over references
16715 in the loop, not its subloops.
16716 (ref_indep_loop_p): Rename to ...
16717 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16718 maintaining a more fine-grained cache.
16719 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16720 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16722 2013-03-22 Richard Biener <rguenther@suse.de>
16724 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16725 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16726 aggregate mem_ref_loc.
16727 (free_mem_ref_locs): Inline into ...
16728 (memref_free): ... this and adjust.
16729 (mem_ref_alloc): Adjust.
16730 (mem_ref_locs_alloc): Remove.
16731 (record_mem_ref_loc): Adjust.
16732 (get_all_locs_in_loop): Rewrite into ...
16733 (for_all_locs_in_loop): ... this iterator.
16734 (rewrite_mem_ref_loc): New functor.
16735 (rewrite_mem_refs): Use for_all_locs_in_loop.
16736 (sm_set_flag_if_changed): New functor.
16737 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16738 (ref_always_accessed): New functor.
16739 (ref_always_accessed_p): Use for_all_locs_in_loop.
16741 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16743 * tree-pass.h (PROP_gimple_lvec): New.
16744 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16745 (init_optimization_passes): Move pass_lower_vector.
16746 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16748 (pass_lower_vector): Provide PROP_gimple_lvec.
16749 (pass_lower_vector_ssa): Likewise.
16750 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16752 2013-03-21 Mark Wielaard <mjw@redhat.com>
16754 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16756 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16758 * config/i386/i386.md (*movdi_internal): Disparage slightly
16759 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16760 conversion alternatives.
16762 2013-03-21 Jakub Jelinek <jakub@redhat.com>
16764 PR middle-end/48087
16765 * diagnostic.def (DK_WERROR): New kind.
16766 * diagnostic.h (werrorcount): Define.
16767 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16768 promoted to DK_ERROR, increment DK_WERROR counter instead of
16770 * toplev.c (toplev_main): Call print_ignored_options even if
16771 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16772 even if just werrorcount is non-zero.
16775 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16777 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16778 (string_cst_pool_decl): New function.
16779 (optimize_one_addr_into_implicit_ptr): New function.
16780 (resolve_addr_in_expr): Optimize DWARF location expression
16781 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16782 which doesn't live in memory, but has DW_AT_location or
16783 DW_AT_const_value, or refers to a string literal, into
16784 DW_OP_GNU_implicit_pointer.
16785 (optimize_location_into_implicit_ptr): New function.
16786 (resolve_addr): If removing DW_AT_location of a variable because
16787 it was DW_OP_addr of address of the variable, but the variable doesn't
16788 live in memory, try to emit const value attribute for the initializer.
16790 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16792 * tree.h (VECTOR_TYPE_P): New macro.
16793 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16794 TYPE_MODE): Use it.
16795 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16796 VEC_COND_EXPR cannot be lvalues.
16797 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16799 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16801 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16802 Restrict the transformation to equal modes.
16804 2013-03-21 Richard Biener <rguenther@suse.de>
16806 PR tree-optimization/39326
16807 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16808 (MEM_ANALYZABLE): Adjust.
16809 (record_mem_ref_loc): Move bitmap ops ...
16810 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16811 unanalyzable refs, do not record locations for it.
16812 (analyze_memory_references): Allocate ref zero as shared
16814 (refs_independent_p): Do not test for unanalyzed mems here.
16815 (ref_indep_loop_p_1): Special-case disambiguation against
16816 the unanalyzed ref.
16817 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
16819 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
16821 * config/arm/arm-protos.h (tune_params): Add
16822 prefer_neon_for_64bits field.
16823 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
16824 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
16825 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
16826 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
16827 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
16828 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
16829 (arm_option_override): Handle -mneon-for-64bits new option.
16830 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
16831 (prefer_neon_for_64bits): Declare new variable.
16832 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
16833 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
16834 (arch_enabled): Handle new arch types. Remove support for onlya8
16836 (one_cmpldi2): Use new arch names.
16837 (zero_extend<mode>di2, extend<mode>di2): Ditto.
16838 * config/arm/arm.opt (mneon-for-64bits): Add option.
16839 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
16840 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
16841 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
16843 * doc/invoke.texi (-mneon-for-64bits): Document.
16845 2013-03-21 Richard Biener <rguenther@suse.de>
16847 PR tree-optimization/39326
16848 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
16849 (sort_bbs_in_loop_postorder_cmp): New function.
16850 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
16852 2013-03-21 Richard Biener <rguenther@suse.de>
16854 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
16855 (vect_insert_into_interleaving_chain): Likewise.
16856 (vect_drs_dependent_in_basic_block): Inline ...
16857 (vect_slp_analyze_data_ref_dependence): ... here. New function,
16859 (vect_analyze_data_ref_dependence): ... here. Simplify.
16860 (vect_check_interleaving): Simplify.
16861 (vect_analyze_data_ref_dependences): Likewise. Split out ...
16862 (vect_slp_analyze_data_ref_dependences): ... this new function.
16863 (dr_group_sort_cmp): New function.
16864 (vect_analyze_data_ref_accesses): Compute data-reference groups
16865 here instead of in vect_analyze_data_ref_dependence. Use
16866 a more efficient algorithm.
16867 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
16868 vect_slp_analyze_data_ref_dependences. Call
16869 vect_analyze_data_ref_accesses earlier.
16870 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
16871 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
16872 (vect_slp_analyze_data_ref_dependences): New prototype.
16874 2013-03-21 Richard Biener <rguenther@suse.de>
16876 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
16877 ref is stored in the loop.
16878 (find_refs_for_sm): Walk only over all stores.
16879 (store_motion_loop): Allocate from lim_bitmap_obstack.
16880 (store_motion): Likewise.
16882 2013-03-21 Richard Biener <rguenther@suse.de>
16884 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
16885 Update virtual SSA form.
16887 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16889 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
16890 * configure: Regenerate.
16891 * config.in: Regenerate.
16892 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
16893 if !HAVE_LD_EH_FRAME_CIEV3.
16895 2013-03-21 Richard Biener <rguenther@suse.de>
16897 * tree-cfg.c (verify_expr_no_block): New function.
16898 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
16899 nor DECL_VALUE_EXPR have locations with associated blocks.
16900 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
16901 (clear_unused_block_pointer): Remove code dealing with
16902 blocks in DECL_DEBUG_EXPR locations.
16904 2013-03-21 Richard Biener <rguenther@suse.de>
16906 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
16907 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
16908 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
16909 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
16910 instead of DECL_DEBUG_EXPR_IS_FROM.
16911 * gimplify.c (gimplify_modify_expr): Likewise.
16912 * tree-cfg.c (verify_expr_location_1): Likewise.
16913 * tree-complex.c (create_one_component_var): Likewise.
16914 * tree-sra.c (create_access_replacement): Likewise.
16915 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
16916 (clear_unused_block_pointer): Likewise.
16917 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
16918 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
16919 * var-tracking.c (var_debug_decl): Likewise.
16920 (track_expr_p): Likewise.
16921 * tree-inline.c (add_local_variables): Likewise. Set
16922 DECL_HAS_DEBUG_EXPR_P after copying it.
16923 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
16924 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
16926 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16929 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
16930 * configure: Regenerate.
16931 * config.in: Regenerate.
16932 * config/i386/i386.md (*movdf_internal): Use
16933 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16934 movd instead of movq mnemonic for interunit moves.
16935 (*movdi_internal): Ditto.
16937 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16939 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
16940 (abd<mode>_3): New pattern.
16941 (aba<mode>_3): New pattern.
16942 (fabd<mode>_3): New pattern.
16944 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16946 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
16947 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
16948 occurrence of REGISTER_PREFIX as its empty string.
16950 2013-03-20 Jeff Law <law@redhat.com>
16952 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
16953 addititional equivalences for equality comparisons between an SSA_NAME
16954 and a constant where the SSA_NAME was set from a widening conversion.
16956 2013-03-20 Walter Lee <walt@tilera.com>
16958 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
16960 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16962 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
16963 depending on type attribute.
16964 (*movti_internal): Ditto.
16965 (*movtf_internal): Ditto.
16966 (*movxf_internal): Ditto.
16967 (*movdf_internal): Ditto.
16968 (*movsf_internal): Ditto.
16970 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16972 * config/i386/i386.md (*movti_internal): Set prefix attribute to
16973 maybe_vex for sselog1 and ssemov types.
16974 (*movdi_internal): Reorder operand constraints.
16975 (*movsi_internal): Ditto. Set prefix attribute to
16976 maybe_vex for sselog1 and ssemov types.
16977 (*movtf_internal): Set prefix attribute to maybe_vex
16978 for sselog1 and ssemov types.
16979 (*movdf_internal): Ditto. Set prefix_data16 attribute for
16980 DImode ssemov types. Reorder operand constraints.
16981 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
16982 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
16983 attribute for SImode ssemov types. Reorder operand constraints.
16985 2013-03-20 Martin Jambor <mjambor@suse.cz>
16987 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
16988 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
16990 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
16992 * config/rs6000/predicates.md (indexed_address, update_address_mem
16993 update_indexed_address_mem): New predicates.
16994 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
16995 attribute for load/store instructions.
16996 * config/rs6000/dfp.md (movsd_store): Likewise.
16997 (movsd_load): Likewise.
16998 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
16999 (unnamed HI->DI extend define_insn): Likewise.
17000 (unnamed SI->DI extend define_insn): Likewise.
17001 (unnamed QI->SI extend define_insn): Likewise.
17002 (unnamed QI->HI extend define_insn): Likewise.
17003 (unnamed HI->SI extend define_insn): Likewise.
17004 (unnamed HI->SI extend define_insn): Likewise.
17005 (extendsfdf2_fpr): Likewise.
17006 (movsi_internal1): Likewise.
17007 (movsi_internal1_single): Likewise.
17008 (movhi_internal): Likewise.
17009 (movqi_internal): Likewise.
17010 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
17011 attribute for load/store instructions.
17012 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
17014 (mov<mode>_softfloat): Likewise.
17015 (mov<mode>_hardfloat32): Likewise.
17016 (mov<mode>_hardfloat64): Likewise.
17017 (mov<mode>_softfloat64): Likewise.
17018 (movdi_internal32): Likewise.
17019 (movdi_internal64): Likewise.
17020 (probe_stack_<mode>): Likewise.
17022 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
17024 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
17025 floating point, and decimal floating point to reload iterator.
17027 * config/rs6000/constraints.md (wl constraint): New constraints to
17028 return FLOAT_REGS if certain options are used to reduce the number
17029 of separate patterns that exist in the file.
17030 (wx constraint): Likewise.
17031 (wz constraint): Likewise.
17033 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17034 -mdebug=reg, print wg, wl, wx, and wz constraints.
17035 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17036 Initialize the reload functions for 64-bit binary/decimal floating
17038 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17039 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17040 create the buffer on the stack to overcome not having a 32-bit
17042 (rs6000_emit_move): Likewise.
17043 (rs6000_secondary_memory_needed_rtx): Likewise.
17044 (rs6000_alloc_sdmode_stack_slot): Likewise.
17045 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17046 via xxlxor, just like DFmode 0.0.
17048 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
17049 define as 1 if we are running on a power7 or newer.
17050 (enum r6000_reg_class_enum): Add new constraints.
17052 * config/rs6000/dfp.md (movsd): Delete, combine with binary
17053 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17054 with other moves by using conditional constraits (wg). Use LFIWZX
17055 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
17056 (movsd splitter): Likewise.
17057 (movsd_hardfloat): Likewise.
17058 (movsd_softfloat): Likewise.
17060 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
17061 binary and decimal floating point moves.
17062 (fmove_ok): New attributes to combine binary and decimal floating
17063 point moves, and to combine power6x (mfpgpr) moves along normal
17065 (real_value_to_target): Likewise.
17066 (f32_lr): Likewise.
17067 (f32_lm): Likewise.
17068 (f32_li): Likewise.
17069 (f32_sr): Likewise.
17070 (f32_sm): Likewise.
17071 (f32_si): Likewise.
17072 (movsf): Combine binary and decimal floating point moves. Combine
17073 power6x (mfpgpr) moves with other moves by using conditional
17074 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
17075 (mov<mode> for SFmode/SDmode); Likewise.
17076 (SFmode/SDmode splitters): Likewise.
17077 (movsf_hardfloat): Likewise.
17078 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
17079 (movsf_softfloat): Likewise.
17080 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
17082 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
17083 wx and wz constraints.
17085 * config/rs6000/constraints.md (wg constraint): New constraint to
17086 return FLOAT_REGS if -mmfpgpr (power6x) was used.
17088 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
17091 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17092 -mdebug=reg, print wg, wl, wx, and wz constraints.
17093 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17094 Initialize the reload functions for 64-bit binary/decimal floating
17096 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17097 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17098 create the buffer on the stack to overcome not having a 32-bit
17100 (rs6000_emit_move): Likewise.
17101 (rs6000_secondary_memory_needed_rtx): Likewise.
17102 (rs6000_alloc_sdmode_stack_slot): Likewise.
17103 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17104 via xxlxor, just like DFmode 0.0.
17106 * config/rs6000/dfp.md (movdd): Delete, combine with binary
17107 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17108 with other moves by using conditional constraits (wg). Use LFIWZX
17109 and STFIWX for loading SDmode on power7.
17110 (movdd splitters): Likewise.
17111 (movdd_hardfloat32): Likewise.
17112 (movdd_softfloat32): Likewise.
17113 (movdd_hardfloat64_mfpgpr): Likewise.
17114 (movdd_hardfloat64): Likewise.
17115 (movdd_softfloat64): Likewise.
17117 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
17118 64-bit binary and decimal floating point moves.
17119 (FMOVE64X): Likewise.
17120 (movdf): Combine 64-bit binary and decimal floating point moves.
17121 Combine power6x (mfpgpr) moves with other moves by using
17122 conditional constraits (wg).
17123 (mov<mode> for DFmode/DDmode): Likewise.
17124 (DFmode/DDmode splitters): Likewise.
17125 (movdf_hardfloat32): Likewise.
17126 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
17127 (movdf_softfloat32): Likewise.
17128 (movdf_hardfloat64_mfpgpr): Likewise.
17129 (movdf_hardfloat64): Likewise.
17130 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
17131 (movdf_softfloat64): Likewise.
17132 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
17133 (reload_<mode>_load): Move to later in the file so they aren't in
17134 the middle of the floating point move insns.
17135 (reload_<mode>_store): Likewise.
17137 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
17140 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
17141 constraint if -mdebug=reg.
17142 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
17143 Enable using dd reload support if needed.
17145 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
17146 binary and decimal floating point moves in rs6000.md.
17147 (movtd_internal): Likewise.
17149 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
17150 decimal floating point moves.
17152 (movtf_internal): Likewise.
17153 (mov<mode>_internal, TDmode/TFmode): Likewise.
17154 (movtf_softfloat): Likewise.
17155 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
17157 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
17158 movdi_internal64, using wg constraint for move direct operations.
17159 (movdi_internal64): Likewise.
17161 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
17162 MODES_TIEABLE_P for selected modes. Print the numerical value of
17163 the various virtual registers. Use GPR/FPR first/last values,
17164 instead of hard coding the register numbers. Print which modes
17165 have reload functions registered.
17166 (rs6000_option_override_internal): If -mdebug=reg, trace the options
17167 settings before/after setting cpu, target and subtarget settings.
17168 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
17169 and for secondary reload failures in rs6000_secondary_reload_inner.
17170 (rs6000_secondary_reload_fail): Likewise.
17171 (rs6000_secondary_reload_inner): Likewise.
17173 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
17174 macros for first/last GPR and FPR registers.
17175 (LAST_GPR_REGNO): Likewise.
17176 (FIRST_FPR_REGNO): Likewise.
17177 (LAST_FPR_REGNO): Likewise.
17179 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
17180 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
17181 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
17182 (vcond<mode><mode>): Likewise.
17183 (vcondu<mode><mode>): Likewise.
17184 (vector_gtu<mode>): Likewise.
17185 (vector_gte<mode>): Likewise.
17186 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
17187 to prevent the compiler from converting DImode operations to TImode.
17188 (ior<mode>3): Likewise.
17189 (and<mode>3): Likewise.
17190 (one_cmpl<mode>2): Likewise.
17191 (nor<mode>3): Likewise.
17192 (andc<mode>3): Likewise.
17194 * config/rs6000/constraints.md (wt constraint): New constraint
17195 that returns VSX_REGS if TImode is allowed in VSX registers.
17197 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
17198 constant under VSX.
17200 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
17201 similar to TImode, but it is restricted to being in the GPRs.
17203 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
17204 TImode to occupy a single VSX register.
17206 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
17207 -mvsx-timode for power7/power8.
17208 (power7 cpu): Likewise.
17209 (power8 cpu): Likewise.
17211 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
17212 sure that TFmode/TDmode take up two registers if they are ever
17213 allowed in the upper VSX registers.
17214 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
17216 (rs6000_init_hard_regno_mode_ok): Likewise.
17217 (rs6000_debug_reg_global): Add debugging for PTImode and wt
17218 constraint. Print if LRA is turned on.
17219 (rs6000_option_override_internal): Give an error if -mvsx-timode
17220 and VSX is not enabled.
17221 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
17222 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
17223 to reg+offset addressing. Use PTImode when checking offset
17224 addresses for validity.
17225 (reg_offset_addressing_ok_p): Likewise.
17226 (rs6000_legitimate_offset_address_p): Likewise.
17227 (rs6000_legitimize_address): Likewise.
17228 (rs6000_legitimize_reload_address): Likewise.
17229 (rs6000_legitimate_address_p): Likewise.
17230 (rs6000_eliminate_indexed_memrefs): Likewise.
17231 (rs6000_emit_move): Likewise.
17232 (rs6000_secondary_reload): Likewise.
17233 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
17234 reloads to fpr registers to continue to use reg+offset addressing,
17235 but 64-bit reloads to altivec registers need reg+reg addressing.
17236 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
17237 it. Treat LO_SUM like a PLUS operation.
17238 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
17239 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
17240 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
17241 registers to share a register with a smaller sized type, since VSX
17242 puts scalars in the upper 64-bits.
17243 (print_operand): Add support for PTImode.
17244 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
17245 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
17246 registers, but don't have arithmetic support.
17247 (rs6000_memory_move_cost): Add test for VSX.
17248 (rs6000_opt_masks): Add -mvsx-timode.
17250 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
17253 (VSr): Use wt constraint for TImode.
17254 (VSv): Drop TImode support.
17255 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
17256 (vsx_movti_64bit): Likewise.
17257 (vsx_movti_32bit): Likewise.
17258 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
17259 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
17260 one '?' on the appropriate output constraint. Do not allow TImode
17261 logical operations on 32-bit systems.
17262 (vsx_ior<mode>3): Likewise.
17263 (vsx_xor<mode>3): Likewise.
17264 (vsx_one_cmpl<mode>2): Likewise.
17265 (vsx_nor<mode>3): Likewise.
17266 (vsx_andc<mode>3): Likewise.
17267 (vsx_concat_<mode>): Likewise.
17268 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
17270 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
17271 OPTION_MASK_VSX_TIMODE.
17272 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
17273 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
17275 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
17276 (TI2 iterator): New iterator for TImode, PTImode.
17277 (wd mode attribute): Add values for vector types.
17278 (movti_string): Replace TI move operations with operations for TImode
17279 and PTImode. Add support for TImode being allowed in VSX registers.
17280 (mov<mode>_string, TImode/PTImode): Likewise.
17281 (movti_ppc64): Likewise.
17282 (mov<mode>_ppc64, TImode/PTImode): Likewise.
17283 (TI mode splitters): Likewise.
17285 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
17288 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
17290 PR tree-optimization/56355
17291 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
17292 Also handle integers with undefined overflow.
17294 2013-03-20 Catherine Moore <clm@codesourcery.com>
17295 Maciej W. Rozycki <macro@codesourcery.com>
17296 Tom de Vries <tom@codesourcery.com>
17297 Nathan Sidwell <nathan@codesourcery.com>
17298 Iain Sandoe <iain@codesourcery.com>
17299 Nathan Froyd <froydnj@codesourcery.com>
17300 Chao-ying Fu <fu@mips.com>
17302 * doc/extend.texi: (micromips, nomicromips, nocompression):
17303 Document new function attributes.
17304 * doc/invoke.texi (minterlink-compressed, mmicromips,
17305 m14k, m14ke, m14kec): Document new options.
17306 (minterlink-mips16): Update documentation.
17307 * doc/md.texi (ZC, ZD): Document new constraints.
17308 * configure.ac (gcc_cv_as_micromips): Check if linker
17309 supports the .set micromips directive.
17310 * configure: Regenerate.
17311 * config.in: Regenerate.
17312 * config/mips/mips-tables.opt: Regenerate.
17313 * config/mips/micromips.md: New file.
17314 * constraints.md (ZC, ZD): New constraints.
17315 * config/mips/predicates.md (movep_src_register): New predicate.
17316 (movep_src_operand): New predicate.
17317 (non_volatile_mem_operand): New predicate.
17318 * config/mips/mips.md (multimem): New type.
17319 (length): Differentiate between 17-bit and 18-bit branch offsets.
17320 (MOVEP1, MOVEP2): New mode iterator.
17321 (mov_<load>l): Use ZC constraint.
17322 (mov_<load>r): Likewise.
17323 (mov_<store>l): Likewise.
17324 (mov_<store>r): Likewise.
17325 (*branch_equality<mode>_inverted): Add microMIPS support.
17326 (*branch_equality<mode>): Likewise.
17327 (*jump_absolute): Likewise.
17328 (indirect_jump_<mode>): Likewise.
17329 (tablejump_<mode>): Likewise.
17330 (<optab>_internal): Likewise.
17331 (sibcall_internal): Likewise.
17332 (sibcall_value_internal): Likewise.
17333 (prefetch): Use constraint ZD.
17334 * config/mips/mips.opt (minterlink-compressed): New option.
17335 (minterlink-mips16): Now an alias for minterlink-compressed.
17336 (mmicromips): New option.
17337 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17338 (compare_and_swap_12): Likewise.
17339 (sync_add<mode>): Likewise.
17340 (sync_<optab>_12): Likewise.
17341 (sync_old_<optab>_12): Likewise.
17342 (sync_new_<optab>_12): Likewise.
17343 (sync_nand_12): Likewise.
17344 (sync_old_nand_12): Likewise.
17345 (sync_new_nand_12): Likewise.
17346 (sync_sub<mode>): Likewise.
17347 (sync_old_add<mode>): Likewise.
17348 (sync_old_sub<mode>): Likewise.
17349 (sync_new_add<mode>): Likewise.
17350 (sync_new_sub<mode>): Likewise.
17351 (sync_<optab><mode>): Likewise.
17352 (sync_old_<optab><mode>): Likewise.
17353 (sync_new_<optab><mode>): Likewise.
17354 (sync_nand<mode>): Likewise.
17355 (sync_old_nand<mode>): Likewise.
17356 (sync_new_nand<mode>): Likewise.
17357 (sync_lock_test_and_set<mode>): Likewise.
17358 (test_and_set_12): Likewise.
17359 (atomic_compare_and_swap<mode>): Likewise.
17360 (atomic_exchange<mode>_llsc): Likewise.
17361 (atomic_fetch_add<mode>_llsc): Likewise.
17362 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17363 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17364 (umips_save_restore_pattern_p): Likewise.
17365 (umips_load_store_pair_p): Likewise.
17366 (umips_output_load_store_pair): Likewise.
17367 (umips_movep_target_p): Likewise.
17368 (umips_12bit_offset_address_p): Likewise.
17369 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17370 (mips_base_mips16): Rename this...
17371 (mips_base_compression_flags): ...to this. Update all uses.
17372 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17373 (mips_mips16_decl_p): Delete.
17374 (mips_nomips16_decl_p): Delete.
17375 (mips_get_compress_on_flags): New function.
17376 (mips_get_compress_off_flags): New function.
17377 (mips_get_compress_mode): New function.
17378 (mips_get_compress_on_name): New function.
17379 (mips_get_compress_off_name): New function.
17380 (mips_insert_attributes): Support multiple compression types.
17381 (mips_merge_decl_attributes): Likewise.
17382 (umips_12bit_offset_address_p): New function.
17383 (mips_start_function_definition): Emit .set micromips directive.
17384 (mips_call_may_need_jalx_p): New function.
17385 (mips_function_ok_for_sibcall): Add microMIPS support.
17386 (mips_print_operand_punctuation): Support short delay slots and
17388 (umips_swm_mask, umips_swm_encoding): New.
17389 (umips_build_save_restore): New function.
17390 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17391 (was_mips16_p): Remove.
17392 (old_compression_mode): New.
17393 (mips_set_compression_mode): New function.
17394 (mips_set_current_function): Add microMIPS support.
17395 (mips_option_override): Likewise.
17396 (umips_save_restore_pattern_p): New function.
17397 (umips_output_save_restore): New function.
17398 (umips_load_store_pair_p_1): New function.
17399 (umips_load_store_pair_p): New function.
17400 (umips_output_load_store_pair_1): New function.
17401 (umips_output_load_store_pair): New function.
17402 (umips_movep_target_p) New function.
17403 (mips_prepare_pch_save): Add microMIPS support.
17404 * config/mips/mips.h (TARGET_COMPRESSION): New.
17405 (TARGET_CPU_CPP_BUILTINS): Update macro
17406 to use new compression flags and to support microMIPS.
17407 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17408 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17409 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17410 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17411 (ASM_SPEC): Support mmicromips and mno-micromips.
17412 (M16STORE_REG_P): New macro.
17413 (MIPS_CALL): Support TARGET_MICROMIPS.
17414 (MICROMIPS_J): New macro.
17415 (mips_base_mips16): Rename this...
17416 (mips_base_compression_flags): ...to this.
17417 (UMIPS_12BIT_OFFSET_P): New macro.
17418 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17419 (MULTILIB_DIRNAMES): Likewise.
17420 2013-03-20 Richard Biener <rguenther@suse.de>
17422 PR tree-optimization/56661
17423 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
17424 the result does not have to be distinct.
17426 2013-03-20 Richard Biener <rguenther@suse.de>
17428 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
17431 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17432 Steven Bosscher <steven@gcc.gnu.org>
17434 PR rtl-optimization/56605
17435 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
17437 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17440 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
17441 that require movd instead of movq.
17443 2013-03-20 Richard Biener <rguenther@suse.de>
17445 * tree-ssa-structalias.c (struct variable_info): Add pointer
17446 to the first field of an aggregate with sub-vars. Make
17447 this and the pointer to the next subfield its ID.
17448 (vi_next): New function.
17449 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
17450 storedanything_id, integer_id): Increment by one.
17451 (new_var_info, get_call_vi, lookup_call_clobber_vi,
17452 get_call_clobber_vi): Adjust.
17453 (solution_set_expand): Simplify and speedup.
17454 (solution_set_add): Inline into ...
17455 (set_union_with_increment): ... this. Adjust accordingly.
17456 (do_sd_constraint): Likewise.
17457 (do_ds_constraint): Likewise.
17458 (do_complex_constraint): Simplify.
17459 (build_pred_graph): Adjust.
17460 (solve_graph): Likewise. Simplify and speedup.
17461 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
17462 get_constraint_for_component_ref, get_constraint_for_1,
17463 first_vi_for_offset, first_or_preceding_vi_for_offset,
17464 create_function_info_for, create_variable_info_for_1,
17465 create_variable_info_for, intra_create_variable_infos): Adjust.
17466 (init_base_vars): Push NULL for ID zero.
17467 (compute_points_to_sets): Adjust.
17469 2013-03-20 Richard Biener <rguenther@suse.de>
17471 * cfgloop.c (verify_loop_structure): Streamline and avoid
17472 ICEing on corrupt loop tree.
17473 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17476 2013-03-20 Richard Biener <rguenther@suse.de>
17478 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17479 check whether an SSA update is needed.
17481 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17483 * config/mips/constraints.md (T): Rename to...
17487 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17488 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17490 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17492 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17493 (*subsi3_carryin_uxtw): Likewise.
17495 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17497 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17498 (*rorsi3_insn_uxtw): Likewise.
17500 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17502 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17503 (*extrsi5_insn_uxtw): Likewise.
17505 2013-03-19 Richard Biener <rguenther@suse.de>
17507 PR tree-optimization/56273
17508 * passes.c (init_optimization_passes): Move second VRP after DOM.
17510 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
17512 * config/i386/i386.md (*movti_internal): Merge from
17513 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17514 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17515 nox64 isa attributes.
17517 2013-03-18 Richard Biener <rguenther@suse.de>
17519 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17521 (merge_node_constraints): Likewise.
17522 (build_succ_graph): Likewise.
17523 (valid_graph_edge): Inline into single caller.
17524 (unify_nodes): Likewise. Use bitmap_set_bit return value
17526 (scc_visit): Fix formatting and variable use.
17527 (do_sd_constraint): Use gcc_checking_assert.
17528 (do_ds_constraint): Likewise.
17529 (do_complex_constraint): Likewise.
17530 (condense_visit): Likewise. Cleanup.
17531 (dump_pred_graph): New function.
17532 (perform_var_substitution): Dump the pred-graph before
17533 variable substitution.
17534 (find_equivalent_node): Use gcc_checking_assert.
17535 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17537 2013-03-18 Richard Biener <rguenther@suse.de>
17539 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17540 Remove cond_expr_stmt_list argument and do not gimplify the
17542 (vect_loop_versioning): Adjust.
17543 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17544 Cleanup to use less temporaries.
17545 (vect_create_data_ref_ptr): Cleanup.
17547 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17549 PR tree-optimization/56635
17550 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17551 require types_compatible_p types.
17553 2013-03-18 Nick Clifton <nickc@redhat.com>
17555 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
17556 spurious backslash.
17558 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
17559 Add missing line to comment describing function.
17561 2013-03-18 Richard Biener <rguenther@suse.de>
17563 PR tree-optimization/56210
17564 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17565 Handle string / character search functions.
17566 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
17568 2013-03-18 Richard Biener <rguenther@suse.de>
17570 PR middle-end/56483
17571 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
17572 and implement properly.
17573 * gimple.h (gimple_cond_single_var_p): Remove.
17575 2013-03-18 Richard Biener <rguenther@suse.de>
17577 * tree-data-ref.h (find_data_references_in_loop): Declare.
17578 * tree-data-ref.c (get_references_in_stmt): Use a stack
17579 vector pre-allocated in the callers.
17580 (find_data_references_in_stmt): Adjust.
17581 (graphite_find_data_references_in_stmt): Likewise.
17582 (create_rdg_vertices): Likewise.
17583 (find_data_references_in_loop): Export.
17584 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
17585 Compute dependences here...
17586 (vect_analyze_data_refs): ...not here. When we encounter
17587 a non-vectorizable data reference in basic-block vectorization
17588 truncate the data reference vector. Do not bother to
17589 fixup data-dependence information for gather loads.
17590 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
17591 of data references, as reported.
17593 2013-03-18 Richard Biener <rguenther@suse.de>
17595 PR tree-optimization/3713
17596 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
17597 has_constants and expr.
17598 (stmt_has_constants): Properly valueize SSA names when deciding
17599 whether the stmt has constants.
17601 2013-03-18 Richard Biener <rguenther@suse.de>
17603 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
17604 whole function when there is nothing to do.
17605 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
17606 * tree-vectorizer.c (vectorize_loops): Update virtual and
17607 loop-closed SSA once.
17608 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
17610 2013-03-18 Richard Biener <rguenther@suse.de>
17612 PR middle-end/56113
17613 * domwalk.c (bb_postorder): New global static.
17614 (cmp_bb_postorder): New function.
17615 (walk_dominator_tree): Replace scheme imposing an order for
17616 visiting dominator sons by one sorting them at the time they
17617 are pushed on the stack.
17619 2013-03-18 Richard Biener <rguenther@suse.de>
17621 PR tree-optimization/39326
17622 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17623 (struct mem_ref): Replace mem member with ao_ref typed member.
17624 (MEM_ANALYZABLE): Adjust.
17625 (memref_eq): Likewise.
17626 (mem_ref_alloc): Likewise.
17627 (gather_mem_refs_stmt): Likewise.
17628 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17629 (execute_sm_if_changed_flag_set): Adjust.
17630 (execute_sm): Likewise.
17631 (ref_always_accessed_p): Likewise.
17632 (refs_independent_p): Likewise.
17633 (can_sm_ref_p): Likewise.
17635 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17638 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17639 return 1 even for !unsignedp.
17641 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17643 * config/i386/i386.md (isa): Add x64 and nox64.
17644 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17645 (*pushtf): Enable *roF alternative for x64 isa only.
17646 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17647 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17648 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17649 (*movtf_internal): Merge from *movtf_internal_rex64 and
17650 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17651 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17652 nox64 isa attributes.
17653 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17654 nox64 isa attributes.
17655 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17657 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17659 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17661 2013-03-17 Jakub Jelinek <jakub@redhat.com>
17664 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17665 class names. Remove trailing comma after "ALL_REGS".
17667 2013-03-16 Jan Hubicka <jh@suse.cz>
17669 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17670 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17671 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17672 of cgraph_get_create_node.
17673 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17675 2013-03-16 Jason Merrill <jason@redhat.com>
17678 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17679 with DW_AT_default_value.
17681 2013-03-16 Jakub Jelinek <jakub@redhat.com>
17683 * BASE-VER: Set to 4.9.0.
17685 2013-03-14 Andi Kleen <ak@linux.intel.com>
17688 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17689 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17690 Document _x* TSX intrinsics.
17692 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17693 David Holsgrove <david.holsgrove@xilinx.com>
17695 * configure.ac: Add MicroBlaze TLS support detection.
17696 * configure: Regenerate.
17697 * config/microblaze/microblaze-protos.h
17698 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17699 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17700 * config/microblaze/microblaze.c (microblaze_address_type): Add
17701 ADDRESS_TLS and tls_reloc address types.
17702 (microblaze_address_info): Add tls_reloc.
17703 (TARGET_HAVE_TLS): Define.
17704 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17705 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
17706 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17707 load_tls_operand, microblaze_call_tls_get_addr,
17708 microblaze_legitimize_tls_address): New functions.
17709 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17710 (get_base_reg): Use microblaze_tls_symbol_p.
17711 (microblaze_classify_address): Handle TLS.
17712 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17713 label_mentioned_p and microblaze_tls_referenced_p.
17714 (microblaze_legitimize_address): Handle TLS.
17715 (microblaze_address_insns): Handle ADDRESS_TLS.
17716 (pic_address_needs_scratch): Handle TLS.
17717 (print_operand_address): Handle TLS.
17718 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17719 (microblaze_expand_move): Handle TLS.
17720 (microblaze_legitimate_constant_p): Check
17721 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
17722 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17723 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17724 (PIC_OFFSET_TABLE_REGNUM): Set.
17725 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17726 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17727 (addsi3, movsi_internal2, movdf_internal): Update constraints
17728 * config/microblaze/predicates.md (arith_plus_operand): Define
17729 (move_operand): Redefine as move_src_operand,
17730 check microblaze_tls_referenced_p.
17732 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17734 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17735 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17737 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17739 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17742 2013-03-14 Jakub Jelinek <jakub@redhat.com>
17744 PR tree-optimization/53265
17745 * common.opt (Waggressive-loop-optimizations): New option.
17746 * tree-ssa-loop-niter.c: Include tree-pass.h.
17747 (do_warn_aggressive_loop_optimizations): New function.
17748 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17749 if number_of_latch_executions returned constant.
17750 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17751 early. If number_of_latch_executions returned constant, set
17752 nb_iterations_upper_bound back to it.
17753 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17755 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17756 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17758 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17759 (MULTILIB_OSDIRNAMES): Set.
17760 * genmultilib: If defaultosdirname doesn't start with :: , set
17761 defaultosdirname2 instead, clear it and emit two . multilib_raw
17762 entries instead of just one.
17764 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17766 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17767 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17768 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17769 (SUBTARGET_OVERRIDE_OPTIONS): New.
17771 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17774 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17775 (musermode): Convert to Var(TARGET_USERMODE).
17776 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17777 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17778 * config/sh/sh.c (sh_option_override): Use
17779 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17780 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17782 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17784 (udivsi3_i4_single, divsi3_i4_single): Use
17785 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17787 2013-03-13 Dave Korn <dave.korn.cygwin@....>
17789 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17792 2013-03-13 Richard Biener <rguenther@suse.de>
17794 PR tree-optimization/56608
17795 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17796 calls when vectorizing basic-blocks.
17798 2013-03-13 Jakub Jelinek <jakub@redhat.com>
17801 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17804 2013-03-12 Jakub Jelinek <jakub@redhat.com>
17806 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17808 2013-03-11 Jan Hubicka <jh@suse.cz>
17811 * lto-streamer-out.c (output_symbol_p): Skip references from
17812 constructors of external variables.
17814 2013-03-11 Jan Hubicka <jh@suse.cz>
17816 PR middle-end/56571
17817 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
17819 * emit-rtl.c (verify_rtx_sharing): Likewise.
17820 (copy_insn_1): Likewise.
17821 * rtl.c (copy_rtx): Likewise.
17823 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
17826 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
17827 output_operand_lossage message.
17829 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
17832 * arm.c (shift_op): Validate RTL pattern on the fly.
17833 (arm_print_operand, case 'S'): Don't use shift_operator to validate
17836 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17839 * config/pa/pa.md (call_value): Check for calls to powf and direct to
17840 new call patterns that clobber %fr12.
17841 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
17842 split and postreload patterns.
17843 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
17844 registers %fr12 and %fr12R as call used.
17846 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
17848 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
17849 (canon_address, record_store, replace_read, check_mem_read_rtx,
17850 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
17851 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
17852 rest_of_handle_dse): Likewise.
17854 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
17856 PR middle-end/56524
17857 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
17859 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
17860 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
17861 (save_optabs_if_changed): Replace with...
17862 (init_tree_optimization_optabs): ...this.
17863 * optabs.c (save_optabs_if_changed): Rename to...
17864 (init_tree_optimization_optabs): ...this. Take the optimization node
17865 as argument. Do nothing if the base optabs are already correct.
17866 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
17867 to recompute optabs.
17868 * function.h (function): Remove optabs field.
17869 * function.c (invoke_set_current_function_hook): Call
17870 init_tree_optimization_optabs. Use the result to initialize
17873 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
17875 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
17876 if GTMA_HAS_NO_INSTRUMENTATION.
17877 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
17878 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
17879 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
17880 * gimple-pretty-print.c (dump_gimple_transaction): Handle
17881 GTMA_HAS_NO_INSTRUMENTATION.
17883 2013-03-08 Jakub Jelinek <jakub@redhat.com>
17885 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
17888 2013-03-08 Marek Polacek <polacek@redhat.com>
17889 Jakub Jelinek <jakub@redhat.com>
17891 PR tree-optimization/56478
17892 * predict.c (is_comparison_with_loop_invariant_p): Change the
17893 type of loop_step to tree.
17894 (predict_loops): Adjust.
17895 (predict_iv_comparison): Perform the computations on double_ints.
17897 2013-03-08 Richard Biener <rguenther@suse.de>
17899 PR tree-optimization/56570
17900 * tree-cfg.c (verify_expr_location_1): Verify locations for
17902 * tree-sra.c (create_access_replacement): Strip locations
17903 from DECL_DEBUG_EXPRs.
17905 2013-03-08 Richard Biener <rguenther@suse.de>
17907 * tree-inline.c (expand_call_inline): Do not associate
17908 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
17909 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
17911 2013-03-08 Richard Biener <rguenther@suse.de>
17913 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
17914 or block changes with -Og. Fix for location / block encoding
17915 changes and PHI arguments with locations.
17917 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
17919 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
17921 (struct output_info): Likewise.
17922 (register_overhead): Remove bad gcc_assert.
17923 (bitmap_find_bit): If there is only a single bitmap element, do not
17924 count a miss as a search.
17925 (print_statistics): Update for counter type changes.
17926 (dump_bitmap_statistics): Likewise. Print headers such that they
17927 are properly lined up with the printed counters.
17929 2013-03-07 Jakub Jelinek <jakub@redhat.com>
17931 PR tree-optimization/56559
17932 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
17933 check that it has only a single use.
17935 2013-03-07 Richard Biener <rguenther@suse.de>
17937 * doc/invoke.texi (fwhole-program): Discourage use in combination
17940 2013-03-06 Jakub Jelinek <jakub@redhat.com>
17942 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
17944 PR tree-optimization/56539
17945 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
17946 instead of GSI_CONTINUE_LINKING as last argument to
17947 force_gimple_operand_gsi. Adjust function comment.
17949 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
17952 PR middle-end/56548
17953 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
17954 promoted mode, convert the result back to the original mode.
17956 2013-03-06 Richard Biener <rguenther@suse.de>
17958 PR middle-end/56294
17959 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
17960 (insert_updated_phi_nodes_compare_uids): New function.
17961 (update_ssa): Sort symbols_to_rename after UID before
17962 traversing it to insert PHI nodes.
17964 2013-03-06 Richard Biener <rguenther@suse.de>
17966 PR middle-end/50494
17967 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
17968 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
17971 2013-02-13 Richard Biener <rguenther@suse.de>
17974 * varasm.c (output_constant_def_1): Get the decl representing
17975 the constant as argument.
17976 (output_constant_def): Wrap output_constant_def_1.
17977 (make_decl_rtl): Use output_constant_def_1 with the decl
17978 representing the constant.
17979 (build_constant_desc): Optionally re-use a decl already
17980 representing the constant.
17981 (tree_output_constant_def): Adjust.
17983 2013-03-06 Joey Ye <joey.ye@arm.com>
17986 * gcc.c (convert_white_space): New function.
17987 (main): Handles white space in function name.
17989 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
17992 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
17993 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
17994 to SH_DIV_CALL_TABLE for TARGET_SH2.
17995 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
17997 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
17998 call-table options.
18000 2013-03-05 Sterling Augustine <saugustine@google.com>
18001 Cary Coutant <ccoutant@google.com>
18004 * dwarf2out.c (resolve_addr): Don't call
18005 remove_loc_list_addr_table_entries a second time for the same
18008 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18011 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
18012 (avoid_complex_debug_insns): New function.
18013 (expand_debug_locations): Call it.
18015 PR rtl-optimization/56484
18016 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
18017 lifetimes of hard registers on small register class machines.
18019 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18021 * config/microblaze/microblaze-protos.h: Rename
18022 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
18023 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
18025 (microblaze_fast_interrupt_function_p): New function.
18026 (microblaze_is_interrupt_handler): Rename to
18027 microblaze_is_interrupt_variant and add fast_interrupt check.
18028 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
18029 (save_restore_insns): Likewise.
18030 (compute_frame_size): Likewise.
18031 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
18032 (microblaze_globalize_label): Likewise.
18033 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
18034 * config/microblaze/microblaze.md: Use wrapper
18035 microblaze_is_interrupt_variant.
18037 2013-03-05 Kai Tietz <ktietz@redhat.com>
18039 * sdbout.c (sdbout_one_type): Switch to current function's section
18040 supporting cold/hot.
18042 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18044 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
18047 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18049 PR middle-end/56461
18050 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
18051 if VALGRIND_GET_VBITS is defined, temporarily make object
18052 memory all defined, and restore previous valgrind addressability
18053 and definability afterwards. Free this_object at the end.
18055 PR middle-end/56461
18056 * lra.c (lra): Call lra_clear_live_ranges if live_p,
18057 right before calling lra_create_live_ranges, also call it
18058 when clearing live_p. Only call lra_clear_live_ranges
18059 at the end if live_p.
18061 PR middle-end/56461
18062 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
18064 2013-03-05 Richard Biener <rguenther@suse.de>
18066 PR tree-optimization/56521
18067 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
18070 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
18073 * except.h (remove_unreachable_eh_regions): New prototype.
18074 * except.c (remove_eh_handler_splicer): New function, split out
18075 of remove_eh_handler.
18076 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
18077 warning about running it on many EH regions one at a time.
18078 (remove_unreachable_eh_regions_worker): New function, walk the
18079 EH tree in depth-first order and remove non-marked regions.
18080 (remove_unreachable_eh_regions): New function.
18081 * tree-eh.c (mark_reachable_handlers): New function, split out
18082 from remove_unreachable_handlers.
18083 (remove_unreachable_handlers): Use mark_reachable_handlers and
18084 remove_unreachable_eh_regions.
18085 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
18086 and remove_unreachable_eh_regions.
18088 2013-03-05 Richard Biener <rguenther@suse.de>
18090 PR middle-end/56525
18091 * loop-init.c (fix_loop_structure): Remove loops in two stages,
18092 not freeing them until the end.
18094 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18096 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
18098 2013-03-05 Richard Biener <rguenther@suse.de>
18100 PR tree-optimization/56270
18101 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
18102 of loads after scheduling an SLP instance.
18104 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18106 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
18108 (check_gcc_parallelize): Run guality.exp as a separate job from
18109 vect.exp with unsorted.exp and $(dg_target_exps) separately from
18110 struct-layout-1.exp with stackalign.exp.
18112 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
18114 PR middle-end/56461
18115 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
18116 load_index sbitmap even if some bit in it isn't set.
18118 PR middle-end/56461
18119 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
18120 (discover_iteration_bound_by_body_walk): Change queues to
18121 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
18122 spelling in comment. Call safe_push on queues[bound_index] directly.
18123 Release queues[queue_index] in every iteration unconditionally.
18124 Release bounds vector.
18126 PR middle-end/56461
18127 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
18128 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
18129 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
18132 2013-03-05 Richard Biener <rguenther@suse.de>
18135 * tree-inline.c (remap_blocks_to_null): New function.
18136 (expand_call_inline): When expanding a call stmt without
18137 an associated block inline remap all callee blocks to NULL.
18139 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18141 PR rtl-optimization/56494
18142 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
18143 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
18144 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
18146 PR middle-end/56461
18147 * sel-sched-ir.c (free_sched_pools): Release
18148 succs_info_pool.stack[succs_info_pool.max_top] vectors too
18149 if succs_info_pool.max_top isn't -1.
18152 * opts.c (opts_obstack, opts_concat): Moved to...
18153 * opts-common.c (opts_obstack, opts_concat): ... here.
18155 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18157 PR middle-end/56461
18158 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
18160 2013-03-04 Martin Jambor <mjambor@suse.cz>
18162 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
18163 all appropriate places.
18165 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
18167 PR tree-optimization/56424
18168 * ipa-split.c (split_function): Do not set the RSO flag if result is
18169 not by reference and its type is a register type.
18171 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
18173 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
18174 (microblaze_legitimate_pic_operand): Likewise
18175 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
18176 new function microblaze_legitimate_pic_operand
18177 * config/microblaze/microblaze-protos.h
18178 (microblaze_legitimate_pic_operand): Declare.
18180 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18182 * config/microblaze/predicates.md (call_insn_simple_operand):
18183 New predicate for supported rtx code types.
18184 * config/microblaze/microblaze.md (call_internal1): Use
18185 call_insn_simple_operand predicate.
18187 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18189 PR middle-end/56461
18190 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
18191 partitions.ordered_remove.
18193 PR middle-end/56461
18194 * tree-vect-stmts.c (vectorizable_conversion): Don't call
18195 vec_oprnds0.create (1) for modifier == NONE.
18197 PR middle-end/56461
18198 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
18199 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
18200 vec_oprnds1 right before pushing anything to it for
18203 PR middle-end/56461
18204 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
18205 set nbbs to 0 instead of having separate code path.
18206 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
18207 instead of false as last argument if returning NULL.
18209 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
18211 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
18212 the attribute is now called "target" instead of "option".
18213 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
18214 * doc/tm.texi.in (Target Attributes): Likewise document the correct
18215 attribute/pragma name for TARGET_OPTION_VALID_P and
18216 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
18217 * doc/tm.texi: Regenerated.
18219 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
18221 * config/microblaze/microblaze.c:
18222 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
18223 * config/microblaze/microblaze.h: Add -mxl-reorder to
18225 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
18226 instructions emitted if TARGET_REORDER.
18227 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
18228 or 0 for -m/-mno case, but initialises as 2 to detect default use case
18231 2013-03-01 Xinliang David Li <davidxl@google.com>
18233 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
18236 2013-03-01 Jakub Jelinek <jakub@redhat.com>
18238 PR middle-end/56461
18239 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
18240 vector even when returning true. Fix up function comment formatting.
18242 PR middle-end/56461
18243 * ira-build.c (ira_loop_nodes_count): New variable.
18244 (create_loop_tree_nodes): Initialize it.
18245 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
18247 PR middle-end/56461
18248 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
18249 method on dr_chain and result_chain.
18250 * tree-vect-stmts.c (vectorizable_store): Only call
18251 result_chain.create if j == 0.
18253 PR middle-end/56461
18254 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
18255 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
18256 before overwriting it.
18258 2013-03-01 Tobias Burnus <burnus@net-b.de>
18260 * doc/extended.texi (C Extensions): Change order in @menu
18262 (Other MIPS Built-in Functions): Move last MIPS entry before
18263 "picoChip Built-in Functions".
18264 (SH Built-in Functions): Move after RX Built-in Functions.
18265 * doc/gcc.texi (Introduction): Change order in @menu
18267 * doc/md.texi (Constraints): Ditto.
18268 * gty.texi (Type Information): Ditto.
18269 (User-provided marking routines for template types): Make
18271 * doc/invoke.texi (AArch64 Options): Move before
18272 "Adapteva Epiphany Options".
18274 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
18275 Jakub Jelinek <jakub@redhat.com>
18278 * asan.c (gate_asan): Lookup no_sanitize_address instead of
18279 no_address_safety_analysis attribute.
18280 * doc/extend.texi (no_address_safety_attribute): Rename to
18281 no_sanitize_address attribute, mention no_address_safety_analysis
18282 attribute as deprecated alias.
18284 2013-02-28 Jakub Jelinek <jakub@redhat.com>
18286 PR middle-end/56461
18287 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
18288 type to vec<vec<tree> > *.
18289 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
18290 to be vec<tree> instead of vec<tree> *, set vec_defs
18291 to vNULL and call vec_defs.create (number_of_vects), adjust other
18293 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
18294 vectorizable_condition): Adjust vect_get_slp_defs callers.
18296 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18298 * config/aarch64/aarch64.c
18299 (aarch64_float_const_representable): Remove unused variable.
18301 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18303 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
18305 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18307 * config/aarch64/aarch64-builtins.c
18308 (aarch64_init_simd_builtins): Make static.
18310 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18312 * config/aarch64/aarch64.c
18313 (aarch64_simd_make_constant): Make static.
18315 2013-02-28 Martin Jambor <mjambor@suse.cz>
18317 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
18318 with no initialization to the RHS of debug statements.
18320 2013-02-28 Martin Jambor <mjambor@suse.cz>
18322 PR tree-optimization/56294
18323 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
18325 (get_access_replacement): Do not call create_access_replacement.
18326 Assert a replacement exists.
18327 (get_repl_default_def_ssa_name): Create the replacement declaration
18330 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18332 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
18333 final_end_function.
18335 2013-02-28 Marek Polacek <polacek@redhat.com>
18337 PR rtl-optimization/56466
18338 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
18339 if we're changing a loop.
18340 (peel_loops_completely): Likewise.
18342 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
18345 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
18347 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18350 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
18351 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
18352 INTX_FTYPE_FX, FX_FTYPE_INTX.
18353 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
18355 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18357 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
18358 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18359 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
18360 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
18361 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
18362 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
18363 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
18364 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
18365 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
18366 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
18367 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
18368 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
18369 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
18370 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
18371 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
18372 (avrxmega6): Increase max flash segments from 5 to 6.
18373 * config/avr/t-multilib: Regenerate.
18374 * config/avr/avr-tables.opt: Regenerate.
18375 * doc/avr-mmcu.texi: Regenerate.
18377 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18379 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
18380 (avr_device_to_arch): Rename to avr_device_to_ld.
18381 (avr_device_to_as): New prototype.
18382 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
18383 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
18384 * config/avr/driver-avr.c (avr_device_to_as): New.
18385 (avr_device_to_arch): Rename to avr_device_to_ld.
18387 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18389 PR middle-end/56461
18390 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
18391 method on dr_chain and result_chain.
18393 PR middle-end/56461
18394 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
18395 pointer_set_destroy on not_executed_last_iteration.
18397 PR middle-end/56461
18398 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
18400 PR middle-end/56461
18401 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
18402 FOR_EACH_DEFINED_FUNCTION when freeing state.
18404 PR middle-end/56461
18405 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
18407 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
18410 PR middle-end/56461
18411 * ipa-cp.c (decide_whether_version_node): Call vec_free on
18412 known_aggs[i].items and release known_aggs vector.
18414 PR middle-end/56461
18415 * ipa-reference.c (propagate): Free node_info even for alias nodes.
18417 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18419 * config/microblaze/microblaze.c (microblaze_emit_compare):
18420 Use xor for EQ/NE comparisions.
18421 * config/microblaze/microblaze.md (cstoresf4): Add constraints
18422 (cbranchsf4): Adjust operator to comparison_operator.
18424 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18426 PR middle-end/56461
18427 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
18429 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
18430 vec_safe_push, always update *slot.
18431 (redirect_edge_var_map_clear): Use vec_free.
18432 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
18433 (free_var_map_entry): Use vec_free.
18434 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
18435 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
18437 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
18439 PR middle-end/45472
18440 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
18441 when the may_trap_p bit of the exprs being merged differs.
18442 Reorder tests for speculativeness in the logical and operator.
18444 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18446 * incpath.c (add_standard_paths): Use reconcat instead of concat
18447 where appropriate and avoid leaking memory.
18449 * opts.h: Include obstack.h.
18450 (opts_concat): New prototype.
18451 (opts_obstack): New declaration.
18452 * opts.c (opts_concat): New function.
18453 (opts_obstack): New variable.
18454 (init_options_struct): Call gcc_init_obstack on opts_obstack.
18455 (finish_options): Use opts_concat instead of concat
18456 and XOBNEWVEC instead of XNEWVEC.
18457 * opts-common.c (generate_canonical_option, decode_cmdline_option,
18458 generate_option): Likewise.
18459 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
18460 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
18463 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18464 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18466 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18468 PR middle-end/56461
18469 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18471 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18473 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18474 (arm_block_move_unaligned_straight): Likewise.
18475 (arm_adjust_block_mem): Likewise.
18477 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18480 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18481 temp, cond and label.
18482 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18485 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18486 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18489 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18490 of prologue/epilogue functions.
18493 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18494 Remove unused variable cfa_offset.
18495 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18498 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18499 type promotion to unsigned.
18502 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18503 for HOST_WIDE_INT of 32 bit / same size as int.
18504 (arm_block_move_unaligned_straight): Likewise.
18505 (arm_adjust_block_mem): Likewise.
18508 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18511 2013-02-26 Marek Polacek <polacek@redhat.com>
18513 PR tree-optimization/56426
18514 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
18516 2013-02-26 Richard Biener <rguenther@suse.de>
18519 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18520 unused variable loops.
18522 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18524 PR tree-optimization/56448
18525 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18526 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18527 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18528 later operands of the references, or even first operand for
18529 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18531 PR tree-optimization/56443
18532 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18533 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18534 to type_for_mode langhook.
18536 2013-02-25 Matt Turner <mattst88@gmail.com>
18538 * doc/invoke.texi: Document r4700.
18540 2013-02-25 Richard Biener <rguenther@suse.de>
18542 PR tree-optimization/56175
18543 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18545 (simplify_bitwise_binary): ... here. Also guard the conversion
18546 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18548 2013-02-25 Catherine Moore <clm@codesourcery.com>
18551 2013-02-24 Catherine Moore <clm@codesourcery.com>
18552 Maciej W. Rozycki <macro@codesourcery.com>
18553 Tom de Vries <tom@codesourcery.com>
18554 Nathan Sidwell <nathan@codesourcery.com>
18555 Iain Sandoe <iain@codesourcery.com>
18556 Nathan Froyd <froydnj@codesourcery.com>
18557 Chao-ying Fu <fu@mips.com>
18559 * doc/extend.texi: (micromips, nomicromips, nocompression):
18560 Document new function attributes.
18561 * doc/invoke.texi (minterlink-compressed, mmicromips,
18562 m14k, m14ke, m14kec): Document new options.
18563 (minterlink-mips16): Update documentation.
18564 * doc/md.texi (ZC, ZD): Document new constraints.
18565 * configure.ac (gcc_cv_as_micromips): Check if linker
18566 supports the .set micromips directive.
18567 * configure: Regenerate.
18568 * config.in: Regenerate.
18569 * config/mips/mips-tables.opt: Regenerate.
18570 * config/mips/micromips.md: New file.
18571 * constraints.md (ZC, AD): New constraints.
18572 * config/mips/predicates.md (movep_src_register): New predicate.
18573 (movep_src_operand): New predicate.
18574 (non_volatile_mem_operand): New predicate.
18575 * config/mips/mips.md (multimem): New type.
18576 (length): Differentiate between 17-bit and 18-bit branch offsets.
18577 (MOVEP1, MOVEP2): New mode iterator.
18578 (mov_<load>l): Use ZC constraint.
18579 (mov_<load>r): Likewise.
18580 (mov_<store>l): Likewise.
18581 (mov_<store>r): Likewise.
18582 (*branch_equality<mode>_inverted): Add microMIPS support.
18583 (*branch_equality<mode>): Likewise.
18584 (*jump_absolute): Likewise.
18585 (indirect_jump_<mode>): Likewise.
18586 (tablejump_<mode>): Likewise.
18587 (<optab>_internal): Likewise.
18588 (sibcall_internal): Likewise.
18589 (sibcall_value_internal): Likewise.
18590 (prefetch): Use constraint ZD.
18591 * config/mips/mips.opt (minterlink-compressed): New option.
18592 (minterlink-mips16): Now an alias for minterlink-compressed.
18593 (mmicromips): New option.
18594 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18595 (compare_and_swap_12): Likewise.
18596 (sync_add<mode>): Likewise.
18597 (sync_<optab>_12): Likewise.
18598 (sync_old_<optab>_12): Likewise.
18599 (sync_new_<optab>_12): Likewise.
18600 (sync_nand_12): Likewise.
18601 (sync_old_nand_12): Likewise.
18602 (sync_new_nand_12): Likewise.
18603 (sync_sub<mode>): Likewise.
18604 (sync_old_add<mode>): Likewise.
18605 (sync_old_sub<mode>): Likewise.
18606 (sync_new_add<mode>): Likewise.
18607 (sync_new_sub<mode>): Likewise.
18608 (sync_<optab><mode>): Likewise.
18609 (sync_old_<optab><mode>): Likewise.
18610 (sync_new_<optab><mode>): Likewise.
18611 (sync_nand<mode>): Likewise.
18612 (sync_old_nand<mode>): Likewise.
18613 (sync_new_nand<mode>): Likewise.
18614 (sync_lock_test_and_set<mode>): Likewise.
18615 (test_and_set_12): Likewise.
18616 (atomic_compare_and_swap<mode>): Likewise.
18617 (atomic_exchange<mode>_llsc): Likewise.
18618 (atomic_fetch_add<mode>_llsc): Likewise.
18619 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18620 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18621 (umips_save_restore_pattern_p): Likewise.
18622 (umips_load_store_pair_p): Likewise.
18623 (umips_output_load_store_pair): Likewise.
18624 (umips_movep_target_p): Likewise.
18625 (umips_12bit_offset_address_p): Likewise.
18626 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18627 (mips_base_mips16): Rename this...
18628 (mips_base_compression_flags): ...to this. Update all uses.
18629 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18630 (mips_mips16_decl_p): Delete.
18631 (mips_nomips16_decl_p): Delete.
18632 (mips_get_compress_on_flags): New function.
18633 (mips_get_compress_off_flags): New function.
18634 (mips_get_compress_mode): New function.
18635 (mips_get_compress_on_name): New function.
18636 (mips_get_compress_off_name): New function.
18637 (mips_insert_attributes): Support multiple compression types.
18638 (mips_merge_decl_attributes): Likewise.
18639 (umips_12bit_offset_address_p): New function.
18640 (mips_start_function_definition): Emit .set micromips directive.
18641 (mips_call_may_need_jalx_p): New function.
18642 (mips_function_ok_for_sibcall): Add microMIPS support.
18643 (mips_print_operand_punctuation): Support short delay slots and
18645 (umips_swm_mask, umips_swm_encoding): New.
18646 (umips_build_save_restore): New function.
18647 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18648 (was_mips16_p): Remove.
18649 (old_compression_mode): New.
18650 (mips_set_compression_mode): New function.
18651 (mips_set_current_function): Add microMIPS support.
18652 (mips_option_override): Likewise.
18653 (umips_save_restore_pattern_p): New function.
18654 (umips_output_save_restore): New function.
18655 (umips_load_store_pair_p_1): New function.
18656 (umips_load_store_pair_p): New function.
18657 (umips_output_load_store_pair_1): New function.
18658 (umips_output_load_store_pair): New function.
18659 (umips_movep_target_p) New function.
18660 (mips_prepare_pch_save): Add microMIPS support.
18661 * config/mips/mips.h (TARGET_COMPRESSION): New.
18662 (TARGET_CPU_CPP_BUILTINS): Update macro
18663 to use new compression flags and to support microMIPS.
18664 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18665 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18666 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18667 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18668 (ASM_SPEC): Support mmicromips and mno-micromips.
18669 (M16STORE_REG_P): New macro.
18670 (MIPS_CALL): Support TARGET_MICROMIPS.
18671 (MICROMIPS_J): New macro.
18672 (mips_base_mips16): Rename this...
18673 (mips_base_compression_flags): ...to this.
18674 (UMIPS_12BIT_OFFSET_P): New macro.
18675 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18676 (MULTILIB_DIRNAMES): Likewise.
18678 2013-02-25 Tom de Vries <tom@codesourcery.com>
18680 PR rtl-optimization/56131
18681 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18682 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18683 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18685 2013-02-25 Tobias Burnus <burnus@net-b.de>
18687 * doc/invoke.texi (-fsanitize=): Move from optimization
18688 to debugging options.
18690 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18692 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18694 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18695 Alexander Monakov <amonakov@ispras.ru>
18697 PR middle-end/56077
18698 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18699 flush pending lists also on non-jumps. Adjust comment.
18701 2013-02-24 Catherine Moore <clm@codesourcery.com>
18702 Maciej W. Rozycki <macro@codesourcery.com>
18703 Tom de Vries <tom@codesourcery.com>
18704 Nathan Sidwell <nathan@codesourcery.com>
18705 Iain Sandoe <iain@codesourcery.com>
18706 Nathan Froyd <froydnj@codesourcery.com>
18707 Chao-ying Fu <fu@mips.com>
18709 * doc/extend.texi: (micromips, nomicromips, nocompression):
18710 Document new function attributes.
18711 * doc/invoke.texi (minterlink-compressed, mmicromips,
18712 m14k, m14ke, m14kec): Document new options.
18713 (minterlink-mips16): Update documentation.
18714 * doc/md.texi (ZC, ZD): Document new constraints.
18715 * configure.ac (gcc_cv_as_micromips): Check if linker
18716 supports the .set micromips directive.
18717 * configure: Regenerate.
18718 * config.in: Regenerate.
18719 * config/mips/mips-tables.opt: Regenerate.
18720 * config/mips/micromips.md: New file.
18721 * constraints.md (ZC, AD): New constraints.
18722 * config/mips/predicates.md (movep_src_register): New predicate.
18723 (movep_src_operand): New predicate.
18724 (non_volatile_mem_operand): New predicate.
18725 * config/mips/mips.md (multimem): New type.
18726 (length): Differentiate between 17-bit and 18-bit branch offsets.
18727 (MOVEP1, MOVEP2): New mode iterator.
18728 (mov_<load>l): Use ZC constraint.
18729 (mov_<load>r): Likewise.
18730 (mov_<store>l): Likewise.
18731 (mov_<store>r): Likewise.
18732 (*branch_equality<mode>_inverted): Add microMIPS support.
18733 (*branch_equality<mode>): Likewise.
18734 (*jump_absolute): Likewise.
18735 (indirect_jump_<mode>): Likewise.
18736 (tablejump_<mode>): Likewise.
18737 (<optab>_internal): Likewise.
18738 (sibcall_internal): Likewise.
18739 (sibcall_value_internal): Likewise.
18740 (prefetch): Use constraint ZD.
18741 * config/mips/mips.opt (minterlink-compressed): New option.
18742 (minterlink-mips16): Now an alias for minterlink-compressed.
18743 (mmicromips): New option.
18744 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18745 (compare_and_swap_12): Likewise.
18746 (sync_add<mode>): Likewise.
18747 (sync_<optab>_12): Likewise.
18748 (sync_old_<optab>_12): Likewise.
18749 (sync_new_<optab>_12): Likewise.
18750 (sync_nand_12): Likewise.
18751 (sync_old_nand_12): Likewise.
18752 (sync_new_nand_12): Likewise.
18753 (sync_sub<mode>): Likewise.
18754 (sync_old_add<mode>): Likewise.
18755 (sync_old_sub<mode>): Likewise.
18756 (sync_new_add<mode>): Likewise.
18757 (sync_new_sub<mode>): Likewise.
18758 (sync_<optab><mode>): Likewise.
18759 (sync_old_<optab><mode>): Likewise.
18760 (sync_new_<optab><mode>): Likewise.
18761 (sync_nand<mode>): Likewise.
18762 (sync_old_nand<mode>): Likewise.
18763 (sync_new_nand<mode>): Likewise.
18764 (sync_lock_test_and_set<mode>): Likewise.
18765 (test_and_set_12): Likewise.
18766 (atomic_compare_and_swap<mode>): Likewise.
18767 (atomic_exchange<mode>_llsc): Likewise.
18768 (atomic_fetch_add<mode>_llsc): Likewise.
18769 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18770 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18771 (umips_save_restore_pattern_p): Likewise.
18772 (umips_load_store_pair_p): Likewise.
18773 (umips_output_load_store_pair): Likewise.
18774 (umips_movep_target_p): Likewise.
18775 (umips_12bit_offset_address_p): Likewise.
18776 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18777 (mips_base_mips16): Rename this...
18778 (mips_base_compression_flags): ...to this. Update all uses.
18779 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18780 (mips_mips16_decl_p): Delete.
18781 (mips_nomips16_decl_p): Delete.
18782 (mips_get_compress_on_flags): New function.
18783 (mips_get_compress_off_flags): New function.
18784 (mips_get_compress_mode): New function.
18785 (mips_get_compress_on_name): New function.
18786 (mips_get_compress_off_name): New function.
18787 (mips_insert_attributes): Support multiple compression types.
18788 (mips_merge_decl_attributes): Likewise.
18789 (umips_12bit_offset_address_p): New function.
18790 (mips_start_function_definition): Emit .set micromips directive.
18791 (mips_call_may_need_jalx_p): New function.
18792 (mips_function_ok_for_sibcall): Add microMIPS support.
18793 (mips_print_operand_punctuation): Support short delay slots and
18795 (umips_swm_mask, umips_swm_encoding): New.
18796 (umips_build_save_restore): New function.
18797 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18798 (was_mips16_p): Remove.
18799 (old_compression_mode): New.
18800 (mips_set_compression_mode): New function.
18801 (mips_set_current_function): Add microMIPS support.
18802 (mips_option_override): Likewise.
18803 (umips_save_restore_pattern_p): New function.
18804 (umips_output_save_restore): New function.
18805 (umips_load_store_pair_p_1): New function.
18806 (umips_load_store_pair_p): New function.
18807 (umips_output_load_store_pair_1): New function.
18808 (umips_output_load_store_pair): New function.
18809 (umips_movep_target_p) New function.
18810 (mips_prepare_pch_save): Add microMIPS support.
18811 * config/mips/mips.h (TARGET_COMPRESSION): New.
18812 (TARGET_CPU_CPP_BUILTINS): Update macro
18813 to use new compression flags and to support microMIPS.
18814 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18815 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18816 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18817 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18818 (ASM_SPEC): Support mmicromips and mno-micromips.
18819 (M16STORE_REG_P): New macro.
18820 (MIPS_CALL): Support TARGET_MICROMIPS.
18821 (MICROMIPS_J): New macro.
18822 (mips_base_mips16): Rename this...
18823 (mips_base_compression_flags): ...to this.
18824 (UMIPS_12BIT_OFFSET_P): New macro.
18825 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18826 (MULTILIB_DIRNAMES): Likewise.
18828 2013-02-24 Jakub Jelinek <jakub@redhat.com>
18831 * target-globals.c (save_target_globals): For init_reg_sets and
18832 target_reinit remporarily set this_fn_optabs to this_target_optabs.
18834 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
18836 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
18837 * config/aarch64/t-aarch64
18838 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
18840 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
18842 PR inline-asm/56148
18843 * lra-constraints.c (process_alt_operands): Reload operand
18844 conflicting with earlier clobber only if no more other conflicting
18847 2013-02-22 Jakub Jelinek <jakub@redhat.com>
18850 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
18851 if not linking a shared library.
18853 2013-02-22 Seth LaForge <sethml@google.com>
18855 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
18857 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
18859 * config/arm/arm.md (split for extendsidi): Update condition.
18860 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
18861 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
18862 (qhs_zextenddi_cstr): Likewise.
18864 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18866 PR middle-end/56420
18867 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
18868 avoid signed wrapping.
18869 (expand_mult): Handle properly multiplication by
18870 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
18871 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
18872 in the compiler if coeff is HOST_WIDE_INT_MIN.
18873 (expand_divmod): Don't make ext_op1 static, change it's type to
18874 uhwi. Avoid undefined behavior in -INTVAL (op1).
18876 PR rtl-optimization/50339
18877 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
18879 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
18880 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
18881 into splitting_ashiftrt field.
18882 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
18884 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
18887 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
18889 PR middle-end/56108
18890 * trans-mem.c (execute_tm_mark): Do not expand transactions that
18891 are sure to go irrevocable.
18893 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
18895 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
18896 scalars are valid operands.
18898 2013-02-21 Martin Jambor <mjambor@suse.cz>
18900 PR tree-optimization/56310
18901 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
18902 only matching indices and non-negative final offsets.
18903 (intersect_aggregates_with_edge): Pass src_idx to
18904 agg_replacements_to_vector. Pass src_idx insstead of index to
18905 intersect_with_agg_replacements.
18907 2013-02-21 Martin Jambor <mjambor@suse.cz>
18909 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
18910 instead of hard-wired defaults.
18912 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
18914 * doc/invoke.texi (MIPS Options): Update documentation of the
18915 floating-point multiply-accumulate instruction restrictions.
18917 2013-02-21 Kostya Serebryany <kcc@google.com>
18919 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
18920 asan_shadow_offset on x86_64 linux.
18922 2013-02-21 Richard Biener <rguenther@suse.de>
18924 PR tree-optimization/56415
18926 2013-02-11 Richard Biener <rguenther@suse.de>
18928 PR tree-optimization/56273
18929 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18932 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18935 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
18938 PR inline-asm/56405
18939 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
18940 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
18942 2013-02-20 Jan Hubicka <jh@suse.cz>
18944 PR tree-optimization/56265
18945 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
18946 when target is referenced for first time.
18948 2013-02-20 Richard Biener <rguenther@suse.de>
18950 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
18951 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
18952 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
18953 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
18954 not return anything.
18955 (rename_ssa_copies): Do not remove unused locals.
18956 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
18957 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
18958 * passes.c (execute_function_todo): Do not schedule unused locals
18959 removal if cleanup_tree_cfg did something.
18960 * tree-ssa-live.c (remove_unused_locals): Dump statistics
18961 about the number of removed locals.
18963 2013-02-20 Richard Biener <rguenther@suse.de>
18965 PR tree-optimization/56398
18966 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
18968 2013-02-20 Martin Jambor <mjambor@suse.cz>
18970 PR tree-optimization/55334
18971 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
18972 restricted pointers to arrays.
18974 2013-02-20 Richard Biener <rguenther@suse.de>
18975 Jakub Jelinek <jakub@redhat.com>
18977 PR tree-optimization/56396
18978 * tree-ssa-ccp.c (n_const_val): New static variable.
18979 (get_value): Return NULL for SSA names we don't have a lattice
18981 (ccp_initialize): Initialize n_const_val.
18982 * tree-ssa-copy.c (n_copy_of): New static variable.
18983 (init_copy_prop): Initialize n_copy_of.
18984 (get_value): Return NULL_TREE for SSA names we don't have a
18987 2013-02-20 Martin Jambor <mjambor@suse.cz>
18989 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
18991 2013-02-20 Richard Biener <rguenther@suse.de>
18993 * genpreds.c (write_lookup_constraint): Do not compare first
18994 letter of the constraint again.
18996 2013-02-20 Richard Biener <rguenther@suse.de>
18998 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
19000 (get_use_iv_cost): Terminate hashtable walk when coming across
19003 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
19005 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
19006 reassociation for avx2 targets.
19008 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19010 * config/microblaze/microblaze.c: microblaze_has_clz = 0
19011 Add version check for v8.10.a to enable microblaze_has_clz
19012 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
19013 version and TARGET_PATTERN_COMPARE check
19014 * config/microblaze/microblaze.md: New clzsi2 instruction
19016 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19018 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
19019 function before branching.
19021 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19023 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
19025 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
19027 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19029 PR middle-end/55889
19030 * sel-sched.c: Include ira.h.
19031 (implicit_clobber_conflict_p): New function.
19032 (moveup_expr): Use it.
19033 * Makefile.in (sel-sched.o): Depend on ira.h.
19035 2013-02-19 Richard Biener <rguenther@suse.de>
19037 PR tree-optimization/56384
19038 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
19039 (vn_hash_type): Split out from ...
19040 (vn_hash_constant_with_type): ... here.
19041 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
19042 (vn_phi_eq): Compare types from vn_phi_s structure.
19043 (vn_phi_lookup): Populate vn_phi_s type.
19044 (vn_phi_insert): Likewise.
19046 2013-02-19 Jakub Jelinek <jakub@redhat.com>
19048 PR tree-optimization/56350
19049 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
19050 if haven't found reduction or nested cycle operand, rather than
19051 asserting we must find it.
19053 PR tree-optimization/56381
19054 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
19057 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
19058 Jakub Jelinek <jakub@redhat.com>
19061 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
19062 (swap_optab_enable): Same.
19063 (init_all_optabs): Use argument instead of global.
19064 * tree.h (struct tree_optimization_option): New field target_optabs.
19065 * expr.h (init_all_optabs): Add argument to prototype.
19066 (TREE_OPTIMIZATION_OPTABS): New.
19067 (save_optabs_if_changed): Protoize.
19068 * optabs.h: Declare this_fn_optabs.
19069 * optabs.c (save_optabs_if_changed): New.
19070 Declare this_fn_optabs.
19071 (init_optabs): Add argument to init_all_optabs() call.
19072 * function.c (invoke_set_current_function_hook): Handle per
19074 * function.h (struct function): New field optabs.
19075 * config/mips/mips.c (mips_set_mips16_mode): Handle when
19076 optimization_current_node has changed.
19077 * target-globals.h (save_target_globals_default_opts): Protoize.
19078 * target-globals.c (save_target_globals_default_opts): New.
19080 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19083 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
19084 registers %fr12 and %fr12R as call used.
19087 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
19088 and HImode, require all displacements to be an integer multiple of
19090 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
19091 only allow QImode and HImode when reload is in progress and strict is
19092 true. Likewise for symbolic addresses. Use base14_operand to check
19093 displacements in REG+BASE addresses.
19095 2013-02-18 Richard Biener <rguenther@suse.de>
19097 PR tree-optimization/56366
19098 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19099 handle sign-conversion of outer-loop initial induction value.
19101 2013-02-18 Richard Biener <rguenther@suse.de>
19103 PR middle-end/56349
19104 * cfghooks.c (merge_blocks): If we merge a latch into another
19105 block adjust references to it.
19106 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
19107 (verify_loop_structure): Verify that a recorded latch is in fact
19110 2013-02-18 Richard Biener <rguenther@suse.de>
19112 PR tree-optimization/56321
19113 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
19114 order SSA name release and virtual operand unlinking.
19116 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19118 * config/microblaze/microblaze.md (save_stack_block): Define.
19119 (restore_stack_block): Likewise.
19121 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19123 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
19124 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
19125 * config/microblaze/microblaze.c (microblaze_option_override):
19126 Bail out early for PIC modes when target does not support PIC.
19128 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19130 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
19131 Replace with a microblaze version.
19132 (microblaze_trampoline_init): Adapt for microblaze.
19133 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
19136 2013-02-16 Jakub Jelinek <jakub@redhat.com>
19137 Dodji Seketeli <dodji@redhat.com>
19140 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
19141 (instrument_mem_region_access): Do not forget to always put
19142 instrumentation of the of 'base' and 'base + len' in a "if (len !=
19143 0) statement, even for cases where either 'base' or 'base + len'
19144 are not instrumented -- because they have been previously
19145 instrumented. Simplify the logic by putting all the statements
19146 instrument 'base + len' inside a sequence, and then insert that
19147 sequence right before the current insertion point. Then, to
19148 instrument 'base + len', just get an iterator on that statement.
19149 And do not forget to update the pointer to iterator the function
19150 received as argument.
19152 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
19154 PR rtl-optimization/56348
19155 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
19157 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
19159 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
19160 (clean_graph_dump_file): Pass base to start_graph_dump.
19162 2013-02-14 Richard Henderson <rth@redhat.com>
19165 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
19167 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
19169 * collect2-aix.h: Define F_LOADONLY.
19171 2013-02-14 Richard Biener <rguenther@suse.de>
19174 * varasm.c (output_constant_def_1): Get the decl representing
19175 the constant as argument.
19176 (output_constant_def): Wrap output_constant_def_1.
19177 (make_decl_rtl): Use output_constant_def_1 with the decl
19178 representing the constant.
19179 (build_constant_desc): Optionally re-use a decl already
19180 representing the constant.
19181 (tree_output_constant_def): Adjust.
19183 2013-02-14 Dodji Seketeli <dodji@redhat.com>
19186 * asan.c (instrument_builtin_call): Really put the length of the
19187 second source argument into src1_len.
19189 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19191 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
19192 argument. If it is false, don't create edge from then_bb to
19194 (insert_if_then_before_iter): Pass true to it.
19195 (build_check_stmt): Pass false to it.
19196 (transform_statements): Flush hash table only on extended basic
19197 block boundaries, rather than at the beginning of every bb.
19198 Don't flush hash table on nonfreeing_call_p calls.
19199 * tree-flow.h (nonfreeing_call_p): New prototype.
19200 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
19202 2013-02-13 David S. Miller <davem@davemloft.net>
19204 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
19206 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
19209 * ira.c (max_regno_before_ira): Move from ...
19211 (fix_reg_equiv_init): Use max_regno_before_ira instead of
19214 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19216 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
19218 2013-02-13 Richard Biener <rguenther@suse.de>
19221 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
19222 globals in MEM_REFs.
19224 2013-02-13 Richard Biener <rguenther@suse.de>
19226 * loop-init.c (loop_optimizer_init): Clear loop state when
19227 re-initializing preserved loops.
19228 * loop-unswitch.c (unswitch_single_loop): Return whether
19229 we unswitched the loop. Do not verify loop state here.
19230 (unswitch_loops): When we unswitched a loop discover new loops.
19232 2013-02-13 Kostya Serebryany <kcc@google.com>
19234 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
19236 * sanitizer.def: Rename __asan_init to __asan_init_v1.
19238 2013-02-12 Dodji Seketeli <dodji@redhat.com>
19240 Avoid instrumenting duplicated memory access in the same basic block
19241 * Makefile.in (asan.o): Add new dependency on hash-table.h
19242 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
19243 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
19244 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
19245 (free_mem_ref_resources, has_mem_ref_been_instrumented)
19246 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
19247 (get_mem_ref_of_assignment): New functions.
19248 (get_mem_refs_of_builtin_call): Extract from
19249 instrument_builtin_call and tweak a little bit to make it fit with
19251 (instrument_builtin_call): Use the new
19252 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
19253 of is_gimple_builtin_call.
19254 (instrument_derefs, instrument_mem_region_access): Insert the
19255 instrumented memory reference into the hash table.
19256 (maybe_instrument_assignment): Renamed instrument_assignment into
19257 this, and change it to advance the iterator when instrumentation
19258 actually happened and return true in that case. This makes it
19259 homogeneous with maybe_instrument_assignment, and thus give a
19260 chance to callers to be more 'regular'.
19261 (transform_statements): Clear the memory reference hash table
19262 whenever we enter a new BB, when we cross a function call, or when
19263 we are done transforming statements. Use
19264 maybe_instrument_assignment instead of instrumentation. No more
19265 need to special case maybe_instrument_assignment and advance the
19266 iterator after calling it; it's now handled just like
19267 maybe_instrument_call. Update comment.
19269 2013-02-13 Richard Biener <rguenther@suse.de>
19271 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
19272 Fix loop discovery code.
19274 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
19276 PR inline-asm/56148
19277 * lra-constraints.c (process_alt_operands): Match early clobber
19278 operand with itself. Check conflicts with earlyclobber only if
19279 the operand is not reloaded. Prefer to reload conflicting operand
19280 if earlyclobber and matching operands are the same.
19282 2013-02-12 Richard Biener <rguenther@suse.de>
19285 * lto-streamer-out.c (write_symbol): Do not output symbols
19286 for hard register variables.
19288 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
19291 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
19292 (umulsidi3_insn, mulsidi3_insn): New insns.
19294 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
19296 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
19297 (struct tune_params): Add vec_costs field.
19298 * config/arm/arm.c (arm_builtin_vectorization_cost)
19299 (arm_add_stmt_cost): New functions.
19300 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
19301 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
19302 (arm_default_vec_cost): New struct of type cpu_vec_costs.
19303 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
19304 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
19305 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
19306 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
19308 2013-02-12 Richard Biener <rguenther@suse.de>
19311 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
19312 decls again if possible.
19314 2013-02-12 Richard Biener <rguenther@suse.de>
19316 PR middle-end/56288
19317 * tree-ssa.c (verify_ssa_name): Fix check, move
19318 SSA_NAME_IN_FREE_LIST check up.
19320 2013-02-12 Jakub Jelinek <jakub@redhat.com>
19321 Steven Bosscher <steven@gcc.gnu.org>
19323 PR rtl-optimization/56151
19324 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
19325 equal to op0 or op1, and last_insn pattern is CODE operation
19326 with MEM dest and one of the operands matches that MEM.
19328 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
19330 * doc/extend.texi: Document Function Multiversioning and "default"
19331 parameter string to target attribute.
19332 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
19333 target attribute parameter is "default".
19334 (ix86_compare_version_priority): Remove checks for target attribute.
19335 (ix86_mangle_function_version_assembler_name): Change error to sorry.
19336 Remove check for target attribute equal to NULL. Add assert.
19337 (ix86_generate_version_dispatcher_body): Change error to sorry.
19339 2013-02-11 Iain Sandoe <iain@codesourcery.com>
19340 Jack Howarth <howarth@bromo.med.uc.edu>
19341 Patrick Marlier <patrick.marlier@gmail.com>
19344 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
19345 define ENDFILE_SPEC as TM_DESTRUCTOR.
19346 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
19348 2013-02-11 Alexander Potapenko <glider@google.com>
19349 Jack Howarth <howarth@bromo.med.uc.edu>
19350 Jakub Jelinek <jakub@redhat.com>
19353 * config/darwin.c (cdtor_record): Rename ctor_record.
19354 (sort_cdtor_records): Rename sort_ctor_records.
19355 (finalize_dtors): New routine to sort destructors by
19356 priority before use in assemble_integer.
19357 (machopic_asm_out_destructor): Use finalize_dtors if needed.
19359 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
19361 PR rtl-optimization/56275
19362 * simplify-rtx.c (avoid_constant_pool_reference): Check that
19363 offset is non-negative and less than cmode size before
19364 calling simplify_subreg.
19366 2013-02-11 Richard Biener <rguenther@suse.de>
19368 PR tree-optimization/56264
19369 * cfgloop.h (fix_loop_structure): Adjust prototype.
19370 * loop-init.c (fix_loop_structure): Return the number of
19371 newly discovered loops.
19372 * tree-cfgcleanup.c (repair_loop_structures): When new loops
19373 are discovered, do a full loop-closed SSA rewrite.
19375 2013-02-11 Richard Biener <rguenther@suse.de>
19377 PR tree-optimization/56273
19378 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19380 (check_array_ref): Fix missing newline in dumps.
19381 (search_for_addr_array): Likewise.
19383 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
19385 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
19387 2013-02-09 Jakub Jelinek <jakub@redhat.com>
19390 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
19392 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
19394 PR rtl-optimization/56246
19395 * lra-constraints.c (simplify_operand_subreg): Try to reuse
19397 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
19398 constraints are satisfied.
19400 2013-02-08 Jeff Law <law@redhat.com>
19403 * emit-rtl.c (reg_is_parm_p): New function.
19404 * regs.h (reg_is_parm_p): New prototype.
19405 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
19406 callee-clobbered registers.
19408 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
19411 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
19412 If there is no implicit builtin declaration, just return NULL.
19414 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19416 * config/i386/sse.md (FMAMODEM): New mode iterator.
19417 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
19418 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
19420 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19422 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
19423 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
19424 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
19426 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19428 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
19429 (microblaze*-*-elf): Likewise.
19430 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
19432 * config/microblaze/microblaze-c.c: Add builtin defines for
19433 _LITTLE_ENDIAN and _BIG_ENDIAN.
19434 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
19435 add to TARGET_DEFAULT flags.
19436 Expand ASM_SPEC and LINK_SPEC.
19437 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
19438 * config/microblaze/microblaze.md: Update extendsidi2 and
19439 movdi_internal instructions to use low-order / high-order reg
19441 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
19442 options and inversemask / mask of LITTLE_ENDIAN.
19443 * config/microblaze/t-microblaze: Expand multilib options to
19444 include mlittle-endian (le) and update exceptions patterns.
19446 2013-02-08 Jakub Jelinek <jakub@redhat.com>
19448 PR rtl-optimization/56195
19449 * lra-constraints.c (get_reload_reg): Don't reuse regs
19450 if they have smaller mode than requested, if they have
19451 wider mode than requested, try to return a SUBREG.
19453 PR tree-optimization/56250
19454 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
19455 if type is unsigned and code isn't MULT_EXPR.
19457 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19459 PR tree-optimization/56064
19460 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
19461 bits according to mode.
19462 * fixed-value.h (fixed_from_double_int)
19463 (const_fixed_from_double_int): Adjust comments.
19465 2013-02-08 Richard Biener <rguenther@suse.de>
19468 * lto-streamer.h (struct data_in): Remove current_file, current_line
19469 and current_col members.
19470 * lto-streamer-out.c (lto_output_location): Stream changed bits
19471 en-block for efficiency.
19472 * lto-streamer-in.c (clear_line_info): Remove.
19473 (lto_input_location): Cache current file, line and column
19474 globally via local statics. Read changed bits en-block.
19475 (input_function): Do not call clear_line_info.
19476 (lto_read_body): Likewise.
19477 (lto_input_toplevel_asms): Likewise.
19479 2013-02-08 Michael Matz <matz@suse.de>
19481 PR tree-optimization/52448
19482 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19483 (nt_call_phase): New static.
19484 (add_or_mark_expr): Only mark accesses with newer phase than any
19486 (nonfreeing_call_p): New.
19487 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19488 (nt_fini_block): Keep blocks marked as visited.
19489 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19491 2013-02-08 Richard Biener <rguenther@suse.de>
19493 * ira.c (ira): Free broken dominator information.
19495 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19497 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19499 2013-02-08 Marek Polacek <polacek@redhat.com>
19501 * cfgloop.c (verify_loop_structure): Add more checking of headers.
19503 2013-02-08 Richard Biener <rguenther@suse.de>
19505 PR middle-end/56181
19506 * cfgloop.h (flow_loops_find): Adjust.
19507 (bb_loop_header_p): Declare.
19508 * cfgloop.c (bb_loop_header_p): New function split out from ...
19509 (flow_loops_find): ... here. Adjust function signature,
19510 support incremental loop structure update.
19511 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19512 * cfgloopmanip.c (fix_loop_structure): Move ...
19513 * loop-init.c (fix_loop_structure): ... here.
19514 (apply_loop_flags): Split out from ...
19515 (loop_optimizer_init): ... here.
19516 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19517 in incremental mode, only remove dead loops here.
19519 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19522 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19523 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19524 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19525 (*round<mode>3.libgcc): New insns for fixed-modes.
19526 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19527 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19528 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19529 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19530 implementations. Define to __builtin_avr_absFX,
19531 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19532 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19533 __builtin_avr_countlsFX, respectively.
19534 * config/avr/avr-c.c (target.h): Include it.
19535 (enum avr_builtin_id): New enum.
19536 (avr_resolve_overloaded_builtin): New static function.
19537 (avr_register_target_pragmas): Use it to set
19538 targetm.resolve_overloaded_builtin.
19539 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19540 tree nodes used by DEF_BUILTIN.
19541 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19542 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19543 <AVR_BUILTIN_xxBITS>: Same.
19545 2013-02-08 Richard Biener <rguenther@suse.de>
19547 * cfgloop.c (verify_loop_structure): Properly handle
19548 a loop exiting to another loop header.
19549 * ira-int.h (ira_loops): Remove.
19550 * ira.c (ira_loops): Remove.
19551 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19552 (do_reload): Use loop_optimizer_finalize.
19553 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19554 number_of_loops to access the loop tree.
19555 (more_one_region_p): Likewise.
19556 (finish_loop_tree_nodes): Likewise.
19557 (rebuild_regno_allocno_maps): Likewise.
19558 (mark_loops_for_removal): Likewise.
19559 (mark_all_loops_for_removal): Likewise.
19560 (remove_unnecessary_regions): Likewise.
19561 (ira_build): Likewise.
19562 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19564 2013-02-08 Richard Biener <rguenther@suse.de>
19566 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
19567 * ipa-pure-const.c (analyze_function): Avoid calling
19568 mark_irreducible_loops twice.
19569 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
19571 2013-02-07 David S. Miller <davem@davemloft.net>
19573 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
19575 * var-tracking.c (vt_add_function_parameter): Test the presence of
19576 HAVE_window_save properly and do not remap argument registers when
19577 we have a leaf function.
19579 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
19582 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
19584 * config/i386/i386.c (ix86_print_operand): Ditto.
19586 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19588 * lra-constraints.c (process_alt_operands): Fix recently added comment.
19590 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19592 PR rtl-optimization/56225
19593 * lra-constraints.c (process_alt_operands): Check that reload hard
19594 reg can hold value for strict_low_part.
19596 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19599 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
19600 dwarf2out_end_function.
19601 (in_first_function_p, maybe_at_text_label_p,
19602 first_loclabel_num_not_at_text_label): New variables.
19603 (dwarf2out_var_location): In the first function find out
19604 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
19605 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
19608 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
19610 PR rtl-optimization/56178
19611 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
19612 SUBREG of a register. Tidy up related block of code.
19613 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
19614 note if the source is a register or a SUBREG of a register.
19616 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19619 * config/rs6000/rs6000.md (ptrm): New mode attr.
19620 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19621 call_value_indirect_aix<pttrsize>,
19622 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19625 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19627 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19628 if -bnortl. Convert to strcmp and strncmp.
19630 2013-02-07 Alan Modra <amodra@gmail.com>
19633 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19634 addresses won't wrap when offsetting.
19635 (rs6000_secondary_reload): Provide secondary reloads needed for
19636 wrapping LO_SUM addresses.
19638 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19640 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19641 MACH, just __MACH__.
19643 2013-02-06 Richard Biener <rguenther@suse.de>
19645 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19646 instead of calling fix_loop_structure.
19648 2013-02-06 Jakub Jelinek <jakub@redhat.com>
19650 PR middle-end/56217
19651 * omp-low.c (use_pointer_for_field): Return false if
19652 lower_send_shared_vars doesn't generate any copy-out code.
19654 2013-02-06 Tom de Vries <tom@codesourcery.com>
19656 PR rtl-optimization/56131
19657 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19658 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19659 of the label is NULL. Add comment.
19661 2013-02-05 Jakub Jelinek <jakub@redhat.com>
19663 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19666 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19667 (STATIC_LIBTSAN_LIBS): Likewise.
19668 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19669 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19670 is defined, don't add anything else beyond that.
19671 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19672 (LINK_COMMAND_SPEC): Use them.
19674 PR tree-optimization/56205
19675 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19676 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19677 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19679 2013-02-05 Richard Biener <rguenther@suse.de>
19681 PR tree-optimization/53342
19682 PR tree-optimization/53185
19683 * tree-vectorizer.h (vect_check_strided_load): Remove.
19684 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19685 not disallow peeling for vectorized strided loads.
19686 (vect_check_strided_load): Make static and simplify.
19687 (vect_analyze_data_refs): Adjust.
19688 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19689 correctly when vectorizing strided loads.
19691 2013-02-05 Richard Biener <rguenther@suse.de>
19693 * doc/install.texi: Refer to ISL, not PPL.
19695 2013-02-05 Jan Hubicka <jh@suse.cz>
19697 PR tree-optimization/55789
19698 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19700 2013-02-05 Jan Hubicka <jh@suse.cz>
19702 PR tree-optimization/55789
19703 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19704 the dead call anyway.
19706 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19709 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19711 2013-02-04 Alexander Potapenko <glider@google.com>
19712 Jack Howarth <howarth@bromo.med.uc.edu>
19713 Jakub Jelinek <jakub@redhat.com>
19716 * config/darwin.c (sort_ctor_records): Stabilized qsort
19717 on constructor priority by using original position.
19718 (finalize_ctors): New routine to sort constructors by
19719 priority before use in assemble_integer.
19720 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19722 2013-02-04 Jakub Jelinek <jakub@redhat.com>
19725 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19726 about visibility on artificial decls.
19727 * config/sol2.c (solaris_assemble_visibility): Likewise.
19729 2013-02-04 Kai Tietz <ktietz@redhat.com>
19732 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19733 argument and improve checking of return-argument types for 16-byte
19735 (ix86_function_value_1): Add additional valtype argument on call
19736 of function_value_64.
19737 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19738 handling infunction_value_64 function.
19740 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19742 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
19744 2013-02-04 Richard Biener <rguenther@suse.de>
19746 PR tree-optimization/56188
19747 * tree-ssa-structalias.c (label_visit): Consider case with
19748 initially non-empty points-to set.
19749 (perform_var_substitution): Dump node mapping and clean up.
19751 2013-02-04 Richard Guenther <rguenther@suse.de>
19754 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19755 node prevail as last resort.
19756 (lto_symtab_merge_decls): Remove guard on LTRANS here.
19757 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
19759 2013-02-04 Richard Biener <rguenther@suse.de>
19761 PR tree-optimization/56113
19762 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19764 (equiv_class_lookup_or_add): ... this.
19765 (label_visit): Adjust and fix error in previous patch.
19766 (perform_var_substitution): Adjust.
19768 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19770 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19771 * config/sh/sh4-300.md: Likewise.
19772 * config/sh/sh4a.md: Likewise.
19773 * config/sh/constraints.md: Likewise.
19774 * config/sh/sh.md: Likewise.
19775 * config/sh/netbsd-elf.h: Likewise.
19776 * config/sh/predicates.md: Likewise.
19777 * config/sh/sh-protos.h: Likewise.
19778 * config/sh/ushmedia.h: Likewise.
19779 * config/sh/linux.h: Likewise.
19780 * config/sh/sh.c: Likewise.
19781 * config/sh/superh.h: Likewise.
19782 * config/sh/elf.h: Likewise.
19783 * config/sh/sh4.md: Likewise.
19784 * config/sh/sh.h: Likewise.
19786 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19788 * config/pa/constraints.md: Adjust unused letters. Change "T"
19789 constraint to match_test floating_point_store_memory_operand().
19790 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19791 (base14_operand): New.
19792 (floating_point_store_memory_operand): New.
19793 (integer_store_memory_operand): Revise to use base14_operand and
19794 reg_plus_base_memory_operand.
19795 (move_dest_operand): Allow symbolic_memory_operands.
19796 (symbolic_memory_operand): Check for LO_SOM.
19797 (symbolic_operand): Change default case to break.
19798 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19799 CONST_DOUBLE values to be reloaded by putting them into memory when
19800 the destination is a floating point register.
19801 (movdf): Remove code to handle CONST_DOUBLE.
19803 (reload_indf_r1): New.
19804 (reload_insf_r1): New.
19805 Consistently use "Q" and "T" constraints with integer and floating
19806 point move instructions, respectively.
19807 (movdi): Remove FAIL.
19808 Change predicate for source operand unamed DImode move from
19809 general_operand to move_src_operand.
19810 (umulsidi3): Change predicate for destination operand to
19812 Likewise for similar unamed patterns.
19813 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19814 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19815 (hppa_legitimize_address): Simplify mask calculation.
19816 (pa_emit_move_sequence): Revised handling of secondary reloads from
19817 REG+D addresses for floating point loads and stores. Directly handle
19818 loading CONST0_RTX (mode) to a floating point register.
19819 (pa_secondary_reload): Handle reloading DF and SFmode constant values
19820 to floating point registers. Don't restrict secondary reloads to
19821 floating point registers to integer modes. Revise some comments and
19823 (TARGET_LEGITIMATE_ADDRESS_P): Define.
19824 (pa_legitimate_address_p): New.
19825 (pa_legitimize_reload_address): New.
19826 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
19827 (STRICT_REG_OK_FOR_BASE_P): New.
19828 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
19829 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
19831 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
19832 Andrew Dixie <andrewd@gentrack.com>
19834 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
19837 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
19839 * expmed.c (extract_bit_field_1): Pass the full width of the
19840 structure to get_best_reg_extraction_insn.
19842 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19845 * configure.ac (use_cxa_atexit): Add AIX.
19846 * configure: Regenerate.
19848 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
19850 2013-02-01 Jakub Jelinek <jakub@redhat.com>
19853 * final.c (need_profile_function): New variable.
19854 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
19855 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
19856 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
19857 notes, targetm.asm_out.function_prologue doesn't emit anything,
19858 HAVE_prologue and profiler should be emitted before prologue,
19859 set need_profile_function instead of emitting it.
19860 (final_scan_insn): If need_profile_function, emit
19861 profile_function on the first NOTE_INSN_BASIC_BLOCK or
19862 NOTE_INSN_FUNCTION_BEG note.
19864 2013-02-01 Richard Henderson <rth@redhat.com>
19866 * config/rs6000/rs6000.md (smulditi3): New.
19869 * config/alpha/alpha.md (umulditi3): New.
19871 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19873 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
19874 (ASM_OUTPUT_ALIGNED_LOCAL): New.
19876 2013-02-01 Richard Biener <rguenther@suse.de>
19878 PR tree-optimization/56113
19879 * tree-ssa-structalias.c (label_visit): Reduce work for
19880 single-predecessor nodes.
19882 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
19884 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
19885 range isn't testing for zero.
19887 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
19889 PR middle-end/56113
19890 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
19892 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
19893 Nick Clifton <nickc@redhat.com>
19895 * config/v850/constraints.md (Q): Define as a memory constraint.
19896 * config/v850/predicates.md (label_ref_operand): New predicate.
19897 (e3v5_shift_operand): New predicate.
19898 (ior_operator): New predicate.
19899 * config/v850/t-v850: Add e3v5 multilib.
19900 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
19901 (v850_gen_movdi): Prototype.
19902 * config/v850/v850.c: Add support for e3v5 architecture.
19903 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
19905 (construct_save_jarl): Add e3v5 long JARL support.
19906 (v850_adjust_insn_length): New function. Adjust length of call
19907 insns when using e3v5 instructions.
19908 (v850_gen_movdi): New function: Generate instructions to move a
19910 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
19911 (CPP_SPEC): Define __v850e3v5__ as appropriate.
19912 (TARGET_USE_FPU): Enable for e3v5.
19913 (CONST_OK_FOR_W): New macro.
19914 (ADJUST_INSN_LENGTH): Define.
19915 * config/v850/v850.md (UNSPEC_LOOP): Define.
19916 (attr cpu): Add v850e3v5.
19917 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
19918 (movdi): New pattern.
19919 (movdi_internal): New pattern.
19920 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
19921 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
19922 (cstoresf4): Likewise.
19923 (cstoredf4): Likewise.
19924 (insv): New pattern.
19925 (rotlso3_a): New pattern.
19926 (rotlsi3_b): New pattern
19927 (rotlsi3_v850e3v5): New pattern.
19928 (doloop_begin): New pattern.
19929 (fix_loop_counter): New pattern.
19930 (doloop_end): New pattern.
19931 (branch_normal): Add e3v5 long branch support.
19932 (branch_invert): Likewise.
19933 (branch_z_normal): Likewise.
19934 (branch_z_invert): Likewise.
19935 (branch_nz_normal): Likewise.
19936 (branch_nz_invert): Likewise.
19937 (call_internal_short): Add e3v5 register-indirect JARL support.
19938 (call_internal_long): Likewise.
19939 (call_value_internal_short): Likewise.
19940 (call_value_internal_long): Likewise.
19941 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
19942 (mloop): New option.
19943 * config.gcc: Add support for configuring v840e3v5 target.
19944 * doc/invoke.texi: Document new v850 specific command line options.
19946 2013-01-31 Paul Koning <ni1d@arrl.net>
19950 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
19951 children if parent is a class.
19952 (prune_unused_types_prune): Don't add DW_AT_declaration.
19954 2013-01-31 Richard Biener <rguenther@suse.de>
19956 PR tree-optimization/56157
19957 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
19958 match up operand with SLP child.
19960 2013-01-31 Jason Merrill <jason@redhat.com>
19963 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
19964 parameters the first time.
19965 (gen_scheduled_generic_parms_dies): Check completeness here.
19967 2013-01-31 Richard Biener <rguenther@suse.de>
19969 PR middle-end/53073
19970 * common.opt (faggressive-loop-optimizations): New flag,
19971 enabled by default.
19972 * doc/invoke.texi (faggressive-loop-optimizations): Document.
19973 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
19974 infer_loop_bounds_from_undefined by it.
19976 2013-01-31 Richard Biener <rguenther@suse.de>
19978 PR tree-optimization/56150
19979 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
19980 visit virtual operands.
19981 (find_uses_to_rename_bb): Likewise.
19983 2013-01-31 Richard Biener <rguenther@suse.de>
19985 PR tree-optimization/56150
19986 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
19987 mixed store non-store stmts.
19989 2013-01-30 Jakub Jelinek <jakub@redhat.com>
19992 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
19993 LIBASAN_EARLY_SPEC is defined.
19994 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
19995 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
19997 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
20000 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
20001 invalid args instead of ICEing on it.
20002 (ix86_valid_target_attribute_tree): Return error_mark_node if
20003 ix86_valid_target_attribute_inner_p failed.
20004 (ix86_valid_target_attribute_p): Return false only if
20005 ix86_valid_target_attribute_tree returned error_mark_node. Allow
20006 target("default") attribute.
20007 (sorted_attr_string): Change argument from const char * to tree,
20008 merge in all target attribute arguments rather than just one.
20009 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
20010 instead of free. Avoid using strcat.
20011 (ix86_mangle_function_version_assembler_name): Mangle
20012 target("default") as if no target attribute is present. Adjust
20013 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
20014 instead of xmalloc and XDELETEVEC instead of free.
20015 (ix86_function_versions): Don't return true if one of the decls
20016 doesn't have target attribute. If they don't and one of the decls
20017 is DECL_FUNCTION_VERSIONED, report an error. Adjust
20018 sorted_attr_string caller. Use XDELETEVEC instead of free.
20019 (ix86_supports_function_versions): Remove.
20020 (make_name): Fix up formatting.
20021 (make_dispatcher_decl): Remove resolver_name and its initialization.
20022 Avoid leaking memory.
20023 (is_function_default_version): Return true if there is
20024 target("default") attribute rather than no target attribute at all.
20025 (make_resolver_func): Avoid leaking memory.
20026 (ix86_generate_version_dispatcher_body): Likewise.
20027 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
20028 * target.def (supports_function_versions): Remove.
20029 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
20030 * doc/tm.texi: Regenerated.
20032 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
20034 PR rtl-optimization/56144
20035 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
20036 for values with side effects.
20038 2013-01-30 Richard Biener <rguenther@suse.de>
20040 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
20041 (sparseset_pop): Likewise.
20042 * cfganal.c (compute_idf): Likewise. Increase work-stack size
20043 to be able to use quick_push in the worker loop.
20045 2013-01-30 Marek Polacek <polacek@redhat.com>
20047 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
20049 2013-01-30 Richard Biener <rguenther@suse.de>
20052 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
20054 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
20056 PR tree-optimization/56064
20057 * fixed-value.c (fixed_from_double_int): New function.
20058 * fixed-value.h (fixed_from_double_int): New prototype.
20059 (const_fixed_from_double_int): New static inline function.
20060 * fold-const.c (native_interpret_fixed): New static function.
20061 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
20062 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
20063 (native_encode_fixed): New static function.
20064 (native_encode_expr) <FIXED_CST>: Use it.
20065 (native_interpret_int): Move double_int worker code to...
20066 * double-int.c (double_int::from_buffer): ...this new static method.
20067 * double-int.h (double_int::from_buffer): Prototype it.
20069 2013-01-30 Richard Biener <rguenther@suse.de>
20071 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
20072 New pointer-map and obstack.
20073 (init_alias_vars): Allocate pointer-map and obstack.
20074 (delete_points_to_sets): Free them.
20075 (find_what_var_points_to): Cache result.
20076 (find_what_p_points_to): Adjust for changed interface of
20077 find_what_var_points_to.
20078 (compute_points_to_sets): Likewise.
20079 (ipa_pta_execute): Likewise.
20081 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20083 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
20084 * configure: Regenerate.
20085 * config.in: Regenerate.
20086 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
20087 #nobits/#progbits if supported.
20089 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
20092 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
20093 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
20094 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
20096 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20098 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
20099 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
20101 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20103 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
20104 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
20106 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20108 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
20110 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
20111 * config/arm/cortex-a7.md: New bypasses using
20112 arm_mac_accumulator_is_result.
20114 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20116 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
20117 (cortex_a7_neon_mla): Likewise.
20118 (cortex_a7_fpfmad): New reservation.
20119 (cortex_a7_fpmacs): Use ffmas and update required units.
20120 (cortex_a7_fpmuld): Update required units and latency.
20121 (cortex_a7_fpmacd): Likewise.
20122 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
20123 (cortex_a7_neon). Likewise.
20124 (bypass) Update participating units.
20126 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20128 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
20129 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
20131 * config/arm/vfp11.md (vfp_farith): Use ffmas.
20132 (vfp_fmul): Use ffmad.
20133 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
20134 (cortex_r4_fmacd): Use ffmad.
20135 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
20136 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
20137 (cortex_a9_fmacd): Use ffmad.
20138 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
20139 (cortex_a8_vfp_macd): Use ffmad.
20140 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
20141 (cortex_a5_fpmacd): Use ffmad.
20142 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
20143 (cortex_a15_vfp_macd): Use ffmad.
20144 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
20146 2013-01-29 Jason Merrill <jason@redhat.com>
20149 * varasm.c (default_assemble_visibility): Don't warn about
20150 visibility on artificial decls.
20152 2013-01-29 Richard Biener <rguenther@suse.de>
20154 PR tree-optimization/56113
20155 * tree-ssa-structalias.c (equiv_class_lookup): Also return
20157 (label_visit): Free duplicate bitmaps and record the leader instead.
20158 (perform_var_substitution): Adjust.
20160 2013-01-29 Richard Biener <rguenther@suse.de>
20162 PR tree-optimization/55270
20163 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
20164 the CFG, schedule loops for fixup.
20166 2013-01-29 Nick Clifton <nickc@redhat.com>
20168 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
20171 2013-01-28 Leif Ekblad <leif@rdos.net>
20173 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
20174 * config/i386/i386.h (TARGET_RDOS): New macro.
20175 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
20176 * config/i386/i386.c (ix86_option_override_internal): For 64bit
20177 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
20178 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
20179 DEFAULT_LARGE_SECTION_THRESHOLD.
20180 * config/i386/i386.md (R14_REG, R15_REG): New constants.
20181 * config/i386/rdos.h: New file.
20182 * config/i386/rdos64.h: New file.
20184 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
20187 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
20190 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20192 PR rtl-optimization/56117
20193 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
20194 call cselib_lookup_from_insn on the MEM before calling
20195 add_insn_mem_dependence.
20197 2013-01-28 Richard Biener <rguenther@suse.de>
20199 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
20200 to a stmt that didn't have one.
20201 (copy_phis_for_bb): Likewise for PHI arguments.
20202 (copy_debug_stmt): Likewise for debug stmts.
20204 2013-01-28 Richard Biener <rguenther@suse.de>
20206 PR tree-optimization/56034
20207 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
20208 (partition_builtin_p): Adjust.
20209 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
20210 it is the last partition.
20211 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
20212 up the vertex for the definition.
20213 (classify_partition): Classify whether a partition is a
20214 PKIND_REDUCTION, thus has uses outside of the loop.
20215 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
20216 Merge all PKIND_REDUCTION partitions into the last partition.
20217 (tree_loop_distribution): Seed partitions from reductions as well.
20219 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20221 PR tree-optimization/56125
20222 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
20223 pow(x,c) into sqrt(x) * powi(x, n/2) or
20224 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
20225 optimizing for size.
20226 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
20227 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
20230 PR tree-optimization/56094
20231 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
20232 to UNKNOWN_LOCATION while gimplifying expr.
20234 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
20237 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
20238 operand 0 in movabs insn template for -masm=intel asm alternative.
20239 (*movabs<mode>_2): Ditto for operand 1.
20241 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
20244 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
20247 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20249 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
20252 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20254 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
20255 Undef to avoid warning.
20257 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20259 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
20260 * configure: Regenerate.
20262 2013-01-25 Jakub Jelinek <jakub@redhat.com>
20264 PR tree-optimization/56098
20265 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
20266 for stmts with volatile ops.
20267 (cond_store_replacement): Don't optimize if assign has volatile ops.
20268 (cond_if_else_store_replacement_1): Don't optimize if either
20269 then_assign or else_assign have volatile ops.
20270 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
20273 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20275 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
20277 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20279 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
20280 missing ':' in asm example.
20282 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
20284 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
20285 entries into lane and laneq entries.
20286 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
20287 Remove AdvSIMD scalar modes.
20288 (aarch64_sq<r>dmulh_laneq<mode>): New.
20289 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
20291 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
20292 builtin implementations to relfect changes in RTL in aarch64-simd.md.
20293 * config/aarch64/iterators.md (VCOND): New.
20296 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20299 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
20300 Add NULL LIBNAME argument to existing definitions.
20301 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
20302 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
20303 * config/avr/avr.c (DEF_BUILTIN): Same.
20304 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
20305 (avr_expand_builtin): Expand to a vanilla call if a libgcc
20306 implementation is available (DECL_ASSEMBLER_NAME is set).
20307 (avr_fold_absfx): New static function.
20308 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
20309 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
20310 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
20311 AVR_BUILTIN_ABSLLK.
20312 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
20313 (abshk, absk, abslk, absllk): Provide as static inline functions.
20315 2013-01-25 Marek Polacek <polacek@redhat.com>
20317 PR tree-optimization/56035
20318 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
20320 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
20322 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
20323 (*movtf_internal_rex64): Add (!o,C) alternative
20324 (*movxf_internal_rex64): Ditto.
20325 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
20327 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
20329 * doc/invoke.texi: fix typo.
20330 * doc/objc.texi: fix typo.
20332 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
20334 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
20335 for the first two alternatives.
20337 2013-01-24 Diego Novillo <dnovillo@google.com>
20339 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
20340 (ggc-zone.o): Remove.
20341 * configure.ac: Remove option --with-gc.
20342 * configure: Re-generate.
20343 * doc/install.texi: Remove documentation for --with-gc.
20344 * gengtype.c (write_enum_defn): Remove. Update all users.
20345 (write_Types_process_field): Remove generation of gt_e_* argument.
20346 (output_type_enum): Remove. Update all users.
20347 (write_enum_defn): Remove. Update all users.
20348 (enum alloc_zone): Remove. Update all users.
20349 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
20350 * ggc-common.c (ggc_splay_alloc): Remove first argument.
20351 Update all callers.
20352 (struct ptr_data): Remove field TYPE. Update all users.
20353 (gt_pch_note_object): Remove argument TYPE. Update all users.
20354 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
20356 * ggc-none.c (ggc_alloc_typed_stat): Remove.
20357 (struct alloc_zone): Remove.
20358 (ggc_internal_alloc_zone_stat): Remove.
20359 (ggc_internal_cleared_alloc_zone_stat): Remove.
20360 * ggc-page.c (ggc_alloc_typed_stat): Remove.
20361 (ggc_pch_count_object): Remove last argument. Update all users.
20362 (ggc_pch_alloc_object): Remove last argument. Update all users.
20363 (struct alloc_zone): Remove.
20364 * ggc-zone.c: Remove.
20365 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
20366 (struct alloc_zone): Remove.
20367 (ggc_alloc_typed_stat): Remove.
20368 (ggc_alloc_typed): Remove.
20369 (ggc_splay_alloc): Remove first argument.
20370 (rtl_zone): Remove. Update all users.
20371 (tree_zone): Remove. Update all users.
20372 (tree_id_zone): Remove. Update all users.
20373 (ggc_internal_zone_alloc_stat): Remove. Update all users.
20374 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
20375 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
20376 * tree-ssanames.c: Remove references to zone allocator in comments.
20378 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
20380 * config/avr/avr.c (avr_out_fract): Make register numbers that
20381 might be outside of source operand signed.
20383 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
20385 * config/i386/constraints.md (Yf): New constraint.
20386 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
20387 of f constraint to conditionaly disable x87 register preferences.
20388 (*movdf_internal): Ditto.
20389 (*movsf_internal): Ditto.
20391 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
20393 PR inline-asm/55934
20394 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
20395 that have operands with impossible constraints.
20396 Add a FIXME for a speed-up opportunity.
20397 * lra-constraints.c (process_alt_operands): Verify that a class
20398 selected from constraints on asms is valid for the operand mode.
20399 (curr_insn_transform): Remove incorrect comment.
20401 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
20403 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
20404 TOC operand is a valid symbol ref in the constant pool.
20406 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20408 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
20410 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
20413 * config/avr/stdfix.h: New file.
20414 * t-avr (stdfix-gcc.h): New rule to build it.
20415 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
20417 2013-01-23 Kostya Serebryany <kcc@google.com>
20419 * config/darwin.h: remove dependency on
20420 CoreFoundation (asan on Mac OS).
20422 2013-01-23 Jakub Jelinek <jakub@redhat.com>
20425 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
20426 instead of cmpdi_operand for first comparison operand.
20427 Don't assert that comparison operands aren't both constants.
20429 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20431 * doc/install.texi (Downloading the Source): Update references to
20432 downloading separate components.
20434 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20436 * doc/extend.texi (__int128): Improve grammar.
20438 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
20441 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
20442 alternative to (o,r).
20443 (*movdi_internal_rex64): Remove (!o,n) alternative.
20444 (DImode immediate->memory splitter): Remove.
20445 (DImode immediate->memory peephole2): Remove.
20446 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
20447 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
20448 alternative to (!o,*r).
20449 (*movtf_internal_sse): New pattern.
20450 (*movxf_internal_rex64): New pattern.
20451 (*movxf_internal): Disable for TARGET_64BIT.
20452 (*movdf_internal_rex64): Remove (!o,F) alternative.
20454 2013-01-22 Jakub Jelinek <jakub@redhat.com>
20456 PR middle-end/56074
20457 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
20458 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
20459 * tree-vect-loop-manip.c (find_loop_location): Also ignore
20460 stmt locations where LOCATION_LOCUS of the stmt location is
20461 UNKNOWN_LOCATION or BUILTINS_LOCATION.
20464 * config/i386/i386.md (UNSPEC_STOS): New.
20465 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20466 *strsetqi_1): Add UNSPEC_STOS.
20468 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20471 * doc/invoke.texi: Remove left over -Wsynth example.
20473 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20475 PR tree-optimization/56051
20476 * fold-const.c (fold_binary_loc): Don't fold
20477 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20478 a narrowing conversion, or widening conversion from signed
20481 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20483 PR rtl-optimization/56023
20484 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20485 dependent on debug instruction.
20487 2013-01-21 Martin Jambor <mjambor@suse.cz>
20489 PR middle-end/56022
20490 * function.c (allocate_struct_function): Call
20491 invoke_set_current_function_hook earlier.
20493 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20495 * reload1.c (init_reload): Only initialize reload_obstack
20496 during the first call.
20498 2013-01-21 Marek Polacek <polacek@redhat.com>
20500 * cfgloop.c (verify_loop_structure): Fix up grammar.
20502 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20504 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20505 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20507 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20510 * config/arm/marvell-pj4.md: Update copyright year.
20511 Fix up use of alu to alu_reg and simple_alu_imm.
20513 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20515 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20517 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20520 * lra-constraints.c (curr_insn_transform): Don't reuse original
20521 insn for secondary memory move when memory mode should be different.
20523 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20525 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20526 atomic_storedi_1): New patterns.
20528 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20530 btver2 pipeline descriptions.
20531 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20533 * config/i386/i386.md (btver2_decode): New type attributes.
20534 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
20536 * config/i386/btver2.md: New file describing btver2 pipelines.
20538 2013-01-19 Andrew Pinski <apinski@cavium.com>
20540 PR tree-optimization/52631
20541 * tree-ssa-sccvn (visit_use): Before looking up the original
20542 statement, try looking up the simplified expression.
20544 2013-01-19 Anthony Green <green@moxielogic.com>
20546 * config/moxie/moxie.c (moxie_expand_prologue): Set
20547 current_function_static_stack_size.
20549 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20551 PR tree-optimization/56029
20552 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20553 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20555 2013-01-18 Sharad Singhai <singhai@google.com>
20557 PR tree-optimization/55995
20558 * dumpfile.c (dump_loc): Print location only if available.
20559 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
20561 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
20564 * lra-constraints.c (curr_insn_transform): Reuse original insn for
20565 secondary memory move.
20566 (inherit_reload_reg): Use rclass instead of cl for
20567 check_secondary_memory_needed_p.
20569 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20571 PR middle-end/56015
20572 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
20573 the case where writing real complex part of target modifies op1.
20575 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
20577 * config/aarch64/aarch64-simd.md
20578 (aarch64_vcond_internal<mode>): Handle unordered cases.
20579 * config/aarch64/iterators.md (v_cmp_result): New.
20581 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
20582 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20584 * config/arm/marvell-pj4.md: New file.
20585 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
20586 * config/arm/arm.md (generic_sched): Add marvell_pj4.
20587 (generic_vfp): Likewise.
20588 * config/arm/arm-cores.def: Add marvell-pj4.
20589 * config/arm/arm-tune.md: Regenerate.
20590 * config/arm/arm-tables.opt: Regenerate.
20591 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
20592 * doc/invoke.texi: Document marvell-pj4.
20594 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
20596 * config/aarch64/arm_neon.h: Map scalar types to standard types.
20598 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20603 * var-tracking.c (negative_power_of_two_p): New.
20604 (global_get_addr_cache, local_get_addr_cache): New.
20605 (get_addr_from_global_cache, get_addr_from_local_cache): New.
20606 (vt_canonicalize_addr): Rewrite using the above. Adjust the
20608 (vt_stack_offset_p): Remove.
20609 (vt_canon_true_dep): Always canonicalize loc's address.
20610 (clobber_overlapping_mems): Make sure we have a MEM.
20611 (local_get_addr_clear_given_value): New.
20612 (val_reset): Clear local cached entries.
20613 (compute_bb_dataflow): Create and release the local cache.
20614 Disable duplicate MEMs clobbering.
20615 (emit_notes_in_bb): Clobber MEMs likewise.
20616 (vt_emit_notes): Create and release the local cache.
20617 (vt_initialize, vt_finalize): Create and release the global
20618 cache, respectively.
20619 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
20621 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20623 PR libmudflap/53359
20624 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20625 not found in the symtab.
20627 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20630 PR rtl-optimization/55547
20631 PR rtl-optimization/53827
20634 * alias.c (offset_overlap_p): New, factored out of...
20635 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20636 the conservative special case for symbolic constants. Don't
20637 adjust zero sizes on alignment.
20639 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20641 PR rtl-optimization/52573
20642 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20643 REG_UNUSED for the same register.
20645 2013-01-17 Richard Biener <rguenther@suse.de>
20646 Marek Polacek <polacek@redhat.com>
20648 PR rtl-optimization/55833
20649 * loop-unswitch.c (unswitch_loops): Move loop verification...
20650 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20651 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20652 Set it to true when we're removing a loop from hierarchy tree in
20653 an irreducible region.
20654 (fix_bb_placements): Adjust caller.
20655 (fix_loop_placements): Likewise.
20657 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
20659 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20660 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20661 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20662 Remove ID. Adjust comments.
20663 * config/avr/avr-c.c (avr_builtin_name): Remove.
20664 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20665 * config/avr/avr.c (avr_tolower): New static function.
20666 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20667 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20668 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20671 2013-01-17 Jan Hubicka <jh@suse.cz>
20673 PR tree-optimization/55273
20674 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20676 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
20679 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20680 store through atomic_store<mode>_1.
20681 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20683 2013-01-17 Martin Jambor <mjambor@suse.cz>
20685 PR tree-optimizations/55264
20686 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20687 for virtual methods.
20688 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20689 virtual methods before inlining is over.
20690 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20692 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20695 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20697 PR rtl-optimization/56005
20698 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20699 pending reads for prefetch.
20701 2013-01-16 Ian Bolton <ian.bolton@arm.com>
20703 * config/aarch64/aarch64.md
20704 (*cstoresi_neg_uxtw): New pattern.
20705 (*cmovsi_insn_uxtw): New pattern.
20706 (*<optab>si3_uxtw): New pattern.
20707 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20708 (*<optab>si3_insn_uxtw): New pattern.
20709 (*bswapsi2_uxtw): New pattern.
20711 2013-01-16 Richard Biener <rguenther@suse.de>
20713 * tree-inline.c (tree_function_versioning): Remove set but
20714 never used variable.
20716 2013-01-16 Richard Biener <rguenther@suse.de>
20718 PR tree-optimization/55964
20719 * tree-flow.h (rename_variables_in_loop): Remove.
20720 (rename_variables_in_bb): Likewise.
20721 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20722 (copy_loop_before): Adjust and delete update-ssa status.
20723 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
20724 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
20725 (rename_variables_in_loop): Remove.
20726 (slpeel_update_phis_for_duplicate_loop): Likewise.
20727 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20728 use available cfg machinery instead of duplicating it.
20729 Update PHI nodes and perform poor-mans SSA update here.
20730 (slpeel_tree_peel_loop_to_edge): Adjust.
20732 2013-01-16 Richard Biener <rguenther@suse.de>
20734 PR tree-optimization/54767
20735 PR tree-optimization/53465
20736 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20737 (vrp_visit_phi_node): For PHI arguments coming via backedges
20738 drop all symbolical range information.
20739 (execute_vrp): Compute backedges.
20741 2013-01-16 Richard Biener <rguenther@suse.de>
20743 * doc/install.texi: Update CLooG and ISL requirements to
20746 2013-01-16 Christian Bruel <christian.bruel@st.com>
20749 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20750 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20751 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20753 2013-01-16 DJ Delorie <dj@redhat.com>
20755 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20756 (UNSPECV_SP_SWITCH_E): New.
20757 (sp_switch_1): Change to an unspec.
20758 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20761 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
20763 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20764 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20765 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20766 (expand_mem_thread_fence): Ditto.
20767 (expand_mem_signal_fence): Ditto.
20768 (expand_atomic_load): Ditto.
20769 (expand_atomic_store): Ditto.
20771 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
20773 PR rtl-optimization/55547
20774 PR rtl-optimization/53827
20777 * alias.c (memrefs_conflict_p): Set sizes to negative after
20780 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20783 * function.c (thread_prologue_and_epilogue_insns): Always
20784 add crtl->drap_reg to set_up_by_prologue.set, even if
20785 stack_realign_drap is false.
20787 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20789 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20790 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20791 *call): Fix indention.
20793 2013-01-15 Tom de Vries <tom@codesourcery.com>
20796 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20799 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20801 PR rtl-optimization/55153
20802 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20804 2013-01-15 Martin Jambor <mjambor@suse.cz>
20806 PR tree-optimization/55920
20807 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20808 accesses as grp_to_be_debug_replaced.
20810 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20812 PR tree-optimization/55920
20813 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20814 there is non-useless type conversion needed from debug rhs to lhs,
20815 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20817 2013-01-15 Joseph Myers <joseph@codesourcery.com>
20818 Mikael Pettersson <mikpe@it.uu.se>
20821 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
20823 (ASM_OUTPUT_CASE_LABEL): Remove.
20824 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
20825 * final.c (shorten_branches): Update alignment of labels before
20826 jump tables if CASE_VECTOR_SHORTEN_MODE.
20828 2013-01-15 Richard Biener <rguenther@suse.de>
20831 * system.h: Do not include gmp.h for building host tools.
20833 2013-01-15 Richard Biener <rguenther@suse.de>
20835 PR middle-end/55882
20836 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
20837 account for bitpos when computing alignment.
20839 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
20841 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
20842 (ix86_target_macros_internal): Likewise.
20844 * config/i386/i386.c (m_CORE2I7): Removed.
20845 (m_CORE_HASWELL): New macro.
20846 (m_CORE_ALL): Likewise.
20847 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
20848 (initial_ix86_arch_features): Likewise.
20849 (processor_target_table): Initializations for Core avx2.
20850 (cpu_names): New names "core-avx2".
20851 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
20852 PROCESSOR_CORE_HASWELL.
20853 (ix86_issue_rate): New case.
20854 (ia32_multipass_dfa_lookahead): Likewise.
20855 (ix86_sched_init_global): Likewise.
20857 * config/i386/i386.h (TARGET_HASWELL): New macro.
20858 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
20859 (processor_type): New PROCESSOR_HASWELL.
20861 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20863 PR tree-optimization/55955
20864 * tree-vect-loop.c (vectorizable_reduction): Give up early on
20865 *SHIFT_EXPR and *ROTATE_EXPR codes.
20867 PR tree-optimization/48766
20868 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
20869 -ftrapv disable -fwrapv.
20871 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20874 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
20875 etc. to 1 and not to __flash.
20876 Use LL suffix for __INT24_MAX__ with -mint8.
20877 Use ULL suffix for __UINT24_MAX__ with -mint8.
20879 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20881 * config/avr/avr-arch.h
20882 (struct base_arch_s): Use typedef avr_arch_t instead.
20883 (struct arch_info_s): Use typedef avr_arch_info_t instead.
20884 (struct mcu_type_s): Use typedef avr_mcu_t instead.
20885 * config/avr/avr.c: Same.
20886 * config/avr/avr-devices.c: Same.
20887 * config/avr/driver-avr.c: Same.
20888 * config/avr/gen-avr-mmcu-texi.c: Same.
20889 * config/avr/avr-mcus.def: Adjust comment.
20891 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
20893 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
20894 * config/aarch64/iterators.md (VALLDI): New.
20896 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
20897 Andi Kleen <ak@linux.intel.com>
20900 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
20901 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
20904 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20906 * config/avr/avr-stdint.h: Remove trailing blanks.
20907 * config/avr/avr-log.h: Same.
20908 * config/avr/avr-arch.h: Same.
20909 * config/avr/avr-devices.c: Same.
20910 * config/avr/avr-dimode.md: Same.
20911 * config/avr/predicates.md: Same.
20912 * config/avr/avr-c.c: Same. And fix typo.
20914 * config/avr/avr-protos.h: Same. And:
20915 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
20916 (init_cumulative_args): Rename to avr_init_cumulative_args.
20917 (expand_prologue): Rename to avr_expand_prologue.
20918 (expand_epilogue): Rename to avr_expand_epilogue.
20919 (adjust_insn_length): Rename to avr_adjust_insn_length.
20920 (notice_update_cc): Rename to avr_notice_update_cc.
20921 (final_prescan_insn): Rename to avr_final_prescan_insn.
20922 * config/avr/avr.c: Same.
20923 * config/avr/avr.h: Same.
20924 * config/avr/avr.md: Remove trailing blanks.
20925 (prologue): Use avr_expand_prologue.
20926 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
20928 2013-01-14 Richard Biener <rguenther@suse.de>
20930 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
20931 verify_location, collect_subblocks): New functions.
20932 (verify_gimple_in_cfg): Verify that locations only reference
20933 BLOCKs in the functions BLOCK tree.
20935 2013-01-14 Richard Biener <rguenther@suse.de>
20937 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
20939 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
20941 (insert_out_of_ssa_copy_on_edge): Likewise.
20942 (rewrite_close_phi_out_of_ssa): Likewise.
20943 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
20945 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
20946 propagated constants.
20947 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
20950 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20952 * config/avr/avr-modes.def: Add GPL copyright notice.
20954 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
20956 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
20957 MEMMODEL_MASK to determine memory model.
20958 (atomic_store<mode>): Ditto from operands[2].
20959 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
20961 2013-01-13 Jakub Jelinek <jakub@redhat.com>
20964 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
20965 (fold_gimple_assign): Don't call unshare_expr here.
20966 (fold_ctor_reference): Call unshare_expr.
20968 2013-01-13 Terry Guo <terry.guo@arm.com>
20970 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
20971 * doc/fragments.texi: Document MULTILIB_REUSE.
20972 * gcc.c (multilib_reuse): New internal spec.
20973 (set_multilib_dir): Also search multilib from multilib_reuse.
20974 * genmultilib (tmpmultilib3): Refactor code.
20975 (tmpmultilib4): Ditto.
20976 (multilib_reuse): New multilib argument.
20978 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
20980 * Makefile.in: Update copyright.
20982 2013-01-12 Tom de Vries <tom@codesourcery.com>
20984 PR middle-end/55890
20985 * calls.c (expand_call): Check if arg_nr is valid.
20987 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
20989 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
20990 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
20991 documentation. Add missing '__' in front of
20992 __builtin_ia32_packssdw256.
20994 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20997 * config/s390/s390.c (s390_preferred_reload_class): Do not return
20998 NO_REGS for larl operands.
20999 (s390_reload_larl_operand): Use s390_load_address instead of
21002 2013-01-11 Richard Biener <rguenther@suse.de>
21004 * tree-cfg.c (verify_node_sharing_1): Split out from ...
21005 (verify_node_sharing): ... here.
21006 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
21008 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
21010 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
21011 Substitute TREECHECKING.
21012 * configure: Regenerate.
21013 * Makefile.in (TREECHECKING): New.
21015 2013-01-11 Richard Guenther <rguenther@suse.de>
21017 PR tree-optimization/44061
21018 * tree-vrp.c (extract_range_basic): Compute zero as
21019 value-range for __builtin_constant_p of function parameters.
21021 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
21023 Update copyright years.
21025 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
21027 PR rtl-optimization/55672
21028 * lra-eliminations.c (mark_not_eliminable): Permit addition with
21029 const to be eliminable.
21031 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
21033 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
21034 * configure: Regenerate.
21036 2013-01-10 Richard Biener <rguenther@suse.de>
21038 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
21040 2013-01-10 Richard Biener <rguenther@suse.de>
21043 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
21044 locations for virtual PHI arguments.
21045 (rewrite_update_phi_arguments): Likewise.
21047 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
21049 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
21052 2013-01-10 Jakub Jelinek <jakub@redhat.com>
21054 PR tree-optimization/55921
21055 * tree-complex.c (expand_complex_asm): New function.
21056 (expand_complex_operations_1): Call it for GIMPLE_ASM.
21058 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21061 * config/s390/s390.c (s390_symref_operand_p)
21062 (s390_loadrelative_operand_p): Merge the two functions.
21063 (s390_check_qrst_address, print_operand_address): Add parameters
21064 to s390_loadrelative_operand_p invokation.
21065 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
21066 (s390_reload_larl_operand, s390_secondary_reload): Use
21067 s390_loadrelative_operand_p instead of s390_symref_operand_p.
21068 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
21070 2013-01-09 Mike Stump <mikestump@comcast.net>
21072 * dse.c (record_store): Remove unnecessary assert.
21074 2013-01-09 Jan Hubicka <jh@suse.cz>
21076 PR tree-optimization/55569
21077 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
21078 * cfgloop.h (scale_loop_profile): Likewise.
21080 2013-01-09 Jan Hubicka <jh@suse.cz>
21083 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
21085 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
21087 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
21089 PR middle-end/55114
21090 * expr.h (maybe_emit_group_store): Declare.
21091 * expr.c (maybe_emit_group_store): New function.
21092 * builtins.c (expand_builtin_int_roundingfn): Call it.
21093 (expand_builtin_int_roundingfn_2): Likewise.
21095 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
21097 PR rtl-optimization/55829
21098 * lra-constraints.c (match_reload): Add code for absent output.
21099 (curr_insn_transform): Add code for reloads of matched inputs
21102 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
21104 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
21105 attribute of movddup insn to DF.
21106 (*vec_interleave_lowv2df): Ditto.
21107 (vec_dupv2df): Ditto.
21109 2013-01-09 Jan Hubicka <jh@suse.cz>
21111 PR tree-optimiation/55875
21112 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
21113 EVERY_ITERATION parameter.
21114 (number_of_iterations_exit): Check if exit is executed every iteration.
21115 (idx_infer_loop_bounds): Similarly here.
21116 (n_of_executions_at_most): Simplify
21117 to only test for cases where statement is dominated by the
21118 particular bound; handle correctly the "postdominance" test.
21119 (scev_probably_wraps_p): Use max loop iterations info
21120 as a global bound first.
21122 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
21123 Nick Clifton <nickc@redhat.com>
21125 * config/v850/v850.md (cbranchsf4): New pattern.
21126 (cstoresf4): New pattern.
21127 (cbranchdf4): New pattern.
21128 (cstoredf4): New pattern.
21129 (movsicc): Disallow floating point comparisons.
21130 (cmpsf_le_insn): Fix order of operators.
21131 (cmpsf_lt_insn): Likewise.
21132 (cmpsf_eq_insn): Likewise.
21133 (cmpdf_le_insn): Likewise.
21134 (cmpdf_lt_insn): Likewise.
21135 (cmpdf_eq_insn): Likewise.
21136 (cmpsf_ge_insn): Use LE comparison.
21137 (cmpdf_ge_insn): Likewise.
21138 (cmpsf_gt_insn): Use LT comparison.
21139 (cmpdf_gt_insn): Likewise.
21140 (cmpsf_ne_insn): Delete pattern.
21141 (cmpdf_ne_insn): Delete pattern.
21142 * config/v850/v850.c (v850_gen_float_compare): Use
21143 gen_cmpdf_eq_insn for NE comparison.
21144 (v850_float_z_comparison_operator)
21145 (v850_float_nz_comparison_operator): Move from here ...
21146 * config/v850/predicates.md: ... to here. Move GT and GE
21147 comparisons into v850_float_z_comparison_operator.
21148 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
21150 (v850_float_nz_comparison_operator): Likewise.
21152 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21154 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
21155 with calls to gen_insvsi/gen_insvdi.
21157 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21159 * config/i386/i386.c (initial_ix86_tune_features): Set up
21160 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
21162 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
21163 Jakub Jelinek <jakub@redhat.com>
21165 PR tree-optimization/48189
21166 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
21167 If nitercst is 0, don't predict the exit edge.
21169 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21171 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
21172 in asm_fprintf with reg_names.
21173 (aarch64_print_operand_address): Likewise.
21174 (aarch64_return_addr): Likewise.
21175 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
21177 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21179 * config/pa/pa.h (VAL_U6_BITS_P): Define.
21180 (INT_U6_BITS): Likewise.
21181 * config/pa/predicates.md (uint6_operand): New predicate.
21182 (shift5_operand, shift6_operand): Likewise.
21183 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
21185 (lshrdi3): Use shift6_operand.
21186 (shrpsi4, shrpdi4): New insn patterns.
21187 (extzv): Delete expander.
21188 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
21189 predicates in unamed zero extract patterns. Tighten common constraint.
21190 (extv): Delete expander.
21191 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
21192 predicates in unamed sign extract patterns. Tighten common constraint.
21193 (insv): Delete expander.
21194 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
21195 predicates in unamed insert patterns. Tighten common constraint.
21196 Change uint32_operand predicate to uint6_operand predicate in unamed
21197 DImode pattern to insert constant values of type 1...1xxxx.
21199 2013-01-04 Jan Hubicka <jh@suse.cz>
21201 PR tree-optimization/55823
21202 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
21205 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21206 Uros Bizjak <ubizjak@gmail.com>
21208 PR rtl-optimization/55845
21209 * df-problems.c (can_move_insns_across): Stop scanning at
21210 volatile_insn_p source instruction or give up if
21211 across_from .. across_to range contains any volatile_insn_p
21214 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21216 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
21217 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
21219 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
21220 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
21222 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21225 * asan.c (asan_clear_shadow): New function.
21226 (asan_emit_stack_protection): Use it.
21228 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21230 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
21231 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
21232 with tab instead of space.
21234 2013-01-08 Nick Clifton <nickc@redhat.com>
21236 * config/rl78/rl78.c (rl78_expand_prologue): Always select
21237 register bank 0 at the start of an interrupt handler.
21238 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
21241 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21243 * config/aarch64/aarch64-simd.md
21244 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
21245 (aarch64_simd_bsl): Likewise.
21246 (aarch64_vcond_internal<mode>): Likewise.
21247 (vcond<mode><mode>): Likewise.
21248 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
21249 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
21251 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21253 * config/aarch64/aarch64-builtins.c
21254 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
21256 2013-01-08 Martin Jambor <mjambor@suse.cz>
21259 * tree-sra.c (analyze_access_subtree): Return true also after
21260 potentially creating a debug-only replacement.
21262 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21264 PR middle-end/55890
21265 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
21267 PR tree-optimization/54120
21268 * tree-vrp.c (range_fits_type_p): Don't allow
21269 src_precision < precision from signed vr to unsigned_p
21270 if vr->min or vr->max is negative.
21271 (simplify_float_conversion_using_ranges): Test can_float_p
21272 against CODE_FOR_nothing.
21274 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21275 Richard Biener <rguenther@suse.de>
21277 PR middle-end/55851
21278 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
21279 types instead of just INTEGER_TYPE types.
21281 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
21283 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
21286 2013-01-07 Steve Ellcey <sellcey@mips.com>
21289 * config/mips/mips.opt: Change mad to mmad to match documentation.
21291 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21294 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
21295 .progmemx.data now.
21297 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21300 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
21301 (avr_addrspace_t): Add .section_name field.
21302 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
21304 (avr_addrspace): Same. Initialize .section_name. Remove last
21305 NULL entry. Put __memx into .progmemx.data.
21306 (progmem_section_prefix): Remove.
21307 (avr_asm_init_sections): No need to initialize progmem_section.
21308 (avr_asm_named_section): Use avr_addrspace[].section_name to get
21309 section name prefix.
21310 (avr_asm_select_section): Ditto. And use get_unnamed_section to
21311 retrieve the progmem section.
21312 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
21313 boundary to run over avr_addrspace[].
21314 (avr_register_target_pragmas): Ditto.
21316 2013-01-06 Jakub Jelinek <jakub@redhat.com>
21318 * varasm.c (output_constant_def_contents): For asan_protect_global
21319 protected strings, adjust DECL_ALIGN if needed, before testing for
21321 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
21322 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
21324 (output_object_block): For asan protected decls, emit asan padding
21325 after their contents.
21326 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
21327 (asan_finish_file): Test it here instead.
21329 2013-01-07 Nick Clifton <nickc@redhat.com>
21330 Matthias Klose <doko@debian.org>
21331 Doug Kwan <dougkwan@google.com>
21332 H.J. Lu <hongjiu.lu@intel.com>
21335 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
21337 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
21339 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
21341 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
21343 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
21345 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21348 * doc/install.texi (Cross-Compiler-Specific Options): Document
21351 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
21353 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
21354 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
21355 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
21356 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
21357 vqmovun_high_s64): Fix source operand number and update copyright.
21359 2013-01-07 Richard Biener <rguenther@suse.de>
21361 PR middle-end/55890
21362 * gimple.h (gimple_call_builtin_p): New overload.
21363 * gimple.c (validate_call): New function.
21364 (gimple_call_builtin_p): Likewise.
21365 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21366 Use gimple_call_builtin_p.
21367 (find_func_clobbers): Likewise.
21368 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
21369 (strlen_optimize_stmt): Likewise.
21371 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21373 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
21374 (vld1q_dup_*): Likewise.
21375 (vld1_*): Likewise.
21376 (vld1q_*): Likewise.
21377 (vld1_lane_*): Likewise.
21378 (vld1q_lane_*): Likewise.
21380 2013-01-07 Richard Biener <rguenther@suse.de>
21382 * lto-streamer.h (LTO_minor_version): Bump to 2.
21384 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21386 * config/aarch64/aarch64-protos.h
21387 (aarch64_const_double_zero_rtx_p): Rename to...
21388 (aarch64_float_const_zero_rtx_p): ...this.
21389 (aarch64_float_const_representable_p): New.
21390 (aarch64_output_simd_mov_immediate): Likewise.
21391 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
21392 move immediate case.
21393 * config/aarch64/aarch64.c
21394 (aarch64_const_double_zero_rtx_p): Rename to...
21395 (aarch64_float_const_zero_rtx_p): ...this.
21396 (aarch64_print_operand): Allow printing of new constants.
21397 (aarch64_valid_floating_const): New.
21398 (aarch64_legitimate_constant_p): Check for valid floating-point
21400 (aarch64_simd_valid_immediate): Likewise.
21401 (aarch64_vect_float_const_representable_p): New.
21402 (aarch64_float_const_representable_p): Likewise.
21403 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
21404 (aarch64_output_simd_mov_immediate): New.
21405 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
21406 (*movdf_aarch64): Likewise.
21407 * config/aarch64/constraints.md (Ufc): New.
21408 (Y): call aarch64_float_const_zero_rtx.
21409 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
21411 2013-01-07 Richard Biener <rguenther@suse.de>
21413 PR tree-optimization/55888
21414 PR tree-optimization/55862
21415 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
21416 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
21417 not if it is contained therein.
21419 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21421 * config/avr/t-avr: Typo.
21423 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21426 * config/avr/t-avr: Don't automatically rebuild
21427 $(srcdir)/config/avr/t-multilib
21428 $(srcdir)/config/avr/avr-tables.opt
21429 $(srcdir)/doc/avr-mmcu.texi
21430 (avr-mcus): New phony target to build them on request.
21431 (s-avr-mlib, s-avr-mmcu-texi): Remove.
21432 * avr/avr-mcus.def: Adjust comments.
21434 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
21436 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
21438 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21440 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
21442 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21444 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
21446 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
21448 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
21449 to generate profiling.
21450 * config/rs6000/aix64.h (LIB_SPEC): Same.
21452 2013-01-04 Andrew Pinski <apinski@cavium.com>
21454 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
21456 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
21458 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
21460 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
21462 (ix86_expand_move): Ditto.
21463 (ix86_zero_extend_to_Pmode): Ditto.
21464 (ix86_expand_call): Ditto.
21465 (ix86_expand_special_args_builtin): Ditto.
21466 (ix86_expand_builtin): Ditto.
21468 2013-01-04 Richard Biener <rguenther@suse.de>
21470 PR tree-optimization/55862
21471 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21472 translating them through PHI nodes.
21474 2013-01-04 Martin Jambor <mjambor@suse.cz>
21476 PR tree-optimization/55755
21477 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21478 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21480 2013-01-04 Marek Polacek <polacek@redhat.com>
21482 PR middle-end/55859
21483 * opts.c (default_options_optimization): Clarify error message.
21485 2013-01-04 Richard Biener <rguenther@suse.de>
21487 PR middle-end/55863
21488 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21491 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21494 * config/pa/pa.md (movsi): Revert previous change.
21495 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21498 2013-01-03 Richard Henderson <rth@redhat.com>
21500 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21501 after eliminating TLS symbols.
21503 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
21506 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21507 * graphite-poly.c (debug_gmp_value): Likewise.
21509 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
21512 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21513 selected code model, define __code_mode_small__, __code_model_medium__,
21514 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21515 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21516 xchg temporary register with %k. Declare temporary register as
21518 [__x86_64__]: For medium and large code models, preserve %rbx register.
21520 2013-01-03 Richard Biener <rguenther@suse.de>
21522 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
21523 (dump_subscript): Adjust.
21524 (finalize_ddr_dependent): Do not dump redundant info.
21525 (analyze_siv_subscript): Adjust.
21526 (subscript_dependence_tester): Likewise.
21527 (compute_affine_dependence): Likewise.
21529 2013-01-03 Richard Biener <rguenther@suse.de>
21532 2013-01-03 Richard Biener <rguenther@suse.de>
21534 PR tree-optimization/55857
21535 * tree-vect-stmts.c (vectorizable_load): Do not setup
21536 re-alignment for invariant loads.
21538 2013-01-02 Richard Biener <rguenther@suse.de>
21540 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21541 invariant load do not generate a vector load from the scalar location.
21543 2013-01-03 Richard Biener <rguenther@suse.de>
21545 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21546 for not vectorizing.
21547 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21548 not build INDIRECT_REFs, call get_name once only.
21549 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21550 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21552 2013-01-03 Richard Biener <rguenther@suse.de>
21554 PR tree-optimization/55857
21555 * tree-vect-stmts.c (vectorizable_load): Do not setup
21556 re-alignment for invariant loads.
21558 2013-01-03 Richard Biener <rguenther@suse.de>
21561 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
21562 prefer a built-in decl.
21564 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21566 * gcc.c (process_command): Update copyright notice dates.
21567 * gcov.c (print_version): Likewise.
21568 * gcov-dump.c (print_version): Likewise.
21570 PR rtl-optimization/55838
21571 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
21572 iv0.step, iv1.step and step.
21574 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21575 Marc Glisse <marc.glisse@inria.fr>
21577 PR tree-optimization/55832
21578 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
21579 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
21580 integer_{one,zero}_node.
21582 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21585 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
21586 * var-tracking.c (reverse_op): Don't add reverse ops to
21587 VALUEs that have already
21588 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
21590 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
21592 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
21594 2013-01-02 Teresa Johnson <tejohnson@google.com>
21596 * dumpfile.c (dump_loc): Print filename with location.
21597 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
21598 new location_t parameter to emit complete unroll message with
21599 new dump framework.
21600 (canonicalize_loop_induction_variables): Compute loops location
21601 and pass to try_unroll_loop_completely.
21602 * loop-unroll.c (report_unroll_peel): New function.
21603 (peel_loops_completely): Use new dump format with location
21604 for main dumpfile message, and invoke report_unroll_peel on success.
21605 (decide_unrolling_and_peeling): Ditto.
21606 (decide_peel_once_rolling): Remove old dumpfile message subsumed
21607 by report_unroll_peel.
21608 (decide_peel_completely): Ditto.
21609 (decide_unroll_constant_iterations): Ditto.
21610 (decide_unroll_runtime_iterations): Ditto.
21611 (decide_peel_simple): Ditto.
21612 (decide_unroll_stupid): Ditto.
21613 * cfgloop.c (get_loop_location): New function.
21614 * cfgloop.h (get_loop_location): Declare.
21616 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21618 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21621 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21623 PR middle-end/55198
21624 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21625 BLKmode objects when EXPAND_MEMORY is specified.
21627 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21629 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21631 (fold_builtin_cpu): Do not share cpu model decls across statements.
21633 2013-01-02 Jason Merrill <jason@redhat.com>
21636 * tree.c (build_array_type_1): Revert earlier change.
21638 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21640 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21642 * config/aarch64/aarch64-tune.md: Re-generate.
21644 2013-01-02 Richard Biener <rguenther@suse.de>
21646 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21647 invariant load do not generate a vector load from the scalar location.
21649 2013-01-02 Richard Biener <rguenther@suse.de>
21652 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21653 * configure: Regenerate.
21655 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21657 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21658 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21659 (expand_builtin_int_roundingfn_2): Keep the original target around
21660 for the fallback case.
21662 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21664 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21665 to be clear for sign changes.
21667 2013-01-01 Jan Hubicka <jh@suse.cz>
21669 * ipa-inline-analysis.c: Fix formatting.
21671 2013-01-01 Jakub Jelinek <jakub@redhat.com>
21673 PR tree-optimization/55831
21674 * tree-vect-loop.c (get_initial_def_for_induction): Use
21675 gsi_after_labels instead of gsi_start_bb.
21677 Copyright (C) 2013 Free Software Foundation, Inc.
21679 Copying and distribution of this file, with or without modification,
21680 are permitted in any medium without royalty provided the copyright
21681 notice and this notice are preserved.