1 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
4 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
6 pointer elimination with using remove_reg_equal_offset_note.
8 2013-09-12 DJ Delorie <dj@redhat.com>
10 * config/msp430/: New port.
11 * config.gcc (msp430): Added.
12 * doc/invoke.texi: Document MSP430 options.
13 * doc/install.texi: Document msp430-elf
14 * doc/md.texi: Document msp430-elf
15 * doc/contrib.texi: Document msp430-elf
17 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
20 2013-09-12 Martin Jambor <mjambor@suse.cz>
23 * ipa-prop.c (remove_described_reference): Give up if the edge in the
24 reference descriptor is NULL.
25 (ipa_edge_removal_hook): If owning a reference descriptor, set its
28 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
30 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
31 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
32 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
33 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
34 make_temp_ssa_name): move to tree-ssanames.h
35 * tree-ssa-alias.h: Move prototype.
36 * tree-ssa.h: Include tree-ssanames.h.
37 * tree-ssanames.c (FREE_SSANAMES): Move to here.
38 * tree-ssanames.h: New. Move items from tree-flow*.h
39 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
41 2013-09-12 Richard Biener <rguenther@suse.de>
43 PR tree-optimization/58404
44 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
45 propagate non-invariant addresses into dereferences wrapped
46 in component references.
48 2013-09-12 Richard Biener <rguenther@suse.de>
50 PR tree-optimization/58402
51 * passes.def: Move pass_late_warn_uninitialized later.
53 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
55 * tree-ssa.h: New. Move content from tree-flow.h and
57 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
58 Move prototypes belonging to tree-ssa.c.
59 * tree-flow-inline.h (redirect_edge_var_map_def,
60 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
62 * gimple.h: Adjust prototypes.
63 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move to...
64 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
65 * tree.h: Move prototype to tree-ssa.h.
66 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
67 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
68 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
69 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
70 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
71 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
72 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
73 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
74 graphite-blocking.c, graphite-clast-to-gimple.c, graphite-dependences.c,
75 graphite-interchange.c, graphite-optimize-isl.c, graphite-poly.c,
76 graphite-scop-detection.c, graphite-sese-to-poly.c, graphite.c,
77 ipa-cp.c, ipa-inline-analysis.c, ipa-inline-transform.c, ipa-inline.c,
78 ipa-prop.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa-utils.c,
79 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
80 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
81 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
82 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
83 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
84 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
85 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
86 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
87 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
88 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
89 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
90 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
91 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
92 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
93 value-prof.c, var-tracking.c,
94 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
96 2013-09-12 Richard Biener <rguenther@suse.de>
98 PR tree-optimization/58396
99 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
100 (build_rdg): Take a loop-nest parameter, fix memleaks.
101 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
103 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
105 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
108 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
110 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
111 The FMA instruction names should have a 'v' prefix.
113 2013-09-12 Richard Biener <rguenther@suse.de>
115 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
116 (dot_rdg): Use popen instead of system in optional code.
117 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
118 (already_processed_vertex_p): Adjust.
119 (has_anti_or_output_dependence, predecessor_has_mem_write,
120 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
121 rdg_flag_uses): Remove.
122 (rdg_flag_vertex): Simplify.
123 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
125 (build_rdg_partition_for_component): Process the first vertex
127 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
129 2013-09-12 Alan Modra <amodra@gmail.com>
131 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
133 2013-09-11 DJ Delorie <dj@redhat.com>
134 Nick Clifton <nickc@redhat.com>
136 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
137 (rl78_stack_based_mem): New.
138 * config/rl78/constraints.md (Iv08): New.
147 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
149 (movsi): Change from expand to insn-and-split.
150 (ashrsi3): Clobber AX.
154 * config/rl78/rl78.md (CC_REG): Fix.
155 (addsi3): Allow memory and immediate operands.
156 (addsi3_internal): Split into...
157 (addsi3_internal_virt): ...new, and ...
158 (addsi3_internal_real): ...new.
160 (subsi3_internal_virt): New.
161 (subsi3_internal_real): New.
162 (mulsi3): Add memory operand.
163 (mulsi3_rl78): Likewise.
164 (mulsi3_g13): Likewise.
165 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
166 (cbranchqi4_real): Add more constraint options.
167 (cbranchhi4_real): Expand pattern.
168 (cbranchhi4_real_signed): New.
169 (cbranchhi4_real_inverted): New.
170 (cbranchsi4_real_lt): New.
171 (cbranchsi4_real_ge): New.
172 (cbranchsi4_real_signed): New.
173 (cbranchsi4_real): New.
175 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
177 (lshrsi3_virt): Likewise.
178 (ashlsi3_virt): Likewise.
179 (cbranchqi4_virt_signed): New.
180 (cbranchhi4_virt_signed): New.
181 (cbranchsi4_virt): New.
182 * config/rl78/rl78.c: Whitespace fixes throughout.
183 (move_elim_pass): New.
184 (pass_data_rl78_move_elim): New.
185 (pass_rl78_move_elim): New.
186 (make_pass_rl78_move_elim): New.
187 (rl78_devirt_info): Run devirt earlier.
188 (rl78_move_elim_info): New.
189 (rl78_asm_file_start): Register it.
190 (rl78_split_movsi): New.
191 (rl78_as_legitimate_address): Allow virtual base registers when
193 (rl78_addr_space_convert): Remove spurious debug stuff.
194 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
195 (rl78_print_operand): More cases for not printing '#'.
196 (rl78_expand_compare): Remove most of the logic.
197 (content_memory): New.
198 (clear_content_memory): New.
199 (get_content_index): New.
200 (get_content_name): New.
201 (display_content_memory): New.
202 (update_content): New.
203 (record_content): New.
204 (already_contains): New.
205 (insn_ok_now): Re-recog insns with virtual registers.
206 (add_postponed_content_update): New.
207 (process_postponed_content_update): New.
208 (gen_and_emit_move): New.
209 (transcode_memory_rtx): Record new location content.
210 Use gen_and_emit_move.
211 (force_into_acc): New.
212 (move_to_acc): Use gen_and_emit_move.
213 (move_from_acc): Likewise.
214 (move_acc_to_reg): Likewise.
215 (move_to_x): Likewise.
216 (move_to_hl): Likewise.
217 (move_to_de): Likewise.
218 (rl78_alloc_physical_registers_op1): Record location content.
219 (has_constraint): New.
220 (rl78_alloc_physical_registers_op2): Record location content.
222 (rl78_alloc_physical_registers_ro1): Likewise.
223 (rl78_alloc_physical_registers_cmp): Likewise.
224 (rl78_alloc_physical_registers_umul): Likewise.
225 (rl78_alloc_address_registers_macax): New.
226 (rl78_alloc_physical_registers): Initialize and set location
227 content memory as needed.
228 (rl78_reorg): Make sure split2 is called.
229 (rl78_rtx_costs): New.
231 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
233 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
234 for (not (neg ...)) and (neg (not ...)) cases.
236 2013-09-11 Richard Biener <rguenther@suse.de>
239 * passes.def: Split critical edges before late uninit warning passes.
240 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
242 2013-09-11 Jakub Jelinek <jakub@redhat.com>
244 PR tree-optimization/58385
245 * fold-const.c (build_range_check): If both low and high are NULL,
246 use omit_one_operand_loc to preserve exp side-effects.
248 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
250 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
252 2013-09-11 Richard Biener <rguenther@suse.de>
254 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
255 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
256 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
257 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
258 stmts_from_loop, known_dependences_p, build_empty_rdg,
259 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
260 * tree-loop-distribution.c: ... here.
261 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
262 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
263 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
264 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
265 * tree-loop-distribution.c: ... here.
266 * tree-loop-distribution.c: Include gimple-pretty-print.h.
267 (struct partition_s): Add loops member.
268 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
269 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
270 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
272 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
273 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
274 Sergey Lega <sergey.s.lega@intel.com>
275 Anna Tikhonova <anna.tikhonova@intel.com>
276 Ilya Tocar <ilya.tocar@intel.com>
277 Andrey Turetskiy <andrey.turetskiy@intel.com>
278 Ilya Verbin <ilya.verbin@intel.com>
279 Kirill Yukhin <kirill.yukhin@intel.com>
280 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
282 * config/i386/constraints.md (k): New.
284 * config/i386/i386.c (const regclass_map): Add new mask registers.
285 (dbx_register_map): Ditto.
286 (dbx64_register_map): Ditto.
287 (svr4_dbx_register_map): Ditto.
288 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
290 (ix86_preferred_reload_class): Disable constants for mask registers.
291 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
292 (ix86_hard_regno_mode_ok): Support new mask registers.
293 (x86_order_regs_for_local_alloc): Ditto.
294 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
295 (FIXED_REGISTERS): Add new mask registers.
296 (CALL_USED_REGISTERS): Ditto.
297 (REG_ALLOC_ORDER): Ditto.
298 (VALID_MASK_REG_MODE): New.
299 (FIRST_MASK_REG): Ditto.
300 (LAST_MASK_REG): Ditto.
301 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
302 (MAYBE_MASK_CLASS_P): New.
303 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
304 (REG_CLASS_CONTENTS): Ditto.
306 (ANY_MASK_REG_P): Ditto.
307 (HI_REGISTER_NAMES): Add new mask registers.
308 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
309 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
311 (attribute "type"): Add mskmov, msklog.
312 (attribute "length_immediate"): Support them.
313 (attribute "memory"): Ditto.
314 (attribute "prefix_0f"): Ditto.
315 (*movhi_internal): Support new mask registers.
316 (*movqi_internal): Ditto.
317 (define_split): Split out clobber pattern is a logic
318 insn on mask registers.
319 (*k<logic><mode>): New.
320 (*andhi_1): Extend to support mask regs.
321 (*andqi_1): Extend to support mask regs.
323 (define_split): Split and-not to and and not if operands
325 (*<code><mode>_1): Separate HI mode to new pattern...
327 (*<code>qi_1): Extend to support mask regs.
332 (*one_cmpl<mode>2_1): Remove HImode and handle it...
333 (*one_cmplhi2_1): ...Here, now with mask registers support.
334 (*one_cmplqi2_1): Support new mask registers.
335 (HI/QImode arithmetics splitter): Don't split if mask registers
337 (HI/QImode not splitter): Ditto.
338 * config/i386/predicated.md (mask_reg_operand): New.
339 (general_reg_operand): Ditto.
341 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
343 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
344 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
346 2013-09-10 Jeff Law <law@redhat.com>
348 PR tree-optimization/58380
349 * tree-ssa-threadupdate.c (thread_block): Recognize another case
350 of threading through a buried loop header.
352 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
353 return value for single successor case.
355 2013-09-10 Jan Hubicka <jh@suse.cz>
357 * ipa-devirt.c (ipa_devirt): Enable with LTO.
359 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
362 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
363 support conditional execution.
364 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
366 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
368 * lra.c (lra): Clear lra_optional_reload_pseudos before every
370 * lra-constraints.c (curr_insn_transform): Switch on optional
371 reloads. Check destination too to check move insn.
372 (undo_optional_reloads): Add check that the original peudo did not
373 changed its allocation and the optional reload was inherited on
374 last inheritance pass. Break loop after deciding to keep optional
376 (lra_undo_inheritance): Add check that inherited pseudo still in
379 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
381 * config/aarch64/aarch64.md (generic_sched): New.
382 * config/aarch64/aarch64-generic.md (load): Make conditional
383 on generic_sched attribute.
386 2013-09-10 Jan Hubicka <jh@suse.cz>
388 * lto-cgraph.c: Include ipa-utils.h.
389 (compute_ltrans_boundary): Also add possible targets into the boundary.
391 2013-09-10 Jan Hubicka <jh@suse.cz>
393 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
394 VAR_DECL of vtable rather than full expression.
396 2013-09-10 Jan Hubicka <jh@suse.cz>
397 Paolo Carlini <paolo.carlini@oracle.com>
399 * cgraphunit.c (analyze_functions): Save input_location, set it
400 to UNKNOWN_LOCATION and restore it at the end.
402 2013-09-10 Martin Jambor <mjambor@suse.cz>
404 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
405 represented by a thunk.
407 2013-09-10 Jeff Law <law@redhat.com>
409 PR tree-optimization/58343
410 * tree-ssa-threadupdate.c (thread_block): Identify and disable
411 jump threading requests through loop headers buried in the middle
412 of a jump threading path.
414 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
415 in return value/type.
417 2013-09-10 Jakub Jelinek <jakub@redhat.com>
419 PR rtl-optimization/58365
420 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
421 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
424 2013-09-10 Richard Biener <rguenther@suse.de>
426 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
427 * tree-data-ref.c (create_rdg_vertices): Collect all data
428 references, signal failure to the caller, use data-ref API.
429 (build_rdg): Compute data references only once. Maintain lifetime
430 of data references and data dependences from within RDG.
431 (free_rdg): Free dependence relations.
432 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
433 inventing extra dependences.
434 (distribute_loop): Update for RDG API changes.
436 2013-09-10 Kai Tietz <ktietz@redhat.com>
438 * doc/invoke.texi (fms-extensions): Document changed
439 behavior for ms-abi targets.
440 * config/i386/i386.c (ix86_option_override_internal):
441 Set default value of option -fms-extension for ms-abi targets.
443 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
445 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
447 2013-09-10 Alan Modra <amodra@gmail.com>
450 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
452 2013-09-10 Alan Modra <amodra@gmail.com>
454 * config/rs6000/predicates.md (add_cint_operand): New.
455 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
456 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
457 using add_cint_operand.
458 (largetoc_high_plus_aix): Likewise.
460 2013-09-09 Jakub Jelinek <jakub@redhat.com>
462 PR tree-optimization/58364
463 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
464 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
465 the current range can't be an unconditional true or false.
467 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
469 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
471 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
473 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
475 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
478 * doc/invoke.texi (-Wdelete-incomplete): Document it.
480 2013-09-09 Ian Bolton <ian.bolton@arm.com>
482 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
483 NO_REGS for immediate that can't be moved directly into FP_REGS.
485 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
487 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
488 comparison with negated operand.
489 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
492 2013-09-09 Richard Biener <rguenther@suse.de>
495 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
496 of a subloop record all its block as affecting loop-closed SSA form.
498 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
500 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
501 of an rtx/bitpos pair.
502 (store_fixed_bit_field): Update accordingly.
504 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
506 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
508 * builtins.c (expand_errno_check): Likewise.
509 * dwarf2cfi.c (init_return_column_size): Likewise.
510 * except.c (sjlj_mark_call_sites): Likewise.
511 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
512 * lra-constraints.c (emit_inc): Likewise.
513 * ree.c (combine_set_extension): Likewise.
514 * regmove.c (fixup_match_2): Likewise.
515 * reload1.c (inc_for_reload): Likewise.
517 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
519 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
520 (force_to_mode, simplify_shift_const_1, simplify_comparison):
521 Use gen_int_mode with the mode of the associated simplify_* call.
522 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
523 * expmed.c (expand_shift_1): Likewise.
524 * function.c (instantiate_virtual_regs_in_insn): Likewise.
525 * loop-iv.c (iv_number_of_iterations): Likewise.
526 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
527 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
529 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
531 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
532 of the associated expand_* call.
533 (asan_emit_stack_protection): Likewise.
534 * builtins.c (round_trampoline_addr): Likewise.
535 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
536 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
537 (emit_store_flag): Likewise.
538 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
539 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
541 * function.c (instantiate_virtual_regs_in_insn): Likewise.
542 * ifcvt.c (noce_try_store_flag_constants): Likewise.
543 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
544 * modulo-sched.c (generate_prolog_epilog): Likewise.
545 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
546 (expand_ctz, expand_ffs, expand_unop): Likewise.
548 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
550 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
551 of the associated gen_rtx_* call.
552 * caller-save.c (init_caller_save): Likewise.
553 * combine.c (find_split_point, make_extraction): Likewise.
554 (make_compound_operation): Likewise.
555 * dwarf2out.c (mem_loc_descriptor): Likewise.
556 * explow.c (plus_constant, probe_stack_range): Likewise.
557 * expmed.c (expand_mult_const): Likewise.
558 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
559 * reload1.c (init_reload): Likewise.
560 * valtrack.c (cleanup_auto_inc_dec): Likewise.
561 * var-tracking.c (adjust_mems): Likewise.
562 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
563 rather than gen_rtx_fmt_ee.
565 2013-09-09 Jan Hubicka <jh@suse.cz>
568 * value-prof.c (gimple_ic): Copy also abnormal edges.
570 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
572 * asan.c (asan_shadow_cst): Use gen_int_mode.
574 2013-09-08 Jan Hubicka <jh@suse.cz>
576 * ipa-profile.c: Add toplevel comment.
577 (ipa_propagate_frequency_1): Be more conservative when profile is read.
578 (contains_hot_call_p): New function.
579 (ipa_propagate_frequency): Set frequencies based on counts when
581 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
582 profile; do not tamper with profile after inlining if it is read.
584 2013-09-08 Jan Hubicka <jh@suse.cz>
586 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
589 2013-09-08 Jan Hubicka <jh@suse.cz>
591 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
594 2013-09-08 Jeff Law <law@redhat.com>
597 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
600 2013-09-08 Andi Kleen <ak@linux.intel.com>
602 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
604 2013-09-08 Jan Hubicka <jh@suse.cz>
606 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
608 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
610 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
612 * lra.c (new_insn_reg): Take the containing insn as a parameter.
613 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
614 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
617 2013-09-08 Jan Hubicka <jh@suse.cz>
619 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
620 targets and devirtualize to BUILT_IN_UNREACHABLE.
621 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
622 * ipa.c (walk_polymorphic_call_targets): New function.
623 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
624 functions; use the new timevar.
625 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
626 was removed from the program.
627 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
628 not consider it in the walk when its vtable is dead.
629 (possible_polymorphic_call_targets_1): Pass anonymous flag to
631 (devirt_variable_node_removal_hook): New function.
632 (possible_polymorphic_call_targets): Also register
633 devirt_variable_node_removal_hook.
634 (ipa_devirt): Do not do non-speculative devirtualization.
635 (gate_ipa_devirt): One execute if devirtualizing speculatively.
637 2013-09-08 Jan Hubicka <jh@suse.cz>
639 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
640 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
641 varpool_remove_variable_insertion_hook): Declare.
642 * varpool.c (varpool_node_hook_list): New structure.
643 (first_varpool_node_removal_hook,
644 first_varpool_variable_insertion_hook): New variables.
645 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
646 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
647 varpool_remove_variable_insertion_hook,
648 varpool_call_variable_insertion_hooks): New functions.
649 (varpool_remove_node): Use it.
651 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
654 * diagnostic.c (diagnostic_build_prefix): When s.file is
655 "<built-in>" don't output line and column numbers.
657 2013-09-06 Jan Hubicka <jh@suse.cz>
659 * cgraphunit.c (expand_thunk): Get body before touching arguments.
660 * lto-streamer-out.c: Stream thunks, too.
661 * lto-streamer-in.c (input_function): Pop cfun here
662 (lto_read_body): Instead of here.
664 2013-09-06 Caroline Tice <cmtice@google.com>
666 * doc/install.texi: Add documentation for the --enable-vtable-verify
667 and the --disable-libvtv configure options.
669 2013-09-06 Jeff Law <law@redhat.com>
671 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
672 edge implied equivalences into successor phis.
674 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
676 * resource.c (mark_referenced_resources): Handle COND_EXEC.
678 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
680 * resource.c (mark_target_live_regs): Compute resources taking
681 into account if a call is predicated or not.
683 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
685 * toplev.c (output_stack_usage): Be prepared for suffixes created by
686 the compiler in the function names.
688 2013-09-06 Jan Hubicka <jh@suse.cz>
691 * ipa-inline.c (has_caller_p): New function.
692 (want_inline_function_to_all_callers_p): Use it.
693 (sum_callers, inline_to_all_callers): Break out from ...
694 (ipa_inline): ... here.
696 2013-09-06 Jan Hubicka <jh@suse.cz>
698 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
699 only when AVX is enabled.
701 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
703 * config/aarch64/aarch64.md
704 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
705 is fpsimd_<load/store>2.
706 (load_pair<mode>): Likewise.
707 (store_pair<mode>): Likewise.
709 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
711 * config/arm/types.md (type): Add "mrs" type.
712 * config/aarch64/aarch64.md
713 (aarch64_load_tp_hard): Make type "mrs".
715 (load_tp_hard): Make type "mrs".
716 * config/arm/cortex-a15.md: Update with new attributes.
717 * config/arm/cortex-a5.md: Update with new attributes.
718 * config/arm/cortex-a53.md: Update with new attributes.
719 * config/arm/cortex-a7.md: Update with new attributes.
720 * config/arm/cortex-a8.md: Update with new attributes.
721 * config/arm/cortex-a9.md: Update with new attributes.
722 * config/arm/cortex-m4.md: Update with new attributes.
723 * config/arm/cortex-r4.md: Update with new attributes.
724 * config/arm/fa526.md: Update with new attributes.
725 * config/arm/fa606te.md: Update with new attributes.
726 * config/arm/fa626te.md: Update with new attributes.
727 * config/arm/fa726te.md: Update with new attributes.
729 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
731 * config/aarch64/aarch64.md
732 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
733 (*movtf_aarch64): Likewise.
735 (thumb1_movdi_insn): Use "multiple" for type where more than one
736 instruction is used for a move.
737 (*arm32_movhf): Likewise.
738 (*thumb_movdf_insn): Likewise.
740 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
742 * config/arm/types.md (type): Rename fcpys to fmov.
744 (*arm_movsi_vfp): Rename type fcpys as fmov.
745 (*thumb2_movsi_vfp): Likewise
746 (*movhf_vfp_neon): Likewise
747 (*movhf_vfp): Likewise
748 (*movsf_vfp): Likewise
749 (*thumb2_movsf_vfp): Likewise
750 (*movsfcc_vfp): Likewise
751 (*thumb2_movsfcc_vfp): Likewise
752 * config/aarch64/aarch64-simd.md
753 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
754 * config/aarch64/aarch64.md
755 (*movsi_aarch64): Replace type mov_reg with fmovs.
756 (*movdi_aarch64): Likewise
757 (*movsf_aarch64): Likewise
758 (*movdf_aarch64): Likewise
760 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
761 * config/arm/iwmmxt.md
762 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
763 * config/arm/arm1020e.md: Update with new attributes.
764 * config/arm/cortex-a15-neon.md: Update with new attributes.
765 * config/arm/cortex-a5.md: Update with new attributes.
766 * config/arm/cortex-a53.md: Update with new attributes.
767 * config/arm/cortex-a7.md: Update with new attributes.
768 * config/arm/cortex-a8-neon.md: Update with new attributes.
769 * config/arm/cortex-a9.md: Update with new attributes.
770 * config/arm/cortex-m4-fpu.md: Update with new attributes.
771 * config/arm/cortex-r4f.md: Update with new attributes.
772 * config/arm/marvell-pj4.md: Update with new attributes.
773 * config/arm/vfp11.md: Update with new attributes.
775 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
777 * config/aarch64/aarch64.md
778 (*madd<mode>): Fix type attribute.
779 (*maddsi_uxtw): Likewise.
780 (*msub<mode>): Likewise.
781 (*msubsi_uxtw): Likewise.
782 (<su_optab>maddsidi4): Likewise.
783 (<su_optab>msubsidi4): Likewise.
785 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
787 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
788 * config/arm/arm.md (core_cycles): Remove fdiv.
790 (*sqrtsf2_vfp): Update for attribute changes.
791 (*sqrtdf2_vfp): Likewise.
792 * config/aarch64/aarch64.md:
793 (sqrt<mode>2): Update for attribute changes.
794 * config/arm/arm1020e.md: Update with new attributes.
795 * config/arm/cortex-a15-neon.md: Update with new attributes.
796 * config/arm/cortex-a5.md: Update with new attributes.
797 * config/arm/cortex-a53.md: Update with new attributes.
798 * config/arm/cortex-a7.md: Update with new attributes.
799 * config/arm/cortex-a8-neon.md: Update with new attributes.
800 * config/arm/cortex-a9.md: Update with new attributes.
801 * config/arm/cortex-m4-fpu.md: Update with new attributes.
802 * config/arm/cortex-r4f.md: Update with new attributes.
803 * config/arm/marvell-pj4.md: Update with new attributes.
804 * config/arm/vfp11.md: Update with new attributes.
806 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
808 * config/arm/types.md
809 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
810 * config/aarch64/aarch64.md
811 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
813 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
814 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
815 (float<GPI:mode><GPF:mode>2): Likewise.
817 (*truncsisf2_vfp): Update with new attributes.
818 (*truncsidf2_vfp): Likewise.
819 (fixuns_truncsfsi2): Likewise.
820 (fixuns_truncdfsi2): Likewise.
821 (*floatsisf2_vfp): Likewise.
822 (*floatsidf2_vfp): Likewise.
823 (floatunssisf2): Likewise.
824 (floatunssidf2): Likewise.
825 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
826 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
827 * config/arm/arm1020e.md: Update with new attributes.
828 * config/arm/cortex-a15-neon.md: Update with new attributes.
829 * config/arm/cortex-a5.md: Update with new attributes.
830 * config/arm/cortex-a53.md: Update with new attributes.
831 * config/arm/cortex-a7.md: Update with new attributes.
832 * config/arm/cortex-a8-neon.md: Update with new attributes.
833 * config/arm/cortex-a9.md: Update with new attributes.
834 * config/arm/cortex-m4-fpu.md: Update with new attributes.
835 * config/arm/cortex-r4f.md: Update with new attributes.
836 * config/arm/marvell-pj4.md: Update with new attributes.
837 * config/arm/vfp11.md: Update with new attributes.
839 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
841 * config/aarch64/arm_neon.h
842 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
843 (vqtbx<1,2,3,4><q>_s8): Likewise.
845 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
847 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
848 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
849 (add<mode>3): Add type attribute.
850 (add<mode>3): Likewise.
851 (usadd<mode>3): Likewise.
852 (ssadd<mode>3): Likewise.
853 (sub<mode>3): Likewise.
854 (sub<mode>3): Likewise.
855 (ussub<mode>3): Likewise.
856 (sssub<mode>3): Likewise.
857 (ssmulsa3): Likewise.
858 (usmulusa3): Likewise.
859 (arm_usatsihi): Likewise.
861 (*movdi_vfp): Add types for all instructions.
862 (*movdi_vfp_cortexa8): Likewise.
863 (*movhf_vfp_neon): Likewise.
864 (*movhf_vfp): Likewise.
865 (*movdf_vfp): Likewise.
866 (*thumb2_movdf_vfp): Likewise.
867 (*thumb2_movdfcc_vfp): Likewise.
868 * config/arm/arm.md: Add type attribute to all insn patterns.
869 (*thumb1_adddi3): Add type attribute.
870 (*arm_adddi3): Likewise.
871 (*adddi_sesidi_di): Likewise.
872 (*adddi_zesidi_di): Likewise.
873 (*thumb1_addsi3): Likewise.
874 (addsi3_compare0): Likewise.
875 (*addsi3_compare0_scratch): Likewise.
876 (*compare_negsi_si): Likewise.
877 (cmpsi2_addneg): Likewise.
878 (*addsi3_carryin_<optab>): Likewise.
879 (*addsi3_carryin_alt2_<optab>): Likewise.
880 (*addsi3_carryin_clobercc_<optab>): Likewise.
881 (*subsi3_carryin): Likewise.
882 (*subsi3_carryin_const): Likewise.
883 (*subsi3_carryin_compare): Likewise.
884 (*subsi3_carryin_compare_const): Likewise.
885 (*arm_subdi3): Likewise.
886 (*thumb_subdi3): Likewise.
887 (*subdi_di_zesidi): Likewise.
888 (*subdi_di_sesidi): Likewise.
889 (*subdi_zesidi_di): Likewise.
890 (*subdi_sesidi_di): Likewise.
891 (*subdi_zesidi_ze): Likewise.
892 (thumb1_subsi3_insn): Likewise.
893 (*arm_subsi3_insn): Likewise.
894 (*anddi3_insn): Likewise.
895 (*anddi_zesidi_di): Likewise.
896 (*anddi_sesdi_di): Likewise.
897 (*ne_zeroextracts): Likewise.
898 (*ne_zeroextracts): Likewise.
899 (*ite_ne_zeroextr): Likewise.
900 (*ite_ne_zeroextr): Likewise.
901 (*anddi_notdi_di): Likewise.
902 (*anddi_notzesidi): Likewise.
903 (*anddi_notsesidi): Likewise.
904 (andsi_notsi_si): Likewise.
905 (thumb1_bicsi3): Likewise.
906 (*iordi3_insn): Likewise.
907 (*iordi_zesidi_di): Likewise.
908 (*iordi_sesidi_di): Likewise.
909 (*thumb1_iorsi3_insn): Likewise.
910 (*xordi3_insn): Likewise.
911 (*xordi_zesidi_di): Likewise.
912 (*xordi_sesidi_di): Likewise.
913 (*arm_xorsi3): Likewise.
914 (*andsi_iorsi3_no): Likewise.
916 (*smax_m1): Likewise.
917 (*arm_smax_insn): Likewise.
919 (*arm_smin_insn): Likewise.
920 (*arm_umaxsi3): Likewise.
921 (*arm_uminsi3): Likewise.
922 (*minmax_arithsi): Likewise.
923 (*minmax_arithsi_): Likewise.
924 (*satsi_<SAT:code>): Likewise.
925 (arm_ashldi3_1bit): Likewise.
926 (arm_ashrdi3_1bit): Likewise.
927 (arm_lshrdi3_1bit): Likewise.
928 (*arm_negdi2): Likewise.
929 (*thumb1_negdi2): Likewise.
930 (*arm_negsi2): Likewise.
931 (*thumb1_negsi2): Likewise.
932 (*negdi_extendsid): Likewise.
933 (*negdi_zero_extend): Likewise.
934 (*arm_abssi2): Likewise.
935 (*thumb1_abssi2): Likewise.
936 (*arm_neg_abssi2): Likewise.
937 (*thumb1_neg_abss): Likewise.
938 (one_cmpldi2): Likewise.
939 (extend<mode>di2): Likewise.
940 (*compareqi_eq0): Likewise.
941 (*arm_extendhisi2addsi): Likewise.
942 (*arm_movdi): Likewise.
943 (*thumb1_movdi_insn): Likewise.
944 (*arm_movt): Likewise.
945 (*thumb1_movsi_insn): Likewise.
946 (pic_add_dot_plus_four): Likewise.
947 (pic_add_dot_plus_eight): Likewise.
948 (tls_load_dot_plus_eight): Likewise.
949 (*thumb1_movhi_insn): Likewise.
950 (*thumb1_movsf_insn): Likewise.
951 (*movdf_soft_insn): Likewise.
952 (*thumb_movdf_insn): Likewise.
953 (cbranchsi4_insn): Likewise.
954 (cbranchsi4_scratch): Likewise.
955 (*negated_cbranchsi4): Likewise.
956 (*tbit_cbranch): Likewise.
957 (*tlobits_cbranch): Likewise.
958 (*tstsi3_cbranch): Likewise.
959 (*cbranchne_decr1): Likewise.
960 (*addsi3_cbranch): Likewise.
961 (*addsi3_cbranch_scratch): Likewise.
962 (*arm_cmpdi_insn): Likewise.
963 (*arm_cmpdi_unsig): Likewise.
964 (*arm_cmpdi_zero): Likewise.
965 (*thumb_cmpdi_zero): Likewise.
966 (*deleted_compare): Likewise.
967 (*mov_scc): Likewise.
968 (*mov_negscc): Likewise.
969 (*mov_notscc): Likewise.
970 (*cstoresi_eq0_thumb1_insn): Likewise.
971 (cstoresi_nltu_thumb1): Likewise.
972 (cstoresi_ltu_thu): Likewise.
973 (thumb1_addsi3_addgeu): Likewise.
974 (*arm_jump): Likewise.
975 (*thumb_jump): Likewise.
976 (*check_arch2): Likewise.
977 (arm_casesi_internal): Likewise.
978 (thumb1_casesi_dispatch): Likewise.
979 (*arm_indirect_jump): Likewise.
980 (*thumb1_indirect_jump): Likewise.
982 (*and_scc): Likewise.
983 (*ior_scc): Likewise.
984 (*compare_scc): Likewise.
985 (*cond_move): Likewise.
986 (*cond_arith): Likewise.
987 (*cond_sub): Likewise.
988 (*cmp_ite0): Likewise.
989 (*cmp_ite1): Likewise.
990 (*cmp_and): Likewise.
991 (*cmp_ior): Likewise.
992 (*ior_scc_scc): Likewise.
993 (*ior_scc_scc_cmp): Likewise.
994 (*and_scc_scc): Likewise.
995 (*and_scc_scc_cmp): Likewise.
996 (*and_scc_scc_nod): Likewise.
998 (movcond_addsi): Likewise.
1000 (*ifcompare_plus_move): Likewise.
1001 (*if_plus_move): Likewise.
1002 (*ifcompare_move_plus): Likewise.
1003 (*if_move_plus): Likewise.
1004 (*ifcompare_arith_arith): Likewise.
1005 (*if_arith_arith): Likewise.
1006 (*ifcompare_arith_move): Likewise.
1007 (*if_arith_move): Likewise.
1008 (*ifcompare_move_arith): Likewise.
1009 (*if_move_arith): Likewise.
1010 (*ifcompare_move_not): Likewise.
1011 (*if_move_not): Likewise.
1012 (*ifcompare_not_move): Likewise.
1013 (*if_not_move): Likewise.
1014 (*ifcompare_shift_move): Likewise.
1015 (*if_shift_move): Likewise.
1016 (*ifcompare_move_shift): Likewise.
1017 (*if_move_shift): Likewise.
1018 (*ifcompare_shift_shift): Likewise.
1019 (*ifcompare_not_arith): Likewise.
1020 (*ifcompare_arith_not): Likewise.
1021 (*if_arith_not): Likewise.
1022 (*ifcompare_neg_move): Likewise.
1023 (*if_neg_move): Likewise.
1024 (*ifcompare_move_neg): Likewise.
1025 (*if_move_neg): Likewise.
1026 (prologue_thumb1_interwork): Likewise.
1027 (*cond_move_not): Likewise.
1028 (*sign_extract_onebit): Likewise.
1029 (*not_signextract_onebit): Likewise.
1030 (stack_tie): Likewise.
1031 (align_4): Likewise.
1032 (align_8): Likewise.
1033 (consttable_end): Likewise.
1034 (consttable_1): Likewise.
1035 (consttable_2): Likewise.
1036 (consttable_4): Likewise.
1037 (consttable_8): Likewise.
1038 (consttable_16): Likewise.
1039 (*thumb1_tablejump): Likewise.
1040 (prefetch): Likewise.
1041 (force_register_use): Likewise.
1042 (thumb_eh_return): Likewise.
1043 (load_tp_hard): Likewise.
1044 (load_tp_soft): Likewise.
1045 (tlscall): Likewise.
1046 (*arm_movtas_ze): Likewise.
1047 (*arm_rev): Likewise.
1048 (*arm_revsh): Likewise.
1049 (*arm_rev16): Likewise.
1050 * config/arm/thumb2.md
1051 (*thumb2_smaxsi3): Likewise.
1052 (*thumb2_sminsi3): Likewise.
1053 (*thumb32_umaxsi3): Likewise.
1054 (*thumb2_uminsi3): Likewise.
1055 (*thumb2_negdi2): Likewise.
1056 (*thumb2_abssi2): Likewise.
1057 (*thumb2_neg_abss): Likewise.
1058 (*thumb2_movsi_insn): Likewise.
1059 (tls_load_dot_plus_four): Likewise.
1060 (*thumb2_movhi_insn): Likewise.
1061 (*thumb2_mov_scc): Likewise.
1062 (*thumb2_mov_negs): Likewise.
1063 (*thumb2_mov_negs): Likewise.
1064 (*thumb2_mov_nots): Likewise.
1065 (*thumb2_mov_nots): Likewise.
1066 (*thumb2_movsicc_): Likewise.
1067 (*thumb2_movsfcc_soft_insn): Likewise.
1068 (*thumb2_indirect_jump): Likewise.
1069 (*thumb2_and_scc): Likewise.
1070 (*thumb2_ior_scc): Likewise.
1071 (*thumb2_ior_scc_strict_it): Likewise.
1072 (*thumb2_cond_move): Likewise.
1073 (*thumb2_cond_arith): Likewise.
1074 (*thumb2_cond_ari): Likewise.
1075 (*thumb2_cond_sub): Likewise.
1076 (*thumb2_negscc): Likewise.
1077 (*thumb2_movcond): Likewise.
1078 (thumb2_casesi_internal): Likewise.
1079 (thumb2_casesi_internal_pic): Likewise.
1080 (*thumb2_alusi3_short): Likewise.
1081 (*thumb2_mov<mode>_shortim): Likewise.
1082 (*thumb2_addsi_short): Likewise.
1083 (*thumb2_subsi_short): Likewise.
1084 (thumb2_addsi3_compare0): Likewise.
1085 (*thumb2_cbz): Likewise.
1086 (*thumb2_cbnz): Likewise.
1087 (*thumb2_one_cmplsi2_short): Likewise.
1088 (*thumb2_negsi2_short): Likewise.
1089 (*orsi_notsi_si): Likewise.
1090 * config/arm/arm1020e.md: Update with new attributes.
1091 * config/arm/arm1026ejs.md: Update with new attributes.
1092 * config/arm/arm1136jfs.md: Update with new attributes.
1093 * config/arm/arm926ejs.md: Update with new attributes.
1094 * config/arm/cortex-a15.md: Update with new attributes.
1095 * config/arm/cortex-a5.md: Update with new attributes.
1096 * config/arm/cortex-a53.md: Update with new attributes.
1097 * config/arm/cortex-a7.md: Update with new attributes.
1098 * config/arm/cortex-a8.md: Update with new attributes.
1099 * config/arm/cortex-a9.md: Update with new attributes.
1100 * config/arm/cortex-m4.md: Update with new attributes.
1101 * config/arm/cortex-r4.md: Update with new attributes.
1102 * config/arm/fa526.md: Update with new attributes.
1103 * config/arm/fa606te.md: Update with new attributes.
1104 * config/arm/fa626te.md: Update with new attributes.
1105 * config/arm/fa726te.md: Update with new attributes.
1107 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1109 * config/aarch64/aarch64-simd.md
1110 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
1111 <vwx> iterator to ensure correct register choice.
1112 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1113 (aarch64_sqdmull_n<mode>): Likewise.
1114 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1115 * config/aarch64/arm_neon.h
1116 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
1117 (vml<as><q>_n_<su>16): Likewise.
1118 (vml<as>l_high_lane<q>_<su>16): Likewise.
1119 (vml<as>l_high_n_<su>16): Likewise.
1120 (vml<as>l_lane<q>_<su>16): Likewise.
1121 (vml<as>l_n_<su>16): Likewise.
1122 (vmul<q>_lane<q>_<su>16): Likewise.
1123 (vmul<q>_n_<su>16): Likewise.
1124 (vmull_lane<q>_<su>16): Likewise.
1125 (vmull_n_<su>16): Likewise.
1126 (vmull_high_lane<q>_<su>16): Likewise.
1127 (vmull_high_n_<su>16): Likewise.
1128 (vqrdmulh<q>_n_s16): Likewise.
1130 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
1132 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
1133 have the correct lane parameter.
1135 2013-09-06 Richard Biener <rguenther@suse.de>
1137 * cfganal.c (control_dependences::~control_dependences):
1138 Properly free all of the vector.
1140 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
1143 * config/i386/i386.c (ix86_conditional_register_usage):
1144 Proper initialize extended SSE registers.
1146 2013-09-06 Jan Hubicka <jh@suse.cz>
1148 PR tree-optimization/58311
1149 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
1151 2013-09-06 Jan Hubicka <jh@suse.cz>
1153 * Makefile.in (tree-sra.o): Update dependencies.
1154 * tree-sra.c: Include ipa-utils.h
1155 (scan_function): Use recursive_call_p.
1156 (has_caller_p): New function.
1157 (cgraph_for_node_and_aliases): Count also callers of aliases.
1159 2013-09-06 Jan Hubicka <jh@suse.cz>
1162 * cgraph.h (symtab_semantically_equivalent_p): Declare.
1163 * tree-tailcall.c: Include ipa-utils.h.
1164 (find_tail_calls): Use it.
1165 * ipa-pure-const.c (check_call): Likewise.
1166 * ipa-utils.c (recursive_call_p): New function.
1167 * ipa-utils.h (recursive_call_p): Dclare.
1168 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
1169 (symtab_semantically_equivalent_p): New function.
1170 * Makefile.in (tree-tailcall.o): Update dependencies.
1172 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1174 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
1177 2013-09-06 Richard Biener <rguenther@suse.de>
1179 * lto-streamer.h (lto_global_var_decls): Remove.
1180 * Makefile.in (OBJS): Remove lto-symtab.o.
1181 (lto-symtab.o): Remove.
1182 (GTFILES): Remove lto-symtab.c
1183 * lto-symtab.c: Move to lto/
1185 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1187 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
1188 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
1189 (UNSPEC_FPINT_RINT): New constant definitions.
1190 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
1191 definition with 2 attributes.
1192 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
1193 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
1196 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1198 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
1199 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
1200 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
1201 attribute to "z196_alone".
1202 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
1205 2013-09-06 Richard Biener <rguenther@suse.de>
1207 * basic-block.h (class control_dependences): New.
1208 * tree-ssa-dce.c (control_dependence_map): Remove.
1210 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
1211 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
1212 find_pdom, find_control_dependence, find_all_control_dependences):
1214 (mark_control_dependent_edges_necessary,
1215 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
1216 tree_dce_done, perform_tree_ssa_dce): Adjust.
1217 * cfganal.c (set_control_dependence_map_bit,
1218 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
1219 find_all_control_dependences): Move from tree-ssa-dce.c and
1220 implement as methods of control_dependences class.
1221 (control_dependences::control_dependences): New.
1222 (control_dependences::~control_dependences): Likewise.
1223 (control_dependences::get_edges_dependent_on): Likewise.
1224 (control_dependences::get_edge): Likewise.
1226 2013-09-04 Jan Hubicka <jh@suse.cz>
1228 * tree.c (types_same_for_odr): Drop overactive check.
1229 * ipa-devirt.c (hash_type_name): Likewise.
1231 2013-09-04 Jan Hubicka <jh@suse.cz>
1233 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
1234 (analyze_functions): ... here.
1236 2013-09-04 Jan Hubicka <jh@suse.cz>
1239 * cgraphunit.c (analyze_functions): Clear AUX fields
1240 after processing; initialize assembler name has.
1242 2013-09-05 Jeff Law <law@redhat.com>
1244 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
1245 from thread_around_empty_block. Record threading path into PATH.
1246 Recurse if threading through the initial block is successful.
1247 (thread_across_edge): Corresponding changes to slightly simplify.
1249 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1251 * config/aarch64/aarch64.md
1252 (type): Remove frecpe, frecps, frecpx.
1253 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
1254 fix to be a TARGET_SIMD instruction.
1255 (aarch64_frecps): Remove.
1256 * config/aarch64/aarch64-simd.md
1257 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
1258 (aarch64_frecps<mode>): Handle all float/vector of float modes.
1260 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1261 Sofiane Naci <sofiane.naci@arm.com>
1263 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
1264 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
1265 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
1266 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
1267 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
1268 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
1269 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
1270 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
1271 into "clz, "rbit". Rename "shift" to "shift_imm".
1272 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
1273 changes. Update for attribute changes all occurrences of arlo_* and
1275 * config/arm/arm-fixed.md: Update for attribute changes
1276 all occurrences of arlo_* types.
1277 * config/arm/thumb2.md: Update for attribute changes all occurrences
1279 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
1280 (cortexa7_older_only): Likewise.
1281 (cortexa7_younger): Likewise.
1282 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
1283 (1020alu_shift_op): Likewise.
1284 (1020alu_shift_reg_op): Likewise.
1285 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
1286 (alu_shift_op): Likewise.
1287 (alu_shift_reg_op): Likewise.
1288 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
1289 (11_alu_shift_op): Likewise.
1290 (11_alu_shift_reg_op): Likewise.
1291 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
1292 (9_alu_shift_reg_op): Likewise.
1293 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
1295 (cortex_a15_alu_shift): Likewise.
1296 (cortex_a15_alu_shift_reg): Likewise.
1297 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
1299 (cortex_a5_alu_shift): Likewise.
1300 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
1302 (cortex_a53_alu_shift): Likewise.
1303 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
1305 (cortex_a7_alu_reg): Likewise.
1306 (cortex_a7_alu_shift): Likewise.
1307 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
1309 (cortex_a8_alu_shift): Likewise.
1310 (cortex_a8_alu_shift_reg): Likewise.
1311 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
1312 (cortex_a9_dp_shift): Likewise.
1313 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
1315 * config/arm/cortex-r4.md
1316 (cortex_r4_alu): Update for attribute changes.
1317 (cortex_r4_mov): Likewise.
1318 (cortex_r4_alu_shift_reg): Likewise.
1319 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
1320 (526_alu_shift_op): Likewise.
1321 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
1322 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
1323 (626te_alu_shift_op): Likewise.
1324 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
1325 (726te_alu_shift_op): Likewise.
1326 (726te_alu_shift_reg_op): Likewise.
1327 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
1328 (mp626_alu_shift_op): Likewise.
1329 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
1330 (pj4_alu_conds): Likewise.
1331 (pj4_shift): Likewise.
1332 (pj4_shift_conds): Likewise.
1333 (pj4_alu_shift): Likewise.
1334 (pj4_alu_shift_conds): Likewise.
1335 * config/aarch64/aarch64.md: Update for attribute change
1336 all occurrences of arlo_* and shift* types.
1338 2013-09-05 Mike Stump <mikestump@comcast.net>
1340 * tree.h: Move documentation for tree_function_decl to tree-core.h
1341 with the declaration.
1343 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
1346 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
1347 looking for widest mode.
1349 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
1351 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
1353 2013-09-05 Richard Biener <rguenther@suse.de>
1355 PR tree-optimization/58137
1356 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
1357 Do not create vectors of pointers.
1358 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
1359 types for the components of the vector initializer.
1360 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
1361 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
1363 2013-09-05 Martin Jambor <mjambor@suse.cz>
1365 * ipa-prop.c (remove_described_reference): Accept missing references,
1366 return false if that hppens, otherwise return true.
1367 (cgraph_node_for_jfunc): New function.
1368 (try_decrement_rdesc_refcount): Likewise.
1369 (try_make_edge_direct_simple_call): Use them.
1370 (ipa_edge_removal_hook): Remove references from rdescs.
1371 (ipa_edge_duplication_hook): Clone rdescs and their references
1372 when the new edge has the same caller as the old one.
1373 * cgraph.c (cgraph_resolve_speculation): Remove speculative
1374 reference before removing any edges.
1376 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
1378 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
1380 * thumb2.md (thumb2_storewb_parisi): New pattern.
1382 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
1384 * config/aarch64/aarch64-option-extensions.def: Add
1385 AARCH64_OPT_EXTENSION of 'crc'.
1386 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
1387 (AARCH64_ISA_CRC): Ditto.
1388 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
1389 description of the CRC extension.
1391 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
1393 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
1394 * config/rs6000/linux.h: Ditto.
1395 * alpha/linux.h: Ditto.
1396 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
1397 no_c99_libc_has_function.
1398 * config/c6x/uclinux-elf.h: Ditto.
1399 * config/lm32/uclinux-elf.h: Ditto.
1400 * config/m68k/uclinux.h: Ditto.
1401 * config/moxie/uclinux.h: Ditto.
1402 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
1403 (crisv32-*-linux*, cris-*-linux*): Ditto.
1404 * config/bfin/bfin.c: Include "tm_p.h".
1406 2013-09-05 Richard Biener <rguenther@suse.de>
1408 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
1409 check for a definition without a basic-block.
1411 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1412 Sofiane Naci <sofiane.naci@arm.com>
1414 * config/aarch64/aarch64.md
1415 (*movti_aarch64): Rename r_2_f and f_2_r.
1416 (*movsf_aarch64): Likewise.
1417 (*movdf_aarch64): Likewise.
1418 (*movtf_aarch64): Likewise.
1419 (aarch64_movdi_<mode>low): Likewise.
1420 (aarch64_movdi_<mode>high): Likewise.
1421 (aarch64_mov<mode>high_di): Likewise.
1422 (aarch64_mov<mode>low_di): Likewise.
1423 (aarch64_movtilow_tilow): Likewise.
1424 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
1425 values to config/arm/types.md
1426 (attribute "conds"): Update for attribute change.
1427 (anddi3_insn): Likewise.
1428 (iordi3_insn): Likewise.
1429 (xordi3_insn): Likewise.
1430 (one_cmpldi2): Likewise.
1431 * config/arm/types.md (type): Add Neon types.
1432 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
1433 use "type" attribute.
1434 (movmisalign<mode>_neon_store): Likewise.
1435 (movmisalign<mode>_neon_load): Likewise.
1436 (vec_set<mode>_internal): Likewise.
1437 (vec_setv2di_internal): Likewise.
1438 (vec_extract<mode>): Likewise.
1439 (vec_extractv2di): Likewise.
1440 (add<mode>3_neon): Likewise.
1441 (adddi3_neon): Likewise.
1442 (sub<mode>3_neon): Likewise.
1443 (subdi3_neon): Likewise.
1444 (mul<mode>3_neon): Likewise.
1445 (mul<mode>3add<mode>_neon): Likewise.
1446 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
1447 (fma<VCVTF:mode>4)): Likewise.
1448 (fma<VCVTF:mode>4_intrinsic): Likewise.
1449 (fmsub<VCVTF:mode>4)): Likewise.
1450 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1451 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1452 (ior<mode>3): Likewise.
1453 (and<mode>3): Likewise.
1454 (anddi3_neon): Likewise.
1455 (orn<mode>3_neon): Likewise.
1456 (orndi3_neon): Likewise.
1457 (bic<mode>3_neon): Likewise.
1458 (bicdi3_neon): Likewise.
1459 (xor<mode>3): Likewise.
1460 (one_cmpl<mode>2): Likewise.
1461 (abs<mode>2): Likewise.
1462 (neg<mode>2): Likewise.
1463 (umin<mode>3_neon): Likewise.
1464 (umax<mode>3_neon): Likewise.
1465 (smin<mode>3_neon): Likewise.
1466 (smax<mode>3_neon): Likewise.
1467 (vashl<mode>3): Likewise.
1468 (vashr<mode>3_imm): Likewise.
1469 (vlshr<mode>3_imm): Likewise.
1470 (ashl<mode>3_signed): Likewise.
1471 (ashl<mode>3_unsigned): Likewise.
1472 (neon_load_count): Likewise.
1473 (ashldi3_neon_noclobber): Likewise.
1474 (signed_shift_di3_neon): Likewise.
1475 (unsigned_shift_di3_neon): Likewise.
1476 (ashrdi3_neon_imm_noclobber): Likewise.
1477 (lshrdi3_neon_imm_noclobber): Likewise.
1478 (widen_ssum<mode>3): Likewise.
1479 (widen_usum<mode>3): Likewise.
1480 (quad_halves_<code>v4si): Likewise.
1481 (quad_halves_<code>v4sf): Likewise.
1482 (quad_halves_<code>v8hi): Likewise.
1483 (quad_halves_<code>v16qi): Likewise.
1484 (reduc_splus_v2di): Likewise.
1485 (neon_vpadd_internal<mode>): Likewise.
1486 (neon_vpsmin<mode>): Likewise.
1487 (neon_vpsmax<mode>): Likewise.
1488 (neon_vpumin<mode>): Likewise.
1489 (neon_vpumax<mode>): Likewise.
1490 (ss_add<mode>_neon): Likewise.
1491 (us_add<mode>_neon): Likewise.
1492 (ss_sub<mode>_neon): Likewise.
1493 (us_sub<mode>_neon): Likewise.
1494 (neon_vadd<mode>_unspec): Likewise.
1495 (neon_vaddl<mode>): Likewise.
1496 (neon_vaddw<mode>): Likewise.
1497 (neon_vhadd<mode>): Likewise.
1498 (neon_vqadd<mode>): Likewise.
1499 (neon_vaddhn<mode>): Likewise.
1500 (neon_vmul<mode>): Likewise.
1501 (neon_vmla<mode>): Likewise.
1502 (neon_vmlal<mode>): Likewise.
1503 (neon_vmls<mode>): Likewise.
1504 (neon_vmlsl<mode>): Likewise.
1505 (neon_vqdmulh<mode>): Likewise.
1506 (neon_vqdmlal<mode>): Likewise.
1507 (neon_vqdmlsl<mode>): Likewise.
1508 (neon_vmull<mode>): Likewise.
1509 (neon_vqdmull<mode>): Likewise.
1510 (neon_vsub<mode>_unspec): Likewise.
1511 (neon_vsubl<mode>): Likewise.
1512 (neon_vsubw<mode>): Likewise.
1513 (neon_vqsub<mode>): Likewise.
1514 (neon_vhsub<mode>): Likewise.
1515 (neon_vsubhn<mode>): Likewise.
1516 (neon_vceq<mode>): Likewise.
1517 (neon_vcge<mode>): Likewise.
1518 (neon_vcgeu<mode>): Likewise.
1519 (neon_vcgt<mode>): Likewise.
1520 (neon_vcgtu<mode>): Likewise.
1521 (neon_vcle<mode>): Likewise.
1522 (neon_vclt<mode>): Likewise.
1523 (neon_vcage<mode>): Likewise.
1524 (neon_vcagt<mode>): Likewise.
1525 (neon_vtst<mode>): Likewise.
1526 (neon_vabd<mode>): Likewise.
1527 (neon_vabdl<mode>): Likewise.
1528 (neon_vaba<mode>): Likewise.
1529 (neon_vabal<mode>): Likewise.
1530 (neon_vmax<mode>): Likewise.
1531 (neon_vmin<mode>): Likewise.
1532 (neon_vpaddl<mode>): Likewise.
1533 (neon_vpadal<mode>): Likewise.
1534 (neon_vpmax<mode>): Likewise.
1535 (neon_vpmin<mode>): Likewise.
1536 (neon_vrecps<mode>): Likewise.
1537 (neon_vrsqrts<mode>): Likewise.
1538 (neon_vqabs<mode>): Likewise.
1539 (neon_vqneg<mode>): Likewise.
1540 (neon_vcls<mode>): Likewise.
1541 (clz<mode>2): Likewise.
1542 (popcount<mode>2): Likewise.
1543 (neon_vrecpe): Likewise.
1544 (neon_vrsqrte): Likewise.
1545 (neon_vget_lane<mode>_sext_internal): Likewise.
1546 (neon_vget_lane<mode>_zext_internal): Likewise.
1547 (neon_vdup_n<mode>): Likewise.
1548 (neon_vdup_nv2di): Likewise.
1549 (neon_vdpu_lane<mode>_internal): Likewise.
1550 (neon_vswp<mode>): Likewise.
1551 (float<mode><V_cvtto>2): Likewise.
1552 (floatuns<mode><V_cvtto>2): Likewise.
1553 (fix_trunc<mode><V_cvtto>)2): Likewise
1554 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1555 (neon_vcvt<mode>): Likewise.
1556 (neon_vcvtv4sfv4hf): Likewise.
1557 (neon_vcvtv4hfv4sf): Likewise.
1558 (neon_vcvt_n<mode>): Likewise.
1559 (neon_vmovn<mode>): Likewise.
1560 (neon_vqmovn<mode>): Likewise.
1561 (neon_vqmovun<mode>): Likewise.
1562 (neon_vmovl<mode>): Likewise.
1563 (neon_vmul_lane<mode>): Likewise.
1564 (neon_vmull_lane<mode>): Likewise.
1565 (neon_vqdmull_lane<mode>): Likewise.
1566 (neon_vqdmulh_lane<mode>): Likewise.
1567 (neon_vmla_lane<mode>): Likewise.
1568 (neon_vmlal_lane<mode>): Likewise.
1569 (neon_vqdmlal_lane<mode>): Likewise.
1570 (neon_vmls_lane<mode>): Likewise.
1571 (neon_vmlsl_lane<mode>): Likewise.
1572 (neon_vqdmlsl_lane<mode>): Likewise.
1573 (neon_vext<mode>): Likewise.
1574 (neon_vrev64<mode>): Likewise.
1575 (neon_vrev32<mode>): Likewise.
1576 (neon_vrev16<mode>): Likewise.
1577 (neon_vbsl<mode>_internal): Likewise.
1578 (neon_vshl<mode>): Likewise.
1579 (neon_vqshl<mode>): Likewise.
1580 (neon_vshr_n<mode>): Likewise.
1581 (neon_vshrn_n<mode>): Likewise.
1582 (neon_vqshrn_n<mode>): Likewise.
1583 (neon_vqshrun_n<mode>): Likewise.
1584 (neon_vshl_n<mode>): Likewise.
1585 (neon_vqshl_n<mode>): Likewise.
1586 (neon_vqshlu_n<mode>): Likewise.
1587 (neon_vshll_n<mode>): Likewise.
1588 (neon_vsra_n<mode>): Likewise.
1589 (neon_vsri_n<mode>): Likewise.
1590 (neon_vsli_n<mode>): Likewise.
1591 (neon_vtbl1v8qi): Likewise.
1592 (neon_vtbl2v8qi): Likewise.
1593 (neon_vtbl3v8qi): Likewise.
1594 (neon_vtbl4v8qi): Likewise.
1595 (neon_vtbx1v8qi): Likewise.
1596 (neon_vtbx2v8qi): Likewise.
1597 (neon_vtbx3v8qi): Likewise.
1598 (neon_vtbx4v8qi): Likewise.
1599 (neon_vtrn<mode>_internal): Likewise.
1600 (neon_vzip<mode>_internal): Likewise.
1601 (neon_vuzp<mode>_internal): Likewise.
1602 (neon_vld1<mode>): Likewise.
1603 (neon_vld1_lane<mode>): Likewise.
1604 (neon_vld1_dup<mode>): Likewise.
1605 (neon_vld1_dupv2di): Likewise.
1606 (neon_vst1<mode>): Likewise.
1607 (neon_vst1_lane<mode>): Likewise.
1608 (neon_vld2<mode>): Likewise.
1609 (neon_vld2_lane<mode>): Likewise.
1610 (neon_vld2_dup<mode>): Likewise.
1611 (neon_vst2<mode>): Likewise.
1612 (neon_vst2_lane<mode>): Likewise.
1613 (neon_vld3<mode>): Likewise.
1614 (neon_vld3qa<mode>): Likewise.
1615 (neon_vld3qb<mode>): Likewise.
1616 (neon_vld3_lane<mode>): Likewise.
1617 (neon_vld3_dup<mode>): Likewise.
1618 (neon_vst3<mode>): Likewise.
1619 (neon_vst3qa<mode>): Likewise.
1620 (neon_vst3qb<mode>): Likewise.
1621 (neon_vst3_lane<mode>): Likewise.
1622 (neon_vld4<mode>): Likewise.
1623 (neon_vld4qa<mode>): Likewise.
1624 (neon_vld4qb<mode>): Likewise.
1625 (neon_vld4_lane<mode>): Likewise.
1626 (neon_vld4_dup<mode>): Likewise.
1627 (neon_vst4<mode>): Likewise.
1628 (neon_vst4qa<mode>): Likewise.
1629 (neon_vst4qb<mode>): Likewise.
1630 (neon_vst4_lane<mode>): Likewise.
1631 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1632 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1633 (neon_vec_<US>mult_lo_<mode>): Likewise.
1634 (neon_vec_<US>mult_hi_<mode>): Likewise.
1635 (neon_vec_<US>shiftl_<mode>): Likewise.
1636 (neon_unpack<US>_<mode>): Likewise.
1637 (neon_vec_<US>mult_<mode>): Likewise.
1638 (vec_pack_trunc_<mode>): Likewise.
1639 (neon_vec_pack_trunk_<mode>): Likewise.
1640 (neon_vabd<mode>_2): Likewise.
1641 (neon_vabd<mode>_3): Likewise.
1642 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1643 (thumb2_movsi_vfp): Likewise.
1644 (movdi_vfp): Likewise.
1645 (movdi_vfp_cortexa8): Likewise.
1646 (movhf_vfp_neon): Likewise.
1647 (movhf_vfp): Likewiwse.
1648 (movsf_vfp): Likewiwse.
1649 (thumb2_movsf_vfp): Likewiwse.
1650 (movdf_vfp): Likewise.
1651 (thumb2_movdf_vfp): Likewise.
1652 (movsfcc_vfp): Likewise.
1653 (thumb2_movsfcc_vfp): Likewise.
1654 (movdfcc_vfp): Likewise.
1655 (thumb2_movdfcc_vfp): Likewise.
1656 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1657 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1658 (v10_v2c): Likewise.
1659 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1661 (cortex_a15_neon_int_2): Likewise.
1662 (cortex_a15_neon_int_3): Likewise.
1663 (cortex_a15_neon_int_4): Likewise.
1664 (cortex_a15_neon_int_5): Likewise.
1665 (cortex_a15_neon_vqneg_vqabs): Likewise.
1666 (cortex_a15_neon_vmov): Likewise.
1667 (cortex_a15_neon_vaba): Likewise.
1668 (cortex_a15_neon_vaba_qqq): Likewise.
1669 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1670 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1671 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1672 scalar_64_32_long_scalar): Likewise.
1673 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1674 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1675 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1676 lotype_qdd_64_32_long): Likewise.
1677 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1678 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1679 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1680 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1681 (cortex_a15_neon_shift_1): Likewise.
1682 (cortex_a15_neon_shift_2): Likewise.
1683 (cortex_a15_neon_shift_3): Likewise.
1684 (cortex_a15_neon_vshl_ddd): Likewise.
1685 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1686 (cortex_a15_neon_vsra_vrsra): Likewise.
1687 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1688 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1689 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1690 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1691 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1692 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1693 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1694 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1695 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1696 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1697 (cortex_a15_neon_bp_simple): Likewise.
1698 (cortex_a15_neon_bp_2cycle): Likewise.
1699 (cortex_a15_neon_bp_3cycle): Likewise.
1700 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1701 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1702 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1703 (cortex_a15_neon_vld2_4_regs): Likewise.
1704 (cortex_a15_neon_vld3_vld4): Likewise.
1705 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1706 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1707 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1708 (cortex_a15_neon_vst3_vst4): Likewise.
1709 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1710 (cortex_a15_neon_vld3_vld4_lane" 10
1711 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1712 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1713 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1714 (cortex_a15_neon_ldm_2): Likewise.0
1715 (cortex_a15_neon_stm_2): Likewise.
1716 (cortex_a15_neon_mcr): Likewise.
1717 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1718 (cortex_a15_neon_mrc): Likewise.
1719 (cortex_a15_neon_mrrc): Likewise.
1720 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1722 (cortex_a15_alu_shift): Likewise.
1723 (cortex_a15_alu_shift_reg): Likewise.
1724 (cortex_a15_mult32): Likewise.
1725 (cortex_a15_mult64): Likewise.
1726 (cortex_a15_block): Likewise.
1727 (cortex_a15_branch): Likewise.
1728 (cortex_a15_load1): Likewise.
1729 (cortex_a15_load3): Likewise.
1730 (cortex_a15_store1): Likewise.
1731 (cortex_a15_store3): Likewise.
1732 (cortex_a15_call): Likewise.
1733 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
1734 (cortex_a5_f2r): Likewise.
1735 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1737 (cortex_a53_f2r): Likewise.
1738 * config/arm/cortex-a7.md
1739 (cortex_a7_branch): Update for attribute change.
1740 (cortex_a7_call): Likewise.
1741 (cortex_a7_alu_imm): Likewise.
1742 (cortex_a7_alu_reg): Likewise.
1743 (cortex_a7_alu_shift): Likewise.
1744 (cortex_a7_mul): Likewise.
1745 (cortex_a7_load1): Likewise.
1746 (cortex_a7_store1): Likewise.
1747 (cortex_a7_load2): Likewise.
1748 (cortex_a7_store2): Likewise.
1749 (cortex_a7_load3): Likewise.
1750 (cortex_a7_store3): Likewise.
1751 (cortex_a7_load4): Likewise.
1752 (cortex_a7_store4): Likewise.
1753 (cortex_a7_fpalu): Likewise.
1754 (cortex_a7_fconst): Likewise.
1755 (cortex_a7_fpmuls): Likewise.
1756 (cortex_a7_neon_mul): Likewise.
1757 (cortex_a7_fpmacs): Likewise.
1758 (cortex_a7_neon_mla: Likewise.
1759 (cortex_a7_fpmuld: Likewise.
1760 (cortex_a7_fpmacd: Likewise.
1761 (cortex_a7_fpfmad: Likewise.
1762 (cortex_a7_fdivs: Likewise.
1763 (cortex_a7_fdivd: Likewise.
1764 (cortex_a7_r2f: Likewise.
1765 (cortex_a7_f2r: Likewise.
1766 (cortex_a7_f_flags: Likewise.
1767 (cortex_a7_f_loads: Likewise.
1768 (cortex_a7_f_loadd: Likewise.
1769 (cortex_a7_f_stores: Likewise.
1770 (cortex_a7_f_stored: Likewise.
1771 (cortex_a7_neon): Likewise.
1772 * config/arm/cortex-a8-neon.md
1773 (cortex_a8_neon_mrc): Update for attribute change.
1774 (cortex_a8_neon_mrrc): Likewise.
1775 (cortex_a8_neon_int_1): Likewise.
1776 (cortex_a8_neon_int_2): Likewise.
1777 (cortex_a8_neon_int_3): Likewise.
1778 (cortex_a8_neon_int_4): Likewise.
1779 (cortex_a8_neon_int_5): Likewise.
1780 (cortex_a8_neon_vqneg_vqabs): Likewise.
1781 (cortex_a8_neon_vmov): Likewise.
1782 (cortex_a8_neon_vaba): Likewise.
1783 (cortex_a8_neon_vaba_qqq): Likewise.
1784 (cortex_a8_neon_vsma): Likewise.
1785 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1786 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1787 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1788 long_scalar): Likewise.
1789 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1790 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1791 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1792 long_scalar_qdd_64_32_long): Likewise.
1793 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1794 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1795 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1796 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1797 (cortex_a8_neon_shift_1): Likewise.
1798 (cortex_a8_neon_shift_2): Likewise.
1799 (cortex_a8_neon_shift_3): Likewise.
1800 (cortex_a8_neon_vshl_ddd): Likewise.
1801 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1802 (cortex_a8_neon_vsra_vrsra): Likewise.
1803 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1804 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1805 (cortex_a8_neon_fp_vsum): Likewise.
1806 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1807 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1808 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1809 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1810 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1811 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1812 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1813 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1814 (cortex_a8_neon_bp_simple): Likewise.
1815 (cortex_a8_neon_bp_2cycle): Likewise.
1816 (cortex_a8_neon_bp_3cycle): Likewise.
1817 (cortex_a8_neon_ldr): Likewise.
1818 (cortex_a8_neon_str): Likewise.
1819 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1820 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1821 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1822 (cortex_a8_neon_vld2_4_regs): Likewise.
1823 (cortex_a8_neon_vld3_vld4): Likewise.
1824 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1825 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1826 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1827 (cortex_a8_neon_vst3_vst4): Likewise.
1828 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1829 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1830 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1831 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1832 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1833 (cortex_a8_neon_mcr): Likewise.
1834 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1835 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
1836 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
1838 (ca9_neon_mrrc): Likewise.
1839 (cortex_a9_neon_int_1): Likewise.
1840 (cortex_a9_neon_int_2): Likewise.
1841 (cortex_a9_neon_int_3): Likewise.
1842 (cortex_a9_neon_int_4): Likewise.
1843 (cortex_a9_neon_int_5): Likewise.
1844 (cortex_a9_neon_vqneg_vqabs): Likewise.
1845 (cortex_a9_neon_vmov): Likewise.
1846 (cortex_a9_neon_vaba): Likewise.
1847 (cortex_a9_neon_vaba_qqq): Likewise.
1848 (cortex_a9_neon_vsma): Likewise.
1849 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1850 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1851 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1852 long_scalar): Likewise.
1853 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1854 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1855 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1856 long_scalar_qdd_64_32_long): Likewise.
1857 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1858 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1859 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1860 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1861 (cortex_a9_neon_shift_1): Likewise.
1862 (cortex_a9_neon_shift_2): Likewise.
1863 (cortex_a9_neon_shift_3): Likewise.
1864 (cortex_a9_neon_vshl_ddd): Likewise.
1865 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1866 (cortex_a9_neon_vsra_vrsra): Likewise.
1867 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1868 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1869 (cortex_a9_neon_fp_vsum): Likewise.
1870 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1871 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1872 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1873 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1874 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1875 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1876 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1877 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1878 (cortex_a9_neon_bp_simple): Likewise.
1879 (cortex_a9_neon_bp_2cycle): Likewise.
1880 (cortex_a9_neon_bp_3cycle): Likewise.
1881 (cortex_a9_neon_ldr): Likewise.
1882 (cortex_a9_neon_str): Likewise.
1883 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1884 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1885 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1886 (cortex_a9_neon_vld2_4_regs): Likewise.
1887 (cortex_a9_neon_vld3_vld4): Likewise.
1888 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1889 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1890 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1891 (cortex_a9_neon_vst3_vst4): Likewise.
1892 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1893 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1894 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1895 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1896 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1897 (cortex_a9_neon_mcr): Likewise.
1898 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1899 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
1900 (cortex_a9_fps): Likewise.
1901 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
1903 (cortex_m4_fmuls): Likewise.
1904 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
1906 (cortex_r4_mrc): Likewise.
1907 * config/arm/iterators.md: Update comment referring to neon_type.
1908 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
1909 (iwmmxt_movsi_insn): Likewise.
1910 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
1912 (pj4_core_to_vfp): Likewise.
1913 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
1915 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
1916 (vfp_fstore): Likewise.
1917 * doc/md.texi: Change references to neon_type to refer to type.
1919 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1921 * tree.h (DECL_BUILT_IN): Fix typo in comment.
1923 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
1925 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
1928 2013-09-04 Easwaran Raman <eraman@google.com>
1931 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
1932 (build_and_add_sum): Use it.
1933 (appears_later_in_bb): Simplify code.
1935 2013-09-04 Teresa Johnson <tejohnson@google.com>
1937 * dumpfile.c (dump_finish): Don't close stderr/stdout.
1939 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
1941 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
1943 2013-09-04 Jan Hubicka <jh@suse.cz>
1945 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
1946 * ipa-devirt.c: Include diganostic.h
1947 (odr_type_d): Add types and types_set.
1948 (hash_type_name): Work for types with vtables during LTO.
1949 (odr_hasher::remove): Fix comment; destroy types_set.
1950 (add_type_duplicate): New function,
1951 (get_odr_type): Use it.
1952 (dump_type_inheritance_graph): Dump type duplicates.
1953 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
1955 * tree.c (types_same_for_odr): Give exact answers on types with
1958 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1960 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
1961 explaining their differences.
1963 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
1965 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
1967 2013-09-03 Jeff Law <law@redhat.com>
1969 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
1970 when not threading through a joiner block. Pass joiner/no joiner
1971 state to register_jump_thread.
1972 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
1973 state from argument rather than implying on path length.
1974 Dump the entire jump thread path into debugging dump.
1975 * tree-flow.h (register_jump_thread): Update prototype.
1977 2013-08-29 Xinliang David Li <davidxl@google.com>
1979 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1980 Remove a trivial gcc_assert.
1982 2013-08-29 Xinliang David Li <davidxl@google.com>
1984 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
1985 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
1986 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1987 Delay base decl alignment adjustment.
1988 * tree-vectorizer.c (vect_destroy_datarefs): New function.
1989 * tree-vectorizer.h: New data structure.
1990 (set_dr_misalignment): New function.
1991 (dr_misalignment): Ditto.
1992 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
1993 (vectorizable_load): Ditto.
1994 (ensure_base_align): New function.
1995 (vectorize_loops): Add dbg_cnt support.
1996 (execute_vect_slp): Ditto.
1997 * dbgcnt.def: New debug counter.
1998 * Makefile: New dependency.
2000 2013-09-03 Meador Inge <meadori@codesourcery.com>
2004 2013-08-30 Meador Inge <meadori@codesourcery.com>
2006 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2008 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
2010 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
2011 function descriptor.
2013 2013-09-03 Richard Biener <rguenther@suse.de>
2015 * tree-affine.c (add_elt_to_tree): Fix association issue,
2016 avoid useless converts and make sure to always return a
2017 properly typed result.
2019 2013-09-03 Richard Biener <rguenther@suse.de>
2022 * fold-const.c (negate_expr_p): Fix division case.
2023 (negate_expr): Likewise.
2025 2013-09-03 Richard Biener <rguenther@suse.de>
2028 * tree-streamer-out.c: Include tm.h.
2029 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
2031 2013-09-03 Jan Hubicka <jh@suse.cz>
2033 * tree-profile.c (tree_profiling): Cleanup CFG when done.
2035 2013-09-03 Alan Modra <amodra@gmail.com>
2037 * config.gcc (powerpc*-*-linux*): Add support for little-endian
2038 multilibs to big-endian target and vice versa.
2039 * config/rs6000/t-linux64: Use := assignment on all vars.
2040 (MULTILIB_EXTRA_OPTS): Remove fPIC.
2041 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
2042 * config/rs6000/t-linux64le: New file.
2043 * config/rs6000/t-linux64bele: New file.
2044 * config/rs6000/t-linux64lebe: New file.
2046 2013-09-02 Jan Hubicka <jh@suse.cz>
2048 * ipa-inline-transform.c (inline_transform): Do not
2049 optimize_inline_calls when not optimizing.
2051 2013-09-02 Jan Hubicka <jh@suse.cz>
2053 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
2054 duplicated nodes for assembler names.
2055 * symtab.c (symtab_unregister_node): Do not attempt to unlink
2056 hard registers from assembler name hash.
2058 2013-09-02 Jan Hubicka <jh@suse.cz>
2060 * ipa-split.c (execute_split_functions): Split externally visible
2061 functions called once.
2063 2013-09-02 Martin Jambor <mjambor@suse.cz>
2066 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
2067 linked list. When finding the correct duplicate, also consider also
2068 the caller in additon to its inlined_to node.
2070 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
2072 * config/aarch64/aarch64-simd-builtins.def
2073 (dup_lane_scalar): Remove.
2074 * config/aarch64/aarch64-simd.md
2075 (aarch64_simd_dup): Add 'w->w' alternative.
2076 (aarch64_dup_lane<mode>): Allow for VALL.
2077 (aarch64_dup_lane_scalar<mode>): Remove.
2078 (aarch64_dup_lane_<vswap_width_name><mode>): New.
2079 (aarch64_get_lane_signed<mode>): Add w->w altenative.
2080 (aarch64_get_lane_unsigned<mode>): Likewise.
2081 (aarch64_get_lane<mode>): Likewise.
2082 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
2083 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
2084 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
2085 (VCON): Change container of V2SF.
2086 (vswap_width_name): Likewise.
2087 * config/aarch64/arm_neon.h
2088 (__aarch64_vdup_lane_any): New.
2089 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2090 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
2091 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2093 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
2096 * expr.c (emit_move_complex): Do not move complex FP values as parts if
2097 the source or the destination is a single hard register.
2099 2013-09-02 Richard Biener <rguenther@suse.de>
2102 * tree-scalar-evolution.c (instantiate_scev_name): Allow
2105 2013-09-02 Richard Biener <rguenther@suse.de>
2107 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
2108 arithmetic to sizetype.
2110 2013-09-02 Bin Cheng <bin.cheng@arm.com>
2112 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
2113 Find auto-increment use both before and after candidate.
2115 2013-09-02 Marek Polacek <polacek@redhat.com>
2117 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
2119 2013-09-01 Jan Hubicka <jh@suse.cz>
2121 * Makefile.in: Add ipa-profile.o
2122 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
2123 * cgraph.c (struct cgraph_propagate_frequency_data,
2124 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
2125 ipa-profile.c; replace cgraph_ by ipa_ prefix.
2126 * cgraph.h (cgraph_propagate_frequency): Remove.
2127 * ipa-inline-analysis.c: Include ipa-utils.h;
2128 drop duplicated cfgloop.h.
2129 (inline_update_callee_summaries): Update.
2130 * ipa-profile.c: New file.
2131 * ipa-utils.h (ipa_propagate_frequency): Declare.
2132 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
2133 data-streamer.h, value-prof.h.
2134 (symtab_remove_unreachable_nodes): Update profile.
2135 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
2136 account_time_size, cmp_counts, dump_histogram,
2137 ipa_profile_generate_summary, ipa_profile_write_summary,
2138 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
2139 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
2140 Move to ipa-profile.c.
2142 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
2144 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
2146 2013-09-01 Jan Hubicka <jh@suse.cz>
2148 * common.opt (fdevirtualize-speculatively): New function.
2149 * invoke.texi (fdevirtualize-speculatively): Document.
2150 * ipa-devirt.c: Include ipa-inline.h
2151 (likely_target_p): New function.
2152 (ipa_devirt): New function.
2153 (gate_ipa_devirt): New function.
2154 (pass_data_ipa_devirt): New static var.
2155 (pass_ipa_devirt): Likewise.
2156 (make_pass_ipa_devirt): New function.
2157 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
2158 (common_handle_option): Disable devirtualization when
2159 value range profiling is available.
2160 * passes.def (pass_ipa_devirt): Add.
2161 * timever.def (TV_IPA_DEVIRT): New timevar.
2162 * tree-pass.h (make_pass_ipa_devirt):
2164 2013-09-01 Iain Sandoe <iain@codesourcery.com>
2166 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
2167 include sanitize(undefined).
2169 2013-08-31 Diego Novillo <dnovillo@google.com>
2171 * Makefile.in (TREE_CORE_H): Define.
2173 (GTFILES): Add tree-core.h.
2174 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
2176 * tree-core.h: New file.
2177 Move all data structures, enum, typedefs, global
2178 declarations and constants from ...
2181 2013-08-31 Jan Hubicka <jh@suse.cz>
2183 * bulitins.c (expand_builtin): Do not early exit for gcov
2184 instrumented functions.
2186 2013-08-31 Marek Polacek <polacek@redhat.com>
2188 * ubsan.c: Include tm_p.h.
2190 2013-08-31 Jan Hubicka <jh@suse.cz>
2192 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
2195 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
2196 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
2197 * tree-cfg.c (verify_gimple_label): ... here.
2198 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
2199 (ipa_merge_profiles): New function.
2200 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
2201 (lto_input_function_body): Likewise.
2202 * ipa-utils.h (ipa_merge_profiles): Declare.
2203 * lto-streamer.h (lto_input_function_body): Update prototype.
2204 (emit_label_in_global_context_p): Remove.
2205 * lto-symtab.c: Include ipa-utils.h
2206 (lto_cgraph_replace_node): Use ipa_merge_profiles.
2208 2013-08-31 Jan Hubicka <jh@suse.cz>
2210 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
2212 2013-08-31 Jan Hubicka <jh@suse.cz>
2214 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
2216 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
2218 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
2219 "cmp" RTX before signed_comparison_operator check to account
2222 2013-08-30 Jan Hubicka <jh@suse.cz>
2224 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
2225 (detect_type_change_1): Rename to ...
2226 (detect_type_change): ... this one; early return on non-polymorphic
2228 (detect_type_change_ssa): Add comp_type parameter; update
2229 use of detect_type_change.
2230 (compute_complex_assign_jump_func): Add param_type parameter;
2231 update use of detect_type_change_ssa.
2232 (compute_complex_ancestor_jump_func): Likewise.
2233 (ipa_get_callee_param_type): New function.
2234 (ipa_compute_jump_functions_for_edge): Compute parameter type;
2235 update calls to the jump function computation functions.
2237 2013-08-30 Teresa Johnson <tejohnson@google.com>
2238 Steven Bosscher <steven@gcc.gnu.org>
2240 * cfgrtl.c (fixup_new_cold_bb): New routine.
2241 (commit_edge_insertions): Invoke fixup_partitions.
2242 (find_partition_fixes): New routine.
2243 (fixup_partitions): Ditto.
2244 (verify_hot_cold_block_grouping): Update comments.
2245 (rtl_verify_edges): Invoke find_partition_fixes.
2246 (rtl_verify_bb_pointers): Update comments.
2247 (rtl_verify_bb_layout): Ditto.
2248 * basic-block.h (probably_never_executed_edge_p): Declare.
2249 (fixup_partitions): Ditto.
2250 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
2251 * bb-reorder.c (sanitize_hot_paths): New function.
2252 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
2254 * predict.c (probably_never_executed_edge_p): New routine.
2255 * cfg.c (check_bb_profile): Add partition insanity warnings.
2257 2013-08-30 Meador Inge <meadori@codesourcery.com>
2259 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2261 2013-08-30 Marek Polacek <polacek@redhat.com>
2263 * Makefile.in (ubsan.o): Add.
2264 (c-family/c-ubsan.o): Add.
2265 (builtins.o): Add ubsan.h dependency.
2266 * ubsan.h: New file.
2267 * ubsan.c: New file.
2268 * common.opt: Add -fsanitize=undefined option.
2269 (flag_sanitize): Add variable.
2270 (fsanitize=): Add option. Add Driver.
2271 (fsanitize=thread): Remove option.
2272 (fsanitize=address): Likewise.
2273 (static-libubsan): New option.
2274 * doc/invoke.texi: Document the new flag and -static-libubsan.
2275 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
2276 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
2277 * builtin-attrs.def (ATTR_COLD): Define.
2278 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2279 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
2280 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
2281 * flag-types.h (sanitize_code): New enum.
2282 * opts.c (common_handle_option): Parse command line arguments
2283 of -fsanitize=. Add -fsanitize=unreachable option.
2284 * varasm.c (get_variable_section): Adjust.
2285 (assemble_noswitch_variable): Likewise.
2286 (assemble_variable): Likewise.
2287 (output_constant_def_contents): Likewise.
2288 (categorize_decl_for_section): Likewise.
2289 (place_block_symbol): Likewise.
2290 (output_object_block): Likewise.
2291 * builtins.def: Likewise.
2292 * toplev.c (compile_file): Likewise.
2293 (process_options): Likewise.
2294 * cppbuiltin.c: Likewise.
2295 * tsan.c (tsan_pass): Likewise.
2296 (tsan_gate): Likewise.
2297 (tsan_gate_O0): Likewise.
2298 * cfgexpand.c (partition_stack_vars): Likewise.
2299 (expand_stack_vars): Likewise.
2300 (defer_stack_allocation): Likewise.
2301 (expand_used_vars): Likewise.
2302 * cfgcleanup.c (old_insns_match_p): Likewise.
2303 * asan.c (asan_finish_file): Likewise.
2304 (asan_instrument): Likewise.
2305 (gate_asan): Likewise.
2306 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
2307 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2308 (asan_global_struct): Use pointer_sized_int_node instead
2309 calling build_nonstandard_integer_type.
2310 (initialize_sanitizer_builtins): Likewise.
2311 (asan_finish_file): Likewise.
2312 * gcc.c: Document %{%:function(args):X}.
2313 (static_spec_functions): Add sanitize.
2314 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
2315 store funcval != NULL there.
2316 (do_spec_1): Adjust handle_spec_function caller.
2317 (handle_braces): Allow %:function(args) as condition.
2318 (sanitize_spec_function): New function.
2319 (ADD_STATIC_LIBUBSAN_LIBS): Define.
2320 (LIBUBSAN_SPEC): Likewise.
2321 (LIBUBSAN_EARLY_SPEC): Likewise.
2322 (SANITIZER_SPEC): Handle libubsan.
2323 (SANITIZER_EARLY_SPEC): Likewise.
2324 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
2325 instead of fsanitize=address.
2326 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
2327 instead of fsanitize=address*.
2328 * builtins.c: Include ubsan.h.
2329 (fold_builtin_0): Instrument __builtin_unreachable.
2330 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
2331 instead of flag_asan.
2332 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
2333 (pointer_sized_int_node): Define.
2334 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
2336 2013-08-30 Mike Stump <mikestump@comcast.net>
2338 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
2341 2013-08-29 Jan Hubicka <jh@suse.cz>
2343 * cgraph.c (cgraph_function_body_availability): Handle weakref
2345 * passes.def: Remove pass_fixup_cfg.
2346 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
2347 track when we need to remove functions.
2348 (gate_ipa_inline): Execute inlining always; add comment why.
2349 (pass_data_ipa_inline): Remove TODO_remove_functions.
2350 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
2351 do not produce summaries.
2352 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
2353 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
2354 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
2355 constant pool and vtable.
2357 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
2359 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
2360 New arm_neon.h's internal macros to specify 64-bit constants.
2361 Avoid using stdint.h's macros.
2363 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
2365 * recog.c (verify_changes): Verify that changes[i].old is non-zero
2366 before applying REG_P.
2368 2013-08-30 Jakub Jelinek <jakub@redhat.com>
2370 PR tree-optimization/58277
2371 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
2372 after seeing too many stmts with vdef in between dombb and current
2373 bb, invalidate everything.
2375 2013-08-30 Richard Biener <rguenther@suse.de>
2377 * fold-const.c (fold_single_bit_test): Fix overflow test.
2379 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
2381 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
2382 and which can live in a register, always retrieve the value on entry.
2383 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
2384 passed by invisible reference specially.
2385 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
2386 (vt_add_function_parameter): Correctly deal with a parameter passed by
2387 invisible reference.
2389 2013-08-30 Jan Hubicka <jh@suse.cz>
2391 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
2393 2013-08-30 Richard Biener <rguenther@suse.de>
2395 PR tree-optimization/58228
2396 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
2397 allow invariant loads in nested loop vectorization.
2399 2013-08-30 Richard Biener <rguenther@suse.de>
2401 PR tree-optimization/58223
2402 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
2403 (has_anti_or_output_dependence): ... this and adjust to also
2404 look for output dependences.
2405 (mark_nodes_having_upstream_mem_writes): Adjust.
2406 (rdg_flag_uses): Likewise.
2408 2013-08-30 Richard Biener <rguenther@suse.de>
2410 PR tree-optimization/58010
2411 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
2412 assert that we have a loop-closed PHI.
2414 2013-08-29 Jan Hubicka <jh@suse.cz>
2416 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
2417 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
2418 * lto-section-in.c (lto_free_function_in_decl_state): New function.
2419 (lto_free_function_in_decl_state_for_node): New function.
2421 2013-08-29 Xinliang David Li <davidxl@google.com>
2423 * loop-unroll.c (report_unroll_peel): Minor message change.
2424 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
2425 Emit alignment peeling message with default -fopt-info.
2426 (vect_loop_versioning): Emit loop version info message.
2427 * tree-vectorizer.c (vectorize_loops): Minor message change.
2428 (execute_vect_slp): Ditto.
2430 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
2432 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
2433 of the clone from the DECL_NAME of the original function.
2435 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
2437 * passes.c (register_pass): Add overload.
2438 * tree-pass.h (register_pass): Forward declare it. Add comment.
2440 2013-08-29 Jan Hubicka <jh@suse.cz>
2442 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
2443 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
2444 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
2445 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2446 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
2447 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
2448 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2449 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
2451 2013-08-29 Teresa Johnson <tejohnson@google.com>
2453 * dumpfile.c (dump_loc): Output column number.
2454 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
2455 * doc/invoke.texi: Document optall -fopt-info flag.
2456 * profile.c (read_profile_edge_counts): Use new dump framework.
2457 (compute_branch_probabilities): Ditto.
2458 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
2459 when pass not in any opt group.
2460 * pass_manager.h (pass_manager::get_pass_profile): New method.
2461 * value-prof.c (check_counter): Use new dump framework.
2462 (check_ic_target): Ditto.
2463 * coverage.c (get_coverage_counts): Ditto.
2464 (coverage_init): Setup new dump framework.
2466 2013-08-29 Richard Biener <rguenther@suse.de>
2468 PR tree-optimization/58246
2469 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
2470 handle the dominance check inside a basic-block.
2472 2013-08-29 Richard Biener <rguenther@suse.de>
2475 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
2476 of default defs that appear in abnormal PHI nodes.
2478 2013-08-29 Richard Biener <rguenther@suse.de>
2480 PR tree-optimization/57685
2481 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2482 single-use operands to avoid exponential complexity.
2484 2013-08-28 Dehao Chen <dehao@google.com>
2486 * ipa-inline.c (edge_badness): Fix integer underflow.
2488 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
2490 * gtm-builtins.def (_ITM_free): Declare leaf.
2492 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2495 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2496 (*tls_local_dynamic_base_64_largepic): Likewise.
2497 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2498 Remove predicate from call operand.
2499 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2500 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2502 2013-08-28 Jeff Law <law@redhat.com>
2504 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2505 checks for the number of predecessors and successors allowed.
2506 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2507 which require copying a joiner block if there is a request which
2508 is a subpath that requires no joiner block copying.
2510 2013-08-28 Jan Hubicka <jh@suse.cz>
2512 * lto-streamer-out.c (DFS_write_tree_body): Drop
2513 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2514 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2515 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2516 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2517 Do not read DECL_ERROR_ISSUED.
2518 (unpack_ts_decl_with_vis_value_fields): Do not read
2520 (lto_input_ts_binfo_tree_pointers): Do not read
2521 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
2522 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2523 write DECL_ERROR_ISSUED..
2524 (pack_ts_decl_with_vis_value_fields): Do not write
2526 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
2527 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
2528 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2529 * tree.h (tree_decl_common): Update comment.
2530 (DECL_ERROR_ISSUED): Remove.
2532 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2535 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2537 2013-08-28 Jan Hubicka <jh@suse.cz>
2539 * builtins.def (free): Declare leaf.
2541 2013-08-27 David Malcolm <dmalcolm@redhat.com>
2544 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2545 * configure: Regenerate.
2547 2013-08-27 Martin Jambor <mjambor@suse.cz>
2549 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2550 (ipa_ancestor_jf_data): Likewise.
2551 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2552 (ipa_get_jf_pass_through_type_preserved): New function.
2553 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2554 (ipa_get_jf_ancestor_type_preserved): New function.
2555 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
2556 (ipa_get_jf_ancestor_result): Likewise.
2557 (propagate_vals_accross_pass_through): Use
2558 ipa_get_jf_pass_through_result to do all the value mappings.
2559 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2560 type_preserved flag.
2561 (ipa_set_jf_cst_copy): New function.
2562 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2563 (ipa_set_jf_arith_pass_through): Likewise.
2564 (ipa_set_ancestor_jf): Likewise.
2565 (compute_complex_assign_jump_func): Set type_preserved instead of
2567 (ipa_compute_jump_functions_for_edge): Likewise.
2568 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2569 (update_jump_functions_after_inlining): Update type_preserved.
2570 Explicitely create jump functions when combining one with pass_through.
2571 (ipa_write_jump_function): Stream the type_preserved flags.
2572 (ipa_read_jump_function): Likewise.
2574 2013-08-27 Jakub Jelinek <jakub@redhat.com>
2575 Aldy Hernandez <aldyh@redhat.com>
2577 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
2578 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
2579 * function.h (struct function): Add has_force_vect_loops and
2581 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
2582 * gimple.c (gimple_build_omp_critical): Add KIND argument and
2584 * gimple.def: Update CLAUSES comments.
2585 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
2586 (gimple_build_omp_for): Add argument to prototype.
2587 (gimple_omp_for_kind): New.
2588 (gimple_omp_for_set_kind): New.
2589 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
2590 GOVD_DATA_SHARE_CLASS.
2591 (enum omp_region_type): Add ORT_SIMD.
2592 (gimple_add_tmp_var): Handle ORT_SIMD.
2593 (gimplify_var_or_parm_decl): Same.
2594 (is_gimple_stmt): Same.
2595 (omp_firstprivatize_variable): Same.
2596 (omp_add_variable): Only use splay_tree_insert if lookup failed.
2597 (omp_notice_variable): Handle ORT_SIMD.
2598 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
2599 (omp_check_private): Handle ORT_SIMD.
2600 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2602 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
2603 Handle OMP_CLAUSE_LASTPRIVATE.
2604 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2606 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
2607 (gimplify_expr): Handle OMP_SIMD.
2608 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
2609 (expand_GOMP_SIMD_VF): New.
2610 (expand_GOMP_SIMD_LAST_LANE): New.
2611 * internal-fn.def (GOMP_SIMD_LANE): New.
2612 (GOMP_SIMD_VF): New.
2613 (GOMP_SIMD_LAST_LANE): New.
2614 * omp-low.c: Include target.h.
2615 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
2617 (check_omp_nesting_restrictions): Same.
2619 (lower_rec_simd_input_clauses): New.
2620 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
2622 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
2623 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
2624 (expand_omp_build_assign): New.
2625 (expand_omp_for_init_counts): New.
2626 (expand_omp_for_init_vars): New.
2627 (extract_omp_for_update_vars): New.
2628 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
2629 and rewrite accordingly.
2630 (expand_omp_simd): New.
2631 (expand_omp_for): Use expand_omp_simd.
2632 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2633 (lower_omp_for): Do not lower the body.
2634 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2636 * tree-flow.h (find_omp_clause): Remove prototype.
2637 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2638 forcing vectorization of the loop, or if flag_tree_vectorize.
2639 (gate_tree_if_conversion): Similarly.
2640 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2641 gimple_build_omp_for.
2642 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2643 * tree-parloops (create_parallel_loop): Pass kind argument to
2644 gimple_build_omp_for.
2645 * tree-pretty-print.c (dump_omp_clause): Add cases for
2646 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2647 OMP_CLAUSE__SIMDUID_.
2648 (dump_generic_node): Handle OMP_SIMD.
2649 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2650 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2651 unroll OMP_SIMD loops here.
2652 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
2653 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
2655 (vect_analyze_data_refs): Handle simd loops.
2656 * tree-vect-loop.c (vectorizable_live_operation): Handle
2658 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
2659 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2660 (vectorizable_load): Same.
2661 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
2662 (struct simduid_to_vf): New.
2663 (simduid_to_vf::hash): New.
2664 (simduid_to-vf::equal): New.
2665 (struct simd_array_to_simduid): New.
2666 (simd_array_to_simduid::hash): New.
2667 (simd_array_to_simduid::equal): New.
2668 (adjust_simduid_builtins): New.
2669 (struct note_simd_array_uses_struct): New.
2670 (note_simd_array_uses_cb): New.
2671 (note_simd_array_uses): New.
2672 (vectorize_loops): Handle simd hints and adjust simd builtins
2674 * tree-vectorizer.h (struct _stmt_vec_info): Add
2675 simd_lane_access_p field.
2676 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2677 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2678 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2679 (omp_clause_code_name): Same.
2680 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2681 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2682 * tree.def (OMP_SIMD): New entry.
2683 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2684 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
2685 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2686 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2687 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2688 (OMP_CLAUSE_LINEAR_STEP): New.
2689 (OMP_CLAUSE_SAFELEN_EXPR): New.
2690 (OMP_CLAUSE__SIMDUID__DECL): New.
2691 (find_omp_clause): New prototype.
2693 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2695 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2696 Haswell processor detection.
2698 2013-08-27 Christian Widmer <shadow@umbrox.de>
2701 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2702 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2703 AVX2 capable processors.
2705 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2707 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2708 of vget_low_* with implementations in terms of other intrinsics.
2710 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
2713 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2714 values to -1, 0 and 1.
2716 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2718 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2719 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2720 (<optab><mode>3_insn): Remove.
2721 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2722 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2723 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2724 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2725 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2726 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2727 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2728 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2729 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2730 (ror<mode>3_insn): Likewise.
2731 * config/aarch64/predicates.md (aarch64_simd_register): New.
2733 2013-08-27 Richard Biener <rguenther@suse.de>
2735 PR tree-optimization/57521
2736 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2737 one edge is non-critical.
2738 (find_phi_replacement_condition): Make sure to use a non-critical
2739 edge. Cleanup and remove old bug workarounds.
2740 (bb_postdominates_preds): Remove.
2741 (if_convertible_loop_p_1): Do not compute post-dominators.
2742 (combine_blocks): Do not free post-dominators.
2743 (main_tree_if_conversion): Likewise.
2744 (pass_data_if_conversion): Add TODO_verify_ssa.
2746 2013-08-27 DJ Delorie <dj@redhat.com>
2748 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2750 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2752 * function.c (assign_parm_find_data_types): Set passed_mode and
2753 nominal_mode to the TYPE_MODE of nominal_type for the built
2754 pointer type in case of the struct-pass-by-reference.
2756 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2758 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2759 (UINT16_TYPE): Change default to "unsigned int".
2761 * config/avr/avr.opt (mfract-convert-truncate): New option.
2762 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2763 is set, round negative fractional integers according to n1169
2764 when converting to integer types.
2766 2013-08-26 Jan Hubicka <jh@suse.cz>
2768 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2769 methods can not be called indirectly when their address is not taken.
2771 2013-08-26 Jan Hubicka <jh@suse.cz>
2773 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2776 2013-08-26 Jan Hubicka <jh@suse.cz>
2778 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2781 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2783 * reload.c (find_valid_class): Allow classes that do not include
2784 FIRST_PSEUDO_REGISTER - 1.
2786 2013-08-26 Jan Hubicka <jh@suse.cz>
2788 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2789 fix edge count/frequency when speculation failed; fix type check
2790 for the direct call.
2792 2013-08-26 Jan Hubicka <jh@suse.cz>
2794 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2796 2013-08-26 Jan Hubicka <jh@suse.cz>
2798 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2799 to be changed by edge redirection.
2801 2013-08-26 Jan Hubicka <jh@suse.cz>
2803 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2804 formating; add sanity check.
2805 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2806 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2807 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2808 (resolve_noninline_speculation): Update callee keys, too.
2810 2013-08-26 Jan Hubicka <jh@suse.cz>
2812 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2813 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2815 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2817 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2820 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
2822 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2823 (x86_64_elf_select_section): Ditto.
2824 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2826 (ix86_pass_by_reference): Ditto.
2827 (output_set_got): Ditto.
2828 (ix86_unary_operator_ok): Ditto.
2829 (ix86_expand_builtin): Ditto.
2831 2013-08-23 Jan Hubicka <jh@suse.cz>
2833 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
2835 2013-08-23 Jan Hubicka <jh@suse.cz>
2837 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
2838 (tree_decl_with_vis): Add FINAL field.
2840 2013-08-23 Jeff Law <law@redhat.com>
2842 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
2843 the debugging dump when the expression is fully redundant.
2845 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2847 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
2848 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
2849 * pretty-print.c (pp_formatted_text_data): Likewise.
2850 (pp_write_text_to_stream): Likewise.
2851 (pp_write_text_as_dot_label_to_stream): Likewise.
2852 (pp_append_r): Likewise.
2853 (pp_format): Likewise.
2854 (pp_flush): Likewise.
2855 (pp_clear_output_area): Likewise.
2856 (pp_append_text): Likewise.
2857 (pp_formatted_text): Likewise.
2858 (pp_remaining_character_count_for_line): Likewise.
2859 (pp_newline): Likewise.
2860 (pp_character): Likewise.
2861 (output_buffer::~output_buffer): Define.
2862 (pretty_printer::~pretty_printer): Destruct output buffer.
2863 * pretty-print.h (output_buffer::~output_buffer): Declare.
2864 (pretty_printer::~pretty_printer): Declare virtual.
2866 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
2869 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
2870 HOST_WIDE_INT_M1U): New macros.
2871 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
2872 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
2873 unsigned -1 for lshift.
2874 * cse.c (cse_insn): Likewise.
2875 * double-int.c (rshift_double, lshift_double): Likewise.
2876 * builtins.c (fold_builtin_bitop): Likewise.
2877 * combine.c (force_to_mode): Likewise.
2878 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
2879 * simplify-rtx.c (simplify_const_unary_operation,
2880 simplify_const_binary_operation): Likewise.
2881 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
2882 check_va_list_escapes): Likewise.
2883 * rtlanal.c (nonzero_bits1): Likewise.
2884 * expmed.c (expand_smod_pow2): Likewise.
2885 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
2887 2013-08-23 Jan Hubicka <jh@suse.cz>
2889 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
2890 as having address taken.
2892 2013-08-23 Jan Hubicka <jh@suse.cz>
2894 * ipa-utils.h (method_class_type): Declare.
2895 * ipa-devirt.c (method_class_type): Export.
2897 * cgraphunit.c (analyze_functions): Do basic devirtualization;
2898 do not walk base classes of anonymous types.
2900 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
2902 PR rtl-optimization/58220
2904 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
2905 handle SEQUENCE insns properly.
2907 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2909 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
2911 (pp_newline_and_indent): Likewise.
2912 (pp_separate_with): Likewise.
2913 * pretty-print.c (pp_newline_and_flush): Define.
2914 (pp_newline_and_indent): Likewise.
2915 (pp_separate_with): Likewise.
2917 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2920 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
2921 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
2923 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
2925 * config/i386/predicates.md (ext_sse_reg_operand): New.
2926 * config/i386/i386.md (*movti_internal): Use
2927 predicate to determine if EVEX is needed.
2928 (*movsi_internal): Ditto.
2929 (*movdf_internal): Ditto.
2930 (*movsf_internal): Ditto.
2931 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2933 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2935 PR tree-optimization/58209
2936 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
2937 (find_tail_calls): Give up for pointer result types if m is non-NULL.
2938 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
2939 emit POINTER_PLUS_EXPR.
2940 (create_tailcall_accumulator): For pointer result type accumulate in
2943 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
2945 * configure.ac: Add backslashes missing from the last change.
2946 * configure: Regenerate.
2948 2013-08-22 Jan Hubicka <jh@susue.cz>
2950 * ipa.c (function_and_variable_visibility): First remember function
2951 was global and then make it local.
2953 2013-08-22 Julian Brown <julian@codesourcery.com>
2955 * configure.ac: Add aarch64 to list of arches which use "nop" in
2957 * configure: Regenerate.
2959 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2961 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
2962 gnu_libc_has_function.
2963 * config/s390/tpf.h: Likewise.
2965 2013-08-22 Jan Hubicka <jh@susue.cz>
2967 * timevar.c (validate_phases): Add cast.
2969 2013-08-22 Jan Hubicka <jh@susue.cz>
2971 * timevar.c (validate_phases): Use size_t for memory.
2972 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
2974 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
2976 * pretty-print.h (output_buffer::output_buffer): Declare.
2977 (pretty_printer::pretty_printer): Likewise.
2978 (pp_construct): Remove.
2979 * pretty-print.c (output_buffer::output_buffer): Define.
2980 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
2981 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
2982 (print_gimple_expr): Likewise.
2983 (print_gimple_seq): Likewise.
2984 (gimple_dump_bb): Likewise.
2985 * sched-vis.c (dump_value_slim): Likewise.
2986 (dump_insn_slim): Likewise.
2987 (dump_rtl_slim): Likewise.
2988 (str_pattern_slim): Likewise.
2989 * tree-mudflap.c (mf_varname_tree): Likewise.
2990 * graph.c (print_graph_cfg): Likewise.
2991 (start_graph_dump): Likewise.
2992 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
2994 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
2995 pretty printer initialization.
2996 * coretypes.h (diagnostic_context): Remove superflous type alias
2998 (pretty_printer): Likewise. Declare directly as a class.
2999 (pretty_print_info): Remove declaration as class.
3000 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
3001 and pp_clear_output_area.
3002 (asan_add_global): Likewise.
3004 2013-08-22 Jan Hubicka <jh@suse.cz>
3006 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
3007 * ipa-utils.h (update_type_inheritance_graph): Declare.
3008 (possible_polymorphic_call_target_p): Declare.
3009 (possible_polymorphic_call_target_p): New.
3010 * ipa-devirt.c: Update toplevel comments.
3011 (cached_polymorphic_call_targets): Move up.
3012 (odr_type_d): Move ID down.
3013 (polymorphic_type_binfo_p): Update comment.
3014 (odr_hasher::remove): Likewise;
3015 (get_odr_type): Set anonymous_namespace.
3016 (dump_odr_type): Dump it.
3017 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
3018 (maybe_record_node): Record node in cached_polymorphic_call_targets.
3019 (record_binfo): Add comment.
3020 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
3022 (devirt_node_removal_hook): Do not iCE when cache is freed.
3023 (possible_polymorphic_call_target_p): New predicate.
3024 (update_type_inheritance_graph): New function.
3026 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
3027 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3028 Sergey Lega <sergey.s.lega@intel.com>
3029 Anna Tikhonova <anna.tikhonova@intel.com>
3030 Ilya Tocar <ilya.tocar@intel.com>
3031 Andrey Turetskiy <andrey.turetskiy@intel.com>
3032 Ilya Verbin <ilya.verbin@intel.com>
3033 Kirill Yukhin <kirill.yukhin@intel.com>
3034 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3036 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
3037 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
3038 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
3039 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
3040 (OPTION_MASK_ISA_AVX2_UNSET): Update.
3041 (OPTION_MASK_ISA_AVX512F_UNSET): New.
3042 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
3043 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
3044 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
3045 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
3046 OPT_mavx512pf, OPT_mavx512er cases.
3047 * config/i386/constraints.md (v): New constraint.
3048 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
3049 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
3050 (bit_AVX512CD): New.
3051 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3052 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
3053 * config/i386/i386-c.c (ix86_target_macros_internal):
3054 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
3056 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
3057 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
3058 * config/i386/i386.c (regclass_map, dbx_register_map)
3059 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
3060 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
3061 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
3063 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
3064 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
3065 -mavx512pf options. Fix formatting.
3066 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
3067 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
3068 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
3069 -mavx512cd, -mavx512pf options.
3070 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
3071 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
3072 (ix86_preferred_output_reload_class): Replace SSE_REGS with
3074 (ix86_hard_regno_mode_ok): Support 512-bit registers.
3075 (ix86_set_reg_reg_cost): Ditto.
3076 (x86_order_regs_for_local_alloc): Ditto.
3077 (MAX_VECT_LEN): Extend to 64-byte.
3078 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
3079 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
3080 (TARGET_AVX512ER, TARGET_AVX512CD): New.
3081 (BIGGEST_ALIGNMENT): Extend to 512-bits.
3082 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
3083 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
3084 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
3085 (SSE_REG_MODE_P): Support new modes.
3086 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
3087 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
3088 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
3089 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
3090 (REG_CLASS_CONTENTS): Add new registers.
3091 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
3092 (EXT_REX_SSE_REGNO_P): New.
3093 (HI_REGISTER_NAMES): Add new registers.
3094 * config/i386/i386.md: Define constants for new registers.
3095 (mode): Add new 512-bit modes.
3096 (prefix): Support evex prefix.
3097 (isa): Support avx512f, noavx512f, fma_avx512f.
3098 (ssemodesuffix): Add new 512-bit modes.
3100 (*movxi_internal_avx512f): Ditto.
3101 (*movdi_internal): Replace constraint "x" with the new constraint "v".
3103 (*movsi_internal): Likewise.
3104 (*movdf_internal): Likewise.
3105 (*movsf_internal): Likewise.
3106 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
3107 (<code><mode>3): Likewise.
3108 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
3110 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
3111 with the new constraint "v".
3112 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
3114 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
3115 with the new constraint "v".
3116 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
3117 (<sse2>_storedqu<avxsizesuffix>): Likewise.
3118 (*<plusminus_insn><mode>3): Likewise.
3119 (<sse>_vm<plusminus_insn><mode>3): Likewise.
3120 (*mul<mode>3): Likewise.
3121 (<sse>_vmmul<mode>3): Likewise.
3122 (<sse>_div<mode>3): Likewise.
3123 (<sse>_vmdiv<mode>3): Likewise.
3124 (<sse>_sqrt<mode>2): Likewise.
3125 (<sse>_vmsqrt<mode>2): Likewise.
3126 (*<code><mode>3_finite): Likewise.
3127 (*<code><mode>3) <smaxmin>: Likewise.
3128 (<sse>_vm<code><mode>3): Likewise.
3129 (*<code><mode>3) <any_logic>: Likewise.
3130 (*fma_fmadd_<mode>): Likewise.
3131 (*fma_fmsub_<mode>): Likewise.
3132 (*fma_fnmadd_<mode>): Likewise.
3133 (*fma_fnmsub_<mode>): Likewise.
3134 (*fma_fmaddsub_<mode>): Likewise.
3135 (*fma_fmsubadd_<mode>): Likewise.
3136 (*fmai_fmadd_<mode>): Likewise.
3137 (*fmai_fmsub_<mode>): Likewise.
3138 (*fmai_fnmadd_<mode>): Likewise.
3139 (*fmai_fnmsub_<mode>): Likewise.
3140 (sse_cvtsi2ss): Likewise.
3141 (sse_cvtsi2ssq): Likewise.
3142 (sse_cvtss2si): Likewise.
3143 (sse_cvtss2si_2): Likewise.
3144 (sse_cvtss2siq): Likewise.
3145 (sse_cvtss2siq_2): Likewise.
3146 (sse_cvttss2si): Likewise.
3147 (sse_cvtss2siq_2): Likewise.
3148 (float<sseintvecmodelower><mode>2): Likewise.
3149 (sse2_cvtsd2si_2): Likewise.
3150 (sse2_cvtsd2siq_2): Likewise.
3151 (*<plusminus_insn><mode>3): Likewise.
3152 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
3153 (*<sse4_1_avx2>_mul<mode>3): Likewise.
3154 (ashr<mode>3): Likewise.
3155 (<shift_insn><mode>3): Likewise.
3156 (avx2_<code><mode>3): Likewise.
3157 (*avx2_<code><mode>3): Likewise.
3158 (*andnot<mode>3): Likewise.
3159 (*<code><mode>3) <any_logic>: Likewise.
3160 (abs<mode>2): Likewise.
3161 (avx2_permvar<mode>): Likewise.
3162 (avx2_perm<mode>_1): Likewise.
3163 (*avx_vpermilp<mode>): Likewise.
3164 (avx_vpermilvar<mode>3): Likewise.
3165 (avx2_ashrv<mode>): Likewise.
3166 (avx2_<shift_insn>v<mode>): Likewise.
3167 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
3169 * doc/rtl.texi: Document XImode.
3171 2013-08-21 Jeff Law <law@redhat.com>
3173 * tree-flow.h (register_jump_thread): Pass vector of edges
3174 instead of each important edge.
3175 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
3176 thread path into a vector and pass that to register_jump_thread.
3177 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
3178 passed in edge vector to the current 3-edge form.
3181 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3183 * dce.c (fini_dce): Call df_analyze again just in case
3184 delete_unmarked_insns removed anything.
3186 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
3188 * reload.h (struct reg_equivs): Rename to ..
3189 (struct reg_equivs_s): .. this.
3191 2013-08-20 Martin Liska <marxin.liska@gmail.com>
3193 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
3195 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3197 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
3199 2013-08-21 Jeff Law <law@redhat.com>
3201 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
3202 simplify assignments too. If the RHS collapses to a singleton
3203 range, then return the value for the range.
3205 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
3207 * config/i386/sse.md (V16): Rename to...
3209 (mov<mode>): Update iterator name.
3210 (*mov<mode>_internal): Ditto.
3211 (push<mode>1): Ditto.
3212 (movmisalign<mode>): Ditto.
3214 2013-08-20 Jan Hubicka <jh@suse.cz>
3217 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
3218 entry for direct edges.
3219 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
3221 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3223 Revert my last two changes, r201865 and r201864:
3226 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3228 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3229 instances can own GC refs.
3231 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3232 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3233 (gcc::context::gt_pch_nx): Likewise.
3234 (gcc::context::gt_pch_nx): Likewise.
3235 * ggc.h (gt_ggc_mx <T>): New.
3236 (gt_pch_nx_with_op <T>): New.
3237 (gt_pch_nx <T>): New.
3238 * passes.c (opt_pass::gt_ggc_mx): New.
3239 (opt_pass::gt_pch_nx): New.
3240 (opt_pass::gt_pch_nx_with_op): New.
3241 (pass_manager::gt_ggc_mx): New.
3242 (pass_manager::gt_pch_nx): New.
3243 (pass_manager::gt_pch_nx_with_op): New.
3244 (pass_manager::operator new): Use
3245 ggc_internal_cleared_alloc_stat rather than xcalloc.
3246 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3247 (pass_manager::gt_ggc_mx): New.
3248 (pass_manager::gt_pch_nx): New.
3249 (pass_manager::gt_pch_nx_with_op): New.
3250 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3251 (opt_pass::operator new): New.
3252 (opt_pass::gt_ggc_mx): New.
3253 (opt_pass::gt_pch_nx): New.
3254 (opt_pass::gt_pch_nx_with_op): New.
3257 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3259 * Makefile.in (GTFILES): Add context.h.
3260 * context.c (gcc::context::operator new): New.
3261 (gcc::context::gt_ggc_mx): New.
3262 (gcc::context::gt_pch_nx): New.
3263 (gcc::context::gt_pch_nx): New.
3264 * context.h (gcc::context): Add GTY((user)) marking.
3265 (gcc::context::operator new): New.
3266 (gcc::context::gt_ggc_mx): New.
3267 (gcc::context::gt_pch_nx): New.
3268 (gcc::context::gt_pch_nx): New.
3269 (g): Add GTY marking.
3270 (gt_ggc_mx (gcc::context *)): New.
3271 (gt_pch_nx (gcc::context *)): New.
3272 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3273 void *cookie)): New.
3274 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3276 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3278 * dce.c (fini_dce): Call df_analyze again just in case
3279 delete_unmarked_insns removed anything.
3281 2013-08-20 Teresa Johnson <tejohnson@google.com>
3283 PR rtl-optimizations/57451
3284 * final.c (reemit_insn_block_notes): Prevent lexical blocks
3285 from crossing split section boundaries.
3287 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
3289 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
3290 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
3291 with MULTLIB_DEFAULTS.
3293 2013-08-20 Nick Clifton <nickc@redhat.com>
3295 * target.def (narrow_volatile_bitfield): Note that the default
3296 value is false, not !TARGET_STRICT_ALIGN.
3297 * doc/tm.texi: Regenerate.
3299 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
3301 Fix LIB_SPEC for systems without libpthread.
3303 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
3304 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
3306 * config/i386/linux-common.h: Likewise.
3307 * config/mips/linux-common.h: Likewise.
3309 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
3311 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
3314 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3316 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3317 instances can own GC refs.
3319 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3320 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3321 (gcc::context::gt_pch_nx): Likewise.
3322 (gcc::context::gt_pch_nx): Likewise.
3323 * ggc.h (gt_ggc_mx <T>): New.
3324 (gt_pch_nx_with_op <T>): New.
3325 (gt_pch_nx <T>): New.
3326 * passes.c (opt_pass::gt_ggc_mx): New.
3327 (opt_pass::gt_pch_nx): New.
3328 (opt_pass::gt_pch_nx_with_op): New.
3329 (pass_manager::gt_ggc_mx): New.
3330 (pass_manager::gt_pch_nx): New.
3331 (pass_manager::gt_pch_nx_with_op): New.
3332 (pass_manager::operator new): Use
3333 ggc_internal_cleared_alloc_stat rather than xcalloc.
3334 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3335 (pass_manager::gt_ggc_mx): New.
3336 (pass_manager::gt_pch_nx): New.
3337 (pass_manager::gt_pch_nx_with_op): New.
3338 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3339 (opt_pass::operator new): New.
3340 (opt_pass::gt_ggc_mx): New.
3341 (opt_pass::gt_pch_nx): New.
3342 (opt_pass::gt_pch_nx_with_op): New.
3344 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3346 * Makefile.in (GTFILES): Add context.h.
3347 * context.c (gcc::context::operator new): New.
3348 (gcc::context::gt_ggc_mx): New.
3349 (gcc::context::gt_pch_nx): New.
3350 (gcc::context::gt_pch_nx): New.
3351 * context.h (gcc::context): Add GTY((user)) marking.
3352 (gcc::context::operator new): New.
3353 (gcc::context::gt_ggc_mx): New.
3354 (gcc::context::gt_pch_nx): New.
3355 (gcc::context::gt_pch_nx): New.
3356 (g): Add GTY marking.
3357 (gt_ggc_mx (gcc::context *)): New.
3358 (gt_pch_nx (gcc::context *)): New.
3359 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3360 void *cookie)): New.
3361 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3363 2013-08-20 Alan Modra <amodra@gmail.com>
3366 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
3367 (rs6000_emit_epilogue): Likewise.
3369 2013-08-19 Dehao Chen <dehao@google.com>
3371 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
3373 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
3374 Jakub Jelinek <jakub@redhat.com>
3376 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
3377 (BUILT_IN_FABSD64): Likewise.
3378 (BUILT_IN_FABSD128): Likewise.
3379 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
3380 (fold_builtin_1): Likewise.
3381 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
3382 destination and source operands.
3383 (*abstd2_fpr): Likewise.
3384 (*nabstd2_fpr): Likewise.
3386 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
3388 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
3391 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
3393 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
3396 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
3398 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
3399 * builtins.c (default_libc_has_function): New.
3400 (gnu_libc_has_function): Ditto.
3401 (no_c99_libc_has_function): Ditto.
3402 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
3403 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
3404 (fold_builtin_sincos): Likewise.
3405 (fold_builtin_cexp): Likewise.
3406 * builtins.def (DEF_C94_BUILTIN): Likewise.
3407 (DEF_C99_BUILTIN): Likewise.
3408 (DEF_C99_C90RES_BUILTIN): Likewise.
3409 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
3410 definitions to using this define.
3411 * config/darwin-protos.h (darwin_libc_has_function): New.
3412 * config/darwin.c (darwin_libc_has_function): Ditto.
3413 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
3414 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
3415 * config/darwin.h: Ditto.
3416 * config/elfos.h: Ditto.
3417 * config/freebsd.h: Ditto.
3418 * config/i386/cygming.h: Ditto.
3419 * config/i386/djgpp.h: Ditto.
3420 * config/i386/i386-interix.h: Ditto.
3421 * config/microblaze/microblaze.h: Ditto.
3422 * config/mmix/mmix.h: Ditto.
3423 * config/gnu-user.h: Ditto.
3424 * config/ia64/hpux.h: Ditto.
3425 * config/pa/pa-hpux.h: Ditto.
3426 * config/pdp11/pdp11.h: Ditto.
3427 * config/picochip/picochip.h: Ditto.
3428 * config/linux.h: Ditto.
3429 * config/netbsd.h: Ditto.
3430 * config/openbsd.h: Ditto.
3431 * config/rs6000/aix43.h: Ditto.
3432 * config/rs6000/aix51.h: Ditto.
3433 * config/rs6000/aix52.h: Ditto.
3434 * config/rs6000/aix53.h: Ditto.
3435 * config/rs6000/aix61.h: Ditto.
3436 * config/rs6000/darwin.h: Ditto.
3437 * config/rs6000/linux.h: Ditto.
3438 * config/rs6000/linux64.h: Ditto.
3439 * config/s390/tpf.h: Ditto.
3440 * config/sol2-10.h: Ditto.
3441 * config/sol2.h: Ditto.
3442 * config/vms/vms.h: Ditto.
3443 * config/vxworks.h: Ditto.
3444 * config/linux-android.c (linux_android_libc_has_function):
3445 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
3446 * config/linux-protos.h (linux_android_libc_has_function):
3448 * config/i386/i386.c (ix86_libc_has_function): New.
3449 * config/i386/i386-protos.h
3450 (ix86_libc_has_function): New declaration.
3451 * config/i386/i386.md
3452 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
3453 ("isinf<mode>2): Likewise.
3454 * convert.c (convert_to_integer): Using new target hook
3455 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
3456 TARGET_C99_FUNCTIONS.
3457 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
3458 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
3459 * coretypes.h (function_class): New enum for different
3460 classes of functions.
3461 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
3462 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
3463 (TARGET_HAS_SINCOS): Likewise.
3464 (TARGET_LIBC_HAS_FUNCTION): New.
3465 * doc/tm.texi: Regenerated.
3466 * targhooks.h (default_libc_has_function): New declaration.
3467 (no_c99_libc_has_function): Ditto.
3468 (gnu_libc_has_function): Ditto.
3469 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
3470 and TARGET_HAS_SINCOS.
3472 2013-08-18 Jan Hubicka <jh@suse.cz>
3474 * Makeifle-in (ipa-devirt.o): New.
3475 (GTFILES): Add ipa-utils.h and ipa-devirt.c
3476 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
3477 (analyze_functions): Look into possible targets of polymorphic call.
3478 * dumpfile.c (dump_files): Add type-inheritance dump.
3479 * dumpfile.h (TDI_inheritance): New.
3480 * ipa-devirt.c: New file.
3481 * ipa-utils.h (odr_type_d): Forward declare.
3482 (odr_type): New type.
3483 (build_type_inheritance_graph): Declare.
3484 (possible_polymorphic_call_targets): Declare and introduce inline
3485 variant when only edge is pased.
3486 (dump_possible_polymorphic_call_targets): Likewise.
3487 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3488 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3489 (types_same_for_odr): ... here.
3490 * tree.h (type_in_anonymous_namespace_p): Declare.
3492 2013-08-18 Jakub Jelinek <jakub@redhat.com>
3494 PR tree-optimization/58006
3495 * tree-parloops.c (take_address_of): Don't ICE if get_name
3497 (eliminate_local_variables_stmt): Remove clobber stmts.
3499 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3501 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3502 error message is issued for an alias to undefined symbol.
3504 2013-08-18 Jan Hubicka <jh@suse.cz>
3506 * cgraph.c (cgraph_create_indirect_edge): Discover
3507 polymorphic calls and record basic info into indirect_info.
3508 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3509 devirtualization, ignore objc function calls.
3510 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3511 call with no parm index info.
3512 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3513 * tree.c (virtual_method_call_p): New function.
3514 * tree.h (virtual_method_call_p): Declare.
3516 2013-08-16 Jan Hubicka <jh@suse.cz>
3519 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3521 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
3523 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3526 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3528 * gengtype.c (type_for_name): Add special-case support for
3529 locating types within the "gcc::" namespace.
3530 (open_base_files): Emit a "using namespace gcc" directive.
3532 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3535 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3536 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3538 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3539 array instead of each individual operand as a separate argument.
3540 (emit_fusion_gpr_load): Likewise.
3541 (expand_fusion_gpr_load): Add new function declaration.
3543 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3544 signature to have the operands passed as an array, instead of as
3545 separate arguments. Allow ZERO_EXTEND to be in the memory
3546 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3547 depend on the register live/dead flags when peepholes are run.
3548 (expand_fusion_gpr_load): New function to be called from the
3549 peephole2 pass, to change the register that addis sets to be the
3551 (emit_fusion_gpr_load): Change the calling signature to have the
3552 operands passed as an array, instead of as separate arguments.
3553 Allow ZERO_EXTEND to be in the memory address, and also
3554 SIGN_EXTEND if -mpower8-fusion-sign.
3556 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3558 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3559 adjust the register addis loads up in the peephole2 pass. Do not
3560 depend on the register live/dead state when the peephole pass is done.
3562 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3564 * gengtype.c (create_user_defined_type): Ensure that the kind
3565 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3566 declaration is seen before the GTY((user)) marking.
3568 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
3571 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
3573 2013-08-16 Jan Hubicka <jh@suse.cz>
3575 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
3576 arugment expected_type.
3577 (gimple_fold_call): Use it.
3578 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
3579 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
3580 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
3581 (try_make_edge_direct_virtual_call): Likewise.
3582 * tree.c (obj_type_ref_class): New.
3583 * tree.h (obj_type_ref_class): Use it.
3585 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
3587 * sched-vis.c (rtl_slim_pp_initialized): Remove.
3588 (rtl_slim_pp): Likewise.
3589 (init_rtl_slim_pretty_print): Likewise.
3590 (dump_value_slim): Don't call it. Use local pretty printer.
3591 (dump_insn_slim): Likewise.
3592 (dump_rtl_slim): Likewise.
3593 (str_pattern_slim): Likewise.
3594 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
3597 2013-08-16 Jakub Jelinek <jakub@redhat.com>
3599 PR tree-optimization/58164
3600 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
3601 walk gimple_goto_dest of GIMPLE_GOTO.
3603 PR tree-optimization/58165
3604 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
3605 bi_call must be the last stmt in a bb, don't split_block, instead
3606 use fallthru edge from it and give up if there is none.
3607 Release conds vector when returning early.
3609 2013-08-14 Xinliang David Li <davidxl@google.com>
3611 * config/i386/i386.c (ix86_option_override_internal):
3612 Remove unused variable and field.
3614 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3617 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
3618 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
3619 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
3620 and Linux, correct BLKmode alignment when 128-bit alignment is
3621 required and compatibility flag is not set.
3622 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
3623 for zero-size arguments when compatibility flag is not set.
3625 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3627 PR tree-optimization/58145
3628 * tree-sra.c (build_ref_for_offset): If prev_base has
3629 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
3631 2013-08-14 Xinliang David Li <davidxl@google.com>
3633 * config/i386/i386.c (ix86_option_override_internal):
3634 Fix uninitialized variable error.
3636 2013-08-14 Xinliang David Li <davidxl@google.com>
3638 * config/i386/i386.opt: Define two new options.
3639 * config/i386/x86-tune.def: Add arch selector field in macros.
3640 * config/i386/i386.h: Adjust macro definition.
3641 * config/i386/i386.c (ix86_option_override_internal):
3643 (parse_mtune_ctrl_str): New function.
3644 (set_ix86_tune_features): New function.
3645 (ix86_function_specific_restore): Call the new helper function.
3647 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
3649 PR rtl-optimization/57662
3650 * sel-sched.c (code_motion_process_successors): When the current insn
3651 is removed after the recursive traversal, break from the loop.
3652 Add comments and debug printouts.
3654 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3655 Alexandre Oliva <aoliva@redhat.com>
3658 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3659 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3660 there also UNSPEC_PLTOFF.
3662 2013-08-14 Marek Polacek <polacek@redhat.com>
3664 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3665 NUM_CONDITIONS bit positions.
3667 2013-08-13 Cary Coutant <ccoutant@google.com>
3669 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3670 (attr_checksum): Hash vector contents instead of pointer.
3671 (attr_checksum_ordered): Likewise.
3673 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
3675 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3676 when Pmode != word_mode. Add length_address attribute.
3677 (sse3_monitor_<mode>): Merge from sse3_monitor and
3678 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3679 Pmode != word_mode. Update insn length attribute.
3680 * config/i386/i386.c (ix86_option_override_internal): Update
3681 ix86_gen_monitor selection for merged sse3_monitor insn.
3683 2013-08-13 Julian Brown <julian@codesourcery.com>
3685 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3686 perform invalid legitimization on greater-than-word-size modes for
3689 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3691 * ira.c (setup_class_translate_array): Use aclass instead of cl
3692 for classes not fully covered by allocno classes.
3694 2013-08-13 Jakub Jelinek <jakub@redhat.com>
3696 PR tree-optimization/57661
3697 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3698 * tree-inline.c (tree_function_versioning): Initialize it.
3699 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3700 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3701 that is not being copied.
3704 * asan.c (instrument_strlen_call): Fix typo in comment.
3705 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3707 2013-08-13 Steve Ellcey <sellcey@mips.com>
3709 * config/mips/mips.md (prefetch): Use lw instead of ld on
3710 loongson in 32bit mode.
3712 2013-08-13 Nick Clifton <nickc@redhat.com>
3714 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3716 2013-08-13 Jan Hubicka <jh@suse.cz>
3718 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3719 introduced edge; fix typo in sanity check.
3720 (cgraph_resolve_speculation): Export; improve diagnostic.
3721 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3722 speculation at type mismatch.
3723 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3724 (cgraph_resolve_speculation): Declare.
3725 (symtab_can_be_discarded): New function.
3726 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3727 * ipa-inline-transform.c (speculation_removed): New global var.
3728 (clone_inlined_nodes): See if speculation can be removed.
3729 (inline_call): If speculations was removed, we growths may not match.
3730 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3731 (speculation_useful_p): New function.
3732 (resolve_noninline_speculation): New function.
3733 (inline_small_functions): Resolve useless speculations.
3734 * ipa-inline.h (speculation_useful_p): Declare
3735 * ipa.c (can_replace_by_local_alias): Simplify.
3736 (ipa_profile): Produce speculative calls in non-lto, too;
3737 add simple cost model; produce local aliases.
3739 2013-08-13 David Malcolm <dmalcolm@redhat.com>
3741 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
3744 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3746 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3748 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
3750 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3751 iterator instead of X87MODEF.
3753 2013-08-12 Perez Read <netfirewall@gmail.com>
3756 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3757 operand 0 for intel asm alternative.
3758 (*movabs<mode>_2): Ditto for operand 1.
3760 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3762 * config/aarch64/arm_none.h
3763 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3765 2013-08-12 Nick Clifton <nickc@redhat.com>
3767 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3769 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3771 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3772 expand for QI/HImode operand to produce more effictive code for
3773 unsigned char(short) --> float(double) conversion.
3775 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
3777 * doc/invoke.texi: Mention that -ftls-model does not force the final
3780 2013-08-12 Marek Polacek <polacek@redhat.com>
3781 Marc Glisse <marc.glisse@inria.fr>
3783 PR tree-optimization/57980
3784 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3785 when creating -1 constant.
3787 2013-08-10 Jan Hubicka <jh@suse.cz>
3789 Workaround binutils PR14342.
3790 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3791 (gimple_init_edge_profiler): Likewise.
3792 (gimple_gen_ic_func_profiler): Likewise.
3794 2013-08-09 Jan Hubicka <jh@suse.cz>
3796 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3798 2013-08-09 Xinliang David Li <davidxl@google.com>
3800 * config/i386/stringop.def: New file.
3801 * config/i386/stringop.opt: New file.
3802 * config/i386/i386-opts.h: Include stringopt.def.
3803 * config/i386/i386.opt: Include stringopt.opt.
3804 * config/i386/i386.c (ix86_option_override_internal):
3805 Override default size based stringop inline strategies with options.
3806 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3809 2013-08-09 Jan Hubicka <jh@suse.cz>
3811 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
3813 2013-08-09 Jan Hubicka <jh@suse.cz>
3815 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3817 (dump_cgraph_node): Dump profile-id.
3818 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3819 and common_target_probability.
3820 * lto-cgraph.c (lto_output_edge): Stream common targets.
3821 (lto_output_node): Stream profile ids.
3822 (input_node): Stream profile ids.
3823 (input_edge): Stream common targets.
3824 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3825 * ipa.c: Include value-prof.h
3826 (ipa_profile_generate_summary): Turn indirect call statement histograms
3827 into common targets.
3828 (ipa_profile): Turn common targets into speculative edges.
3830 2013-08-09 Jan Hubicka <jh@suse.cz>
3832 * cgraph.h (cgraph_node): Add profile_id.
3833 * value-prof.c (cgraph_node_map): Turn into pointer_map.
3834 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
3835 (del_node_map): Update.
3836 (find_func_by_funcdef_no): Replace by ...
3837 (find_func_by_profile_id): ... this one.
3838 (gimple_ic_transform): Do not remove useful histograms when
3839 speculation is not done; dump info when indirect call removal
3841 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
3842 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
3843 (__gcov_indirect_call_profiler_v2): .. this one.
3844 * profile.h (init_node_map): Update.
3845 * coverage.c (coverage_compute_profile_id): New function.
3846 * coverage.h (coverage_compute_profile_id): Declare.
3847 * tree-profile.c (init_ic_make_global_vars): Make
3848 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
3849 (gimple_init_edge_profiler): Update prototype of
3850 __gcov_indirect_call_profiler.
3851 (gimple_gen_ic_func_profiler): Simplify.
3852 (tree_profiling): Use init_node_map
3854 2013-08-09 Jan Hubicka <jh@suse.cz>
3856 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
3857 non-speculative refs.
3858 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
3859 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
3860 (cgraph_set_call_stmt): Likewise.
3861 (cgraph_create_edge_1): Fix release checking compilatoin;
3863 (cgraph_free_edge): Free indirect info.
3864 (cgraph_turn_edge_to_speculative): New function.
3865 (cgraph_speculative_call_info): New function.
3866 (cgraph_make_edge_direct): Return direct edge; handle speculation.
3867 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
3868 (dump_cgraph_node): Dump speculation.
3869 (verify_edge_count_and_frequency): Accept speculative edges.
3870 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
3871 (verify_cgraph_node): Handle speculation.
3872 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
3873 (cgraph_set_call_stmt): Update prototype.
3874 (cgraph_make_edge_direct): Update prototype.
3875 (cgraph_speculative_call_info): Declare.
3876 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
3877 to change; update call of ipa_find_references.
3878 * ipa-ref.c (ipa_record_reference): Fix return value; clear
3879 lto_stmt_uid and speculative flags.
3880 (ipa_dump_references): Dump speculation.
3881 (ipa_clone_references): Clone speculative flag.
3882 (ipa_clone_referring): Likewise.
3883 (ipa_clone_ref): New function.
3884 (ipa_find_reference): Look into lto_stmt_uids
3885 (ipa_clear_stmts_in_references): Do not clear speculative calls.
3886 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
3887 (ipa_find_reference): Update declaration.
3888 (ipa_clone_ref): Declare.
3889 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
3890 stream speculative flag.
3891 (lto_output_ref): Stream statements uids and speculation.
3892 (input_ref): Likewise.
3893 (input_edge): Stream speuclation.
3894 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
3895 (cgraph_set_call_stmt_including_clones): Handle speculation.
3896 * ipa-inline.c (heap_edge_removal_hook): New function.
3897 (inline_small_functions): Register it.
3898 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
3899 also initialize refs.
3900 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
3902 (try_make_edge_direct_simple_call): Likewise.
3903 (try_make_edge_direct_simple_call): Likewise.
3904 (update_indirect_edges_after_inlining): Likewise.
3905 (remove_described_reference): Look proper lto_stmt_uid.
3906 (propagate_controlled_uses): Likewise.
3907 (propagate_controlled_uses): Liekwise.
3908 * tree-inline.c (copy_bb): Copy speculative edges.
3909 (redirect_all_calls): New function.
3910 (copy_cfg_body): Do redirection after loop info is updated.
3911 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
3913 2013-08-09 Jan Hubicka <jh@suse.cz>
3915 * lto-streamer-out.c (output_function): Renumber PHIs.
3916 * lto-streamer-in.c (input_function): Likewise.
3918 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
3920 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
3921 (get_lane_unsigned): Likewise.
3922 (dup_lane_scalar): Likewise.
3923 (get_lane): enable for VALL.
3924 * config/aarch64/aarch64-simd.md
3925 (aarch64_dup_lane_scalar<mode>): Remove.
3926 (aarch64_get_lane_signed<mode>): Likewise.
3927 (aarch64_get_lane_unsigned<mode>): Likewise.
3928 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
3929 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
3930 (aarch64_get_lane<mode>): Enable for all vector modes.
3931 (aarch64_get_lanedi): Remove misleading constraints.
3932 * config/aarch64/arm_neon.h
3933 (__aarch64_vget_lane_any): Define.
3934 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
3935 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
3936 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
3937 * config/aarch64/iterators.md (VDQQH): New.
3941 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
3943 * configure.ac: Add GAS check for LEON instructions on SPARC.
3944 * configure: Regenerate.
3945 * config.in: Likewise.
3946 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
3948 * config/sparc/sparc.opt (LEON, LEON3): New masks.
3949 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
3951 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
3952 (AS_LEON_FLAG): New macro.
3953 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
3954 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
3955 Deal with LEON and LEON3 for the memory model.
3956 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
3957 (atomic_compare_and_swap<mode>_1): Likewise.
3958 (*atomic_compare_and_swap<mode>_1): Likewise.
3960 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3962 * config/arm/neon.md (vcond): Fix floating-point vector
3963 comparisons against 0.
3965 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
3967 * lra-constraints.c (emit_spill_move): Remove assert.
3968 (process_alt_operands): Add more debugging
3969 output. Increase reject for spilling into memory. Decrease
3970 reject for reloading scratch.
3971 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
3973 2013-08-08 Steve Ellcey <sellcey@mips.com>
3975 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
3976 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
3977 micromips incompatible. Add nan2008.
3978 (MULTILIB_DIRNAMES): Add nan2008.
3979 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3980 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
3981 and micromips incompatible. Add nan2008.
3982 (MULTILIB_DIRNAMES): Add nan2008.
3983 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3985 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
3987 PR rtl-optimization/58079
3988 * combine.c (combine_simplify_rtx): Avoid using SUBST if
3989 simplify_comparison has widened a comparison with an integer.
3991 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3993 * config/arm/neon.md (movmisalign<mode>): Disable when we
3994 don't allow unaligned accesses.
3995 (*movmisalign<mode>_neon_store): Likewise.
3996 (*movmisalign<mode>_neon_load): Likewise.
3997 (*movmisalign<mode>_neon_store): Likewise.
3998 (*movmisalign<mode>_neon_load): Likewise.
4000 2013-08-08 Jan Hubicka <jh@suse.cz>
4002 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
4003 (make_pass_rebuild_cgraph_edges): Also clear references.
4004 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
4005 * ipa-inline-transform.c (inline_transform): Remove all references
4007 * cgraphunit.c (expand_function): Remove all references after
4009 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
4010 (ipa_find_reference): Rewrite to iterator.
4011 (remove_stmt_references): Likewise.
4012 (ipa_clear_stmts_in_references): New function.
4013 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
4014 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
4016 * ipa-split.c (split_function): Remove references in split function.
4018 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4021 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
4022 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
4024 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4027 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
4028 suggested mode for the assignment isn't compatible with the
4031 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
4034 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
4036 2013-08-07 Xinliang David Li <davidxl@google.com>
4038 * config/i386/i386.opt: New option -mtune-ctrl=.
4039 * config/i386/x86-tune.def: New file.
4040 * config/i386/i386.h: include x86-tune.def.
4041 * config/i386/i386.c (ix86_option_override_internal):
4042 Parsing -mtune-ctrl= option and set tune features.
4044 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
4047 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
4048 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
4051 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4053 * rtl.h (update_alignments): Declare.
4054 * final.c (grow_label_align): New function extracted from...
4055 (shorten_branches): ...here. Call it.
4056 (update_alignments): New function.
4057 * reorg.c (sibling_labels): New variable.
4058 (get_label_before): Add SIBLING parameter. If it is non-zero, push
4059 the new label along with it onto the sibling_labels vector.
4060 (fill_simple_delay_slots): Adjust call to get_label_before.
4061 (fill_slots_from_thread): Likewise.
4062 (relax_delay_slots): Likewise.
4063 (make_return_insns): Likewise.
4064 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
4066 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4068 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
4069 document its semantics.
4070 (diagnostic_report_diagnostic): Adjust accordingly.
4072 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4074 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
4075 (sparc_option_override): ...and port to new C++ pass API.
4076 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
4078 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
4080 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
4082 2013-08-06 Caroline Tice <cmtice@google.com>
4084 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4085 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
4086 * tree-pass.h: Add pass_vtable_verify.
4087 * varasm.c (assemble_variable): Add code to properly set the comdat
4088 section and name for the .vtable_map_vars section.
4089 (assemble_vtyv_preinit_initializer): New function.
4090 (default_sectin_type_flags): Make sure .vtable_map_vars section has
4092 * output.h: Add function decl for assemble_vtv_preinit_initializer.
4093 * vtable-verify.c: New file.
4094 * vtable-verify.h: New file.
4095 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
4096 initialiation levels.
4097 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
4098 * passes.def: Insert pass_vtable_verify.
4099 * aclocal.m4: Reorder includes.
4100 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
4101 -fvtv-counts options.
4102 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
4103 as appropriate, if -fvtable-verify=... is used.
4104 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
4105 -fvtable-verify=... is used.
4106 * Makefile.in (OBJS): Add vtable-verify.o to list.
4107 (vtable-verify.o): Add new build rule.
4108 (GTFILES): Add vtable-verify.c to list.
4109 * common.opt (fvtable-verify=): New flag.
4110 (vtv_priority): Values for fvtable-verify= flag.
4111 (fvtv-counts): New flag.
4112 (fvtv-debug): New flag.
4113 * tree.h (save_vtable_map_decl): New extern function decl.
4115 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4117 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
4118 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
4119 (pass_data_rl78_devirt): ...new pass_data instance and...
4120 (make_pass_rl78_devirt): ...new function.
4121 (rl78_asm_file_start): Port pass registration to new C++ API.
4123 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4125 * coretypes.h (rtl_opt_pass): Add.
4126 (gcc::context): Add.
4127 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
4128 (epiphany_init): Port to new C++ pass API.
4129 (epiphany_optimize_mode_switching): Likewise.
4130 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
4131 (pass_manager::get_pass_mode_switching): New.
4132 (pass_manager::get_pass_peephole2): New.
4133 * mode-switching.c (pass_mode_switching): Add clone method.
4134 * recog.c (pass_peephole2): Add clone method.
4135 (pass_split_all_insns): Add clone method.
4137 2013-08-06 David Malcolm <dmalcolm@redhat.com>
4139 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
4140 (mips_option_override): ...here, porting to new C++ API for passes.
4142 2013-08-06 Jan Hubicka <jh@suse.cz>
4144 * cgraph.c (cgraph_get_body): New function based on lto.c
4146 * cgraph.h (cgraph_get_body): Declare.
4147 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
4149 * cgraphunit.c (expand_function): Get body prior expanding.
4150 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
4151 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
4153 * passes.c (do_per_function_toporder): Get body.
4154 * tree-inline.c (expand_call_inline): Get body prior inlining it.
4155 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
4157 2013-08-06 Martin Jambor <mjambor@suse.cz>
4160 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
4161 is not re-finalized. Rename second parameter to no_collect.
4163 2013-08-06 Martin Jambor <mjambor@suse.cz>
4166 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
4167 MEM_REF has proper alignment information.
4169 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
4172 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
4174 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
4175 rtx (*) (rtx, ...) with insn_gen_fn.
4176 * genoutput.c (output_insn_data): Cast gen_? function pointers to
4177 insn_gen_fn::stored_funcptr. Add initializer braces.
4179 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4181 Rewrite how instances of passes are cloned to remove assumptions
4182 about their sizes (thus allowing pass subclasses to have
4183 additional data fields, albeit non-GC-managed ones at this point).
4185 * passes.c (make_pass_instance): Now that passes have clone
4186 methods, rewrite this function to eliminate XNEW and memcpy
4187 calls that used hardcoded sizes. Since this function no longer
4188 creates pass instances, rename it to...
4189 (add_pass_instance): ...this. Document the old way that passes were
4190 numbered and flagged, and rework this function to continue using it.
4191 (next_pass_1): Add an initial_pass argument for use by
4193 (position_pass): When adding multiple instances of a pass, use
4194 the pass's clone method, rather than relying on the XNEW/memcpy
4195 within the former make_pass_instance (now add_pass_instance).
4196 (pass_manager::pass_manager): When invoking next_pass_1, also supply
4197 the initial instance of the current pass within the pass manager.
4199 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4201 This is the automated part of the conversion of passes from C
4202 structs to C++ classes.
4204 Patch autogenerated by refactor_passes.py from
4205 https://github.com/davidmalcolm/gcc-refactoring-scripts
4206 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
4208 * asan.c (pass_asan): Convert from a global struct to a subclass of
4209 gimple_opt_pass along with...
4210 (pass_data_asan): ...new pass_data instance and...
4211 (make_pass_asan): ...new function.
4212 (pass_asan_O0): Convert from a global struct to a subclass of
4213 gimple_opt_pass along with...
4214 (pass_data_asan_O0): ...new pass_data instance and...
4215 (make_pass_asan_O0): ...new function.
4216 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
4217 subclass of rtl_opt_pass along with...
4218 (pass_data_inc_dec): ...new pass_data instance and...
4219 (make_pass_inc_dec): ...new function.
4220 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
4221 a subclass of rtl_opt_pass along with...
4222 (pass_data_reorder_blocks): ...new pass_data instance and...
4223 (make_pass_reorder_blocks): ...new function.
4224 (pass_duplicate_computed_gotos): Convert from a global struct to a
4225 subclass of rtl_opt_pass along with...
4226 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
4227 (make_pass_duplicate_computed_gotos): ...new function.
4228 (pass_partition_blocks): Convert from a global struct to a subclass of
4229 rtl_opt_pass along with...
4230 (pass_data_partition_blocks): ...new pass_data instance and...
4231 (make_pass_partition_blocks): ...new function.
4232 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
4233 struct to a subclass of rtl_opt_pass along with...
4234 (pass_data_branch_target_load_optimize1): ...new pass_data instance
4236 (make_pass_branch_target_load_optimize1): ...new function.
4237 (pass_branch_target_load_optimize2): Convert from a global struct to a
4238 subclass of rtl_opt_pass along with...
4239 (pass_data_branch_target_load_optimize2): ...new pass_data instance
4241 (make_pass_branch_target_load_optimize2): ...new function.
4242 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
4243 of rtl_opt_pass along with...
4244 (pass_data_jump): ...new pass_data instance and...
4245 (make_pass_jump): ...new function.
4246 (pass_jump2): Convert from a global struct to a subclass of
4247 rtl_opt_pass along with...
4248 (pass_data_jump2): ...new pass_data instance and...
4249 (make_pass_jump2): ...new function.
4250 * cfgexpand.c (pass_expand): Convert from a global struct to a
4251 subclass of rtl_opt_pass along with...
4252 (pass_data_expand): ...new pass_data instance and...
4253 (make_pass_expand): ...new function.
4254 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
4255 of rtl_opt_pass along with...
4256 (pass_data_free_cfg): ...new pass_data instance and...
4257 (make_pass_free_cfg): ...new function.
4258 (pass_into_cfg_layout_mode): Convert from a global struct to a
4259 subclass of rtl_opt_pass along with...
4260 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
4261 (make_pass_into_cfg_layout_mode): ...new function.
4262 (pass_outof_cfg_layout_mode): Convert from a global struct to a
4263 subclass of rtl_opt_pass along with...
4264 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
4265 (make_pass_outof_cfg_layout_mode): ...new function.
4266 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
4267 struct to a subclass of gimple_opt_pass along with...
4268 (pass_data_build_cgraph_edges): ...new pass_data instance and...
4269 (make_pass_build_cgraph_edges): ...new function.
4270 (pass_rebuild_cgraph_edges): Convert from a global struct to a
4271 subclass of gimple_opt_pass along with...
4272 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
4273 (make_pass_rebuild_cgraph_edges): ...new function.
4274 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
4275 subclass of gimple_opt_pass along with...
4276 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
4278 (make_pass_remove_cgraph_callee_edges): ...new function.
4279 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
4280 struct to a subclass of rtl_opt_pass along with...
4281 (pass_data_stack_adjustments): ...new pass_data instance and...
4282 (make_pass_stack_adjustments): ...new function.
4283 * combine.c (pass_combine): Convert from a global struct to a subclass
4284 of rtl_opt_pass along with...
4285 (pass_data_combine): ...new pass_data instance and...
4286 (make_pass_combine): ...new function.
4287 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
4288 global struct to a subclass of rtl_opt_pass along with...
4289 (pass_data_compare_elim_after_reload): ...new pass_data instance
4291 (make_pass_compare_elim_after_reload): ...new function.
4292 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
4293 of rtl_opt_pass along with...
4294 (pass_data_rtl_cprop): ...new pass_data instance and...
4295 (make_pass_rtl_cprop): ...new function.
4296 * cse.c (pass_cse): Convert from a global struct to a subclass of
4297 rtl_opt_pass along with...
4298 (pass_data_cse): ...new pass_data instance and...
4299 (make_pass_cse): ...new function.
4300 (pass_cse2): Convert from a global struct to a subclass of
4301 rtl_opt_pass along with...
4302 (pass_data_cse2): ...new pass_data instance and...
4303 (make_pass_cse2): ...new function.
4304 (pass_cse_after_global_opts): Convert from a global struct to a
4305 subclass of rtl_opt_pass along with...
4306 (pass_data_cse_after_global_opts): ...new pass_data instance and...
4307 (make_pass_cse_after_global_opts): ...new function.
4308 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
4309 of rtl_opt_pass along with...
4310 (pass_data_ud_rtl_dce): ...new pass_data instance and...
4311 (make_pass_ud_rtl_dce): ...new function.
4312 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
4313 rtl_opt_pass along with...
4314 (pass_data_fast_rtl_dce): ...new pass_data instance and...
4315 (make_pass_fast_rtl_dce): ...new function.
4316 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
4317 a subclass of rtl_opt_pass along with...
4318 (pass_data_df_initialize_opt): ...new pass_data instance and...
4319 (make_pass_df_initialize_opt): ...new function.
4320 (pass_df_initialize_no_opt): Convert from a global struct to a
4321 subclass of rtl_opt_pass along with...
4322 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
4323 (make_pass_df_initialize_no_opt): ...new function.
4324 (pass_df_finish): Convert from a global struct to a subclass of
4325 rtl_opt_pass along with...
4326 (pass_data_df_finish): ...new pass_data instance and...
4327 (make_pass_df_finish): ...new function.
4328 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
4329 rtl_opt_pass along with...
4330 (pass_data_rtl_dse1): ...new pass_data instance and...
4331 (make_pass_rtl_dse1): ...new function.
4332 (pass_rtl_dse2): Convert from a global struct to a subclass of
4333 rtl_opt_pass along with...
4334 (pass_data_rtl_dse2): ...new pass_data instance and...
4335 (make_pass_rtl_dse2): ...new function.
4336 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
4337 subclass of rtl_opt_pass along with...
4338 (pass_data_dwarf2_frame): ...new pass_data instance and...
4339 (make_pass_dwarf2_frame): ...new function.
4340 * except.c (pass_set_nothrow_function_flags): Convert from a global
4341 struct to a subclass of rtl_opt_pass along with...
4342 (pass_data_set_nothrow_function_flags): ...new pass_data instance
4344 (make_pass_set_nothrow_function_flags): ...new function.
4345 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
4346 subclass of rtl_opt_pass along with...
4347 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
4349 (make_pass_convert_to_eh_region_ranges): ...new function.
4350 * final.c (pass_compute_alignments): Convert from a global struct to a
4351 subclass of rtl_opt_pass along with...
4352 (pass_data_compute_alignments): ...new pass_data instance and...
4353 (make_pass_compute_alignments): ...new function.
4354 (pass_final): Convert from a global struct to a subclass of
4355 rtl_opt_pass along with...
4356 (pass_data_final): ...new pass_data instance and...
4357 (make_pass_final): ...new function.
4358 (pass_shorten_branches): Convert from a global struct to a subclass of
4359 rtl_opt_pass along with...
4360 (pass_data_shorten_branches): ...new pass_data instance and...
4361 (make_pass_shorten_branches): ...new function.
4362 (pass_clean_state): Convert from a global struct to a subclass of
4363 rtl_opt_pass along with...
4364 (pass_data_clean_state): ...new pass_data instance and...
4365 (make_pass_clean_state): ...new function.
4366 * function.c (pass_instantiate_virtual_regs): Convert from a global
4367 struct to a subclass of rtl_opt_pass along with...
4368 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
4369 (make_pass_instantiate_virtual_regs): ...new function.
4370 (pass_leaf_regs): Convert from a global struct to a subclass of
4371 rtl_opt_pass along with...
4372 (pass_data_leaf_regs): ...new pass_data instance and...
4373 (make_pass_leaf_regs): ...new function.
4374 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
4375 subclass of rtl_opt_pass along with...
4376 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
4378 (make_pass_thread_prologue_and_epilogue): ...new function.
4379 (pass_match_asm_constraints): Convert from a global struct to a
4380 subclass of rtl_opt_pass along with...
4381 (pass_data_match_asm_constraints): ...new pass_data instance and...
4382 (make_pass_match_asm_constraints): ...new function.
4383 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
4384 subclass of rtl_opt_pass along with...
4385 (pass_data_rtl_fwprop): ...new pass_data instance and...
4386 (make_pass_rtl_fwprop): ...new function.
4387 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
4388 rtl_opt_pass along with...
4389 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
4390 (make_pass_rtl_fwprop_addr): ...new function.
4391 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
4392 rtl_opt_pass along with...
4393 (pass_data_rtl_pre): ...new pass_data instance and...
4394 (make_pass_rtl_pre): ...new function.
4395 (pass_rtl_hoist): Convert from a global struct to a subclass of
4396 rtl_opt_pass along with...
4397 (pass_data_rtl_hoist): ...new pass_data instance and...
4398 (make_pass_rtl_hoist): ...new function.
4399 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
4400 subclass of gimple_opt_pass along with...
4401 (pass_data_lower_cf): ...new pass_data instance and...
4402 (make_pass_lower_cf): ...new function.
4403 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
4404 from a global struct to a subclass of gimple_opt_pass along with...
4405 (pass_data_strength_reduction): ...new pass_data instance and...
4406 (make_pass_strength_reduction): ...new function.
4407 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
4408 of rtl_opt_pass along with...
4409 (pass_data_rtl_ifcvt): ...new pass_data instance and...
4410 (make_pass_rtl_ifcvt): ...new function.
4411 (pass_if_after_combine): Convert from a global struct to a subclass of
4412 rtl_opt_pass along with...
4413 (pass_data_if_after_combine): ...new pass_data instance and...
4414 (make_pass_if_after_combine): ...new function.
4415 (pass_if_after_reload): Convert from a global struct to a subclass of
4416 rtl_opt_pass along with...
4417 (pass_data_if_after_reload): ...new pass_data instance and...
4418 (make_pass_if_after_reload): ...new function.
4419 * init-regs.c (pass_initialize_regs): Convert from a global struct to
4420 a subclass of rtl_opt_pass along with...
4421 (pass_data_initialize_regs): ...new pass_data instance and...
4422 (make_pass_initialize_regs): ...new function.
4423 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
4424 of ipa_opt_pass_d along with...
4425 (pass_data_ipa_cp): ...new pass_data instance and...
4426 (make_pass_ipa_cp): ...new function.
4427 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
4428 global struct to a subclass of gimple_opt_pass along with...
4429 (pass_data_inline_parameters): ...new pass_data instance and...
4430 (make_pass_inline_parameters): ...new function.
4431 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
4432 subclass of gimple_opt_pass along with...
4433 (pass_data_early_inline): ...new pass_data instance and...
4434 (make_pass_early_inline): ...new function.
4435 (pass_ipa_inline): Convert from a global struct to a subclass of
4436 ipa_opt_pass_d along with...
4437 (pass_data_ipa_inline): ...new pass_data instance and...
4438 (make_pass_ipa_inline): ...new function.
4439 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
4440 struct to a subclass of gimple_opt_pass along with...
4441 (pass_data_local_pure_const): ...new pass_data instance and...
4442 (make_pass_local_pure_const): ...new function.
4443 (pass_ipa_pure_const): Convert from a global struct to a subclass of
4444 ipa_opt_pass_d along with...
4445 (pass_data_ipa_pure_const): ...new pass_data instance and...
4446 (make_pass_ipa_pure_const): ...new function.
4447 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
4448 to a subclass of ipa_opt_pass_d along with...
4449 (pass_data_ipa_reference): ...new pass_data instance and...
4450 (make_pass_ipa_reference): ...new function.
4451 * ipa-split.c (pass_split_functions): Convert from a global struct to
4452 a subclass of gimple_opt_pass along with...
4453 (pass_data_split_functions): ...new pass_data instance and...
4454 (make_pass_split_functions): ...new function.
4455 (pass_feedback_split_functions): Convert from a global struct to a
4456 subclass of gimple_opt_pass along with...
4457 (pass_data_feedback_split_functions): ...new pass_data instance and...
4458 (make_pass_feedback_split_functions): ...new function.
4459 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
4460 global struct to a subclass of simple_ipa_opt_pass along with...
4461 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
4463 (make_pass_ipa_function_and_variable_visibility): ...new function.
4464 (pass_ipa_free_inline_summary): Convert from a global struct to a
4465 subclass of simple_ipa_opt_pass along with...
4466 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
4467 (make_pass_ipa_free_inline_summary): ...new function.
4468 (pass_ipa_whole_program_visibility): Convert from a global struct to a
4469 subclass of ipa_opt_pass_d along with...
4470 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
4472 (make_pass_ipa_whole_program_visibility): ...new function.
4473 (pass_ipa_profile): Convert from a global struct to a subclass of
4474 ipa_opt_pass_d along with...
4475 (pass_data_ipa_profile): ...new pass_data instance and...
4476 (make_pass_ipa_profile): ...new function.
4477 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
4478 ipa_opt_pass_d along with...
4479 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4480 (make_pass_ipa_cdtor_merge): ...new function.
4481 * ira.c (pass_ira): Convert from a global struct to a subclass of
4482 rtl_opt_pass along with...
4483 (pass_data_ira): ...new pass_data instance and...
4484 (make_pass_ira): ...new function.
4485 (pass_reload): Convert from a global struct to a subclass of
4486 rtl_opt_pass along with...
4487 (pass_data_reload): ...new pass_data instance and...
4488 (make_pass_reload): ...new function.
4489 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4490 subclass of rtl_opt_pass along with...
4491 (pass_data_cleanup_barriers): ...new pass_data instance and...
4492 (make_pass_cleanup_barriers): ...new function.
4493 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4494 of rtl_opt_pass along with...
4495 (pass_data_loop2): ...new pass_data instance and...
4496 (make_pass_loop2): ...new function.
4497 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4498 rtl_opt_pass along with...
4499 (pass_data_rtl_loop_init): ...new pass_data instance and...
4500 (make_pass_rtl_loop_init): ...new function.
4501 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4502 rtl_opt_pass along with...
4503 (pass_data_rtl_loop_done): ...new pass_data instance and...
4504 (make_pass_rtl_loop_done): ...new function.
4505 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4506 subclass of rtl_opt_pass along with...
4507 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4508 (make_pass_rtl_move_loop_invariants): ...new function.
4509 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4510 rtl_opt_pass along with...
4511 (pass_data_rtl_unswitch): ...new pass_data instance and...
4512 (make_pass_rtl_unswitch): ...new function.
4513 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4514 subclass of rtl_opt_pass along with...
4515 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4517 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4518 (pass_rtl_doloop): Convert from a global struct to a subclass of
4519 rtl_opt_pass along with...
4520 (pass_data_rtl_doloop): ...new pass_data instance and...
4521 (make_pass_rtl_doloop): ...new function.
4522 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4523 a subclass of rtl_opt_pass along with...
4524 (pass_data_lower_subreg): ...new pass_data instance and...
4525 (make_pass_lower_subreg): ...new function.
4526 (pass_lower_subreg2): Convert from a global struct to a subclass of
4527 rtl_opt_pass along with...
4528 (pass_data_lower_subreg2): ...new pass_data instance and...
4529 (make_pass_lower_subreg2): ...new function.
4530 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4531 struct to a subclass of ipa_opt_pass_d along with...
4532 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4533 (make_pass_ipa_lto_gimple_out): ...new function.
4534 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4535 of ipa_opt_pass_d along with...
4536 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4537 (make_pass_ipa_lto_finish_out): ...new function.
4538 * mode-switching.c (pass_mode_switching): Convert from a global struct
4539 to a subclass of rtl_opt_pass along with...
4540 (pass_data_mode_switching): ...new pass_data instance and...
4541 (make_pass_mode_switching): ...new function.
4542 * modulo-sched.c (pass_sms): Convert from a global struct to a
4543 subclass of rtl_opt_pass along with...
4544 (pass_data_sms): ...new pass_data instance and...
4545 (make_pass_sms): ...new function.
4546 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4547 subclass of gimple_opt_pass along with...
4548 (pass_data_expand_omp): ...new pass_data instance and...
4549 (make_pass_expand_omp): ...new function.
4550 (pass_lower_omp): Convert from a global struct to a subclass of
4551 gimple_opt_pass along with...
4552 (pass_data_lower_omp): ...new pass_data instance and...
4553 (make_pass_lower_omp): ...new function.
4554 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4555 of gimple_opt_pass along with...
4556 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4557 (make_pass_diagnose_omp_blocks): ...new function.
4558 * passes.c (pass_early_local_passes): Convert from a global struct to
4559 a subclass of simple_ipa_opt_pass along with...
4560 (pass_data_early_local_passes): ...new pass_data instance and...
4561 (make_pass_early_local_passes): ...new function.
4562 (pass_all_early_optimizations): Convert from a global struct to a
4563 subclass of gimple_opt_pass along with...
4564 (pass_data_all_early_optimizations): ...new pass_data instance and...
4565 (make_pass_all_early_optimizations): ...new function.
4566 (pass_all_optimizations): Convert from a global struct to a subclass
4567 of gimple_opt_pass along with...
4568 (pass_data_all_optimizations): ...new pass_data instance and...
4569 (make_pass_all_optimizations): ...new function.
4570 (pass_all_optimizations_g): Convert from a global struct to a subclass
4571 of gimple_opt_pass along with...
4572 (pass_data_all_optimizations_g): ...new pass_data instance and...
4573 (make_pass_all_optimizations_g): ...new function.
4574 (pass_rest_of_compilation): Convert from a global struct to a subclass
4575 of rtl_opt_pass along with...
4576 (pass_data_rest_of_compilation): ...new pass_data instance and...
4577 (make_pass_rest_of_compilation): ...new function.
4578 (pass_postreload): Convert from a global struct to a subclass of
4579 rtl_opt_pass along with...
4580 (pass_data_postreload): ...new pass_data instance and...
4581 (make_pass_postreload): ...new function.
4582 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
4583 subclass of rtl_opt_pass along with...
4584 (pass_data_gcse2): ...new pass_data instance and...
4585 (make_pass_gcse2): ...new function.
4586 * postreload.c (pass_postreload_cse): Convert from a global struct to
4587 a subclass of rtl_opt_pass along with...
4588 (pass_data_postreload_cse): ...new pass_data instance and...
4589 (make_pass_postreload_cse): ...new function.
4590 * predict.c (pass_profile): Convert from a global struct to a subclass
4591 of gimple_opt_pass along with...
4592 (pass_data_profile): ...new pass_data instance and...
4593 (make_pass_profile): ...new function.
4594 (pass_strip_predict_hints): Convert from a global struct to a subclass
4595 of gimple_opt_pass along with...
4596 (pass_data_strip_predict_hints): ...new pass_data instance and...
4597 (make_pass_strip_predict_hints): ...new function.
4598 * recog.c (pass_peephole2): Convert from a global struct to a subclass
4599 of rtl_opt_pass along with...
4600 (pass_data_peephole2): ...new pass_data instance and...
4601 (make_pass_peephole2): ...new function.
4602 (pass_split_all_insns): Convert from a global struct to a subclass of
4603 rtl_opt_pass along with...
4604 (pass_data_split_all_insns): ...new pass_data instance and...
4605 (make_pass_split_all_insns): ...new function.
4606 (pass_split_after_reload): Convert from a global struct to a subclass
4607 of rtl_opt_pass along with...
4608 (pass_data_split_after_reload): ...new pass_data instance and...
4609 (make_pass_split_after_reload): ...new function.
4610 (pass_split_before_regstack): Convert from a global struct to a
4611 subclass of rtl_opt_pass along with...
4612 (pass_data_split_before_regstack): ...new pass_data instance and...
4613 (make_pass_split_before_regstack): ...new function.
4614 (pass_split_before_sched2): Convert from a global struct to a subclass
4615 of rtl_opt_pass along with...
4616 (pass_data_split_before_sched2): ...new pass_data instance and...
4617 (make_pass_split_before_sched2): ...new function.
4618 (pass_split_for_shorten_branches): Convert from a global struct to a
4619 subclass of rtl_opt_pass along with...
4620 (pass_data_split_for_shorten_branches): ...new pass_data instance
4622 (make_pass_split_for_shorten_branches): ...new function.
4623 * ree.c (pass_ree): Convert from a global struct to a subclass of
4624 rtl_opt_pass along with...
4625 (pass_data_ree): ...new pass_data instance and...
4626 (make_pass_ree): ...new function.
4627 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
4628 subclass of rtl_opt_pass along with...
4629 (pass_data_stack_regs): ...new pass_data instance and...
4630 (make_pass_stack_regs): ...new function.
4631 (pass_stack_regs_run): Convert from a global struct to a subclass of
4632 rtl_opt_pass along with...
4633 (pass_data_stack_regs_run): ...new pass_data instance and...
4634 (make_pass_stack_regs_run): ...new function.
4635 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4636 subclass of rtl_opt_pass along with...
4637 (pass_data_cprop_hardreg): ...new pass_data instance and...
4638 (make_pass_cprop_hardreg): ...new function.
4639 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4640 subclass of rtl_opt_pass along with...
4641 (pass_data_reginfo_init): ...new pass_data instance and...
4642 (make_pass_reginfo_init): ...new function.
4643 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4644 of rtl_opt_pass along with...
4645 (pass_data_regmove): ...new pass_data instance and...
4646 (make_pass_regmove): ...new function.
4647 * regrename.c (pass_regrename): Convert from a global struct to a
4648 subclass of rtl_opt_pass along with...
4649 (pass_data_regrename): ...new pass_data instance and...
4650 (make_pass_regrename): ...new function.
4651 * reorg.c (pass_delay_slots): Convert from a global struct to a
4652 subclass of rtl_opt_pass along with...
4653 (pass_data_delay_slots): ...new pass_data instance and...
4654 (make_pass_delay_slots): ...new function.
4655 (pass_machine_reorg): Convert from a global struct to a subclass of
4656 rtl_opt_pass along with...
4657 (pass_data_machine_reorg): ...new pass_data instance and...
4658 (make_pass_machine_reorg): ...new function.
4659 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4660 of rtl_opt_pass along with...
4661 (pass_data_sched): ...new pass_data instance and...
4662 (make_pass_sched): ...new function.
4663 (pass_sched2): Convert from a global struct to a subclass of
4664 rtl_opt_pass along with...
4665 (pass_data_sched2): ...new pass_data instance and...
4666 (make_pass_sched2): ...new function.
4667 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4668 to a subclass of rtl_opt_pass along with...
4669 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4670 (make_pass_stack_ptr_mod): ...new function.
4671 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4672 to a subclass of rtl_opt_pass along with...
4673 (pass_data_rtl_store_motion): ...new pass_data instance and...
4674 (make_pass_rtl_store_motion): ...new function.
4675 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4676 of gimple_opt_pass along with...
4677 (pass_data_tracer): ...new pass_data instance and...
4678 (make_pass_tracer): ...new function.
4679 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4680 to a subclass of gimple_opt_pass along with...
4681 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4682 (make_pass_diagnose_tm_blocks): ...new function.
4683 (pass_lower_tm): Convert from a global struct to a subclass of
4684 gimple_opt_pass along with...
4685 (pass_data_lower_tm): ...new pass_data instance and...
4686 (make_pass_lower_tm): ...new function.
4687 (pass_tm_init): Convert from a global struct to a subclass of
4688 gimple_opt_pass along with...
4689 (pass_data_tm_init): ...new pass_data instance and...
4690 (make_pass_tm_init): ...new function.
4691 (pass_tm_mark): Convert from a global struct to a subclass of
4692 gimple_opt_pass along with...
4693 (pass_data_tm_mark): ...new pass_data instance and...
4694 (make_pass_tm_mark): ...new function.
4695 (pass_tm_edges): Convert from a global struct to a subclass of
4696 gimple_opt_pass along with...
4697 (pass_data_tm_edges): ...new pass_data instance and...
4698 (make_pass_tm_edges): ...new function.
4699 (pass_tm_memopt): Convert from a global struct to a subclass of
4700 gimple_opt_pass along with...
4701 (pass_data_tm_memopt): ...new pass_data instance and...
4702 (make_pass_tm_memopt): ...new function.
4703 (pass_ipa_tm): Convert from a global struct to a subclass of
4704 simple_ipa_opt_pass along with...
4705 (pass_data_ipa_tm): ...new pass_data instance and...
4706 (make_pass_ipa_tm): ...new function.
4707 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4708 subclass of gimple_opt_pass along with...
4709 (pass_data_call_cdce): ...new pass_data instance and...
4710 (make_pass_call_cdce): ...new function.
4711 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4712 subclass of gimple_opt_pass along with...
4713 (pass_data_build_cfg): ...new pass_data instance and...
4714 (make_pass_build_cfg): ...new function.
4715 (pass_split_crit_edges): Convert from a global struct to a subclass of
4716 gimple_opt_pass along with...
4717 (pass_data_split_crit_edges): ...new pass_data instance and...
4718 (make_pass_split_crit_edges): ...new function.
4719 (pass_warn_function_return): Convert from a global struct to a
4720 subclass of gimple_opt_pass along with...
4721 (pass_data_warn_function_return): ...new pass_data instance and...
4722 (make_pass_warn_function_return): ...new function.
4723 (pass_warn_function_noreturn): Convert from a global struct to a
4724 subclass of gimple_opt_pass along with...
4725 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4726 (make_pass_warn_function_noreturn): ...new function.
4727 (pass_warn_unused_result): Convert from a global struct to a subclass
4728 of gimple_opt_pass along with...
4729 (pass_data_warn_unused_result): ...new pass_data instance and...
4730 (make_pass_warn_unused_result): ...new function.
4731 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4732 a subclass of gimple_opt_pass along with...
4733 (pass_data_merge_phi): ...new pass_data instance and...
4734 (make_pass_merge_phi): ...new function.
4735 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4736 a subclass of gimple_opt_pass along with...
4737 (pass_data_lower_complex): ...new pass_data instance and...
4738 (make_pass_lower_complex): ...new function.
4739 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4740 gimple_opt_pass along with...
4741 (pass_data_lower_complex_O0): ...new pass_data instance and...
4742 (make_pass_lower_complex_O0): ...new function.
4743 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4744 subclass of gimple_opt_pass along with...
4745 (pass_data_lower_eh): ...new pass_data instance and...
4746 (make_pass_lower_eh): ...new function.
4747 (pass_refactor_eh): Convert from a global struct to a subclass of
4748 gimple_opt_pass along with...
4749 (pass_data_refactor_eh): ...new pass_data instance and...
4750 (make_pass_refactor_eh): ...new function.
4751 (pass_lower_resx): Convert from a global struct to a subclass of
4752 gimple_opt_pass along with...
4753 (pass_data_lower_resx): ...new pass_data instance and...
4754 (make_pass_lower_resx): ...new function.
4755 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4756 of gimple_opt_pass along with...
4757 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4758 (make_pass_lower_eh_dispatch): ...new function.
4759 (pass_cleanup_eh): Convert from a global struct to a subclass of
4760 gimple_opt_pass along with...
4761 (pass_data_cleanup_eh): ...new pass_data instance and...
4762 (make_pass_cleanup_eh): ...new function.
4763 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4764 to a subclass of simple_ipa_opt_pass along with...
4765 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4766 (make_pass_ipa_lower_emutls): ...new function.
4767 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4768 a subclass of gimple_opt_pass along with...
4769 (pass_data_if_conversion): ...new pass_data instance and...
4770 (make_pass_if_conversion): ...new function.
4771 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4772 subclass of gimple_opt_pass along with...
4773 (pass_data_build_ssa): ...new pass_data instance and...
4774 (make_pass_build_ssa): ...new function.
4775 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4776 global struct to a subclass of gimple_opt_pass along with...
4777 (pass_data_loop_distribution): ...new pass_data instance and...
4778 (make_pass_loop_distribution): ...new function.
4779 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4780 subclass of gimple_opt_pass along with...
4781 (pass_data_mudflap_1): ...new pass_data instance and...
4782 (make_pass_mudflap_1): ...new function.
4783 (pass_mudflap_2): Convert from a global struct to a subclass of
4784 gimple_opt_pass along with...
4785 (pass_data_mudflap_2): ...new pass_data instance and...
4786 (make_pass_mudflap_2): ...new function.
4787 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4788 subclass of gimple_opt_pass along with...
4789 (pass_data_mudflap_1): ...new pass_data instance and...
4790 (make_pass_mudflap_1): ...new function.
4791 (pass_mudflap_2): Convert from a global struct to a subclass of
4792 gimple_opt_pass along with...
4793 (pass_data_mudflap_2): ...new pass_data instance and...
4794 (make_pass_mudflap_2): ...new function.
4795 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4796 gimple_opt_pass along with...
4797 (pass_data_nrv): ...new pass_data instance and...
4798 (make_pass_nrv): ...new function.
4799 (pass_return_slot): Convert from a global struct to a subclass of
4800 gimple_opt_pass along with...
4801 (pass_data_return_slot): ...new pass_data instance and...
4802 (make_pass_return_slot): ...new function.
4803 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4804 to a subclass of gimple_opt_pass along with...
4805 (pass_data_object_sizes): ...new pass_data instance and...
4806 (make_pass_object_sizes): ...new function.
4807 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4808 global struct to a subclass of gimple_opt_pass along with...
4809 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4811 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4812 (pass_fixup_cfg): Convert from a global struct to a subclass of
4813 gimple_opt_pass along with...
4814 (pass_data_fixup_cfg): ...new pass_data instance and...
4815 (make_pass_fixup_cfg): ...new function.
4816 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4817 (make_pass_mudflap_1): ...new function.
4818 (pass_mudflap_2): Replace declaration with that of...
4819 (make_pass_mudflap_2): ...new function.
4820 (pass_asan): Replace declaration with that of...
4821 (make_pass_asan): ...new function.
4822 (pass_asan_O0): Replace declaration with that of...
4823 (make_pass_asan_O0): ...new function.
4824 (pass_tsan): Replace declaration with that of...
4825 (make_pass_tsan): ...new function.
4826 (pass_tsan_O0): Replace declaration with that of...
4827 (make_pass_tsan_O0): ...new function.
4828 (pass_lower_cf): Replace declaration with that of...
4829 (make_pass_lower_cf): ...new function.
4830 (pass_refactor_eh): Replace declaration with that of...
4831 (make_pass_refactor_eh): ...new function.
4832 (pass_lower_eh): Replace declaration with that of...
4833 (make_pass_lower_eh): ...new function.
4834 (pass_lower_eh_dispatch): Replace declaration with that of...
4835 (make_pass_lower_eh_dispatch): ...new function.
4836 (pass_lower_resx): Replace declaration with that of...
4837 (make_pass_lower_resx): ...new function.
4838 (pass_build_cfg): Replace declaration with that of...
4839 (make_pass_build_cfg): ...new function.
4840 (pass_early_tree_profile): Replace declaration with that of...
4841 (make_pass_early_tree_profile): ...new function.
4842 (pass_cleanup_eh): Replace declaration with that of...
4843 (make_pass_cleanup_eh): ...new function.
4844 (pass_sra): Replace declaration with that of...
4845 (make_pass_sra): ...new function.
4846 (pass_sra_early): Replace declaration with that of...
4847 (make_pass_sra_early): ...new function.
4848 (pass_early_ipa_sra): Replace declaration with that of...
4849 (make_pass_early_ipa_sra): ...new function.
4850 (pass_tail_recursion): Replace declaration with that of...
4851 (make_pass_tail_recursion): ...new function.
4852 (pass_tail_calls): Replace declaration with that of...
4853 (make_pass_tail_calls): ...new function.
4854 (pass_tree_loop): Replace declaration with that of...
4855 (make_pass_tree_loop): ...new function.
4856 (pass_tree_loop_init): Replace declaration with that of...
4857 (make_pass_tree_loop_init): ...new function.
4858 (pass_lim): Replace declaration with that of...
4859 (make_pass_lim): ...new function.
4860 (pass_tree_unswitch): Replace declaration with that of...
4861 (make_pass_tree_unswitch): ...new function.
4862 (pass_predcom): Replace declaration with that of...
4863 (make_pass_predcom): ...new function.
4864 (pass_iv_canon): Replace declaration with that of...
4865 (make_pass_iv_canon): ...new function.
4866 (pass_scev_cprop): Replace declaration with that of...
4867 (make_pass_scev_cprop): ...new function.
4868 (pass_empty_loop): Replace declaration with that of...
4869 (make_pass_empty_loop): ...new function.
4870 (pass_record_bounds): Replace declaration with that of...
4871 (make_pass_record_bounds): ...new function.
4872 (pass_graphite): Replace declaration with that of...
4873 (make_pass_graphite): ...new function.
4874 (pass_graphite_transforms): Replace declaration with that of...
4875 (make_pass_graphite_transforms): ...new function.
4876 (pass_if_conversion): Replace declaration with that of...
4877 (make_pass_if_conversion): ...new function.
4878 (pass_loop_distribution): Replace declaration with that of...
4879 (make_pass_loop_distribution): ...new function.
4880 (pass_vectorize): Replace declaration with that of...
4881 (make_pass_vectorize): ...new function.
4882 (pass_slp_vectorize): Replace declaration with that of...
4883 (make_pass_slp_vectorize): ...new function.
4884 (pass_complete_unroll): Replace declaration with that of...
4885 (make_pass_complete_unroll): ...new function.
4886 (pass_complete_unrolli): Replace declaration with that of...
4887 (make_pass_complete_unrolli): ...new function.
4888 (pass_parallelize_loops): Replace declaration with that of...
4889 (make_pass_parallelize_loops): ...new function.
4890 (pass_loop_prefetch): Replace declaration with that of...
4891 (make_pass_loop_prefetch): ...new function.
4892 (pass_iv_optimize): Replace declaration with that of...
4893 (make_pass_iv_optimize): ...new function.
4894 (pass_tree_loop_done): Replace declaration with that of...
4895 (make_pass_tree_loop_done): ...new function.
4896 (pass_ch): Replace declaration with that of...
4897 (make_pass_ch): ...new function.
4898 (pass_ccp): Replace declaration with that of...
4899 (make_pass_ccp): ...new function.
4900 (pass_phi_only_cprop): Replace declaration with that of...
4901 (make_pass_phi_only_cprop): ...new function.
4902 (pass_build_ssa): Replace declaration with that of...
4903 (make_pass_build_ssa): ...new function.
4904 (pass_build_alias): Replace declaration with that of...
4905 (make_pass_build_alias): ...new function.
4906 (pass_build_ealias): Replace declaration with that of...
4907 (make_pass_build_ealias): ...new function.
4908 (pass_dominator): Replace declaration with that of...
4909 (make_pass_dominator): ...new function.
4910 (pass_dce): Replace declaration with that of...
4911 (make_pass_dce): ...new function.
4912 (pass_dce_loop): Replace declaration with that of...
4913 (make_pass_dce_loop): ...new function.
4914 (pass_cd_dce): Replace declaration with that of...
4915 (make_pass_cd_dce): ...new function.
4916 (pass_call_cdce): Replace declaration with that of...
4917 (make_pass_call_cdce): ...new function.
4918 (pass_merge_phi): Replace declaration with that of...
4919 (make_pass_merge_phi): ...new function.
4920 (pass_split_crit_edges): Replace declaration with that of...
4921 (make_pass_split_crit_edges): ...new function.
4922 (pass_pre): Replace declaration with that of...
4923 (make_pass_pre): ...new function.
4924 (pass_profile): Replace declaration with that of...
4925 (make_pass_profile): ...new function.
4926 (pass_strip_predict_hints): Replace declaration with that of...
4927 (make_pass_strip_predict_hints): ...new function.
4928 (pass_lower_complex_O0): Replace declaration with that of...
4929 (make_pass_lower_complex_O0): ...new function.
4930 (pass_lower_complex): Replace declaration with that of...
4931 (make_pass_lower_complex): ...new function.
4932 (pass_lower_vector): Replace declaration with that of...
4933 (make_pass_lower_vector): ...new function.
4934 (pass_lower_vector_ssa): Replace declaration with that of...
4935 (make_pass_lower_vector_ssa): ...new function.
4936 (pass_lower_omp): Replace declaration with that of...
4937 (make_pass_lower_omp): ...new function.
4938 (pass_diagnose_omp_blocks): Replace declaration with that of...
4939 (make_pass_diagnose_omp_blocks): ...new function.
4940 (pass_expand_omp): Replace declaration with that of...
4941 (make_pass_expand_omp): ...new function.
4942 (pass_expand_omp_ssa): Replace declaration with that of...
4943 (make_pass_expand_omp_ssa): ...new function.
4944 (pass_object_sizes): Replace declaration with that of...
4945 (make_pass_object_sizes): ...new function.
4946 (pass_strlen): Replace declaration with that of...
4947 (make_pass_strlen): ...new function.
4948 (pass_fold_builtins): Replace declaration with that of...
4949 (make_pass_fold_builtins): ...new function.
4950 (pass_stdarg): Replace declaration with that of...
4951 (make_pass_stdarg): ...new function.
4952 (pass_early_warn_uninitialized): Replace declaration with that of...
4953 (make_pass_early_warn_uninitialized): ...new function.
4954 (pass_late_warn_uninitialized): Replace declaration with that of...
4955 (make_pass_late_warn_uninitialized): ...new function.
4956 (pass_cse_reciprocals): Replace declaration with that of...
4957 (make_pass_cse_reciprocals): ...new function.
4958 (pass_cse_sincos): Replace declaration with that of...
4959 (make_pass_cse_sincos): ...new function.
4960 (pass_optimize_bswap): Replace declaration with that of...
4961 (make_pass_optimize_bswap): ...new function.
4962 (pass_optimize_widening_mul): Replace declaration with that of...
4963 (make_pass_optimize_widening_mul): ...new function.
4964 (pass_warn_function_return): Replace declaration with that of...
4965 (make_pass_warn_function_return): ...new function.
4966 (pass_warn_function_noreturn): Replace declaration with that of...
4967 (make_pass_warn_function_noreturn): ...new function.
4968 (pass_cselim): Replace declaration with that of...
4969 (make_pass_cselim): ...new function.
4970 (pass_phiopt): Replace declaration with that of...
4971 (make_pass_phiopt): ...new function.
4972 (pass_forwprop): Replace declaration with that of...
4973 (make_pass_forwprop): ...new function.
4974 (pass_phiprop): Replace declaration with that of...
4975 (make_pass_phiprop): ...new function.
4976 (pass_tree_ifcombine): Replace declaration with that of...
4977 (make_pass_tree_ifcombine): ...new function.
4978 (pass_dse): Replace declaration with that of...
4979 (make_pass_dse): ...new function.
4980 (pass_nrv): Replace declaration with that of...
4981 (make_pass_nrv): ...new function.
4982 (pass_rename_ssa_copies): Replace declaration with that of...
4983 (make_pass_rename_ssa_copies): ...new function.
4984 (pass_sink_code): Replace declaration with that of...
4985 (make_pass_sink_code): ...new function.
4986 (pass_fre): Replace declaration with that of...
4987 (make_pass_fre): ...new function.
4988 (pass_check_data_deps): Replace declaration with that of...
4989 (make_pass_check_data_deps): ...new function.
4990 (pass_copy_prop): Replace declaration with that of...
4991 (make_pass_copy_prop): ...new function.
4992 (pass_vrp): Replace declaration with that of...
4993 (make_pass_vrp): ...new function.
4994 (pass_uncprop): Replace declaration with that of...
4995 (make_pass_uncprop): ...new function.
4996 (pass_return_slot): Replace declaration with that of...
4997 (make_pass_return_slot): ...new function.
4998 (pass_reassoc): Replace declaration with that of...
4999 (make_pass_reassoc): ...new function.
5000 (pass_rebuild_cgraph_edges): Replace declaration with that of...
5001 (make_pass_rebuild_cgraph_edges): ...new function.
5002 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
5003 (make_pass_remove_cgraph_callee_edges): ...new function.
5004 (pass_build_cgraph_edges): Replace declaration with that of...
5005 (make_pass_build_cgraph_edges): ...new function.
5006 (pass_local_pure_const): Replace declaration with that of...
5007 (make_pass_local_pure_const): ...new function.
5008 (pass_tracer): Replace declaration with that of...
5009 (make_pass_tracer): ...new function.
5010 (pass_warn_unused_result): Replace declaration with that of...
5011 (make_pass_warn_unused_result): ...new function.
5012 (pass_diagnose_tm_blocks): Replace declaration with that of...
5013 (make_pass_diagnose_tm_blocks): ...new function.
5014 (pass_lower_tm): Replace declaration with that of...
5015 (make_pass_lower_tm): ...new function.
5016 (pass_tm_init): Replace declaration with that of...
5017 (make_pass_tm_init): ...new function.
5018 (pass_tm_mark): Replace declaration with that of...
5019 (make_pass_tm_mark): ...new function.
5020 (pass_tm_memopt): Replace declaration with that of...
5021 (make_pass_tm_memopt): ...new function.
5022 (pass_tm_edges): Replace declaration with that of...
5023 (make_pass_tm_edges): ...new function.
5024 (pass_split_functions): Replace declaration with that of...
5025 (make_pass_split_functions): ...new function.
5026 (pass_feedback_split_functions): Replace declaration with that of...
5027 (make_pass_feedback_split_functions): ...new function.
5028 (pass_strength_reduction): Replace declaration with that of...
5029 (make_pass_strength_reduction): ...new function.
5030 (pass_ipa_lower_emutls): Replace declaration with that of...
5031 (make_pass_ipa_lower_emutls): ...new function.
5032 (pass_ipa_function_and_variable_visibility): Replace declaration with
5034 (make_pass_ipa_function_and_variable_visibility): ...new function.
5035 (pass_ipa_tree_profile): Replace declaration with that of...
5036 (make_pass_ipa_tree_profile): ...new function.
5037 (pass_early_local_passes): Replace declaration with that of...
5038 (make_pass_early_local_passes): ...new function.
5039 (pass_ipa_whole_program_visibility): Replace declaration with that
5041 (make_pass_ipa_whole_program_visibility): ...new function.
5042 (pass_ipa_lto_gimple_out): Replace declaration with that of...
5043 (make_pass_ipa_lto_gimple_out): ...new function.
5044 (pass_ipa_increase_alignment): Replace declaration with that of...
5045 (make_pass_ipa_increase_alignment): ...new function.
5046 (pass_ipa_inline): Replace declaration with that of...
5047 (make_pass_ipa_inline): ...new function.
5048 (pass_ipa_free_lang_data): Replace declaration with that of...
5049 (make_pass_ipa_free_lang_data): ...new function.
5050 (pass_ipa_free_inline_summary): Replace declaration with that of...
5051 (make_pass_ipa_free_inline_summary): ...new function.
5052 (pass_ipa_cp): Replace declaration with that of...
5053 (make_pass_ipa_cp): ...new function.
5054 (pass_ipa_reference): Replace declaration with that of...
5055 (make_pass_ipa_reference): ...new function.
5056 (pass_ipa_pure_const): Replace declaration with that of...
5057 (make_pass_ipa_pure_const): ...new function.
5058 (pass_ipa_pta): Replace declaration with that of...
5059 (make_pass_ipa_pta): ...new function.
5060 (pass_ipa_lto_finish_out): Replace declaration with that of...
5061 (make_pass_ipa_lto_finish_out): ...new function.
5062 (pass_ipa_tm): Replace declaration with that of...
5063 (make_pass_ipa_tm): ...new function.
5064 (pass_ipa_profile): Replace declaration with that of...
5065 (make_pass_ipa_profile): ...new function.
5066 (pass_ipa_cdtor_merge): Replace declaration with that of...
5067 (make_pass_ipa_cdtor_merge): ...new function.
5068 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
5070 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5071 (pass_init_datastructures): Replace declaration with that of...
5072 (make_pass_init_datastructures): ...new function.
5073 (pass_fixup_cfg): Replace declaration with that of...
5074 (make_pass_fixup_cfg): ...new function.
5075 (pass_expand): Replace declaration with that of...
5076 (make_pass_expand): ...new function.
5077 (pass_instantiate_virtual_regs): Replace declaration with that of...
5078 (make_pass_instantiate_virtual_regs): ...new function.
5079 (pass_rtl_fwprop): Replace declaration with that of...
5080 (make_pass_rtl_fwprop): ...new function.
5081 (pass_rtl_fwprop_addr): Replace declaration with that of...
5082 (make_pass_rtl_fwprop_addr): ...new function.
5083 (pass_jump): Replace declaration with that of...
5084 (make_pass_jump): ...new function.
5085 (pass_jump2): Replace declaration with that of...
5086 (make_pass_jump2): ...new function.
5087 (pass_lower_subreg): Replace declaration with that of...
5088 (make_pass_lower_subreg): ...new function.
5089 (pass_cse): Replace declaration with that of...
5090 (make_pass_cse): ...new function.
5091 (pass_fast_rtl_dce): Replace declaration with that of...
5092 (make_pass_fast_rtl_dce): ...new function.
5093 (pass_ud_rtl_dce): Replace declaration with that of...
5094 (make_pass_ud_rtl_dce): ...new function.
5095 (pass_rtl_dce): Replace declaration with that of...
5096 (make_pass_rtl_dce): ...new function.
5097 (pass_rtl_dse1): Replace declaration with that of...
5098 (make_pass_rtl_dse1): ...new function.
5099 (pass_rtl_dse2): Replace declaration with that of...
5100 (make_pass_rtl_dse2): ...new function.
5101 (pass_rtl_dse3): Replace declaration with that of...
5102 (make_pass_rtl_dse3): ...new function.
5103 (pass_rtl_cprop): Replace declaration with that of...
5104 (make_pass_rtl_cprop): ...new function.
5105 (pass_rtl_pre): Replace declaration with that of...
5106 (make_pass_rtl_pre): ...new function.
5107 (pass_rtl_hoist): Replace declaration with that of...
5108 (make_pass_rtl_hoist): ...new function.
5109 (pass_rtl_store_motion): Replace declaration with that of...
5110 (make_pass_rtl_store_motion): ...new function.
5111 (pass_cse_after_global_opts): Replace declaration with that of...
5112 (make_pass_cse_after_global_opts): ...new function.
5113 (pass_rtl_ifcvt): Replace declaration with that of...
5114 (make_pass_rtl_ifcvt): ...new function.
5115 (pass_into_cfg_layout_mode): Replace declaration with that of...
5116 (make_pass_into_cfg_layout_mode): ...new function.
5117 (pass_outof_cfg_layout_mode): Replace declaration with that of...
5118 (make_pass_outof_cfg_layout_mode): ...new function.
5119 (pass_loop2): Replace declaration with that of...
5120 (make_pass_loop2): ...new function.
5121 (pass_rtl_loop_init): Replace declaration with that of...
5122 (make_pass_rtl_loop_init): ...new function.
5123 (pass_rtl_move_loop_invariants): Replace declaration with that of...
5124 (make_pass_rtl_move_loop_invariants): ...new function.
5125 (pass_rtl_unswitch): Replace declaration with that of...
5126 (make_pass_rtl_unswitch): ...new function.
5127 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
5128 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5129 (pass_rtl_doloop): Replace declaration with that of...
5130 (make_pass_rtl_doloop): ...new function.
5131 (pass_rtl_loop_done): Replace declaration with that of...
5132 (make_pass_rtl_loop_done): ...new function.
5133 (pass_web): Replace declaration with that of...
5134 (make_pass_web): ...new function.
5135 (pass_cse2): Replace declaration with that of...
5136 (make_pass_cse2): ...new function.
5137 (pass_df_initialize_opt): Replace declaration with that of...
5138 (make_pass_df_initialize_opt): ...new function.
5139 (pass_df_initialize_no_opt): Replace declaration with that of...
5140 (make_pass_df_initialize_no_opt): ...new function.
5141 (pass_reginfo_init): Replace declaration with that of...
5142 (make_pass_reginfo_init): ...new function.
5143 (pass_inc_dec): Replace declaration with that of...
5144 (make_pass_inc_dec): ...new function.
5145 (pass_stack_ptr_mod): Replace declaration with that of...
5146 (make_pass_stack_ptr_mod): ...new function.
5147 (pass_initialize_regs): Replace declaration with that of...
5148 (make_pass_initialize_regs): ...new function.
5149 (pass_combine): Replace declaration with that of...
5150 (make_pass_combine): ...new function.
5151 (pass_if_after_combine): Replace declaration with that of...
5152 (make_pass_if_after_combine): ...new function.
5153 (pass_ree): Replace declaration with that of...
5154 (make_pass_ree): ...new function.
5155 (pass_partition_blocks): Replace declaration with that of...
5156 (make_pass_partition_blocks): ...new function.
5157 (pass_match_asm_constraints): Replace declaration with that of...
5158 (make_pass_match_asm_constraints): ...new function.
5159 (pass_regmove): Replace declaration with that of...
5160 (make_pass_regmove): ...new function.
5161 (pass_split_all_insns): Replace declaration with that of...
5162 (make_pass_split_all_insns): ...new function.
5163 (pass_fast_rtl_byte_dce): Replace declaration with that of...
5164 (make_pass_fast_rtl_byte_dce): ...new function.
5165 (pass_lower_subreg2): Replace declaration with that of...
5166 (make_pass_lower_subreg2): ...new function.
5167 (pass_mode_switching): Replace declaration with that of...
5168 (make_pass_mode_switching): ...new function.
5169 (pass_sms): Replace declaration with that of...
5170 (make_pass_sms): ...new function.
5171 (pass_sched): Replace declaration with that of...
5172 (make_pass_sched): ...new function.
5173 (pass_ira): Replace declaration with that of...
5174 (make_pass_ira): ...new function.
5175 (pass_reload): Replace declaration with that of...
5176 (make_pass_reload): ...new function.
5177 (pass_clean_state): Replace declaration with that of...
5178 (make_pass_clean_state): ...new function.
5179 (pass_branch_prob): Replace declaration with that of...
5180 (make_pass_branch_prob): ...new function.
5181 (pass_value_profile_transformations): Replace declaration with that
5183 (make_pass_value_profile_transformations): ...new function.
5184 (pass_postreload_cse): Replace declaration with that of...
5185 (make_pass_postreload_cse): ...new function.
5186 (pass_gcse2): Replace declaration with that of...
5187 (make_pass_gcse2): ...new function.
5188 (pass_split_after_reload): Replace declaration with that of...
5189 (make_pass_split_after_reload): ...new function.
5190 (pass_branch_target_load_optimize1): Replace declaration with that
5192 (make_pass_branch_target_load_optimize1): ...new function.
5193 (pass_thread_prologue_and_epilogue): Replace declaration with that
5195 (make_pass_thread_prologue_and_epilogue): ...new function.
5196 (pass_stack_adjustments): Replace declaration with that of...
5197 (make_pass_stack_adjustments): ...new function.
5198 (pass_peephole2): Replace declaration with that of...
5199 (make_pass_peephole2): ...new function.
5200 (pass_if_after_reload): Replace declaration with that of...
5201 (make_pass_if_after_reload): ...new function.
5202 (pass_regrename): Replace declaration with that of...
5203 (make_pass_regrename): ...new function.
5204 (pass_cprop_hardreg): Replace declaration with that of...
5205 (make_pass_cprop_hardreg): ...new function.
5206 (pass_reorder_blocks): Replace declaration with that of...
5207 (make_pass_reorder_blocks): ...new function.
5208 (pass_branch_target_load_optimize2): Replace declaration with that
5210 (make_pass_branch_target_load_optimize2): ...new function.
5211 (pass_leaf_regs): Replace declaration with that of...
5212 (make_pass_leaf_regs): ...new function.
5213 (pass_split_before_sched2): Replace declaration with that of...
5214 (make_pass_split_before_sched2): ...new function.
5215 (pass_compare_elim_after_reload): Replace declaration with that of...
5216 (make_pass_compare_elim_after_reload): ...new function.
5217 (pass_sched2): Replace declaration with that of...
5218 (make_pass_sched2): ...new function.
5219 (pass_stack_regs): Replace declaration with that of...
5220 (make_pass_stack_regs): ...new function.
5221 (pass_stack_regs_run): Replace declaration with that of...
5222 (make_pass_stack_regs_run): ...new function.
5223 (pass_df_finish): Replace declaration with that of...
5224 (make_pass_df_finish): ...new function.
5225 (pass_compute_alignments): Replace declaration with that of...
5226 (make_pass_compute_alignments): ...new function.
5227 (pass_duplicate_computed_gotos): Replace declaration with that of...
5228 (make_pass_duplicate_computed_gotos): ...new function.
5229 (pass_variable_tracking): Replace declaration with that of...
5230 (make_pass_variable_tracking): ...new function.
5231 (pass_free_cfg): Replace declaration with that of...
5232 (make_pass_free_cfg): ...new function.
5233 (pass_machine_reorg): Replace declaration with that of...
5234 (make_pass_machine_reorg): ...new function.
5235 (pass_cleanup_barriers): Replace declaration with that of...
5236 (make_pass_cleanup_barriers): ...new function.
5237 (pass_delay_slots): Replace declaration with that of...
5238 (make_pass_delay_slots): ...new function.
5239 (pass_split_for_shorten_branches): Replace declaration with that of...
5240 (make_pass_split_for_shorten_branches): ...new function.
5241 (pass_split_before_regstack): Replace declaration with that of...
5242 (make_pass_split_before_regstack): ...new function.
5243 (pass_convert_to_eh_region_ranges): Replace declaration with that
5245 (make_pass_convert_to_eh_region_ranges): ...new function.
5246 (pass_shorten_branches): Replace declaration with that of...
5247 (make_pass_shorten_branches): ...new function.
5248 (pass_set_nothrow_function_flags): Replace declaration with that of...
5249 (make_pass_set_nothrow_function_flags): ...new function.
5250 (pass_dwarf2_frame): Replace declaration with that of...
5251 (make_pass_dwarf2_frame): ...new function.
5252 (pass_final): Replace declaration with that of...
5253 (make_pass_final): ...new function.
5254 (pass_rtl_seqabstr): Replace declaration with that of...
5255 (make_pass_rtl_seqabstr): ...new function.
5256 (pass_release_ssa_names): Replace declaration with that of...
5257 (make_pass_release_ssa_names): ...new function.
5258 (pass_early_inline): Replace declaration with that of...
5259 (make_pass_early_inline): ...new function.
5260 (pass_inline_parameters): Replace declaration with that of...
5261 (make_pass_inline_parameters): ...new function.
5262 (pass_update_address_taken): Replace declaration with that of...
5263 (make_pass_update_address_taken): ...new function.
5264 (pass_convert_switch): Replace declaration with that of...
5265 (make_pass_convert_switch): ...new function.
5266 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
5267 to a subclass of simple_ipa_opt_pass along with...
5268 (pass_data_ipa_tree_profile): ...new pass_data instance and...
5269 (make_pass_ipa_tree_profile): ...new function.
5270 * tree-sra.c (pass_sra_early): Convert from a global struct to a
5271 subclass of gimple_opt_pass along with...
5272 (pass_data_sra_early): ...new pass_data instance and...
5273 (make_pass_sra_early): ...new function.
5274 (pass_sra): Convert from a global struct to a subclass of
5275 gimple_opt_pass along with...
5276 (pass_data_sra): ...new pass_data instance and...
5277 (make_pass_sra): ...new function.
5278 (pass_early_ipa_sra): Convert from a global struct to a subclass of
5279 gimple_opt_pass along with...
5280 (pass_data_early_ipa_sra): ...new pass_data instance and...
5281 (make_pass_early_ipa_sra): ...new function.
5282 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
5283 subclass of gimple_opt_pass along with...
5284 (pass_data_ccp): ...new pass_data instance and...
5285 (make_pass_ccp): ...new function.
5286 (pass_fold_builtins): Convert from a global struct to a subclass of
5287 gimple_opt_pass along with...
5288 (pass_data_fold_builtins): ...new pass_data instance and...
5289 (make_pass_fold_builtins): ...new function.
5290 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
5291 subclass of gimple_opt_pass along with...
5292 (pass_data_copy_prop): ...new pass_data instance and...
5293 (make_pass_copy_prop): ...new function.
5294 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
5295 global struct to a subclass of gimple_opt_pass along with...
5296 (pass_data_rename_ssa_copies): ...new pass_data instance and...
5297 (make_pass_rename_ssa_copies): ...new function.
5298 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
5299 subclass of gimple_opt_pass along with...
5300 (pass_data_dce): ...new pass_data instance and...
5301 (make_pass_dce): ...new function.
5302 (pass_dce_loop): Convert from a global struct to a subclass of
5303 gimple_opt_pass along with...
5304 (pass_data_dce_loop): ...new pass_data instance and...
5305 (make_pass_dce_loop): ...new function.
5306 (pass_cd_dce): Convert from a global struct to a subclass of
5307 gimple_opt_pass along with...
5308 (pass_data_cd_dce): ...new pass_data instance and...
5309 (make_pass_cd_dce): ...new function.
5310 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
5311 subclass of gimple_opt_pass along with...
5312 (pass_data_dominator): ...new pass_data instance and...
5313 (make_pass_dominator): ...new function.
5314 (pass_phi_only_cprop): Convert from a global struct to a subclass of
5315 gimple_opt_pass along with...
5316 (pass_data_phi_only_cprop): ...new pass_data instance and...
5317 (make_pass_phi_only_cprop): ...new function.
5318 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
5319 subclass of gimple_opt_pass along with...
5320 (pass_data_dse): ...new pass_data instance and...
5321 (make_pass_dse): ...new function.
5322 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
5323 a subclass of gimple_opt_pass along with...
5324 (pass_data_forwprop): ...new pass_data instance and...
5325 (make_pass_forwprop): ...new function.
5326 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
5327 struct to a subclass of gimple_opt_pass along with...
5328 (pass_data_tree_ifcombine): ...new pass_data instance and...
5329 (make_pass_tree_ifcombine): ...new function.
5330 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
5331 subclass of gimple_opt_pass along with...
5332 (pass_data_ch): ...new pass_data instance and...
5333 (make_pass_ch): ...new function.
5334 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
5335 subclass of gimple_opt_pass along with...
5336 (pass_data_tree_loop): ...new pass_data instance and...
5337 (make_pass_tree_loop): ...new function.
5338 (pass_tree_loop_init): Convert from a global struct to a subclass of
5339 gimple_opt_pass along with...
5340 (pass_data_tree_loop_init): ...new pass_data instance and...
5341 (make_pass_tree_loop_init): ...new function.
5342 (pass_lim): Convert from a global struct to a subclass of
5343 gimple_opt_pass along with...
5344 (pass_data_lim): ...new pass_data instance and...
5345 (make_pass_lim): ...new function.
5346 (pass_tree_unswitch): Convert from a global struct to a subclass of
5347 gimple_opt_pass along with...
5348 (pass_data_tree_unswitch): ...new pass_data instance and...
5349 (make_pass_tree_unswitch): ...new function.
5350 (pass_predcom): Convert from a global struct to a subclass of
5351 gimple_opt_pass along with...
5352 (pass_data_predcom): ...new pass_data instance and...
5353 (make_pass_predcom): ...new function.
5354 (pass_vectorize): Convert from a global struct to a subclass of
5355 gimple_opt_pass along with...
5356 (pass_data_vectorize): ...new pass_data instance and...
5357 (make_pass_vectorize): ...new function.
5358 (pass_graphite): Convert from a global struct to a subclass of
5359 gimple_opt_pass along with...
5360 (pass_data_graphite): ...new pass_data instance and...
5361 (make_pass_graphite): ...new function.
5362 (pass_graphite_transforms): Convert from a global struct to a subclass
5363 of gimple_opt_pass along with...
5364 (pass_data_graphite_transforms): ...new pass_data instance and...
5365 (make_pass_graphite_transforms): ...new function.
5366 (pass_check_data_deps): Convert from a global struct to a subclass of
5367 gimple_opt_pass along with...
5368 (pass_data_check_data_deps): ...new pass_data instance and...
5369 (make_pass_check_data_deps): ...new function.
5370 (pass_iv_canon): Convert from a global struct to a subclass of
5371 gimple_opt_pass along with...
5372 (pass_data_iv_canon): ...new pass_data instance and...
5373 (make_pass_iv_canon): ...new function.
5374 (pass_scev_cprop): Convert from a global struct to a subclass of
5375 gimple_opt_pass along with...
5376 (pass_data_scev_cprop): ...new pass_data instance and...
5377 (make_pass_scev_cprop): ...new function.
5378 (pass_record_bounds): Convert from a global struct to a subclass of
5379 gimple_opt_pass along with...
5380 (pass_data_record_bounds): ...new pass_data instance and...
5381 (make_pass_record_bounds): ...new function.
5382 (pass_complete_unroll): Convert from a global struct to a subclass of
5383 gimple_opt_pass along with...
5384 (pass_data_complete_unroll): ...new pass_data instance and...
5385 (make_pass_complete_unroll): ...new function.
5386 (pass_complete_unrolli): Convert from a global struct to a subclass of
5387 gimple_opt_pass along with...
5388 (pass_data_complete_unrolli): ...new pass_data instance and...
5389 (make_pass_complete_unrolli): ...new function.
5390 (pass_parallelize_loops): Convert from a global struct to a subclass
5391 of gimple_opt_pass along with...
5392 (pass_data_parallelize_loops): ...new pass_data instance and...
5393 (make_pass_parallelize_loops): ...new function.
5394 (pass_loop_prefetch): Convert from a global struct to a subclass of
5395 gimple_opt_pass along with...
5396 (pass_data_loop_prefetch): ...new pass_data instance and...
5397 (make_pass_loop_prefetch): ...new function.
5398 (pass_iv_optimize): Convert from a global struct to a subclass of
5399 gimple_opt_pass along with...
5400 (pass_data_iv_optimize): ...new pass_data instance and...
5401 (make_pass_iv_optimize): ...new function.
5402 (pass_tree_loop_done): Convert from a global struct to a subclass of
5403 gimple_opt_pass along with...
5404 (pass_data_tree_loop_done): ...new pass_data instance and...
5405 (make_pass_tree_loop_done): ...new function.
5406 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
5407 struct to a subclass of gimple_opt_pass along with...
5408 (pass_data_cse_reciprocals): ...new pass_data instance and...
5409 (make_pass_cse_reciprocals): ...new function.
5410 (pass_cse_sincos): Convert from a global struct to a subclass of
5411 gimple_opt_pass along with...
5412 (pass_data_cse_sincos): ...new pass_data instance and...
5413 (make_pass_cse_sincos): ...new function.
5414 (pass_optimize_bswap): Convert from a global struct to a subclass of
5415 gimple_opt_pass along with...
5416 (pass_data_optimize_bswap): ...new pass_data instance and...
5417 (make_pass_optimize_bswap): ...new function.
5418 (pass_optimize_widening_mul): Convert from a global struct to a
5419 subclass of gimple_opt_pass along with...
5420 (pass_data_optimize_widening_mul): ...new pass_data instance and...
5421 (make_pass_optimize_widening_mul): ...new function.
5422 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
5423 subclass of gimple_opt_pass along with...
5424 (pass_data_phiopt): ...new pass_data instance and...
5425 (make_pass_phiopt): ...new function.
5426 (pass_cselim): Convert from a global struct to a subclass of
5427 gimple_opt_pass along with...
5428 (pass_data_cselim): ...new pass_data instance and...
5429 (make_pass_cselim): ...new function.
5430 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
5431 subclass of gimple_opt_pass along with...
5432 (pass_data_phiprop): ...new pass_data instance and...
5433 (make_pass_phiprop): ...new function.
5434 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
5435 subclass of gimple_opt_pass along with...
5436 (pass_data_pre): ...new pass_data instance and...
5437 (make_pass_pre): ...new function.
5438 (pass_fre): Convert from a global struct to a subclass of
5439 gimple_opt_pass along with...
5440 (pass_data_fre): ...new pass_data instance and...
5441 (make_pass_fre): ...new function.
5442 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
5443 subclass of gimple_opt_pass along with...
5444 (pass_data_reassoc): ...new pass_data instance and...
5445 (make_pass_reassoc): ...new function.
5446 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
5447 subclass of gimple_opt_pass along with...
5448 (pass_data_sink_code): ...new pass_data instance and...
5449 (make_pass_sink_code): ...new function.
5450 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
5451 subclass of gimple_opt_pass along with...
5452 (pass_data_strlen): ...new pass_data instance and...
5453 (make_pass_strlen): ...new function.
5454 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
5455 struct to a subclass of gimple_opt_pass along with...
5456 (pass_data_build_alias): ...new pass_data instance and...
5457 (make_pass_build_alias): ...new function.
5458 (pass_build_ealias): Convert from a global struct to a subclass of
5459 gimple_opt_pass along with...
5460 (pass_data_build_ealias): ...new pass_data instance and...
5461 (make_pass_build_ealias): ...new function.
5462 (pass_ipa_pta): Convert from a global struct to a subclass of
5463 simple_ipa_opt_pass along with...
5464 (pass_data_ipa_pta): ...new pass_data instance and...
5465 (make_pass_ipa_pta): ...new function.
5466 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
5467 subclass of gimple_opt_pass along with...
5468 (pass_data_uncprop): ...new pass_data instance and...
5469 (make_pass_uncprop): ...new function.
5470 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
5471 global struct to a subclass of gimple_opt_pass along with...
5472 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
5473 (make_pass_late_warn_uninitialized): ...new function.
5474 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
5475 to a subclass of gimple_opt_pass along with...
5476 (pass_data_init_datastructures): ...new pass_data instance and...
5477 (make_pass_init_datastructures): ...new function.
5478 (pass_early_warn_uninitialized): Convert from a global struct to a
5479 subclass of gimple_opt_pass along with...
5480 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5481 (make_pass_early_warn_uninitialized): ...new function.
5482 (pass_update_address_taken): Convert from a global struct to a
5483 subclass of gimple_opt_pass along with...
5484 (pass_data_update_address_taken): ...new pass_data instance and...
5485 (make_pass_update_address_taken): ...new function.
5486 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5487 struct to a subclass of gimple_opt_pass along with...
5488 (pass_data_release_ssa_names): ...new pass_data instance and...
5489 (make_pass_release_ssa_names): ...new function.
5490 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5491 subclass of gimple_opt_pass along with...
5492 (pass_data_stdarg): ...new pass_data instance and...
5493 (make_pass_stdarg): ...new function.
5494 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5495 global struct to a subclass of gimple_opt_pass along with...
5496 (pass_data_convert_switch): ...new pass_data instance and...
5497 (make_pass_convert_switch): ...new function.
5498 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5499 to a subclass of gimple_opt_pass along with...
5500 (pass_data_tail_recursion): ...new pass_data instance and...
5501 (make_pass_tail_recursion): ...new function.
5502 (pass_tail_calls): Convert from a global struct to a subclass of
5503 gimple_opt_pass along with...
5504 (pass_data_tail_calls): ...new pass_data instance and...
5505 (make_pass_tail_calls): ...new function.
5506 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5507 struct to a subclass of gimple_opt_pass along with...
5508 (pass_data_lower_vector): ...new pass_data instance and...
5509 (make_pass_lower_vector): ...new function.
5510 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5511 gimple_opt_pass along with...
5512 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5513 (make_pass_lower_vector_ssa): ...new function.
5514 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5515 to a subclass of gimple_opt_pass along with...
5516 (pass_data_slp_vectorize): ...new pass_data instance and...
5517 (make_pass_slp_vectorize): ...new function.
5518 (pass_ipa_increase_alignment): Convert from a global struct to a
5519 subclass of simple_ipa_opt_pass along with...
5520 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5521 (make_pass_ipa_increase_alignment): ...new function.
5522 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5523 gimple_opt_pass along with...
5524 (pass_data_vrp): ...new pass_data instance and...
5525 (make_pass_vrp): ...new function.
5526 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5527 subclass of simple_ipa_opt_pass along with...
5528 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5529 (make_pass_ipa_free_lang_data): ...new function.
5530 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5531 gimple_opt_pass along with...
5532 (pass_data_tsan): ...new pass_data instance and...
5533 (make_pass_tsan): ...new function.
5534 (pass_tsan_O0): Convert from a global struct to a subclass of
5535 gimple_opt_pass along with...
5536 (pass_data_tsan_O0): ...new pass_data instance and...
5537 (make_pass_tsan_O0): ...new function.
5538 * var-tracking.c (pass_variable_tracking): Convert from a global
5539 struct to a subclass of rtl_opt_pass along with...
5540 (pass_data_variable_tracking): ...new pass_data instance and...
5541 (make_pass_variable_tracking): ...new function.
5542 * web.c (pass_web): Convert from a global struct to a subclass of
5543 rtl_opt_pass along with...
5544 (pass_data_web): ...new pass_data instance and...
5545 (make_pass_web): ...new function.
5546 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5547 declaration with that of...
5548 (make_pass_mode_switch_use): ...new function.
5549 (pass_resolve_sw_modes): Replace declaration with that of...
5550 (make_pass_resolve_sw_modes): ...new function.
5551 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5552 from a global struct to a subclass of rtl_opt_pass along with...
5553 (pass_data_mode_switch_use): ...new pass_data instance and...
5554 (make_pass_mode_switch_use): ...new function.
5555 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5556 from a global struct to a subclass of rtl_opt_pass along with...
5557 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5558 (make_pass_resolve_sw_modes): ...new function.
5559 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5560 struct to a subclass of rtl_opt_pass along with...
5561 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5562 (make_pass_insert_vzeroupper): ...new function.
5563 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5564 global struct to a subclass of rtl_opt_pass along with...
5565 (pass_data_work_around_errata): ...new pass_data instance and...
5566 (make_pass_work_around_errata): ...new function.
5567 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5568 struct to a subclass of rtl_opt_pass along with...
5569 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
5570 (make_pass_mips_machine_reorg2): ...new function.
5572 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5574 * passes.c (pass_manager::operator new): New.
5576 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5578 Handwritten part of conversion of passes to C++ classes.
5580 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
5581 (toplev.o): Add dep on PASS_MANAGER_H.
5582 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
5583 of early local pases to reflect this moving from a global to a
5584 member of gcc::pass_manager.
5585 (cgraph_add_new_function): Likewise.
5586 * lto-cgraph.c (lto_output_node): Update for conversion of
5587 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
5588 * passes.c (opt_pass::clone): New.
5589 (opt_pass::gate): New.
5590 (opt_pass::execute): New.
5591 (opt_pass::opt_pass): New.
5592 (pass_manager::execute_early_local_passes): New.
5593 (pass_manager::execute_pass_mode_switching): new.
5594 (finish_optimization_passes): Convert to...
5595 (pass_manager::finish_optimization_passes): ...this.
5596 (finish_optimization_passes): Update for conversion of passes to
5598 (register_dump_files_1): Use has_gate since we cannot portably
5599 check a vtable entry against NULL.
5600 (dump_one_pass): Likewise.
5601 (ipa_write_summaries_2): Likewise.
5602 (ipa_write_optimization_summaries_1): Likewise.
5603 (ipa_read_summaries_1): Likewise.
5604 (ipa_read_optimization_summaries_1): Likewise.
5605 (execute_ipa_stmt_fixups): Likewise.
5606 (pass_manager::pass_manager): Rewrite pass-creation, invoking
5607 pass-creation functions rather than wiring up globals, and
5608 storing the results in fields of pass_manager generated using
5610 (pass_manager::dump_profile_report): Update for conversion of
5611 passes to C++ classes.
5612 (pass_manager::execute_ipa_summary_passes): Likewise.
5613 (execute_one_ipa_transform_pass): Likewise.
5614 (execute_one_pass): Use has_gate and has_execute since we cannot
5615 portably check a vtable entry against NULL.
5616 * pass_manager.h (pass_manager::finish_optimization_passes): New.
5617 (pass_manager): Use pass-instances.def to add fields for the
5618 various pass instances.
5619 * toplev.c (finalize): Update for move of
5620 finish_optimization_passes to a method of gcc::pass_manager.
5621 * toplev.h (finish_optimization_passes): Move to method of class
5623 * tree-pass.h (struct pass_data): New.
5624 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
5625 (opt_pass::gate): Convert to virtual function.
5626 (opt_pass::~opt_pass): New.
5627 (opt_pass::clone): New.
5628 (opt_pass::execute): Convert to virtual function.
5629 (opt_pass::opt_pass): New.
5630 (opt_pass::ctxt_): new.
5631 (gimple_opt_pass): Convert to subclass of opt_pass.
5632 (gimple_opt_pass::gimple_opt_pass): New.
5633 (rtl_opt_pass): Convert to subclass of opt_pass.
5634 (rtl_opt_pass::rtl_opt_pass): New.
5635 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5636 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5637 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5638 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5639 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5640 invocation of pass_mode_switching to reflect this moving from a
5641 global to a member of gcc::pass_manager.
5642 (ix86_option_override): Rework how pass_insert_vzeroupper is
5643 added to the pass_manager to reflect autogenerated changes.
5644 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
5646 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
5648 PR rtl-optimization/57708
5649 * recog.c (peep2_find_free_register): Validate all regs in a
5652 2013-08-05 Jan Hubicka <jh@suse.cz>
5655 * cgraph.c (verify_cgraph_node): Accept local flags from other
5657 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5658 (function_and_variable_visibility): Likewise.
5659 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5661 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5663 * graph.c (init_graph_slim_pretty_print): Remove.
5664 (print_graph_cfg): Do not call it. Use local pretty printer.
5665 (start_graph_dump): Likewise.
5667 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5669 * gimple-pretty-print.c (buffer): Remove.
5670 (initialized): Likewise.
5671 (maybe_init_pretty_print): Likewise.
5672 (print_gimple_stmt): Do not call it. Use non-static local
5673 pretty_printer variable.
5674 (print_gimple_expr): Likewise.
5675 (print_gimple_seq): Likewise.
5676 (gimple_dump_bb): Likewise.
5678 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5680 * asan.c (asan_pp): Remove.
5681 (asan_pp_initialized): Likewise.
5682 (asan_pp_initialize): Likewise.
5683 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5684 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5685 (asan_add_global): Likewise.
5687 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5689 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5690 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5691 * pretty-print.c (pp_formatted_text_data): Likewise.
5692 (pp_indent): Rename from pp_base_indent.
5693 (pp_format): Rename from pp_base_format.
5694 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5695 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5696 (pp_flush): Rename from pp_base_flush.
5697 (pp_set_line_maximum_length): Rename from
5698 pp_base_set_line_maximum_length.
5699 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5700 (pp_set_prefix): Rename from pp_base_set_prefix.
5701 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5702 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5703 (pp_append_text): Rename from pp_base_append_text.
5704 (pp_formatted_text): Rename from pp_base_formatted_text.
5705 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5706 (pp_remaining_character_count_for_line): Rename from
5707 pp_base_remaining_character_count_for_line.
5708 (pp_newline): Rename from pp_base_newline.
5709 (pp_character): Rename from pp_base_character.
5710 (pp_string): Rename from pp_base_string.
5711 (pp_maybe_space): Rename from pp_base_maybe_space.
5712 * asan.c (asan_pp_string): Adjust.
5713 (asan_emit_stack_protection): Likewise.
5714 (asan_add_global): Likewise.
5715 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5716 * tree-mudflap.c (mf_varname_tree): Likewise.
5717 * tree-pretty-print.c (pp_tree_identifier): Rename from
5718 pp_base_tree_identifier.
5719 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5720 Declare as function.
5722 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5724 * pretty-print.h (pp_bar_bar): New.
5725 (pp_ampersand_ampersand): Likewise.
5726 (pp_less_equal): Likewise.
5727 (pp_greater_equal): Likewise.
5728 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5729 printer functions instead of pp_string or operators and punctuators.
5730 (dump_gimple_call): Likewise.
5731 (dump_gimple_omp_for): Likewise.
5732 (dump_gimple_transaction): Likewise.
5733 (dump_gimple_phi): Likewise.
5734 (pp_gimple_stmt_1): Likewise.
5735 * sched-vis.c (print_insn): Likewise.
5736 * tree-mudflap.c (mf_varname_tree): Likewise.
5737 * tree-pretty-print.c (dump_block_node): Likewise.
5738 (dump_generic_node): Likewise.
5740 2013-08-02 Jan Hubicka <jh@suse.cz>
5742 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5744 * lto-streamer-out.c (tree_is_indexable): Results decls and
5745 parm decls are not indexable.
5746 (DFS_write_tree_body): Do not follow args and results.
5747 (hash_tree): Likewise.
5748 (output_functions): Rearrange so struct function is needed
5749 only when real body is output; be able to also ouptut abstract
5750 functions; output DECL_ARGUMENTS and DECL_RESULT.
5751 (lto_output): When not in WPA, ale store abstract functions.
5752 (write_symbol): Do not care about RESULT_DECL.
5753 (output_symbol_p): Handle correctly sbtract decls.
5754 * lto-streamer-in.c (input_function): Rearrange so struct
5755 function can be NULL at entry; allow streaming of
5756 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5757 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5758 sanity check during LTO.
5759 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5760 RESULT_DECl and DECL_ARGUMENTS.
5761 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5764 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5766 * pretty-print.h (pp_underscore): New.
5768 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5769 printer functions instead of pp_character.
5770 (dump_binary_rhs): Likewise.
5771 (dump_ternary_rhs): Likewise.
5772 (dump_gimple_call_args): Likewise.
5773 (pp_points_to_solution): Likewise.
5774 (dump_gimple_call): Likewise.
5775 (dump_gimple_switch): Likewise.
5776 (dump_gimple_cond): Likewise.
5777 (dump_gimple_bind): Likewise.
5778 (dump_gimple_try): Likewise.
5779 (dump_gimple_omp_for): Likewise.
5780 (dump_gimple_omp_continue): Likewise.
5781 (dump_gimple_omp_single): Likewise.
5782 (dump_gimple_omp_sections): Likewise.
5783 (dump_gimple_omp_block): Likewise.
5784 (dump_gimple_omp_critical): Likewise.
5785 (dump_gimple_transaction): Likewise.
5786 (dump_gimple_asm): Likewise.
5787 (dump_gimple_phi): Likewise.
5788 (dump_gimple_omp_parallel): Likewise.
5789 (dump_gimple_omp_task): Likewise.
5790 (dump_gimple_omp_atomic_load): Likewise.
5791 (dump_gimple_omp_atomic_store): Likewise.
5792 (dump_gimple_mem_ops): Likewise.
5793 (pp_gimple_stmt_1): Likewise.
5794 (pp_cfg_jump): Likewise.
5795 (dump_implicit_edges): Likewise.
5796 (gimple_dump_bb_for_graph): Likewise.
5797 * graph.c (draw_cfg_node): Likewise.
5798 * langhooks.c (lhd_print_error_function): Likewise.
5799 * sched-vis.c (print_exp): Likewise.
5800 (print_value): Likewise.
5801 (print_pattern): Likewise.
5802 (print_insn): Likewise.
5803 (rtl_dump_bb_for_graph): Likewise.
5804 * tree-pretty-print.c (dump_function_declaration): Likewise.
5805 (dump_array_domain): Likewise.
5806 (dump_omp_clause): Likewise.
5807 (dump_location): Likewise.
5808 (dump_generic_node): Likewise.
5809 (print_struct_decl): Likewise.
5810 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5812 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5814 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5815 candidate table when replacing a candidate statement.
5816 (replace_rhs_if_not_dup): Likewise.
5817 (replace_one_candidate): Likewise.
5819 2013-08-02 Jan Hubicka <jh@suse.cz>
5820 Martin Liska <marxin.liska@gmail.com>
5822 * cgraphunit.c (add_new_function): Fix logic when adding from
5824 (assemble_thunk): Rename to ...
5825 (expand_thunk); .. this one; export; get it working with
5826 general functions; make produced gimple valid.
5827 * cgraph.h (expand_thunk): Declare.
5829 2013-08-02 Jan Hubicka <jh@suse.cz>
5831 * ipa-cp.c (gather_context_independent_values): Use
5832 ipa_get_param_move_cost.
5833 (get_replacement_map): Remove PARAM; move parameter folding
5835 (create_specialized_node): Update.
5836 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
5837 assert that we have gimple body; update move_cost.
5838 (count_formal_params): Assert that we have gimple body.
5839 (ipa_dump_param): New function.
5840 (ipa_alloc_node_params): Break out from ...
5841 (ipa_initialize_node_params): ... here.
5842 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
5843 (ipa_write_node_info): Stream move costs.
5844 (ipa_read_node_info): Read move costs.
5845 (ipa_update_after_lto_read): Do not recompute node params.
5846 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
5847 (ipa_get_param): Check we are not in WPA.
5848 (ipa_get_param_move_cost): New.
5849 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
5850 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
5851 parm numbers to be present.
5853 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5855 PR rtl-optimization/58048
5856 * lra-constraints.c (process_alt_operands): Don't check asm
5857 operand on register.
5859 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
5861 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
5862 the implied StoreLoad barrier for atomic operations if before.
5864 2013-08-02 Jan Hubicka <jh@suse.cz>
5865 Martin Liska <marxin.liska@gmail.com>
5867 * cgraph.c (cgraph_function_body_availability): Do not check
5869 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
5870 symtab_node_availability): Declare.
5871 * ipa.c (can_replace_by_local_alias): New.
5872 (function_and_variable_visibility): Use it.
5873 * symtab.c (symtab_for_node_and_aliases,
5874 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
5876 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5878 PR rtl-optimization/57963
5879 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
5880 (lra_constraints): Use them.
5882 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
5884 * config/arm/types.md (define_attr "type"): Add "load_acq"
5886 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
5888 (cortex_a53_store1): Likewise.
5890 2013-08-01 Jan Hubicka <jh@suse.cz>
5892 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
5893 partitions are not needed.
5895 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
5897 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
5898 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
5899 MAYBE_NON_Q_CLASS_P where appropriate.
5901 2013-08-01 Jan Hubicka <jh@suse.cz>
5903 * cgraph.h (release_function_body): Declare.
5904 * tree.c (free_lang_data_in_decl): Free, parameters and return values
5905 of unused delcarations.
5907 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5909 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
5910 RTL form when subtracting a constant.
5912 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5914 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
5915 Generate canonical plus rtx with negated immediate instead of minus
5917 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
5919 2013-08-01 Jan Hubicka <jh@suse.cz>
5921 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
5922 (cgraph_release_function_body): Likewise.
5923 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5924 * cgraph.h (cgrpah_node): Rename abstract_and_needed
5925 to used_as_abstract_origin.
5926 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
5927 symbols used as abstract origins.
5928 * cgraphunit.c (analyze_functions): Update.
5929 * ipa.c (symtab_remove_unreachable_nodes): Recompute
5930 used_as_abstract_origin.
5931 * tree-inline.c (tree_function_versioning): Update
5932 used_as_abstract_origin; be ready for DECL_RESULT and
5933 DECL_ARGUMENTS to be NULL.
5935 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
5936 for abstract functions.
5937 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
5940 2013-08-01 Jan Hubicka <jh@suse.cz>
5942 * profile.c (compute_value_histograms): Fix thinko.
5944 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5946 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
5947 aarch-common-protos.h to extra_headers.
5948 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5949 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
5950 * config/aarch64/t-aarch64 (aarch-common.o): Define.
5952 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5954 * config/aarch64/aarch64.md (define_attr "type"): Delete.
5955 Include "../arm/types.md". Define "type" attribute for all patterns.
5956 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
5959 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
5961 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
5962 to support power8 load fusion.
5963 (fusion_gpr_mem_load): Likewise.
5965 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
5967 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
5968 declarations for power8 load fusion.
5969 (emit_fusion_gpr_load): Likewise.
5971 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
5972 tuning for power8, turn on fusion mode by default. Turn on sign
5973 extending fusion mode if normal fusion mode is on, and we are at
5975 (fusion_gpr_load_p): New function, return true if we can fuse an
5976 addis instruction with a dependent load to a GPR.
5977 (emit_fusion_gpr_load): Emit the instructions for power8 load
5980 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
5981 (VSX load fusion peepholes): New peepholes to fuse together an
5982 addi instruction with a VSX load instruction.
5984 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
5985 peepholes to fuse an addis instruction with a load to a GPR base
5986 register. If we are supporting sign extending fusions, convert
5987 sign extending loads to zero extending loads and add an explicit
5990 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5992 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
5993 aarch-common-protos.h to extra_headers.
5994 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5995 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
5996 (arm_early_store_addr_dep): Likewise.
5997 (arm_no_early_alu_shift_dep): Likewise.
5998 (arm_no_early_alu_shift_value_dep): Likewise.
5999 (arm_no_early_mul_dep): Likewise.
6000 (arm_no_early_store_addr_dep): Likewise.
6001 (arm_mac_accumulator_is_mul_result): Likewise.
6002 (arm_mac_accumulator_is_result): Likewise.
6003 * config/arm/aarch-common.c: ... here. New file.
6004 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
6006 (arm_early_store_addr_dep): Likewise.
6007 (arm_no_early_alu_shift_dep): Likewise.
6008 (arm_no_early_alu_shift_value_dep): Likewise.
6009 (arm_no_early_mul_dep): Likewise.
6010 (arm_no_early_store_addr_dep): Likewise.
6011 (arm_mac_accumulator_is_mul_result): Likewise.
6012 (arm_mac_accumulator_is_result): Likewise.
6013 * config/arm/aarch-common-protos.h: ... here. New file.
6014 * config/arm/t-arm (aarch-common.o): Define.
6016 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6018 * config/arm/arm.md: Include new file "types.md".
6019 (define_attr "type"): Move from here to ...
6020 (define_attr "mul32"): Likewise.
6021 (define_attr "mul64"): Likewise.
6022 * config/arm/types.md: ... here. New file.
6024 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
6026 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
6027 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
6029 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6031 * gen-pass-instances.awk: Fix offset of substr().
6033 2013-07-31 David Malcolm <dmalcolm@redhat.com>
6035 * Makefile.in (pass-instances.def): New.
6036 (passes.o): Replace dependency on passes.def with one on
6039 * gen-pass-instances.awk: New.
6041 * passes.c (pass_manager::pass_manager): Use pass-instances.def
6042 rather than passes.def, updating local definition of NEXT_PASS
6043 macro to add an extra NUM parameter (currently unused).
6045 2013-07-30 David Malcolm <dmalcolm@redhat.com>
6047 * Makefile.in (PASS_MANAGER_H): New.
6048 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
6049 (passes.o): Likewise.
6050 (statistics.o): Likewise.
6051 (cgraphunit.o): Likewise.
6052 (context.o): Depend on PASS_MANAGER_H.
6054 * pass_manager.h: New.
6056 * cgraphunit.c (cgraph_add_new_function): Update for moves
6057 of globals to fields of pass_manager.
6058 (analyze_function): Likewise.
6059 (expand_function): Likewise.
6060 (ipa_passes): Likewise.
6061 (compile): Likewise.
6063 * context.c (context::context): New.
6064 * context.h (context::context): New.
6065 (context::get_passes): New.
6066 (context::passes_): New.
6068 * lto-cgraph.c (input_node): Update for moves of globals to
6069 fields of pass_manager.
6071 * passes.c (all_passes): Remove, in favor of a field of the
6072 same name within the new class pass_manager.
6073 (all_small_ipa_passes): Likewise.
6074 (all_lowering_passes): Likewise.
6075 (all_regular_ipa_passes): Likewise.
6076 (all_late_ipa_passes): Likewise.
6077 (all_lto_gen_passes): Likewise.
6078 (passes_by_id): Likewise.
6079 (passes_by_id_size): Likewise.
6080 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6081 the new class pass_manager.
6082 (set_pass_for_id): Convert to...
6083 (pass_manager::set_pass_for_id): ...method.
6084 (get_pass_for_id): Convert to...
6085 (pass_manager::get_pass_for_id): ...method.
6086 (register_one_dump_file): Move body of implementation into...
6087 (pass_manager::register_one_dump_file): ...here.
6088 (register_dump_files_1): Convert to...
6089 (pass_manager::register_dump_files_1): ...method.
6090 (register_dump_files): Convert to...
6091 (pass_manager::register_dump_files): ...method.
6092 (create_pass_tab): Update for moves of globals to fields of
6094 (dump_passes): Move body of implementation into...
6095 (pass_manager::dump_passes): ...here.
6096 (register_pass): Move body of implementation into...
6097 (pass_manager::register_pass): ...here.
6098 (init_optimization_passes): Convert into...
6099 (pass_manager::pass_manager): ...constructor for new
6100 pass_manager class, and initialize the pass_lists array.
6101 (check_profile_consistency): Update for moves of globals to
6102 fields of pass_manager.
6103 (dump_profile_report): Move body of implementation into...
6104 (pass_manager::dump_profile_report): ...here.
6105 (ipa_write_summaries_1): Update for moves of pass lists from
6106 being globals to fields of pass_manager.
6107 (ipa_write_optimization_summaries): Likewise.
6108 (ipa_read_summaries): Likewise.
6109 (ipa_read_optimization_summaries): Likewise.
6110 (execute_all_ipa_stmt_fixups): Likewise.
6112 * statistics.c (statistics_fini): Update for moves of globals to
6113 fields of pass_manager.
6115 * toplev.c (general_init): Replace call to
6116 init_optimization_passes with construction of the pass_manager
6119 * tree-pass.h (all_passes): Remove, in favor of a field of the
6120 same name within the new class pass_manager.
6121 (all_small_ipa_passes): Likewise.
6122 (all_lowering_passes): Likewise.
6123 (all_regular_ipa_passes): Likewise.
6124 (all_lto_gen_passes): Likewise.
6125 (all_late_ipa_passes): Likewise.
6126 (passes_by_id): Likewise.
6127 (passes_by_id_size): Likewise.
6128 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6129 the new class pass_manager.
6130 (get_pass_for_id): Remove.
6132 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6134 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
6137 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6139 * arm.md (mulhi3): New expand pattern.
6141 2013-07-30 Jan Hubicka <jh@suse.cz>
6142 Martin Liska <marxin.liska@gmail.com>
6144 * profile.c (compute_value_histograms): Do not ICE when
6145 there is mismatch only on some counters.
6147 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6149 PR rtl-optimization/57637
6150 * function.c (move_insn_for_shrink_wrap): Also check the
6151 GEN set of the LIVE problem for the liveness analysis
6152 if it exists, otherwise give up.
6154 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6156 PR tree-optimization/57993
6157 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
6158 replaced statement in the candidate table.
6159 (phi_add_costs): Return infinite cost when the hidden basis does
6160 not dominate all phis on which the candidate is dependent.
6161 (replace_one_candidate): Record replaced statement in the
6164 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
6166 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
6167 (ashlv2si3): New expander.
6168 (*ashlv2si3_i): New define_insn_and_split.
6169 * predicates.md (float_operation): Allow patterns with three
6172 PR rtl-optimization/58021
6173 * mode-switching.c (create_pre_exit): Always split off preceding
6174 insns if we are not at the basic block head.
6176 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
6178 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
6179 (UCLIBC_DYNAMIC_LINKER): New macro.
6180 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
6182 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
6183 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
6185 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
6186 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
6187 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
6188 for SF and DF modes. Use ieee_quad_format for TF mode.
6189 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
6190 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
6191 (mips_option_override): Handle `-mnan=legacy'.
6192 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
6193 `-mabs=2008' and `-mnan=2008'.
6194 (OPTION_DEFAULT_SPECS): Add "nan" default.
6195 (ASM_SPEC): Handle `-mnan='.
6196 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
6197 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
6198 comment accordingly.
6199 (neg<mode>2): Likewise.
6200 * config/mips/mips.opt (mabs, mnan): New options.
6201 * doc/install.texi (Configuration): Document `--with-nan=' option.
6202 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
6204 (MIPS Options): Document them.
6205 * config.gcc <mips*-*-*>: Handle `--with-nan='.
6206 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
6207 * configure: Regenerate.
6208 * config.in: Regenerate.
6210 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6212 * config/i386/i386.md (float post-reload splitters): Do not check
6213 for subregs of SSE registers.
6215 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6216 H.J. Lu <hongjiu.lu@intel.com>
6220 * config/i386/i386.md (post-reload splitter
6221 to avoid partial SSE reg dependency stalls): New pattern.
6223 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
6225 * config/s390/s390.md ("movcc"): Swap load and store instructions.
6227 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
6229 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
6230 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
6232 2013-07-26 Cary Coutant <ccoutant@google.com>
6234 * dwarf2out.c (die_checksum_ordered): Don't include template
6235 instantiations in signature.
6236 (is_template_parameter): New function.
6237 (is_template_instantiation): New function.
6238 (generate_skeleton_bottom_up): Don't include template instantiations
6240 (generate_skeleton): Likewise.
6241 (break_out_comdat_types): Move recursive call to break out nested
6243 (prune_unused_types_mark_generic_parms_dies): Call
6244 is_template_parameter.
6246 2013-07-26 Ian Bolton <ian.bolton@arm.com>
6248 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
6249 uses vector registers.
6250 * config/aarch64/iterators.md: Add attributes rtn and vas.
6252 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6253 Richard Earnshaw <richard.earnshaw@arm.com>
6255 * combine.c (simplify_comparison): Re-canonicalize operands
6257 * config/arm/arm.md (movcond_addsi): New splitter.
6259 2013-07-25 Sterling Augustine <saugustine@google.com>
6261 * dwarf2out.c (size_of_pubnames): Move code to...
6262 (include_pubname_in_output): ...here. New.
6263 (want_pubnames): Rearrange.
6264 (output_pubnames): Call include_pubname_in_output. Move assertion.
6266 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6268 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
6270 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6273 * doc/extend.texi: Remove obsolete builtins. Fix
6274 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
6276 2013-07-25 Jan Hubicka <jh@suse.cz>
6278 * cgraph.c (release_function_body): Break out from ...
6279 (cgraph_release_function_body): ... this one; also release DECL_RESULT
6281 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
6282 old_tree in the map.
6283 (create_specialized_node): Update.
6284 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
6286 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
6287 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
6288 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
6289 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
6292 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6294 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
6295 addsi3_carryin_alt2_<optab>): Correct output template.
6297 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6299 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
6300 Adjust for arm_restrict_it.
6301 Remove trailing whitespace.
6303 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
6305 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
6306 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
6308 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
6310 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
6312 PR rtl-optimization/57960
6313 * lra-constraints.c (process_alt_operands): Use the right mode
6314 when checking strict_low.
6316 2013-07-25 Jan Hubicka <jh@suse.cz>
6318 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
6319 * cgraph.c (cgraph_remove_node): Do not release function body
6320 when in cgraph streaming.
6321 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
6322 in other partitions are not considered reachable; fix handling of
6325 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6327 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
6329 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6334 * config/arm/arm.md ("*sibcall_insn): Replace use of
6335 Ss with US. Adjust output for v5 and v4t.
6336 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
6338 * config/arm/constraints.md ("Ss"): Rename to US.
6340 2013-07-25 Terry Guo <terry.guo@arm.com>
6342 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
6343 shift_add/shift_sub0/shift_sub1 RTXs.
6345 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
6346 Anton Blanchard <anton@au1.ibm.com>
6348 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
6349 (altivec_vpks<VI_char>ss): Likewise.
6350 (altivec_vpks<VI_char>us): Likewise.
6351 (altivec_vpku<VI_char>us): Likewise.
6352 (altivec_vpku<VI_char>um): Likewise.
6354 2013-07-24 David Malcolm <dmalcolm@redhat.com>
6356 Introduce context class.
6358 * Makefile.in (CONTEXT_H): New.
6359 (OBJS): Add context.o.
6360 (toplev.o): Add CONTEXT_H to dependencies.
6363 * toplev.c (general_init): Create the singleton gcc::context instance.
6369 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
6371 PR rtl-optimization/57968
6372 * mode-switching.c (create_pre_exit): Allow instructions that
6373 don't set a return register to need a non-exit mode.
6375 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6376 Anton Blanchard <anton@au1.ibm.com>
6378 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
6379 operands to vperm for little endian.
6380 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
6381 of lvsl to create the control mask for a vperm for little endian.
6383 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6384 Anton Blanchard <anton@au1.ibm.com>
6386 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6387 two operands for little-endian.
6389 2013-07-23 Steve Ellcey <sellcey@mips.com>
6391 * config/mips/mips.c (mips_case_values_threshold): New.
6392 (TARGET_CASE_VALUES_THRESHOLD): Define.
6394 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6395 Anton Blanchard <anton@au1.ibm.com>
6397 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
6398 selection of field for vector splat in little endian mode.
6400 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6402 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
6403 expanders to rs6000.md.
6404 (ior<mode>3): Likewise.
6405 (and<mode>3): Likewise.
6406 (one_cmpl<mode>2): Likewise.
6407 (nor<mode>3): Likewise.
6408 (andc<mode>3): Likewise.
6409 (eqv<mode>3): Likewise.
6410 (nand<mode>3): Likewise.
6411 (orc<mode>3): Likewise.
6413 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
6416 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
6417 to split multi-word logical operations.
6418 (rs6000_split_logical_di): Likewise.
6419 (rs6000_split_logical): Likewise.
6421 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
6422 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
6423 and allow TImode operations in 32-bit.
6424 (vsx_and<mode>3_64bit): Likewise.
6425 (vsx_ior<mode>3_32bit): Likewise.
6426 (vsx_ior<mode>3_64bit): Likewise.
6427 (vsx_xor<mode>3_32bit): Likewise.
6428 (vsx_xor<mode>3_64bit): Likewise.
6429 (vsx_one_cmpl<mode>2_32bit): Likewise.
6430 (vsx_one_cmpl<mode>2_64bit): Likewise.
6431 (vsx_nor<mode>3_32bit): Likewise.
6432 (vsx_nor<mode>3_64bit): Likewise.
6433 (vsx_andc<mode>3_32bit): Likewise.
6434 (vsx_andc<mode>3_64bit): Likewise.
6435 (vsx_eqv<mode>3_32bit): Likewise.
6436 (vsx_eqv<mode>3_64bit): Likewise.
6437 (vsx_nand<mode>3_32bit): Likewise.
6438 (vsx_nand<mode>3_64bit): Likewise.
6439 (vsx_orc<mode>3_32bit): Likewise.
6440 (vsx_orc<mode>3_64bit): Likewise.
6442 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
6443 logical types in GPRs.
6445 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
6446 logical insns to rs6000.md, and allow TImode operations in
6448 (altivec_ior<mode>3): Likewise.
6449 (altivec_xor<mode>3): Likewise.
6450 (altivec_one_cmpl<mode>2): Likewise.
6451 (altivec_nor<mode>3): Likewise.
6452 (altivec_andc<mode>3): Likewise.
6454 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
6455 attributes for moving the 128-bit logical operations into
6457 (BOOL_REGS_OUTPUT): Likewise.
6458 (BOOL_REGS_OP1): Likewise.
6459 (BOOL_REGS_OP2): Likewise.
6460 (BOOL_REGS_UNARY): Likewise.
6461 (BOOL_REGS_AND_CR0): Likewise.
6462 (one_cmpl<mode>2): Add support for DI logical operations on
6463 32-bit, splitting the operations to 32-bit.
6467 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
6468 changes to combine the 32/64-bit code, allow logical operations on
6469 TI mode in 32-bit, and to use similar match_operator patterns like
6470 scalar mode uses. Combine the Altivec and VSX code for logical
6471 operations, and move it here.
6472 (ior<mode>3, 128-bit types): Likewise.
6473 (xor<mode>3, 128-bit types): Likewise.
6474 (one_cmpl<mode>3, 128-bit types): Likewise.
6475 (nor<mode>3, 128-bit types): Likewise.
6476 (andc<mode>3, 128-bit types): Likewise.
6477 (eqv<mode>3, 128-bit types): Likewise.
6478 (nand<mode>3, 128-bit types): Likewise.
6479 (orc<mode>3, 128-bit types): Likewise.
6480 (and<mode>3_internal): Likewise.
6481 (bool<mode>3_internal): Likewise.
6482 (boolc<mode>3_internal1): Likewise.
6483 (boolc<mode>3_internal2): Likewise.
6484 (boolcc<mode>3_internal1): Likewise.
6485 (boolcc<mode>3_internal2): Likewise.
6486 (eqv<mode>3_internal1): Likewise.
6487 (eqv<mode>3_internal2): Likewise.
6488 (one_cmpl1<mode>3_internal): Likewise.
6490 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6492 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6493 Rename flag_stack_usage to flag_stack_usage_info.
6495 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6497 * config/microblaze/sync.md: New file.
6498 * config/microblaze/microblaze.md: Include sync.md
6499 * config/microblaze/microblaze.c: Add print_operand 'y'.
6500 * config/microblaze/constraints.md: Add memory_contraint
6501 'Q' which is a single register.
6503 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6505 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6507 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6509 * reload.c (find_reloads): Exit loop once we find this operand
6510 cannot be reloaded somehow for this alternative.
6512 * reload.c (find_reloads): Exit loop once we find a hard register.
6514 * rtlanal.c (computed_jump_p): Exit loop once we find label
6517 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6519 * cfgloopmanip.c (remove_path): Exit loop after setting
6522 * gensupport.c (subst_dup): Avoid loop if code is not
6523 MATCH_DUP nor MATCH_OP_DUP.
6525 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6527 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6529 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6531 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6532 true for SP_REGNUM if mode == ptr_mode.
6533 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6534 with value R0_REGNUM + 31.
6536 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6538 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6539 pad pointer-typed argument downward.
6541 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6543 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6544 and __ILP32__ when the ILP32 model is in use.
6546 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6548 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6549 (aarch64_load_symref_appropriately): In the case of
6550 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6551 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6552 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6553 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6554 if 'mode' doesn't equal to 'ptr_mode'.
6555 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6556 'vcall_offset'; change to call aarch64_emit_move differently depending
6557 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6558 to calculate the upper bound of 'vcall_offset'.
6559 (aarch64_cannot_force_const_mem): Change to also return true if
6561 (aarch64_legitimize_reload_address): In the case of large
6562 displacements, add new local variable 'xmode' and an assertion
6563 based on it; change to use 'xmode' to generate the new rtx and
6565 (aarch64_asm_trampoline_template): Change to generate the template
6566 differently depending on TARGET_ILP32 or not; change to use
6567 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6568 (aarch64_trampoline_size): Removed.
6569 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
6570 and replace immediate literals with it. Change to use 'ptr_mode'
6571 instead of 'DImode' and call convert_memory_address if the mode
6572 of 'fnaddr' doesn't equal to 'ptr_mode'.
6573 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
6575 (aarch64_elf_asm_destructor): Likewise.
6576 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
6577 on TARGET_ILP32 instead of aarch64_trampoline_size.
6578 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
6579 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
6580 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
6581 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
6582 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
6583 (storewb_pair<GPI:mode>_<P:mode>): ... this.
6584 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
6585 depending on the value of 'mode'.
6586 (add_losym_<mode>): New.
6587 (ldr_got_small_<mode>): New, based on ldr_got_small.
6588 (ldr_got_small): Remove.
6589 (ldr_got_small_sidi): New.
6590 * config/aarch64/iterators.md (P): New.
6591 (PTR): Change to 'ptr_mode' in the condition.
6593 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6595 * config.gcc (aarch64*-*-*): Support --with-abi.
6596 (aarch64*-*-elf): Support --with-multilib-list.
6597 (aarch64*-*-linux*): Likewise.
6598 (supported_defaults): Add abi to aarch64*-*-*.
6599 * configure.ac: Mention AArch64 for --with-multilib-list.
6600 * configure: Re-generated.
6601 * config/aarch64/biarchilp32.h: New file.
6602 * config/aarch64/biarchlp64.h: New file.
6603 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
6605 (MULTILIB_DEFAULTS): Ditto.
6606 (DRIVER_SELF_SPECS): Ditto.
6607 (ASM_SPEC): Update to also substitute -mabi.
6608 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
6609 file whose name depends on -mabi= and -mbig-endian.
6610 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
6612 (POINTER_SIZE): New define.
6613 (POINTERS_EXTEND_UNSIGNED): Ditto.
6614 (enum aarch64_abi_type): New enumeration tag.
6615 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
6616 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
6617 (TARGET_ILP32): New define.
6618 * config/aarch64/aarch64.opt (mabi): New.
6620 (ilp32, lp64): New values for -mabi.
6621 * config/aarch64/t-aarch64 (comma): New define.
6622 (MULTILIB_OPTIONS): Ditto.
6623 (MULTILIB_DIRNAMES): Ditto.
6624 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
6625 * doc/invoke.texi: Document -mabi for AArch64.
6627 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
6629 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
6631 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6632 Anton Blanchard <anton@au1.ibm.com>
6634 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6635 endianness when selecting field to splat.
6637 2013-07-22 Eric Christopher <echristo@gmail.com>
6639 * dwarf2out.c (die_odr_checksum): New function to use
6640 CHECKSUM_ macros and ULEB128 for DIE tag.
6641 (generate_type_signature): Use.
6643 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6645 * config.gcc (sparc*-*-*): Accept leon3 processor.
6646 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6647 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6648 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6649 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6650 (mfix-ut699): Adjust comment.
6651 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6652 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6653 (CPP_CPU_SPEC): Likewise.
6654 (ASM_CPU_SPEC): Likewise.
6655 * config/sparc/sparc.c (leon3_cost): New constant.
6656 (sparc_option_override): Add leon3 support.
6657 (mem_ref): New function.
6658 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6659 (sparc_do_work_around_errata): Look into the instruction in the delay
6660 slot and adjust accordingly. Add fix for the data cache nullify issues
6661 of the UT699. Change insertion position for the NOP.
6662 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6663 (leon3_load): New reservation.
6664 (leon_store): Bump latency to 2.
6665 (grfpu): New automaton.
6666 (grfpu_alu): New unit.
6667 (grfpu_ds): Likewise.
6668 (leon_fp_alu): Adjust.
6669 (leon_fp_mult): Delete.
6670 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6671 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6672 * config/sparc/sparc.md (cpu): Add leon3.
6673 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6675 (atomic_test_and_set): Likewise.
6678 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6680 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6681 default for R5900 targets.
6682 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6683 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6684 * config/mips/mips.c (mips_option_override): Report an error for
6685 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6686 for -march=r5900 -mhard-float.
6688 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6690 * df-problems.c (can_move_insns_across): Exit loop once we
6691 find a non-fixed, non-global register.
6693 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6696 * omega.c (omega_eliminate_red): Break after setting red_found.
6697 (omega_problem_has_red_equations): Similarly after setting found.
6698 (omega_query_variable): Similarly after setting coupled.
6700 2013-07-22 Marek Polacek <polacek@redhat.com>
6702 * gimplify.c: Don't include gimple.h twice.
6704 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6706 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6707 instead of TARGET_THUMB1.
6708 (Pz): New constraint.
6709 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6711 (compare_negsi_si): Likewise.
6712 (compare_addsi2_op0): Likewise.
6713 (compare_addsi2_op1): Likewise.
6714 (addsi3_carryin_<optab>): Likewise.
6715 (addsi3_carryin_alt2_<optab>): Likewise.
6716 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6717 for arm_restrict_it.
6718 (subsi3_carryin): Likewise.
6719 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6720 (minmax_arithsi): Disable for arm_restrict_it.
6721 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6722 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6723 (satsi_<SAT:code>_shift): Likewise.
6724 (arm_shiftsi3): Add alternative for 16-bit encoding.
6725 (arm32_movhf): Disable for arm_restrict_it.
6726 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6727 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6729 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6731 * config/arm/arm.md (attribute "insn"): Delete.
6732 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6733 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6734 (not_shiftsi): Update for attribute change.
6735 (not_shiftsi_compare0): Likewise.
6736 (not_shiftsi_compare0_scratch): Likewise.
6737 (arm_one_cmplsi2): Likewise.
6738 (thumb1_one_cmplsi2): Likewise.
6739 (notsi_compare0): Likewise.
6740 (notsi_compare0_scratch): Likewise.
6741 (thumb1_movdi_insn): Likewise.
6742 (arm_movsi_insn): Likewise.
6743 (movhi_insn_arch4): Likewise.
6744 (movhi_bytes): Likewise.
6745 (arm_movqi_insn): Likewise.
6746 (thumb1_movqi_insn): Likewise.
6747 (arm32_movhf): Likewise.
6748 (thumb1_movhf): Likewise.
6749 (arm_movsf_soft_insn): Likewise.
6750 (thumb1_movsf_insn): Likewise.
6751 (thumb_movdf_insn): Likewise.
6752 (movsicc_insn): Likewise.
6753 (movsfcc_soft_insn): Likewise.
6754 (and_scc): Likewise.
6755 (cond_move): Likewise.
6756 (if_move_not): Likewise.
6757 (if_not_move): Likewise.
6758 (if_shift_move): Likewise.
6759 (if_move_shift): Likewise.
6760 (if_shift_shift): Likewise.
6761 (if_not_arith): Likewise.
6762 (if_arith_not): Likewise.
6763 (cond_move_not): Likewise.
6764 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6765 (neon_mov<mode>): Likewise.
6766 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6767 (thumb2_movsi_vfp): Likewise.
6768 (movsf_vfp): Likewise.
6769 (thumb2_movsf_vfp): Likewise.
6770 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6772 (cortexa7_older_only): Likewise.
6773 (cortexa7_younger): Likewise.
6774 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6775 (1020alu_shift_op): Likewise.
6776 (1020alu_shift_reg_op): Likewise.
6777 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6778 (alu_shift_op): Likewise.
6779 (alu_shift_reg_op): Likewise.
6780 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6781 (11_alu_shift_op): Likewise.
6782 (11_alu_shift_reg_op): Likewise.
6783 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6784 (9_alu_shift_reg_op): Likewise.
6785 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6787 (cortex_a15_alu_shift): Likewise.
6788 (cortex_a15_alu_shift_reg): Likewise.
6789 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6790 (cortex_a5_alu_shift): Likewise.
6791 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6793 (cortex_a53_alu_shift): Likewise.
6794 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6796 (cortex_a7_alu_reg): Likewise.
6797 (cortex_a7_alu_shift): Likewise.
6798 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6799 (cortex_a8_alu_shift): Likewise.
6800 (cortex_a8_alu_shift_reg): Likewise.
6801 (cortex_a8_mov): Likewise.
6802 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6803 (cortex_a9_dp_shift): Likewise.
6804 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6805 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6806 (cortex_r4_mov): Likewise.
6807 (cortex_r4_alu_shift): Likewise.
6808 (cortex_r4_alu_shift_reg): Likewise.
6809 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6810 (526_alu_shift_op): Likewise.
6811 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6812 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6813 (626te_alu_shift_op): Likewise.
6814 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6815 (726te_alu_op): Likewise.
6816 (726te_alu_shift_op): Likewise.
6817 (726te_alu_shift_reg_op): Likewise.
6818 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6819 (mp626_alu_shift_op): Likewise.
6820 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6821 (pj4_alu_e1_conds): Likewise.
6822 (pj4_alu): Likewise.
6823 (pj4_alu_conds): Likewise.
6824 (pj4_shift): Likewise.
6825 (pj4_shift_conds): Likewise.
6826 (pj4_alu_shift): Likewise.
6827 (pj4_alu_shift_conds): Likewise.
6829 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6831 * config/arm/predicates.md (shiftable_operator_strict_it):
6833 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
6834 Disable cond_exec version for arm_restrict_it.
6835 (thumb2_smaxsi3): Convert to generate cond_exec.
6836 (thumb2_sminsi3): Likewise.
6837 (thumb32_umaxsi3): Likewise.
6838 (thumb2_uminsi3): Likewise.
6839 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
6840 (thumb2_neg_abssi2): Likewise.
6841 (thumb2_mov_scc): Add alternative for 16-bit encoding.
6842 (thumb2_movsicc_insn): Adjust alternatives.
6843 (thumb2_mov_negscc): Disable for arm_restrict_it.
6844 (thumb2_mov_negscc_strict_it): New pattern.
6845 (thumb2_mov_notscc_strict_it): New pattern.
6846 (thumb2_mov_notscc): Disable for arm_restrict_it.
6847 (thumb2_ior_scc): Likewise.
6848 (thumb2_ior_scc_strict_it): New pattern.
6849 (thumb2_cond_move): Adjust for arm_restrict_it.
6850 (thumb2_cond_arith): Disable for arm_restrict_it.
6851 (thumb2_cond_arith_strict_it): New pattern.
6852 (thumb2_cond_sub): Adjust for arm_restrict_it.
6853 (thumb2_movcond): Likewise.
6854 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
6855 (thumb2_zero_extendhisi2_v6): Likewise.
6856 (thumb2_zero_extendqisi2_v6): Likewise.
6857 (orsi_notsi_si): Likewise.
6858 (orsi_not_shiftsi_si): Likewise.
6860 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
6862 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
6863 instruction sequence is 1 byte shorter.
6865 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
6867 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
6868 it is not needed after split.
6870 2013-07-20 Iain Sandoe <iain@codesourcery.com>
6873 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
6874 second label for nonlocal goto receivers. Don't output pic base labels
6875 unless we're producing PIC; mark that action unreachable().
6876 (ix86_save_reg): If the function contains a nonlocal label, save the
6878 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
6879 * config/darwin.c (emitted_pic_label_num): New GTY.
6880 (update_pic_label_number_if_needed): New.
6881 (machopic_output_function_base_name): Adjust for nonlocal receiver
6883 (machopic_should_output_picbase_label): New.
6884 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
6885 (nonlocal_goto_receiver): New insn and split.
6887 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
6889 * config/aarch64/aarch64-builtins.c
6890 (aarch64_fold_builtin): Fold abs in all modes.
6891 * config/aarch64/aarch64-simd-builtins.def
6892 (abs): Enable for all modes.
6893 * config/aarch64/arm_neon.h
6894 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
6895 (vabs_f64): Add missing intrinsic.
6897 2013-07-19 Ian Bolton <ian.bolton@arm.com>
6899 * config/aarch64/arm_neon.h (vabs_s64): New function
6901 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
6904 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
6905 * config/avr/avr.md (adjust_len): Add `round'.
6906 * config/avr/avr-protos.h (avr_out_round): New prototype.
6907 (avr_out_plus): Add `out_label' argument.
6908 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
6909 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
6910 Handle the case where `insn' is just a pattern.
6911 (avr_out_bitop): Handle the case where `insn' is just a pattern.
6912 (avr_out_round): New function.
6913 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
6915 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
6917 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6918 Add check for flag_stack_usage to handle -fstack-usage support
6920 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
6922 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
6923 interaction for new Power8 flags and VSX.
6925 2013-07-18 Sriraman Tallam <tmsriram@google.com>
6928 * tree-inline.c (expand_call_inline): Emit errors during
6929 early_inlining only if optimization is not turned on.
6931 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6935 * passes.c (init_optimization_passes): Move the construction of
6936 the pass hierarchy into a new passes.def file.
6938 * Makefile.in (passes.o): Add dependency on passes.def.
6940 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6942 * passes.c (init_optimization_passes): Introduce macros for
6943 constructing the tree of passes (INSERT_PASSES_AFTER,
6944 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
6945 TERMINATE_PASS_LIST).
6947 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
6948 Wei Mi <wmi@google.com>
6950 PR rtl-optimization/57878
6951 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
6953 (reload_pseudo_compare_func): Check nregs first for reload
6956 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6958 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
6960 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
6962 * read-rtl.c (validate_const_int): Once an invalid character is
6963 seen, quit the loop.
6965 * gengtype.c (write_roots): Similarly once we find the "deletable"
6966 or "if_marked" option.
6968 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6970 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
6971 "xtab" and "sat". Move value "clz" from here to ...
6972 (attriubte "type"): ... here.
6973 (satsi_<SAT:code>): Delete "insn" attribute.
6974 (satsi_<SAT:code>_shift): Likewise.
6975 (arm_zero_extendqisi2addsi): Likewise.
6976 (arm_extendqisi2addsi): Likewise.
6977 (clzsi2): Update for attribute changes.
6978 (rbitsi2): Likewise.
6979 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
6981 (arm_usatsihi): Likewise.
6982 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6984 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6986 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
6987 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
6988 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
6989 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
6990 in alphabetical order.
6991 (attribute "core_cycles"): Update for attribute changes.
6992 (arm_addsi3): Likewise.
6993 (addsi3_compare0): Likewise.
6994 (addsi3_compare0_scratch): Likewise.
6995 (addsi3_compare_op1): Likewise.
6996 (addsi3_compare_op2): Likewise.
6997 (compare_addsi2_op0): Likewise.
6998 (compare_addsi2_op1): Likewise.
6999 (addsi3_carryin_shift_<optab>): Likewise.
7000 (subsi3_carryin_shift): Likewise.
7001 (rsbsi3_carryin_shift): Likewise.
7002 (arm_subsi3_insn): Likewise.
7003 (subsi3_compare0): Likewise.
7004 (subsi3_compare): Likewise.
7005 (arm_andsi3_insn): Likewise.
7006 (thumb1_andsi3_insn): Likewise.
7007 (andsi3_compare0): Likewise.
7008 (andsi3_compare0_scratch): Likewise.
7009 (zeroextractsi_compare0_scratch
7010 (andsi_not_shiftsi_si): Likewise.
7011 (iorsi3_insn): Likewise.
7012 (iorsi3_compare0): Likewise.
7013 (iorsi3_compare0_scratch): Likewise.
7014 (arm_xorsi3): Likewise.
7015 (thumb1_xorsi3_insn): Likewise.
7016 (xorsi3_compare0): Likewise.
7017 (xorsi3_compare0_scratch): Likewise.
7018 (satsi_<SAT:code>_shift): Likewise.
7020 (arm_shiftsi3): Likewise.
7021 (shiftsi3_compare0): Likewise.
7022 (not_shiftsi): Likewise.
7023 (not_shiftsi_compare0): Likewise.
7024 (not_shiftsi_compare0_scratch): Likewise.
7025 (arm_one_cmplsi2): Likewise.
7026 (thumb_one_complsi2): Likewise.
7027 (notsi_compare0): Likewise.
7028 (notsi_compare0_scratch): Likewise.
7029 (thumb1_zero_extendhisi2): Likewise.
7030 (arm_zero_extendhisi2): Likewise.
7031 (arm_zero_extendhisi2_v6): Likewise.
7032 (arm_zero_extendhisi2addsi): Likewise.
7033 (thumb1_zero_extendqisi2): Likewise.
7034 (thumb1_zero_extendqisi2_v6): Likewise.
7035 (arm_zero_extendqisi2): Likewise.
7036 (arm_zero_extendqisi2_v6): Likewise.
7037 (arm_zero_extendqisi2addsi): Likewise.
7038 (thumb1_extendhisi2): Likewise.
7039 (arm_extendhisi2): Likewise.
7040 (arm_extendhisi2_v6): Likewise.
7041 (arm_extendqisi): Likewise.
7042 (arm_extendqisi_v6): Likewise.
7043 (arm_extendqisi2addsi): Likewise.
7044 (thumb1_extendqisi2): Likewise.
7045 (thumb1_movdi_insn): Likewise.
7046 (arm_movsi_insn): Likewise.
7047 (movsi_compare0): Likewise.
7048 (movhi_insn_arch4): Likewise.
7049 (movhi_bytes): Likewise.
7050 (arm_movqi_insn): Likewise.
7051 (thumb1_movqi_insn): Likewise.
7052 (arm32_movhf): Likewise.
7053 (thumb1_movhf): Likewise.
7054 (arm_movsf_soft_insn): Likewise.
7055 (thumb1_movsf_insn): Likewise.
7056 (movdf_soft_insn): Likewise.
7057 (thumb_movdf_insn): Likewise.
7058 (arm_cmpsi_insn): Likewise.
7059 (cmpsi_shiftsi): Likewise.
7060 (cmpsi_shiftsi_swp): Likewise.
7061 (arm_cmpsi_negshiftsi_si): Likewise.
7062 (movsicc_insn): Likewise.
7063 (movsfcc_soft_insn): Likewise.
7064 (arith_shiftsi): Likewise.
7065 (arith_shiftsi_compare0
7066 (arith_shiftsi_compare0_scratch
7067 (sub_shiftsi): Likewise.
7068 (sub_shiftsi_compare0
7069 (sub_shiftsi_compare0_scratch
7070 (and_scc): Likewise.
7071 (cond_move): Likewise.
7072 (if_plus_move): Likewise.
7073 (if_move_plus): Likewise.
7074 (if_move_not): Likewise.
7075 (if_not_move): Likewise.
7076 (if_shift_move): Likewise.
7077 (if_move_shift): Likewise.
7078 (if_shift_shift): Likewise.
7079 (if_not_arith): Likewise.
7080 (if_arith_not): Likewise.
7081 (cond_move_not): Likewise.
7082 (thumb1_ashlsi3): Set type attribute.
7083 (thumb1_ashrsi3): Likewise.
7084 (thumb1_lshrsi3): Likewise.
7085 (thumb1_rotrsi3): Likewise.
7086 (shiftsi3_compare0_scratch): Likewise.
7087 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
7088 (neon_mov<mode>): Likewise.
7089 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
7091 (thumb2_movsi_insn): Likewise.
7092 (thumb2_cmpsi_neg_shiftsi): Likewise.
7093 (thumb2_extendqisi_v6): Likewise.
7094 (thumb2_zero_extendhisi2_v6): Likewise.
7095 (thumb2_zero_extendqisi2_v6): Likewise.
7096 (thumb2_shiftsi3_short): Likewise.
7097 (thumb2_addsi3_compare0_scratch): Likewise.
7098 (orsi_not_shiftsi_si): Likewise.
7099 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7100 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
7102 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7103 (1020alu_shift_op): Likewise.
7104 (1020alu_shift_reg_op): Likewise.
7105 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7106 (alu_shift_op): Likewise.
7107 (alu_shift_reg_op): Likewise.
7108 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7109 (11_alu_shift_op): Likewise.
7110 (11_alu_shift_reg_op): Likewise.
7111 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7112 (9_alu_shift_reg_op): Likewise.
7113 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7115 (cortex_a15_alu_shift): Likewise.
7116 (cortex_a15_alu_shift_reg): Likewise.
7117 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
7119 (cortex_a5_alu_shift): Likewise.
7120 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
7122 (cortex_a53_alu_shift): Likewise.
7123 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7125 (cortex_a7_alu_reg): Likewise.
7126 (cortex_a7_alu_shift): Likewise.
7127 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
7129 (cortex_a8_alu_shift): Likewise.
7130 (cortex_a8_alu_shift_reg): Likewise.
7131 (cortex_a8_mov): Likewise.
7132 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7133 (cortex_a9_dp_shift): Likewise.
7134 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
7136 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
7138 (cortex_r4_mov): Likewise.
7139 (cortex_r4_alu_shift): Likewise.
7140 (cortex_r4_alu_shift_reg): Likewise.
7141 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7142 (526_alu_shift_op): Likewise.
7143 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7144 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7145 (626te_alu_shift_op): Likewise.
7146 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
7147 (726te_alu_op): Likewise.
7148 (726te_alu_shift_op): Likewise.
7149 (726te_alu_shift_reg_op): Likewise.
7150 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7151 (mp626_alu_shift_op): Likewise.
7152 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
7153 (pj4_alu_e1_conds): Likewise.
7154 (pj4_alu): Likewise.
7155 (pj4_alu_conds): Likewise.
7156 (pj4_shift): Likewise.
7157 (pj4_shift_conds): Likewise.
7158 (pj4_alu_shift): Likewise.
7159 (pj4_alu_shift_conds): Likewise.
7160 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7162 (cortexa7_older_only): Likewise.
7163 (cortexa7_younger): Likewise.
7165 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7167 * ipa-pure-const.c (generate_summary): Rename to...
7168 (pure_const_generate_summary): ... this.
7170 2013-07-17 Iain Sandoe <iain@codesourcery.com>
7172 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
7174 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
7177 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
7180 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7182 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
7183 enabled without -march=zEC12.
7184 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
7187 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7189 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
7190 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
7191 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
7192 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
7193 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
7194 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
7195 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
7197 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
7198 and ISA_HAS_NMADD3_NMSUB3.
7199 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
7200 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
7201 (nmsub4<mode>, nmsub3<mode>): Likewise.
7202 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
7204 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7206 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
7207 TARGET_MIPS5400 checking.
7209 2013-07-16 Jakub Jelinek <jakub@redhat.com>
7210 Peter Bergner <bergner@vnet.ibm.com>
7212 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
7213 registers in the comment.
7214 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
7215 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
7216 rather than FIRST_PSEUDO_REGISTERS.
7218 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
7220 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
7221 enable extra ISA flags with TARGET_HTM.
7223 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7225 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
7228 2013-07-15 Cong Hou <congh@google.com>
7230 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
7231 in compare function for sorting.
7233 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
7235 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
7236 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
7237 * config/rs6000/rs6000.opt: Add -mhtm option.
7238 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
7239 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
7240 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7241 __HTM__ if the HTM instructions are available.
7242 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
7243 htm_spr_reg_operand): New define_predicates.
7244 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
7245 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
7247 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
7248 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
7249 HTM builtin functions.
7250 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
7251 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
7252 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
7253 (rs6000_builtin_mask_calculate): Likewise.
7254 (rs6000_option_override_internal): Likewise.
7255 (bdesc_htm): Add new HTM builtin support.
7256 (htm_spr_num): New function.
7257 (htm_spr_regno): Likewise.
7258 (rs6000_htm_spr_icode): Likewise.
7259 (htm_expand_builtin): Likewise.
7260 (htm_init_builtins): Likewise.
7261 (rs6000_expand_builtin): Add support for HTM builtin functions.
7262 (rs6000_init_builtins): Likewise.
7263 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
7265 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
7266 (TARGET_HTM, MASK_HTM): Define macros.
7267 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
7268 (FIXED_REGISTERS): Likewise.
7269 (CALL_USED_REGISTERS): Likewise.
7270 (CALL_REALLY_USED_REGISTERS): Likewise.
7271 (REG_ALLOC_ORDER): Likewise.
7272 (enum reg_class): Likewise.
7273 (REG_CLASS_NAMES): Likewise.
7274 (REG_CLASS_CONTENTS): Likewise.
7275 (REGISTER_NAMES): Likewise.
7276 (ADDITIONAL_REGISTER_NAMES): Likewise.
7277 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
7278 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
7279 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
7280 * config/rs6000/htm.md: New file.
7281 * config/rs6000/htmintrin.h: New file.
7282 * config/rs6000/htmxlintrin.h: New file.
7284 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
7286 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7287 Define SYMBOL_TINY_GOT, update comment.
7288 * config/aarch64/aarch64.c
7289 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
7290 (aarch64_expand_mov_immediate): Likewise.
7291 (aarch64_print_operand): Likewise.
7292 (aarch64_classify_symbol): Likewise.
7293 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
7294 (ldr_got_tiny): Define.
7296 2013-07-13 Tobias Grosser <tobias@grosser.es>
7298 PR tree-optimization/54094
7299 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
7300 scheduling dimension for the parallelism check from the polyhedral
7301 information in the AST.
7302 * graphite-dependences.c (carries_deps): Do not assume the schedule is
7305 2013-07-13 Jason Merrill <jason@redhat.com>
7307 * print-tree.c (debug_vec_tree): Use debug_raw.
7308 (debug_raw (vec<tree, va_gc> &)): New.
7309 (debug_raw (vec<tree, va_gc> *)): New.
7310 * tree.h: Declare them.
7312 2013-07-13 Bin Cheng <bin.cheng@arm.com>
7314 * ifcvt.c (ifcvt_after_combine): New static variable.
7315 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
7317 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
7318 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
7319 rest_of_handle_if_after_reload): Pass new argument for if_convert.
7321 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
7323 * config/mips/mips.c (mips_expand_call): Remove empty statement.
7325 2013-07-12 Michael Matz <matz@suse.de>
7328 * convert.c (convert_to_real): Reject non-float inner types.
7330 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
7332 * config/aarch64/aarch64-protos.h
7333 (aarch64_simd_immediate_valid_for_move): Remove.
7334 * config/aarch64/aarch64.c (simd_immediate_info): New member.
7335 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
7337 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
7339 2013-07-11 Steve Ellcey <sellcey@mips.com>
7341 * config/mips/mips.c (mips_conditional_register_usage): Do not
7342 use t[0-7] registers in MIPS16 mode when optimizing for size.
7344 2013-07-11 Sriraman Tallam <tmsriram@google.com>
7346 * config/i386/i386.c (dispatch_function_versions): Fix array
7347 indexing of function_version_info to match actual_versions.
7349 2013-07-11 Teresa Johnson <tejohnson@google.com>
7351 * vec.h (struct va_gc): Move release out-of-line.
7352 (va_gc::release): Call ggc_free on released vec.
7354 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7356 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
7357 Require GOT register as additional operand in UNSPEC.
7358 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
7359 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7360 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7361 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
7362 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
7363 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
7364 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
7366 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
7369 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
7370 name seen by assembler/linker rather if available.
7372 2013-07-11 Andreas Schwab <schwab@suse.de>
7374 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
7376 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
7378 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
7380 2013-07-10 Joseph Myers <joseph@codesourcery.com>
7382 * doc/tm.texi.in: Move hook documentation to ....
7383 * target.def: ... here.
7385 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
7387 * doc/tm.texi: Regenerate.
7389 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
7392 * doc/invoke.texi: Document -Wconditionally-supported.
7394 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7397 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
7400 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7403 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
7404 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
7405 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
7406 Remove duplicate devices.
7407 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
7408 * config/avr/t-multilib: Regenerate.
7409 * config/avr/avr-tables.opt: Regenerate.
7410 * doc/avr-mmcu.texi: Regenerate.
7412 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7415 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
7417 2013-07-10 Graham Stott <graham.stott@btinternet.com>
7419 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
7420 the cost of MULT when optimizing for size.
7422 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7424 * config/cr16/cr16-protos.h: Don't include target.h.
7426 2013-07-09 Joseph Myers <joseph@codesourcery.com>
7428 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
7429 adjust register size for TDmode and TFmode for VSX registers.
7431 2013-07-08 Kai Tietz <ktietz@redhat.com>
7434 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
7435 hook_bool_const_tree_true.
7437 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7439 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
7440 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
7441 * config/s390/s390.md: Define FPR*_REGNUM constants.
7442 Fix FPR2_REGNUM constant (18 -> 17).
7443 ("*trunc<BFP:mode><DFP_ALL:mode>2")
7444 ("*trunc<DFP_ALL:mode><BFP:mode>2")
7445 ("trunc<BFP:mode><DFP_ALL:mode>2")
7446 ("trunc<DFP_ALL:mode><BFP:mode>2")
7447 ("*extend<BFP:mode><DFP_ALL:mode>2")
7448 ("*extend<DFP_ALL:mode><BFP:mode>2")
7449 ("extend<BFP:mode><DFP_ALL:mode>2")
7450 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
7453 2013-07-08 Graham Stott <graham.stott@btinternet.com>
7455 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
7457 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7459 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
7460 and cfun_fpr_bit_p to cfun_fpr_save_p.
7461 (s390_frame_area, s390_register_info, s390_frame_info)
7462 (s390_emit_prologue, s390_emit_epilogue)
7463 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
7465 * config/s390/s390.h: Define *_REGNUM macros for floating point
7468 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
7470 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
7472 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
7474 PR rtl-optimization/57786
7475 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
7476 and break out of the loop when it is set to false.
7478 2013-07-08 Jakub Jelinek <jakub@redhat.com>
7481 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7482 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7483 (const_int 63)) 0)).
7484 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7485 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7486 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7488 PR rtl-optimization/57829
7489 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7490 mask bits outside of mode are just sign-extension from mode to HWI.
7492 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7494 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7495 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7496 adjust_address instead of change_address to keep info about alignment.
7497 (emit_strmov): Remove.
7498 (emit_memmov): New function.
7499 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7500 (expand_movmem_epilogue): Likewise and return updated rtx for
7502 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7503 destination and source.
7504 (decide_alignment): Refactor, handle vector_loop.
7505 (ix86_expand_movmem): Likewise.
7506 (ix86_expand_setmem): Likewise.
7507 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7509 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
7511 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7512 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7514 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7516 * config/i386/sse.md (sse_movlhps): Change alternative 3
7517 of operand 2 to "m".
7519 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7522 * config/i386/sse.md (iptr): New mode attribute.
7523 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7524 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7525 (<sse>_vmmul<mode>3): Ditto.
7526 (<sse>_vmdiv<mode>3): Ditto.
7527 (sse_vmrcpv4sf2): Ditto.
7528 (<sse>_vmsqrt<mode>2): Ditto.
7529 (sse_vmrsqrtv4sf2): Ditto.
7530 (<sse>_vm<code><mode>3): Ditto.
7531 (avx_vmcmp<mode>3): Ditto.
7532 (<sse>_vmmaskcmp<mode>3): Ditto.
7533 (<sse>_comi): Ditto.
7534 (<sse>_ucomi): Ditto.
7535 (*xop_vmfrcz_<mode>): Ditto.
7536 (*fmai_fmadd_<mode>): Ditto.
7537 (*fmai_fmsub_<mode>): Ditto.
7538 (*fmai_fnmadd_<mode>): Ditto.
7539 (*fmai_fnmsub_<mode>): Ditto.
7540 (*fma4i_vmfmadd_<mode>): Ditto.
7541 (*fma4i_vmfmsub_<mode>): Ditto.
7542 (*fma4i_vmfnmadd_<mode>): Ditto.
7543 (*fma4i_vmfnmsub_<mode>): Ditto.
7544 (*xop_vmfrcz_<mode>): Ditto.
7545 (sse_cvtps2pi): Ditto.
7546 (sse_cvttps2pi): Ditto.
7547 (sse_cvtss2si): Ditto.
7548 (sse_cvtss2si_2): Ditto.
7549 (sse_cvtss2siq_2): Ditto.
7550 (sse_cvttss2si): Ditto.
7551 (sse_cvttss2siq): Ditto.
7552 (sse_cvtsd2si): Ditto.
7553 (sse_cvtsd2si_2): Ditto.
7554 (sse_cvtsd2siq_2): Ditto.
7555 (sse_cvttsd2si): Ditto.
7556 (sse_cvttsd2siq): Ditto.
7557 (sse_cvtsd2ss): Ditto.
7558 (sse_cvtss2sd): Ditto.
7559 (avx2_pbroadcast<mode>): Ditto.
7560 (avx2_pbroadcast<mode>_1): Ditto.
7561 (*avx_vperm_broadcast_v4sf): Ditto.
7563 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7564 (sse_movlhps): Ditto.
7565 (sse_storehps): Ditto.
7566 (sse_loadhps): Ditto.
7567 (sse_storelps): Ditto.
7568 (sse_loadlps): Ditto.
7569 (*vec_concatv4sf): Ditto.
7570 (*vec_interleave_highv2df): Ditto.
7571 (*vec_interleave_lowv2df): Ditto.
7572 (*vec_extractv2df_1_sse): Ditto.
7573 (*vec_extractv2df_0_sse): Ditto.
7574 (sse2_storelpd): Ditto.
7575 (sse2_loadlpd): Ditto.
7576 (sse2_movsd): Ditto.
7577 (*vec_concatv4si): Ditto.
7578 (vec_concatv2di): Ditto.
7580 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
7581 for Intel asm dialect.
7582 (mmx_punpcklwd): Ditto.
7583 (mmx_punpckldq): Ditto.
7585 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
7586 for intel assembler dialect.
7588 2013-07-06 Jakub Jelinek <jakub@redhat.com>
7591 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
7592 for BUILT_IN_C{LZ,LRSB}*.
7593 * tree.h (CASE_INT_FN): Add FN##IMAX case.
7594 * tree-vrp.c (extract_range_basic): Handle
7595 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
7596 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
7597 fall thru to code calling set_value*.
7598 * builtins.c (expand_builtin): Remove *IMAX cases.
7599 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
7600 if width is bigger than 2*HWI.
7602 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
7604 PR rtl-optimization/55342
7605 * lra-int.h (lra_subreg_reload_pseudos): New.
7606 * lra.c: Add undoing optional reloads to the block diagram.
7607 (lra_subreg_reload_pseudos): New.
7608 (lra_optional_reload_pseudos): Change comments.
7609 (lra): Init and clear lra_subreg_reload_pseudos. Clear
7610 lra_optional_reload_pseudos after undo transformations.
7611 * lra-assigns.c (pseudo_prefix_title): New.
7612 (lra_setup_reg_renumber): Use it.
7613 (spill_for): Ditto. Check subreg reload pseudos too.
7614 (assign_by_spills): Consider subreg reload pseudos too.
7615 * lra-constraints.c (simplify_operand_subreg): Use
7616 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
7617 (curr_insn_transform): Recognize and do optional reloads.
7618 (undo_optional_reloads): New.
7619 (lra_undo_inheritance): Call undo_optional_reloads.
7621 2013-07-05 Thomas Quinot <quinot@adacore.com>
7623 * tree-complex.c (expand_complex_operations_1): Fix typo.
7625 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
7627 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
7628 (tune_params): New member 'const vec_costs'.
7629 * config/aarch64/aarch64.c (generic_vector_cost): New.
7630 (generic_tunings): New member 'generic_vector_cost'.
7631 (aarch64_builtin_vectorization_cost): New.
7632 (aarch64_add_stmt_cost): New.
7633 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7634 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7636 2013-07-03 Jakub Jelinek <jakub@redhat.com>
7639 * config/i386/predicates.md (vsib_address_operand): Disallow
7640 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7642 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7645 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7646 expand_builtin_setjmp_receiver.
7647 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7648 with NULL for the label parameter.
7649 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7650 the frame-pointer. Adjust comments.
7651 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7652 only if LABEL is non-NULL.
7654 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7656 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7657 (ARM_ABI_AAPCS64): Ditto.
7659 (ARM_DEFAULT_ABI): Ditto.
7661 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7663 * config/aarch64/aarch64-builtins.c
7664 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7665 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7667 * config/aarch64/aarch64-simd.md
7668 (aarch64_ld1<VALL:mode>): New.
7669 (aarch64_st1<VALL:mode>): Likewise.
7670 * config/aarch64/arm_neon.h
7671 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7673 2013-07-02 Sriraman Tallam <tmsriram@google.com>
7675 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7677 (ix86_option_override_internal):Turn on all -mavx target flags by
7678 default as they are dependent on AVX anyway.
7680 2013-07-02 Cary Coutant <ccoutant@google.com>
7682 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7684 (loc_checksum_ordered): Likewise.
7685 (hash_loc_operands): Remove inline keyword.
7687 2013-07-02 Jakub Jelinek <jakub@redhat.com>
7689 PR tree-optimization/57741
7690 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7691 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7692 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7693 Allow REAL_CST step_exprs if flag_associative_math.
7694 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7696 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7698 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
7700 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7702 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7704 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7706 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7708 (iorsi3_insn): Likewise.
7709 (arm_xorsi3): Likewise.
7711 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7713 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7715 (attribute "type"): ... here, and prefix with "wmmx_".
7716 (attribute "core_cycles"): Update for attribute changes.
7717 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7718 (tbcstv4hi): Likewise.
7719 (tbcstv2si): Likewise.
7720 (iwmmxt_iordi3): Likewise.
7721 (iwmmxt_xordi3): Likewise.
7722 (iwmmxt_anddi3): Likewise.
7723 (iwmmxt_nanddi3): Likewise.
7724 (iwmmxt_arm_movdi): Likewise.
7725 (iwmmxt_movsi_insn): Likewise.
7726 (mov<mode>_internal): Likewise.
7727 (and<mode>3_iwmmxt): Likewise.
7728 (ior<mode>3_iwmmxt): Likewise.
7729 (xor<mode>3_iwmmxt): Likewise.
7730 (add<mode>3_iwmmxt): Likewise.
7731 (ssaddv8qi3): Likewise.
7732 (ssaddv4hi3): Likewise.
7733 (ssaddv2si3): Likewise.
7734 (usaddv8qi3): Likewise.
7735 (usaddv4hi3): Likewise.
7736 (usaddv2si3): Likewise.
7737 (sub<mode>3_iwmmxt): Likewise.
7738 (sssubv8qi3): Likewise.
7739 (sssubv4hi3): Likewise.
7740 (sssubv2si3): Likewise.
7741 (ussubv8qi3): Likewise.
7742 (ussubv4hi3): Likewise.
7743 (ussubv2si3): Likewise.
7744 (mulv4hi3_iwmmxt): Likewise.
7745 (smulv4hi3_highpart): Likewise.
7746 (umulv4hi3_highpart): Likewise.
7747 (iwmmxt_wmacs): Likewise.
7748 (iwmmxt_wmacsz): Likewise.
7749 (iwmmxt_wmacu): Likewise.
7750 (iwmmxt_wmacuz): Likewise.
7751 (iwmmxt_clrdi): Likewise.
7752 (iwmmxt_clrv8qi): Likewise.
7753 (iwmmxt_clr4hi): Likewise.
7754 (iwmmxt_clr2si): Likewise.
7755 (iwmmxt_uavgrndv8qi3): Likewise.
7756 (iwmmxt_uavgrndv4hi3): Likewise.
7757 (iwmmxt_uavgv8qi3): Likewise.
7758 (iwmmxt_uavgv4hi3): Likewise.
7759 (iwmmxt_tinsrb): Likewise.
7760 (iwmmxt_tinsrh): Likewise.
7761 (iwmmxt_tinsrw): Likewise.
7762 (iwmmxt_textrmub): Likewise.
7763 (iwmmxt_textrmsb): Likewise.
7764 (iwmmxt_textrmuh): Likewise.
7765 (iwmmxt_textrmsh): Likewise.
7766 (iwmmxt_textrmw): Likewise.
7767 (iwmxxt_wshufh): Likewise.
7768 (eqv8qi3): Likewise.
7769 (eqv4hi3): Likewise.
7770 (eqv2si3): Likewise.
7771 (gtuv8qi3): Likewise.
7772 (gtuv4hi3): Likewise.
7773 (gtuv2si3): Likewise.
7774 (gtv8qi3): Likewise.
7775 (gtv4hi3): Likewise.
7776 (gtv2si3): Likewise.
7777 (smax<mode>3_iwmmxt): Likewise.
7778 (umax<mode>3_iwmmxt): Likewise.
7779 (smin<mode>3_iwmmxt): Likewise.
7780 (umin<mode>3_iwmmxt): Likewise.
7781 (iwmmxt_wpackhss): Likewise.
7782 (iwmmxt_wpackwss): Likewise.
7783 (iwmmxt_wpackdss): Likewise.
7784 (iwmmxt_wpackhus): Likewise.
7785 (iwmmxt_wpackwus): Likewise.
7786 (iwmmxt_wpackdus): Likewise.
7787 (iwmmxt_wunpckihb): Likewise.
7788 (iwmmxt_wunpckihh): Likewise.
7789 (iwmmxt_wunpckihw): Likewise.
7790 (iwmmxt_wunpckilb): Likewise.
7791 (iwmmxt_wunpckilh): Likewise.
7792 (iwmmxt_wunpckilw): Likewise.
7793 (iwmmxt_wunpckehub): Likewise.
7794 (iwmmxt_wunpckehuh): Likewise.
7795 (iwmmxt_wunpckehuw): Likewise.
7796 (iwmmxt_wunpckehsb): Likewise.
7797 (iwmmxt_wunpckehsh): Likewise.
7798 (iwmmxt_wunpckehsw): Likewise.
7799 (iwmmxt_wunpckelub): Likewise.
7800 (iwmmxt_wunpckeluh): Likewise.
7801 (iwmmxt_wunpckeluw): Likewise.
7802 (iwmmxt_wunpckelsb): Likewise.
7803 (iwmmxt_wunpckelsh): Likewise.
7804 (iwmmxt_wunpckelsw): Likewise.
7805 (ror<mode>3): Likewise.
7806 (ashr<mode>3_iwmmxt): Likewise.
7807 (lshr<mode>3_iwmmxt): Likewise.
7808 (ashl<mode>3_iwmmxt): Likewise.
7809 (ror<mode>3_di): Likewise.
7810 (ashr<mode>3_di): Likewise.
7811 (lshr<mode>3_di): Likewise.
7812 (ashl<mode>3_di): Likewise.
7813 (iwmmxt_wmadds): Likewise.
7814 (iwmmxt_wmaddu): Likewise.
7815 (iwmmxt_tmia): Likewise.
7816 (iwmmxt_tmiaph): Likewise.
7817 (iwmmxt_tmiabb): Likewise.
7818 (iwmmxt_tmiatb): Likewise.
7819 (iwmmxt_tmiabt): Likewise.
7820 (iwmmxt_tmiatt): Likewise.
7821 (iwmmxt_tmovmskb): Likewise.
7822 (iwmmxt_tmovmskh): Likewise.
7823 (iwmmxt_tmovmskw): Likewise.
7824 (iwmmxt_waccb): Likewise.
7825 (iwmmxt_wacch): Likewise.
7826 (iwmmxt_waccw): Likewise.
7827 (iwmmxt_waligni): Likewise.
7828 (iwmmxt_walignr): Likewise.
7829 (iwmmxt_walignr0): Likewise.
7830 (iwmmxt_walignr1): Likewise.
7831 (iwmmxt_walignr2): Likewise.
7832 (iwmmxt_walignr3): Likewise.
7833 (iwmmxt_wsadb): Likewise.
7834 (iwmmxt_wsadh): Likewise.
7835 (iwmmxt_wsadbz): Likewise.
7836 (iwmmxt_wsadhz): Likewise.
7837 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
7838 (iwmmxt_wabsdiffb): Likewise.
7839 (iwmmxt_wabsdiffh): Likewise.
7840 (iwmmxt_wabsdiffw): Likewise.
7841 (iwmmxt_waddsubhx): Likewise
7842 (iwmmxt_wsubaddhx): Likewise.
7843 (addc<mode>3): Likewise.
7844 (iwmmxt_avg4): Likewise.
7845 (iwmmxt_avg4r): Likewise.
7846 (iwmmxt_wmaddsx): Likewise.
7847 (iwmmxt_wmaddux): Likewise.
7848 (iwmmxt_wmaddsn): Likewise.
7849 (iwmmxt_wmaddun): Likewise.
7850 (iwmmxt_wmulwsm): Likewise.
7851 (iwmmxt_wmulwum): Likewise.
7852 (iwmmxt_wmulsmr): Likewise.
7853 (iwmmxt_wmulumr): Likewise.
7854 (iwmmxt_wmulwsmr): Likewise.
7855 (iwmmxt_wmulwumr): Likewise.
7856 (iwmmxt_wmulwl): Likewise.
7857 (iwmmxt_wqmulm): Likewise.
7858 (iwmmxt_wqmulwm): Likewise.
7859 (iwmmxt_wqmulmr): Likewise.
7860 (iwmmxt_wqmulwmr): Likewise.
7861 (iwmmxt_waddbhusm): Likewise.
7862 (iwmmxt_waddbhusl): Likewise.
7863 (iwmmxt_wqmiabb): Likewise.
7864 (iwmmxt_wqmiabt): Likewise.
7865 (iwmmxt_wqmiatb): Likewise.
7866 (iwmmxt_wqmiatt): Likewise.
7867 (iwmmxt_wqmiabbn): Likewise.
7868 (iwmmxt_wqmiabtn): Likewise.
7869 (iwmmxt_wqmiatbn): Likewise.
7870 (iwmmxt_wqmiattn): Likewise.
7871 (iwmmxt_wmiabb): Likewise.
7872 (iwmmxt_wmiabt): Likewise.
7873 (iwmmxt_wmiatb): Likewise.
7874 (iwmmxt_wmiatt): Likewise.
7875 (iwmmxt_wmiabbn): Likewise.
7876 (iwmmxt_wmiabtn): Likewise.
7877 (iwmmxt_wmiatbn): Likewise.
7878 (iwmmxt_wmiattn): Likewise.
7879 (iwmmxt_wmiawbb): Likewise.
7880 (iwmmxt_wmiawbt): Likewise.
7881 (iwmmxt_wmiawtb): Likewise.
7882 (iwmmxt_wmiawtt): Likewise.
7883 (iwmmxt_wmiawbbn): Likewise.
7884 (iwmmxt_wmiawbtn): Likewise.
7885 (iwmmxt_wmiawtbn): Likewise.
7886 (iwmmxt_wmiawttn): Likewise.
7887 (iwmmxt_wmerge): Likewise.
7888 (iwmmxt_tandc<mode>3): Likewise.
7889 (iwmmxt_torc<mode>3): Likewise.
7890 (iwmmxt_torvsc<mode>3): Likewise.
7891 (iwmmxt_textrc<mode>3): Likewise.
7892 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
7893 (wmmxt_pack): Likewise.
7894 (wmmxt_mult_c1): Likewise.
7895 (wmmxt_mult_c2): Likewise.
7896 (wmmxt_alu_c1): Likewise.
7897 (wmmxt_alu_c2): Likewise.
7898 (wmmxt_alu_c3): Likewise.
7899 (wmmxt_transfer_c1): Likewise.
7900 (wmmxt_transfer_c2): Likewise.
7901 (wmmxt_transfer_c3): Likewise.
7902 (marvell_f_iwmmxt_wstr): Likewise.
7903 (marvell_f_iwmmxt_wldr): Likewise.
7905 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
7907 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
7909 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7912 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7913 * lra-constraints.c (need_for_split_p): Check call used hard regs
7914 living through calls.
7916 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
7917 call used regs for call insn.
7919 2013-06-28 Jakub Jelinek <jakub@redhat.com>
7922 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
7923 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
7925 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7927 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
7929 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7931 * lra-constraints.c (need_for_split_p): Check call used hard regs
7932 living through calls.
7934 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
7937 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
7938 to tie with any other modes. Eliminate Altivec vector mode tests,
7939 since these are a subset of ALTIVEC or VSX vector modes. Simplify
7940 code, to return 0 if testing MODE2 for a condition, if we've
7941 already tested MODE1 for the same condition.
7943 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7945 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
7948 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7950 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7951 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
7953 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7955 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
7957 (aarch64_symbolic_constant_p): Remove.
7958 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
7959 static. Fix line length and white space.
7960 (aarch64_symbolic_constant_p): Remove.
7961 * config/aarch64/predicates.md (aarch64_valid_symref):
7962 Use aarch64_classify_symbol_expression.
7964 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7966 * config/arm/constraints.md (Ts): New constraint.
7967 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
7969 (compare_scc): Use "Ts" constraint for operand 0.
7970 (ior_scc_scc): Likewise.
7971 (and_scc_scc): Likewise.
7972 (and_scc_scc_nodom): Likewise.
7973 (ior_scc_scc_cmp): Likewise for operand 7.
7974 (and_scc_scc_cmp): Likewise.
7975 * config/arm/thumb2.md (thumb2_movsi_insn):
7976 Add alternatives for 16-bit encodings.
7977 (thumb2_movhi_insn): Likewise.
7978 (thumb2_movsicc_insn): Likewise.
7979 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
7980 (thumb2_negscc): Use "Ts" constraint.
7981 Move mvn instruction outside cond_exec block.
7982 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
7983 for 16-bit encodings.
7985 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7987 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
7989 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
7990 (mulsi3subsi): Likewise.
7991 (mulsidi3adddi): Likewise.
7992 (mulsidi3_v6): Likewise.
7993 (umulsidi3_v6): Likewise.
7994 (umulsidi3adddi_v6): Likewise.
7995 (smulsi3_highpart_v6): Likewise.
7996 (umulsi3_highpart_v6): Likewise.
7997 (mulhisi3tb): Likewise.
7998 (mulhisi3bt): Likewise.
7999 (mulhisi3tt): Likewise.
8000 (maddhisi4): Likewise.
8001 (maddhisi4tb): Likewise.
8002 (maddhisi4tt): Likewise.
8003 (maddhidi4): Likewise.
8004 (maddhidi4tb): Likewise.
8005 (maddhidi4tt): Likewise.
8006 (zeroextractsi_compare0_scratch): Likewise.
8007 (insv_zero): Likewise.
8008 (insv_t2): Likewise.
8009 (anddi_notzesidi_di): Likewise.
8010 (anddi_notsesidi_di): Likewise.
8011 (andsi_notsi_si): Likewise.
8012 (iordi_zesidi_di): Likewise.
8013 (xordi_zesidi_di): Likewise.
8014 (andsi_iorsi3_notsi): Likewise.
8016 (smax_m1): Likewise.
8018 (not_shiftsi): Likewise.
8019 (unaligned_loadsi): Likewise.
8020 (unaligned_loadhis): Likewise.
8021 (unaligned_loadhiu): Likewise.
8022 (unaligned_storesi): Likewise.
8023 (unaligned_storehi): Likewise.
8024 (extv_reg): Likewise.
8025 (extzv_t2): Likewise.
8027 (udivsi3): Likewise.
8028 (arm_zero_extendhisi2addsi): Likewise.
8029 (arm_zero_extendqisi2addsi): Likewise.
8030 (compareqi_eq0): Likewise.
8031 (arm_extendhisi2_v6): Likewise.
8032 (arm_extendqisi2addsi): Likewise.
8033 (arm_movt): Likewise.
8034 (thumb2_ldrd): Likewise.
8035 (thumb2_ldrd_base): Likewise.
8036 (thumb2_ldrd_base_neg): Likewise.
8037 (thumb2_strd): Likewise.
8038 (thumb2_strd_base): Likewise.
8039 (thumb2_strd_base_neg): Likewise.
8040 (arm_negsi2): Add alternative for 16-bit encoding.
8041 (arm_one_cmplsi2): Likewise.
8043 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8045 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
8046 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
8047 (movdfcc): Likewise.
8048 * config/arm/vfp.md (*thumb2_movsf_vfp):
8049 Disable predication for arm_restrict_it.
8050 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
8051 (*thumb2_movdfcc_vfp): Likewise.
8052 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
8053 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
8054 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
8055 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
8056 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
8057 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
8058 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
8059 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
8060 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
8061 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
8062 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
8063 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
8064 Disable predication for arm_restrict_it.
8066 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
8068 * config/i386/bmiintrin.h (_bextr_u32): New.
8069 (_bextr_u64): Ditto.
8071 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
8073 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
8074 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
8075 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
8076 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
8077 * config/mips/n32-elf.h: ...this new file.
8079 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
8082 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
8083 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
8085 2013-06-27 Catherine Moore <clm@codesourcery.com>
8087 * config/mips/mips-tables.opt: Regenerate.
8088 * config/mips/mips-cpus.def: Add m14ke and m14kec.
8089 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8090 * doc/invoke.texi: Add -m14kc.
8092 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8095 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
8096 constraints of operand 1 and 2.
8099 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
8100 to match RTL canonicalization. Swap predicates and
8101 constraints of operand 1 and 2.
8103 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
8105 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
8106 Process OP_INOUT regs for splitting too.
8108 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8110 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
8111 decl before the loop, initialize to NULL.
8112 (vectorizable_load): Initialize ptr_incr to NULL.
8114 2013-06-27 Martin Jambor <mjambor@suse.cz>
8117 * ipa-ref.h (ipa_maybe_record_reference): Declare.
8118 * ipa-ref.c (ipa_maybe_record_reference): New function.
8119 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
8120 * ipa-cp.c (create_specialized_node): Record potential references from
8122 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
8124 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8126 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
8127 parameter 'mode' of type 'enum machine_mode mode'; change to pass
8128 'mode' to force_reg.
8129 (aarch64_add_offset): Update calls to aarch64_force_temporary.
8130 (aarch64_expand_mov_immediate): Likewise.
8132 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8134 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
8135 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
8137 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8139 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
8140 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8141 (struct machine_function): Add tbegin_p.
8142 (s390_canonicalize_comparison): Fold CC mode compares to
8143 conditional jump if possible.
8144 (s390_emit_jump): Return the emitted jump.
8145 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
8146 Handle CCRAWmode compares.
8147 (s390_option_override): Default to -mhtm if available.
8148 (s390_reg_clobbered_rtx): Handle floating point regs as well.
8149 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
8150 FPRs instead of df_regs_ever_live_p.
8151 (s390_optimize_nonescaping_tx): New function.
8152 (s390_init_frame_layout): Extend clobbered_regs array to cover
8154 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
8155 (s390_expand_tbegin): New function.
8156 (enum s390_builtin): New enum definition.
8157 (code_for_builtin): New array definition.
8158 (s390_init_builtins): New function.
8159 (s390_expand_builtin): New function.
8160 (TARGET_INIT_BUILTINS): Define.
8161 (TARGET_EXPAND_BUILTIN): Define.
8162 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
8163 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
8164 (s390_alc_comparison): Likewise.
8165 * config/s390/s390-modes.def: Add CCRAWmode.
8166 * config/s390/s390.h (processor_flags): Add PF_TX.
8167 (TARGET_CPU_HTM): Define macro.
8168 (TARGET_HTM): Define macro.
8169 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
8170 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
8171 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8172 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
8173 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
8174 (TBEGIN_MASK, TBEGINC_MASK): New constants.
8175 ("*cc_to_int"): Move up.
8176 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
8177 constants other than 0.
8178 ("*ccraw_to_int"): New insn and splitter definition.
8179 ("tbegin", "tbegin_nofloat", "tbegin_retry")
8180 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
8181 ("tx_assist"): New expander.
8182 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
8183 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
8184 * config/s390/s390.opt: Add -mhtm option.
8185 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
8186 * config/s390/htmxlintrin.h: New file.
8187 * config/s390/htmintrin.h: New file.
8188 * config/s390/s390intrin.h: New file.
8189 * doc/extend.texi: Document htm builtins.
8190 * config.gcc: Add the new header files to extra_headers.
8192 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
8194 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
8195 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
8197 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
8198 Pat Haugen <pthaugen@us.ibm.com>
8199 Peter Bergner <bergner@vnet.ibm.com>
8201 * config/rs6000/power8.md: New.
8202 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
8203 setting for power8 entry.
8204 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
8205 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
8206 test for Power4/Power5 only.
8207 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
8209 (force_new_group): Adjust comment.
8210 * config/rs6000/rs6000.md: Include power8.md.
8212 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
8214 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
8215 * config/arm/arm-protos.h (arm_max_conditional_execute): New
8217 (tune_params): Update comment.
8218 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
8219 (arm_max_conditional_execute): New function.
8220 (thumb2_final_prescan_insn): Use max_insn_skipped and
8221 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
8223 2013-06-25 Jakub Jelinek <jakub@redhat.com>
8225 PR tree-optimization/57705
8226 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
8227 SSA_NAME step, provided that it is not defined inside the loop.
8228 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
8229 (get_initial_def_for_induction): Handle SSA_NAME IV step.
8231 2013-06-25 Martin Jambor <mjambor@suse.cz>
8234 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
8235 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
8237 (ipa_note_param_call): Initialize member_ptr flag.
8238 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
8239 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
8240 (ipa_write_indirect_edge_info): Stream member_ptr flag.
8241 (ipa_read_indirect_edge_info): Likewise.
8243 2013-06-25 Richard Biener <rguenther@suse.de>
8246 * passes.c (init_optimization_passes): Move pass_fold_builtins
8247 and pass_dce earlier with -Og.
8249 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
8251 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
8252 <BIT_FIELD_REF>: Remove trailing TAB.
8253 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
8256 2013-06-24 Martin Jambor <mjambor@suse.cz>
8258 PR tree-optimization/57358
8259 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
8260 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
8261 (ipa_analyze_params_uses): Generate pessimistic info when true.
8263 2013-06-24 Martin Jambor <mjambor@suse.cz>
8265 PR tree-optimization/57539
8266 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
8267 global.inlined_to of the new node to it. All callers changed.
8268 * ipa-inline-transform.c (clone_inlined_nodes): New variable
8269 inlining_into, pass it to cgraph_clone_node.
8270 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
8271 ipa_free_edge_args_substructures.
8272 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
8273 rdesc linked list. Do not assert rdesc edges have inlined caller.
8274 Assert we have found an rdesc in the rdesc list.
8276 2013-06-24 Richard Biener <rguenther@suse.de>
8278 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
8279 (pointer_set_lookup): Declare.
8280 (class pointer_map): New template class implementing a
8281 generic pointer to T map.
8282 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
8283 pointer_map<T>::contains, pointer_map<T>::insert,
8284 pointer_map<T>::traverse): New functions.
8285 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
8286 (pointer_set_lookup): New function.
8287 (pointer_set_contains): Use pointer_set_lookup.
8288 (pointer_set_insert): Likewise.
8289 (insert_aux): Remove.
8290 (struct pointer_map_t): Embed a pointer_set_t.
8291 (pointer_map_create): Adjust.
8292 (pointer_map_destroy): Likewise.
8293 (pointer_map_contains): Likewise.
8294 (pointer_map_insert): Likewise.
8295 (pointer_map_traverse): Likewise.
8296 * tree-streamer.h (struct streamer_tree_cache_d): Use a
8297 pointer_map<unsigned> instead of a pointer_map_t.
8298 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
8299 (streamer_tree_cache_lookup): Likewise.
8300 (streamer_tree_cache_create): Likewise.
8301 (streamer_tree_cache_delete): Likewise.
8302 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
8303 pointer_map<unsigned> instead of a pointer_map_t.
8304 (lto_init_tree_ref_encoder): Adjust.
8305 (lto_destroy_tree_ref_encoder): Likewise.
8306 * lto-section-out.c (lto_output_decl_index): Likewise.
8307 (lto_record_function_out_decl_state): Likewise.
8308 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
8310 2013-06-24 Richard Biener <rguenther@suse.de>
8312 PR tree-optimization/57488
8313 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
8315 2013-06-24 Alan Modra <amodra@gmail.com>
8317 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
8318 (gen_easy_altivec_constant): Likewise.
8319 * config/rs6000/predicates.md (easy_vector_constant_add_self,
8320 easy_vector_constant_msb): Likewise.
8322 2013-06-23 Jakub Jelinek <jakub@redhat.com>
8325 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
8326 add missing return true.
8328 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
8331 * config/sh/predicates.md (general_extend_operand): Invoke
8332 general_movsrc_operand for memory operands.
8333 (general_movsrc_operand): Allow reg+reg addressing, do not use
8334 general_operand for memory operands.
8336 2013-06-23 Sriraman Tallam <tmsriram@google.com>
8338 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
8339 when current target options does not apply.
8340 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
8341 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
8342 * config/i386/bmiintrin.h: Pass appropriate target
8343 attributes to header.
8344 * config/i386/mmintrin.h: Ditto.
8345 * config/i386/nmmintrin.h: Ditto.
8346 * config/i386/avx2intrin.h: Ditto.
8347 * config/i386/fxsrintrin.h: Ditto.
8348 * config/i386/tbmintrin.h: Ditto.
8349 * config/i386/xsaveintrin.h: Ditto.
8350 * config/i386/f16cintrin.h: Ditto.
8351 * config/i386/xtestintrin.h: Ditto.
8352 * config/i386/xsaveoptintrin.h: Ditto.
8353 * config/i386/bmi2intrin.h: Ditto.
8354 * config/i386/lzcntintrin.h: Ditto.
8355 * config/i386/smmintrin.h: Ditto.
8356 * config/i386/wmmintrin.h: Ditto.
8357 * config/i386/x86intrin.h: Remove all header include guards.
8358 * config/i386/prfchwintrin.h: Ditto.
8359 * config/i386/pmmintrin.h: Ditto.
8360 * config/i386/tmmintrin.h: Ditto.
8361 * config/i386/xmmintrin.h: Ditto.
8362 * config/i386/popcntintrin.h: Ditto.
8363 * config/i386/rdseedintrin.h: Ditto.
8364 * config/i386/ammintrin.h: Ditto.
8365 * config/i386/emmintrin.h: Ditto.
8366 * config/i386/immintrin.h: Remove all header include guards.
8367 * config/i386/fma4intrin.h: Ditto.
8368 * config/i386/lwpintrin.h: Ditto.
8369 * config/i386/xopintrin.h: Ditto.
8370 * config/i386/ia32intrin.h: Ditto.
8371 * config/i386/avxintrin.h: Ditto.
8372 * config/i386/rtmintrin.h: Ditto.
8373 * config/i386/fmaintrin.h: Ditto.
8374 * config/i386/mm3dnow.h: Ditto.
8376 2013-06-22 Sriraman Tallam <tmsriram@google.com>
8378 * common/config/i386/i386-common.c: Handle LZCNT.
8380 2013-06-22 Andi Kleen <ak@linux.intel.com>
8382 * doc/extend.texi: Use __atomic_store_n instead of
8383 __atomic_store in HLE example.
8385 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
8387 * config/sh/sh.c: Remove <cstdlib> workaround.
8389 2013-06-21 Andi Kleen <ak@linux.intel.com>
8391 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
8393 2013-06-21 Andi Kleen <ak@linux.intel.com>
8395 * doc/extend.texi: Document that __atomic_clear and
8396 __atomic_test_and_set should only be used with bool.
8398 2013-06-20 Jan Hubicka <jh@suse.cz>
8400 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
8402 * tree.c (decls_same_for_odr): New function.
8403 (same_for_edr): New function.
8404 (types_same_for_odr): New function.
8405 (get_binfo_at_offset): Use it.
8406 * tree.h (types_same_for_odr): Declare.
8408 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
8409 Jason Merrill <jason@redhat.com>
8411 * system.h: Include <cstdlib> as well as <stdlib.h>.
8413 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
8416 * config/i386/i386.c (construct_container): Report error if
8417 long double is used with disabled x87 float returns.
8419 2013-06-20 Jan Hubicka <jh@suse.cz>
8421 * lto-cgraph.c (input_symtab): Do not set cgraph state.
8423 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
8425 PR rtl-optimization/57425
8426 PR rtl-optimization/57569
8427 * alias.c (write_dependence_p): Remove parameters mem_mode and
8428 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
8429 Changed all callers.
8430 (canon_anti_dependence): Get comments and semantics in sync.
8431 Add parameter mem_canonicalized. Changed all callers.
8432 * rtl.h (canon_anti_dependence): Update prototype.
8434 2013-06-20 Richard Biener <rguenther@suse.de>
8436 * data-streamer-in.c (streamer_read_uhwi): Optimize single
8437 byte case, inline streamer_read_uchar and defer section
8440 2013-06-20 Richard Biener <rguenther@suse.de>
8442 PR tree-optimization/57584
8443 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
8444 SSA names into the expanded expression that take part in
8445 abnormal coalescing.
8447 2013-06-19 Sharad Singhai <singhai@google.com>
8449 * gcov.c (print_usage): Handle new option.
8450 (process_args): Ditto.
8451 (get_gcov_intermediate_filename): New function.
8452 (output_intermediate_file): New function.
8453 (output_gcov_file): New function
8454 (generate_results): Handle new option.
8455 (release_function): Relase demangled name.
8456 (read_graph_file): Handle demangled name.
8457 (output_lines): Ditto.
8458 * doc/gcov.texi: Document gcov intermediate format.
8460 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
8463 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
8464 (lra_emit_add): Use the functions. Add comment about Y as an
8467 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
8470 * collect2.c (collect_atexit): New.
8471 (collect_exit): Delete.
8472 (main): Register collect_atexit with atexit.
8473 (collect_wait): Change collect_exit to exit.
8475 * collect2.h (collect_exit): Delete.
8476 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
8478 2013-06-19 Wei Mi <wmi@google.com>
8480 PR rtl-optimization/57518
8481 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8482 if regno is used in paradoxical subreg.
8483 (update_equiv_regs): Check pdx_subregs[regno] before
8484 set a reg to be equivalent with a mem.
8486 2013-06-19 Matthias Klose <doko@ubuntu.com>
8489 * file-find.h (find_a_file): Add a mode parameter.
8490 * file-find.c (find_a_file): Likewise.
8491 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8492 with X_OK for the executables.
8493 * collect2.c (main): Call find_a_file with X_OK.
8495 2013-06-19 Steve Ellcey <sellcey@mips.com>
8498 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8499 Use NEXT_INSN instead of next_real_insn.
8501 2013-06-19 Jan Hubicka <jh@suse.cz>
8503 * cgraph.h (const_value_known_p): Replace by ...
8504 (ctor_for_folding): .. this one.
8505 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8506 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8507 * expr.c (expand_expr_real_1): Likewise.
8508 (string_constant): Likewise.
8509 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8510 * ipa.c (process_references): Likewise.
8511 (symtab_remove_unreachable_nodes): Likewise.
8512 * ipa-inline-analysis.c (param_change_prob): Likewise.
8513 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8514 (get_base_constructor): Likwise.
8515 * varpool.c (varpool_remove_node): Likewise.
8516 (varpool_remove_initializer): LIkewise.
8517 (dump_varpool_node): LIkwise.
8518 (const_value_known_p): Rewrite to ...
8519 (ctor_for_folding): ... this one.
8521 2013-06-19 Jakub Jelinek <jakub@redhat.com>
8524 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8525 PERSONALITY in $PATH derived prefixes.
8527 2013-06-19 Jeff Law <law@redhat.com>
8529 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8532 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8533 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8536 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8538 * config/arm/vfp.md: Move VFP instruction classification documentation
8540 * config/arm/arm.md: ... here. Update instruction classification
8543 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
8545 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8546 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8547 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8549 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8552 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8553 with NEXT_INSN. Use tablejump_p to check for jump table data
8556 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8559 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8560 that now in C++ the value is correct per the C++ standards.
8562 2013-06-19 Richard Biener <rguenther@suse.de>
8564 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8567 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8570 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8573 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8574 with next_active_insn.
8576 2013-06-18 Sriraman Tallam <tmsriram@google.com>
8578 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
8579 functions are inlined during failures to flag an error.
8580 * tree-inline.c (expand_call_inline): Allow the error to be flagged
8581 in early inline pass.
8583 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
8585 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
8588 2013-06-18 Julian Brown <julian@codesourcery.com>
8590 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
8591 Permit virtual register pre-reload if !strict.
8592 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
8594 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
8596 * config/arm/neon.md (movmisalign<mode>): Use
8597 neon_perm_struct_or_reg_operand instead of
8598 neon_struct_or_register_operand.
8599 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
8600 neon_permissive_struct_operand instead of neon_struct_operand.
8601 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
8602 neon_vector_mem_operand.
8603 * config/arm/predicates.md (neon_struct_operand): Adjust call to
8604 neon_vector_mem_operand.
8605 (neon_permissive_struct_operand): New.
8606 (neon_struct_or_register_operand): Rename to...
8607 (neon_perm_struct_or_reg_operand): This. Adjust call to
8608 neon_vector_mem_operand.
8610 2013-06-18 Richard Biener <rguenther@suse.de>
8612 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
8613 * lto-streamer.h: Include pointer-set.h.
8614 (struct lto_decl_slot): Remove.
8615 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
8616 Remove next_index entry.
8617 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8618 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8619 (lto_init_tree_ref_encoder): Adjust.
8620 (lto_destroy_tree_ref_encoder): Likewise.
8621 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8622 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8623 (lto_output_decl_index): Adjust.
8624 (lto_new_out_decl_state): Likewise.
8625 (lto_record_function_out_decl_state): Likewise.
8626 * lto-streamer-out.c (copy_function): Likewise.
8628 2013-06-18 Richard Biener <rguenther@suse.de>
8630 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
8631 * cgraphunit.c: Include cfgloop.h.
8632 (init_lowered_empty_function): Initialize the loop tree.
8633 (assemble_thunk): Insert new BBs into loops.
8635 2013-06-18 Richard Biener <rguenther@suse.de>
8637 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8638 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8639 the map from cache entry to cache index optional.
8640 (streamer_tree_cache_replace_tree): Adjust accordingly.
8641 (streamer_tree_cache_append): Likewise.
8642 (streamer_tree_cache_delete): Likewise.
8643 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8644 streamer cache map from cache entry to cache index.
8645 * lto-streamer-out.c (create_output_block): Adjust.
8647 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8649 * config/arm/arm.md (attribute "insn"): Move multiplication and
8650 division attributes to...
8651 (attribute "type"): ... here. Remove mult.
8652 (attribute "mul32"): New attribute.
8653 (attribute "mul64"): Add umaal.
8654 (*arm_mulsi3): Update attributes.
8655 (*arm_mulsi3_v6): Likewise.
8656 (*thumb_mulsi3): Likewise.
8657 (*thumb_mulsi3_v6): Likewise.
8658 (*mulsi3_compare0): Likewise.
8659 (*mulsi3_compare0_v6): Likewise.
8660 (*mulsi_compare0_scratch): Likewise.
8661 (*mulsi_compare0_scratch_v6): Likewise.
8662 (*mulsi3addsi): Likewise.
8663 (*mulsi3addsi_v6): Likewise.
8664 (*mulsi3addsi_compare0): Likewise.
8665 (*mulsi3addsi_compare0_v6): Likewise.
8666 (*mulsi3addsi_compare0_scratch): Likewise.
8667 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8668 (*mulsi3subsi): Likewise.
8669 (*mulsidi3adddi): Likewise.
8670 (*mulsi3addsi_v6): Likewise.
8671 (*mulsidi3adddi_v6): Likewise.
8672 (*mulsidi3_nov6): Likewise.
8673 (*mulsidi3_v6): Likewise.
8674 (*umulsidi3_nov6): Likewise.
8675 (*umulsidi3_v6): Likewise.
8676 (*umulsidi3adddi): Likewise.
8677 (*umulsidi3adddi_v6): Likewise.
8678 (*smulsi3_highpart_nov6): Likewise.
8679 (*smulsi3_highpart_v6): Likewise.
8680 (*umulsi3_highpart_nov6): Likewise.
8681 (*umulsi3_highpart_v6): Likewise.
8682 (mulhisi3): Likewise.
8683 (*mulhisi3tb): Likewise.
8684 (*mulhisi3bt): Likewise.
8685 (*mulhisi3tt): Likewise.
8686 (maddhisi4): Likewise.
8687 (*maddhisi4tb): Likewise.
8688 (*maddhisi4tt): Likewise.
8689 (maddhidi4): Likewise.
8690 (*maddhidi4tb): Likewise.
8691 (*maddhidi4tt): Likewise.
8693 (udivsi3): Likewise.
8694 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8695 (thumb2_mulsi_short_compare0): Likewise.
8696 (thumb2_mulsi_short_compare0_scratch): Likewise.
8697 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8698 (1020mult2): Likewise.
8699 (1020mult3): Likewise.
8700 (1020mult4): Likewise.
8701 (1020mult5): Likewise.
8702 (1020mult6): Likewise.
8703 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8705 (cortex_a15_mult64): Likewise.
8706 (cortex_a15_sdiv): Likewise.
8707 (cortex_a15_udiv): Likewise.
8708 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8714 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8715 (pj4_ir_div): Likewise.
8716 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8717 (11_mult2): Likewise.
8718 (11_mult3): Likewise.
8719 (11_mult4): Likewise.
8720 (11_mult5): Likewise.
8721 (11_mult6): Likewise.
8722 (11_mult7): Likewise.
8723 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8724 (cortex_a8_mla): Likewise.
8725 (cortex_a8_mull): Likewise.
8726 (cortex_a8_smulwy): Likewise.
8727 (cortex_a8_smlald): Likewise.
8728 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8729 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8730 (cortex_r4_mul_3): Likewise.
8731 (cortex_r4_mla_4): Likewise.
8732 (cortex_r4_mla_3): Likewise.
8733 (cortex_r4_smlald): Likewise.
8734 (cortex_r4_mull): Likewise.
8735 (cortex_r4_sdiv): Likewise.
8736 (cortex_r4_udiv): Likewise.
8737 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8738 (cortex_a7_idiv): Likewise.
8739 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8740 (9_mult2): Likewise.
8741 (9_mult3): Likewise.
8742 (9_mult4): Likewise.
8743 (9_mult5): Likewise.
8744 (9_mult6): Likewise.
8745 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8746 (cortex_a53_sdiv): Likewise.
8747 (cortex_a53_udiv): Likewise.
8748 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8749 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8750 (mp626_mult2): Likewise.
8751 (mp626_mult3): Likewise.
8752 (mp626_mult4): Likewise.
8753 * config/arm/fa526.md (526_mult1): Update attribute change.
8754 (526_mult2): Likewise.
8755 * config/arm/arm-generic.md (mult): Update attribute change.
8756 (mult_ldsched_strongarm): Likewise.
8757 (mult_ldsched): Likewise.
8758 (multi_cycle): Likewise.
8759 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8760 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8761 (606te_mult2): Likewise.
8762 (606te_mult3): Likewise.
8763 (606te_mult4): Likewise.
8764 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8765 (cortex_a9_mac16): Likewise.
8766 (cortex_a9_multiply): Likewise.
8767 (cortex_a9_mac): Likewise.
8768 (cortex_a9_multiply_long): Likewise.
8769 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8770 (626te_mult2): Likewise.
8771 (626te_mult3): Likewise.
8772 (626te_mult4): Likewise.
8774 2013-06-18 Richard Biener <rguenther@suse.de>
8777 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8779 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8782 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8783 with next_active_insn.
8785 2013-06-18 Alan Modra <amodra@gmail.com>
8787 * config/rs6000/rs6000.h (enum data_align): New.
8788 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8789 (DATA_ABI_ALIGNMENT): Define.
8790 (CONSTANT_ALIGNMENT): Correct comment.
8791 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8792 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8794 2013-06-17 David Malcolm <dmalcolm@redhat.com>
8796 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8797 ATTRIBUTE_UNUSED marking.
8799 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8801 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8802 alternative and update.
8803 (aarch64_dup_lanedi): Delete.
8804 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8805 * config/aarch64/aarch64-simd-builtins.def: Update.
8807 2013-06-17 Richard Biener <rguenther@suse.de>
8809 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8810 (lto_input_scc): Declare.
8811 (lto_input_tree_1): Likewise.
8812 (struct lto_stats_d): Add num_tree_bodies_output and
8813 num_pickle_refs_output.
8814 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8815 (lto_read_tree_1): Split out from ...
8816 (lto_read_tree): ... this.
8817 (lto_input_scc): New function.
8818 (lto_input_tree_1): Split out from ...
8819 (lto_input_tree): ... this. Handle LTO_tree_scc.
8820 (lto_data_in_create): Create the streamer cache without hashes.
8821 * lto-streamer-out.c (create_output_block): Create the streamer
8822 cache with hashes when not doing WPA.
8823 (lto_write_tree_1): Split out from ...
8824 (lto_write_tree): ... this.
8825 (get_symbol_initial_value): New function.
8826 (lto_output_tree_1): Split out from ...
8827 (lto_output_tree): ... this. Write trees as series of SCCs
8828 using a DFS walk via DFS_write_tree.
8829 (struct sccs, struct scc_entry): New types.
8830 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
8831 (DFS_write_tree_body): New function.
8832 (DFS_write_tree): Likewise.
8833 (hash_tree): Likewise.
8834 (scc_entry_compare): Likewise.
8835 (hash_scc): Likewise.
8836 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
8837 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
8838 TREE_CHAIN as regular reference.
8839 (streamer_read_integer_cst): Remove.
8840 (streamer_get_pickled_tree): Adjust.
8841 * tree-streamer-out.c (streamer_write_chain): Disable streaming
8842 of DECL_EXTERNALs in BLOCK_VARS for now.
8843 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
8845 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
8846 Add hash value argument and record that if hashes are recorded
8848 (streamer_tree_cache_insert_1): Adjust.
8849 (streamer_tree_cache_insert): Likewise.
8850 (streamer_tree_cache_insert_at): Rename to ...
8851 (streamer_tree_cache_replace_tree): ... this and adjust.
8852 (streamer_tree_cache_append): Adjust.
8853 (record_common_node): Likewise.
8854 (streamer_tree_cache_create): Add argument whether to
8855 record hash values together with trees.
8856 (streamer_tree_cache_delete): Adjust.
8857 * tree-streamer.h (struct streamer_tree_cache_d): Add
8859 (streamer_read_integer_cst): Remove.
8860 (streamer_tree_cache_insert): Adjust.
8861 (streamer_tree_cache_append): Likewise.
8862 (streamer_tree_cache_insert_at): Rename to ...
8863 (streamer_tree_cache_replace_tree): ... this and adjust.
8864 (streamer_tree_cache_create): Add argument whether to record hashes.
8865 (streamer_tree_cache_get): Rename to ...
8866 (streamer_tree_cache_get_tree): ... this.
8867 (streamer_tree_cache_get_hash): New function.
8868 * tree.c (cache_integer_cst): New function.
8869 * tree.h (cache_integer_cst): Declare.
8870 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
8871 * lto-symtab.c (lto_varpool_replace_node): Only release
8872 DECL_INITIAL of non-prevailing decls.
8873 * varpool.c (varpool_remove_initializer): Do not release
8874 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
8876 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8878 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
8879 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
8880 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
8881 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
8882 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
8883 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
8884 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
8885 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
8886 instead of TARGET_64BIT.
8887 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
8888 Require ISA_HAS_<D>DIV.
8890 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
8892 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
8893 (mips*-*-linux*): Move default with_llsc setting to where other
8895 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
8897 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
8898 Likewise. Remove default with_tune setting. Move default float
8899 setting to its own block. Handle with_llsc in the same block as above.
8901 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
8903 PR rtl-optimization/57425
8904 PR rtl-optimization/57569
8905 * alias.c (write_dependence_p): Add new parameters mem_mode,
8906 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
8907 Changed all callers.
8908 (canon_anti_dependence): New function.
8909 * cse.c (check_dependence): Use canon_anti_dependence.
8910 * cselib.c (cselib_invalidate_mem): Likewise.
8911 * rtl.h (canon_anti_dependence): Declare.
8913 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8915 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
8916 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
8917 ".set mips3" for 64-bit targets.
8919 2013-06-15 Dehao Chen <dehao@google.com>
8921 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
8922 * gimple-low.c (gimple_check_call_matching_types): Likewise.
8923 (gimple_check_call_args): Likewise.
8924 * value-prof.c (check_ic_target): Likewise.
8925 * ipa-inline.c (early_inliner): Likewise.
8926 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
8927 * cgraph.c (cgraph_create_edge_1): Likewise.
8928 (cgraph_make_edge_direct): Likewise.
8930 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
8933 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
8934 rs6000_output_move_128bit to handle emitting quad memory
8935 operations. Set attribute length to 8 bytes.
8937 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
8939 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
8941 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
8942 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
8943 (aarch64_<su>mlsl<mode>): Likewise.
8945 2013-06-14 Mike Stump <mikestump@comcast.net>
8947 * Makefile.in (TARGET_H): Add insn-codes.h.
8949 2013-06-14 Alan Modra <amodra@gmail.com>
8953 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
8954 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
8955 bitfield expansion when EXPAND_MEMORY.
8956 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
8958 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
8960 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
8961 test for clearing quad memory on 32-bit later.
8963 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8965 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
8966 (fold_negate_expr): Likewise.
8967 (fold_real_zero_addition_p): Handle vectors.
8968 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
8970 2013-06-14 Alan Modra <amodra@gmail.com>
8972 * varasm.c (force_const_mem): Revert 2013-06-07 change.
8974 2013-06-13 Jan Hubicka <jh@suse.cz>
8976 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
8977 Local comdats are not externally visible.
8978 * symtab.c (dump_symtab_base): Dump externally visible.
8979 (verify_symtab_base): Verify back links in the symtab hash.
8981 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8983 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
8984 CONVERT_EXPR as equal nodes.
8986 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8988 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8990 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8992 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
8993 Generalize to complex and vector.
8994 * tree.c (build_all_ones_cst): New function.
8995 * tree.h (build_all_ones_cst): Declare it.
8997 2013-06-13 Alan Modra <amodra@gmail.com>
8999 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
9000 * config/rs6000/rs6000.md (signbittf2): New insn.
9001 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
9002 (abstf2_internal, cmptf_internal2): Likewise.
9003 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
9005 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
9006 Pat Haugen <pthaugen@us.ibm.com>
9007 Peter Bergner <bergner@vnet.ibm.com>
9009 * config/rs6000/rs6000.c (emit_load_locked): Add support for
9010 power8 byte, half-word, and quad-word atomic instructions.
9011 (emit_store_conditional): Likewise.
9012 (rs6000_expand_atomic_compare_and_swap): Likewise.
9013 (rs6000_expand_atomic_op): Likewise.
9015 * config/rs6000/sync.md (larx): Add new modes for power8.
9017 (AINT): New mode iterator to include TImode as well as normal
9018 integer modes on power8.
9019 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
9020 that VSX registers are not considered. Use AINT mode iterator
9021 instead of INT1 to allow inclusion of quad word atomic operations
9023 (load_locked<mode>): Likewise.
9024 (store_conditional<mode>): Likewise.
9025 (atomic_compare_and_swap<mode>): Likewise.
9026 (atomic_exchange<mode>): Likewise.
9027 (atomic_nand<mode>): Likewise.
9028 (atomic_fetch_<fetchop_name><mode>): Likewise.
9029 (atomic_nand_fetch<mode>): Likewise.
9030 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
9032 (ATOMIC): On power8, add QImode, HImode modes.
9033 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
9034 modes that promote to SImode.
9035 (load_lockedti): Convert TImode arguments to PTImode, so that we
9036 get a guaranteed even/odd register pair.
9037 (load_lockedpti): Likewise.
9038 (store_conditionalti): Likewise.
9039 (store_conditionalpti): Likewise.
9041 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
9042 atomic load/store instructions.
9045 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
9047 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
9049 (insn_count): New attribute, with most cases extracted from...
9050 (length): ...here. Redefine most cases in terms of insn_count.
9051 (single_insn): Delete.
9052 (can_delay): Use insn_count to check for single instructions.
9053 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
9054 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
9055 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
9056 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
9057 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
9058 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
9059 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
9060 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
9061 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
9062 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
9063 rather than "length".
9064 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
9065 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
9066 Use "insn_count" rather than "length".
9067 * config/mips/mips-dsp.md
9068 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
9069 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
9072 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
9074 PR tree-optimization/57361
9075 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
9077 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
9079 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
9081 (aarch64_simd_combine<mode>): New instruction expansion.
9082 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
9084 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
9085 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
9087 2013-06-12 Jan Hubicka <jh@suse.cz>
9089 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
9090 decl has when in streaming stage.
9091 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
9092 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
9094 2013-06-12 Roland Stigge <stigge@antcom.de>
9097 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
9099 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9101 PR tree-optimization/57537
9102 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
9103 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
9105 2013-06-12 Richard Biener <rguenther@suse.de>
9107 * data-streamer.h (streamer_write_char_stream): CSE
9108 obs->current_pointer.
9109 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
9110 streamer_write_char_stream manually and optimize the resulting loop.
9111 (streamer_write_hwi_stream): Likewise.
9113 2013-06-12 Jan Hubicka <jh@suse.cz>
9115 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
9116 * cgraph.h (varpool_create_empty_node): Declare.
9117 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
9119 * symtab.c (symtab_unregister_node): Be lax about missin entries
9121 (symtab_get_node): Update comment.
9122 * varpool.c (varpool_create_empty_node): Break out from ...
9123 (varpool_node_for_decl): ... here.
9124 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
9126 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
9128 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
9129 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
9130 part. Use straight-line flow at the end.
9131 <COMPONENT_REF>: Remove superfluous else.
9132 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
9134 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9137 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
9138 target hook even for !TREE_PUBLIC decls. If no resolution info
9139 is available, return false for common and external decls.
9141 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9143 * config/rl78/constraints.md (U): New constraint.
9144 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
9147 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9150 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
9151 to allow returning address to AT_PLATFORM name.
9153 2013-06-11 Jan Hubicka <jh@suse.cz>
9155 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
9156 * cgraph.h (symtab_node_base): Add weakref flag.
9157 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
9158 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
9159 (output_weakrefs): Use weakref flag.
9160 * fold-const.c (simple_operand_p): Handle WEAK.
9161 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
9162 * ipa.c (varpool_externally_visible_p): Drop weakref.
9163 (function_and_variable_visibility): Update comment; fix weakref
9164 sanity checks; do not clear DECL_WEAK on them.
9165 * lto-cgraph.c (lto_output_node): update.
9166 (lto_output_varpool_node): Update.
9167 (input_overwrite_node): Update.
9168 (input_node): Update.
9169 (input_varpool_node): Update.
9170 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
9171 (lto_symtab_merge_symbols): Add sanity check.
9172 (lto_symtab_prevailing_decl): Do not special case weakrefs.
9173 * passes.c (rest_of_decl_compilation): Set static flag, too.
9174 * symtab.c (dump_symtab_base): Dump weakref.
9175 (verify_symtab_base): Sanity check weakrefs.
9176 (symtab_make_decl_local): Remove duplicated code.
9177 (symtab_alias_ultimate_target): Simplify.
9178 * varpool.c (varpool_create_variable_alias): Set weakref flag.
9180 2013-06-11 Tom de Vries <tom@codesourcery.com>
9182 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
9183 sequence_vect == NULL.
9185 2013-06-11 DJ Delorie <dj@redhat.com>
9187 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
9188 (rl78_unwind_word_mode): New.
9190 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9192 * final.c (debug_prefix_maps): Make static.
9194 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9196 * function.c (initial_trampoline): Remove stray copy.
9198 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
9200 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
9202 2013-06-11 Martin Jambor <mjambor@suse.cz>
9204 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
9205 within bounds at the beginning of the function.
9207 2013-06-11 Alan Modra <amodra@gmail.com>
9209 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
9211 (get_named_section): Don't NULL !DECL_P decl.
9213 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
9215 * doc/invoke.texi (core-avx2): Document.
9217 (atom): Updated with MOVBE.
9219 2013-06-11 Richard Biener <rguenther@suse.de>
9221 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
9223 2013-06-11 Anton Blanchard <anton@samba.org>
9225 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
9226 correct shift value in little-endian mode.
9228 2013-06-11 Jakub Jelinek <jakub@redhat.com>
9231 * varasm.c (get_variable_align): Move #endif to the right place.
9233 2013-06-10 Cary Coutant <ccoutant@google.com>
9235 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
9236 for hash so that hash table traversal order is deterministic.
9238 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
9239 Pat Haugen <pthaugen@us.ibm.com>
9240 Peter Bergner <bergner@vnet.ibm.com>
9242 * config/rs6000/vector.md (GPR move splitter): Do not split moves
9243 of vectors in GPRS if they are direct moves or quad word load or
9246 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
9248 (direct_move_p): Likewise.
9249 (quad_load_store_p): Likewise.
9251 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
9252 classes into bins based on the physical register type.
9253 (reg_class_to_reg_type): Likewise.
9254 (IS_STD_REG_TYPE): Likewise.
9255 (IS_FP_VECT_REG_TYPE): Likewise.
9256 (reload_fpr_gpr): Arrays to determine what insn to use if we can
9257 use direct move instructions.
9258 (reload_gpr_vsx): Likewise.
9259 (reload_vsx_gpr): Likewise.
9260 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
9261 information that is a simplification of register classes. Also
9262 precalculate direct move reload helpers.
9263 (direct_move_p): New function to return true if the operation can
9264 be done as a direct move instruciton.
9265 (quad_load_store_p): New function to return true if the operation
9266 is a quad memory operation.
9267 (rs6000_legitimize_address): If quad memory, only allow register
9268 indirect for TImode addresses.
9269 (rs6000_legitimate_address_p): Likewise.
9270 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
9271 (rs6000_reload_register_type): Likewise.
9272 (register_to_reg_type): Return register type.
9273 (rs6000_secondary_reload_simple_move): New helper function for
9274 secondary reload and secondary memory needed to identify anything
9275 that is a simple move, and does not need reloading.
9276 (rs6000_secondary_reload_direct_move): New helper function for
9277 secondary reload to identify cases that can be done with several
9278 instructions via the direct move instructions.
9279 (rs6000_secondary_reload_move): New helper function for secondary
9280 reload to identify moves between register types that can be done.
9281 (rs6000_secondary_reload): Add support for quad memory operations
9282 and for direct move.
9283 (rs6000_secondary_memory_needed): Likewise.
9284 (rs6000_debug_secondary_memory_needed): Change argument names.
9285 (rs6000_output_move_128bit): New function to return the move to
9286 use for 128-bit moves, including knowing about the various
9287 limitations of quad memory operations.
9289 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
9290 memory operations. call rs6000_output_move_128bit for the actual
9291 instruciton(s) to generate.
9292 (vsx_movti_64bit): Likewise.
9294 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
9295 (UNSPEC_P8V_MTVSRWZ): Likewise.
9296 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
9297 (UNSPEC_P8V_MTVSRD): Likewise.
9298 (UNSPEC_P8V_XXPERMDI): Likewise.
9299 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
9300 (UNSPEC_FUSION_GPR): Likewise.
9301 (FMOVE128_GPR): New iterator for direct move.
9302 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
9305 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
9306 loads and direct move instructions.
9307 (zero_extendsidi2_lfiwzx): Likewise.
9308 (extendsidi2_lfiwax): Likewise.
9309 (extendsidi2_nocell): Likewise.
9310 (floatsi<mode>2_lfiwax): Likewise.
9312 (floatunssi<mode>2_lfiwzx): Likewise.
9314 (fix_trunc<mode>_stfiwx): Likewise.
9315 (fixuns_trunc<mode>_stfiwx): Likewise.
9316 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
9317 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
9318 (parity<mode>2_cmpb): Set length/type attr.
9319 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
9320 for 'mr.' to fast_compare.
9321 (bpermd_<mode>): Change type attr to popcnt.
9322 (p8_fmrgow_<mode>): New insns for power8 direct move support.
9323 (p8_mtvsrwz_1): Likewise.
9324 (p8_mtvsrwz_2): Likewise.
9325 (reload_fpr_from_gpr<mode>): Likewise.
9326 (p8_mtvsrd_1): Likewise.
9327 (p8_mtvsrd_2): Likewise.
9328 (p8_xxpermdi_<mode>): Likewise.
9329 (reload_vsx_from_gpr<mode>): Likewise.
9330 (reload_vsx_from_gprsf): Likewise.
9331 (p8_mfvsrd_3_<mode>): LIkewise.
9332 (reload_gpr_from_vsx<mode>): Likewise.
9333 (reload_gpr_from_vsxsf): Likewise.
9334 (p8_mfvsrd_4_disf): Likewise.
9335 (multi-word GPR splits): Do not split direct moves or quad memory
9338 2013-06-10 David Malcolm <dmalcolm@redhat.com>
9340 * tree-into-ssa.c (interesting_blocks): Make static.
9342 2013-06-10 Jakub Jelinek <jakub@redhat.com>
9345 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
9346 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
9347 Use DATA_ABI_ALIGNMENT for that case instead if defined.
9348 (get_variable_align): New function.
9349 (get_variable_section, emit_bss, emit_common,
9350 assemble_variable_contents, place_block_symbol): Use
9351 get_variable_align instead of DECL_ALIGN.
9352 (assemble_noswitch_variable): Add align argument, use it
9353 instead of DECL_ALIGN.
9354 (assemble_variable): Adjust caller. Use get_variable_align
9355 instead of DECL_ALIGN.
9356 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
9358 (DATA_ABI_ALIGNMENT): Define.
9359 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
9360 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
9361 opt is false, only return the psABI mandated alignment increase.
9362 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
9363 (DATA_ABI_ALIGNMENT): ... this.
9364 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
9365 (DATA_ABI_ALIGNMENT): ... this.
9366 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
9367 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
9368 (DATA_ABI_ALIGNMENT): ... this.
9369 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
9370 * doc/tm.texi: Regenerated.
9372 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
9374 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
9375 cmp_code to construct REG_EQUAL note.
9377 2013-06-09 Jakub Jelinek <jakub@redhat.com>
9380 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
9381 that operands[2] doesn't overlap with operands[0].
9383 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
9384 Jan Hubicka <jh@suse.cz>
9386 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
9387 hack to mark symbols as used.
9389 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
9391 PR rtl-optimization/57559
9392 * lra-constraints.c (process_alt_operands): Don't discourage
9393 memory with known offset for offsetable memory constraint.
9394 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
9396 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
9398 * varasm.c (struct oc_local_state): Reorder fields.
9399 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
9400 and adjust accordingly.
9401 (output_constructor): Reorder initialization code and adjust call to
9402 output_constructor_bitfield.
9404 2013-06-07 Jan Hubicka <jh@suse.cz>
9406 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
9408 2013-06-07 David Malcolm <dmalcolm@redhat.com>
9410 * tree-object-size.c (unknown): Make const.
9412 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9414 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
9415 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
9416 for last alternative in the cpu_facility attribute.
9418 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9421 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
9422 (xordi3): Change operand 2 constraint to arm_xordi_operand.
9423 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
9424 * config/arm/constraints.md (Dg): New constraint.
9425 * config/arm/neon.md (xordi3_neon): Remove.
9426 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
9427 * config/arm/predicates.md (arm_xordi_operand): New predicate.
9429 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9431 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
9432 Clean up alternatives.
9434 2013-06-07 Alan Modra <amodra@gmail.com>
9436 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
9439 2013-06-07 Alan Modra <amodra@gmail.com>
9441 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
9443 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9445 * config/arm/constraints.md (Df): New constraint.
9446 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
9447 Correct length attribute for last two alternatives.
9449 2013-06-07 Alan Modra <amodra@gmail.com>
9451 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9452 override user -mfp-in-toc.
9453 (offsettable_ok_by_alignment): Consider just the current access
9454 rather than the whole object, unless BLKmode. Handle
9455 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
9456 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
9457 for -mcmodel=medium.
9458 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
9459 override user -mfp-in-toc or -msum-in-toc. Default to
9460 -mno-fp-in-toc for -mcmodel=medium.
9462 2013-06-06 DJ Delorie <dj@redhat.com>
9464 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
9465 TARGET_VALID_POINTER_MODE.
9467 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
9468 Pat Haugen <pthaugen@us.ibm.com>
9469 Peter Bergner <bergner@vnet.ibm.com>
9471 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9472 Document new power8 builtins.
9474 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
9475 condition code register, to allow 128-bit logical operations to be
9476 done in the VSX or GPR registers.
9477 (nor<mode>3): Use the canonical form for nor.
9478 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9479 vclz*, and vpopcnt* vector instructions.
9480 (nand<mode>3): Likewise.
9481 (orc<mode>3): Likewise.
9482 (clz<mode>2): LIkewise.
9483 (popcount<mode>2): Likewise.
9485 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9486 that only the GPRs are recognized.
9488 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9489 support for new power8 builtins.
9491 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9493 (xscvdpspn): Likewise.
9499 (vpopcnt): Likewise.
9500 (vpopcntb): Likewise.
9501 (vpopcnth): Likewise.
9502 (vpopcntw): Likewise.
9503 (vpopcntd): Likewise.
9508 (eqv_v16qi3): Likewise.
9509 (eqv_v8hi3): Likewise.
9510 (eqv_v4si3): Likewise.
9511 (eqv_v2di3): Likewise.
9512 (eqv_v4sf3): Likewise.
9513 (eqv_v2df3): Likewise.
9515 (nand_v16qi3): Likewise.
9516 (nand_v8hi3): Likewise.
9517 (nand_v4si3): Likewise.
9518 (nand_v2di3): Likewise.
9519 (nand_v4sf3): Likewise.
9520 (nand_v2df3): Likewise.
9522 (orc_v16qi3): Likewise.
9523 (orc_v8hi3): Likewise.
9524 (orc_v4si3): Likewise.
9525 (orc_v2di3): Likewise.
9526 (orc_v4sf3): Likewise.
9527 (orc_v2df3): Likewise.
9529 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9530 allow power8 quad mode in 64-bit.
9531 (rs6000_builtin_vectorized_function): Add support to vectorize
9532 ISA 2.07 count leading zeros, population count builtins.
9533 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9534 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9535 (builtin_function_type): Add vgbbd builtin function which takes an
9537 (altivec_expand_vec_perm_const): Add support for new power8 merge
9540 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9541 that does not include TImdoe for use with 32-bit.
9542 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9544 (UNSPEC_VSX_CVDPSPN): Likewise.
9545 (vsx_xscvdpspn): Likewise.
9546 (vsx_xscvspdpn): Likewise.
9547 (vsx_xscvdpspn_scalar): Likewise.
9548 (vsx_xscvspdpn_directmove): Likewise.
9549 (vsx_and<mode>3): Split logical operations into 32-bit and
9550 64-bit. Add support to do logical operations on TImode as well as
9551 VSX vector types. Allow logical operations to be done in either
9552 VSX registers or in general purpose registers in 64-bit mode. Add
9553 splitters if GPRs were used. For AND, add clobber of CCmode to
9554 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9556 (vsx_and<mode>3_32bit): Likewise.
9557 (vsx_and<mode>3_64bit): Likewise.
9558 (vsx_ior<mode>3): Likewise.
9559 (vsx_ior<mode>3_32bit): Likewise.
9560 (vsx_ior<mode>3_64bit): Likewise.
9561 (vsx_xor<mode>3): Likewise.
9562 (vsx_xor<mode>3_32bit): Likewise.
9563 (vsx_xor<mode>3_64bit): Likewise.
9564 (vsx_one_cmpl<mode>2): Likewise.
9565 (vsx_one_cmpl<mode>2_32bit): Likewise.
9566 (vsx_one_cmpl<mode>2_64bit): Likewise.
9567 (vsx_nor<mode>3): Likewise.
9568 (vsx_nor<mode>3_32bit): Likewise.
9569 (vsx_nor<mode>3_64bit): Likewise.
9570 (vsx_andc<mode>3): Likewise.
9571 (vsx_andc<mode>3_32bit): Likewise.
9572 (vsx_andc<mode>3_64bit): Likewise.
9573 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
9574 and xxlorc instructions.
9575 (vsx_eqv<mode>3_64bit): Likewise.
9576 (vsx_nand<mode>3_32bit): Likewise.
9577 (vsx_nand<mode>3_64bit): Likewise.
9578 (vsx_orc<mode>3_32bit): Likewise.
9579 (vsx_orc<mode>3_64bit): Likewise.
9581 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
9583 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
9585 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
9586 (p8_vmrgow): Likewise.
9587 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
9588 GPRs to be split under VSX.
9589 (p8v_clz<mode>2): Add power8 count leading zero support.
9590 (p8v_popcount<mode>2): Add power8 population count support.
9591 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
9594 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
9597 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
9599 (vec_nand): Likewise.
9600 (vec_vclz): Likewise.
9601 (vec_vclzb): Likewise.
9602 (vec_vclzd): Likewise.
9603 (vec_vclzh): Likewise.
9604 (vec_vclzw): Likewise.
9605 (vec_vgbbd): Likewise.
9606 (vec_vmrgew): Likewise.
9607 (vec_vmrgow): Likewise.
9608 (vec_vpopcnt): Likewise.
9609 (vec_vpopcntb): Likewise.
9610 (vec_vpopcntd): Likewise.
9611 (vec_vpopcnth): Likewise.
9612 (vec_vpopcntw): Likewise.
9614 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9616 PR rtl-optimization/57468
9617 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
9620 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9622 PR rtl-optimization/57459
9623 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
9624 type when setting live regs.
9626 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9628 * config/s390/s390.opt (mlra): New option.
9629 * config/s390/s390.c (s390_decompose_address): Check displacement
9630 for all registers for LRA.
9631 (s390_secondary_reload): Don't used secondary reloads for LRA.
9632 (s390_lra_p): New function.
9633 (TARGET_LRA_P): Define.
9634 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
9635 of attribute cpu_facility to zarch for the last alternative.
9636 (*cmpmem_short): Ditto.
9638 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9640 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9641 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9642 (arm_expand_prologue): Likewise.
9644 2013-06-06 Teresa Johnson <tejohnson@google.com>
9647 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9648 as this is now done by redirect_edge_and_branch_force.
9649 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9650 barriers, and fix interaction with splitting.
9651 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9652 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9653 reflect changes made in the routine.
9654 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9655 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9656 since this is called in cfglayout mode, and replace partition fixup
9657 with assert as that is now done by force_nonfallthru_and_redirect.
9658 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9659 already be marked with region crossing note.
9660 (insert_section_boundary_note): Make non-static, gate on flag
9661 has_bb_partition, rewrite to also check for multiple partitions.
9662 (rest_of_handle_reorder_blocks): Remove call to
9663 insert_section_boundary_note, now done later during free_cfg.
9664 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9665 * bb-reorder.h (insert_section_boundary_note): Declare.
9666 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9667 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9668 invoke insert_section_boundary_note.
9669 (try_redirect_by_replacing_jump): Remove unnecessary
9670 check for region crossing note.
9671 (fixup_partition_crossing): New function.
9672 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9673 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9674 in non-cfglayout mode.
9675 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9676 remove old code that tried to do this. Emit barrier correctly
9677 when we are in cfglayout mode.
9678 (last_bb_in_partition): New function.
9679 (rtl_split_edge): Correctly fixup partition boundaries.
9680 (commit_one_edge_insertion): Remove old code that tried to
9681 fixup region crossing edge since this is now handled in
9682 split_block, and set up insertion point correctly since
9683 block may now end in a jump.
9684 (verify_hot_cold_block_grouping): Guard against checking when not in
9685 linearized RTL mode.
9686 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9688 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9689 rtl_verify_flow_info, so not called in cfglayout mode.
9690 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9691 (fixup_reorder_chain): Remove old code that attempted to fixup region
9692 crossing note as this is now handled in force_nonfallthru_and_redirect.
9693 (duplicate_insn_chain): Don't duplicate switch section notes.
9694 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9696 * basic-block.h (emit_barrier_after_bb): Declare.
9698 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9700 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9701 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9702 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9704 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9706 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9708 * config/arm/ldmstm.md: Regenerate.
9710 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9712 * config/arm/sync.md (atomic_loaddi_1):
9713 Disable predication for arm_restrict_it.
9714 (arm_load_exclusive<mode>): Likewise.
9715 (arm_load_exclusivesi): Likewise.
9716 (arm_load_exclusivedi): Likewise.
9717 (arm_load_acquire_exclusive<mode>): Likewise.
9718 (arm_load_acquire_exclusivesi): Likewise.
9719 (arm_load_acquire_exclusivedi): Likewise.
9720 (arm_store_exclusive<mode>): Likewise.
9721 (arm_store_exclusive<mode>): Likewise.
9722 (arm_store_release_exclusivedi): Likewise.
9723 (arm_store_release_exclusive<mode>): Likewise.
9725 2013-06-06 Richard Biener <rguenther@suse.de>
9727 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9729 (lto_tag_is_tree_code_p): Adjust.
9730 (lto_tag_is_gimple_code_p): Likewise.
9731 (lto_gimple_code_to_tag): Likewise.
9732 (lto_tag_to_gimple_code): Likewise.
9733 (lto_tree_code_to_tag): Likewise.
9734 (lto_tag_to_tree_code): Likewise.
9735 * data-streamer.h (streamer_write_hwi_in_range): Use
9736 uhwi streaming to stream the normalized range.
9737 (streamer_read_hwi_in_range): Likewise.
9739 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9741 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9742 (predicable_short_it): Likewise.
9743 (predicated): Likewise.
9744 (enabled): Handle above.
9745 (define_cond_exec): Set predicated attribute to yes.
9747 2013-06-05 Mike Stump <mikestump@comcast.net>
9749 * gdbinit.in (__FUNCTION__): Add.
9751 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
9753 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9754 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9756 2013-06-05 Jan Hubicka <jh@suse.cz>
9758 * varasm.c (mark_decl_referenced): Revert the removal until targets
9761 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
9763 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9764 instead of mark_decl_referenced.
9766 2013-06-05 Jan Hubicka <jh@suse.cz>
9768 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9769 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9770 and symtab_used_from_object_file_p.
9771 (cgraph_make_node_local_1): Clear forced_by_abi.
9772 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9773 * cgraph.h (symtab_node_base): Add forced_by_abi.
9774 (decide_is_variable_needed): Remove.
9775 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9776 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9777 (decide_is_symbol_needed): ... this one; handle symbols in general;
9778 always analyze virtuals; honnor forced_by_abi.
9779 (cgraph_finalize_function): Update.
9780 (varpool_finalize_decl): Update.
9781 (symbol_defined_and_needed): Remove.
9782 (analyze_functions): Update.
9783 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9784 output_refs, input_overwrite_node): Handle forced_by_abi.
9785 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9786 (address_taken_from_non_vtable_p): ... this one.
9787 (comdat_can_be_unshared_p_1): New function.
9788 (cgraph_comdat_can_be_unshared_p): Rename to ...
9789 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9790 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9791 (function_and_variable_visibility): Clear forced_by_abi as needed.
9792 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9793 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
9794 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9795 * varpool.c (decide_is_variable_needed): Remove.
9797 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9799 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9800 (arm_option_override): Override arm_restrict_it where appropriate.
9801 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9802 * config/arm/arm.opt (mrestrict-it): New command-line option.
9803 * doc/invoke.texi: Document -mrestrict-it.
9805 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9807 * tsan.c (tsan_atomic_table): Make const.
9809 2013-06-05 Richard Biener <rguenther@suse.de>
9811 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9812 index associated with the tree we are supposed to replace.
9813 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9814 TREE_ASM_WRITTEN as zero for everything but SSA names.
9816 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9818 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9820 2013-06-04 Jan Hubicka <jh@suse.cz>
9822 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9823 (input_node, input_varpool_node): Handle correctly external same
9825 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9826 nodes at ltrans stage.
9828 2013-06-04 Jan Hubicka <jh@suse.cz>
9830 * ipa-inline.c (update_caller_keys): Fix availability test.
9831 (update_callee_keys): Likewise.
9832 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
9833 to follow ELF standard.
9835 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
9837 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
9838 (mips64r5900el-*-elf*): New configurations.
9839 * config/mips/mips-cpus.def (r5900): New processor.
9840 * config/mips/mips-tables.opt: Regenerate.
9841 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
9842 (mips_issue_rate): Handle PROCESSOR_R5900.
9843 (mips_reorg_process_insns): Force reorder mode for the R5900.
9844 * config/mips/mips.h (TARGET_MIPS5900): Define.
9845 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
9847 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
9849 * config/mips/mips.md (processor): Add r5900.
9850 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
9852 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9854 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
9855 into function to generate MOVI instruction.
9856 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
9857 (aarch64_preferred_simd_mode): Turn into wrapper.
9858 (aarch64_output_scalar_simd_mov_immediate): New function.
9859 * config/aarch64/aarch64-protos.h: Add prototype for above.
9861 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9863 * config/aarch64/aarch64.c (simd_immediate_info): Remove
9864 element_char member.
9865 (sizetochar): Return signed char.
9866 (aarch64_simd_valid_immediate): Remove elchar and other
9867 unnecessary variables.
9868 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
9869 Calculate element_char as required.
9870 * config/aarch64/aarch64-protos.h: Update and move prototype
9871 for aarch64_output_simd_mov_immediate.
9872 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
9875 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9877 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
9878 information completed by aarch64_simd_valid_immediate.
9879 (aarch64_legitimate_constant_p): Update arguments.
9880 (aarch64_simd_valid_immediate): Work with struct rather than many
9882 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
9883 (aarch64_simd_make_constant): Update arguments.
9884 (aarch64_output_simd_mov_immediate): Work with struct rather than
9885 many pointers. Output immediate directly rather than as operand.
9886 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9888 * config/aarch64/constraints.md (Dn): Update arguments.
9890 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9892 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
9894 (aarch64_simd_immediate_valid_for_move): Remove.
9895 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
9896 (aarch64_simd_make_constant): Update call.
9897 (aarch64_output_simd_mov_immediate): Update call.
9898 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9900 * config/aarch64/constraints.md (Dn): Update call.
9902 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9904 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
9905 return type to bool for prototype.
9906 (aarch64_legitimate_constant_p): Check for true instead of not -1.
9907 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
9908 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
9910 2013-06-04 Catherine Moore <clm@codesourcery.com>
9912 * config/mips/mips.opt (meva): New.
9913 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
9914 (ASM_SPEC): Handle -meva.
9915 * doc/invoke.texi (meva): Document.
9917 2013-06-04 Alan Modra <amodra@gmail.com>
9919 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
9922 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9924 * rtl.def: Add extra fourth optional field to define_cond_exec.
9925 * gensupport.c (process_one_cond_exec): Process attributes from
9927 * doc/md.texi: Document fourth field in define_cond_exec.
9929 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
9931 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
9932 out the processing order as in store_bit_field_1.
9934 2013-06-04 Jan Hubicka <jh@suse.cz>
9937 * cgraphunit.c (cgraph_process_same_body_aliases): Create
9938 non-VAR_DECL node if it does not exist yet.
9940 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
9942 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
9943 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
9944 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
9945 target_cpu_default setting.
9947 2013-06-03 Teresa Johnson <tejohnson@google.com>
9949 * dumpfile.c (opt_info_switch_p): Change -fopt-info
9950 default to -fopt-info=optimized instead of all.
9951 * doc/invoke.texi: Ditto.
9952 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
9953 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
9954 (execute_vect_slp): Emit BB vectorization success under
9955 MSG_OPTIMIZED_LOCATIONS.
9956 * tree-vect-slp.c (vect_slp_transform_bb): Change
9957 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
9958 * tree-vect-loop.c (vect_transform_loop): Ditto.
9960 2013-06-03 Jason Merrill <jason@redhat.com>
9963 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9964 Use TARGET_EXPR for C++.
9966 2013-06-03 Jakub Jelinek <jakub@redhat.com>
9968 PR rtl-optimization/57268
9969 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
9970 if DEBUG_INSN_P (insn).
9973 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9975 PR rtl-optimization/57268
9976 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9977 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9979 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9981 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
9982 (ix86_avoid_lea_for_addr): Likewise.
9983 (exact_dependency_1): Likewise.
9984 (ix86_adjust_cost): Likewise.
9985 (swap_top_of_ready_list): Fix formatting and !reload_completed check
9987 (do_reorder_for_imul): Fix typo, formatting and
9988 !reload_completed check removed.
9989 (ix86_sched_reorder): Fix typo and formatting.
9990 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
9993 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
9995 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
9997 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
9999 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
10000 <STRING_CST>: Likewise.
10001 <VECTOR_CST>: Likewise.
10003 2013-06-01 Janus Weil <janus@gcc.gnu.org>
10004 Mikael Morin <mikael@gcc.gnu.org>
10006 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
10007 * config.in: Regenerated.
10008 * configure: Regenerated.
10010 2013-06-01 Jan Hubicka <jh@suse.cz>
10012 PR middle-end/57366
10013 * cgraphunit.c (compile): When weakref is not supported,
10014 set up transparent aliases before final output pass.
10015 * varasm.c (assemble_alias): Do not try to do it here.
10017 2013-06-01 Jan Hubicka <jh@suse.cz>
10019 PR middle-end/57467
10020 * passes.c (for_per_function): Skip unanalyzed functions.
10022 2013-06-01 Jan Hubicka <jh@suse.cz>
10024 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
10025 (lto_symtab_merge_symbols_1): ... this one.
10026 (lto_symtab_merge_cgraph_nodes): Rename to ...
10027 (lto_symtab_merge_symbols): ... this one; simplify.
10028 * cgraph.c (same_body_aliases_done): Rename to ...
10029 (cpp_implicit_aliases_done): ... this one.
10030 (cgraph_create_function_alias): Update.
10031 (cgraph_same_body_alias): Update.
10032 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
10033 (verify_edge_corresponds_to_fndecl): Simplify.
10034 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
10035 (cgraph_node): Remove same_body_alias.
10036 (varpool_node): Remove alias_of and extra_name_alias.
10037 (same_body_aliases_done): Rename to ..
10038 (cpp_implicit_aliases_done): ... this one.
10039 (symtab_alias_ultimate_target): Add default parameter.
10040 (symtab_resolve_alias): New function.
10041 (fixup_same_cpp_alias_visibility): Declare.
10042 (cgraph_function_node): Add default parameter.
10043 (cgraph_node_asm_name): Likewise.
10044 (cgraph_function_or_thunk_node): Add default parameter; do
10045 not ICE when it is NULL.
10046 (varpool_variable_node): Likewise.
10047 * tree-emutls.c (create_emultls_var): Update.
10048 (ipa_lower_emutls): Update.
10049 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
10050 (cgraph_reset_node): Reset alias info.
10051 (cgraph_finalize_function): Update.
10052 (fixup_same_cpp_alias_visibility): Move to symtab.c.
10053 (analyze_function): Simplify.
10054 (cgraph_process_same_body_aliases): Simplify.
10055 (analyze_functions): Fixup same body aliases.
10056 (handle_alias_pairs): Simplify.
10057 (assemble_thunk): Update.
10058 (assemble_thunks_and_aliases): Update.
10059 (output_weakrefs): Rewrite.
10060 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
10061 (lto_output_varpool_node): Likewise.
10062 (compute_ltrans_boundary): Remve assert.
10063 (get_alias_symbol): New functoin.
10064 (input_node): Rewrite alias handling.
10065 (input_varpool_node): Likewise.
10066 * ipa-pure-const.c (propagate_pure_const): Fix formating.
10067 * ipa.c (process_references): Handle weakrefs correctly.
10068 (symtab_remove_unreachable_nodes): Likewise.
10069 * trans-mem.c (get_cg_data): Update.
10070 (ipa_tm_create_version_alias): Update.
10071 (ipa_tm_execute): Update.
10072 * symtab.c (dump_symtab_base): Dump aliases.
10073 (verify_symtab_base): Verify aliases.
10074 (symtab_node_availability): New function.
10075 (symtab_alias_ultimate_target): Simplify.
10076 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
10077 handle all the fixup cases.
10078 (symtab_resolve_alias): New function.
10079 * passes.c (ipa_write_summaries): Handle weakrefs.
10080 * varpool.c (varpool_analyze_node): Simplify.
10081 (assemble_aliases): Update.
10082 (varpool_create_variable_alias): Simplify.
10083 (varpool_extra_name_alias): Simplify.
10084 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
10085 (lto_symtab_merge_symbols): ... this one.
10087 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
10090 PR rtl-optimization/57268
10091 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10092 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10094 2013-06-01 Tobias Burnus <burnus@net-b.de>
10096 Partially reverted:
10097 2013-05-31 Tobias Burnus <burnus@net-b.de>
10099 PR middle-end/57073
10100 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
10103 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10105 PR rtl-optimization/57268
10106 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10107 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10109 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10111 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
10112 unordered comparison operators when -fno-trapping-math is in effect
10114 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
10115 and implement unordered comparison operators properly on the e500.
10117 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10119 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
10120 for constant scalar integers.
10121 (simplify_relational_operation_1): Likewise.
10123 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
10125 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
10126 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
10129 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10130 Igor Zamyatin <igor.zamyatin@intel.com>
10132 Silvermont (SLM) architecture performance tuning.
10133 * config/i386/i386.h (enum ix86_tune_indices): Add
10134 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
10135 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
10137 * config/i386/i386.c (initial_ix86_tune_features)
10138 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
10139 (ix86_lea_outperforms): Handle Silvermont tuning.
10140 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
10142 (ix86_use_lea_for_mov): Likewise.
10143 (ix86_avoid_lea_for_addr): Likewise.
10144 (ix86_lea_for_add_ok): Likewise.
10145 (exact_dependency_1): New function.
10146 (exact_store_load_dependency): Likewise.
10147 (ix86_adjust_cost): Handle Silvermont tuning.
10148 (do_reoder_for_imul): Likewise.
10149 (swap_top_of_ready_list): New function.
10150 (ix86_sched_reorder): Changed to handle Silvermont tuning.
10152 * config/i386/i386.md (peepholes that split memory operand in fp
10155 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10157 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
10158 Remove un-necessary braces.
10160 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10162 * config/aarch64/aarch64.c (aarch64_classify_symbol):
10163 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
10165 2013-05-31 Tobias Burnus <burnus@net-b.de>
10167 PR middle-end/57073
10168 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
10170 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10173 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
10174 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
10175 * config/arm/neon.md (iordi3_neon): Remove.
10176 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
10177 * config/arm/predicates.md (imm_for_neon_logic_operand):
10178 Move to earlier in the file.
10179 (neon_logic_op2): Likewise.
10180 (arm_iordi_operand_neon): New predicate.
10182 2013-05-31 Richard Biener <rguenther@suse.de>
10184 PR tree-optimization/57478
10185 PR tree-optimization/57453
10186 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
10189 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10191 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
10192 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
10194 2013-05-30 Tobias Burnus <burnus@net-b.de>
10195 Thomas Koenig <tkoenig@gcc.gnu.org>
10197 PR middle-end/57073
10198 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
10199 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
10201 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
10203 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
10205 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
10207 * target.def (register_usage_leveling_p): New hook.
10208 * targhooks.c (default_register_usage_leveling_p): New.
10209 * targhooks.h (default_register_usage_leveling_p): New prototype.
10210 * lra-assigns.c (register_usage_leveling_p): Use the hook.
10211 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
10212 * doc/tm.texi: Update.
10213 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
10215 2013-05-30 Ian Bolton <ian.bolton@arm.com>
10217 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
10218 (*insv_reg<mode>): New define_insn.
10220 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
10222 PR rtl-optimization/57439
10223 * postreload.c (move2add_valid_value_p): Check that we have
10224 a zero subreg_regno_offset when accessing the register in
10225 the requested mode.
10227 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10228 Igor Zamyatin <igor.zamyatin@intel.com>
10230 Silvermont (SLM) architecture pipeline model, tuning and
10232 * config.gcc: Add slm config options and target.
10234 * config/i386/slm.md: New.
10236 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
10238 * config/i386/i386-c.c (ix86_target_macros_internal): New case
10240 (ix86_target_macros_internal): Likewise.
10242 * config/i386/i386.c (slm_cost): New cost.
10243 (m_SLM): New macro flag.
10244 (initial_ix86_tune_features): Set m_SLM.
10245 (x86_accumulate_outgoing_args): Likewise.
10246 (x86_arch_always_fancy_math_387): Likewise.
10247 (processor_target_table): Add slm cost.
10248 (cpu_names): Add slm cpu name.
10249 (x86_option_override_internal): Set SLM ISA.
10250 (ix86_issue_rate): New case PROCESSOR_SLM.
10251 (ia32_multipass_dfa_lookahead): Likewise.
10252 (fold_builtin_cpu): Add slm.
10254 * config/i386/i386.h (TARGET_SLM): New target macro.
10255 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
10256 (processor_type): Add PROCESSOR_SLM.
10258 * config/i386/i386.md (cpu): Add new value "slm".
10259 (slm.md): Include slm.md.
10261 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
10262 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10264 * config/arm/arm-protos.h: Add and update function protos.
10265 * config/arm/arm.c (use_simple_return_p): New added.
10266 (thumb2_expand_return): Check simple_return flag.
10267 * config/arm/arm.md: Add simple_return and conditional simple_return.
10268 * config/arm/iterators.md: Add iterator for return and simple_return.
10270 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10272 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
10273 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
10274 (arm_emit_vfp_multi_reg_pop): Likewise.
10275 (thumb2_emit_ldrd_pop): Likewise.
10276 (arm_expand_epilogue): Add misc REG_CFA notes.
10277 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
10279 2013-05-29 Lawrence Crowl <crowl@google.com>
10281 * config/arm/t-arm: Update for below.
10283 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
10284 Change type to hash_table. Update dependent calls and types.
10286 * config/i386/t-cygming: Update for below.
10288 * config/i386/t-interix: Update for below.
10290 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
10291 Change type to hash_table. Update dependent calls and types.
10292 (i386_find_on_wrapper_list::wrappers): Likewise.
10294 * config/ia64/t-ia64: Update for below.
10296 * config/ia64/ia64.c (bundle_state_table):
10297 Change type to hash_table. Update dependent calls and types.
10299 * config/mips/mips.c (mips_reorg_process_insns::htab):
10300 Change type to hash_table. Update dependent calls and types.
10302 * config/sol2.c (solaris_comdat_htab):
10303 Change type to hash_table. Update dependent calls and types.
10305 * config/t-sol2: Update for above.
10307 2013-05-29 Teresa Johnson <tejohnson@google.com>
10309 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
10310 functions are not yet marked as defined.
10312 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
10313 Pat Haugen <pthaugen@us.ibm.com>
10314 Peter Bergner <bergner@vnet.ibm.com>
10316 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
10320 (vrotl<mode>3): Likewise.
10321 (vashl<mode>3): Likewise.
10322 (vlshr<mode>3): Likewise.
10323 (vashr<mode>3): Likewise.
10325 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10326 support for power8 V2DI builtins.
10328 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
10329 power8 V2DI builtins.
10330 (vupkhsw): Likewise.
10331 (vupklsw): Likewise.
10332 (vaddudm): Likewise.
10333 (vminsd): Likewise.
10334 (vmaxsd): Likewise.
10335 (vminud): Likewise.
10336 (vmaxud): Likewise.
10337 (vpkudum): Likewise.
10338 (vpksdss): Likewise.
10339 (vpkudus): Likewise.
10340 (vpksdus): Likewise.
10345 (vsubudm): Likewise.
10346 (vcmpequd): Likewise.
10347 (vcmpgtsd): Likewise.
10348 (vcmpgtud): Likewise.
10349 (vcmpequd_p): Likewise.
10350 (vcmpgtsd_p): Likewise.
10351 (vcmpgtud_p): Likewise.
10352 (vupkhsw): Likewise.
10353 (vupklsw): Likewise.
10354 (vaddudm): Likewise.
10355 (vmaxsd): Likewise.
10356 (vmaxud): Likewise.
10357 (vminsd): Likewise.
10358 (vminud): Likewise.
10359 (vpksdss): Likewise.
10360 (vpksdus): Likewise.
10361 (vpkudum): Likewise.
10362 (vpkudus): Likewise.
10367 (vsubudm): Likewise.
10369 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
10370 support for power8 V2DI instructions.
10372 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
10373 power8 V2DI instructions. Combine pack and unpack insns to use an
10374 iterator for each mode. Check whether a particular mode supports
10375 Altivec instructions instead of just checking TARGET_ALTIVEC.
10376 (UNSPEC_VPKUWUM): Likewise.
10377 (UNSPEC_VPKSHSS): Likewise.
10378 (UNSPEC_VPKSWSS): Likewise.
10379 (UNSPEC_VPKUHUS): Likewise.
10380 (UNSPEC_VPKSHUS): Likewise.
10381 (UNSPEC_VPKUWUS): Likewise.
10382 (UNSPEC_VPKSWUS): Likewise.
10383 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
10384 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
10385 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
10386 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
10387 (UNSPEC_VUPKHSB): Likewise.
10388 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
10389 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
10390 (UNSPEC_VUPKHSH): Likewise.
10391 (UNSPEC_VUPKLSB): Likewise.
10392 (UNSPEC_VUPKLSH): Likewise.
10394 (VI_char): Likewise.
10395 (VI_scalar): Likewise.
10396 (VI_unit): Likewise.
10398 (VP_small): Likewise.
10399 (VP_small_lc): Likewise.
10400 (VU_char): Likewise.
10401 (add<mode>3): Likewise.
10402 (altivec_vaddcuw): Likewise.
10403 (altivec_vaddu<VI_char>s): Likewise.
10404 (altivec_vadds<VI_char>s): Likewise.
10405 (sub<mode>3): Likewise.
10406 (altivec_vsubcuw): Likewise.
10407 (altivec_vsubu<VI_char>s): Likewise.
10408 (altivec_vsubs<VI_char>s): Likewise.
10409 (altivec_vavgs<VI_char>): Likewise.
10410 (altivec_vcmpbfp): Likewise.
10411 (altivec_eq<mode>): Likewise.
10412 (altivec_gt<mode>): Likewise.
10413 (altivec_gtu<mode>): Likewise.
10414 (umax<mode>3): Likewise.
10415 (smax<mode>3): Likewise.
10416 (umin<mode>3): Likewise.
10417 (smin<mode>3): Likewise.
10418 (altivec_vpkuhum): Likewise.
10419 (altivec_vpkuwum): Likewise.
10420 (altivec_vpkshss): Likewise.
10421 (altivec_vpkswss): Likewise.
10422 (altivec_vpkuhus): Likewise.
10423 (altivec_vpkshus): Likewise.
10424 (altivec_vpkuwus): Likewise.
10425 (altivec_vpkswus): Likewise.
10426 (altivec_vpks<VI_char>ss): Likewise.
10427 (altivec_vpks<VI_char>us): Likewise.
10428 (altivec_vpku<VI_char>us): Likewise.
10429 (altivec_vpku<VI_char>um): Likewise.
10430 (altivec_vrl<VI_char>): Likewise.
10431 (altivec_vsl<VI_char>): Likewise.
10432 (altivec_vsr<VI_char>): Likewise.
10433 (altivec_vsra<VI_char>): Likewise.
10434 (altivec_vsldoi_<mode>): Likewise.
10435 (altivec_vupkhsb): Likewise.
10436 (altivec_vupkhs<VU_char>): Likewise.
10437 (altivec_vupkls<VU_char>): Likewise.
10438 (altivec_vupkhsh): Likewise.
10439 (altivec_vupklsb): Likewise.
10440 (altivec_vupklsh): Likewise.
10441 (altivec_vcmpequ<VI_char>_p): Likewise.
10442 (altivec_vcmpgts<VI_char>_p): Likewise.
10443 (altivec_vcmpgtu<VI_char>_p): Likewise.
10444 (abs<mode>2): Likewise.
10445 (vec_unpacks_hi_v16qi): Likewise.
10446 (vec_unpacks_hi_v8hi): Likewise.
10447 (vec_unpacks_lo_v16qi): Likewise.
10448 (vec_unpacks_hi_<VP_small_lc>): Likewise.
10449 (vec_unpacks_lo_v8hi): Likewise.
10450 (vec_unpacks_lo_<VP_small_lc>): Likewise.
10451 (vec_pack_trunc_v8h): Likewise.
10452 (vec_pack_trunc_v4si): Likewise.
10453 (vec_pack_trunc_<mode>): Likewise.
10455 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
10457 (vec_vmaxsd): Likewise.
10458 (vec_vmaxud): Likewise.
10459 (vec_vminsd): Likewise.
10460 (vec_vminud): Likewise.
10461 (vec_vpksdss): Likewise.
10462 (vec_vpksdus): Likewise.
10463 (vec_vpkudum): Likewise.
10464 (vec_vpkudus): Likewise.
10465 (vec_vrld): Likewise.
10466 (vec_vsld): Likewise.
10467 (vec_vsrad): Likewise.
10468 (vec_vsrd): Likewise.
10469 (vec_vsubudm): Likewise.
10470 (vec_vupkhsw): Likewise.
10471 (vec_vupklsw): Likewise.
10473 2013-05-29 Jan Hubicka <jh@suse.cz>
10475 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
10476 flags; reorder rest of fields in more consistent way.
10477 (varpool_node): Remove analyzed, finalized and alias.
10478 (cgraph_ndoe): Likewise.
10479 (symtab_alias_ultimate_target): New function.
10480 (cgraph_function_node): Move offline.
10481 (cgraph_reset_node): Declare.
10482 (cgraph_comdat_can_be_unshared_p): Remove.
10483 (varpool_remove_initializer): Declare.
10484 (varpool_first_defined_variable, varpool_next_defined_variable
10485 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10486 (cgraph_function_with_gimple_body_p): Update.
10487 (varpool_all_refs_explicit_p): Update.
10488 (symtab_alias_target): New function.
10489 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10490 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
10491 (cgraph_function_or_thunk_node): Simplify using
10492 symtab_alias_ultimate_target.
10493 (varpool_variable_node): Likewise.
10494 * cgraph.c (cgraph_create_function_alias): Update.
10495 (cgraph_add_thunk): Update.
10496 (cgraph_remove_node): Update.
10497 (dump_cgraph_node): Do not dump removed flags.
10498 (cgraph_function_body_availability): Update.
10499 (cgraph_propagate_frequency): Update.
10500 (verify_cgraph_node): Check sanity of local flag.
10501 (cgraph_function_node): Move here from cgraph.h; revamp for
10502 cgraph_function_or_thunk_node.
10503 * lto-symtab.c (lto_varpool_replace_node): Update.
10504 (lto_symtab_resolve_can_prevail_p): Update.
10505 (lto_symtab_merge_cgraph_nodes): Update.
10506 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10507 propagate_constants_accross_call, devirtualization_time_bonus,
10508 ipcp_propagate_stage): Update.
10509 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
10510 * ipa-inline-transform.c (clone_inlined_nodes,
10511 preserve_function_body_p): Update.
10512 * ipa-reference.c (propagate): Update.
10513 (write_node_summary_p): Update.
10514 * toplev.c (wrapup_global_declaration_2): Update.
10515 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10516 (analyze_function) ... this one.
10517 (cgraph_process_new_functions): Update.
10518 (cgraph_reset_node): Export.
10519 (cgraph_finalize_function): Update.
10520 (cgraph_add_new_function): Update.
10521 (process_function_and_variable_attributes): Update.
10522 (varpool_finalize_decl): Update.
10523 (symbol_finalized): Remove.
10524 (symbol_finalized_and_needed): Rename to ...
10525 (symbol_defined_and_needed): ... update.
10526 (cgraph_analyze_functions): Update.
10527 (handle_alias_pairs): Update.
10528 (mark_functions_to_output): Update.
10529 (assemble_thunk): Update.
10530 (output_in_order): Update.
10531 (output_weakrefs): Update.
10532 (finalize_compilation_unit): Update.
10533 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10534 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10535 input_node, input_varpool_node): Update.
10536 * dbxout.c (dbxout_expand_expr): Update.
10537 * cgraphclones.c (cgraph_clone_node): Update.
10538 (cgraph_copy_node_for_versioning): Update.
10539 (cgraph_materialize_clone): Update.
10540 (cgraph_materialize_all_clones): Update.
10541 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10542 propagate_pure_const, propagate_nothrow): Update.
10543 * lto-streamer-out.c (lto_output, write_symbol): Update.
10544 * ipa-utils.c (ipa_reverse_postorder): Update.
10545 * ipa-inline.c (can_inline_edge_p): Update.
10546 (update_caller_keys, ipa_inline): Update.
10547 * dwarf2out.c (reference_to_unused,
10548 premark_types_used_by_global_vars_helper): Update.
10549 * tree-eh.c (tree_could_trap_p): Update.
10550 * ipa-split.c (consider_split, execute_split_functions): Update.
10551 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
10552 has_addr_references_p): Update; move ahead in file for better
10554 (process_references): Simplify.
10555 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10556 bodies are removed.
10557 (cgraph_comdat_can_be_unshared_p): Make static.
10558 (cgraph_externally_visible_p): Update.
10559 (varpool_externally_visible_p): Update.
10560 (function_and_variable_visibility): Update.
10561 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10562 ipa_tm_mark_force_output_node): Update.
10563 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10564 estimate_edge_devirt_benefit, inline_generate_summary,
10565 inline_write_summary): Update.
10566 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10567 * ipa-prop.c (ipa_compute_jump_functions): Update.
10568 (ipa_print_node_params, ipa_prop_read_section,
10569 ipa_update_after_lto_read, read_replacements_section): Update.
10570 * varasm.c (mark_decl_referenced): Update.
10571 (assemble_alias, dump_tm_clone_pairs): Update.
10572 * tree-inline.c (copy_bb): Update.
10573 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
10575 * symtab.c (dump_symtab_base): Print new flags.
10576 (verify_symtab_base): Verify new flags.
10577 (symtab_alias_ultimate_target): New function.
10578 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
10579 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
10581 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
10583 * i386.c (ix86_get_function_versions_dispatcher,
10584 ix86_generate_version_dispatcher_body): Update.
10585 (fold_builtin_cpu): Use varpool_add_new_variable.
10586 * varpool.c (varpool_remove_initializer): Break out from ...
10587 (varpool_remove_node): ... this one.
10588 (dump_varpool_node, varpool_node_for_asm,
10589 cgraph_variable_initializer_availability, varpool_analyze_node,
10590 varpool_assemble_decl, varpool_remove_unreferenced_decls,
10591 varpool_finalize_named_section_flags, varpool_create_variable_alias):
10594 2013-05-29 Jan Hubicka <jh@suse.cz>
10596 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
10598 2013-05-29 Easwaran Raman <eraman@google.com>
10600 PR tree-optimization/57442
10601 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
10602 when control exits the main loop.
10604 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
10606 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
10608 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
10609 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
10610 * rx/t-rx: Add rx100 under multi library matches option for nofpu
10613 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10615 PR tree-optimization/57441
10616 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
10617 Don't limit size of incr_vec to number of candidates.
10619 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
10621 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
10622 and mips16 directories.
10623 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
10624 (MULTILIB_DIRNAMES): Ditto.
10625 (MULTILIB_EXCEPTIONS): Add new exceptions.
10626 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
10627 (MULTILIB_DIRNAMES): Ditto.
10628 (MULTILIB_EXCEPTIONS): Add new exceptions.
10630 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10631 Marcus Shawcroft <marcus.shawcroft@arm.com>
10633 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
10634 SYMBOL_TINY_ABSOLUTE.
10635 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
10636 SYMBOL_TINY_ABSOLUTE.
10637 (aarch64_expand_mov_immediate): Likewise.
10638 (aarch64_classify_symbol): Likewise.
10639 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10640 Permit SYMBOL_TINY_ABSOLUTE.
10641 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10643 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10644 Marcus Shawcroft <marcus.shawcroft@arm.com>
10646 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10647 Refactor if/switch. Replace gcc_assert with if.
10649 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10651 * config/i386/i386.c (initial_ix86_tune_features): Enable
10652 FP Reassociation for AMD bdver1 and bdver2.
10654 2013-05-29 Martin Jambor <mjambor@suse.cz>
10656 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10657 and IMAGPART_EXPR do not occur within other handled_components.
10659 2013-05-29 Richard Biener <rguenther@suse.de>
10661 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10662 access on whether the use is in the BB we currently try to
10664 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10665 vectorize to vect_bb_slp_scalar_cost.
10667 2013-05-29 Richard Biener <rguenther@suse.de>
10669 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10670 computing scalar cost offsetted by stmts that are kept live
10672 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10673 for computation of scalar cost.
10675 2013-05-28 Steve Ellcey <sellcey@mips.com>
10677 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10679 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10681 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10682 array notation built-in reduction functions.
10683 * doc/passes.texi (Passes): Added documentation about changes done
10685 * doc/invoke.texi (C Dialect Options): Added documentation about
10686 the -fcilkplus flag.
10687 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10688 (BUILTINS_DEF): Depend on cilkplus.def.
10689 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10690 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10691 * cilkplus.def: New file.
10693 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
10695 PR rtl-optimization/57439
10696 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10698 2013-05-28 Easwaran Raman <eraman@google.com>
10700 PR tree-optimization/57337
10701 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10702 (find_insert_point): Correctly identify the insertion point
10703 when two statements with the same UID is compared.
10705 2013-05-28 Richard Biener <rguenther@suse.de>
10707 PR tree-optimization/56787
10708 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10709 from the list of data references.
10710 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10712 (vect_analyze_loop_operations): Likewise.
10713 (vect_transform_loop): Remove clobbers.
10715 2013-05-28 Martin Jambor <mjambor@suse.cz>
10717 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10718 and REALPART_EXPRs have scalar type.
10720 2013-05-28 Richard Biener <rguenther@suse.de>
10722 PR tree-optimization/57411
10723 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10725 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10726 virtual operand propagation.
10728 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10730 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10731 destination register for bmasksi_vis.
10732 (vector_init_bshuffle): Likewise.
10733 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10735 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10737 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10738 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10739 mode if the instruction isn't available in the original mode.
10740 * config/sparc/sparc.opt (mfix-ut699): New option.
10741 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10742 (divdf3): Turn into expander.
10743 (divdf3_nofix): New insn.
10744 (divdf3_fix): Likewise.
10745 (divsf3): Disable if -mfix-ut699.
10746 (sqrtdf2): Turn into expander.
10747 (sqrtdf2_nofix): New insn.
10748 (sqrtdf2_fix): Likewise.
10749 (sqrtsf2): Disable if -mfix-ut699.
10751 2013-05-27 Richard Biener <rguenther@suse.de>
10753 PR middle-end/57412
10754 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10755 block for the new loop.
10757 2013-05-27 Richard Biener <rguenther@suse.de>
10759 PR tree-optimization/57343
10760 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10761 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10762 (number_of_iterations_cond): Do not build the folded tree.
10764 2013-05-27 Richard Biener <rguenther@suse.de>
10767 PR middle-end/57381
10768 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10769 OEP_CONSTANT_ADDRESS_OF retained.
10771 PR tree-optimization/57417
10772 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10773 for unchanged base.
10774 (set_ssa_val_to): Compare addresses using
10775 get_addr_base_and_unit_offset.
10777 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
10779 PR rtl-optimization/56833
10780 * postreload.c (move2add_record_mode): New function.
10781 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10782 (move2add_use_add2_insn): Use move2add_record_sym_value.
10783 (move2add_use_add3_insn): Likewise.
10784 (reload_cse_move2add): Use move2add_valid_value_p and
10785 move2add_record_mode. Invalidate call-clobbered and REG_INC
10786 affected regs by setting reg_mode to VOIDmode.
10787 (move2add_note_store): Don't pretend the inside of a SUBREG is
10788 the actual destination. Invalidate single/leading registers by
10789 setting reg_mode to VOIDmode.
10790 Use move2add_record_sym_value, move2add_valid_value_p and
10791 move2add_record_mode.
10793 2013-05-27 Richard Biener <rguenther@suse.de>
10795 PR tree-optimization/57396
10796 * tree-affine.c (double_int_constant_multiple_p): Properly
10797 return false for val == 0 and div != 0.
10799 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10801 * config/mips/mips.h: Use #elif in preprocessor conditions.
10803 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10806 * config/mips/constraints.md (kl): New constraint.
10807 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10808 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10809 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10810 from LO for MIPS16.
10811 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10813 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10816 * config/mips/mips.c (mips_can_inline_p): New function.
10817 (TARGET_CAN_INLINE_P): Define.
10819 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10821 * sched-int.h (ds_t, dw_t): Make unsigned int.
10822 Fix documentation that describes how all the ds_t bits are used.
10823 Reserve the last bit for delayed-branch scheduling.
10824 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10825 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10826 (gen_dep_weak_1): Remove prototype.
10827 * sched-deps.c (get_dep_weak_1): Make static.
10828 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10829 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10830 * doc/tm.texi: Regenerate.
10831 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
10833 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
10836 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
10838 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
10839 Sandra Loosemore <sandra@codesourcery.com>
10841 * config.gcc (powerpc-*): Allow native for with-cpu.
10843 2013-05-24 Jeff Law <law@redhat.com>
10845 PR tree-optimization/57124
10846 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
10847 conversion feeding a condition if the range has an overflow
10848 if -fstrict-overflow. Add warnings for when we do make the
10851 2013-05-24 Dehao Chen <dehao@google.com>
10853 * tree-cfg.c (locus_discrim_map): Fix the typo.
10854 (locus_discrim_hasher): Likewise.
10855 (locus_discrim_hasher::hash): Likewise.
10856 (locus_discrim_hasher::equal): Likewise.
10858 2013-05-24 Martin Jambor <mjambor@suse.cz>
10860 PR tree-optimization/57294
10861 * cgraph.h (ipa_record_stmt_references): Declare.
10862 * cgraphbuild.c (ipa_record_stmt_references): New function.
10863 (build_cgraph_edges): Use ipa_record_stmt_references.
10864 (rebuild_cgraph_edges): Likewise.
10865 (cgraph_rebuild_references): Likewise.
10866 * ipa-prop.c (ipa_modify_call_arguments): Discard references
10867 associated with the old statement and build references from the
10868 newly built statements.
10869 * ipa-ref.c (ipa_remove_stmt_references): New function.
10870 * ipa-ref.h (ipa_remove_stmt_references): Declare.
10872 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
10874 * lra-constraints.c (emit_spill_move): Use smaller mode for
10876 (check_and_process_move): Consider mem-reg moves for secondary
10878 (curr_insn_transform): Don't lose insns emitted before for
10879 secondary memory moves.
10880 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
10881 reg set up in the current insn.
10883 2013-05-24 Dehao Chen <dehao@google.com>
10885 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
10887 (locus_descrim_hasher::equal): Likewise.
10888 (build_gimple_cfg): New discriminator assignment algorithm.
10889 (make_edges): Likewise.
10890 (next_discriminator_for_locus): Likewise.
10891 (same_line_p): Likewise.
10892 (assign_discriminators): Likewise.
10893 (make_cond_expr_edges): Likewise.
10894 (make_gimple_switch_edges): Likewise.
10895 (make_goto_expr_edges): Likewise.
10896 (make_gimple_asm_edges): Likewise.
10898 2013-05-24 Ian Bolton <ian.bolton@arm.com>
10900 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
10901 X format specifier to only display bottom 16 bits.
10902 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
10903 immediate to match for operand 2, since it will be masked.
10905 2013-05-24 Richard Biener <rguenther@suse.de>
10907 PR tree-optimization/57287
10908 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
10909 all SSA names that occur in abnormal PHIs.
10911 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
10914 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
10915 that index is not negative.
10917 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
10919 PR rtl-optimization/55177
10920 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
10921 (simplify_byte_swapping_operation): New.
10922 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
10923 (simplify_relational_operation_1): Deal with BSWAP.
10925 2013-05-23 Richard Henderson <rth@redhat.com>
10928 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
10929 (ix86_reorg): Call it.
10931 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
10934 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
10935 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
10936 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
10938 2013-05-23 Christian Bruel <christian.bruel@st.com>
10941 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
10943 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10944 Marcus Shawcroft <marcus.shawcroft@arm.com>
10946 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
10947 * config/aarch64/constraints.md (Usa): Remove.
10948 * doc/md.texi (AArch64 Usa): Remove.
10950 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10951 Marcus Shawcroft <marcus.shawcroft@arm.com>
10953 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
10954 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
10955 * config/aarch64/predicates.md (aarch64_const_address): Remove.
10956 (aarch64_mov_operand): Use aarch64_mov_operand_p.
10958 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
10960 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
10961 instruction (AdvSIMD).
10962 * config/aarch64/aarch64-builtins.c
10963 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
10964 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
10966 2013-05-23 Martin Jambor <mjambor@suse.cz>
10968 PR middle-end/57347
10969 * tree.h (contains_bitfld_component_ref_p): Declare.
10970 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
10971 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
10973 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
10974 not access a bit-field. Assert all final offsets are byte-aligned.
10976 2013-05-23 Richard Biener <rguenther@suse.de>
10978 PR tree-optimization/57380
10979 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
10980 least one invariant or re-used load.
10981 * passes.c (init_optimization_passes): Move pass_phiprop before
10984 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
10986 * config/aarch64/aarch64-simd.md
10987 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
10989 2013-05-23 Richard Biener <rguenther@suse.de>
10991 PR middle-end/57381
10992 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10993 OEP_CONSTANT_ADDRESS_OF retained.
10995 2013-05-23 Jakub Jelinek <jakub@redhat.com>
10997 PR middle-end/57344
10998 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
10999 don't lower unit. Handle unit not being always BITS_PER_WORD.
11001 2013-05-23 Richard Biener <rguenther@suse.de>
11003 PR rtl-optimization/57341
11004 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
11005 instead of true_dependence.
11007 2013-05-22 David Malcolm <dmalcolm@redhat.com>
11009 * bb-reorder.c (branch_threshold): Make const.
11010 (exec_threshold): Ditto.
11012 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11013 Pat Haugen <pthaugen@us.ibm.com>
11014 Peter Bergner <bergner@vnet.ibm.com>
11016 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
11017 documentation for the power8 crypto builtins.
11019 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
11021 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
11022 macros for defining power8 builtin functions.
11023 (BU_P8V_AV_2): Likewise.
11024 (BU_P8V_AV_P): Likewise.
11025 (BU_P8V_VSX_1): Likewise.
11026 (BU_P8V_OVERLOAD_1): Likewise.
11027 (BU_P8V_OVERLOAD_2): Likewise.
11028 (BU_CRYPTO_1): Likewise.
11029 (BU_CRYPTO_2): Likewise.
11030 (BU_CRYPTO_3): Likewise.
11031 (BU_CRYPTO_OVERLOAD_1): Likewise.
11032 (BU_CRYPTO_OVERLOAD_2): Likewise.
11033 (XSCVSPDP): Fix typo, point to the correct instruction.
11034 (VCIPHER): Add power8 crypto builtins.
11035 (VCIPHERLAST): Likewise.
11036 (VNCIPHER): Likewise.
11037 (VNCIPHERLAST): Likewise.
11038 (VPMSUMB): Likewise.
11039 (VPMSUMH): Likewise.
11040 (VPMSUMW): Likewise.
11041 (VPERMXOR_V2DI): Likewise.
11042 (VPERMXOR_V4SI: Likewise.
11043 (VPERMXOR_V8HI: Likewise.
11044 (VPERMXOR_V16QI: Likewise.
11045 (VSHASIGMAW): Likewise.
11046 (VSHASIGMAD): Likewise.
11047 (VPMSUM): Likewise.
11048 (VPERMXOR): Likewise.
11049 (VSHASIGMA): Likewise.
11051 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11052 __CRYPTO__ if the crypto instructions are available.
11053 (altivec_overloaded_builtins): Add support for overloaded power8
11056 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
11057 support for power8 crypto builtins.
11058 (builtin_function_type): Likewise.
11059 (altivec_init_builtins): Add support for builtins that take vector
11060 long long (V2DI) arguments.
11062 * config/rs6000/crypto.md: New file, define power8 crypto
11065 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11066 Pat Haugen <pthaugen@us.ibm.com>
11067 Peter Bergner <bergner@vnet.ibm.com>
11069 * doc/invoke.texi (Option Summary): Add power8 options.
11070 (RS/6000 and PowerPC Options): Likewise.
11072 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
11073 constraints.md instead of rs6000.h. Reorder w* constraints. Add
11074 wm, wn, wr documentation.
11076 * config/rs6000/constraints.md (wm): New constraint for VSX
11077 registers if direct move instructions are enabled.
11078 (wn): New constraint for no registers.
11079 (wq): New constraint for quad word even GPR registers.
11080 (wr): New constraint if 64-bit instructions are enabled.
11081 (wv): New constraint if power8 vector instructions are enabled.
11082 (wQ): New constraint for quad word memory locations.
11084 * config/rs6000/predicates.md (const_0_to_15_operand): New
11085 constraint for 0..15 for crypto instructions.
11086 (gpc_reg_operand): If VSX allow registers in VSX registers as well
11087 as GPR and floating point registers.
11088 (int_reg_operand): New predicate to match only GPR registers.
11089 (base_reg_operand): New predicate to match base registers.
11090 (quad_int_reg_operand): New predicate to match even GPR registers
11091 for quad memory operations.
11092 (vsx_reg_or_cint_operand): New predicate to allow vector logical
11093 operations in both GPR and VSX registers.
11094 (quad_memory_operand): New predicate for quad memory operations.
11095 (reg_or_indexed_operand): New predicate for direct move support.
11097 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
11098 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
11099 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
11100 (POWERPC_MASKS): Add power8 options.
11101 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
11104 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11105 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
11107 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
11108 (-mpower8-fusion): New power8 options.
11109 (-mpower8-fusion-sign): Likewise.
11110 (-mpower8-vector): Likewise.
11111 (-mcrypto): Likewise.
11112 (-mdirect-move): Likewise.
11113 (-mquad-memory): Likewise.
11115 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
11116 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
11118 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
11119 (rs6000_debug_vector_unit): Add p8_vector.
11120 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
11121 definitions. Also print fusion state.
11122 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
11123 (rs6000_builtin_mask_calculate): Add power8 builtin support.
11124 (rs6000_option_override_internal): Add support for power8.
11125 (rs6000_common_init_builtins): Add debugging for skipped builtins
11126 if -mdebug=builtin.
11127 (rs6000_adjust_cost): Add power8 support.
11128 (rs6000_issue_rate): Likewise.
11129 (insn_must_be_first_in_group): Likewise.
11130 (insn_must_be_last_in_group): Likewise.
11131 (force_new_group): Likewise.
11132 (rs6000_register_move_cost): Likewise.
11133 (rs6000_opt_masks): Likewise.
11135 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
11136 power8 capable assembler, default to power7 options.
11137 (TARGET_DIRECT_MOVE): Likewise.
11138 (TARGET_CRYPTO): Likewise.
11139 (TARGET_P8_VECTOR): Likewise.
11140 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
11141 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
11142 (VECTOR_MEM_P8_VECTOR_P): Likewise.
11143 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
11144 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
11145 (TARGET_XSCVDPSPN): Likewise.
11146 (TARGET_XSCVSPDPN): Likewsie.
11147 (TARGET_SYNC_HI_QI): Likewise.
11148 (TARGET_SYNC_TI): Likewise.
11149 (MASK_CRYPTO): Likewise.
11150 (MASK_DIRECT_MOVE): Likewise.
11151 (MASK_P8_FUSION): Likewise.
11152 (MASK_P8_VECTOR): Likewise.
11153 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
11154 temporary used by some of the direct move instructions to get two FP
11155 temporary registers does not force creation of a stack frame.
11156 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
11157 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
11158 that any VSX registers are tieable, even if they are also an
11159 Altivec vector mode.
11160 (r6000_reg_class_enum): Add wm, wr, wv constraints.
11161 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
11162 (RS6000_BTM_CRYPTO): Likewise.
11163 (RS6000_BTM_COMMON): Likewise.
11165 * config/rs6000/rs6000.md (cpu attribute): Add power8.
11166 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
11167 (enum rs6000_vector): Add power8 vector support.
11169 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11173 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
11174 (any_sibcall_could_use_r3): this and handle indirect calls.
11175 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
11177 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11179 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
11181 2013-05-22 Richard Biener <rguenther@suse.de>
11183 PR middle-end/57349
11184 * profile.c (branch_prob): Do not split blocks that are
11185 abnormally receiving from ECF_RETURNS_TWICE functions.
11187 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11189 * recog.c (offsettable_address_addr_space_p): Fix calculation of
11190 address mode. Move pointer mode initialization to the same place.
11192 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11194 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
11195 while it has any effect.
11197 2013-05-21 Easwaran Raman <eraman@google.com>
11199 PR tree-optimization/57322
11200 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
11201 UID of the statement added to the BB to be 1.
11203 2013-05-21 Jakub Jelinek <jakub@redhat.com>
11205 PR tree-optimization/57331
11206 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
11207 of conversion from pointer type to integral type with integer.
11209 2013-05-21 Martin Jambor <mjambor@suse.cz>
11212 * ipa-prop.c (ipa_read_node_info): Process param_used and
11213 controlled_uses in the same order as when writing.
11215 2013-05-21 Magnus Granberg <baldrick@free.fr>
11218 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
11220 2013-05-21 Richard Biener <rguenther@suse.de>
11222 PR tree-optimization/57318
11223 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
11224 estimate stmts with side-effects as likely eliminated.
11226 2013-05-21 Richard Biener <rguenther@suse.de>
11228 PR tree-optimization/57330
11229 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
11230 preserve the call stmts fntype.
11232 2013-05-21 Richard Biener <rguenther@suse.de>
11234 PR tree-optimization/57303
11235 * tree-ssa-sink.c (statement_sink_location): Improve killing
11236 stmt detection and properly handle self-assignments.
11238 2013-05-21 Christian Bruel <christian.bruel@st.com>
11240 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
11241 spanning registers. LEAF_REG_REMAP is supported only for contiguous
11242 registers. Set register size out of the PARALLEL loop.
11244 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
11247 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
11248 (*fmasf4, *fmasf4_media): New insns.
11250 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11252 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
11253 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
11254 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
11255 (mips_idiv_insns): Update the comments to say that the returned
11256 instruction counts are in units of BASE_INSN_LENGTH.
11257 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
11258 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
11259 using 2 rather than 4 as the length of indirect MIPS16 and
11260 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
11261 length of a NOP. Don't divide MIPS16 lengths by 2.
11262 (mips16_split_long_branches): Assume a branch is long if the
11263 length is greater than 4 rather than 8.
11264 * config/mips/mips.md (length): Give MIPS16 lengths directly,
11265 rather than multiplying them by 2. Multiply instruction counts
11266 by BASE_INSN_LENGTH rather than 4.
11267 (*jump_mips16, tls_get_tp_mips16_<mode>)
11268 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
11270 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11272 * config/mips/mips.md (extended_mips16): Remove branch case.
11273 (length): Remove duplicated extended_mips16 test.
11275 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11277 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
11279 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
11281 * recog.h (Recog_data): Rename to...
11282 (recog_data_d): ...this.
11283 (recog_data): Update accordingly.
11284 * recog.c (recog_data): Likewise.
11285 * reload.c (save_recog_data): Likewise.
11286 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
11287 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
11289 2013-05-17 Julian Brown <julian@codesourcery.com>
11291 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
11292 found in a REG_EQUAL note, invalidate it.
11294 2013-05-17 Easwaran Raman <eraman@google.com>
11296 * tree-ssa-reassoc.c (find_insert_point): New function.
11297 (insert_stmt_after): Likewise.
11298 (get_def_stmt): Likewise.
11299 (ensure_ops_are_available): Likewise.
11300 (not_dominated_by): Likewise.
11301 (rewrite_expr_tree): Do not move statements beyond what is
11302 necessary. Remove call to swap_ops_for_binary_stmt...
11303 (reassociate_bb): ... and move it here.
11304 (build_and_add_sum): Assign UIDs for new statements.
11305 (linearize_expr): Likewise.
11306 (do_reassoc): Renumber gimple statement UIDs.
11308 2013-05-17 Jan Hubicka <jh@suse.cz>
11310 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
11312 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
11313 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
11315 (output_weakrefs): Update.
11317 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
11318 Martin Jambor <mjambor@suse.cz>
11320 PR middle-end/57276
11321 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
11322 value that corresponds to the given aggval is found in values vector.
11324 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
11326 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
11327 sse, sse2, sse3, ssse3 and sse4a flags to options.
11329 2013-05-17 David Malcolm <dmalcolm@redhat.com>
11331 * gengtype-state.c: (s_expr_writer): New class, to handle
11332 prettifying of output layout of s-expressions.
11333 (state_writer): New class, to write out gtype.state.
11334 (state_written_type_count): Move this variable into member data of
11336 (s_expr_writer::s_expr_writer): New code: constructor for new class
11337 (state_writer::state_writer(): ditto
11338 (s_expr_writer::write_new_line): New function
11339 (s_expr_writer::write_any_indent): ditto
11340 (s_expr_writer::begin_s_expr): ditto
11341 (s_expr_writer::end_s_expr): ditto
11342 (write_state_fileloc): convert to method of state_writer...
11343 (state_writer:: write_state_fileloc): ...and use methods of
11344 s_expr_writer to write indentation into the gtype.state output file
11345 to visually represent the hierarchical structure of the list
11347 (write_state_fields): ditto, renaming to...
11348 (state_writer::write_state_fields)
11349 (write_state_a_string): ditto, renaming to...
11350 (state_writer::write_state_a_string)
11351 (write_state_string_option): ditto, renaming to...
11352 (state_writer::write_state_string_option)
11353 (write_state_type_option): ditto, renaming to...
11354 (state_writer::write_state_type_option)
11355 (write_state_nested_option): ditto, renaming to...
11356 (state_writer::write_state_nested_option)
11357 (write_state_option): ditto, renaming to...
11358 (state_writer::write_state_option)
11359 (write_state_options): ditto, renaming to...
11360 (state_writer::write_state_options)
11361 (write_state_lang_bitmap): ditto, renaming to...
11362 (state_writer::write_state_lang_bitmap)
11363 (write_state_version): ditto, renaming to...
11364 (state_writer::write_state_version)
11365 (write_state_scalar_type): ditto, renaming to...
11366 (state_writer::write_state_scalar_type)
11367 (write_state_string_type): ditto, renaming to...
11368 (state_writer::write_state_string_type)
11369 (write_state_undefined_type): ditto, renaming to...
11370 (state_writer::write_state_undefined_type)
11371 (write_state_struct_union_type): ditto, renaming to...
11372 (state_writer::write_state_struct_union_type)
11373 (write_state_struct_type): ditto, renaming to...
11374 (state_writer::write_state_struct_type)
11375 (write_state_user_struct_type): ditto, renaming to...
11376 (state_writer::write_state_user_struct_type)
11377 (write_state_lang_struct_type): ditto, renaming to...
11378 (state_writer::write_state_lang_struct_type)
11379 (write_state_param_struct_type): ditto, renaming to...
11380 (state_writer::write_state_param_struct_type)
11381 (write_state_pointer_type): ditto, renaming to...
11382 (state_writer::write_state_pointer_type)
11383 (write_state_array_type): ditto, renaming to...
11384 (state_writer::write_state_array_type)
11385 (write_state_gc_used): ditto, renaming to...
11386 (state_writer::write_state_gc_used)
11387 (write_state_common_type_content): ditto, renaming to...
11388 (state_writer::write_state_common_type_content)
11389 (write_state_type): ditto, renaming to...
11390 (state_writer::write_state_type)
11391 (write_state_pair_list): ditto, renaming to...
11392 (state_writer::write_state_pair_list)
11393 (write_state_pair): ditto, renaming to...
11394 (state_writer::write_state_pair)
11395 (write_state_typedefs): ditto, renaming to...
11396 (state_writer::write_state_typedefs)
11397 (write_state_structures): ditto, renaming to...
11398 (state_writer::write_state_structures)
11399 (write_state_param_structs): ditto, renaming to...
11400 (state_writer::write_state_param_structs)
11401 (write_state_variables): ditto, renaming to...
11402 (state_writer::write_state_variables)
11403 (write_state_srcdir): ditto, renaming to...
11404 (state_writer::write_state_srcdir)
11405 (write_state_files_list): ditto, renaming to...
11406 (state_writer::write_state_files_list)
11407 (write_state_languages): ditto, renaming to...
11408 (state_writer::write_state_languages)
11409 (write_state): create a state_writer instance and use it when
11410 writing out the state file
11412 2013-05-17 Mike Stump <mikestump@comcast.net>
11414 PR rtl-optimization/57304
11415 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
11416 accessing DF_REF_REAL_LOC.
11418 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11420 PR rtl-optimization/57281
11421 PR rtl-optimization/57300
11422 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
11423 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
11424 what the other splitter did if the registers are dead.
11426 2013-05-17 Richard Biener <rguenther@suse.de>
11428 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
11431 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11433 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
11436 2013-05-17 Marek Polacek <polacek@redhat.com>
11438 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
11439 length when doing non-zero store of storing '\0' to '\0'.
11441 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11443 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
11444 vect_external_def oprnd1 with loop_vinfo, try to emit
11445 optional cast, negation and and stmts on the loop preheader
11446 edge instead of into the pattern def seq.
11448 PR tree-optimization/57051
11449 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
11450 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
11452 2013-05-16 Nick Clifton <nickc@redhat.com>
11454 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
11455 (rl78_is_naked_func): New function.
11456 (rl78_expand_prologue): Skip prologue generation for naked functions.
11457 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
11458 * doc/extend.texi (naked): Add RL78 to the list of processors
11459 that supports this attribute.
11461 2013-05-16 Jeff Law <law@redhat.com>
11463 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
11465 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11467 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
11468 cache parameters using detect_caches_amd also for CYRIX,
11469 NSC and TM2 signatures.
11471 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11472 Dzianis Kahanovich <mahatma@eu.by>
11476 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11477 VIA/Centaur processors and determine their cache parameters
11478 using detect_caches_amd.
11480 2013-05-16 Teresa Johnson <tejohnson@google.com>
11482 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11483 (rtl_verify_edges): New function.
11484 (rtl_verify_bb_insns): Ditto.
11485 (rtl_verify_bb_pointers): Ditto.
11486 (rtl_verify_bb_insn_chain): Ditto.
11487 (rtl_verify_fallthru): Ditto.
11488 (rtl_verify_bb_layout): Ditto.
11489 (rtl_verify_flow_info_1): Outline checks into new functions.
11490 (rtl_verify_flow_info): Ditto.
11492 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
11494 * cfghooks.c (copy_bbs): Add update_dominance argument.
11495 * cfghooks.h (copy_bbs): Update prototype.
11496 * tree-cfg.c (gimple_duplicate_sese_region):
11497 Add update_dominance argument.
11498 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11499 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11500 gimple_duplicate_sese_region call.
11501 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11502 Update copy_bbs call.
11503 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11504 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11506 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11508 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11509 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11510 vect_recog_rotate_pattern.
11511 (vect_recog_rotate_pattern): New function.
11513 2013-05-16 Jason Merrill <jason@redhat.com>
11515 * Makefile.in (LLINKER): New variable.
11516 (mostlyclean): Remove link mutex.
11517 * configure.ac: Handle --enable-link-mutex.
11518 * lock-and-run.sh: New script.
11520 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11523 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11526 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11528 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11530 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11532 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11533 * config/arm/arm.c (next_consecutive_mem): New function.
11534 (gen_movmem_ldrd_strd): Likewise.
11535 * config/arm/arm.md (movmemqi): Update condition and code.
11536 (unaligned_loaddi, unaligned_storedi): New patterns.
11538 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11540 * config.gcc: Obsolete *-*-solaris2.9*.
11541 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11543 2013-05-16 Richard Biener <rguenther@suse.de>
11545 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11546 earlier, after GRAPHITE transforms and IV canonicalization.
11548 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11550 * omp-low.c (extract_omp_for_data): For collapsed loops,
11551 if at least one of the loops is known at compile time to
11552 iterate zero times, set count to 0.
11553 (expand_omp_regimplify_p): New function.
11554 (expand_omp_for_generic): For collapsed loops, if at least
11555 one of the loops isn't known to iterate at least once,
11556 add runtime check with setting count to 0.
11557 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11558 For unsigned types if it isn't known at compile time that
11559 the loop will iterate at least once, add runtime check to bypass
11560 the whole loop if initial condition isn't true.
11562 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11564 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11566 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
11568 PR middle-end/57286
11569 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
11570 transformations to avoid an infinite loop.
11572 2013-05-16 Marek Polacek <polacek@redhat.com>
11574 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
11576 2013-05-15 Leif Ekblad <leif@rdos.net>
11578 * config/i386/i386.c (ix86_decompose_address): Use
11579 DEFAULT_TLS_SEG_REG to access TLS segment register.
11580 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
11581 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
11582 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
11584 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
11587 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
11588 sibling calls to functions that would normally be lazily bound,
11589 unless $gp is call-clobbered.
11591 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
11593 * config/i386/i386.c (ix86_option_override_internal): Update
11594 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
11595 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
11596 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
11597 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
11599 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
11601 2013-05-15 Andreas Schwab <schwab@suse.de>
11603 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
11604 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
11607 2013-05-15 Teresa Johnson <tejohnson@google.com>
11609 * loop-unroll.c (report_unroll_peel): Check decision before
11610 emitting unroll/peel message.
11612 2013-05-15 Teresa Johnson <tejohnson@google.com>
11614 * function.h (has_bb_partition): New rtl_data flag.
11615 (bb_reorder_complete): Ditto.
11616 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
11617 instead of flag_reorder_blocks_and_partition.
11618 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
11619 with some enhancements.
11620 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
11621 * bb-reorder.c (connect_traces): Check for has_bb_partition
11622 instead of flag_reorder_blocks_and_partition.
11623 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
11624 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
11625 verify_hot_cold_block_grouping.
11626 (partition_hot_cold_basic_blocks): Set has_bb_partition.
11628 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11631 * config/arm/predicates.md (call_insn_operand): New predicate.
11632 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
11633 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
11634 if insn is not a tail call.
11635 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11637 * config/arm/arm.h (enum reg_class): New caller save register class.
11638 (REG_CLASS_NAMES): Likewise.
11639 (REG_CLASS_CONTENTS): Likewise.
11640 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11643 2013-05-15 Richard Biener <rguenther@suse.de>
11645 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11646 of MSG_OPTIMIZED_LOCATIONS.
11647 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11648 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11650 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11651 of MSG_OPTIMIZED_LOCATIONS.
11652 (execute_vect_slp): Likewise.
11653 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11654 (vect_create_cond_for_alias_checks): Likewise.
11655 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11656 (vect_recog_widen_mult_pattern): Likewise.
11657 (vect_recog_widen_sum_pattern): Likewise.
11658 (vect_recog_over_widening_pattern): Likewise.
11659 (vect_recog_widen_shift_pattern): Likewise.
11660 (vect_recog_vector_vector_shift_pattern): Likewise.
11661 (vect_recog_divmod_pattern): Likewise.
11662 (vect_recog_mixed_size_cond_pattern): Likewise.
11663 (vect_recog_bool_pattern): Likewise.
11664 (vect_pattern_recog_1): Likewise.
11666 2013-05-15 Martin Jambor <mjambor@suse.cz>
11668 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11669 non-functions to builtin_unreachable.
11670 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11671 correct when new direct edges were discovered.
11673 2013-05-15 Martin Jambor <mjambor@suse.cz>
11675 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11676 header, print symbol order instead of node uid, print more information
11677 about indirect edge targets.
11678 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11680 (ipa_make_edge_direct_to_target): Likewise.
11681 (remove_described_reference): Likewise.
11682 (propagate_controlled_uses): Likewise.
11683 (ipa_print_node_params): Also print symbol order.
11684 (ipcp_transform_function): Print symbol order instead of node uids.
11685 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11686 (cgraph_get_create_real_symbol_node): Likewise.
11687 * ipa-cp.c (print_lattice): Likewise.
11688 (print_all_lattices): Likewise.
11689 (determine_versionability): Likewise.
11690 (initialize_node_lattices): Likewise.
11691 (estimate_local_effects): Likewise.
11692 (update_profiling_info): Likewise.
11693 (create_specialized_node): Likewise.
11694 (perhaps_add_new_callers): Likewise.
11695 (decide_about_value): Likewise.
11696 (decide_whether_version_node): Likewise.
11697 (identify_dead_nodes): Likewise.
11698 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11699 (dump_inline_summary): Likewise.
11700 (estimate_node_size_and_time): Likewise.
11701 (inline_analyze_function): Likewise.
11702 * ipa-inline.c (report_inline_failed_reason): Likewise.
11703 (want_early_inline_function_p): Likewise.
11704 (edge_badness): Likewise.
11705 (update_edge_key): Likewise.
11706 (inline_small_functions): Likewise. Add dumping of order to two other
11708 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11709 instead of node uids.
11710 (propagate_pure_const): Likewise.
11711 (propagate_pure_const): Likewise.
11712 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11713 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11714 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11716 * tree-pretty-print.c (dump_function_header): Likewise.
11717 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11718 Print symbol order instead of node uids.
11720 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11722 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11723 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11725 2013-05-15 Richard Biener <rguenther@suse.de>
11727 PR tree-optimization/57275
11728 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11729 return value for fail to do runtime alias checks for gather loads.
11731 2013-05-15 Jan Hubicka <jh@suse.cz>
11735 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11736 weakrefs are not external.
11737 (lto_symtab_merge_decls): Fix thinko when dealing with
11738 non-lto_symtab decls.
11739 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11740 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11741 * varpool.c (dump_varpool_node): Dump more flags.
11743 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11745 * config/i386/i386.c (processor_alias_table): Add instruction
11746 FSGSBASE for AMD bdver3 architecture.
11748 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11750 * tree.c (warn_deprecated_use): Print file:line using locus color.
11751 * diagnostic.c (diagnostic_report_current_module): Print file:line
11752 and file:line:column using locus color.
11754 2013-05-14 Mike Stump <mikestump@comcast.net>
11756 * gdbinit.in: Add __null.
11758 2013-05-14 Mike Stump <mikestump@comcast.net>
11760 * recog.h: Rename struct recog_data to Recog_data.
11761 * recog.c: Likewise.
11762 * reload.c (can_reload_into): Likewise.
11763 * config/picochip/picochip.c: Likewise.
11765 2013-05-14 Mike Stump <mikestump@comcast.net>
11767 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11769 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11771 * resource.h (struct resources): Remove unch_memory member.
11772 (CLEAR_RESOURCE): Don't clear unch_memory.
11773 * resource.c (mark_referenced_resources): Don't set it.
11774 (mark_set_resources): Likewise.
11775 (mark_target_live_regs): Don't clear it.
11776 (init_resource_info): Likewise.
11777 * reorg.c (resource_conflicts_p): Don't compare it.
11778 (redundant_insn): Don't set it.
11780 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11782 * emit-rtl.c (next_label): Remove unused function.
11783 (skip_consecutive_labels, link_cc0_insns): Move to ...
11784 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11785 only place where these functions are used, and make them static.
11787 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11789 * fold-const.c (fold_negate_expr): Handle vectors.
11790 (fold_truth_not_expr): Make it static.
11791 (fold_invert_truthvalue): New static function.
11792 (invert_truthvalue_loc): Handle vectors. Do not call
11793 fold_truth_not_expr directly.
11794 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11795 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11796 (fold_comparison): Handle vector constants.
11797 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11798 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11799 * tree.h (fold_truth_not_expr): Remove declaration.
11801 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11803 * config/aarch64/aarch64-simd.md
11804 (aarch64_vcond_internal<mode>): Rename to...
11805 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11806 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11807 float modes. Clarify all iterator modes.
11808 (vcond<mode><mode>): Use new name for vcond expanders.
11809 (vcond<v_cmp_result><mode>): Likewise.
11810 (vcondu<mode><mode>: Likewise.
11811 * config/aarch64/iterators.md (VDQF_COND): New.
11813 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11816 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11817 variable for the shift amount. Check that we shift by non-negative
11820 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11823 * config/arm/arm.h (EPILOGUE_USES): Only return true
11824 for LR_REGNUM after epilogue_completed.
11826 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
11828 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
11829 is error_mark_node.
11831 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11834 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
11835 and Solaris 11+/x86 with gld.
11836 * configure: Regenerate.
11838 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11840 * expmed.c (expand_shift_1): Canonicalize rotates by
11841 constant bitsize / 2 to bitsize - 1.
11842 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
11843 case ROTATERT>: Likewise.
11846 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11848 * config/i386/i386.md (rotateinv): New code attr.
11849 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11850 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11851 roll $31, %eax, etc.
11853 2013-05-14 Richard Biener <rguenther@suse.de>
11855 PR middle-end/57235
11856 * tree-eh.c (sink_clobbers): Give up for successors with
11857 multiple predecessors and no virtual uses.
11859 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
11861 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
11862 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
11864 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11866 PR middle-end/57251
11867 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
11868 the case when both op0 and op1 have VOIDmode.
11870 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11872 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
11873 in multiply-accumulate mode.
11875 2013-05-13 Guozhi Wei <carrot@google.com>
11877 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
11879 2013-05-13 Kai Tietz <ktietz@redhat.com>
11882 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
11883 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
11884 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
11885 * config/i386/i386.c (ix86_option_override_internal): Likewise.
11886 (ix86_expand_prologue): Likewise.
11887 (ix86_expand_split_stack_prologue): Likewise.
11888 (legitimate_pic_address_disp_p): Likewise.
11889 (legitimize_pic_address): Likewise.
11890 (legitimize_tls_address): Likewise.
11891 (legitimize_pe_coff_symbol): Likewise.
11892 (output_pic_addr_const): Likewise.
11893 (construct_plt_address): Likewise.
11894 (ix86_expand_call): Likewise.
11895 (x86_output_mi_thunk): Likewise.
11896 (x86_function_profiler): Likewise.
11898 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
11900 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
11901 similar switch cases.
11902 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
11903 (aarch64_simd_mov_to_<mode>low): Delete.
11904 (aarch64_simd_mov_to_<mode>high): Delete.
11905 (move_lo_quad_<mode>): Add w<-r alternative.
11906 (aarch64_simd_move_hi_quad_<mode>): Likewise.
11907 (aarch64_simd_mov_from_*): Update type attribute.
11908 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
11911 2013-05-13 Jan Hubicka <jh@suse.cz>
11913 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
11914 * config/i386/i386.c (ix86_compute_frame_layout,
11915 ix86_expand_epilogue, emit_i387_cw_initialization,
11916 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
11917 ix86_local_alignment): Fix use of size/speed predicates.
11919 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11921 PR tree-optimization/45216
11922 PR tree-optimization/57157
11923 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
11924 the (-Y) & (B - 1) variant if OP is |.
11925 * expmed.c (expand_shift_1): For rotations by const0_rtx just
11926 return shifted. Use (-op1) & (prec - 1) as other_amount
11927 instead of prec - op1.
11929 2013-05-13 Martin Jambor <mjambor@suse.cz>
11931 PR middle-end/42371
11932 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
11933 (ipa_constant_data): New type.
11934 (ipa_jump_func): Use ipa_constant_data to hold information about
11935 constant jump functions.
11936 (ipa_get_jf_constant): Adjust to jump function type changes.
11937 (ipa_get_jf_constant_rdesc): New function.
11938 (ipa_param_descriptor): New field controlled_uses.
11939 (ipa_get_controlled_uses): New function.
11940 (ipa_set_controlled_uses): Likewise.
11941 * ipa-ref.h (ipa_find_reference): Declare.
11942 * ipa-prop.c (ipa_cst_ref_desc): New type.
11943 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
11945 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
11946 New parameter cs. Adjust all callers.
11947 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
11948 (remove_described_reference): New function.
11949 (jfunc_rdesc_usable): Likewise.
11950 (try_make_edge_direct_simple_call): Decrement controlled use count,
11951 attempt to remove reference if it hits zero.
11952 (combine_controlled_uses_counters): New function.
11953 (propagate_controlled_uses): Likewise.
11954 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
11955 (ipa_edge_duplication_hook): Duplicate reference descriptions.
11956 (ipa_print_node_params): Print described use counter.
11957 (ipa_write_jump_function): Adjust to jump function type changes.
11958 (ipa_read_jump_function): New parameter CS, pass it to
11959 ipa_set_jf_constant. Adjust caller.
11960 (ipa_write_node_info): Stream controlled use count
11961 (ipa_read_node_info): Likewise.
11962 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
11964 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
11965 count. Remove cloning-added reference if it reaches zero.
11966 * ipa-ref.c (ipa_find_reference): New function.
11968 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11970 * config/i386/i386.c (processor_target_table): Modified default
11971 alignment values for AMD BD and BT architectures.
11973 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
11975 * tree-vect-generic.c (uniform_vector_p): Move ...
11976 * tree.c (uniform_vector_p): ... here.
11977 * tree.h (uniform_vector_p): Declare it.
11978 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
11981 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11983 PR tree-optimization/57230
11984 * tree-ssa-strlen.c (handle_char_store): Record length for
11985 array store from STRING_CST.
11987 PR tree-optimization/57230
11988 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
11991 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
11993 * config/epiphany/epiphany.c (epiphany_init): Check size of
11994 NUM_MODES_FOR_MODE_SWITCHING.
11995 (epiphany_expand_prologue):
11996 Remove CONFIG_REGNUM initial value handling code.
11997 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
11998 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
11999 (emit_set_fp_mode, epiphany_mode_after): Likewise.
12000 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
12001 Don't return 1 for FP_MODE_NONE.
12002 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
12003 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
12004 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
12005 * config/epiphany/epiphany.md (save_config): New pattern.
12007 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
12009 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
12011 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
12013 * config/i386/i386.md (memory): Handle sseishft1.
12014 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
12015 (*vec_extractv2di_1): Ditto.
12017 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
12019 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
12022 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
12024 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
12025 Add mthumb/march=armv7-a multilib.
12026 Add mthumb/march=armv7-r multilib.
12027 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
12029 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
12031 * config/v850/t-rtems: Add more multilibs.
12033 2013-05-10 Richard Biener <rguenther@suse.de>
12035 PR tree-optimization/57214
12036 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
12037 not propagate from SSA names that occur in abnormal PHI nodes.
12039 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
12041 * stor-layout.c (element_precision): New function.
12042 * machmode.h (element_precision): Declare it.
12043 * tree.c (build_minus_one_cst): New function.
12044 (element_precision): Likewise.
12045 * tree.h (build_minus_one_cst): Declare new function.
12046 (element_precision): Likewise.
12047 * fold-const.c (operand_equal_p): Use element_precision.
12048 (fold_binary_loc): Handle vector types.
12049 * convert.c (convert_to_integer): Use element_precision.
12050 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
12053 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
12055 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
12056 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
12057 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
12058 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12059 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
12060 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
12061 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
12062 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12063 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
12064 (Uuw8): New constraints.
12065 (Usb4): Move into alphabetical order.
12066 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
12067 (sd8_operand, ub8_operand, uw8_operand): New predicates.
12068 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
12069 previously unnamed patterns.
12070 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
12071 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
12072 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
12073 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
12074 of set_attr_alternative/if_then_else. Use extended_mips16 instead
12075 of specific lengths.
12077 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12079 * config/i386/i386.md (rotateinv): New code attr.
12080 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12081 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12082 roll $31, %eax, etc.
12084 PR tree-optimization/45216
12085 PR tree-optimization/57157
12086 * tree-ssa-forwprop.c (simplify_rotate): New function.
12087 (ssa_forward_propagate_and_combine): Call it.
12089 2013-05-10 Richard Biener <rguenther@suse.de>
12091 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
12092 disable peeling when we version for aliasing.
12093 (vector_alignment_reachable_p): Honor explicit user alignment.
12094 (vect_supportable_dr_alignment): Likewise.
12095 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
12096 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
12097 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
12098 then peeling to arrange for the cost-model check to come first.
12100 2013-05-10 Alan Modra <amodra@gmail.com>
12102 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
12103 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
12104 * configure: Regenerate.
12106 2013-05-10 Alan Modra <amodra@gmail.com>
12109 * varasm.c (default_elf_select_section): Move !DECL_P check..
12110 (get_named_section): ..to here before calling get_section_name.
12112 (default_section_type_flags): Add DECL_P check.
12113 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
12114 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
12116 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
12118 * config/epiphany/epiphany.c (epiphany_expand_prologue):
12119 When using gen_stack_adjust_str with a register offset, add a
12120 REG_FRAME_RELATED_EXPR note.
12122 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
12124 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
12125 (*vec_extractv4si_zext_mem): Ditto.
12126 (*vec_extractv2di): Add 0->x and x->x alternatives.
12127 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
12128 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
12130 2013-05-09 Jason Merrill <jason@redhat.com>
12132 N3639 C++1y VLA support
12133 * gimplify.c (gimplify_vla_decl): Don't touch an existing
12136 * tree.c (build_constructor_va): New.
12137 * tree.h: Declare it.
12139 2013-05-09 Martin Jambor <mjambor@suse.cz>
12142 * gimple-fold.c (canonicalize_constructor_val): Call
12143 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
12145 2013-05-09 Jan Hubicka <jh@suse.cz>
12146 Richard Biener <rguenther@suse.de>
12149 * symtab.c (symtab_make_decl_local): Do not add private names.
12151 2013-05-09 Jan Hubicka <jh@suse.cz>
12154 * symtab.c (insert_to_assembler_name_hash): Handle clones.
12155 (unlink_from_assembler_name_hash): Likewise.
12156 (symtab_prevail_in_asm_name_hash, symtab_register_node,
12157 symtab_unregister_node, symtab_initialize_asm_name_hash,
12158 change_decl_assembler_name): Update.
12160 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
12162 * config/aarch64/aarch64.md: New movtf split.
12163 (*movtf_aarch64): Update.
12164 (aarch64_movdi_tilow): Handle TF modes and rename to
12165 aarch64_movdi_<mode>low.
12166 (aarch64_movdi_tihigh): Handle TF modes and rename to
12167 aarch64_movdi_<mode>high
12168 (aarch64_movtihigh_di): Handle TF modes and rename to
12169 aarch64_mov<mode>high_di
12170 (aarch64_movtilow_di): Handle TF modes and rename to
12171 aarch64_mov<mode>low_di
12172 (aarch64_movtilow_tilow): Remove spurious whitespace.
12173 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
12175 (aarch64_print_operand): Update.
12177 2013-05-09 Alan Modra <amodra@gmail.com>
12179 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
12181 * configure: Regenerate.
12183 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12185 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
12186 splitter preparation statements.
12187 * config/i386/sse.md (*vec_extract* splitters): Ditto.
12188 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
12191 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12193 * gimple-ssa-strength-reduction.c (count_candidates): Change
12194 return value to int.
12195 (analyze_candidates_and_replace): Change type of length to int.
12197 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12199 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
12200 (*vec_extract<mode>): Use VI12_128 mode iterator.
12201 (*vec_extract<mode>_mem): Ditto.
12202 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
12205 2013-05-08 Diego Novillo <dnovillo@google.com>
12210 2012-08-17 Diego Novillo <dnovillo@google.com>
12213 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
12214 * config.in: Regenerate.
12215 * configure: Regenerate.
12216 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
12218 2013-05-08 Jan Hubicka <jh@suse.cz>
12221 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
12222 * cgraph.h (symtab_node_base): Add unique_name.
12223 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12224 input_overwrite_node, input_varpool_node): Stream unique_name.
12225 * cgraphclones.c (cgraph_create_virtual_clone,
12226 cgraph_function_versioning): Set unique_name.
12227 * ipa.c (function_and_variable_visibility): Set unique_name.
12229 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12231 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
12232 (alloc_cand_and_find_basis): Restrict conditional candidate
12233 processing to CAND_MULTs.
12235 2013-05-08 Jan Hubicka <jh@suse.cz>
12238 lto-symtab.c (lto_symtab_symbol_p): New function.
12239 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
12240 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
12241 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
12242 Skip static symbols.
12244 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
12246 PR tree-optimization/57200
12247 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
12248 Only call inform if the preceding warning_at returns true.
12250 2013-05-07 Han Shen <shenhan@google.com>
12252 * cfgexpand.c (record_or_union_type_has_array_p): New function.
12253 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
12254 * common.opt (fstack-protector-strong): New option.
12255 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
12256 * doc/invoke.texi (Optimization Options): Document
12257 "-fstack-protector-strong".
12258 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
12260 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12262 * config/mips/mips.c (mips_machine_reorg2): Return 0.
12264 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
12266 * ira.c (update_equiv_regs): Add insn having equiv memory even if
12267 it is not lhs of the insn.
12268 (setup_reg_equiv): Remove insn having equiv memory which it is not
12270 * lra-constraints.c (process_address): Try to improve generation
12271 code for address base + disp.
12272 (lra_constraints): Make correct the code for checking insn setting
12273 up backward equivalence. Remove insn only if it is in the init
12275 * lra-eliminations.c (update_reg_eliminate): Change return value.
12276 (lra_eliminate): Use the result.
12278 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
12280 * config/i386/sse.md (ssescalarnummask): New mode attribute.
12281 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
12282 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
12283 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
12284 register target operands.
12285 (*vec_extractv8hi_sse2): New pattern.
12286 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
12287 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
12288 (*vec_extract<mode>_mem): New insn and split pattern.
12290 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
12292 * config/arm/arm.c (arm_asan_shadow_offset): New function.
12293 (TARGET_ASAN_SHADOW_OFFSET): Define.
12294 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
12295 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
12297 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12299 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
12300 (incr_vec_index): Return -1 if increment not found.
12301 (create_add_on_incoming_edge): Assert if increment not found.
12302 (record_increment): Limit number of increments recorded.
12303 (all_phi_incrs_profitable): Return false if an increment not found.
12304 (replace_profitable_candidates): Don't process increments that were
12306 (analyze_candidates_and_replace): Limit size of incr_vec.
12308 2013-05-07 Richard Biener <rguenther@suse.de>
12310 * calls.c (special_function_p): setjmp-like functions are leaf.
12311 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
12312 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
12314 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
12316 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
12317 (aarch64_simd_mov<mode>): New expander.
12318 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
12319 (aarch64_simd_mov_to_<mode>high): Likewise.
12320 (aarch64_simd_mov_from_<mode>low): Likewise.
12321 (aarch64_simd_mov_from_<mode>high): Likewise.
12322 (aarch64_dup_lane<mode>): Update.
12323 (aarch64_dup_lanedi): New instruction pattern.
12324 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
12325 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
12327 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12329 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
12330 (replace_mult_candidate): Remove unnecessary argument; remove
12331 unnecessary parameter from call to introduce_cast_before_cand.
12332 (replace_unconditional_candidate): Remove unnecessary parameter
12333 from call to replace_mult_candidate.
12334 (replace_conditional_candidate): Likewise.
12335 (insert_initializers): Use make_temp_ssa_name.
12336 (introduce_cast_before_cand): Remove unnecessary argument; use
12337 make_temp_ssa_name.
12338 (replace_one_candidate): Remove unnecessary argument; remove
12339 unnecessary parameter from calls to introduce_cast_before_cand.
12340 (replace_profitable_candidates): Remove unnecessary parameters
12341 from calls to replace_one_candidate.
12343 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12345 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
12346 phi def as possibly hiding a basis for a CAND_ADD whose operands
12347 have been commuted in the analysis.
12348 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
12350 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12352 * config/aarch64/aarch64.md
12353 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
12354 shift value between 0-4.
12356 2013-05-07 Richard Biener <rguenther@suse.de>
12358 * double-int.h (rshift): New overload.
12359 * double-int.c (rshift): New function.
12360 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
12361 (create_reference_ops_from_ref): Remove.
12362 (vn_reference_insert): Use shared ops for constructing the
12363 reference and copy it.
12365 2013-05-07 Richard Biener <rguenther@suse.de>
12367 PR middle-end/57190
12368 * tree-eh.c (sink_clobbers): Properly propagate
12369 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
12371 2013-05-07 Jakub Jelinek <jakub@redhat.com>
12373 PR tree-optimization/57149
12374 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
12375 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
12376 collect_phi_def_edges, execute_late_warn_uninitialized): Use
12377 uninit_undefined_value_p instead of ssa_undefined_value_p.
12380 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
12381 for modifier == EXPAND_INITIALIZER.
12383 2013-05-07 Anton Blanchard <anton@samba.org>
12385 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
12386 for powerpc64 little endian.
12387 * configure: Regenerate.
12389 2013-05-06 Graham Stott <grahams@btinternet.com>
12391 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
12392 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
12393 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
12394 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
12396 2013-05-06 Graham Stott <grahams@btinternet.com>
12398 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
12399 codes which allow non-lvalues.
12401 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
12403 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
12404 components are all 1s.
12405 (integer_minus_onep): New function.
12406 * tree.h (integer_minus_onep): Declare it.
12407 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
12408 integer_minus_onep instead of integer_all_onesp.
12410 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12413 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
12414 variations of these patterns.
12416 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12418 * config/i386/i386.md (isa): Add x64_sse4 member.
12419 (enabled): Handle x64_sse4.
12420 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
12421 instruction for 64bit SSE4_1 targets. Update insn attributes.
12422 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
12423 instruction for SSE4_1 targets. Update insn attributes.
12424 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12425 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
12426 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
12428 (*vec_extractv4si): Rename from *sse4_1_pextrd.
12429 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
12430 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
12432 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12435 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
12437 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
12439 * final.c (do_assembler_dialects): Don't handle curly braces and
12440 vertical bar escaped by % as dialect delimiters.
12441 (output_asm_insn): Print curly braces and vertical bar if escaped
12442 by % and ASSEMBLER_DIALECT defined.
12443 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
12444 * doc/tm.texi: Regenerated.
12446 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12448 * config/mips/mips.c: Include tree-pass.h.
12449 (mips_reorg): Split in pre- and post-dbr_schedule parts.
12450 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
12451 (pass_mips_machine_reorg2): New machine specific pass.
12452 (insert_pass_mips_machine_reorg2): New pass plugin definition.
12453 (mips_option_override): Register the new pass.
12454 * rtl.h (cleanup_barriers): Remove prototype.
12455 (dbr_schedule): Likewise.
12456 * jump.c (cleanup_barriers): Make static.
12457 * reorg.c (dbr_schedule): Likewise.
12459 2013-05-06 Richard Biener <rguenther@suse.de>
12461 PR tree-optimization/57185
12462 * tree-parloops.c (add_field_for_reduction): Handle anonymous
12463 SSA names properly.
12465 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12468 * config/i386/i386.c (add_parameter_dependencies): Add dependence
12469 between "first_arg" and "insn", not "last" and "insn".
12471 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12473 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
12474 (find_candidates_in_block): Re-enable slsr_process_phi.
12475 (create_phi_basis): Fix double counting of candidate adjustment.
12477 2013-05-06 Richard Biener <rguenther@suse.de>
12479 PR middle-end/57147
12480 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12481 the edge is also fallthru, preserve it and just clear the
12483 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12484 also complex, preserve that and just clear the fallthru flag.
12485 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12486 update virtual operands.
12488 2013-05-06 Alan Modra <amodra@gmail.com>
12490 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12491 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12492 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12493 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12494 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12495 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12497 2013-05-06 Alan Modra <amodra@gmail.com>
12499 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12500 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12501 (DEFAULT_ASM_ENDIAN): Define.
12502 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12503 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12504 Update -K PIC clause from sysv4.h.
12505 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12506 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12508 2013-05-06 Alan Modra <amodra@gmail.com>
12510 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12511 twice for little-endian.
12512 (ashrdi3_no_power, ashrdi3): Support little-endian.
12514 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12517 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12518 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12520 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12521 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12522 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12525 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12526 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12528 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12529 * config/pa/pa.opt: Make mbig-switch a no-op.
12530 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12531 (CASE_VECTOR_MODE): Always return SImode.
12532 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12533 for the !TARGET_BIG_SWITCH case.
12534 * config/pa/pa-linux.h: Likewise.
12535 * config/pa/pa-openbsd.h: Likewise.
12536 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12537 * config/pa/pa.md (short_jump): Remove define_insn.
12538 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12539 (casesi0): Remove define_insn.
12540 (type): Remove btable_branch.
12541 (pa_combine_type): Likewise.
12542 (in_nullified_branch_delay): Likewise.
12543 (in_call_delay): Likewise.
12544 (define_delay): Likewise.
12545 (define_insn_reservation "Z3"): Likewise.
12546 (define_insn_reservation "Z4"): Likewise.
12547 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12548 (pa_adjust_insn_length): Remove adjustment for btable branches.
12549 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12552 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
12554 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12555 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12556 Add m->r,x alternatives.
12557 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12558 splitters using SWI48x mode iterator.
12559 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12560 TARGET_64BIT. Add m->x alternative.
12561 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12562 Add o->x alternative. Enable for TARGET_SSE.
12563 (sse_storeq): Remove expander.
12564 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12565 with memory input operand.
12566 (*vec_extractv2di_1 splitter): New.
12567 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12568 * config/i386/i386.md (ssevecmodelower): New mode attribute.
12570 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12572 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
12573 (INT_LOWPART): Delete.
12574 (extract_MB): Adjust.
12575 (extract_ME): Adjust.
12576 (print_operand): Adjust.
12578 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12580 * config/rs6000/predicates.md (reg_or_add_cint_operand,
12581 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12582 (reg_or_logical_cint_operand, easy_fp_constant,
12583 logical_const_operand): Delete "CONST_DOUBLE" case.
12584 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
12585 "HOST_BITS_PER_WIDE_INT == 64" test.
12586 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
12587 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
12588 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
12589 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
12591 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
12592 CONST_DOUBLE DImode/VOIDmode case.
12593 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
12594 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12595 CONST_DOUBLE VOIDmode case.
12596 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12597 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
12598 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
12599 Delete CONST_DOUBLE case.
12600 (splitters for mov FMOVE64 const_double): Delete
12601 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12602 "HOST_BITS_PER_WIDE_INT >= 64" test.
12603 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
12605 (mov DI const_double): Delete.
12607 2013-05-04 Jakub Jelinek <jakub@redhat.com>
12609 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
12610 on op shows all bits zero in mode of a lowpart subreg, return zero.
12612 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
12615 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
12616 to save TFmode registers and DImode to save TImode registers for
12617 caller save operations.
12618 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
12619 mark being partially clobbered since they only use the first
12622 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
12623 and TDmode only use the upper 64-bits of each VSX register.
12625 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12627 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
12628 (find_candidates_in_block): Disable slsr_process_phi.
12630 2013-05-03 Guozhi Wei <carrot@google.com>
12632 * coverage.c (coverage_obj_init): Move the construction of gcov
12634 (build_init_ctor): ... here.
12636 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12638 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12639 (slsr_cand_d): Redefine def_phi.
12640 (stride_status, phi_adjust_status, count_phis_status): New enums.
12641 (find_phi_def): New.
12642 (find_basis_for_base_expr): New.
12643 (find_basis_for_candidate): Handle hidden bases.
12644 (alloc_cand_and_find_basis): Handle phi candidates.
12645 (slsr_process_phi): New.
12646 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12647 (create_mul_imm_cand): Likewise.
12648 (create_add_ssa_cand): Exclude phi base candidates.
12649 (create_add_imm_cand): Likewise.
12650 (slsr_process_cast): Likewise.
12651 (slsr_process_copy): Likewise.
12652 (find_candidates_in_block): Handle phi candidates.
12653 (dump_candidate): Likewise.
12654 (unconditional_cands): Delete.
12655 (unconditional_cands_with_known_stride_p): Delete.
12656 (phi_dependent_cand_p): New.
12657 (cand_increment): Handle phi-dependent candidates.
12658 (replace_dependent): Delete.
12659 (replace_mult_candidate): New.
12660 (replace_unconditional_candidate): New.
12661 (incr_vec_index): Move to avoid forward reference.
12662 (create_add_on_incoming_edge): New.
12663 (create_phi_basis): New.
12664 (replace_dependents): Delete.
12665 (replace_conditional_candidate): New.
12666 (phi_add_costs): New.
12667 (replace_uncond_cands_and_profitable_phis): New.
12668 (record_increment): Handle phi adjustments.
12669 (record_phi_increments): New.
12670 (record_increments): Handle phi adjustments.
12671 (phi_incr_cost): New.
12672 (lowest_cost_path): Handle phis.
12673 (total_savings): Likewise.
12674 (analyze_increments): Likewise.
12675 (ncd_with_phi): New.
12676 (ncd_of_cand_and_phis): New.
12677 (nearest_common_dominator_for_cands): Handle phi increments.
12678 (all_phi_incrs_profitable): New.
12679 (replace_profitable_candidates): Handle phi-dependent candidates.
12680 (analyze_candidates_and_replace): Likewise.
12682 2013-05-03 Teresa Johnson <tejohnson@google.com>
12685 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12686 do not exceed REG_BR_PROB_BASE.
12688 2013-05-03 Jeff Law <law@redhat.com>
12690 PR tree-optimization/57144
12691 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12692 operand of the condition will bit into the new type when eliminating
12693 a cast feeding a condition.
12695 2013-05-03 Jakub Jelinek <jakub@redhat.com>
12697 PR rtl-optimization/57130
12698 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12699 of COMPARE as in_code to the recursive call if needed.
12701 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
12703 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12704 (enabled): Handle new members.
12705 * config/i386/sse.md (*vec_concatv2si): Merge from
12706 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12707 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12709 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
12711 PR tree-optimization/57027
12712 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12713 for fnms opportunity, check we got the prerequisite kind
12714 of tree / gimple before using accessor functions.
12716 2013-05-03 Richard Biener <rguenther@suse.de>
12718 * double-int.h (lshift): New overload without precision
12719 and arith argument.
12720 (operator *=, operator +=, operator -=): Move ...
12721 * double-int.c (operator *=, operator +=, operator -=): ... here
12722 and implement more efficiently.
12723 (mul_double_with_sign): Remove.
12724 (lshift_double): Adjust to take unsinged shift argument, push
12725 dispatching code to callers.
12726 (mul_double_wide_with_sign): Add early out for callers that
12727 are not interested in high parts or overflow.
12728 (lshift): New function.
12729 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12730 dispatch code here.
12731 (lrotate, rrotate): Use logical shifts.
12732 * expr.c (get_inner_reference): Use lshift.
12733 * fixed-value.c (do_fixed_divide): Likewise.
12734 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12735 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12736 (indirect_refs_may_alias_p): Likewise.
12737 (stmt_kills_ref_p_1): Likewise.
12739 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12741 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12743 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12745 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12746 scalar form of FABD instruction.
12748 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12750 * lra-constraints.c (process_alt_operands): Add checking alt
12751 number to choose the best alternative.
12753 2013-05-02 Richard Biener <rguenther@suse.de>
12755 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12756 bitmap and its handling.
12757 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12759 2013-05-02 Richard Biener <rguenther@suse.de>
12761 PR middle-end/57140
12762 * tree-inline.c (copy_loops): Properly handle removed loops.
12763 (copy_cfg_body): Mark destination loops for fixup if source
12764 loops needed fixup.
12766 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12769 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12770 generating simple_return for naked functions.
12772 2013-05-02 Martin Jambor <mjambor@suse.cz>
12774 PR middle-end/56988
12775 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12776 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12778 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12779 ipa_agg_replacement_value structures.
12780 (known_aggs_to_agg_replacement_list): Likewise.
12781 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12782 (read_agg_replacement_chain): Likewise.
12783 (ipcp_transform_function): Also check that by_ref flags match.
12785 2013-05-02 Richard Biener <rguenther@suse.de>
12787 * graphds.h (struct graph): Add obstack member.
12788 * graphds.c (new_graph): Initialize obstack and allocate
12790 (add_edge): Allocate edge from the obstack.
12791 (free_graph): Free the obstack instead of all edges and vertices.
12793 2013-05-02 Teresa Johnson <tejohnson@google.com>
12795 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12797 * cfg.c (update_bb_profile_for_threading): Ditto.
12798 * tree-inline.c (copy_bb): Ditto.
12799 (copy_edges_for_bb): Ditto.
12800 (initialize_cfun): Ditto.
12801 (copy_cfg_body): Ditto.
12802 (expand_call_inline): Ditto.
12803 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12804 (estimate_node_size_and_time): Ditto.
12805 (inline_merge_summary): Ditto.
12806 * cgraphclones.c (cgraph_clone_edge): Ditto.
12807 (cgraph_clone_node): Ditto.
12808 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12809 (compute_trg_info): Ditto.
12811 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12813 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12814 S reg when fp attribute set.
12815 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12817 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12819 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12821 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12822 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12823 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12825 2013-05-02 Richard Biener <rguenther@suse.de>
12827 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12828 (struct instantiate_cache_entry): New type.
12829 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12830 (struct instantiate_cache_type): New type.
12831 (set_instantiated_value, get_instantiated_value): Remove.
12832 (get_instantiated_value_entry): New function.
12833 (instantiate_scev_name): Use the new cache and adjust.
12834 (instantiate_scev_poly): Adjust.
12835 (instantiate_scev_binary): Likewise.
12836 (instantiate_array_ref): Likewise.
12837 (instantiate_scev_convert): Likewise.
12838 (instantiate_scev_not): Likewise.
12839 (instantiate_scev_3): Likewise.
12840 (instantiate_scev_2): Likewise.
12841 (instantiate_scev_r): Likewise.
12842 (instantiate_scev): Likewise.
12843 (resolve_mixers): Likewise.
12845 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
12848 * lra-constraints.c (best_small_class_operands_num): Remove.
12849 (process_alt_operands): Remove small_class_operands_num. Take
12850 small classes operands into losers and only if the operand is not
12851 matched. Modify debugging output.
12852 (curr_insn_transform): Remove best_small_class_operands_num.
12855 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12857 * config/aarch64/aarch64-builtins.c
12858 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
12859 * config/aarch64/aarch64-simd-builtins.def
12860 (reduc_splus_): Add new modes.
12861 (reduc_uplus_): New.
12862 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
12863 (reduc_uplus_v4sf): Likewise.
12864 (reduc_splus_v4sf): Likewise.
12865 (aarch64_addv<mode>): Likewise.
12866 (reduc_uplus_<mode>): Likewise.
12867 (reduc_splus_<mode>): Likewise.
12868 (aarch64_addvv2di): Likewise.
12869 (reduc_uplus_v2di): Likewise.
12870 (reduc_splus_v2di): Likewise.
12871 (aarch64_addvv2si): Likewise.
12872 (reduc_uplus_v2si): Likewise.
12873 (reduc_splus_v2si): Likewise.
12874 (reduc_<sur>plus_<mode>): New.
12875 (reduc_<sur>plus_v2di): Likewise.
12876 (reduc_<sur>plus_v2si): Likewise.
12877 (reduc_<sur>plus_v4sf): Likewise.
12878 (aarch64_addpv4sf): Likewise.
12879 * config/aarch64/arm_neon.h
12880 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
12881 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
12882 add UNSPEC_SADDV, UNSPEC_UADDV.
12884 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
12886 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12888 * config/aarch64/arm_neon.h
12889 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
12891 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12893 * config/aarch64/aarch64-builtins
12894 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
12896 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12898 * config/aarch64/aarch64-simd-builtins.def
12899 (reduc_smax_): New.
12900 (reduc_smin_): Likewise.
12901 (reduc_umax_): Likewise.
12902 (reduc_umin_): Likewise.
12903 (reduc_smax_nan_): Likewise.
12904 (reduc_smin_nan_): Likewise.
12907 (smax): Update for V2SF, V4SF and V2DF modes.
12910 (smin_nan): Likewise.
12911 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
12912 (<su><maxmin><mode>3): ...This, refactor.
12913 (s<maxmin><mode>3): New.
12914 (<maxmin_uns><mode>3): Likewise.
12915 (reduc_<maxmin_uns>_<mode>): Refactor.
12916 (reduc_<maxmin_uns>_v4sf): Likewise.
12917 (reduc_<maxmin_uns>_v2si): Likewise.
12918 (aarch64_<fmaxmin><mode>: Remove.
12919 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
12921 (vmin<q>_f<32,64>): Likewise.
12922 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12924 (su): Add mappings for smax, smin, umax, umin.
12926 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12927 (FMAXMIN): Rename as...
12928 (FMAXMIN_UNS): ...This.
12930 (fmaxminv): Likewise.
12931 (fmaxmin): Likewise.
12933 (maxmin_uns_op): Likewise.
12935 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12937 * config/aarch64/arm_neon.h
12938 (vac<ge, gt><sd>_f<32, 64>): Rename to...
12939 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
12940 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
12942 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12944 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
12945 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
12947 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12949 * config/aarch64/aarch64-simd.md
12950 (vcond<mode>_internal): Handle special cases for constant masks.
12951 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
12952 (vcondu<mode><mode>): Likewise.
12953 (vcond<v_cmp_result><mode>): New.
12955 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12957 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
12958 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
12959 * config/aarch64/aarch64-simd-builtins.def
12960 (cmeq): Update to BUILTIN_VALLDI.
12965 * config/aarch64/arm_neon.h
12966 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
12967 to builtins or C as appropriate.
12969 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12971 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
12973 (cmhi): Rename to...
12975 * config/aarch64/aarch64-simd.md
12976 (simd_mode): Add SF.
12977 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
12978 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
12979 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
12980 (cstore<mode>_neg): ...This.
12981 * config/aarch64/iterators.md
12983 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
12984 (COMPARISONS): New.
12985 (UCOMPARISONS): Likewise.
12986 (optab): Add missing comparisons.
12992 (VCMP_S): Likewise.
12993 (VCMP_U): Likewise.
12994 (V_cmp_result): Add DF, SF modes.
12995 (v_cmp_result): Likewise.
12997 (vmtype): Likewise.
12998 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
13000 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
13002 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
13003 define_insn to define_insn_and_split.
13004 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
13005 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
13006 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
13007 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
13008 (thumb2_negscc): Likewise.
13010 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13012 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
13014 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13016 * config/arm/thumb2.md: Remove trailing whitespaces.
13018 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13020 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
13021 Use gen_int_mode rather than GEN_INT.
13023 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
13025 * value-prof.c (stream_in_histogram_value): Remove the strayed
13028 2013-04-30 Richard Biener <rguenther@suse.de>
13030 PR middle-end/57122
13031 * cfghooks.c (split_edge): Properly check for the loop latch edge.
13033 2013-04-30 Richard Biener <rguenther@suse.de>
13035 PR middle-end/57107
13036 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
13038 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
13040 PR rtl-optimization/56957
13041 PR rtl-optimization/57105
13042 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
13043 variable. Use just INSN_UID for determining whether an insn
13044 should be only disconnected from the insn stream.
13045 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
13047 2013-04-30 Jakub Jelinek <jakub@redhat.com>
13049 PR tree-optimization/57104
13050 * tsan.c (instrument_expr): Don't instrument accesses to
13051 DECL_HARD_REGISTER VAR_DECLs.
13053 2013-04-30 Richard Biener <rguenther@suse.de>
13055 * function.h (loops_for_fn): New inline function.
13056 (set_loops_for_fn): Likewise.
13057 * cfgloop.h (place_new_loop): Add struct function parameter.
13058 (get_loop): Likewise.
13059 (get_loops): Likewise.
13060 (number_of_loops): Likewise.
13061 (fel_next): Adjust.
13062 (fel_init): Likewise.
13063 * cfg.c (get_loop_copy): Adjust.
13064 * cfgloop.c (flow_loops_dump): Likewise.
13065 (record_loop_exits): Likewise.
13066 (verify_loop_structure): Likewise.
13067 * cfgloopanal.c (mark_irreducible_loops): Likewise.
13068 (estimate_reg_pressure_cost): Likewise.
13069 (mark_loop_exit_edges): Likewise.
13070 * cfgloopmanip.c (place_new_loop): Likewise.
13071 (add_loop): Likewise.
13072 (duplicate_loop): Likewise.
13073 * graph.c (draw_cfg_nodes): Likewise.
13074 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
13075 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
13076 (extract_affine_chrec): Likewise.
13077 (build_scop_iteration_domain): Likewise.
13078 * graphite.c (graphite_initialize): Likewise.
13079 * ira-build.c (create_loop_tree_nodes): Likewise.
13080 (more_one_region_p): Likewise.
13081 (rebuild_regno_allocno_maps): Likewise.
13082 (mark_loops_for_removal): Likewise.
13083 (mark_all_loops_for_removal): Likewise.
13084 (remove_unnecessary_regions): Likewise.
13085 (ira_build): Likewise.
13086 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
13087 * loop-init.c (fix_loop_structure): Likewise.
13088 (gate_rtl_move_loop_invariants): Likewise.
13089 (gate_rtl_unswitch): Likewise.
13090 (gate_rtl_unroll_and_peel_loops): Likewise.
13091 (rtl_doloop): Likewise.
13092 * lto-streamer-in.c (input_cfg): Likewise.
13093 * lto-streamer-out.c (output_cfg): Likewise.
13094 * modulo-sched.c (sms_schedule): Likewise.
13095 * predict.c (tree_estimate_probability): Likewise.
13096 (tree_estimate_probability_driver): Likewise.
13097 (estimate_loops): Likewise.
13098 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
13099 (move_sese_region_to_fn): Likewise.
13100 (debug_loop_num): Likewise.
13101 * tree-chrec.c (chrec_evaluate): Likewise.
13102 (hide_evolution_in_other_loops_than_loop): Likewise.
13103 (chrec_component_in_loop_num): Likewise.
13104 (reset_evolution_in_loop): Likewise.
13105 (evolution_function_is_invariant_rec_p): Likewise.
13106 * tree-if-conv.c (main_tree_if_conversion): Likewise.
13107 * tree-inline.c (copy_loops): Likewise.
13108 (copy_cfg_body): Likewise.
13109 (tree_function_versioning): Likewise.
13110 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
13111 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
13113 (add_to_evolution_1): Likewise.
13114 (scev_const_prop): Likewise.
13115 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
13116 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
13117 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
13118 (tree_ssa_lim_initialize): Likewise.
13119 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
13120 (verify_loop_closed_ssa): Likewise.
13121 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
13122 (tree_ssa_loop_im): Likewise.
13123 (tree_ssa_loop_unswitch): Likewise.
13124 (tree_vectorize): Likewise.
13125 (check_data_deps): Likewise.
13126 (tree_ssa_loop_ivcanon): Likewise.
13127 (tree_ssa_loop_bounds): Likewise.
13128 (tree_complete_unroll): Likewise.
13129 (tree_complete_unroll_inner): Likewise.
13130 (tree_parallelize_loops): Likewise.
13131 (tree_ssa_loop_prefetch): Likewise.
13132 (tree_ssa_loop_ivopts): Likewise.
13133 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
13134 * tree-vectorizer.c (vectorize_loops): Likewise.
13136 2013-04-29 Mike Frysinger <vapier@gentoo.org>
13138 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
13139 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
13140 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
13141 with EABI_LINK_SPEC.
13143 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13146 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
13149 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
13152 * lra-constraints.c (process_alt_operands): Discourage a bit more
13153 using memory for pseudos. Print cost dump for alternatives.
13154 Modify cost values for conflicts with early clobbers.
13155 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
13157 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13160 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
13162 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13164 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
13165 from/to S register.
13166 (movdi_aarch64): Support LDR/STR from/to D register.
13168 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13170 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
13171 or higher by default.
13173 2013-04-29 Richard Biener <rguenther@suse.de>
13175 PR middle-end/57075
13176 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
13177 even if not adding abnormal edges for calls that can make
13180 2013-04-29 Richard Biener <rguenther@suse.de>
13182 PR middle-end/57103
13183 * tree-cfg.c (move_stmt_op): Fix condition under which to update
13185 (move_stmt_r): Remove redundant checking.
13187 2013-04-29 Teresa Johnson <tejohnson@google.com>
13190 * basic-block.h (apply_scale): New function.
13191 (apply_probability): Use apply_scale.
13192 * gimple-streamer-in.c (input_bb): Ditto.
13193 * lto-streamer-in.c (input_cfg): Ditto.
13194 * lto-cgraph.c (merge_profile_summaries): Ditto.
13195 * tree-optimize.c (execute_fixup_cfg): Ditto.
13196 * tree-inline.c (copy_bb): Update comment to use apply_scale.
13197 (copy_edges_for_bb): Ditto.
13198 (copy_cfg_body): Ditto.
13200 2013-04-29 Tom de Vries <tom@codesourcery.com>
13202 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
13203 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
13204 (tail_merge_optimize): Handle current_loops == NULL.
13206 2013-04-26 Jeff Law <law@redhat.com>
13208 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
13209 (simplify_cond_using_ranges): Generalize code to simplify
13210 COND_EXPRs where one argument is a constant and the other
13211 is an SSA_NAME created by an integral type conversion.
13213 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13215 * config/arm/arm.md (store_minmaxsi): Use only when
13216 optimize_insn_for_size_p.
13218 2013-04-29 Christian Bruel <christian.bruel@st.com>
13221 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
13223 2013-04-29 Richard Biener <rguenther@suse.de>
13225 PR middle-end/57089
13226 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
13227 loop tree make sure to schedule a fixup for the child as well.
13228 (expand_omp_for_generic): Properly add loops.
13229 (expand_omp_for_static_nochunk): Likewise.
13230 (expand_omp_for_static_chunk): Likewise.
13231 (expand_omp_for): For the degenerate case fixup loops.
13232 (expand_omp_sections): Fix default bb placement in loops.
13233 (expand_omp_atomic_pipeline): Properly add loops.
13235 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13237 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
13239 2013-04-29 Tom de Vries <tom@codesourcery.com>
13241 * tree-ssa-tail-merge.c: Update header comment.
13243 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13245 * config/aarch64/arm_neon.h
13246 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
13247 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
13248 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
13249 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
13250 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
13251 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
13252 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
13253 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
13255 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13257 * config/aarch64/aarch64-simd.md
13258 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
13259 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
13260 fix_trunc, fixuns_trunc.
13261 (ftrunc<VDQF:mode>2): New.
13262 * config/aarch64/iterators.md (optab): Add fix, fixuns.
13263 (fix_trunc_optab): New.
13265 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13267 * config/aarch64/aarch64-builtins.c
13268 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
13269 iceilf, lround, iroundf.
13271 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13274 * config/i386/i386.h (enum ix86_tune_indices)
13275 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
13276 New, split from X86_TUNE_INTER_UNIT_MOVES.
13277 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
13278 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
13279 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
13280 (TARGET_INTER_UNIT_MOVES): Remove.
13281 * config/i386/i386.c (initial_ix86_tune_features): Update.
13282 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
13283 (ix86_expand_convert_uns_didf_sse): Use
13284 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13285 (ix86_expand_vector_init_one_nonzero): Ditto.
13286 (ix86_expand_vector_init_interleave): Ditto.
13287 (inline_secondary_memory_needed): Return true for moves from SSE class
13288 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
13289 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13290 * config/i386/constraints.md (Yi, Ym): Depend on
13291 TARGET_INTER_UNIT_MOVES_TO_VEC.
13292 (Yj, Yn): New constraints.
13293 * config/i386/i386.md (*movdi_internal): Change constraints of
13294 operand 1 from Yi to Yj and from Ym to Yn.
13295 (*movsi_internal): Ditto.
13296 (*movdf_internal): Ditto.
13297 (*movsf_internal): Ditto.
13298 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
13299 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13300 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
13301 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
13302 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
13303 * config/i386/sse.md (movdi_to_sse): Ditto.
13304 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
13305 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
13306 TARGET_INTER_UNIT_MOVES.
13307 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
13308 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
13309 instead of TARGET_INTER_UNIT_MOVES.
13310 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
13311 operand 1 from Yi to Yj and from Ym to Yn.
13313 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13315 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
13316 (float_truncate_hi_): Likewise.
13317 (float_extend_lo_): Likewise.
13318 (float_truncate_lo_): Likewise.
13319 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
13320 (aarch64_float_extend_lo_v2df): Likewise.
13321 (vec_unpacks_hi_v4sf): Likewise.
13322 (aarch64_float_truncate_lo_v2sf): Likewise.
13323 (aarch64_float_truncate_hi_v4sf): Likewise.
13324 (vec_pack_trunc_v2df): Likewise.
13325 (vec_pack_trunc_df): Likewise.
13327 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13329 * config/aarch64/aarch64-builtins.c
13330 (aarch64_fold_builtin): Fold float conversions.
13331 * config/aarch64/aarch64-simd-builtins.def
13332 (floatv2si, floatv4si, floatv2di): New.
13333 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
13334 * config/aarch64/aarch64-simd.md
13335 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
13336 * config/aarch64/iterators.md (FLOATUORS): New.
13337 (optab): Add float, floatuns.
13338 (su_optab): Likewise.
13340 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13342 * config/aarch64/aarch64-builtins.c
13343 (aarch64_builtin_vectorized_function): Use new names for
13345 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
13346 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
13347 (fcvtzu): Split as...
13348 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
13349 (fcvtas): Split as...
13350 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
13351 (fcvtau): Split as...
13352 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
13353 (fcvtps): Split as...
13354 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
13355 (fcvtpu): Split as...
13356 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
13357 (fcvtms): Split as...
13358 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
13359 (fcvtmu): Split as...
13360 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
13361 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
13362 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
13363 (lfrintnusf, lfrintnudf): Likewise.
13364 * config/aarch64/aarch64-simd.md
13365 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
13367 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
13368 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
13369 (fcvt_pattern): Likewise.
13371 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13373 * config/aarch64/aarch64-simd.md
13374 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
13375 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
13377 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13379 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
13380 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
13381 (vrnd<a,m,n,p>_f32): Implement using builtins.
13382 (vrnd<i,x><q>_f<32, 64>): New.
13384 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13386 * config/aarch64/aarch64-builtins.c
13387 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
13388 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
13389 (frintz): Rename to...
13391 (frintp): Rename to...
13393 (frintm): Rename to...
13395 (frinti): Rename to...
13396 (nearbyint): ...this.
13397 (frintx): Rename to...
13399 (frinta): Rename to...
13401 * config/aarch64/aarch64-simd.md
13402 (aarch64_frint<frint_suffix><mode>): Delete.
13403 (<frint_pattern><mode>2): Convert to insn.
13404 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
13405 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
13406 (frint_pattern): Likewise.
13407 (frint_suffix): Likewise.
13409 2013-04-29 Richard Biener <rguenther@suse.de>
13411 PR tree-optimization/57081
13412 * loop-init.c: Include tree-flow.h.
13413 (loop_optimizer_finalize): Free number of iteration estimates.
13414 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
13416 2013-04-29 Jakub Jelinek <jakub@redhat.com>
13418 PR tree-optimization/57083
13419 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
13420 non-singleton shift count range, zero extend low_bound for uns case.
13422 * config/i386/predicates.md (general_vector_operand): New predicate.
13423 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
13424 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
13425 if they aren't nonimmediate operands. If their original values
13426 satisfy const_vector_equal_evenodd_p, don't shift them.
13427 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
13428 predicates. For the SSE4.1 case force operands[{1,2}] into registers
13429 if not nonimmediate_operand.
13430 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
13431 instead of register_operand.
13432 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
13434 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
13436 * stor-layout.c (finalize_size_functions): Allocate a structure and
13437 reset cfun before dumping the functions.
13439 2013-04-27 Jakub Jelinek <jakub@redhat.com>
13441 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
13444 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
13445 use xop_pmacsdqh if uns_p.
13446 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
13447 the immediate rotate count.
13449 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
13451 * rtl.h (struct rtx_def): Add comment for field jump.
13452 (LRA_SUBREG_P): New macro.
13453 * recog.c (register_operand): Check LRA_SUBREG_P.
13454 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13456 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13457 (lra_final_code_change): Skip subreg change for operators.
13458 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13459 if there are no operand changes.
13460 * lra-constraints.c (curr_insn_set): New.
13461 (match_reload): Set LRA_SUBREG_P.
13462 (emit_spill_move): Ditto.
13463 (check_and_process_move): Use curr_insn_set. Process only single
13464 set insns. Don't initialize sec_mem_p and change_p.
13465 (simplify_operand_subreg): Use LRA_SUBREG_P.
13466 (reg_in_class_p): New function.
13467 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13468 of #ifdef. Add code to remove cycling.
13469 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13470 non-null disp. Reload inner instead of disp when base and index
13471 are null. Try to put lo_sum into register.
13472 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13473 (check_and_process_move): Move code for move cost check to
13474 simple_move_p. Remove equiv_substitution.
13475 (simple_move_p): New function.
13476 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13477 curr_insn_set. Call check_and_process_move only for single set
13478 insns. Use the new function. Move call of check_and_process_move
13479 after operand equiv substitution and address process.
13481 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13484 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13485 with nonlocal goto receivers or returns twice calls, ignore
13486 unininitialized values from abnormal edges to nl goto receiver
13487 or returns twice call.
13489 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13491 PR tree-optimization/57051
13492 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13493 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13496 2013-04-26 Richard Biener <rguenther@suse.de>
13498 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13499 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13500 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13501 (fixup_loop_arrays_after_move): New function.
13502 (move_sese_region_to_fn): Properly outline the loop tree parts
13503 of the SESE region.
13505 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
13507 * config/i386/i386.md (type, unit): Fix long lines.
13509 2013-04-26 Richard Biener <rguenther@suse.de>
13511 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13512 (lto-streamer-out.o): Likewise.
13513 * cfgloop.c (init_loops_structure): Export, add struct function
13514 argument and adjust.
13515 (flow_loops_find): Adjust.
13516 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13517 (init_loops_structure): Declare.
13518 * lto-streamer-in.c: Include cfgloop.h.
13519 (input_cfg): Input the loop tree.
13520 * lto-streamer-out.c: Include cfgloop.h.
13521 (output_cfg): Output the loop tree.
13522 (output_struct_function_base): Do not drop PROP_loops.
13524 2013-03-26 Richard Biener <rguenther@suse.de>
13526 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13527 (pass_build_cfg): Provide PROP_loops.
13528 (move_sese_region_to_fn): Remove loops that are outlined into fn
13530 * tree-inline.c: Include cfgloop.h.
13531 (initialize_cfun): Do not drop PROP_loops.
13532 (copy_loops): New function.
13533 (copy_cfg_body): Copy loop structure.
13534 (tree_function_versioning): Initialize destination loop tree.
13535 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13536 (pass_parallelize_loops): Do IL verification.
13537 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13538 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13539 the CFG make sure we fixup loops as well.
13540 * tree-ssa-tail-merge.c: Include cfgloop.h.
13541 (replace_block_by): When merging loop latches mark loops for fixup.
13542 * lto-streamer-out.c (output_struct_function_base): Drop
13543 PROP_loops for now.
13544 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13545 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13546 * ipa-split.c: Include cfgloop.h.
13547 (split_function): Add the new return block to the loop tree root.
13548 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13549 whether we have removed the forwarder block.
13550 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13551 * cfgloop.h (place_new_loop): Declare.
13552 * cfgloopmanip.c (place_new_loop): Export.
13553 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13554 (tree-switch-conversion.o): Likewise.
13555 (tree-complex.o): Likewise.
13556 (tree-inline.o): Likewise.
13557 (tree-ssa-tailmerge.o): Likewise.
13558 (ipa-split.o): Likewise.
13559 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13560 (tree-ssa-copy.o): Likewise.
13561 * tree-switch-conversion.c: Include cfgloop.h
13562 (process_switch): If we emit a bit-test cascade, schedule loops
13564 * tree-complex.c: Include cfgloop.h.
13565 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13566 * asan.c: Include cfgloop.h.
13567 (create_cond_insert_point): Properly add new basic-blocks to
13568 loops, schedule loop fixup.
13569 * cfgloop.c (verify_loop_structure): Check that looks are not
13571 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
13573 (expand_omp_for_generic): Likewise.
13574 (expand_omp_sections): Likewise.
13575 (expand_omp_atomic_pipeline): Schedule loops for fixup.
13576 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
13577 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
13578 is initialized, not when loops are present.
13579 * tree-parloops.c (parallelize_loops): Remove checking here.
13580 * passes.c (init_optimization_passes): Schedule a copy-propagation
13581 pass before complete unrolling of inner loops.
13583 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13585 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
13586 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
13587 (colorize_init): Add argument to _WIN32 version.
13588 * toplev.c: Include diagnostic-color.h.
13589 (process_options): Default to -fdiagnostics-color=auto if
13590 GCC_COLORS env var is in the environment.
13591 * common.opt (fdiagnostics-color=): Add Var and Init.
13592 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
13593 env var is in the environment, the default is auto rather than never.
13595 * diagnostic.h (file_name_as_prefix): Add context argument.
13596 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
13597 the string as locus.
13598 * langhooks.c (lhd_print_error_function): Adjust caller.
13600 2013-04-25 Lawrence Crowl <crowl@google.com>
13602 * var-tracking.c (shared_hash_def::htab):
13603 Change type to hash_table. Update dependent calls and types.
13605 2013-04-25 Lawrence Crowl <crowl@google.com>
13607 * Makefile.in: Update as needed below.
13609 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
13610 Move declaration to after the type's method definitons.
13612 * attribs.c (htab_t scoped_attributes::attribute_hash):
13613 Change type to hash_table. Update dependent calls and types.
13615 * bitmap.c (htab_t bitmap_desc_hash):
13616 Change type to hash_table. Update dependent calls and types.
13618 * cselib.c (htab_t cselib_hash_table):
13619 Change type to hash_table. Update dependent calls and types.
13621 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
13622 (hash_string_slot_node): Move implementation into lto-streamer.h
13623 struct string_slot_hasher.
13624 (eq_string_slot_node): Likewise.
13626 * data-streamer-out.c: Update output_block::string_hash_table
13627 dependent calls and types.
13629 * dwarf2cfi.c (htab_t trace_index):
13630 Change type to hash_table. Update dependent calls and types.
13632 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
13633 Change type to hash_table. Update dependent calls and types.
13634 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
13635 (htab_t optimize_external_refs::map): Likewise.
13636 (htab_t output_comp_unit::extern_map): Likewise.
13637 (htab_t output_comdat_type_unit::extern_map): Likewise.
13638 (htab_t output_macinfo::macinfo_htab): Likewise.
13639 (htab_t optimize_location_lists::htab): Likewise.
13640 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13642 * except.c (htab_t ehspec_hash_type):
13643 Change type to hash_table. Update dependent calls and types.
13644 (assign_filter_values::ttypes): Likewise.
13645 (assign_filter_values::ehspec): Likewise.
13646 (sjlj_assign_call_site_values::ar_hash): Likewise.
13647 (convert_to_eh_region_ranges::ar_hash): Likewise.
13649 * gcse.c (htab_t pre_ldst_table):
13650 Change type to hash_table. Update dependent calls and types.
13652 * ggc-common.c (htab_t saving_htab):
13653 Change type to hash_table. Update dependent calls and types.
13654 (htab_t loc_hash): Likewise.
13655 (htab_t ptr_hash): Likewise.
13656 (call_count): Rename ggc_call_count.
13657 (call_alloc): Rename ggc_call_alloc.
13658 (loc_descriptor): Rename make_loc_descriptor.
13659 (add_statistics): Rename ggc_add_statistics.
13661 * ggc-common.c (saving_htab):
13662 Change type to hash_table. Update dependent calls and types.
13664 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13665 (push_gimplify_context): Likewise.
13666 (pop_gimplify_context): Likewise.
13667 (struct gimple_temp_hash_elt): Added.
13668 (struct gimplify_hasher): Likewise.
13669 (struct gimplify_ctx.temp_htab):
13670 Change type to hash_table. Update dependent calls and types.
13672 * gimple-fold.c: Include gimplify-ctx.h.
13674 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13675 Change type to hash_table. Update dependent calls and types.
13676 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13677 avoid potential global name collision.
13679 * gimplify.c: Include gimplify-ctx.h.
13680 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13681 (htab_t gimplify_ctx::temp_htab):
13682 Update dependent calls and types for new type hash_table.
13683 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13684 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13686 * gimplify-ctx.h: New.
13687 (struct gimple_temp_hash_elt): Move from gimplify.c.
13688 (class gimplify_hasher): New.
13689 (struct gimplify_ctx): Move from gimple.h.
13690 (htab_t gimplify_ctx::temp_htab):
13691 Change type to hash_table. Update dependent calls and types.
13693 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13694 (htab_t ivs_params::newivs_index):
13695 Change type to hash_table. Update dependent calls and types.
13696 (htab_t ivs_params::params_index): Likewise.
13697 (htab_t print_generated_program::params_index): Likewise.
13698 (htab_t gloog::newivs_index): Likewise.
13699 (htab_t gloog::params_index): Likewise.
13701 * graphite.c: Include graphite-htab.h.
13702 4htab_t graphite_transform_loops::bb_pbb_mapping):
13703 Change type to hash_table. Update dependent calls and types.
13705 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13706 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13707 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13709 * graphite-dependences.c: Include graphite-htab.h.
13710 (loop_is_parallel_p): Change hash table type of parameter.
13712 * graphite-htab.h: New.
13713 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13714 (extern find_pbb_via_hash): Move from graphite-poly.h.
13715 (extern loop_is_parallel_p): Move from graphite-poly.h.
13716 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13718 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13719 (extern loop_is_parallel_p): Move to graphite-htab.h.
13720 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13722 * haifa-sched.c (htab_t delay_htab):
13723 Change type to hash_table. Update dependent calls and types.
13724 (htab_t delay_htab_i2): Likewise.
13726 * ira-color.c (htab_t allocno_hard_regs_htab):
13727 Change type to hash_table. Update dependent calls and types.
13729 * ira-costs.c (htab_t cost_classes_htab):
13730 Change type to hash_table. Update dependent calls and types.
13732 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13733 Change type to hash_table. Update dependent calls and types.
13735 * loop-iv.c (htab_t bivs):
13736 Change type to hash_table. Update dependent calls and types.
13738 * loop-unroll.c (htab_t opt_info::insns_to_split):
13739 Change type to hash_table. Update dependent calls and types.
13740 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13742 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13743 (struct string_slot_hasher): New.
13744 (htab_t output_block::string_hash_table):
13745 Change type to hash_table. Update dependent calls and types.
13747 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13748 (htab_t file_name_hash_table):
13749 Change type to hash_table. Update dependent calls and types.
13751 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13754 * lto-streamer.c (htab_t tree_htab):
13755 Change type to hash_table. Update dependent calls and types.
13757 * omp-low.c: Include gimplify-ctx.h.
13759 * passes.c (htab_t name_to_pass_map):
13760 Change type to hash_table. Update dependent calls and types.
13761 (pass_traverse): Rename to passes_pass_traverse.
13763 * plugin.c (htab_t event_tab):
13764 Change type to hash_table. Update dependent calls and types.
13766 * postreload-gcse.c (htab_t expr_table):
13767 Change type to hash_table. Update dependent calls and types.
13768 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13770 * sese.c (debug_rename_map_1): Make extern.
13771 (htab_t copy_bb_and_scalar_dependences::rename_map):
13772 Change type to hash_table. Update dependent calls and types.
13774 * sese.h (extern debug_rename_map): Move to .c file.
13776 * store-motion.c (htab_t store_motion_mems_table):
13777 Change type to hash_table. Update dependent calls and types.
13779 * trans-mem.c (htab_t tm_new_mem_hash):
13780 Change type to hash_table. Update dependent calls and types.
13782 * tree-browser.c (htab_t TB_up_ht):
13783 Change type to hash_table. Update dependent calls and types.
13785 * tree-cfg.c (htab_t discriminator_per_locus):
13786 Change type to hash_table. Update dependent calls and types.
13788 * tree-complex.c: Include tree-hasher.h
13789 (htab_t complex_variable_components):
13790 Change type to hash_table. Update dependent calls and types.
13792 * tree-eh.c (htab_t finally_tree):
13793 Change type to hash_table. Update dependent calls and types.
13795 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13796 struct int_tree_hasher.
13797 (extern int_tree_map_eq): Likewise.
13798 (uid_decl_map_hash): Removed.
13799 (extern decl_tree_map_eq): Likewise.
13801 * tree-hasher.h: New.
13802 (struct int_tree_hasher): New.
13803 (typedef int_tree_htab_type): New.
13805 * tree-inline.c: Include gimplify-ctx.h.
13807 * tree-mudflap.c: Include gimplify-ctx.h.
13809 * tree-parloops.c: Include tree-hasher.h.
13810 (htab_t eliminate_local_variables_stmt::decl_address):
13811 Change type to hash_table. Update dependent calls and types.
13812 (htab_t separate_decls_in_region::decl_copies): Likewise.
13814 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13815 Change type to hash_table. Update dependent calls and types.
13817 * tree-sra.c (candidates):
13818 Change type to hash_table. Update dependent calls and types.
13820 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13822 (int_tree_map_hash): Likewise.
13824 * tree-ssa-dom.c (htab_t avail_exprs):
13825 Change type to hash_table. Update dependent calls and types.
13827 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13828 Change type to hash_table. Update dependent calls and types.
13830 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
13831 Change type to hash_table. Update dependent calls and types.
13833 * tree-ssa-phiopt.c (seen_ssa_names):
13834 Change type to hash_table. Update dependent calls and types.
13836 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
13837 Change type to hash_table. Update dependent calls and types.
13839 * tree-ssa-uncprop.c (equiv):
13840 Change type to hash_table. Update dependent calls and types.
13842 2013-04-25 Jakub Jelinek <jakub@redhat.com>
13844 PR rtl-optimization/57003
13845 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
13846 call note_stores with kill_clobbered_value callback again after
13847 killing regs_invalidated_by_call.
13849 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13851 * config/aarch64/aarch64-simd.md
13852 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
13853 (aarch64_simd_bsl<mode>): Likewise.
13854 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
13856 2013-04-25 Marek Polacek <polacek@redhat.com>
13858 PR tree-optimization/57066
13859 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
13861 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
13863 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
13865 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13867 * config/aarch64/aarch64-builtins.c
13868 (aarch64_fold_builtin): New.
13869 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
13870 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
13871 * config/aarch64/aarch64-simd-builtins.def (abs): New.
13872 * config/aarch64/arm_neon.h
13873 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
13875 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13876 Tejas Belagod <tejas.belagod@arm.com>
13878 * config/aarch64/aarch64-builtins.c
13879 (aarch64_gimple_fold_builtin): New.
13880 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
13881 * config/aarch64/aarch64-simd-builtins.def (addv): New.
13882 * config/aarch64/aarch64-simd.md (addpv4sf): New.
13883 (addvv4sf): Update.
13884 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
13886 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13888 * config/aarch64/aarch64.md
13889 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
13891 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13893 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
13894 (*ngcsi_uxtw): New pattern.
13896 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13897 Julian Brown <julian@codesourcery.com>
13899 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
13900 (TB_DREG): Add T_V4HF.
13901 (v4hf_UP): New macro.
13902 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13903 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13904 Handle initialisation of V4HF. Adjust initialisation of reinterpret
13906 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13907 (arm_vector_mode_supported_p): Handle V4HF.
13908 (arm_mangle_map): Handle V4HFmode.
13909 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
13910 * config/arm/arm_neon_builtins.def: Add entries for
13911 vcvtv4hfv4sf, vcvtv4sfv4hf.
13912 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
13913 (neon_vcvtv4hfv4sf): Likewise.
13914 * config/arm/neon-gen.ml: Handle half-precision floating point
13916 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
13917 * config/arm/arm_neon.h: Regenerate.
13918 * config/arm/neon.ml (type elts): Add F16.
13919 (type vectype): Add T_float16x4, T_floatHF.
13920 (type vecmode): Add V4HF.
13921 (type features): Add Requires_FP_bit feature.
13922 (elt_width): Handle F16.
13923 (elt_class): Likewise.
13924 (elt_of_class_width): Likewise.
13925 (mode_of_elt): Refactor.
13926 (type_for_elt): Handle F16, fix error messages.
13927 (vectype_size): Handle T_float16x4.
13928 (vcvt_sh): New function.
13929 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
13930 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
13931 (string_of_mode): Handle V4HF.
13932 * doc/arm-neon-intrinsics.texi: Regenerate.
13934 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13936 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
13937 format specifier in 'X' case.
13939 2013-04-25 Alan Modra <amodra@gmail.com>
13942 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
13943 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
13944 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
13945 Repeat for many other rotate/shift and mask patterns using subregs.
13946 Name lshiftrt insns.
13947 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
13948 on WORDS_BIG_ENDIAN.
13950 2013-04-25 Alan Modra <amodra@gmail.com>
13952 * config.gcc: Support little-endian powerpc-linux targets.
13953 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
13954 (LINK_OS_LINUX_SPEC): Define.
13955 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
13956 Preserve MASK_LITTLE_ENDIAN.
13957 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
13958 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
13959 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
13960 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
13961 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
13962 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
13963 Correct fp word order for little-endian. Don't shift toc entries
13964 smaller than a word for little-endian.
13965 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
13966 (bswapdi2 splits): Correct low-part subreg for little-endian.
13967 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
13968 low/high where such is correct only for be.
13969 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
13970 little-endian for -mcall-aixdesc.
13972 2013-04-25 Alan Modra <amodra@gmail.com>
13974 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
13975 replace_equiv_address_nv.
13977 2013-04-25 Alan Modra <amodra@gmail.com>
13979 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
13981 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13984 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13985 * rtl.h (struct rtx_def): ...
13987 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13989 PR rtl-optimizations/57046
13990 * lra-constraints (split_reg): Set up lra_risky_transformations_p
13991 for multi-reg splits.
13993 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
13995 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
13997 2013-04-24 Sterling Augustine <saugustine@google.com>
13999 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
14000 (comp_dir_string, debug_str_dwo_section): New.
14001 (DEBUG_STR_DWO_SECTION): Rename to ...
14002 (DEBUG_DWO_STR_SECTION): ... this.
14003 (DEBUG_NORM_STR_SECTION): Delete.
14004 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
14005 (DEBUG_STR_DWO_SECTION_FLAGS): New.
14006 (find_AT_string): Move most logic to ...
14007 (find_AT_string_in_table): ... here. New.
14008 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
14009 add_skeleton_AT_string. Delete logic.
14010 (output_skeleton_debug_sections): Remove call to
14011 add_top_level_skeleton_die_attrs.
14012 (add_comp_dir_attribute): Move logic to comp_dir_string.
14013 (dwarf2out_init): Initialize debug_str_dwo_section.
14014 (output_indirect_string): Call find_string_form.
14015 (output_indirect_strings): Rewrite.
14016 (prune_unused_types): Empty skeleton_debug_str_hash.
14017 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
14018 (dwarf2out_finish): Call output_indirect_strings.
14020 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14022 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
14024 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14026 * rtl.h (struct rtx_def): Add comment for field jump.
14027 (LRA_SUBREG_P): New macro.
14028 * recog.c (register_operand): Check LRA_SUBREG_P.
14029 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14031 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14032 (lra_final_code_change): Skip subreg change for operators.
14033 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14034 if there are no operand changes.
14035 * lra-constraints.c (curr_insn_set): New.
14036 (match_reload): Set LRA_SUBREG_P.
14037 (emit_spill_move): Ditto.
14038 (check_and_process_move): Use curr_insn_set. Process only single
14039 set insns. Don't initialize sec_mem_p and change_p.
14040 (simplify_operand_subreg): Use LRA_SUBREG_P.
14041 (reg_in_class_p): New function.
14042 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14043 of #ifdef. Add code to remove cycling.
14044 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14045 non-null disp. Reload inner instead of disp when base and index
14046 are null. Try to put lo_sum into register.
14047 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14048 (check_and_process_move): Move code for move cost check to
14049 simple_move_p. Remove equiv_substitution.
14050 (simple_move_p): New function.
14051 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14052 curr_insn_set. Call check_and_process_move only for single set
14053 insns. Use the new function. Move call of check_and_process_move
14054 after operand equiv substitution and address process.
14056 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
14058 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
14059 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
14060 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
14062 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14064 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
14066 2013-04-24 Marek Polacek <polacek@redhat.com>
14068 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
14069 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
14070 (select_loops_exit_conditions): Likewise.
14071 (number_of_iterations_for_all_loops): Likewise.
14072 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
14073 (scev_analysis): Likewise.
14075 2013-04-02 Catherine Moore <clm@codesourcery.com>
14076 Chao-ying Fu <fu@mips.com>
14078 * config/mips/micromips.md (jraddiusp): New pattern.
14079 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
14080 instruction if possible.
14082 2013-04-24 Alan Modra <amodra@gmail.com>
14084 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
14086 2013-04-24 Julian Brown <julian@codesourcery.com>
14087 Chung-Lin Tang <cltang@codesourcery.com>
14089 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
14090 dependency behavior in enumeration type DIE generation. Add TODO note
14091 to comments about future DW_FORM_sdata/udata re-work of related code.
14093 2013-04-23 Lawrence Crowl <crowl@google.com>
14095 * Makefile.in: Update as needed below.
14097 * hash-table.h (class hash_table):
14098 Correct many methods with parameter types compare_type to the correct
14099 value_type. (Correct code was unlikely to notice the change.)
14100 (hash_table::elements_with_deleted) New.
14101 (class hashtable::iterator): New.
14102 (hashtable::begin()): New.
14103 (hashtable::end()): New.
14104 (FOR_EACH_HASH_TABLE_ELEMENT): New.
14106 * statistics.c (statistics_hashes):
14107 Change type to hash_table. Update dependent calls and types.
14109 * tree-into-ssa.c (var_infos):
14110 Change type to hash_table. Update dependent calls and types.
14112 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
14113 Change type to hash_table. Update dependent calls and types.
14115 * tree-ssa-loop-im.c (struct mem_ref.refs):
14116 Change type to hash_table. Update dependent calls and types.
14118 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
14119 Change type to hash_table. Update dependent calls and types.
14121 * tree-ssa-sccvn.c (vn_tables_s::nary):
14122 Change type to hash_table. Update dependent calls and types.
14123 (vn_tables_s::phis): Likewise.
14124 (vn_tables_s::references): Likewise.
14126 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
14127 (vn_reference_eq): Update parameter and return types.
14129 * tree-ssa-structalias.c (pointer_equiv_class_table):
14130 Change type to hash_table. Update dependent calls and types.
14131 (location_equiv_class_table): Likewise.
14133 * tree-vect-data-refs.c: Consequential changes for making
14134 peeling a hash_table.
14136 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
14137 (destroy_loop_vec_info): Dependent hash_table update.
14139 * tree-vectorizer.h (peeling_htab):
14140 Change type to hash_table. Update dependent calls and types.
14142 2013-04-23 Shiva Chen <shiva0217@gmail.com>
14144 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
14145 to check the register content is equal or not.
14146 * lra-constraints.c (match_reload): Use lra_assign_reg_val
14147 to assign register content record.
14148 * lra-eliminations.c (update_reg_eliminate): Use
14149 lra_update_reg_val_offset to update register content offset.
14150 * lra-int.h (struct lra_reg): Add offset member.
14151 (lra_reg_val_equal_p): New static inline function.
14152 (lra_update_reg_val_offset): New static inline function.
14153 (lra_assign_reg_val): New static inline function.
14154 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
14155 to assign register content record.
14156 (initialize_lra_reg_info_element): Initial offset to zero.
14158 2013-04-23 Catherine Moore <clm@codesourcery.com>
14160 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
14161 operands. Record compression.
14163 2013-04-23 Xinliang David Li <davidxl@google.com>
14165 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
14167 2013-04-23 Richard Biener <rguenther@suse.de>
14169 PR middle-end/57036
14170 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
14171 parameter, only add abnormal goto edges from the copied body
14172 if the call could perform abnormal gotos.
14173 (copy_cfg_body): Adjust.
14175 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
14177 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
14179 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
14181 * coretypes.h (gimple_stmt_iterator): Add struct to make
14184 2013-04-23 Richard Biener <rguenther@suse.de>
14186 PR tree-optimization/57026
14187 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
14188 from SSA names occuring in abnormal PHI nodes.
14190 2013-04-22 Andi Kleen <ak@linux.intel.com>
14192 * lto/lto.c (print_lto_report_1): Fix LTO report names.
14194 2013-04-22 Andi Kleen <ak@linux.intel.com>
14196 * lto/lto.c (print_lto_report_1): Declare early.
14197 (read_cgraph_and_symbols): Call print_lto_report_1 early.
14199 2013-04-22 Andi Kleen <ak@linux.intel.com>
14201 * common.opt (-flto-report-wpa): Add.
14202 * doc/invoke.texi (-flto-report-wpa): Add.
14203 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
14206 2013-04-22 Xinliang David Li <davidxl@google.com>
14208 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
14209 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
14210 * Makefile.in: New dependency
14212 David Daney <ddaney.cavm@gmail.com>
14214 * configure.ac (gcc_cv_as_micromips_support): Use the
14215 --fatal-warnings option.
14216 * configure: Regenerate.
14218 2013-04-22 Marek Polacek <polacek@redhat.com>
14221 * tsan.c (instrument_expr): Don't instrument expression
14222 in case its size is zero.
14224 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
14228 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14230 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14232 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14234 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
14235 (gimple_stmt_iterator): New typedef.
14236 * gimple.h (gimple_stmt_iterator): Rename to...
14237 (gimple_stmt_iterator_d): ... This.
14238 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
14239 trees be valid for GIMPLE and GENERIC.
14240 (TARGET_GIMPLE_FOLD_BUILTIN): New.
14241 * gimple-fold.c (gimple_fold_call): Call target hook
14242 gimple_fold_builtin.
14243 * hooks.c (hook_bool_gsiptr_false): New.
14244 * hooks.h (hook_bool_gsiptr_false): New.
14245 * target.def (fold_stmt): New.
14246 * doc/tm.texi: Regenerate.
14248 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
14251 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
14252 a set sp if no stack realignment.
14254 2013-04-22 Nick Clifton <nickc@redhat.com>
14256 * config.gcc (tilegx-linux): Extend extra_objs rather than
14258 (tilepro-linux): Likewise.
14260 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14262 * config/aarch64/aarch64-builtins.c
14264 (CF0, CF1, CF2, CF3, CF4, CF10): New.
14265 (VAR<1-12>): Add MAP parameter.
14266 (BUILTIN_*): Likewise.
14267 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
14268 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
14269 (aarch64_ushl_n<mode>): Likewise.
14270 (aarch64_sshr_n<mode>): Likewise.
14271 (aarch64_ushr_n<mode>): Likewise.
14272 (aarch64_<maxmin><mode>): Likewise.
14273 (aarch64_sqrt<mode>): Likewise.
14274 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
14275 (vshr<q>_n_*): Likewise.
14277 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14279 * config/aarch64/aarch64-builtins.c
14280 (aarch64_simd_builtin_type_mode): Handle SF types.
14282 (BUILTIN_GPF): Define.
14283 (aarch64_init_simd_builtins): Handle SF types.
14284 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
14285 (frecps): Likewise.
14286 (frecpx): Likewise.
14287 * config/aarch64/aarch64-simd.md
14288 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
14289 (aarch64_frecpe<mode>): New.
14290 (aarch64_frecps<mode>): Likewise.
14291 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
14292 (v8type): Add frecp<esx>.
14293 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
14294 (aarch64_frecps<mode>): Likewise.
14295 * config/aarch64/iterators.md (FRECP): New.
14296 (frecp_suffix): Likewise.
14297 * config/aarch64/arm_neon.h
14298 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
14300 2013-04-22 Christian Bruel <christian.bruel@st.com>
14303 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
14304 (REG_CLASS_NAMES): Idem.
14305 (REG_CLASS_CONTENTS): Idem.
14306 (REGCLASS_HAS_FP_REG): Idem.
14307 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
14308 (sh_conditional_register_usage): Idem.
14310 2013-04-21 Jeff Law <law@redhat.com>
14312 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
14313 (ssa_forward_propagate_and_combine): Use it.
14315 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14317 * lra.c: Update the flow chart diagram.
14319 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14321 PR rtl-optimization/56847
14322 * lra-constraints.c (process_alt_operands): Discourage alternative
14323 with non-matche doffsettable memory constraint fro memory with
14326 2013-04-19 Richard Biener <rguenther@suse.de>
14328 PR tree-optimization/56982
14329 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
14331 * gimplify.c (gimplify_call_expr): Notice special calls.
14332 (gimplify_modify_expr): Likewise.
14333 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
14334 abnormal control flow receivers.
14335 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
14336 in the same way as cfun->has_nonlocal_labels.
14337 (gimple_purge_dead_abnormal_call_edges): Likewise.
14338 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
14339 receivers start a basic-block.
14341 2013-04-19 Richard Biener <rguenther@suse.de>
14343 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
14345 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
14346 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
14347 (SLP_TREE_LOAD_PERMUTATION): Add.
14348 (vect_transform_slp_perm_load): Adjust prototype.
14349 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
14350 (vect_free_slp_instance): Likewise.
14351 (vect_create_new_slp_node): Likewise.
14352 (vect_supported_slp_permutation_p): Remove.
14353 (vect_slp_rearrange_stmts): Adjust.
14354 (vect_supported_load_permutation_p): Likewise. Inline
14355 vect_supported_slp_permutation_p here.
14356 (vect_analyze_slp_instance): Compute load permutations per
14357 slp node instead of per instance.
14358 (vect_get_slp_defs): Adjust.
14359 (vect_transform_slp_perm_load): Likewise.
14360 (vect_schedule_slp_instance): Remove redundant code.
14361 (vect_schedule_slp): Remove hack for PR56270, add it ...
14362 * tree-vect-stmts.c (vectorizable_load): ... here, do not
14363 CSE loads for SLP. Adjust.
14365 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14367 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
14368 spelling in two comments.
14370 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14373 * config/arm/arm.c (load_multiple_sequence): Require SP
14374 as base register for loads if SP is in the register list.
14376 2013-04-19 Martin Jambor <mjambor@suse.cz>
14378 PR tree-optimization/56718
14379 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
14380 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
14381 and made public. Adjusted all callers.
14382 (ipa_intraprocedural_devirtualization): New function.
14383 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
14384 (ipa_intraprocedural_devirtualization): Likewise.
14385 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
14387 2013-04-19 Richard Biener <rguenther@suse.de>
14389 PR tree-optimization/57000
14390 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
14392 2013-04-19 Terry Guo <terry.guo@arm.com>
14394 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
14396 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
14397 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
14398 (cortex_m4_fmacs): Use new reservations.
14399 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
14401 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
14403 PR rtl-optimization/56999
14404 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
14406 (lra_coalesce): Remove split_origin_bitmap and related code.
14407 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
14408 ranges if necessary.
14410 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
14412 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
14414 (ix86_expand_call): Remove clobbered_registers array and use
14415 x86_64_ms_sysv_extra_clobbered_registers instead.
14416 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
14418 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
14420 * config/i386/i386.md (*call_rex64_ms_sysv): Use
14421 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
14422 (*call_value_rex64_ms_sysv): Ditto.
14424 2013-04-18 Cary Coutant <ccoutant@google.com>
14426 * dwarf2out.c (output_pubnames): Check die_perennial_p of
14427 parent instead of die_mark.
14429 2013-04-18 Diego Novillo <dnovillo@google.com>
14431 * gimple.c (create_gimple_tmp): New.
14432 (get_expr_type): New.
14433 (build_assign): New.
14434 (build_type_cast): New.
14435 * gimple.h (enum ssa_mode): Define.
14436 (gimple_seq_set_location): New.
14437 * asan.c (build_check_stmt): Change some gimple_build_* calls
14438 to use build_assign and build_type_cast.
14440 2013-04-18 Richard Biener <rguenther@suse.de>
14442 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
14443 handle negative step. Remove redundant checks.
14444 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
14445 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
14446 for negative step and grouped loads fail to vectorize.
14448 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14450 * emit-rtl.c (reset_insn_used_flags): New function.
14451 (reset_all_used_flags): Use it.
14452 (verify_insn_sharing): New function.
14453 (verify_rtl_sharing): Fix verification for SEQUENCEs.
14455 2013-04-18 Jakub Jelinek <jakub@redhat.com>
14457 PR tree-optimization/56984
14458 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
14459 and (x >> M) >= N don't register any assertion if N << M is the
14462 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14464 * lower-subreg.c (resolve_simple_move): If called self-recursive,
14465 do not delete_insn insns that have not yet been emitted, only
14466 unlink them with remove_insn.
14467 * df-scan.c (df_insn_delete): Revert r197492.
14469 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
14471 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
14472 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
14474 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14476 * config/arm/arm.md (movsicc_insn): Convert define_insn into
14477 define_insn_and_split.
14478 (and_scc,ior_scc,negscc): Likewise.
14479 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14481 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14483 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14484 can benefit from using a sequence of LDRD instructions in epilogue
14485 instead of a single LDM instruction.
14487 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14490 * doc/extend.texi: Fix typo.
14492 2013-04-17 Richard Biener <rguenther@suse.de>
14494 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14495 (vect_build_slp_tree): ... here.
14496 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14497 match. Remove special-casing of mismatched complex loads.
14498 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14499 re-try the match with swapped commutative operands.
14500 (vect_supported_load_permutation_p): Remove special-casing of
14501 mismatched complex loads.
14502 (vect_analyze_slp_instance): Adjust.
14504 2013-04-17 Richard Biener <rguenther@suse.de>
14506 PR rtl-optimization/56921
14507 * cfgloop.h (struct loop): Add simple_loop_desc member.
14508 (struct niter_desc): Mark with GTY(()).
14509 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14510 * loop-iv.c (get_simple_loop_desc): Likewise.
14511 (free_simple_loop_desc): Likewise.
14514 2013-04-16 Richard Biener <rguenther@suse.de>
14516 PR rtl-optimization/56921
14517 * loop-init.c (pass_rtl_move_loop_invariants): Add
14518 TODO_do_not_ggc_collect to todo_flags_finish.
14519 (pass_rtl_unswitch): Same.
14520 (pass_rtl_unroll_and_peel_loops): Same.
14521 (pass_rtl_doloop): Same.
14523 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14525 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14526 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14527 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14529 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14530 * tree-streamer.c (record_common_node): Adjust reference in comment.
14532 2013-04-17 Terry Guo <terry.guo@arm.com>
14534 * config/arm/cortex-m4.md: Add a new bypass.
14536 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14538 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14540 (*subs_<optab><mode>_multp2): New pattern.
14541 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14542 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14544 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14546 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14547 (*subs_mul_imm_<mode>): New pattern.
14549 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
14552 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14553 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14554 (vsx_movti_32bit): Change j->wa to O->wa.
14556 2013-04-16 Richard Biener <rguenther@suse.de>
14558 PR rtl-optimization/56921
14559 * loop-init.c (pass_rtl_move_loop_invariants): Add
14560 TODO_do_not_ggc_collect to todo_flags_finish.
14561 (pass_rtl_unswitch): Same.
14562 (pass_rtl_unroll_and_peel_loops): Same.
14563 (pass_rtl_doloop): Same.
14565 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
14567 * config/arm/arm.c (emit_multi_reg_push): New declaration
14568 for an existing function.
14569 (arm_emit_strd_push): New function.
14570 (arm_expand_prologue): Used here.
14571 (arm_emit_ldrd_pop): New function.
14572 (arm_expand_epilogue): Used here.
14573 (arm_get_frame_offsets): Update condition.
14574 (arm_emit_multi_reg_pop): Add a special case for load of a single
14575 register with writeback.
14577 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
14579 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
14582 2013-04-16 Richard Biener <rguenther@suse.de>
14584 PR tree-optimization/56756
14585 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
14586 (first_mem_ref_loc): New.
14587 (execute_sm): Place the load temporarily before a previous
14588 access instead of in the latch edge to ensure its SSA dependencies
14589 are defined at points dominating the load.
14591 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
14593 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
14594 correct fix by moving header and footer insn to the footer of
14595 the merged basic block. Clear BB_END of the merged-away block.
14597 PR middle-end/43631
14598 * emit-rtl.c (make_note_raw): New function.
14599 (link_insn_into_chain): New static inline function.
14600 (add_insn): Use it.
14601 (add_insn_before, add_insn_after): Factor insn chain linking code...
14602 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
14603 using link_insn_into_chain.
14604 (note_outside_basic_block_p): New helper function for emit_note_after
14605 and emit_note_before.
14606 (emit_note_after): Use nobb variant of add_insn_after if the note
14607 should not be contained in a basic block.
14608 (emit_note_before): Use nobb variant of add_insn_before if the note
14609 should not be contained in a basic block.
14610 (emit_note_copy): Use make_note_raw.
14611 (emit_note): Likewise.
14612 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
14613 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
14614 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
14615 the moved barrier the tail of the basic block it follows.
14616 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
14618 2013-04-15 Jakub Jelinek <jakub@redhat.com>
14620 PR tree-optimization/56962
14621 * gimple-ssa-strength-reduction.c (record_increment): Only set
14622 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
14623 either rhs1 or rhs2 is equal to c->base_expr.
14625 2013-04-15 Richard Biener <rguenther@suse.de>
14627 PR tree-optimization/56933
14628 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
14630 (GROUP_READ_WRITE_DEPENDENCE): Remove.
14631 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
14632 * tree-vect-data-refs.c (vect_analyze_group_access): Move
14633 dependence check ...
14634 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14636 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14637 GROUP_READ_WRITE_DEPENDENCE.
14639 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14641 * emit-rtl.c (reset_all_used_flags): New function.
14642 (verify_rtl_sharing): Call reset_all_used_flags before and after
14643 performing the checks.
14645 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14647 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14648 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14649 * config/arm/constraints.md (De): New constraint.
14650 * config/arm/neon.md (anddi3_neon): Delete.
14651 (neon_vand<mode>): Expand to standard anddi3 pattern.
14652 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14653 Move earlier in the file.
14654 (neon_inv_logic_op2): Likewise.
14655 (arm_anddi_operand_neon): New predicate.
14657 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14659 * configure.ac (gcc_cv_ld_as_needed): Set
14660 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14661 Use -z ignore, -z record on *-*-solaris2*.
14662 (HAVE_LD_AS_NEEDED): Update comment.
14663 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14664 * configure: Regenerate.
14665 * config.in: Regenerate.
14666 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14667 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14668 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14669 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14670 equivalents. Fix markup.
14671 * doc/tm.texi: Regenerate.
14673 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14675 * config/i386/i386.opt: New option mstack-protector-guard=.
14676 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14677 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14678 TARGET_SSP_TLS_GUARD.
14679 * config/i386/i386.c (ix86_option_override_internal): Set
14680 ix86_stack_protector_guard.
14681 * config/i386/i386.md (stack_protect_set): Enable for
14682 TARGET_SSP_TLS_GUARD only.
14683 (stack_protect_set_<mode>): Ditto.
14684 (stack_protect_test): Ditto.
14685 (stack_protect_test_<mode>): Ditto.
14686 * doc/invoke.texi (i386 Option): Document.
14688 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14691 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14692 (S_MODES): Set H_MODE bit.
14693 (SF_MODES): Set only S_MODE and SF_MODE bits.
14694 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14695 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14696 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14697 <MODE_FLOAT>: Likewise.
14699 2013-04-15 Joey Ye <joey.ye@arm.com>
14701 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14703 2013-04-15 Joey Ye <joey.ye@arm.com>
14705 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14707 (thumb_far_jump_used_p): Count instruction size and set
14710 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14712 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14713 * resource.c (mark_target_live_regs): Reformat conditional block.
14715 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14717 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14718 notes, they are emitted only just before final.
14719 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14721 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14723 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14724 * cfgrtl.c (delete_insn): Call it here instead.
14725 * lra-spills.c (lra_final_code_change): Use delete_insn.
14726 * haifa-sched.c (sched_remove_insn): Likewise.
14727 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14728 returning to the nop pool.
14729 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14730 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14732 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14734 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14735 * doc/tm.texi: Regenerated.
14737 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
14739 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14742 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14744 * df-core.c (df_find_def): Compare register numbers.
14745 (df_find_use): Likewise.
14747 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14750 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14751 lra_in_progress for return.
14753 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14755 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14756 define_insn into define_insn_and_split and emit movsicc patterns.
14758 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14760 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14762 2013-04-12 Richard Biener <rguenther@suse.de>
14764 * tree-pass.h (TODO_do_not_ggc_collect): New.
14765 * passes.c (execute_one_ipa_transform_pass): Honor
14766 TODO_do_not_ggc_collect.
14767 (execute_one_pass): Likewise.
14770 2013-04-10 Richard Biener <rguenther@suse.de>
14772 * passes.c (init_optimization_passes): Remove reload pass.
14773 * ira.c (do_reload): Merge into ...
14775 (rest_of_handle_reload): Remove.
14776 (pass_reload): Likewise.
14777 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14778 of reload for vzeroupper pass placement.
14780 2013-04-12 Jakub Jelinek <jakub@redhat.com>
14782 PR tree-optimization/56918
14783 PR tree-optimization/56920
14784 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14785 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14786 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14787 use wide_mul_with_sign method.
14789 2013-04-12 Richard Biener <rguenther@suse.de>
14791 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14792 not be considered a gimple constant.
14794 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
14796 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14797 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14800 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14801 Jakub Jelinek <jakub@redhat.com>
14803 * opts.c: Include diagnostic-color.h.
14804 (common_handle_option): Handle OPT_fdiagnostics_color_.
14805 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14806 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14807 (diagnostic-color.o): New.
14808 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14809 (diagnostic_color_rule): New enum.
14810 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14811 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14812 the location string.
14813 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14814 either NULL, or color kind.
14815 * diagnostic-color.c: New file.
14816 * diagnostic-color.h: New file.
14817 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14819 * doc/invoke.texi (-fdiagnostics-color): Document.
14820 * pretty-print.h (pp_show_color): Define.
14821 (struct pretty_print_info): Add show_color field.
14822 * diagnostic.c: Include diagnostic-color.h.
14823 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14824 macros. Colorize error:, warning: etc. strings and also the location
14826 (diagnostic_show_locus): Colorize the caret line.
14827 * pretty-print.c: Include diagnostic-color.h.
14828 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14829 inside of %< %> quotes or quoted through q format modifier.
14831 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14833 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
14835 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14837 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
14838 code in CC_NZ mode.
14839 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
14842 2013-04-11 Marek Polacek <polacek@redhat.com>
14844 PR tree-optimization/48184
14845 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
14847 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
14849 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
14850 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
14851 (skip_simple_arithmetic): Tidy up.
14852 * tree.h (skip_simple_constant_arithmetic): Declare.
14854 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14856 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
14858 2013-04-11 Richard Biener <rguenther@suse.de>
14860 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14861 generate vector constants.
14863 2013-04-11 Richard Biener <rguenther@suse.de>
14865 PR tree-optimization/56878
14866 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
14867 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
14869 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
14870 Prefer to align the DR with the most invariant base address.
14872 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14874 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
14877 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
14879 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
14880 floating-point vector comparisons against 0.
14882 2013-04-11 Jakub Jelinek <jakub@redhat.com>
14884 PR tree-optimization/56899
14885 * fold-const.c (extract_muldiv_1): Apply distributive law
14886 only if TYPE_OVERFLOW_WRAPS (ctype).
14888 2013-04-11 Bin Cheng <bin.cheng@arm.com>
14891 * ira-costs.c (scan_one_insn): Check whether the source rtx of
14892 loading has side effect.
14894 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
14896 * config/sparc/sparc.c: Include tree-pass.h.
14897 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
14898 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
14899 head of file. Change return type. Split off gate function.
14900 (sparc_gate_work_around_errata): New function.
14901 (pass_work_around_errata): New pass definition.
14902 (insert_pass_work_around_errata) New pass insert definition to
14903 insert pass_work_around_errata just after delayed-branch scheduling.
14904 (sparc_option_override): Insert the pass.
14905 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
14907 2013-04-10 David S. Miller <davem@davemloft.net>
14909 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
14910 or -mcpu=hypersparc.
14912 * target.def (cstore_mode): New hook.
14913 * target.h: Include insn-codes.h
14914 * targhooks.c: Likewise.
14915 (default_cstore_mode): New function.
14916 * targhooks.h: Declare it.
14917 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
14918 * doc/tm.texi: Rebuild.
14919 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
14920 target hook, rather than inspecting the insn_data.
14921 * config/sparc/sparc.c (sparc_cstore_mode): New function.
14922 (TARGET_CSTORE_MODE): Redefine.
14923 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
14925 * config/sparc/predicates.md (cstore_result_operand): New special
14927 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
14928 Use it for operand 0.
14929 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
14930 (*snesi_special): Likewise.
14931 (*snesi_zero): Likewise.
14932 (*seqsi_zero): Likewise.
14933 (*sltu_insn): Likewise.
14934 (*sgeu_insn): Likewise.
14935 (*seqdi_special): Make operand 0 and comparison operation be of
14937 (*snedi_special): Likewise.
14938 (*snedi_special_vis3): Likewise.
14939 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
14940 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
14941 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
14942 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
14943 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
14944 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
14945 (*sltu_extend_sp64): Likewise.
14946 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
14947 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
14948 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
14949 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
14950 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
14952 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
14954 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
14955 (aarch64_start_file): Use the new function.
14957 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14958 Jason Merrill <jason@redhat.com>
14960 * common.opt: Add -gdwarf.
14961 * opts.c (common_handle_option): Handle it.
14962 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
14964 2013-04-10 Richard Biener <rguenther@suse.de>
14966 * passes.c (execute_todo): Do not call ggc_collect conditional here.
14967 (execute_one_ipa_transform_pass): But unconditionally here.
14968 (execute_one_pass): And here.
14969 (init_optimization_passes): Remove reload pass.
14970 * tree-pass.h (TODO_ggc_collect): Remove.
14971 (pass_reload): Likewise.
14972 * ira.c (do_reload): Merge into ...
14974 (rest_of_handle_reload): Remove.
14975 (pass_reload): Likewise.
14976 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14977 of reload for vzeroupper pass placement.
14978 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
14979 and todo_flags_finish of all passes.
14981 2013-04-10 Richard Biener <rguenther@suse.de>
14983 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
14984 first_const_oprnd field, rename first_def_type to first_op_type.
14985 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
14986 (vect_get_and_check_slp_defs): Always use the type of the
14987 operand. Allow mixed vect_external_def, vect_constant_def types.
14988 (vect_get_constant_vectors): Handle mixed vect_external_def,
14989 vect_constant_def types.
14991 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
14993 PR tree-optimization/55524
14994 * tree-ssa-math-opts.c
14995 (convert_mult_to_fma): Don't use an fms construct
14996 when we don't have an fms operation, but fnma, and it looks
14997 likely that we'll be able to use the latter.
14999 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
15001 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
15003 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
15004 inline fail caused by overwritable functions.
15006 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
15008 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
15009 unnecessary bits in the constant power of two case.
15011 2013-04-10 Richard Biener <rguenther@suse.de>
15013 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
15014 broken code swapping operands.
15015 (vect_build_slp_tree): Do not compute load permutations here.
15016 (vect_analyze_slp_instance): Compute load permutations here,
15017 after building the SLP tree.
15019 2013-04-09 Christian Bruel <christian.bruel@st.com>
15021 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
15022 of next/prev_real_insn.
15024 2013-04-09 Jan Hubicka <jh@suse.cz>
15026 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15027 Drop aliased parameter.
15028 (function_and_variable_visibility): Do not handle alias pairs.
15029 * cgraph.c (varpool_externally_visible_p): Update prototype.
15030 * varpool.c (varpool_add_new_variable): Update.
15032 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15034 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
15036 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
15038 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
15040 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
15042 2013-04-09 Marek Polacek <polacek@redhat.com>
15044 PR tree-optimization/48762
15045 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
15047 2013-04-09 Richard Biener <rguenther@suse.de>
15049 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
15051 (vect_build_slp_tree): Likewise.
15052 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
15053 calculating the cost of a SLP instance.
15054 (vect_analyze_slp_instance): Use it from here, after building
15057 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15059 PR middle-end/56883
15060 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15061 expand_omp_for_static_chunk): Use simple_p = true in
15062 force_gimple_operand_gsi calls when assigning to addressable decls.
15064 2013-04-09 Jeff Law <law@redhat.com>
15066 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
15067 when the boolean was created by converting a wider object which
15068 had a boolean range.
15070 2013-04-09 Richard Biener <rguenther@suse.de>
15072 * tree-vectorizer.h (slp_void_p): Remove.
15073 (slp_tree): Typedef before _slp_tree declaration.
15074 (struct _slp_tree): Use a vector of slp_tree as children.
15075 (vect_get_place_in_interleaving_chain): Remove.
15076 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
15078 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
15080 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
15081 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
15082 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
15083 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
15084 Use slp_node instead of slp_void_p and adjust.
15086 2013-04-09 Richard Biener <rguenther@suse.de>
15088 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
15089 work that is not necessary.
15091 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15093 PR tree-optimization/56854
15094 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
15095 forward into clobber stmts if it would change MEM_REF lhs into
15098 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
15100 * tree.c (type_hash_lookup, type_hash_add): Make static.
15101 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
15103 2013-04-09 Richard Biener <rguenther@suse.de>
15105 * tree.h (unsave_expr_now): Remove.
15106 * tree-inline.c (mark_local_for_remap_r): Remove.
15107 (unsave_expr_1): Likewise.
15108 (unsave_r): Likewise.
15109 (unsave_expr_now): Likewise.
15110 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
15111 (propagate_tree_value): Likewise.
15113 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
15115 * doc/rtl.texi (sequence): Rewrite documentation to match the
15116 current use of SEQUENCE rtl objects.
15117 * rtl.def (SEQUENCE): Likewise.
15119 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
15120 Update documentation.
15121 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
15122 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
15124 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
15126 2013-04-08 Teresa Johnson <tejohnson@google.com>
15128 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
15129 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
15131 (estimate_edge_size_and_time): Add comment to suggest using rounding
15133 (estimate_node_size_and_time): Ditto.
15134 (remap_edge_change_prob): Use helper rounding divide methods.
15135 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
15136 (gimple_mod_pow2_value_transform): Ditto.
15137 (gimple_mod_subtract_transform): Ditto.
15138 (gimple_ic_transform): Ditto.
15139 (gimple_stringops_transform): Ditto.
15140 * stmt.c (conditional_probability): Ditto.
15141 (emit_case_dispatch_table): Ditto.
15142 * lto-cgraph.c (merge_profile_summaries): Ditto.
15143 * tree-optimize.c (execute_fixup_cfg): Ditto.
15144 * cfgcleanup.c (try_forward_edges): Ditto.
15145 * cfgloopmanip.c (scale_loop_profile): Ditto.
15147 (duplicate_loop_to_header_edge): Ditto.
15148 (lv_adjust_loop_entry_edge): Ditto.
15149 * tree-vect-loop.c (vect_transform_loop): Ditto.
15150 * profile.c (compute_branch_probabilities): Ditto.
15151 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
15152 * lto-streamer-in.c (input_cfg): Ditto.
15153 * gimple-streamer-in.c (input_bb): Ditto.
15154 * ipa-cp.c (update_profiling_info): Ditto.
15155 (update_specialized_profile): Ditto.
15156 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
15157 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
15159 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15160 (compute_trg_info): Ditto.
15161 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
15162 (purge_dead_edges): Ditto.
15163 * loop-unswitch.c (unswitch_loop): Ditto.
15164 * cgraphclones.c (cgraph_clone_edge): Ditto.
15165 (cgraph_clone_node): Ditto.
15166 * tree-inline.c (copy_bb): Ditto.
15167 (copy_edges_for_bb): Ditto.
15168 (initialize_cfun): Ditto.
15169 (copy_cfg_body): Ditto.
15170 (expand_call_inline): Ditto.
15172 2013-04-08 Kai Tietz <ktietz@redhat.com>
15174 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
15175 TARGET_CYGWIN64 by TARGET_64BIT.
15177 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15179 * config/epiphany/epiphany.md (GPR_1): New constant.
15180 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
15181 * config/epiphany/epiphany.c (gen_compare_reg):
15182 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
15183 is already in place.
15184 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
15185 Don't require being called during rtl expansion; If y operlaps r0,
15187 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
15188 (epiphany_expand_epilogue): Likewise.
15190 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
15191 Don't use CC_FPmode for ORDERED / UNORDERED.
15192 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
15194 * config/epiphany/constraints.md (CnL): New constraint.
15195 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
15196 * config/epiphany/predicates.md (add_operand): Allow 1024.
15198 * config/epiphany/epiphany.md (logical_op): New code iterator.
15199 (op_mnc): New code attribute.
15200 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
15201 (mov_f+1, mov_f+2): New peephole2 patterns.
15203 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
15204 (cstoresi4): Also allow re-use of zero result when doing a NE
15205 comparison to a non-zero operand.
15206 Use (clobber (scratch)) for first insn if the gpr output is not needed.
15208 * config/epiphany/epiphany.md (<insn_opname>v2si3):
15209 Use gen_addsi3_i / gen_subsi3_i.
15211 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15215 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
15216 contain anything but clobbers, at most one __builtin_stack_restore,
15217 optionally debug stmts and final resx, and if it has at least one
15218 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
15219 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
15220 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
15221 which isn't defaut definition, remove them.
15222 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
15223 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
15224 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
15225 with MEM_REF LHS with SSA_NAME address.
15227 2013-04-08 Jeff Law <law@redhat.com>
15229 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
15231 2013-04-08 Richard Biener <rguenther@suse.de>
15233 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
15235 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
15236 determined vector type.
15237 (vect_analyze_data_refs): Likewise.
15238 (vect_get_new_vect_var): Adjust.
15239 (vect_create_destination_var): Preserve SSA name versions.
15240 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
15241 not dump anything here.
15243 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15245 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
15246 Add member lr_slot_known.
15247 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
15249 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15250 Remove code that sets lr_slot_offset according to what a previous
15251 version of epiphany_emit_save_restore used to do.
15252 (epiphany_emit_save_restore): When doing an lr save or restore,
15253 set/verify lr_slot_known and lr_slot_offset.
15255 2013-04-08 Xinyu Qi <xyqi@marvell.com>
15258 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
15261 2013-04-08 Richard Biener <rguenther@suse.de>
15263 * alias.c (find_base_term): Fix thinko in previous change.
15265 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15267 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
15268 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
15269 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
15270 if possible to compute val.
15271 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
15272 For QImode integers don't require anything about precision. Use
15273 const_with_all_bytes_same to find out if the constant doesn't have
15274 repeated bytes in it.
15276 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15278 * config/s390/s390.c (s390_expand_insv): Only accept insertions
15281 2013-04-08 Marek Polacek <polacek@redhat.com>
15283 PR rtl-optimization/48182
15284 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
15287 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15290 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
15291 nuses, make sure we have a label.
15293 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15296 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
15297 (rs6000_emit_swdiv_low_precision): Remove.
15298 (rs6000_emit_swdiv): Rewrite to handle between one and four
15299 iterations of Newton-Raphson generally; modify required number of
15300 iterations for some cases.
15301 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
15303 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
15305 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
15306 set-but-unused variable.
15308 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
15309 basic blocks of released function bodies garbage-collectable.
15311 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
15312 (struct rtl_opt_pass): Add TODO_df_finish.
15314 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
15316 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15318 * config/arm/constraints.md (q): New constraint.
15319 * config/arm/ldrdstrd.md: New file.
15320 * config/arm/arm.md (ldrdstrd.md) New include.
15321 (arm_movdi): Use "q" instead of "r" constraint
15322 for double-word memory access.
15323 (movdf_soft_insn): Likewise.
15324 * config/arm/vfp.md (movdi_vfp): Likewise.
15325 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
15326 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
15327 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
15328 (mem_ok_for_ldrd_strd): Likewise.
15329 (output_move_double): Update assertion.
15331 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15333 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
15335 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15337 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
15338 define_insn_and_split.
15339 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
15341 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15343 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
15344 define_insn_and_split.
15345 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
15346 (shiftsi3_compare): New pattern.
15347 (rrx): New pattern.
15348 * config/arm/unspecs.md (UNSPEC_RRX): New.
15350 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15352 * config/arm/arm.md (negdi_extendsidi): New pattern.
15353 (negdi_zero_extendsidi): Likewise.
15355 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15357 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
15358 define_insn_and_split.
15359 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
15360 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
15362 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15364 * config/arm/arm.md (arm_subdi3): Convert define_insn into
15365 define_insn_and_split.
15366 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
15367 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
15369 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15371 * config/arm/arm.md (subsi3_carryin): New pattern.
15372 (subsi3_carryin_const): Likewise.
15373 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
15374 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
15376 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15378 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
15380 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15382 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
15383 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
15385 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15387 * config/arm/arm.c (arm_expand_builtin): Change fcode
15388 type to unsigned int.
15390 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15392 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
15394 2013-04-04 Ian Lance Taylor <iant@google.com>
15396 * doc/standards.texi (Standards): The Go frontend supports the Go 1
15399 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
15401 PR middle-end/56729
15402 * df-scan.c (df_insn_delete): Disable failing assert.
15404 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15406 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
15407 New function prototype.
15408 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
15409 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
15410 (arm_builtin_vectorized_function): New function.
15412 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15414 * config/arm/arm_neon_builtins.def: New file.
15415 * config/arm/arm.c (neon_builtin_data): Move contents to
15416 arm_neon_builtins.def.
15417 (enum arm_builtins): Include neon builtin definitions.
15418 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
15419 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
15421 2013-04-04 Marek Polacek <polacek@redhat.com>
15423 PR tree-optimization/48186
15424 * predict.c (maybe_hot_frequency_p): Return false if
15425 HOT_BB_FREQUENCY_FRACTION is 0.
15426 (cgraph_maybe_hot_edge_p): Likewise.
15428 2013-04-04 Richard Biener <rguenther@suse.de>
15430 PR tree-optimization/56826
15431 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
15434 2013-04-04 Richard Biener <rguenther@suse.de>
15436 PR tree-optimization/56213
15437 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
15438 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
15440 2013-04-04 Richard Biener <rguenther@suse.de>
15442 PR tree-optimization/56837
15443 * tree-loop-distribution.c (classify_partition): For non-zero
15444 values require that the value has the same precision as its
15445 mode to be useful as memset value.
15447 2013-04-03 Nick Clifton <nickc@redhat.com>
15449 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
15450 (fmssf4): Use fmsf.s on E3V5 architectures.
15451 (fnmasf4): Use fnmaf.s on E3V5 architectures.
15452 (fnmssf4): Use fnmsf.s on E3V5 architectures.
15454 2013-04-03 Jeff Law <law@redhat.com>
15456 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
15457 (lra-eliminations.o): Likewise.
15459 2013-04-03 Teresa Johnson <tejohnson@google.com>
15461 * gcov-io.c (compute_working_sets): Moved most of body of old
15462 compute_working_sets here from profile.c.
15463 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
15464 (gcov_working_set_t): Moved typedef here from basic-block.h
15465 (compute_working_set): Declare.
15466 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
15467 (get_working_sets): Renamed from compute_working_set,
15468 replace most of body with call to new compute_working_sets.
15469 (get_exec_counts): Replace call to compute_working_sets
15470 to get_working_sets.
15471 * profile.h (get_working_sets): Renamed from compute_working_set.
15472 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
15473 to get_working_sets.
15474 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
15475 * gcov-dump.c (dump_working_sets): New function.
15477 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
15479 * hwint.c (sext_hwi, zext_hwi): New functions.
15480 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15481 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15482 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15483 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15484 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15485 (sext_hwi, zext_hwi): New functions.
15487 2013-04-03 Jeff Law <law@redhat.com>
15489 PR tree-optimization/56799
15490 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15491 back test for widening conversion erroneously dropped in prior change.
15493 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15496 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15497 instead of next_real_insn.
15499 2013-04-03 Marek Polacek <polacek@redhat.com>
15502 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
15504 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15507 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15509 (thumb1_output_casesi): Likewise.
15510 (thumb2_output_casesi): Likewise.
15512 2013-04-03 Richard Biener <rguenther@suse.de>
15514 PR tree-optimization/56817
15515 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15517 (tree_unroll_loops_completely_1): ... new function to manually
15518 walk the loop tree, properly defering outer loops of unrolled
15519 loops to later iterations.
15521 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15523 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15524 (vectorizable_load): Likewise.
15525 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15526 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15528 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15530 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15533 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15535 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15537 2013-04-03 Bin Cheng <bin.cheng@arm.com>
15539 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15541 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15543 PR tree-optimization/56790
15544 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15547 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15549 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15551 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15552 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15555 2013-04-03 Jakub Jelinek <jakub@redhat.com>
15558 * tree.h (force_folding_builtin_constant_p): New decl.
15559 * builtins.c (force_folding_builtin_constant_p): New variable.
15560 (fold_builtin_constant_p): Fold immediately also if
15561 force_folding_builtin_constant_p.
15563 2013-04-03 Richard Biener <rguenther@suse.de>
15565 PR tree-optimization/56812
15566 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15567 DRs of the same interleaving chain are independent.
15569 2013-04-02 Jason Merrill <jason@redhat.com>
15571 * gdbinit.in (pbb): Use debug fn.
15573 2013-04-02 Lawrence Crowl <crowl@google.com>
15575 * sese.h (struct ivtype_map_elt_s): Remove unused.
15576 (extern debug_ivtype_map): Remove unused.
15577 (extern eq_ivtype_map_elts): Remove unused.
15578 * sese.c (debug_ivtype_map): Removed unused.
15579 (debug_ivtype_map_1): Removed unused.
15580 (debug_ivtype_elt): Remove unused.
15581 (eq_ivtype_map_elts): Remove unused.
15584 2013-04-02 Kai Tietz <ktietz@redhat.com>
15587 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
15588 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
15589 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
15591 (legitimize_pe_coff_symbol): Likewise.
15592 (is_imported_p): New helper-function.
15593 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
15594 for Windows x64 targets.
15595 (ix86_expand_prologue): Optimize for pe-coff targets.
15596 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
15597 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
15598 medium/large code-model.
15599 (legitimize_pic_address): Likewise.
15600 (legitimize_tls_address): Likewise.
15601 (ix86_expand_call): Likewise.
15602 (x86_output_mi_thunk): Likewise.
15603 (get_dllimport_decl): Add new beimport argument.
15604 (construct_plt_address): Don't assert for x64 pe-coff targets.
15605 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
15607 (SYMBOL_FLAG_STUBVAR): New macro.
15608 (SYMBOL_REF_STUBVAR_P): Likewise.
15609 * config/i386/winnt.c (stub_list): New structure.
15610 (stub_head): New local variable.
15611 (i386_pe_record_stub): New function.
15612 (i386_pe_file_end): Emit refptr-stubs.
15614 2013-04-02 Jakub Jelinek <jakub@redhat.com>
15616 PR rtl-optimization/56745
15617 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
15618 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
15621 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
15622 and both of them are MEM_REFs, just compare first argument for
15623 equality and attempt to deal even with differing offsets.
15626 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
15627 of gimple_clobber_p to be MEM_REF.
15628 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
15629 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
15630 after gimplification.
15631 * asan.c (get_mem_ref_of_assignment): Don't instrument
15632 gimple_clobber_p stmts.
15633 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
15634 gimple_clobber_p stmt if they have MEM_REF lhs and
15635 are dead because of another gimple_clobber_p stmt.
15636 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15637 gimple_clobber_p stmts like debug stmts.
15638 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15639 that refer to unused VAR_DECLs or uninitialized values.
15640 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15641 gimple_clobber_p stmts if they refer to removed parameters.
15642 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15645 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
15647 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15648 using SWI48 mode attribute.
15650 2013-04-02 Wei Mi <wmi@google.com>
15652 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15653 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15654 *<rotate_insn><mode>3_mask in i386.md.
15656 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15658 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15660 2013-04-02 Richard Biener <rguenther@suse.de>
15662 PR tree-optimization/56778
15663 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15664 Runtime alias tests are not supported for gather loads.
15665 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15666 stmts referenced from SSA operands before updating SSA form.
15668 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15669 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15671 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
15672 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15673 * config/arm/cortex-a53.md: New file.
15674 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15675 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15676 * config/arm/arm.c (arm_issue_rate): Likewise.
15677 * config/arm/arm-tune.md: Regenerate
15678 * config/arm/arm-tables.opt: Regenerate.
15679 * config/arm/arm-cores.def: Add cortex-a53.
15681 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15683 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15686 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15688 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15689 scalar load/store operations using B/H registers.
15690 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15692 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15694 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15696 * config/aarch64/aarch64.c
15697 (aarch64_simd_scalar_immediate_valid_for_move): New.
15698 * config/aarch64/aarch64-protos.h
15699 (aarch64_simd_scalar_immediate_valid_for_move): New.
15700 * config/aarch64/constraints.md (Dh, Dq): New.
15701 * config/aarch64/iterators.md (hq): New.
15703 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15705 * reorg.c (get_branch_condition): Deal with conditional returns.
15706 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15708 2013-04-01 Wei Mi <wmi@google.com>
15710 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15711 Truncate operand 2 using %b asm operand modifier.
15712 (*<shift_insn><mode>3_mask): Ditto.
15713 (*<rotate_insn><mode>3_mask): Ditto.
15715 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15717 PR middle-end/56798
15718 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15720 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15722 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15724 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15726 2013-03-30 Lawrence Crowl <crowl@google.com>
15728 * dse.c (clear_alias_sets): Remove never set.
15729 (disqualified_clear_alias_sets): Remove never set.
15730 (clear_alias_mode_pool): Remove never set.
15731 (dse_step0): Remove condition that is never true.
15732 (canon_address): Remove condition that is never true.
15733 (dse_step7): Remove condition that is never true.
15734 (rest_of_handle_dse): Remove condition that is never true.
15735 (rest_of_handle_dse::did_global): Remove never read from above.
15736 (dse_step2_spill): Remove never called from above.
15737 (dse_step5_spill): Remove never called from above.
15739 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15741 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15742 JUMP_TABLE_DATA changes.
15743 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15744 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15745 (Insns) <jump_table_data>: New entry.
15746 * doc/tm.texi: Regenerate.
15748 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15750 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15751 for table jump at the end of a basic block using tablejump_p.
15752 * targhooks.c (default_invalid_within_doloop): Likewise.
15753 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15754 target hook implementation that is identical to the default hook.
15755 (rs6000_invalid_within_doloop): Remove.
15757 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15758 unused variable from tablejump_p call.
15760 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15761 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15762 (INSN_DELETED_P): Likewise.
15763 (emit_jump_table_data): New prototype.
15764 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15765 after 4th as unused.
15766 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15767 * sched-vis.c (print_insn): Likewise.
15768 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15769 insn for compatibility with back ends that use next_active_insn to
15770 identify jump table data.
15771 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15772 (remove_insn): Likewise.
15773 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15775 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15776 (emit_jump_table_data): New function.
15778 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15779 basic block, a JUMP_TABLE_DATA never is.
15780 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15781 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15782 off from code handling real insns.
15783 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15784 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15785 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15786 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15787 is not a NONDEBUG_INSN_P.
15788 * ira-costs.c (scan_one_insn): Likewise.
15789 * jump.c (mark_all_labels): Likewise.
15790 (mark_jump_label_1): Likewise.
15791 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15792 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15794 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15795 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15796 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15797 JUMP_TABLE_DATA_P insns.
15798 (calculate_elim_costs_all_insns): Likewise.
15799 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15800 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15801 (delete_output_reload): Code style fixups.
15802 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15803 insn flags on this non-insn.
15804 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15805 as scheduling barriers, for pre-change compatibility.
15806 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15807 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15809 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15810 redundant JUMP_TABLE_DATA_P test.
15811 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15812 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15813 (frv_for_each_packet): Likewise.
15814 * config/i386/i386.c (min_insn_size): Likewise.
15815 (ix86_avoid_jump_mispredicts): Likewise.
15816 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15817 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15818 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15819 (mips16_insn_length): Robustify.
15820 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15821 (mips16_split_long_branches): Likewise.
15822 * config/pa/pa.c (pa_combine_instructions): Likewise.
15823 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15824 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15825 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15826 as contributing to pool range lengths.
15827 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15828 Remove redundant JUMP_TABLE_DATA_P test.
15829 (sh_loop_align): Likewise.
15830 (split_branches): Likewise.
15831 (sh_insn_length_adjustment): Likewise.
15832 * config/spu/spu.c (get_branch_target): Likewise.
15834 2013-03-29 Jan Hubicka <jh@suse.cz>
15836 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
15837 gcov streaming; stream hot bb threshold to ltrans.
15838 * predict.c (get_hot_bb_threshold): Break out from ....
15839 (maybe_hot_count_p): ... here.
15840 (set_hot_bb_threshold): New function.
15841 * lto-section-in.c (lto_section_name): Add profile.
15842 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
15843 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
15844 and data-streamer.h
15845 (histogram_entry): New structure.
15846 (histogram, histogram_pool): New global vars.
15847 (histogram_hash): New structure.
15848 (histogram_hash::hash): New method.
15849 (histogram_hash::equal): Likewise.
15850 (account_time_size): New function.
15851 (cmp_counts): New function.
15852 (dump_histogram): New function.
15853 (ipa_profile_generate_summary): New function.
15854 (ipa_profile_write_summary): New function.
15855 (ipa_profile_read_summary): New function.
15856 (ipa_profile): Decide on threshold.
15857 (pass_ipa_profile): Add ipa_profile_write_summary and
15858 ipa_profile_read_summary.
15859 * Makefile.in (ipa.o): Update dependencies.
15860 * lto-streamer.h (LTO_section_ipa_profile): New section.
15862 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
15864 * tree.h (VAR_P): New.
15866 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
15869 * doc/invoke.texi ([-fwhole-program]): Fix typo.
15871 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
15873 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
15874 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15875 (control_flow_insn_p): Likewise.
15876 * cfgrtl.c (duplicate_insn_chain): Likewise.
15877 * final.c (get_attr_length_1): Likewise.
15878 (shorten_branches): Likewise.
15879 (final_scan_insn): Likewise.
15880 * function.c (instantiate_virtual_regs): Likewise.
15881 * gcse.c (insert_insn_end_basic_block): Likewise.
15882 * ira-costs.c (scan_one_insn): Likewise.
15883 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15884 * lra.c (check_rtl): Likewise.
15885 * reload1.c (elimination_costs_in_insn): Likewise.
15886 * reorg.c (follow_jumps): Likewise.
15888 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
15889 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15890 (thumb_far_jump_used_p): Likewise.
15891 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
15892 (workaround_speculation): Likewise.
15893 (add_sched_insns_for_speculation): Likewise.
15894 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
15895 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15896 (frv_for_each_packet): Likewise.
15897 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
15898 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
15899 (final_emit_insn_group_barriers): Likewise.
15900 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15901 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15902 (mips16_insn_length): Likewise.
15903 * config/pa/pa.c (pa_reorg): Likewise.
15904 (pa_combine_instructions): Likewise.
15905 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
15906 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
15907 (sh_reorg): Likewise.
15908 (split_branches): Likewise.
15909 * config/spu/spu.c (get_branch_target): Likewise.
15911 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
15914 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
15916 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
15917 Fix declaration name.
15919 2013-03-28 Lawrence Crowl <crowl@google.com>
15921 * graphds.h (struct graph.indicies): Remove unused.
15922 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
15923 (SCOP_ORIGINAL_PDDRS): Remove unused.
15924 * sese.h (extern insert_loop_close_phis): Removed unused.
15925 (extern insert_guard_phis): Removed unused.
15926 (extern ivtype_map_elt_info): Removed unused.
15927 (new_ivtype_map_elt): Removed unused.
15928 * sese.c (ivtype_map_elt_info): Removed unused.
15930 2013-03-28 Lawrence Crowl <crowl@google.com>
15932 * Makefile.in: Add several missing include dependences.
15934 (test-dump.o): New. This object is not added to any executable,
15935 but is present for ad-hoc testing.
15937 (debug (const bitmap_head_def &)): New.
15938 (debug (const bitmap_head_def *)): New.
15940 (extern debug (const bitmap_head_def &)): New.
15941 (extern debug (const bitmap_head_def *)): New.
15943 (debug (edge_def &)): New.
15944 (debug (edge_def *)): New.
15946 (debug (basic_block_def &)): New.
15947 (debug (basic_block_def *)): New.
15949 (dump_node (const_tree, int, FILE *)): Correct source file.
15951 (debug (die_struct &)): New.
15952 (debug (die_struct *)): New.
15954 (extern debug (die_struct &)): New.
15955 (extern debug (die_struct *)): New.
15956 * gimple-pretty-print.c
15957 (debug (gimple_statement_d &)): New.
15958 (debug (gimple_statement_d *)): New.
15959 * gimple-pretty-print.h
15960 (extern debug (gimple_statement_d &)): New.
15961 (extern debug (gimple_statement_d *)): New.
15963 (debug (ira_allocno_copy &)): New.
15964 (debug (ira_allocno_copy *)): New.
15965 (debug (ira_allocno &)): New.
15966 (debug (ira_allocno *)): New.
15968 (extern debug (ira_allocno_copy &)): New.
15969 (extern debug (ira_allocno_copy *)): New.
15970 (extern debug (ira_allocno &)): New.
15971 (extern debug (ira_allocno *)): New.
15973 (debug (live_range &)): New.
15974 (debug (live_range *)): New.
15976 (debug (lra_live_range &)): New.
15977 (debug (lra_live_range *)): New.
15979 (debug (lra_live_range &)): New.
15980 (debug (lra_live_range *)): New.
15982 (debug (omega_pb_d &)): New.
15983 (debug (omega_pb_d *)): New.
15985 (extern debug (omega_pb_d &)): New.
15986 (extern debug (omega_pb_d *)): New.
15988 (debug (const rtx_def &)): New.
15989 (debug (const rtx_def *)): New.
15991 (debug_tree (tree): Move within file.
15992 (debug_raw (const tree_node &)): New.
15993 (debug_raw (const tree_node *)): New.
15994 (dump_tree_via_hooks (const tree_node *, int)): New.
15995 (debug (const tree_node &)): New.
15996 (debug (const tree_node *)): New.
15997 (debug_verbose (const tree_node &)): New.
15998 (debug_verbose (const tree_node *)): New.
15999 (debug_head (const tree_node &)): New.
16000 (debug_head (const tree_node *)): New.
16001 (debug_body (const tree_node &)): New.
16002 (debug_body (const tree_node *)): New.
16003 (debug_vec_tree (tree): Move and reimplement in terms of dump.
16004 (debug (vec<tree, va_gc> &)): New.
16005 (debug (vec<tree, va_gc> *)): New.
16007 (extern debug (const rtx_def &)): New.
16008 (extern debug (const rtx_def *)): New.
16010 (debug_raw (simple_bitmap_def &)): New.
16011 (debug_raw (simple_bitmap_def *)): New.
16012 (debug (simple_bitmap_def &)): New.
16013 (debug (simple_bitmap_def *)): New.
16015 (extern debug (simple_bitmap_def &)): New.
16016 (extern debug (simple_bitmap_def *)): New.
16017 (extern debug_raw (simple_bitmap_def &)): New.
16018 (extern debug_raw (simple_bitmap_def *)): New.
16020 (debug (vinsn_def &)): New.
16021 (debug (vinsn_def *)): New.
16022 (debug_verbose (vinsn_def &)): New.
16023 (debug_verbose (vinsn_def *)): New.
16024 (debug (expr_def &)): New.
16025 (debug (expr_def *)): New.
16026 (debug_verbose (expr_def &)): New.
16027 (debug_verbose (expr_def *)): New.
16028 (debug (vec<rtx> &)): New.
16029 (debug (vec<rtx> *)): New.
16031 (extern debug (vinsn_def &)): New.
16032 (extern debug (vinsn_def *)): New.
16033 (extern debug_verbose (vinsn_def &)): New.
16034 (extern debug_verbose (vinsn_def *)): New.
16035 (extern debug (expr_def &)): New.
16036 (extern debug (expr_def *)): New.
16037 (extern debug_verbose (expr_def &)): New.
16038 (extern debug_verbose (expr_def *)): New.
16039 (extern debug (vec<rtx> &)): New.
16040 (extern debug (vec<rtx> *)): New.
16042 (_list_iter_cond_expr): Make inline instead of static.
16044 (debug (sreal &)): New.
16045 (debug (sreal *)): New.
16047 (extern debug (sreal &)): New.
16048 (extern debug (sreal *)): New.
16050 (extern debug_raw (const tree_node &)): New.
16051 (extern debug_raw (const tree_node *)): New.
16052 (extern debug (const tree_node &)): New.
16053 (extern debug (const tree_node *)): New.
16054 (extern debug_verbose (const tree_node &)): New.
16055 (extern debug_verbose (const tree_node *)): New.
16056 (extern debug_head (const tree_node &)): New.
16057 (extern debug_head (const tree_node *)): New.
16058 (extern debug_body (const tree_node &)): New.
16059 (extern debug_body (const tree_node *)): New.
16060 (extern debug (vec<tree, va_gc> &)): New.
16061 (extern debug (vec<tree, va_gc> *)): New.
16063 (debug (struct loop &)): New.
16064 (debug (struct loop *)): New.
16065 (debug_verbose (struct loop &)): New.
16066 (debug_verbose (struct loop *)): New.
16067 * tree-dump.c: Add header dependence.
16069 (extern debug (struct loop &)): New.
16070 (extern debug (struct loop *)): New.
16071 (extern debug_verbose (struct loop &)): New.
16072 (extern debug_verbose (struct loop *)): New.
16074 (debug (data_reference &)): New.
16075 (debug (data_reference *)): New.
16076 (debug (vec<data_reference_p> &)): New.
16077 (debug (vec<data_reference_p> *)): New.
16078 (debug (vec<ddr_p> &)): New.
16079 (debug (vec<ddr_p> *)): New.
16081 (extern debug (data_reference &)): New.
16082 (extern debug (data_reference *)): New.
16083 (extern debug (vec<data_reference_p> &)): New.
16084 (extern debug (vec<data_reference_p> *)): New.
16085 (extern debug (vec<ddr_p> &)): New.
16086 (extern debug (vec<ddr_p> *)): New.
16088 (debug (pt_solution &)): New.
16089 (debug (pt_solution *)): New.
16091 (extern debug (pt_solution &)): New.
16092 (extern debug (pt_solution *)): New.
16094 (debug (_var_map &)): New.
16095 (debug (_var_map *)): New.
16096 (debug (tree_live_info_d &)): New.
16097 (debug (tree_live_info_d *)): New.
16099 (extern debug (_var_map &)): New.
16100 (extern debug (_var_map *)): New.
16101 (extern debug (tree_live_info_d &)): New.
16102 (extern debug (tree_live_info_d *)): New.
16104 2013-03-28 Jan Hubicka <jh@suse.cz>
16106 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
16108 2013-03-28 Ian Bolton <ian.bolton@arm.com>
16110 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
16111 record only when desired or required.
16113 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
16115 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
16116 *vec_extractv2di_1_rex64. Use x64 isa attribute.
16118 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16120 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
16121 (*andsi3_compare0_uxtw): New pattern.
16122 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
16123 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
16125 2013-03-28 Jan Hubicka <jh@suse.cz>
16127 * data-streamer-in.c (streamer_read_gcov_count): New function.
16128 * gimple-streamer-out.c: Include value-prof.h.
16129 (output_gimple_stmt): Output histogram.
16130 (output_bb): Use streamer_write_gcov_count.
16131 * value-prof.c: Include data-streamer.h
16132 (dump_histogram_value): Add HIST_TYPE_MAX.
16133 (stream_out_histogram_value): New function.
16134 (stream_in_histogram_value): New function.
16135 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
16136 (stream_out_histogram_value, stream_in_histogram_value): Declare.
16137 * data-streamer-out.c (streamer_write_gcov_count): New function.
16138 (streamer_write_gcov_count_stream): New function.
16139 * lto-cgraph.c (lto_output_edge): Update counter streaming.
16140 (lto_output_node): Likewise.
16141 (input_node, input_edge): Likewise.
16142 * lto-streamer-out.c (output_cfg): Update streaming.
16143 * lto-streamer-in.c (input_cfg): Likewise.
16144 * data-streamer.h (streamer_write_gcov_count,
16145 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
16146 * gimple-streamer-in.c: Include value-prof.h
16147 (input_gimple_stmt): Input histograms.
16148 (input_bb): Update profile streaming.
16150 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16152 * genmodes.c (emit_max_int): New function.
16153 (emit_insn_modes_h): Added call to emit_max_function.
16154 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
16156 * machmode.def: Fixed comment.
16158 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16160 * combine.c (try_combine): Removed useless assert.
16161 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
16163 2013-03-28 Marek Polacek <polacek@redhat.com>
16164 Richard Biener <rguenther@suse.de>
16166 PR tree-optimization/56695
16167 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
16168 build signed result of a vector comparison.
16169 * tree-cfg.c (verify_gimple_comparison): Check that a result
16170 of a vector comparison has signed type.
16172 2013-03-28 Richard Biener <rguenther@suse.de>
16174 PR tree-optimization/37021
16175 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
16176 do not restrict gaps between groups.
16177 * tree-vect-stmts.c (vectorizable_load): Properly account for
16178 a gap between groups.
16180 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
16182 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
16183 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
16186 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
16188 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
16189 * doc/extend.texi (Named Address Spaces): Ditto.
16190 (Variable Attributes): Ditto.
16192 2013-03-27 Kai Tietz <ktietz@redhat.com>
16194 * config.build: Add support for cygwin x64 target.
16195 * config.gcc: Likewise.
16196 * config.host: Likewise.
16197 * configure.ac: Likewise
16198 * configure: Regenerated.
16200 2013-03-27 Kai Tietz <ktietz@redhat.com>
16202 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
16203 * config/i386/t-cygwin-w64: New file.
16204 * config/i386/cygwin-w64.h: New file.
16205 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
16206 and add support for x64-cygwin target.
16207 (CPP_SPEC): Likewise.
16208 (CXX_WRAP_SPEC_LIST): Undefine before define.
16209 (LIBGCJ_SONAME): Use 15 as version.
16211 2013-03-27 Richard Biener <rguenther@suse.de>
16213 PR tree-optimization/56716
16214 * tree-ssa-structalias.c (perform_var_substitution): Adjust
16215 dumping for ref nodes.
16217 2013-03-27 Martin Jambor <mjambor@suse.cz>
16219 PR tree-optimization/55334
16220 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
16221 restricted pointers to arrays.
16223 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
16225 * Makefile.in (.SUFFIXES): Add .cc.
16226 (.c.o): Apply same recipe for implicit rule .cc.o.
16228 2013-03-27 Richard Biener <rguenther@suse.de>
16230 PR tree-optimization/37021
16231 * tree-vect-data-refs.c (vect_check_strided_load): Allow
16232 REALPART/IMAGPART_EXPRs around the supported refs.
16233 * tree-ssa-structalias.c (find_func_aliases): Assume that
16234 floating-point values are not used to transfer pointers.
16236 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
16238 * target.def (TARGET_HAS_IFUNC_P): New target hook.
16239 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
16240 * doc/tm.texi: Regenerate.
16241 * targhooks.h (default_has_ifunc_p): New.
16242 * targhooks.c (default_has_ifunc_p): Ditto.
16243 * config/linux-protos.h: New file.
16244 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
16245 hook for linux which disables support of indirect functions in android.
16246 * config/linux-android.c: New file.
16247 * config/t-linux-android.c: Ditto.
16248 * config.gcc: Added new object file linux-android.o.
16249 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
16250 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
16251 * varasm.c (do_assemble_alias): Likewise.
16252 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
16253 doesn't support indirect functions.
16254 * configure: Regenerate.
16256 2013-03-27 Bin Cheng <bin.cheng@arm.com>
16259 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
16260 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
16263 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16265 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
16267 2013-03-27 Terry Guo <terry.guo@arm.com>
16269 * config/arm/arm-cores.def: Added core cortex-r7.
16270 * config/arm/arm-tune.md: Regenerated.
16271 * config/arm/arm-tables.opt: Regenerated.
16272 * doc/invoke.texi: Added entry for core cortex-r7.
16274 2013-03-27 Walter Lee <walt@tilera.com>
16276 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
16277 double-decrement of next_scratch_regno.
16279 2013-03-27 Walter Lee <walt@tilera.com>
16281 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
16283 (insn_v1mulus): Ditto.
16284 (insn_v2muls): Ditto.
16286 2013-03-27 Walter Lee <walt@tilera.com>
16288 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
16289 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
16291 2013-03-27 Walter Lee <walt@tilera.com>
16293 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
16294 (*sibcall_value): Ditto.
16296 2013-03-27 Walter Lee <walt@tilera.com>
16298 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
16299 (insn_mnz_v8qi): ... this ...
16300 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
16302 (insn_v<n>mnz): Replaced by ...
16303 (insn_v1mnz): ... this ...
16304 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
16306 (insn_mz_<mode>): Replaced by ...
16307 (insn_mz_v8qi): ... this ...
16308 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
16310 (insn_v<n>mz): Replaced by ...
16311 (insn_v1mz): ... this ...
16312 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
16315 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
16317 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
16319 2013-03-26 Roland McGrath <mcgrathr@google.com>
16321 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
16322 than fprintf with a non-constant, non-format string.
16324 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16326 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
16327 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
16328 operand 0 predicate.
16329 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
16330 attribute. Use general_x64nomem_operand as operand 1 predicate.
16331 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
16332 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16333 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
16334 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16335 (mov<mode>_insv_1): Remove expander. Merge insn with
16336 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
16337 Use general_x64nomem_operand as operand 1 predicate.
16338 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
16339 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
16340 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16341 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
16342 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
16343 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
16344 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16345 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
16346 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
16347 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
16348 (general_x64nomem_operand): Ditto.
16350 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
16352 * config/rtems.opt: Add -pthread option.
16354 2013-03-26 Richard Biener <rguenther@suse.de>
16356 * alias.c (find_base_term): Avoid redundant and not used recursion.
16357 (base_alias_check): Get the initial base term from the caller.
16358 (true_dependence_1): Compute and pass base terms to base_alias_check.
16359 (write_dependence_p): Likewise.
16360 (may_alias_p): Likewise.
16362 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
16364 * config/aarch64/aarch64.c (aarch64_classify_address): Support
16365 PC-relative load in SI modes and above only.
16367 2013-03-26 Xinyu Qi <xyqi@marvell.com>
16369 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
16370 * config/arm/iwmmxt.md (WCGR0): Update.
16371 (WCGR1, WCGR2, WCGR3): Likewise.
16373 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16375 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
16376 Use x64 and nox64 isa attributes.
16378 2013-03-26 Richard Biener <rguenther@suse.de>
16380 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
16381 alignment computations and rely on get_object_alignment_1
16382 for the !TYPE_P case.
16383 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
16385 2013-03-26 Walter Lee <walt@tilera.com>
16387 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
16388 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
16390 2013-03-25 Jeff Law <law@redhat.com>
16392 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
16393 check for INTEGRAL_TYPE_P that was missing due to checking in
16394 wrong version of prior patch.
16396 2013-03-25 Walter Lee <walt@tilera.com>
16398 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
16399 TILEGX_INSN_SHUFFLEBYTES1.
16400 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
16402 (tilegx_builtins): Ditto.
16403 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
16405 2013-03-25 Walter Lee <walt@tilera.com>
16407 * config/tilegx/tilegx.md (floatsisf2): New pattern.
16408 (floatunssisf2): New pattern.
16409 (floatsidf2): New pattern.
16410 (floatunssidf2): New pattern.
16412 2013-03-25 Walter Lee <walt@tilera.com>
16414 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
16415 tests for constraint J, K, N, P.
16417 2013-03-25 Walter Lee <walt@tilera.com>
16419 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
16420 Use indirect/pcrel encoding.
16421 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
16424 2013-03-25 Steve Ellcey <sellcey@mips.com>
16426 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
16427 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
16428 * config/mips/mips.c (mips_option_override): Set IMADD default.
16429 * config/mips/mips.h (PTF_AVOID_IMADD): New.
16430 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
16431 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
16432 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
16433 * doc/invoke.texi (-mimadd/-mno-imadd): New.
16435 2013-03-25 Jeff Law <law@redhat.com>
16437 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
16438 slightly to avoid creating and folding useless trees. Simplify
16439 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
16441 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
16443 * config/i386/i386.md (*zero_extendsidi2): Merge with
16444 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
16445 * config/i386/predicates.md (x86_64_zext_operand): Rename from
16446 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
16447 targets. Clarify comment.
16449 2013-03-25 Martin Jambor <mjambor@suse.cz>
16451 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
16452 pass-through jump functions differently.
16453 (ipa_read_jump_function): Likewise. Also use setter functions to set
16456 2013-03-25 Martin Jambor <mjambor@suse.cz>
16458 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
16459 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
16461 (ipa_get_indirect_edge_target): New function.
16462 (devirtualization_time_bonus): New parameter known_aggs, pass it to
16463 ipa_get_indirect_edge_target. Update all callers.
16464 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
16465 ipa_get_indirect_edge_target_1 instead of calling
16466 ipa_get_indirect_edge_target.
16467 (create_specialized_node): Pass aggvlas to
16468 ipcp_discover_new_direct_edges.
16470 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16472 * config/arm/arm.md (f_sels, f_seld): New types.
16473 (*cmov<mode>): New pattern.
16474 * config/arm/predicates.md (arm_vsel_comparison_operator): New
16477 2013-03-25 Kai Tietz <ktietz@redhat.com>
16479 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16480 POSIX-printf for mingw-hosted builds.
16482 2013-03-25 Richard Biener <rguenther@suse.de>
16484 PR middle-end/56694
16485 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16486 must-not-throw stmt location.
16488 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16490 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16491 Emit load-acquire versions when acq is true.
16492 (arm_emit_store_exclusive): Add rel parameter.
16493 Emit store-release versions when rel is true.
16494 (arm_split_compare_and_swap): Use acquire-release instructions
16496 of barriers when appropriate.
16497 (arm_split_atomic_op): Likewise.
16498 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16499 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16500 (VUNSPEC_SLX): Likewise.
16501 (VUNSPEC_LDA): Likewise.
16502 (VUNSPEC_STL): Likewise.
16503 * config/arm/sync.md (atomic_load<mode>): New pattern.
16504 (atomic_store<mode>): Likewise.
16505 (arm_load_acquire_exclusive<mode>): Likewise.
16506 (arm_load_acquire_exclusivesi): Likewise.
16507 (arm_load_acquire_exclusivedi): Likewise.
16508 (arm_store_release_exclusive<mode>): Likewise.
16510 2013-03-25 Catherine Moore <clm@codesourcery.com>
16512 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16513 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16514 * config/mip/predicates.md (lwsp_swsp_operand,
16515 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16516 sb16_operand, db4_operand, db7_operand, ib3_operand,
16517 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16518 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16519 andi16_operand): New predicates.
16520 * config/mips/mips.md (compression): New attribute.
16521 (enabled): New attribute.
16522 (length): Consider compression in computing length.
16523 (shift_compression): New code attribute.
16524 (*add<mode>3): New operands. Record compression.
16525 (sub<mode>3): Likewise.
16526 (one_cmpl<mode>2): Likewise.
16527 (*and<mode>3): Likewise.
16528 (*ior<mode>3): Likewise.
16529 (unnamed pattern for xor): Likewise.
16530 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16531 (*<optab><mode>3): Likewise.
16532 (*mov<mode>_internal: Likewise.
16533 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16534 (mips_unsigned_immediate_p): New.
16535 (umips_lwsp_swsp_address_p): New.
16536 (m16_based_address_p): New.
16537 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16538 (mips_unsigned_immediate_p): New prototype.
16539 (lwsp_swsp_address_p): New prototype.
16540 (m16_based_address_p): New prototype.
16541 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16542 (mips_signed_immediate_p): New function.
16543 (m16_based_address_p): New function.
16544 (lwsp_swsp_address_p): New function.
16545 (mips_print_operand_punctuation): Recognize short delay slot insns
16546 for microMIPS.add<mode>3"
16548 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16551 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16552 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16554 2013-03-25 Richard Biener <rguenther@suse.de>
16556 PR tree-optimization/56689
16557 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16560 2013-03-25 Richard Biener <rguenther@suse.de>
16562 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16564 (memory_references): Likewise.
16565 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16566 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16567 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16568 (gather_mem_refs_in_loops): Fold into ...
16569 (analyze_memory_references): ... this. Move initialization
16570 to tree_ssa_lim_initialize.
16571 (fill_always_executed_in): Rename to ...
16572 (fill_always_executed_in_1): ... this.
16573 (fill_always_executed_in): Move contains_call computation to
16574 this new function from ...
16575 (tree_ssa_lim_initialize): ... here.
16576 (tree_ssa_lim): Call fill_always_executed_in.
16578 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16580 * postreload.c (reload_combine): Fix code detecting returns.
16582 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16584 * function.c (emit_use_return_register_into_block): On cc0 targets,
16585 do not emit the sequence between cc0 setter and user.
16587 2013-03-25 Kai Tietz <ktietz@redhat.com>
16589 * config/i386/predicates.md (local_symbolic_operand): Interpret
16590 dll-imported symbols as none-local.
16592 2013-03-25 Richard Biener <rguenther@suse.de>
16594 * tree-ssa-loop-im.c (struct depend): Remove.
16595 (struct lim_aux_data): Make depends a vec of gimples.
16596 (free_lim_aux_data): Adjust.
16597 (add_dependency): Likewise.
16598 (set_level): Likewise.
16600 2013-03-25 Richard Biener <rguenther@suse.de>
16602 PR middle-end/56434
16603 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
16604 the pointer returned by calls with ECF_MALLOC set.
16606 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16608 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
16610 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16612 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
16613 using MMXMODE mode iterator.
16614 (*move<mode>_internal): Merge with *movv2sf_internal and
16615 *movv2sf_internal_rex64 using MMXMODE mode iterator.
16617 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16619 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
16620 (record_last_mem_set_info): Likewise.
16622 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
16623 of XNEWVEC followed by memset.
16624 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
16626 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16628 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
16629 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
16630 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
16631 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
16632 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
16633 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
16634 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16635 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16636 BARRIER_P instead of GET_CODE.
16638 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16640 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16641 inaccuracy in the probing code.
16643 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16644 (ctrapdi4): Likewise.
16646 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16648 * calls.c (expand_call): Add missing guard to code handling return
16649 of non-BLKmode structures in MSB.
16650 * function.c (expand_function_end): Likewise.
16652 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16654 * combine.c (try_combine): Adjust comment. Do not add the set of
16655 insn #0 if the destination indirectly is set or dies in insn #2.
16656 Tidy up code to distribute a new note.
16658 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16660 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16661 also for alternatives 16 and 17.
16663 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16665 * config/i386/sse.md (*mov<mode>_internal): Merge with
16666 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16667 Emit insn template depending on type attribute. Use
16668 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16669 movd instead of movq mnemonic for interunit moves. Rewrite mode
16670 attribute calculation. Remove unit attribute calculation.
16671 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16672 Set prefix_data16 attribute for DImode ssemov types.
16673 Use Ym instead of y for SSE-MMX conversion alternatives.
16674 Reorder operand constraints.
16676 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16678 * df.h (df_insn_delete): Adjust prototype.
16679 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16680 and let it decide whether mark the basic block dirty.
16681 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16682 * df-scan.c (df_insn_info_delete): New helper function, split
16683 off from df_insn_delete.
16684 (df_scan_free_bb_info): Use it.
16685 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16687 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16688 that the insn is actually an insn and it has a non-NULL basic block.
16689 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16691 2013-03-22 Richard Biener <rguenther@suse.de>
16693 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16695 (mem_ref_alloc): Do not allocate them.
16696 (refs_independent_p): Do not query or maintain a cache.
16698 2013-03-22 Richard Biener <rguenther@suse.de>
16700 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16701 (gather_mem_refs_in_loops): Do not compute it.
16702 (analyze_memory_references): Do not allocate it.
16703 (tree_ssa_lim_finalize): Do not free it.
16704 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16706 2013-03-22 Richard Biener <rguenther@suse.de>
16708 * is-a.h (as_a): Use gcc_checking_assert.
16710 2013-03-22 Ian Bolton <ian.bolton@arm.com>
16712 * config/aarch64/aarch64.c (aarch64_print_operand): New
16713 format specifier for printing a constant in hex.
16714 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16715 format specifier for printing second operand.
16717 2013-03-22 Richard Biener <rguenther@suse.de>
16719 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16721 (gather_mem_refs_in_loops): Perform store accumulation here.
16722 (create_vop_ref_mapping_loop): Remove.
16723 (create_vop_ref_mapping): Likewise.
16724 (analyze_memory_references): Initialize refs_stored_in_loop.
16725 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
16726 (record_indep_loop): Remove.
16727 (record_dep_loop): New function.
16728 (ref_indep_loop_p_1): Adjust to only walk over references
16729 in the loop, not its subloops.
16730 (ref_indep_loop_p): Rename to ...
16731 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16732 maintaining a more fine-grained cache.
16733 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16734 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16736 2013-03-22 Richard Biener <rguenther@suse.de>
16738 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16739 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16740 aggregate mem_ref_loc.
16741 (free_mem_ref_locs): Inline into ...
16742 (memref_free): ... this and adjust.
16743 (mem_ref_alloc): Adjust.
16744 (mem_ref_locs_alloc): Remove.
16745 (record_mem_ref_loc): Adjust.
16746 (get_all_locs_in_loop): Rewrite into ...
16747 (for_all_locs_in_loop): ... this iterator.
16748 (rewrite_mem_ref_loc): New functor.
16749 (rewrite_mem_refs): Use for_all_locs_in_loop.
16750 (sm_set_flag_if_changed): New functor.
16751 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16752 (ref_always_accessed): New functor.
16753 (ref_always_accessed_p): Use for_all_locs_in_loop.
16755 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16757 * tree-pass.h (PROP_gimple_lvec): New.
16758 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16759 (init_optimization_passes): Move pass_lower_vector.
16760 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16762 (pass_lower_vector): Provide PROP_gimple_lvec.
16763 (pass_lower_vector_ssa): Likewise.
16764 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16766 2013-03-21 Mark Wielaard <mjw@redhat.com>
16768 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16770 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16772 * config/i386/i386.md (*movdi_internal): Disparage slightly
16773 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16774 conversion alternatives.
16776 2013-03-21 Jakub Jelinek <jakub@redhat.com>
16778 PR middle-end/48087
16779 * diagnostic.def (DK_WERROR): New kind.
16780 * diagnostic.h (werrorcount): Define.
16781 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16782 promoted to DK_ERROR, increment DK_WERROR counter instead of
16784 * toplev.c (toplev_main): Call print_ignored_options even if
16785 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16786 even if just werrorcount is non-zero.
16789 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16791 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16792 (string_cst_pool_decl): New function.
16793 (optimize_one_addr_into_implicit_ptr): New function.
16794 (resolve_addr_in_expr): Optimize DWARF location expression
16795 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16796 which doesn't live in memory, but has DW_AT_location or
16797 DW_AT_const_value, or refers to a string literal, into
16798 DW_OP_GNU_implicit_pointer.
16799 (optimize_location_into_implicit_ptr): New function.
16800 (resolve_addr): If removing DW_AT_location of a variable because
16801 it was DW_OP_addr of address of the variable, but the variable doesn't
16802 live in memory, try to emit const value attribute for the initializer.
16804 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16806 * tree.h (VECTOR_TYPE_P): New macro.
16807 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16808 TYPE_MODE): Use it.
16809 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16810 VEC_COND_EXPR cannot be lvalues.
16811 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16813 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16815 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16816 Restrict the transformation to equal modes.
16818 2013-03-21 Richard Biener <rguenther@suse.de>
16820 PR tree-optimization/39326
16821 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16822 (MEM_ANALYZABLE): Adjust.
16823 (record_mem_ref_loc): Move bitmap ops ...
16824 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16825 unanalyzable refs, do not record locations for it.
16826 (analyze_memory_references): Allocate ref zero as shared
16828 (refs_independent_p): Do not test for unanalyzed mems here.
16829 (ref_indep_loop_p_1): Special-case disambiguation against
16830 the unanalyzed ref.
16831 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
16833 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
16835 * config/arm/arm-protos.h (tune_params): Add
16836 prefer_neon_for_64bits field.
16837 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
16838 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
16839 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
16840 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
16841 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
16842 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
16843 (arm_option_override): Handle -mneon-for-64bits new option.
16844 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
16845 (prefer_neon_for_64bits): Declare new variable.
16846 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
16847 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
16848 (arch_enabled): Handle new arch types. Remove support for onlya8
16850 (one_cmpldi2): Use new arch names.
16851 (zero_extend<mode>di2, extend<mode>di2): Ditto.
16852 * config/arm/arm.opt (mneon-for-64bits): Add option.
16853 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
16854 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
16855 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
16857 * doc/invoke.texi (-mneon-for-64bits): Document.
16859 2013-03-21 Richard Biener <rguenther@suse.de>
16861 PR tree-optimization/39326
16862 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
16863 (sort_bbs_in_loop_postorder_cmp): New function.
16864 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
16866 2013-03-21 Richard Biener <rguenther@suse.de>
16868 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
16869 (vect_insert_into_interleaving_chain): Likewise.
16870 (vect_drs_dependent_in_basic_block): Inline ...
16871 (vect_slp_analyze_data_ref_dependence): ... here. New function,
16873 (vect_analyze_data_ref_dependence): ... here. Simplify.
16874 (vect_check_interleaving): Simplify.
16875 (vect_analyze_data_ref_dependences): Likewise. Split out ...
16876 (vect_slp_analyze_data_ref_dependences): ... this new function.
16877 (dr_group_sort_cmp): New function.
16878 (vect_analyze_data_ref_accesses): Compute data-reference groups
16879 here instead of in vect_analyze_data_ref_dependence. Use
16880 a more efficient algorithm.
16881 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
16882 vect_slp_analyze_data_ref_dependences. Call
16883 vect_analyze_data_ref_accesses earlier.
16884 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
16885 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
16886 (vect_slp_analyze_data_ref_dependences): New prototype.
16888 2013-03-21 Richard Biener <rguenther@suse.de>
16890 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
16891 ref is stored in the loop.
16892 (find_refs_for_sm): Walk only over all stores.
16893 (store_motion_loop): Allocate from lim_bitmap_obstack.
16894 (store_motion): Likewise.
16896 2013-03-21 Richard Biener <rguenther@suse.de>
16898 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
16899 Update virtual SSA form.
16901 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16903 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
16904 * configure: Regenerate.
16905 * config.in: Regenerate.
16906 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
16907 if !HAVE_LD_EH_FRAME_CIEV3.
16909 2013-03-21 Richard Biener <rguenther@suse.de>
16911 * tree-cfg.c (verify_expr_no_block): New function.
16912 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
16913 nor DECL_VALUE_EXPR have locations with associated blocks.
16914 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
16915 (clear_unused_block_pointer): Remove code dealing with
16916 blocks in DECL_DEBUG_EXPR locations.
16918 2013-03-21 Richard Biener <rguenther@suse.de>
16920 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
16921 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
16922 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
16923 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
16924 instead of DECL_DEBUG_EXPR_IS_FROM.
16925 * gimplify.c (gimplify_modify_expr): Likewise.
16926 * tree-cfg.c (verify_expr_location_1): Likewise.
16927 * tree-complex.c (create_one_component_var): Likewise.
16928 * tree-sra.c (create_access_replacement): Likewise.
16929 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
16930 (clear_unused_block_pointer): Likewise.
16931 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
16932 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
16933 * var-tracking.c (var_debug_decl): Likewise.
16934 (track_expr_p): Likewise.
16935 * tree-inline.c (add_local_variables): Likewise. Set
16936 DECL_HAS_DEBUG_EXPR_P after copying it.
16937 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
16938 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
16940 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16943 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
16944 * configure: Regenerate.
16945 * config.in: Regenerate.
16946 * config/i386/i386.md (*movdf_internal): Use
16947 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16948 movd instead of movq mnemonic for interunit moves.
16949 (*movdi_internal): Ditto.
16951 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16953 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
16954 (abd<mode>_3): New pattern.
16955 (aba<mode>_3): New pattern.
16956 (fabd<mode>_3): New pattern.
16958 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16960 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
16961 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
16962 occurrence of REGISTER_PREFIX as its empty string.
16964 2013-03-20 Jeff Law <law@redhat.com>
16966 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
16967 addititional equivalences for equality comparisons between an SSA_NAME
16968 and a constant where the SSA_NAME was set from a widening conversion.
16970 2013-03-20 Walter Lee <walt@tilera.com>
16972 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
16974 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16976 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
16977 depending on type attribute.
16978 (*movti_internal): Ditto.
16979 (*movtf_internal): Ditto.
16980 (*movxf_internal): Ditto.
16981 (*movdf_internal): Ditto.
16982 (*movsf_internal): Ditto.
16984 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16986 * config/i386/i386.md (*movti_internal): Set prefix attribute to
16987 maybe_vex for sselog1 and ssemov types.
16988 (*movdi_internal): Reorder operand constraints.
16989 (*movsi_internal): Ditto. Set prefix attribute to
16990 maybe_vex for sselog1 and ssemov types.
16991 (*movtf_internal): Set prefix attribute to maybe_vex
16992 for sselog1 and ssemov types.
16993 (*movdf_internal): Ditto. Set prefix_data16 attribute for
16994 DImode ssemov types. Reorder operand constraints.
16995 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
16996 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
16997 attribute for SImode ssemov types. Reorder operand constraints.
16999 2013-03-20 Martin Jambor <mjambor@suse.cz>
17001 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
17002 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
17004 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
17006 * config/rs6000/predicates.md (indexed_address, update_address_mem
17007 update_indexed_address_mem): New predicates.
17008 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
17009 attribute for load/store instructions.
17010 * config/rs6000/dfp.md (movsd_store): Likewise.
17011 (movsd_load): Likewise.
17012 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
17013 (unnamed HI->DI extend define_insn): Likewise.
17014 (unnamed SI->DI extend define_insn): Likewise.
17015 (unnamed QI->SI extend define_insn): Likewise.
17016 (unnamed QI->HI extend define_insn): Likewise.
17017 (unnamed HI->SI extend define_insn): Likewise.
17018 (unnamed HI->SI extend define_insn): Likewise.
17019 (extendsfdf2_fpr): Likewise.
17020 (movsi_internal1): Likewise.
17021 (movsi_internal1_single): Likewise.
17022 (movhi_internal): Likewise.
17023 (movqi_internal): Likewise.
17024 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
17025 attribute for load/store instructions.
17026 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
17028 (mov<mode>_softfloat): Likewise.
17029 (mov<mode>_hardfloat32): Likewise.
17030 (mov<mode>_hardfloat64): Likewise.
17031 (mov<mode>_softfloat64): Likewise.
17032 (movdi_internal32): Likewise.
17033 (movdi_internal64): Likewise.
17034 (probe_stack_<mode>): Likewise.
17036 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
17038 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
17039 floating point, and decimal floating point to reload iterator.
17041 * config/rs6000/constraints.md (wl constraint): New constraints to
17042 return FLOAT_REGS if certain options are used to reduce the number
17043 of separate patterns that exist in the file.
17044 (wx constraint): Likewise.
17045 (wz constraint): Likewise.
17047 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17048 -mdebug=reg, print wg, wl, wx, and wz constraints.
17049 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17050 Initialize the reload functions for 64-bit binary/decimal floating
17052 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17053 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17054 create the buffer on the stack to overcome not having a 32-bit
17056 (rs6000_emit_move): Likewise.
17057 (rs6000_secondary_memory_needed_rtx): Likewise.
17058 (rs6000_alloc_sdmode_stack_slot): Likewise.
17059 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17060 via xxlxor, just like DFmode 0.0.
17062 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
17063 define as 1 if we are running on a power7 or newer.
17064 (enum r6000_reg_class_enum): Add new constraints.
17066 * config/rs6000/dfp.md (movsd): Delete, combine with binary
17067 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17068 with other moves by using conditional constraits (wg). Use LFIWZX
17069 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
17070 (movsd splitter): Likewise.
17071 (movsd_hardfloat): Likewise.
17072 (movsd_softfloat): Likewise.
17074 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
17075 binary and decimal floating point moves.
17076 (fmove_ok): New attributes to combine binary and decimal floating
17077 point moves, and to combine power6x (mfpgpr) moves along normal
17079 (real_value_to_target): Likewise.
17080 (f32_lr): Likewise.
17081 (f32_lm): Likewise.
17082 (f32_li): Likewise.
17083 (f32_sr): Likewise.
17084 (f32_sm): Likewise.
17085 (f32_si): Likewise.
17086 (movsf): Combine binary and decimal floating point moves. Combine
17087 power6x (mfpgpr) moves with other moves by using conditional
17088 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
17089 (mov<mode> for SFmode/SDmode); Likewise.
17090 (SFmode/SDmode splitters): Likewise.
17091 (movsf_hardfloat): Likewise.
17092 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
17093 (movsf_softfloat): Likewise.
17094 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
17096 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
17097 wx and wz constraints.
17099 * config/rs6000/constraints.md (wg constraint): New constraint to
17100 return FLOAT_REGS if -mmfpgpr (power6x) was used.
17102 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
17105 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17106 -mdebug=reg, print wg, wl, wx, and wz constraints.
17107 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17108 Initialize the reload functions for 64-bit binary/decimal floating
17110 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17111 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17112 create the buffer on the stack to overcome not having a 32-bit
17114 (rs6000_emit_move): Likewise.
17115 (rs6000_secondary_memory_needed_rtx): Likewise.
17116 (rs6000_alloc_sdmode_stack_slot): Likewise.
17117 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17118 via xxlxor, just like DFmode 0.0.
17120 * config/rs6000/dfp.md (movdd): Delete, combine with binary
17121 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17122 with other moves by using conditional constraits (wg). Use LFIWZX
17123 and STFIWX for loading SDmode on power7.
17124 (movdd splitters): Likewise.
17125 (movdd_hardfloat32): Likewise.
17126 (movdd_softfloat32): Likewise.
17127 (movdd_hardfloat64_mfpgpr): Likewise.
17128 (movdd_hardfloat64): Likewise.
17129 (movdd_softfloat64): Likewise.
17131 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
17132 64-bit binary and decimal floating point moves.
17133 (FMOVE64X): Likewise.
17134 (movdf): Combine 64-bit binary and decimal floating point moves.
17135 Combine power6x (mfpgpr) moves with other moves by using
17136 conditional constraits (wg).
17137 (mov<mode> for DFmode/DDmode): Likewise.
17138 (DFmode/DDmode splitters): Likewise.
17139 (movdf_hardfloat32): Likewise.
17140 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
17141 (movdf_softfloat32): Likewise.
17142 (movdf_hardfloat64_mfpgpr): Likewise.
17143 (movdf_hardfloat64): Likewise.
17144 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
17145 (movdf_softfloat64): Likewise.
17146 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
17147 (reload_<mode>_load): Move to later in the file so they aren't in
17148 the middle of the floating point move insns.
17149 (reload_<mode>_store): Likewise.
17151 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
17154 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
17155 constraint if -mdebug=reg.
17156 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
17157 Enable using dd reload support if needed.
17159 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
17160 binary and decimal floating point moves in rs6000.md.
17161 (movtd_internal): Likewise.
17163 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
17164 decimal floating point moves.
17166 (movtf_internal): Likewise.
17167 (mov<mode>_internal, TDmode/TFmode): Likewise.
17168 (movtf_softfloat): Likewise.
17169 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
17171 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
17172 movdi_internal64, using wg constraint for move direct operations.
17173 (movdi_internal64): Likewise.
17175 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
17176 MODES_TIEABLE_P for selected modes. Print the numerical value of
17177 the various virtual registers. Use GPR/FPR first/last values,
17178 instead of hard coding the register numbers. Print which modes
17179 have reload functions registered.
17180 (rs6000_option_override_internal): If -mdebug=reg, trace the options
17181 settings before/after setting cpu, target and subtarget settings.
17182 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
17183 and for secondary reload failures in rs6000_secondary_reload_inner.
17184 (rs6000_secondary_reload_fail): Likewise.
17185 (rs6000_secondary_reload_inner): Likewise.
17187 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
17188 macros for first/last GPR and FPR registers.
17189 (LAST_GPR_REGNO): Likewise.
17190 (FIRST_FPR_REGNO): Likewise.
17191 (LAST_FPR_REGNO): Likewise.
17193 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
17194 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
17195 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
17196 (vcond<mode><mode>): Likewise.
17197 (vcondu<mode><mode>): Likewise.
17198 (vector_gtu<mode>): Likewise.
17199 (vector_gte<mode>): Likewise.
17200 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
17201 to prevent the compiler from converting DImode operations to TImode.
17202 (ior<mode>3): Likewise.
17203 (and<mode>3): Likewise.
17204 (one_cmpl<mode>2): Likewise.
17205 (nor<mode>3): Likewise.
17206 (andc<mode>3): Likewise.
17208 * config/rs6000/constraints.md (wt constraint): New constraint
17209 that returns VSX_REGS if TImode is allowed in VSX registers.
17211 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
17212 constant under VSX.
17214 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
17215 similar to TImode, but it is restricted to being in the GPRs.
17217 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
17218 TImode to occupy a single VSX register.
17220 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
17221 -mvsx-timode for power7/power8.
17222 (power7 cpu): Likewise.
17223 (power8 cpu): Likewise.
17225 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
17226 sure that TFmode/TDmode take up two registers if they are ever
17227 allowed in the upper VSX registers.
17228 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
17230 (rs6000_init_hard_regno_mode_ok): Likewise.
17231 (rs6000_debug_reg_global): Add debugging for PTImode and wt
17232 constraint. Print if LRA is turned on.
17233 (rs6000_option_override_internal): Give an error if -mvsx-timode
17234 and VSX is not enabled.
17235 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
17236 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
17237 to reg+offset addressing. Use PTImode when checking offset
17238 addresses for validity.
17239 (reg_offset_addressing_ok_p): Likewise.
17240 (rs6000_legitimate_offset_address_p): Likewise.
17241 (rs6000_legitimize_address): Likewise.
17242 (rs6000_legitimize_reload_address): Likewise.
17243 (rs6000_legitimate_address_p): Likewise.
17244 (rs6000_eliminate_indexed_memrefs): Likewise.
17245 (rs6000_emit_move): Likewise.
17246 (rs6000_secondary_reload): Likewise.
17247 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
17248 reloads to fpr registers to continue to use reg+offset addressing,
17249 but 64-bit reloads to altivec registers need reg+reg addressing.
17250 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
17251 it. Treat LO_SUM like a PLUS operation.
17252 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
17253 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
17254 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
17255 registers to share a register with a smaller sized type, since VSX
17256 puts scalars in the upper 64-bits.
17257 (print_operand): Add support for PTImode.
17258 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
17259 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
17260 registers, but don't have arithmetic support.
17261 (rs6000_memory_move_cost): Add test for VSX.
17262 (rs6000_opt_masks): Add -mvsx-timode.
17264 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
17267 (VSr): Use wt constraint for TImode.
17268 (VSv): Drop TImode support.
17269 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
17270 (vsx_movti_64bit): Likewise.
17271 (vsx_movti_32bit): Likewise.
17272 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
17273 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
17274 one '?' on the appropriate output constraint. Do not allow TImode
17275 logical operations on 32-bit systems.
17276 (vsx_ior<mode>3): Likewise.
17277 (vsx_xor<mode>3): Likewise.
17278 (vsx_one_cmpl<mode>2): Likewise.
17279 (vsx_nor<mode>3): Likewise.
17280 (vsx_andc<mode>3): Likewise.
17281 (vsx_concat_<mode>): Likewise.
17282 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
17284 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
17285 OPTION_MASK_VSX_TIMODE.
17286 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
17287 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
17289 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
17290 (TI2 iterator): New iterator for TImode, PTImode.
17291 (wd mode attribute): Add values for vector types.
17292 (movti_string): Replace TI move operations with operations for TImode
17293 and PTImode. Add support for TImode being allowed in VSX registers.
17294 (mov<mode>_string, TImode/PTImode): Likewise.
17295 (movti_ppc64): Likewise.
17296 (mov<mode>_ppc64, TImode/PTImode): Likewise.
17297 (TI mode splitters): Likewise.
17299 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
17302 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
17304 PR tree-optimization/56355
17305 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
17306 Also handle integers with undefined overflow.
17308 2013-03-20 Catherine Moore <clm@codesourcery.com>
17309 Maciej W. Rozycki <macro@codesourcery.com>
17310 Tom de Vries <tom@codesourcery.com>
17311 Nathan Sidwell <nathan@codesourcery.com>
17312 Iain Sandoe <iain@codesourcery.com>
17313 Nathan Froyd <froydnj@codesourcery.com>
17314 Chao-ying Fu <fu@mips.com>
17316 * doc/extend.texi: (micromips, nomicromips, nocompression):
17317 Document new function attributes.
17318 * doc/invoke.texi (minterlink-compressed, mmicromips,
17319 m14k, m14ke, m14kec): Document new options.
17320 (minterlink-mips16): Update documentation.
17321 * doc/md.texi (ZC, ZD): Document new constraints.
17322 * configure.ac (gcc_cv_as_micromips): Check if linker
17323 supports the .set micromips directive.
17324 * configure: Regenerate.
17325 * config.in: Regenerate.
17326 * config/mips/mips-tables.opt: Regenerate.
17327 * config/mips/micromips.md: New file.
17328 * constraints.md (ZC, ZD): New constraints.
17329 * config/mips/predicates.md (movep_src_register): New predicate.
17330 (movep_src_operand): New predicate.
17331 (non_volatile_mem_operand): New predicate.
17332 * config/mips/mips.md (multimem): New type.
17333 (length): Differentiate between 17-bit and 18-bit branch offsets.
17334 (MOVEP1, MOVEP2): New mode iterator.
17335 (mov_<load>l): Use ZC constraint.
17336 (mov_<load>r): Likewise.
17337 (mov_<store>l): Likewise.
17338 (mov_<store>r): Likewise.
17339 (*branch_equality<mode>_inverted): Add microMIPS support.
17340 (*branch_equality<mode>): Likewise.
17341 (*jump_absolute): Likewise.
17342 (indirect_jump_<mode>): Likewise.
17343 (tablejump_<mode>): Likewise.
17344 (<optab>_internal): Likewise.
17345 (sibcall_internal): Likewise.
17346 (sibcall_value_internal): Likewise.
17347 (prefetch): Use constraint ZD.
17348 * config/mips/mips.opt (minterlink-compressed): New option.
17349 (minterlink-mips16): Now an alias for minterlink-compressed.
17350 (mmicromips): New option.
17351 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17352 (compare_and_swap_12): Likewise.
17353 (sync_add<mode>): Likewise.
17354 (sync_<optab>_12): Likewise.
17355 (sync_old_<optab>_12): Likewise.
17356 (sync_new_<optab>_12): Likewise.
17357 (sync_nand_12): Likewise.
17358 (sync_old_nand_12): Likewise.
17359 (sync_new_nand_12): Likewise.
17360 (sync_sub<mode>): Likewise.
17361 (sync_old_add<mode>): Likewise.
17362 (sync_old_sub<mode>): Likewise.
17363 (sync_new_add<mode>): Likewise.
17364 (sync_new_sub<mode>): Likewise.
17365 (sync_<optab><mode>): Likewise.
17366 (sync_old_<optab><mode>): Likewise.
17367 (sync_new_<optab><mode>): Likewise.
17368 (sync_nand<mode>): Likewise.
17369 (sync_old_nand<mode>): Likewise.
17370 (sync_new_nand<mode>): Likewise.
17371 (sync_lock_test_and_set<mode>): Likewise.
17372 (test_and_set_12): Likewise.
17373 (atomic_compare_and_swap<mode>): Likewise.
17374 (atomic_exchange<mode>_llsc): Likewise.
17375 (atomic_fetch_add<mode>_llsc): Likewise.
17376 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17377 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17378 (umips_save_restore_pattern_p): Likewise.
17379 (umips_load_store_pair_p): Likewise.
17380 (umips_output_load_store_pair): Likewise.
17381 (umips_movep_target_p): Likewise.
17382 (umips_12bit_offset_address_p): Likewise.
17383 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17384 (mips_base_mips16): Rename this...
17385 (mips_base_compression_flags): ...to this. Update all uses.
17386 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17387 (mips_mips16_decl_p): Delete.
17388 (mips_nomips16_decl_p): Delete.
17389 (mips_get_compress_on_flags): New function.
17390 (mips_get_compress_off_flags): New function.
17391 (mips_get_compress_mode): New function.
17392 (mips_get_compress_on_name): New function.
17393 (mips_get_compress_off_name): New function.
17394 (mips_insert_attributes): Support multiple compression types.
17395 (mips_merge_decl_attributes): Likewise.
17396 (umips_12bit_offset_address_p): New function.
17397 (mips_start_function_definition): Emit .set micromips directive.
17398 (mips_call_may_need_jalx_p): New function.
17399 (mips_function_ok_for_sibcall): Add microMIPS support.
17400 (mips_print_operand_punctuation): Support short delay slots and
17402 (umips_swm_mask, umips_swm_encoding): New.
17403 (umips_build_save_restore): New function.
17404 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17405 (was_mips16_p): Remove.
17406 (old_compression_mode): New.
17407 (mips_set_compression_mode): New function.
17408 (mips_set_current_function): Add microMIPS support.
17409 (mips_option_override): Likewise.
17410 (umips_save_restore_pattern_p): New function.
17411 (umips_output_save_restore): New function.
17412 (umips_load_store_pair_p_1): New function.
17413 (umips_load_store_pair_p): New function.
17414 (umips_output_load_store_pair_1): New function.
17415 (umips_output_load_store_pair): New function.
17416 (umips_movep_target_p) New function.
17417 (mips_prepare_pch_save): Add microMIPS support.
17418 * config/mips/mips.h (TARGET_COMPRESSION): New.
17419 (TARGET_CPU_CPP_BUILTINS): Update macro
17420 to use new compression flags and to support microMIPS.
17421 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17422 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17423 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17424 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17425 (ASM_SPEC): Support mmicromips and mno-micromips.
17426 (M16STORE_REG_P): New macro.
17427 (MIPS_CALL): Support TARGET_MICROMIPS.
17428 (MICROMIPS_J): New macro.
17429 (mips_base_mips16): Rename this...
17430 (mips_base_compression_flags): ...to this.
17431 (UMIPS_12BIT_OFFSET_P): New macro.
17432 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17433 (MULTILIB_DIRNAMES): Likewise.
17434 2013-03-20 Richard Biener <rguenther@suse.de>
17436 PR tree-optimization/56661
17437 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
17438 the result does not have to be distinct.
17440 2013-03-20 Richard Biener <rguenther@suse.de>
17442 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
17445 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17446 Steven Bosscher <steven@gcc.gnu.org>
17448 PR rtl-optimization/56605
17449 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
17451 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17454 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
17455 that require movd instead of movq.
17457 2013-03-20 Richard Biener <rguenther@suse.de>
17459 * tree-ssa-structalias.c (struct variable_info): Add pointer
17460 to the first field of an aggregate with sub-vars. Make
17461 this and the pointer to the next subfield its ID.
17462 (vi_next): New function.
17463 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
17464 storedanything_id, integer_id): Increment by one.
17465 (new_var_info, get_call_vi, lookup_call_clobber_vi,
17466 get_call_clobber_vi): Adjust.
17467 (solution_set_expand): Simplify and speedup.
17468 (solution_set_add): Inline into ...
17469 (set_union_with_increment): ... this. Adjust accordingly.
17470 (do_sd_constraint): Likewise.
17471 (do_ds_constraint): Likewise.
17472 (do_complex_constraint): Simplify.
17473 (build_pred_graph): Adjust.
17474 (solve_graph): Likewise. Simplify and speedup.
17475 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
17476 get_constraint_for_component_ref, get_constraint_for_1,
17477 first_vi_for_offset, first_or_preceding_vi_for_offset,
17478 create_function_info_for, create_variable_info_for_1,
17479 create_variable_info_for, intra_create_variable_infos): Adjust.
17480 (init_base_vars): Push NULL for ID zero.
17481 (compute_points_to_sets): Adjust.
17483 2013-03-20 Richard Biener <rguenther@suse.de>
17485 * cfgloop.c (verify_loop_structure): Streamline and avoid
17486 ICEing on corrupt loop tree.
17487 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17490 2013-03-20 Richard Biener <rguenther@suse.de>
17492 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17493 check whether an SSA update is needed.
17495 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17497 * config/mips/constraints.md (T): Rename to...
17501 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17502 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17504 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17506 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17507 (*subsi3_carryin_uxtw): Likewise.
17509 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17511 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17512 (*rorsi3_insn_uxtw): Likewise.
17514 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17516 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17517 (*extrsi5_insn_uxtw): Likewise.
17519 2013-03-19 Richard Biener <rguenther@suse.de>
17521 PR tree-optimization/56273
17522 * passes.c (init_optimization_passes): Move second VRP after DOM.
17524 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
17526 * config/i386/i386.md (*movti_internal): Merge from
17527 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17528 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17529 nox64 isa attributes.
17531 2013-03-18 Richard Biener <rguenther@suse.de>
17533 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17535 (merge_node_constraints): Likewise.
17536 (build_succ_graph): Likewise.
17537 (valid_graph_edge): Inline into single caller.
17538 (unify_nodes): Likewise. Use bitmap_set_bit return value
17540 (scc_visit): Fix formatting and variable use.
17541 (do_sd_constraint): Use gcc_checking_assert.
17542 (do_ds_constraint): Likewise.
17543 (do_complex_constraint): Likewise.
17544 (condense_visit): Likewise. Cleanup.
17545 (dump_pred_graph): New function.
17546 (perform_var_substitution): Dump the pred-graph before
17547 variable substitution.
17548 (find_equivalent_node): Use gcc_checking_assert.
17549 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17551 2013-03-18 Richard Biener <rguenther@suse.de>
17553 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17554 Remove cond_expr_stmt_list argument and do not gimplify the
17556 (vect_loop_versioning): Adjust.
17557 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17558 Cleanup to use less temporaries.
17559 (vect_create_data_ref_ptr): Cleanup.
17561 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17563 PR tree-optimization/56635
17564 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17565 require types_compatible_p types.
17567 2013-03-18 Nick Clifton <nickc@redhat.com>
17569 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
17570 spurious backslash.
17572 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
17573 Add missing line to comment describing function.
17575 2013-03-18 Richard Biener <rguenther@suse.de>
17577 PR tree-optimization/56210
17578 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17579 Handle string / character search functions.
17580 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
17582 2013-03-18 Richard Biener <rguenther@suse.de>
17584 PR middle-end/56483
17585 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
17586 and implement properly.
17587 * gimple.h (gimple_cond_single_var_p): Remove.
17589 2013-03-18 Richard Biener <rguenther@suse.de>
17591 * tree-data-ref.h (find_data_references_in_loop): Declare.
17592 * tree-data-ref.c (get_references_in_stmt): Use a stack
17593 vector pre-allocated in the callers.
17594 (find_data_references_in_stmt): Adjust.
17595 (graphite_find_data_references_in_stmt): Likewise.
17596 (create_rdg_vertices): Likewise.
17597 (find_data_references_in_loop): Export.
17598 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
17599 Compute dependences here...
17600 (vect_analyze_data_refs): ...not here. When we encounter
17601 a non-vectorizable data reference in basic-block vectorization
17602 truncate the data reference vector. Do not bother to
17603 fixup data-dependence information for gather loads.
17604 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
17605 of data references, as reported.
17607 2013-03-18 Richard Biener <rguenther@suse.de>
17609 PR tree-optimization/3713
17610 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
17611 has_constants and expr.
17612 (stmt_has_constants): Properly valueize SSA names when deciding
17613 whether the stmt has constants.
17615 2013-03-18 Richard Biener <rguenther@suse.de>
17617 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
17618 whole function when there is nothing to do.
17619 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
17620 * tree-vectorizer.c (vectorize_loops): Update virtual and
17621 loop-closed SSA once.
17622 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
17624 2013-03-18 Richard Biener <rguenther@suse.de>
17626 PR middle-end/56113
17627 * domwalk.c (bb_postorder): New global static.
17628 (cmp_bb_postorder): New function.
17629 (walk_dominator_tree): Replace scheme imposing an order for
17630 visiting dominator sons by one sorting them at the time they
17631 are pushed on the stack.
17633 2013-03-18 Richard Biener <rguenther@suse.de>
17635 PR tree-optimization/39326
17636 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17637 (struct mem_ref): Replace mem member with ao_ref typed member.
17638 (MEM_ANALYZABLE): Adjust.
17639 (memref_eq): Likewise.
17640 (mem_ref_alloc): Likewise.
17641 (gather_mem_refs_stmt): Likewise.
17642 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17643 (execute_sm_if_changed_flag_set): Adjust.
17644 (execute_sm): Likewise.
17645 (ref_always_accessed_p): Likewise.
17646 (refs_independent_p): Likewise.
17647 (can_sm_ref_p): Likewise.
17649 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17652 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17653 return 1 even for !unsignedp.
17655 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17657 * config/i386/i386.md (isa): Add x64 and nox64.
17658 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17659 (*pushtf): Enable *roF alternative for x64 isa only.
17660 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17661 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17662 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17663 (*movtf_internal): Merge from *movtf_internal_rex64 and
17664 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17665 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17666 nox64 isa attributes.
17667 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17668 nox64 isa attributes.
17669 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17671 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17673 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17675 2013-03-17 Jakub Jelinek <jakub@redhat.com>
17678 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17679 class names. Remove trailing comma after "ALL_REGS".
17681 2013-03-16 Jan Hubicka <jh@suse.cz>
17683 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17684 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17685 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17686 of cgraph_get_create_node.
17687 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17689 2013-03-16 Jason Merrill <jason@redhat.com>
17692 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17693 with DW_AT_default_value.
17695 2013-03-16 Jakub Jelinek <jakub@redhat.com>
17697 * BASE-VER: Set to 4.9.0.
17699 2013-03-14 Andi Kleen <ak@linux.intel.com>
17702 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17703 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17704 Document _x* TSX intrinsics.
17706 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17707 David Holsgrove <david.holsgrove@xilinx.com>
17709 * configure.ac: Add MicroBlaze TLS support detection.
17710 * configure: Regenerate.
17711 * config/microblaze/microblaze-protos.h
17712 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17713 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17714 * config/microblaze/microblaze.c (microblaze_address_type): Add
17715 ADDRESS_TLS and tls_reloc address types.
17716 (microblaze_address_info): Add tls_reloc.
17717 (TARGET_HAVE_TLS): Define.
17718 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17719 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
17720 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17721 load_tls_operand, microblaze_call_tls_get_addr,
17722 microblaze_legitimize_tls_address): New functions.
17723 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17724 (get_base_reg): Use microblaze_tls_symbol_p.
17725 (microblaze_classify_address): Handle TLS.
17726 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17727 label_mentioned_p and microblaze_tls_referenced_p.
17728 (microblaze_legitimize_address): Handle TLS.
17729 (microblaze_address_insns): Handle ADDRESS_TLS.
17730 (pic_address_needs_scratch): Handle TLS.
17731 (print_operand_address): Handle TLS.
17732 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17733 (microblaze_expand_move): Handle TLS.
17734 (microblaze_legitimate_constant_p): Check
17735 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
17736 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17737 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17738 (PIC_OFFSET_TABLE_REGNUM): Set.
17739 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17740 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17741 (addsi3, movsi_internal2, movdf_internal): Update constraints
17742 * config/microblaze/predicates.md (arith_plus_operand): Define
17743 (move_operand): Redefine as move_src_operand,
17744 check microblaze_tls_referenced_p.
17746 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17748 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17749 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17751 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17753 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17756 2013-03-14 Jakub Jelinek <jakub@redhat.com>
17758 PR tree-optimization/53265
17759 * common.opt (Waggressive-loop-optimizations): New option.
17760 * tree-ssa-loop-niter.c: Include tree-pass.h.
17761 (do_warn_aggressive_loop_optimizations): New function.
17762 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17763 if number_of_latch_executions returned constant.
17764 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17765 early. If number_of_latch_executions returned constant, set
17766 nb_iterations_upper_bound back to it.
17767 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17769 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17770 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17772 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17773 (MULTILIB_OSDIRNAMES): Set.
17774 * genmultilib: If defaultosdirname doesn't start with :: , set
17775 defaultosdirname2 instead, clear it and emit two . multilib_raw
17776 entries instead of just one.
17778 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17780 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17781 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17782 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17783 (SUBTARGET_OVERRIDE_OPTIONS): New.
17785 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17788 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17789 (musermode): Convert to Var(TARGET_USERMODE).
17790 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17791 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17792 * config/sh/sh.c (sh_option_override): Use
17793 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17794 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17796 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17798 (udivsi3_i4_single, divsi3_i4_single): Use
17799 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17801 2013-03-13 Dave Korn <dave.korn.cygwin@....>
17803 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17806 2013-03-13 Richard Biener <rguenther@suse.de>
17808 PR tree-optimization/56608
17809 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17810 calls when vectorizing basic-blocks.
17812 2013-03-13 Jakub Jelinek <jakub@redhat.com>
17815 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17818 2013-03-12 Jakub Jelinek <jakub@redhat.com>
17820 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17822 2013-03-11 Jan Hubicka <jh@suse.cz>
17825 * lto-streamer-out.c (output_symbol_p): Skip references from
17826 constructors of external variables.
17828 2013-03-11 Jan Hubicka <jh@suse.cz>
17830 PR middle-end/56571
17831 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
17833 * emit-rtl.c (verify_rtx_sharing): Likewise.
17834 (copy_insn_1): Likewise.
17835 * rtl.c (copy_rtx): Likewise.
17837 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
17840 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
17841 output_operand_lossage message.
17843 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
17846 * arm.c (shift_op): Validate RTL pattern on the fly.
17847 (arm_print_operand, case 'S'): Don't use shift_operator to validate
17850 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17853 * config/pa/pa.md (call_value): Check for calls to powf and direct to
17854 new call patterns that clobber %fr12.
17855 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
17856 split and postreload patterns.
17857 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
17858 registers %fr12 and %fr12R as call used.
17860 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
17862 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
17863 (canon_address, record_store, replace_read, check_mem_read_rtx,
17864 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
17865 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
17866 rest_of_handle_dse): Likewise.
17868 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
17870 PR middle-end/56524
17871 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
17873 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
17874 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
17875 (save_optabs_if_changed): Replace with...
17876 (init_tree_optimization_optabs): ...this.
17877 * optabs.c (save_optabs_if_changed): Rename to...
17878 (init_tree_optimization_optabs): ...this. Take the optimization node
17879 as argument. Do nothing if the base optabs are already correct.
17880 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
17881 to recompute optabs.
17882 * function.h (function): Remove optabs field.
17883 * function.c (invoke_set_current_function_hook): Call
17884 init_tree_optimization_optabs. Use the result to initialize
17887 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
17889 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
17890 if GTMA_HAS_NO_INSTRUMENTATION.
17891 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
17892 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
17893 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
17894 * gimple-pretty-print.c (dump_gimple_transaction): Handle
17895 GTMA_HAS_NO_INSTRUMENTATION.
17897 2013-03-08 Jakub Jelinek <jakub@redhat.com>
17899 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
17902 2013-03-08 Marek Polacek <polacek@redhat.com>
17903 Jakub Jelinek <jakub@redhat.com>
17905 PR tree-optimization/56478
17906 * predict.c (is_comparison_with_loop_invariant_p): Change the
17907 type of loop_step to tree.
17908 (predict_loops): Adjust.
17909 (predict_iv_comparison): Perform the computations on double_ints.
17911 2013-03-08 Richard Biener <rguenther@suse.de>
17913 PR tree-optimization/56570
17914 * tree-cfg.c (verify_expr_location_1): Verify locations for
17916 * tree-sra.c (create_access_replacement): Strip locations
17917 from DECL_DEBUG_EXPRs.
17919 2013-03-08 Richard Biener <rguenther@suse.de>
17921 * tree-inline.c (expand_call_inline): Do not associate
17922 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
17923 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
17925 2013-03-08 Richard Biener <rguenther@suse.de>
17927 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
17928 or block changes with -Og. Fix for location / block encoding
17929 changes and PHI arguments with locations.
17931 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
17933 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
17935 (struct output_info): Likewise.
17936 (register_overhead): Remove bad gcc_assert.
17937 (bitmap_find_bit): If there is only a single bitmap element, do not
17938 count a miss as a search.
17939 (print_statistics): Update for counter type changes.
17940 (dump_bitmap_statistics): Likewise. Print headers such that they
17941 are properly lined up with the printed counters.
17943 2013-03-07 Jakub Jelinek <jakub@redhat.com>
17945 PR tree-optimization/56559
17946 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
17947 check that it has only a single use.
17949 2013-03-07 Richard Biener <rguenther@suse.de>
17951 * doc/invoke.texi (fwhole-program): Discourage use in combination
17954 2013-03-06 Jakub Jelinek <jakub@redhat.com>
17956 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
17958 PR tree-optimization/56539
17959 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
17960 instead of GSI_CONTINUE_LINKING as last argument to
17961 force_gimple_operand_gsi. Adjust function comment.
17963 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
17966 PR middle-end/56548
17967 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
17968 promoted mode, convert the result back to the original mode.
17970 2013-03-06 Richard Biener <rguenther@suse.de>
17972 PR middle-end/56294
17973 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
17974 (insert_updated_phi_nodes_compare_uids): New function.
17975 (update_ssa): Sort symbols_to_rename after UID before
17976 traversing it to insert PHI nodes.
17978 2013-03-06 Richard Biener <rguenther@suse.de>
17980 PR middle-end/50494
17981 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
17982 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
17985 2013-02-13 Richard Biener <rguenther@suse.de>
17988 * varasm.c (output_constant_def_1): Get the decl representing
17989 the constant as argument.
17990 (output_constant_def): Wrap output_constant_def_1.
17991 (make_decl_rtl): Use output_constant_def_1 with the decl
17992 representing the constant.
17993 (build_constant_desc): Optionally re-use a decl already
17994 representing the constant.
17995 (tree_output_constant_def): Adjust.
17997 2013-03-06 Joey Ye <joey.ye@arm.com>
18000 * gcc.c (convert_white_space): New function.
18001 (main): Handles white space in function name.
18003 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
18006 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
18007 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
18008 to SH_DIV_CALL_TABLE for TARGET_SH2.
18009 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
18011 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
18012 call-table options.
18014 2013-03-05 Sterling Augustine <saugustine@google.com>
18015 Cary Coutant <ccoutant@google.com>
18018 * dwarf2out.c (resolve_addr): Don't call
18019 remove_loc_list_addr_table_entries a second time for the same
18022 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18025 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
18026 (avoid_complex_debug_insns): New function.
18027 (expand_debug_locations): Call it.
18029 PR rtl-optimization/56484
18030 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
18031 lifetimes of hard registers on small register class machines.
18033 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18035 * config/microblaze/microblaze-protos.h: Rename
18036 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
18037 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
18039 (microblaze_fast_interrupt_function_p): New function.
18040 (microblaze_is_interrupt_handler): Rename to
18041 microblaze_is_interrupt_variant and add fast_interrupt check.
18042 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
18043 (save_restore_insns): Likewise.
18044 (compute_frame_size): Likewise.
18045 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
18046 (microblaze_globalize_label): Likewise.
18047 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
18048 * config/microblaze/microblaze.md: Use wrapper
18049 microblaze_is_interrupt_variant.
18051 2013-03-05 Kai Tietz <ktietz@redhat.com>
18053 * sdbout.c (sdbout_one_type): Switch to current function's section
18054 supporting cold/hot.
18056 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18058 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
18061 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18063 PR middle-end/56461
18064 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
18065 if VALGRIND_GET_VBITS is defined, temporarily make object
18066 memory all defined, and restore previous valgrind addressability
18067 and definability afterwards. Free this_object at the end.
18069 PR middle-end/56461
18070 * lra.c (lra): Call lra_clear_live_ranges if live_p,
18071 right before calling lra_create_live_ranges, also call it
18072 when clearing live_p. Only call lra_clear_live_ranges
18073 at the end if live_p.
18075 PR middle-end/56461
18076 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
18078 2013-03-05 Richard Biener <rguenther@suse.de>
18080 PR tree-optimization/56521
18081 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
18084 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
18087 * except.h (remove_unreachable_eh_regions): New prototype.
18088 * except.c (remove_eh_handler_splicer): New function, split out
18089 of remove_eh_handler.
18090 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
18091 warning about running it on many EH regions one at a time.
18092 (remove_unreachable_eh_regions_worker): New function, walk the
18093 EH tree in depth-first order and remove non-marked regions.
18094 (remove_unreachable_eh_regions): New function.
18095 * tree-eh.c (mark_reachable_handlers): New function, split out
18096 from remove_unreachable_handlers.
18097 (remove_unreachable_handlers): Use mark_reachable_handlers and
18098 remove_unreachable_eh_regions.
18099 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
18100 and remove_unreachable_eh_regions.
18102 2013-03-05 Richard Biener <rguenther@suse.de>
18104 PR middle-end/56525
18105 * loop-init.c (fix_loop_structure): Remove loops in two stages,
18106 not freeing them until the end.
18108 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18110 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
18112 2013-03-05 Richard Biener <rguenther@suse.de>
18114 PR tree-optimization/56270
18115 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
18116 of loads after scheduling an SLP instance.
18118 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18120 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
18122 (check_gcc_parallelize): Run guality.exp as a separate job from
18123 vect.exp with unsorted.exp and $(dg_target_exps) separately from
18124 struct-layout-1.exp with stackalign.exp.
18126 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
18128 PR middle-end/56461
18129 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
18130 load_index sbitmap even if some bit in it isn't set.
18132 PR middle-end/56461
18133 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
18134 (discover_iteration_bound_by_body_walk): Change queues to
18135 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
18136 spelling in comment. Call safe_push on queues[bound_index] directly.
18137 Release queues[queue_index] in every iteration unconditionally.
18138 Release bounds vector.
18140 PR middle-end/56461
18141 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
18142 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
18143 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
18146 2013-03-05 Richard Biener <rguenther@suse.de>
18149 * tree-inline.c (remap_blocks_to_null): New function.
18150 (expand_call_inline): When expanding a call stmt without
18151 an associated block inline remap all callee blocks to NULL.
18153 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18155 PR rtl-optimization/56494
18156 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
18157 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
18158 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
18160 PR middle-end/56461
18161 * sel-sched-ir.c (free_sched_pools): Release
18162 succs_info_pool.stack[succs_info_pool.max_top] vectors too
18163 if succs_info_pool.max_top isn't -1.
18166 * opts.c (opts_obstack, opts_concat): Moved to...
18167 * opts-common.c (opts_obstack, opts_concat): ... here.
18169 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18171 PR middle-end/56461
18172 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
18174 2013-03-04 Martin Jambor <mjambor@suse.cz>
18176 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
18177 all appropriate places.
18179 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
18181 PR tree-optimization/56424
18182 * ipa-split.c (split_function): Do not set the RSO flag if result is
18183 not by reference and its type is a register type.
18185 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
18187 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
18188 (microblaze_legitimate_pic_operand): Likewise
18189 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
18190 new function microblaze_legitimate_pic_operand
18191 * config/microblaze/microblaze-protos.h
18192 (microblaze_legitimate_pic_operand): Declare.
18194 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18196 * config/microblaze/predicates.md (call_insn_simple_operand):
18197 New predicate for supported rtx code types.
18198 * config/microblaze/microblaze.md (call_internal1): Use
18199 call_insn_simple_operand predicate.
18201 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18203 PR middle-end/56461
18204 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
18205 partitions.ordered_remove.
18207 PR middle-end/56461
18208 * tree-vect-stmts.c (vectorizable_conversion): Don't call
18209 vec_oprnds0.create (1) for modifier == NONE.
18211 PR middle-end/56461
18212 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
18213 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
18214 vec_oprnds1 right before pushing anything to it for
18217 PR middle-end/56461
18218 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
18219 set nbbs to 0 instead of having separate code path.
18220 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
18221 instead of false as last argument if returning NULL.
18223 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
18225 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
18226 the attribute is now called "target" instead of "option".
18227 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
18228 * doc/tm.texi.in (Target Attributes): Likewise document the correct
18229 attribute/pragma name for TARGET_OPTION_VALID_P and
18230 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
18231 * doc/tm.texi: Regenerated.
18233 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
18235 * config/microblaze/microblaze.c:
18236 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
18237 * config/microblaze/microblaze.h: Add -mxl-reorder to
18239 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
18240 instructions emitted if TARGET_REORDER.
18241 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
18242 or 0 for -m/-mno case, but initialises as 2 to detect default use case
18245 2013-03-01 Xinliang David Li <davidxl@google.com>
18247 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
18250 2013-03-01 Jakub Jelinek <jakub@redhat.com>
18252 PR middle-end/56461
18253 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
18254 vector even when returning true. Fix up function comment formatting.
18256 PR middle-end/56461
18257 * ira-build.c (ira_loop_nodes_count): New variable.
18258 (create_loop_tree_nodes): Initialize it.
18259 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
18261 PR middle-end/56461
18262 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
18263 method on dr_chain and result_chain.
18264 * tree-vect-stmts.c (vectorizable_store): Only call
18265 result_chain.create if j == 0.
18267 PR middle-end/56461
18268 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
18269 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
18270 before overwriting it.
18272 2013-03-01 Tobias Burnus <burnus@net-b.de>
18274 * doc/extended.texi (C Extensions): Change order in @menu
18276 (Other MIPS Built-in Functions): Move last MIPS entry before
18277 "picoChip Built-in Functions".
18278 (SH Built-in Functions): Move after RX Built-in Functions.
18279 * doc/gcc.texi (Introduction): Change order in @menu
18281 * doc/md.texi (Constraints): Ditto.
18282 * gty.texi (Type Information): Ditto.
18283 (User-provided marking routines for template types): Make
18285 * doc/invoke.texi (AArch64 Options): Move before
18286 "Adapteva Epiphany Options".
18288 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
18289 Jakub Jelinek <jakub@redhat.com>
18292 * asan.c (gate_asan): Lookup no_sanitize_address instead of
18293 no_address_safety_analysis attribute.
18294 * doc/extend.texi (no_address_safety_attribute): Rename to
18295 no_sanitize_address attribute, mention no_address_safety_analysis
18296 attribute as deprecated alias.
18298 2013-02-28 Jakub Jelinek <jakub@redhat.com>
18300 PR middle-end/56461
18301 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
18302 type to vec<vec<tree> > *.
18303 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
18304 to be vec<tree> instead of vec<tree> *, set vec_defs
18305 to vNULL and call vec_defs.create (number_of_vects), adjust other
18307 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
18308 vectorizable_condition): Adjust vect_get_slp_defs callers.
18310 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18312 * config/aarch64/aarch64.c
18313 (aarch64_float_const_representable): Remove unused variable.
18315 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18317 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
18319 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18321 * config/aarch64/aarch64-builtins.c
18322 (aarch64_init_simd_builtins): Make static.
18324 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18326 * config/aarch64/aarch64.c
18327 (aarch64_simd_make_constant): Make static.
18329 2013-02-28 Martin Jambor <mjambor@suse.cz>
18331 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
18332 with no initialization to the RHS of debug statements.
18334 2013-02-28 Martin Jambor <mjambor@suse.cz>
18336 PR tree-optimization/56294
18337 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
18339 (get_access_replacement): Do not call create_access_replacement.
18340 Assert a replacement exists.
18341 (get_repl_default_def_ssa_name): Create the replacement declaration
18344 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18346 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
18347 final_end_function.
18349 2013-02-28 Marek Polacek <polacek@redhat.com>
18351 PR rtl-optimization/56466
18352 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
18353 if we're changing a loop.
18354 (peel_loops_completely): Likewise.
18356 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
18359 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
18361 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18364 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
18365 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
18366 INTX_FTYPE_FX, FX_FTYPE_INTX.
18367 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
18369 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18371 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
18372 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18373 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
18374 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
18375 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
18376 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
18377 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
18378 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
18379 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
18380 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
18381 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
18382 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
18383 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
18384 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
18385 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
18386 (avrxmega6): Increase max flash segments from 5 to 6.
18387 * config/avr/t-multilib: Regenerate.
18388 * config/avr/avr-tables.opt: Regenerate.
18389 * doc/avr-mmcu.texi: Regenerate.
18391 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18393 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
18394 (avr_device_to_arch): Rename to avr_device_to_ld.
18395 (avr_device_to_as): New prototype.
18396 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
18397 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
18398 * config/avr/driver-avr.c (avr_device_to_as): New.
18399 (avr_device_to_arch): Rename to avr_device_to_ld.
18401 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18403 PR middle-end/56461
18404 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
18405 method on dr_chain and result_chain.
18407 PR middle-end/56461
18408 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
18409 pointer_set_destroy on not_executed_last_iteration.
18411 PR middle-end/56461
18412 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
18414 PR middle-end/56461
18415 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
18416 FOR_EACH_DEFINED_FUNCTION when freeing state.
18418 PR middle-end/56461
18419 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
18421 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
18424 PR middle-end/56461
18425 * ipa-cp.c (decide_whether_version_node): Call vec_free on
18426 known_aggs[i].items and release known_aggs vector.
18428 PR middle-end/56461
18429 * ipa-reference.c (propagate): Free node_info even for alias nodes.
18431 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18433 * config/microblaze/microblaze.c (microblaze_emit_compare):
18434 Use xor for EQ/NE comparisions.
18435 * config/microblaze/microblaze.md (cstoresf4): Add constraints
18436 (cbranchsf4): Adjust operator to comparison_operator.
18438 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18440 PR middle-end/56461
18441 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
18443 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
18444 vec_safe_push, always update *slot.
18445 (redirect_edge_var_map_clear): Use vec_free.
18446 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
18447 (free_var_map_entry): Use vec_free.
18448 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
18449 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
18451 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
18453 PR middle-end/45472
18454 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
18455 when the may_trap_p bit of the exprs being merged differs.
18456 Reorder tests for speculativeness in the logical and operator.
18458 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18460 * incpath.c (add_standard_paths): Use reconcat instead of concat
18461 where appropriate and avoid leaking memory.
18463 * opts.h: Include obstack.h.
18464 (opts_concat): New prototype.
18465 (opts_obstack): New declaration.
18466 * opts.c (opts_concat): New function.
18467 (opts_obstack): New variable.
18468 (init_options_struct): Call gcc_init_obstack on opts_obstack.
18469 (finish_options): Use opts_concat instead of concat
18470 and XOBNEWVEC instead of XNEWVEC.
18471 * opts-common.c (generate_canonical_option, decode_cmdline_option,
18472 generate_option): Likewise.
18473 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
18474 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
18477 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18478 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18480 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18482 PR middle-end/56461
18483 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18485 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18487 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18488 (arm_block_move_unaligned_straight): Likewise.
18489 (arm_adjust_block_mem): Likewise.
18491 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18494 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18495 temp, cond and label.
18496 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18499 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18500 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18503 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18504 of prologue/epilogue functions.
18507 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18508 Remove unused variable cfa_offset.
18509 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18512 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18513 type promotion to unsigned.
18516 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18517 for HOST_WIDE_INT of 32 bit / same size as int.
18518 (arm_block_move_unaligned_straight): Likewise.
18519 (arm_adjust_block_mem): Likewise.
18522 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18525 2013-02-26 Marek Polacek <polacek@redhat.com>
18527 PR tree-optimization/56426
18528 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
18530 2013-02-26 Richard Biener <rguenther@suse.de>
18533 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18534 unused variable loops.
18536 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18538 PR tree-optimization/56448
18539 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18540 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18541 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18542 later operands of the references, or even first operand for
18543 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18545 PR tree-optimization/56443
18546 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18547 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18548 to type_for_mode langhook.
18550 2013-02-25 Matt Turner <mattst88@gmail.com>
18552 * doc/invoke.texi: Document r4700.
18554 2013-02-25 Richard Biener <rguenther@suse.de>
18556 PR tree-optimization/56175
18557 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18559 (simplify_bitwise_binary): ... here. Also guard the conversion
18560 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18562 2013-02-25 Catherine Moore <clm@codesourcery.com>
18565 2013-02-24 Catherine Moore <clm@codesourcery.com>
18566 Maciej W. Rozycki <macro@codesourcery.com>
18567 Tom de Vries <tom@codesourcery.com>
18568 Nathan Sidwell <nathan@codesourcery.com>
18569 Iain Sandoe <iain@codesourcery.com>
18570 Nathan Froyd <froydnj@codesourcery.com>
18571 Chao-ying Fu <fu@mips.com>
18573 * doc/extend.texi: (micromips, nomicromips, nocompression):
18574 Document new function attributes.
18575 * doc/invoke.texi (minterlink-compressed, mmicromips,
18576 m14k, m14ke, m14kec): Document new options.
18577 (minterlink-mips16): Update documentation.
18578 * doc/md.texi (ZC, ZD): Document new constraints.
18579 * configure.ac (gcc_cv_as_micromips): Check if linker
18580 supports the .set micromips directive.
18581 * configure: Regenerate.
18582 * config.in: Regenerate.
18583 * config/mips/mips-tables.opt: Regenerate.
18584 * config/mips/micromips.md: New file.
18585 * constraints.md (ZC, AD): New constraints.
18586 * config/mips/predicates.md (movep_src_register): New predicate.
18587 (movep_src_operand): New predicate.
18588 (non_volatile_mem_operand): New predicate.
18589 * config/mips/mips.md (multimem): New type.
18590 (length): Differentiate between 17-bit and 18-bit branch offsets.
18591 (MOVEP1, MOVEP2): New mode iterator.
18592 (mov_<load>l): Use ZC constraint.
18593 (mov_<load>r): Likewise.
18594 (mov_<store>l): Likewise.
18595 (mov_<store>r): Likewise.
18596 (*branch_equality<mode>_inverted): Add microMIPS support.
18597 (*branch_equality<mode>): Likewise.
18598 (*jump_absolute): Likewise.
18599 (indirect_jump_<mode>): Likewise.
18600 (tablejump_<mode>): Likewise.
18601 (<optab>_internal): Likewise.
18602 (sibcall_internal): Likewise.
18603 (sibcall_value_internal): Likewise.
18604 (prefetch): Use constraint ZD.
18605 * config/mips/mips.opt (minterlink-compressed): New option.
18606 (minterlink-mips16): Now an alias for minterlink-compressed.
18607 (mmicromips): New option.
18608 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18609 (compare_and_swap_12): Likewise.
18610 (sync_add<mode>): Likewise.
18611 (sync_<optab>_12): Likewise.
18612 (sync_old_<optab>_12): Likewise.
18613 (sync_new_<optab>_12): Likewise.
18614 (sync_nand_12): Likewise.
18615 (sync_old_nand_12): Likewise.
18616 (sync_new_nand_12): Likewise.
18617 (sync_sub<mode>): Likewise.
18618 (sync_old_add<mode>): Likewise.
18619 (sync_old_sub<mode>): Likewise.
18620 (sync_new_add<mode>): Likewise.
18621 (sync_new_sub<mode>): Likewise.
18622 (sync_<optab><mode>): Likewise.
18623 (sync_old_<optab><mode>): Likewise.
18624 (sync_new_<optab><mode>): Likewise.
18625 (sync_nand<mode>): Likewise.
18626 (sync_old_nand<mode>): Likewise.
18627 (sync_new_nand<mode>): Likewise.
18628 (sync_lock_test_and_set<mode>): Likewise.
18629 (test_and_set_12): Likewise.
18630 (atomic_compare_and_swap<mode>): Likewise.
18631 (atomic_exchange<mode>_llsc): Likewise.
18632 (atomic_fetch_add<mode>_llsc): Likewise.
18633 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18634 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18635 (umips_save_restore_pattern_p): Likewise.
18636 (umips_load_store_pair_p): Likewise.
18637 (umips_output_load_store_pair): Likewise.
18638 (umips_movep_target_p): Likewise.
18639 (umips_12bit_offset_address_p): Likewise.
18640 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18641 (mips_base_mips16): Rename this...
18642 (mips_base_compression_flags): ...to this. Update all uses.
18643 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18644 (mips_mips16_decl_p): Delete.
18645 (mips_nomips16_decl_p): Delete.
18646 (mips_get_compress_on_flags): New function.
18647 (mips_get_compress_off_flags): New function.
18648 (mips_get_compress_mode): New function.
18649 (mips_get_compress_on_name): New function.
18650 (mips_get_compress_off_name): New function.
18651 (mips_insert_attributes): Support multiple compression types.
18652 (mips_merge_decl_attributes): Likewise.
18653 (umips_12bit_offset_address_p): New function.
18654 (mips_start_function_definition): Emit .set micromips directive.
18655 (mips_call_may_need_jalx_p): New function.
18656 (mips_function_ok_for_sibcall): Add microMIPS support.
18657 (mips_print_operand_punctuation): Support short delay slots and
18659 (umips_swm_mask, umips_swm_encoding): New.
18660 (umips_build_save_restore): New function.
18661 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18662 (was_mips16_p): Remove.
18663 (old_compression_mode): New.
18664 (mips_set_compression_mode): New function.
18665 (mips_set_current_function): Add microMIPS support.
18666 (mips_option_override): Likewise.
18667 (umips_save_restore_pattern_p): New function.
18668 (umips_output_save_restore): New function.
18669 (umips_load_store_pair_p_1): New function.
18670 (umips_load_store_pair_p): New function.
18671 (umips_output_load_store_pair_1): New function.
18672 (umips_output_load_store_pair): New function.
18673 (umips_movep_target_p) New function.
18674 (mips_prepare_pch_save): Add microMIPS support.
18675 * config/mips/mips.h (TARGET_COMPRESSION): New.
18676 (TARGET_CPU_CPP_BUILTINS): Update macro
18677 to use new compression flags and to support microMIPS.
18678 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18679 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18680 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18681 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18682 (ASM_SPEC): Support mmicromips and mno-micromips.
18683 (M16STORE_REG_P): New macro.
18684 (MIPS_CALL): Support TARGET_MICROMIPS.
18685 (MICROMIPS_J): New macro.
18686 (mips_base_mips16): Rename this...
18687 (mips_base_compression_flags): ...to this.
18688 (UMIPS_12BIT_OFFSET_P): New macro.
18689 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18690 (MULTILIB_DIRNAMES): Likewise.
18692 2013-02-25 Tom de Vries <tom@codesourcery.com>
18694 PR rtl-optimization/56131
18695 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18696 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18697 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18699 2013-02-25 Tobias Burnus <burnus@net-b.de>
18701 * doc/invoke.texi (-fsanitize=): Move from optimization
18702 to debugging options.
18704 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18706 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18708 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18709 Alexander Monakov <amonakov@ispras.ru>
18711 PR middle-end/56077
18712 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18713 flush pending lists also on non-jumps. Adjust comment.
18715 2013-02-24 Catherine Moore <clm@codesourcery.com>
18716 Maciej W. Rozycki <macro@codesourcery.com>
18717 Tom de Vries <tom@codesourcery.com>
18718 Nathan Sidwell <nathan@codesourcery.com>
18719 Iain Sandoe <iain@codesourcery.com>
18720 Nathan Froyd <froydnj@codesourcery.com>
18721 Chao-ying Fu <fu@mips.com>
18723 * doc/extend.texi: (micromips, nomicromips, nocompression):
18724 Document new function attributes.
18725 * doc/invoke.texi (minterlink-compressed, mmicromips,
18726 m14k, m14ke, m14kec): Document new options.
18727 (minterlink-mips16): Update documentation.
18728 * doc/md.texi (ZC, ZD): Document new constraints.
18729 * configure.ac (gcc_cv_as_micromips): Check if linker
18730 supports the .set micromips directive.
18731 * configure: Regenerate.
18732 * config.in: Regenerate.
18733 * config/mips/mips-tables.opt: Regenerate.
18734 * config/mips/micromips.md: New file.
18735 * constraints.md (ZC, AD): New constraints.
18736 * config/mips/predicates.md (movep_src_register): New predicate.
18737 (movep_src_operand): New predicate.
18738 (non_volatile_mem_operand): New predicate.
18739 * config/mips/mips.md (multimem): New type.
18740 (length): Differentiate between 17-bit and 18-bit branch offsets.
18741 (MOVEP1, MOVEP2): New mode iterator.
18742 (mov_<load>l): Use ZC constraint.
18743 (mov_<load>r): Likewise.
18744 (mov_<store>l): Likewise.
18745 (mov_<store>r): Likewise.
18746 (*branch_equality<mode>_inverted): Add microMIPS support.
18747 (*branch_equality<mode>): Likewise.
18748 (*jump_absolute): Likewise.
18749 (indirect_jump_<mode>): Likewise.
18750 (tablejump_<mode>): Likewise.
18751 (<optab>_internal): Likewise.
18752 (sibcall_internal): Likewise.
18753 (sibcall_value_internal): Likewise.
18754 (prefetch): Use constraint ZD.
18755 * config/mips/mips.opt (minterlink-compressed): New option.
18756 (minterlink-mips16): Now an alias for minterlink-compressed.
18757 (mmicromips): New option.
18758 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18759 (compare_and_swap_12): Likewise.
18760 (sync_add<mode>): Likewise.
18761 (sync_<optab>_12): Likewise.
18762 (sync_old_<optab>_12): Likewise.
18763 (sync_new_<optab>_12): Likewise.
18764 (sync_nand_12): Likewise.
18765 (sync_old_nand_12): Likewise.
18766 (sync_new_nand_12): Likewise.
18767 (sync_sub<mode>): Likewise.
18768 (sync_old_add<mode>): Likewise.
18769 (sync_old_sub<mode>): Likewise.
18770 (sync_new_add<mode>): Likewise.
18771 (sync_new_sub<mode>): Likewise.
18772 (sync_<optab><mode>): Likewise.
18773 (sync_old_<optab><mode>): Likewise.
18774 (sync_new_<optab><mode>): Likewise.
18775 (sync_nand<mode>): Likewise.
18776 (sync_old_nand<mode>): Likewise.
18777 (sync_new_nand<mode>): Likewise.
18778 (sync_lock_test_and_set<mode>): Likewise.
18779 (test_and_set_12): Likewise.
18780 (atomic_compare_and_swap<mode>): Likewise.
18781 (atomic_exchange<mode>_llsc): Likewise.
18782 (atomic_fetch_add<mode>_llsc): Likewise.
18783 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18784 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18785 (umips_save_restore_pattern_p): Likewise.
18786 (umips_load_store_pair_p): Likewise.
18787 (umips_output_load_store_pair): Likewise.
18788 (umips_movep_target_p): Likewise.
18789 (umips_12bit_offset_address_p): Likewise.
18790 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18791 (mips_base_mips16): Rename this...
18792 (mips_base_compression_flags): ...to this. Update all uses.
18793 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18794 (mips_mips16_decl_p): Delete.
18795 (mips_nomips16_decl_p): Delete.
18796 (mips_get_compress_on_flags): New function.
18797 (mips_get_compress_off_flags): New function.
18798 (mips_get_compress_mode): New function.
18799 (mips_get_compress_on_name): New function.
18800 (mips_get_compress_off_name): New function.
18801 (mips_insert_attributes): Support multiple compression types.
18802 (mips_merge_decl_attributes): Likewise.
18803 (umips_12bit_offset_address_p): New function.
18804 (mips_start_function_definition): Emit .set micromips directive.
18805 (mips_call_may_need_jalx_p): New function.
18806 (mips_function_ok_for_sibcall): Add microMIPS support.
18807 (mips_print_operand_punctuation): Support short delay slots and
18809 (umips_swm_mask, umips_swm_encoding): New.
18810 (umips_build_save_restore): New function.
18811 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18812 (was_mips16_p): Remove.
18813 (old_compression_mode): New.
18814 (mips_set_compression_mode): New function.
18815 (mips_set_current_function): Add microMIPS support.
18816 (mips_option_override): Likewise.
18817 (umips_save_restore_pattern_p): New function.
18818 (umips_output_save_restore): New function.
18819 (umips_load_store_pair_p_1): New function.
18820 (umips_load_store_pair_p): New function.
18821 (umips_output_load_store_pair_1): New function.
18822 (umips_output_load_store_pair): New function.
18823 (umips_movep_target_p) New function.
18824 (mips_prepare_pch_save): Add microMIPS support.
18825 * config/mips/mips.h (TARGET_COMPRESSION): New.
18826 (TARGET_CPU_CPP_BUILTINS): Update macro
18827 to use new compression flags and to support microMIPS.
18828 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18829 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18830 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18831 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18832 (ASM_SPEC): Support mmicromips and mno-micromips.
18833 (M16STORE_REG_P): New macro.
18834 (MIPS_CALL): Support TARGET_MICROMIPS.
18835 (MICROMIPS_J): New macro.
18836 (mips_base_mips16): Rename this...
18837 (mips_base_compression_flags): ...to this.
18838 (UMIPS_12BIT_OFFSET_P): New macro.
18839 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18840 (MULTILIB_DIRNAMES): Likewise.
18842 2013-02-24 Jakub Jelinek <jakub@redhat.com>
18845 * target-globals.c (save_target_globals): For init_reg_sets and
18846 target_reinit remporarily set this_fn_optabs to this_target_optabs.
18848 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
18850 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
18851 * config/aarch64/t-aarch64
18852 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
18854 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
18856 PR inline-asm/56148
18857 * lra-constraints.c (process_alt_operands): Reload operand
18858 conflicting with earlier clobber only if no more other conflicting
18861 2013-02-22 Jakub Jelinek <jakub@redhat.com>
18864 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
18865 if not linking a shared library.
18867 2013-02-22 Seth LaForge <sethml@google.com>
18869 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
18871 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
18873 * config/arm/arm.md (split for extendsidi): Update condition.
18874 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
18875 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
18876 (qhs_zextenddi_cstr): Likewise.
18878 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18880 PR middle-end/56420
18881 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
18882 avoid signed wrapping.
18883 (expand_mult): Handle properly multiplication by
18884 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
18885 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
18886 in the compiler if coeff is HOST_WIDE_INT_MIN.
18887 (expand_divmod): Don't make ext_op1 static, change it's type to
18888 uhwi. Avoid undefined behavior in -INTVAL (op1).
18890 PR rtl-optimization/50339
18891 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
18893 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
18894 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
18895 into splitting_ashiftrt field.
18896 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
18898 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
18901 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
18903 PR middle-end/56108
18904 * trans-mem.c (execute_tm_mark): Do not expand transactions that
18905 are sure to go irrevocable.
18907 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
18909 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
18910 scalars are valid operands.
18912 2013-02-21 Martin Jambor <mjambor@suse.cz>
18914 PR tree-optimization/56310
18915 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
18916 only matching indices and non-negative final offsets.
18917 (intersect_aggregates_with_edge): Pass src_idx to
18918 agg_replacements_to_vector. Pass src_idx insstead of index to
18919 intersect_with_agg_replacements.
18921 2013-02-21 Martin Jambor <mjambor@suse.cz>
18923 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
18924 instead of hard-wired defaults.
18926 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
18928 * doc/invoke.texi (MIPS Options): Update documentation of the
18929 floating-point multiply-accumulate instruction restrictions.
18931 2013-02-21 Kostya Serebryany <kcc@google.com>
18933 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
18934 asan_shadow_offset on x86_64 linux.
18936 2013-02-21 Richard Biener <rguenther@suse.de>
18938 PR tree-optimization/56415
18940 2013-02-11 Richard Biener <rguenther@suse.de>
18942 PR tree-optimization/56273
18943 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18946 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18949 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
18952 PR inline-asm/56405
18953 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
18954 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
18956 2013-02-20 Jan Hubicka <jh@suse.cz>
18958 PR tree-optimization/56265
18959 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
18960 when target is referenced for first time.
18962 2013-02-20 Richard Biener <rguenther@suse.de>
18964 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
18965 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
18966 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
18967 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
18968 not return anything.
18969 (rename_ssa_copies): Do not remove unused locals.
18970 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
18971 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
18972 * passes.c (execute_function_todo): Do not schedule unused locals
18973 removal if cleanup_tree_cfg did something.
18974 * tree-ssa-live.c (remove_unused_locals): Dump statistics
18975 about the number of removed locals.
18977 2013-02-20 Richard Biener <rguenther@suse.de>
18979 PR tree-optimization/56398
18980 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
18982 2013-02-20 Martin Jambor <mjambor@suse.cz>
18984 PR tree-optimization/55334
18985 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
18986 restricted pointers to arrays.
18988 2013-02-20 Richard Biener <rguenther@suse.de>
18989 Jakub Jelinek <jakub@redhat.com>
18991 PR tree-optimization/56396
18992 * tree-ssa-ccp.c (n_const_val): New static variable.
18993 (get_value): Return NULL for SSA names we don't have a lattice
18995 (ccp_initialize): Initialize n_const_val.
18996 * tree-ssa-copy.c (n_copy_of): New static variable.
18997 (init_copy_prop): Initialize n_copy_of.
18998 (get_value): Return NULL_TREE for SSA names we don't have a
19001 2013-02-20 Martin Jambor <mjambor@suse.cz>
19003 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
19005 2013-02-20 Richard Biener <rguenther@suse.de>
19007 * genpreds.c (write_lookup_constraint): Do not compare first
19008 letter of the constraint again.
19010 2013-02-20 Richard Biener <rguenther@suse.de>
19012 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
19014 (get_use_iv_cost): Terminate hashtable walk when coming across
19017 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
19019 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
19020 reassociation for avx2 targets.
19022 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19024 * config/microblaze/microblaze.c: microblaze_has_clz = 0
19025 Add version check for v8.10.a to enable microblaze_has_clz
19026 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
19027 version and TARGET_PATTERN_COMPARE check
19028 * config/microblaze/microblaze.md: New clzsi2 instruction
19030 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19032 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
19033 function before branching.
19035 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19037 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
19039 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
19041 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19043 PR middle-end/55889
19044 * sel-sched.c: Include ira.h.
19045 (implicit_clobber_conflict_p): New function.
19046 (moveup_expr): Use it.
19047 * Makefile.in (sel-sched.o): Depend on ira.h.
19049 2013-02-19 Richard Biener <rguenther@suse.de>
19051 PR tree-optimization/56384
19052 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
19053 (vn_hash_type): Split out from ...
19054 (vn_hash_constant_with_type): ... here.
19055 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
19056 (vn_phi_eq): Compare types from vn_phi_s structure.
19057 (vn_phi_lookup): Populate vn_phi_s type.
19058 (vn_phi_insert): Likewise.
19060 2013-02-19 Jakub Jelinek <jakub@redhat.com>
19062 PR tree-optimization/56350
19063 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
19064 if haven't found reduction or nested cycle operand, rather than
19065 asserting we must find it.
19067 PR tree-optimization/56381
19068 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
19071 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
19072 Jakub Jelinek <jakub@redhat.com>
19075 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
19076 (swap_optab_enable): Same.
19077 (init_all_optabs): Use argument instead of global.
19078 * tree.h (struct tree_optimization_option): New field target_optabs.
19079 * expr.h (init_all_optabs): Add argument to prototype.
19080 (TREE_OPTIMIZATION_OPTABS): New.
19081 (save_optabs_if_changed): Protoize.
19082 * optabs.h: Declare this_fn_optabs.
19083 * optabs.c (save_optabs_if_changed): New.
19084 Declare this_fn_optabs.
19085 (init_optabs): Add argument to init_all_optabs() call.
19086 * function.c (invoke_set_current_function_hook): Handle per
19088 * function.h (struct function): New field optabs.
19089 * config/mips/mips.c (mips_set_mips16_mode): Handle when
19090 optimization_current_node has changed.
19091 * target-globals.h (save_target_globals_default_opts): Protoize.
19092 * target-globals.c (save_target_globals_default_opts): New.
19094 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19097 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
19098 registers %fr12 and %fr12R as call used.
19101 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
19102 and HImode, require all displacements to be an integer multiple of
19104 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
19105 only allow QImode and HImode when reload is in progress and strict is
19106 true. Likewise for symbolic addresses. Use base14_operand to check
19107 displacements in REG+BASE addresses.
19109 2013-02-18 Richard Biener <rguenther@suse.de>
19111 PR tree-optimization/56366
19112 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19113 handle sign-conversion of outer-loop initial induction value.
19115 2013-02-18 Richard Biener <rguenther@suse.de>
19117 PR middle-end/56349
19118 * cfghooks.c (merge_blocks): If we merge a latch into another
19119 block adjust references to it.
19120 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
19121 (verify_loop_structure): Verify that a recorded latch is in fact
19124 2013-02-18 Richard Biener <rguenther@suse.de>
19126 PR tree-optimization/56321
19127 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
19128 order SSA name release and virtual operand unlinking.
19130 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19132 * config/microblaze/microblaze.md (save_stack_block): Define.
19133 (restore_stack_block): Likewise.
19135 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19137 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
19138 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
19139 * config/microblaze/microblaze.c (microblaze_option_override):
19140 Bail out early for PIC modes when target does not support PIC.
19142 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19144 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
19145 Replace with a microblaze version.
19146 (microblaze_trampoline_init): Adapt for microblaze.
19147 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
19150 2013-02-16 Jakub Jelinek <jakub@redhat.com>
19151 Dodji Seketeli <dodji@redhat.com>
19154 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
19155 (instrument_mem_region_access): Do not forget to always put
19156 instrumentation of the of 'base' and 'base + len' in a "if (len !=
19157 0) statement, even for cases where either 'base' or 'base + len'
19158 are not instrumented -- because they have been previously
19159 instrumented. Simplify the logic by putting all the statements
19160 instrument 'base + len' inside a sequence, and then insert that
19161 sequence right before the current insertion point. Then, to
19162 instrument 'base + len', just get an iterator on that statement.
19163 And do not forget to update the pointer to iterator the function
19164 received as argument.
19166 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
19168 PR rtl-optimization/56348
19169 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
19171 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
19173 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
19174 (clean_graph_dump_file): Pass base to start_graph_dump.
19176 2013-02-14 Richard Henderson <rth@redhat.com>
19179 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
19181 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
19183 * collect2-aix.h: Define F_LOADONLY.
19185 2013-02-14 Richard Biener <rguenther@suse.de>
19188 * varasm.c (output_constant_def_1): Get the decl representing
19189 the constant as argument.
19190 (output_constant_def): Wrap output_constant_def_1.
19191 (make_decl_rtl): Use output_constant_def_1 with the decl
19192 representing the constant.
19193 (build_constant_desc): Optionally re-use a decl already
19194 representing the constant.
19195 (tree_output_constant_def): Adjust.
19197 2013-02-14 Dodji Seketeli <dodji@redhat.com>
19200 * asan.c (instrument_builtin_call): Really put the length of the
19201 second source argument into src1_len.
19203 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19205 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
19206 argument. If it is false, don't create edge from then_bb to
19208 (insert_if_then_before_iter): Pass true to it.
19209 (build_check_stmt): Pass false to it.
19210 (transform_statements): Flush hash table only on extended basic
19211 block boundaries, rather than at the beginning of every bb.
19212 Don't flush hash table on nonfreeing_call_p calls.
19213 * tree-flow.h (nonfreeing_call_p): New prototype.
19214 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
19216 2013-02-13 David S. Miller <davem@davemloft.net>
19218 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
19220 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
19223 * ira.c (max_regno_before_ira): Move from ...
19225 (fix_reg_equiv_init): Use max_regno_before_ira instead of
19228 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19230 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
19232 2013-02-13 Richard Biener <rguenther@suse.de>
19235 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
19236 globals in MEM_REFs.
19238 2013-02-13 Richard Biener <rguenther@suse.de>
19240 * loop-init.c (loop_optimizer_init): Clear loop state when
19241 re-initializing preserved loops.
19242 * loop-unswitch.c (unswitch_single_loop): Return whether
19243 we unswitched the loop. Do not verify loop state here.
19244 (unswitch_loops): When we unswitched a loop discover new loops.
19246 2013-02-13 Kostya Serebryany <kcc@google.com>
19248 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
19250 * sanitizer.def: Rename __asan_init to __asan_init_v1.
19252 2013-02-12 Dodji Seketeli <dodji@redhat.com>
19254 Avoid instrumenting duplicated memory access in the same basic block
19255 * Makefile.in (asan.o): Add new dependency on hash-table.h
19256 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
19257 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
19258 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
19259 (free_mem_ref_resources, has_mem_ref_been_instrumented)
19260 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
19261 (get_mem_ref_of_assignment): New functions.
19262 (get_mem_refs_of_builtin_call): Extract from
19263 instrument_builtin_call and tweak a little bit to make it fit with
19265 (instrument_builtin_call): Use the new
19266 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
19267 of is_gimple_builtin_call.
19268 (instrument_derefs, instrument_mem_region_access): Insert the
19269 instrumented memory reference into the hash table.
19270 (maybe_instrument_assignment): Renamed instrument_assignment into
19271 this, and change it to advance the iterator when instrumentation
19272 actually happened and return true in that case. This makes it
19273 homogeneous with maybe_instrument_assignment, and thus give a
19274 chance to callers to be more 'regular'.
19275 (transform_statements): Clear the memory reference hash table
19276 whenever we enter a new BB, when we cross a function call, or when
19277 we are done transforming statements. Use
19278 maybe_instrument_assignment instead of instrumentation. No more
19279 need to special case maybe_instrument_assignment and advance the
19280 iterator after calling it; it's now handled just like
19281 maybe_instrument_call. Update comment.
19283 2013-02-13 Richard Biener <rguenther@suse.de>
19285 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
19286 Fix loop discovery code.
19288 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
19290 PR inline-asm/56148
19291 * lra-constraints.c (process_alt_operands): Match early clobber
19292 operand with itself. Check conflicts with earlyclobber only if
19293 the operand is not reloaded. Prefer to reload conflicting operand
19294 if earlyclobber and matching operands are the same.
19296 2013-02-12 Richard Biener <rguenther@suse.de>
19299 * lto-streamer-out.c (write_symbol): Do not output symbols
19300 for hard register variables.
19302 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
19305 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
19306 (umulsidi3_insn, mulsidi3_insn): New insns.
19308 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
19310 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
19311 (struct tune_params): Add vec_costs field.
19312 * config/arm/arm.c (arm_builtin_vectorization_cost)
19313 (arm_add_stmt_cost): New functions.
19314 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
19315 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
19316 (arm_default_vec_cost): New struct of type cpu_vec_costs.
19317 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
19318 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
19319 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
19320 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
19322 2013-02-12 Richard Biener <rguenther@suse.de>
19325 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
19326 decls again if possible.
19328 2013-02-12 Richard Biener <rguenther@suse.de>
19330 PR middle-end/56288
19331 * tree-ssa.c (verify_ssa_name): Fix check, move
19332 SSA_NAME_IN_FREE_LIST check up.
19334 2013-02-12 Jakub Jelinek <jakub@redhat.com>
19335 Steven Bosscher <steven@gcc.gnu.org>
19337 PR rtl-optimization/56151
19338 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
19339 equal to op0 or op1, and last_insn pattern is CODE operation
19340 with MEM dest and one of the operands matches that MEM.
19342 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
19344 * doc/extend.texi: Document Function Multiversioning and "default"
19345 parameter string to target attribute.
19346 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
19347 target attribute parameter is "default".
19348 (ix86_compare_version_priority): Remove checks for target attribute.
19349 (ix86_mangle_function_version_assembler_name): Change error to sorry.
19350 Remove check for target attribute equal to NULL. Add assert.
19351 (ix86_generate_version_dispatcher_body): Change error to sorry.
19353 2013-02-11 Iain Sandoe <iain@codesourcery.com>
19354 Jack Howarth <howarth@bromo.med.uc.edu>
19355 Patrick Marlier <patrick.marlier@gmail.com>
19358 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
19359 define ENDFILE_SPEC as TM_DESTRUCTOR.
19360 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
19362 2013-02-11 Alexander Potapenko <glider@google.com>
19363 Jack Howarth <howarth@bromo.med.uc.edu>
19364 Jakub Jelinek <jakub@redhat.com>
19367 * config/darwin.c (cdtor_record): Rename ctor_record.
19368 (sort_cdtor_records): Rename sort_ctor_records.
19369 (finalize_dtors): New routine to sort destructors by
19370 priority before use in assemble_integer.
19371 (machopic_asm_out_destructor): Use finalize_dtors if needed.
19373 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
19375 PR rtl-optimization/56275
19376 * simplify-rtx.c (avoid_constant_pool_reference): Check that
19377 offset is non-negative and less than cmode size before
19378 calling simplify_subreg.
19380 2013-02-11 Richard Biener <rguenther@suse.de>
19382 PR tree-optimization/56264
19383 * cfgloop.h (fix_loop_structure): Adjust prototype.
19384 * loop-init.c (fix_loop_structure): Return the number of
19385 newly discovered loops.
19386 * tree-cfgcleanup.c (repair_loop_structures): When new loops
19387 are discovered, do a full loop-closed SSA rewrite.
19389 2013-02-11 Richard Biener <rguenther@suse.de>
19391 PR tree-optimization/56273
19392 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19394 (check_array_ref): Fix missing newline in dumps.
19395 (search_for_addr_array): Likewise.
19397 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
19399 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
19401 2013-02-09 Jakub Jelinek <jakub@redhat.com>
19404 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
19406 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
19408 PR rtl-optimization/56246
19409 * lra-constraints.c (simplify_operand_subreg): Try to reuse
19411 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
19412 constraints are satisfied.
19414 2013-02-08 Jeff Law <law@redhat.com>
19417 * emit-rtl.c (reg_is_parm_p): New function.
19418 * regs.h (reg_is_parm_p): New prototype.
19419 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
19420 callee-clobbered registers.
19422 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
19425 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
19426 If there is no implicit builtin declaration, just return NULL.
19428 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19430 * config/i386/sse.md (FMAMODEM): New mode iterator.
19431 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
19432 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
19434 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19436 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
19437 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
19438 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
19440 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19442 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
19443 (microblaze*-*-elf): Likewise.
19444 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
19446 * config/microblaze/microblaze-c.c: Add builtin defines for
19447 _LITTLE_ENDIAN and _BIG_ENDIAN.
19448 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
19449 add to TARGET_DEFAULT flags.
19450 Expand ASM_SPEC and LINK_SPEC.
19451 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
19452 * config/microblaze/microblaze.md: Update extendsidi2 and
19453 movdi_internal instructions to use low-order / high-order reg
19455 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
19456 options and inversemask / mask of LITTLE_ENDIAN.
19457 * config/microblaze/t-microblaze: Expand multilib options to
19458 include mlittle-endian (le) and update exceptions patterns.
19460 2013-02-08 Jakub Jelinek <jakub@redhat.com>
19462 PR rtl-optimization/56195
19463 * lra-constraints.c (get_reload_reg): Don't reuse regs
19464 if they have smaller mode than requested, if they have
19465 wider mode than requested, try to return a SUBREG.
19467 PR tree-optimization/56250
19468 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
19469 if type is unsigned and code isn't MULT_EXPR.
19471 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19473 PR tree-optimization/56064
19474 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
19475 bits according to mode.
19476 * fixed-value.h (fixed_from_double_int)
19477 (const_fixed_from_double_int): Adjust comments.
19479 2013-02-08 Richard Biener <rguenther@suse.de>
19482 * lto-streamer.h (struct data_in): Remove current_file, current_line
19483 and current_col members.
19484 * lto-streamer-out.c (lto_output_location): Stream changed bits
19485 en-block for efficiency.
19486 * lto-streamer-in.c (clear_line_info): Remove.
19487 (lto_input_location): Cache current file, line and column
19488 globally via local statics. Read changed bits en-block.
19489 (input_function): Do not call clear_line_info.
19490 (lto_read_body): Likewise.
19491 (lto_input_toplevel_asms): Likewise.
19493 2013-02-08 Michael Matz <matz@suse.de>
19495 PR tree-optimization/52448
19496 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19497 (nt_call_phase): New static.
19498 (add_or_mark_expr): Only mark accesses with newer phase than any
19500 (nonfreeing_call_p): New.
19501 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19502 (nt_fini_block): Keep blocks marked as visited.
19503 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19505 2013-02-08 Richard Biener <rguenther@suse.de>
19507 * ira.c (ira): Free broken dominator information.
19509 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19511 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19513 2013-02-08 Marek Polacek <polacek@redhat.com>
19515 * cfgloop.c (verify_loop_structure): Add more checking of headers.
19517 2013-02-08 Richard Biener <rguenther@suse.de>
19519 PR middle-end/56181
19520 * cfgloop.h (flow_loops_find): Adjust.
19521 (bb_loop_header_p): Declare.
19522 * cfgloop.c (bb_loop_header_p): New function split out from ...
19523 (flow_loops_find): ... here. Adjust function signature,
19524 support incremental loop structure update.
19525 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19526 * cfgloopmanip.c (fix_loop_structure): Move ...
19527 * loop-init.c (fix_loop_structure): ... here.
19528 (apply_loop_flags): Split out from ...
19529 (loop_optimizer_init): ... here.
19530 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19531 in incremental mode, only remove dead loops here.
19533 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19536 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19537 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19538 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19539 (*round<mode>3.libgcc): New insns for fixed-modes.
19540 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19541 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19542 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19543 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19544 implementations. Define to __builtin_avr_absFX,
19545 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19546 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19547 __builtin_avr_countlsFX, respectively.
19548 * config/avr/avr-c.c (target.h): Include it.
19549 (enum avr_builtin_id): New enum.
19550 (avr_resolve_overloaded_builtin): New static function.
19551 (avr_register_target_pragmas): Use it to set
19552 targetm.resolve_overloaded_builtin.
19553 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19554 tree nodes used by DEF_BUILTIN.
19555 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19556 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19557 <AVR_BUILTIN_xxBITS>: Same.
19559 2013-02-08 Richard Biener <rguenther@suse.de>
19561 * cfgloop.c (verify_loop_structure): Properly handle
19562 a loop exiting to another loop header.
19563 * ira-int.h (ira_loops): Remove.
19564 * ira.c (ira_loops): Remove.
19565 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19566 (do_reload): Use loop_optimizer_finalize.
19567 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19568 number_of_loops to access the loop tree.
19569 (more_one_region_p): Likewise.
19570 (finish_loop_tree_nodes): Likewise.
19571 (rebuild_regno_allocno_maps): Likewise.
19572 (mark_loops_for_removal): Likewise.
19573 (mark_all_loops_for_removal): Likewise.
19574 (remove_unnecessary_regions): Likewise.
19575 (ira_build): Likewise.
19576 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19578 2013-02-08 Richard Biener <rguenther@suse.de>
19580 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
19581 * ipa-pure-const.c (analyze_function): Avoid calling
19582 mark_irreducible_loops twice.
19583 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
19585 2013-02-07 David S. Miller <davem@davemloft.net>
19587 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
19589 * var-tracking.c (vt_add_function_parameter): Test the presence of
19590 HAVE_window_save properly and do not remap argument registers when
19591 we have a leaf function.
19593 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
19596 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
19598 * config/i386/i386.c (ix86_print_operand): Ditto.
19600 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19602 * lra-constraints.c (process_alt_operands): Fix recently added comment.
19604 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19606 PR rtl-optimization/56225
19607 * lra-constraints.c (process_alt_operands): Check that reload hard
19608 reg can hold value for strict_low_part.
19610 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19613 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
19614 dwarf2out_end_function.
19615 (in_first_function_p, maybe_at_text_label_p,
19616 first_loclabel_num_not_at_text_label): New variables.
19617 (dwarf2out_var_location): In the first function find out
19618 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
19619 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
19622 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
19624 PR rtl-optimization/56178
19625 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
19626 SUBREG of a register. Tidy up related block of code.
19627 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
19628 note if the source is a register or a SUBREG of a register.
19630 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19633 * config/rs6000/rs6000.md (ptrm): New mode attr.
19634 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19635 call_value_indirect_aix<pttrsize>,
19636 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19639 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19641 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19642 if -bnortl. Convert to strcmp and strncmp.
19644 2013-02-07 Alan Modra <amodra@gmail.com>
19647 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19648 addresses won't wrap when offsetting.
19649 (rs6000_secondary_reload): Provide secondary reloads needed for
19650 wrapping LO_SUM addresses.
19652 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19654 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19655 MACH, just __MACH__.
19657 2013-02-06 Richard Biener <rguenther@suse.de>
19659 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19660 instead of calling fix_loop_structure.
19662 2013-02-06 Jakub Jelinek <jakub@redhat.com>
19664 PR middle-end/56217
19665 * omp-low.c (use_pointer_for_field): Return false if
19666 lower_send_shared_vars doesn't generate any copy-out code.
19668 2013-02-06 Tom de Vries <tom@codesourcery.com>
19670 PR rtl-optimization/56131
19671 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19672 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19673 of the label is NULL. Add comment.
19675 2013-02-05 Jakub Jelinek <jakub@redhat.com>
19677 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19680 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19681 (STATIC_LIBTSAN_LIBS): Likewise.
19682 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19683 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19684 is defined, don't add anything else beyond that.
19685 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19686 (LINK_COMMAND_SPEC): Use them.
19688 PR tree-optimization/56205
19689 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19690 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19691 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19693 2013-02-05 Richard Biener <rguenther@suse.de>
19695 PR tree-optimization/53342
19696 PR tree-optimization/53185
19697 * tree-vectorizer.h (vect_check_strided_load): Remove.
19698 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19699 not disallow peeling for vectorized strided loads.
19700 (vect_check_strided_load): Make static and simplify.
19701 (vect_analyze_data_refs): Adjust.
19702 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19703 correctly when vectorizing strided loads.
19705 2013-02-05 Richard Biener <rguenther@suse.de>
19707 * doc/install.texi: Refer to ISL, not PPL.
19709 2013-02-05 Jan Hubicka <jh@suse.cz>
19711 PR tree-optimization/55789
19712 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19714 2013-02-05 Jan Hubicka <jh@suse.cz>
19716 PR tree-optimization/55789
19717 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19718 the dead call anyway.
19720 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19723 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19725 2013-02-04 Alexander Potapenko <glider@google.com>
19726 Jack Howarth <howarth@bromo.med.uc.edu>
19727 Jakub Jelinek <jakub@redhat.com>
19730 * config/darwin.c (sort_ctor_records): Stabilized qsort
19731 on constructor priority by using original position.
19732 (finalize_ctors): New routine to sort constructors by
19733 priority before use in assemble_integer.
19734 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19736 2013-02-04 Jakub Jelinek <jakub@redhat.com>
19739 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19740 about visibility on artificial decls.
19741 * config/sol2.c (solaris_assemble_visibility): Likewise.
19743 2013-02-04 Kai Tietz <ktietz@redhat.com>
19746 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19747 argument and improve checking of return-argument types for 16-byte
19749 (ix86_function_value_1): Add additional valtype argument on call
19750 of function_value_64.
19751 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19752 handling infunction_value_64 function.
19754 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19756 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
19758 2013-02-04 Richard Biener <rguenther@suse.de>
19760 PR tree-optimization/56188
19761 * tree-ssa-structalias.c (label_visit): Consider case with
19762 initially non-empty points-to set.
19763 (perform_var_substitution): Dump node mapping and clean up.
19765 2013-02-04 Richard Guenther <rguenther@suse.de>
19768 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19769 node prevail as last resort.
19770 (lto_symtab_merge_decls): Remove guard on LTRANS here.
19771 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
19773 2013-02-04 Richard Biener <rguenther@suse.de>
19775 PR tree-optimization/56113
19776 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19778 (equiv_class_lookup_or_add): ... this.
19779 (label_visit): Adjust and fix error in previous patch.
19780 (perform_var_substitution): Adjust.
19782 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19784 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19785 * config/sh/sh4-300.md: Likewise.
19786 * config/sh/sh4a.md: Likewise.
19787 * config/sh/constraints.md: Likewise.
19788 * config/sh/sh.md: Likewise.
19789 * config/sh/netbsd-elf.h: Likewise.
19790 * config/sh/predicates.md: Likewise.
19791 * config/sh/sh-protos.h: Likewise.
19792 * config/sh/ushmedia.h: Likewise.
19793 * config/sh/linux.h: Likewise.
19794 * config/sh/sh.c: Likewise.
19795 * config/sh/superh.h: Likewise.
19796 * config/sh/elf.h: Likewise.
19797 * config/sh/sh4.md: Likewise.
19798 * config/sh/sh.h: Likewise.
19800 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19802 * config/pa/constraints.md: Adjust unused letters. Change "T"
19803 constraint to match_test floating_point_store_memory_operand().
19804 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19805 (base14_operand): New.
19806 (floating_point_store_memory_operand): New.
19807 (integer_store_memory_operand): Revise to use base14_operand and
19808 reg_plus_base_memory_operand.
19809 (move_dest_operand): Allow symbolic_memory_operands.
19810 (symbolic_memory_operand): Check for LO_SOM.
19811 (symbolic_operand): Change default case to break.
19812 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19813 CONST_DOUBLE values to be reloaded by putting them into memory when
19814 the destination is a floating point register.
19815 (movdf): Remove code to handle CONST_DOUBLE.
19817 (reload_indf_r1): New.
19818 (reload_insf_r1): New.
19819 Consistently use "Q" and "T" constraints with integer and floating
19820 point move instructions, respectively.
19821 (movdi): Remove FAIL.
19822 Change predicate for source operand unamed DImode move from
19823 general_operand to move_src_operand.
19824 (umulsidi3): Change predicate for destination operand to
19826 Likewise for similar unamed patterns.
19827 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19828 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19829 (hppa_legitimize_address): Simplify mask calculation.
19830 (pa_emit_move_sequence): Revised handling of secondary reloads from
19831 REG+D addresses for floating point loads and stores. Directly handle
19832 loading CONST0_RTX (mode) to a floating point register.
19833 (pa_secondary_reload): Handle reloading DF and SFmode constant values
19834 to floating point registers. Don't restrict secondary reloads to
19835 floating point registers to integer modes. Revise some comments and
19837 (TARGET_LEGITIMATE_ADDRESS_P): Define.
19838 (pa_legitimate_address_p): New.
19839 (pa_legitimize_reload_address): New.
19840 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
19841 (STRICT_REG_OK_FOR_BASE_P): New.
19842 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
19843 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
19845 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
19846 Andrew Dixie <andrewd@gentrack.com>
19848 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
19851 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
19853 * expmed.c (extract_bit_field_1): Pass the full width of the
19854 structure to get_best_reg_extraction_insn.
19856 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19859 * configure.ac (use_cxa_atexit): Add AIX.
19860 * configure: Regenerate.
19862 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
19864 2013-02-01 Jakub Jelinek <jakub@redhat.com>
19867 * final.c (need_profile_function): New variable.
19868 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
19869 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
19870 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
19871 notes, targetm.asm_out.function_prologue doesn't emit anything,
19872 HAVE_prologue and profiler should be emitted before prologue,
19873 set need_profile_function instead of emitting it.
19874 (final_scan_insn): If need_profile_function, emit
19875 profile_function on the first NOTE_INSN_BASIC_BLOCK or
19876 NOTE_INSN_FUNCTION_BEG note.
19878 2013-02-01 Richard Henderson <rth@redhat.com>
19880 * config/rs6000/rs6000.md (smulditi3): New.
19883 * config/alpha/alpha.md (umulditi3): New.
19885 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19887 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
19888 (ASM_OUTPUT_ALIGNED_LOCAL): New.
19890 2013-02-01 Richard Biener <rguenther@suse.de>
19892 PR tree-optimization/56113
19893 * tree-ssa-structalias.c (label_visit): Reduce work for
19894 single-predecessor nodes.
19896 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
19898 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
19899 range isn't testing for zero.
19901 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
19903 PR middle-end/56113
19904 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
19906 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
19907 Nick Clifton <nickc@redhat.com>
19909 * config/v850/constraints.md (Q): Define as a memory constraint.
19910 * config/v850/predicates.md (label_ref_operand): New predicate.
19911 (e3v5_shift_operand): New predicate.
19912 (ior_operator): New predicate.
19913 * config/v850/t-v850: Add e3v5 multilib.
19914 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
19915 (v850_gen_movdi): Prototype.
19916 * config/v850/v850.c: Add support for e3v5 architecture.
19917 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
19919 (construct_save_jarl): Add e3v5 long JARL support.
19920 (v850_adjust_insn_length): New function. Adjust length of call
19921 insns when using e3v5 instructions.
19922 (v850_gen_movdi): New function: Generate instructions to move a
19924 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
19925 (CPP_SPEC): Define __v850e3v5__ as appropriate.
19926 (TARGET_USE_FPU): Enable for e3v5.
19927 (CONST_OK_FOR_W): New macro.
19928 (ADJUST_INSN_LENGTH): Define.
19929 * config/v850/v850.md (UNSPEC_LOOP): Define.
19930 (attr cpu): Add v850e3v5.
19931 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
19932 (movdi): New pattern.
19933 (movdi_internal): New pattern.
19934 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
19935 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
19936 (cstoresf4): Likewise.
19937 (cstoredf4): Likewise.
19938 (insv): New pattern.
19939 (rotlso3_a): New pattern.
19940 (rotlsi3_b): New pattern
19941 (rotlsi3_v850e3v5): New pattern.
19942 (doloop_begin): New pattern.
19943 (fix_loop_counter): New pattern.
19944 (doloop_end): New pattern.
19945 (branch_normal): Add e3v5 long branch support.
19946 (branch_invert): Likewise.
19947 (branch_z_normal): Likewise.
19948 (branch_z_invert): Likewise.
19949 (branch_nz_normal): Likewise.
19950 (branch_nz_invert): Likewise.
19951 (call_internal_short): Add e3v5 register-indirect JARL support.
19952 (call_internal_long): Likewise.
19953 (call_value_internal_short): Likewise.
19954 (call_value_internal_long): Likewise.
19955 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
19956 (mloop): New option.
19957 * config.gcc: Add support for configuring v840e3v5 target.
19958 * doc/invoke.texi: Document new v850 specific command line options.
19960 2013-01-31 Paul Koning <ni1d@arrl.net>
19964 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
19965 children if parent is a class.
19966 (prune_unused_types_prune): Don't add DW_AT_declaration.
19968 2013-01-31 Richard Biener <rguenther@suse.de>
19970 PR tree-optimization/56157
19971 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
19972 match up operand with SLP child.
19974 2013-01-31 Jason Merrill <jason@redhat.com>
19977 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
19978 parameters the first time.
19979 (gen_scheduled_generic_parms_dies): Check completeness here.
19981 2013-01-31 Richard Biener <rguenther@suse.de>
19983 PR middle-end/53073
19984 * common.opt (faggressive-loop-optimizations): New flag,
19985 enabled by default.
19986 * doc/invoke.texi (faggressive-loop-optimizations): Document.
19987 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
19988 infer_loop_bounds_from_undefined by it.
19990 2013-01-31 Richard Biener <rguenther@suse.de>
19992 PR tree-optimization/56150
19993 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
19994 visit virtual operands.
19995 (find_uses_to_rename_bb): Likewise.
19997 2013-01-31 Richard Biener <rguenther@suse.de>
19999 PR tree-optimization/56150
20000 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
20001 mixed store non-store stmts.
20003 2013-01-30 Jakub Jelinek <jakub@redhat.com>
20006 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
20007 LIBASAN_EARLY_SPEC is defined.
20008 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
20009 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
20011 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
20014 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
20015 invalid args instead of ICEing on it.
20016 (ix86_valid_target_attribute_tree): Return error_mark_node if
20017 ix86_valid_target_attribute_inner_p failed.
20018 (ix86_valid_target_attribute_p): Return false only if
20019 ix86_valid_target_attribute_tree returned error_mark_node. Allow
20020 target("default") attribute.
20021 (sorted_attr_string): Change argument from const char * to tree,
20022 merge in all target attribute arguments rather than just one.
20023 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
20024 instead of free. Avoid using strcat.
20025 (ix86_mangle_function_version_assembler_name): Mangle
20026 target("default") as if no target attribute is present. Adjust
20027 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
20028 instead of xmalloc and XDELETEVEC instead of free.
20029 (ix86_function_versions): Don't return true if one of the decls
20030 doesn't have target attribute. If they don't and one of the decls
20031 is DECL_FUNCTION_VERSIONED, report an error. Adjust
20032 sorted_attr_string caller. Use XDELETEVEC instead of free.
20033 (ix86_supports_function_versions): Remove.
20034 (make_name): Fix up formatting.
20035 (make_dispatcher_decl): Remove resolver_name and its initialization.
20036 Avoid leaking memory.
20037 (is_function_default_version): Return true if there is
20038 target("default") attribute rather than no target attribute at all.
20039 (make_resolver_func): Avoid leaking memory.
20040 (ix86_generate_version_dispatcher_body): Likewise.
20041 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
20042 * target.def (supports_function_versions): Remove.
20043 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
20044 * doc/tm.texi: Regenerated.
20046 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
20048 PR rtl-optimization/56144
20049 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
20050 for values with side effects.
20052 2013-01-30 Richard Biener <rguenther@suse.de>
20054 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
20055 (sparseset_pop): Likewise.
20056 * cfganal.c (compute_idf): Likewise. Increase work-stack size
20057 to be able to use quick_push in the worker loop.
20059 2013-01-30 Marek Polacek <polacek@redhat.com>
20061 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
20063 2013-01-30 Richard Biener <rguenther@suse.de>
20066 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
20068 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
20070 PR tree-optimization/56064
20071 * fixed-value.c (fixed_from_double_int): New function.
20072 * fixed-value.h (fixed_from_double_int): New prototype.
20073 (const_fixed_from_double_int): New static inline function.
20074 * fold-const.c (native_interpret_fixed): New static function.
20075 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
20076 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
20077 (native_encode_fixed): New static function.
20078 (native_encode_expr) <FIXED_CST>: Use it.
20079 (native_interpret_int): Move double_int worker code to...
20080 * double-int.c (double_int::from_buffer): ...this new static method.
20081 * double-int.h (double_int::from_buffer): Prototype it.
20083 2013-01-30 Richard Biener <rguenther@suse.de>
20085 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
20086 New pointer-map and obstack.
20087 (init_alias_vars): Allocate pointer-map and obstack.
20088 (delete_points_to_sets): Free them.
20089 (find_what_var_points_to): Cache result.
20090 (find_what_p_points_to): Adjust for changed interface of
20091 find_what_var_points_to.
20092 (compute_points_to_sets): Likewise.
20093 (ipa_pta_execute): Likewise.
20095 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20097 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
20098 * configure: Regenerate.
20099 * config.in: Regenerate.
20100 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
20101 #nobits/#progbits if supported.
20103 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
20106 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
20107 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
20108 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
20110 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20112 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
20113 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
20115 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20117 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
20118 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
20120 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20122 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
20124 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
20125 * config/arm/cortex-a7.md: New bypasses using
20126 arm_mac_accumulator_is_result.
20128 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20130 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
20131 (cortex_a7_neon_mla): Likewise.
20132 (cortex_a7_fpfmad): New reservation.
20133 (cortex_a7_fpmacs): Use ffmas and update required units.
20134 (cortex_a7_fpmuld): Update required units and latency.
20135 (cortex_a7_fpmacd): Likewise.
20136 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
20137 (cortex_a7_neon). Likewise.
20138 (bypass) Update participating units.
20140 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20142 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
20143 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
20145 * config/arm/vfp11.md (vfp_farith): Use ffmas.
20146 (vfp_fmul): Use ffmad.
20147 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
20148 (cortex_r4_fmacd): Use ffmad.
20149 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
20150 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
20151 (cortex_a9_fmacd): Use ffmad.
20152 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
20153 (cortex_a8_vfp_macd): Use ffmad.
20154 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
20155 (cortex_a5_fpmacd): Use ffmad.
20156 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
20157 (cortex_a15_vfp_macd): Use ffmad.
20158 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
20160 2013-01-29 Jason Merrill <jason@redhat.com>
20163 * varasm.c (default_assemble_visibility): Don't warn about
20164 visibility on artificial decls.
20166 2013-01-29 Richard Biener <rguenther@suse.de>
20168 PR tree-optimization/56113
20169 * tree-ssa-structalias.c (equiv_class_lookup): Also return
20171 (label_visit): Free duplicate bitmaps and record the leader instead.
20172 (perform_var_substitution): Adjust.
20174 2013-01-29 Richard Biener <rguenther@suse.de>
20176 PR tree-optimization/55270
20177 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
20178 the CFG, schedule loops for fixup.
20180 2013-01-29 Nick Clifton <nickc@redhat.com>
20182 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
20185 2013-01-28 Leif Ekblad <leif@rdos.net>
20187 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
20188 * config/i386/i386.h (TARGET_RDOS): New macro.
20189 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
20190 * config/i386/i386.c (ix86_option_override_internal): For 64bit
20191 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
20192 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
20193 DEFAULT_LARGE_SECTION_THRESHOLD.
20194 * config/i386/i386.md (R14_REG, R15_REG): New constants.
20195 * config/i386/rdos.h: New file.
20196 * config/i386/rdos64.h: New file.
20198 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
20201 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
20204 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20206 PR rtl-optimization/56117
20207 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
20208 call cselib_lookup_from_insn on the MEM before calling
20209 add_insn_mem_dependence.
20211 2013-01-28 Richard Biener <rguenther@suse.de>
20213 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
20214 to a stmt that didn't have one.
20215 (copy_phis_for_bb): Likewise for PHI arguments.
20216 (copy_debug_stmt): Likewise for debug stmts.
20218 2013-01-28 Richard Biener <rguenther@suse.de>
20220 PR tree-optimization/56034
20221 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
20222 (partition_builtin_p): Adjust.
20223 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
20224 it is the last partition.
20225 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
20226 up the vertex for the definition.
20227 (classify_partition): Classify whether a partition is a
20228 PKIND_REDUCTION, thus has uses outside of the loop.
20229 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
20230 Merge all PKIND_REDUCTION partitions into the last partition.
20231 (tree_loop_distribution): Seed partitions from reductions as well.
20233 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20235 PR tree-optimization/56125
20236 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
20237 pow(x,c) into sqrt(x) * powi(x, n/2) or
20238 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
20239 optimizing for size.
20240 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
20241 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
20244 PR tree-optimization/56094
20245 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
20246 to UNKNOWN_LOCATION while gimplifying expr.
20248 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
20251 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
20252 operand 0 in movabs insn template for -masm=intel asm alternative.
20253 (*movabs<mode>_2): Ditto for operand 1.
20255 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
20258 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
20261 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20263 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
20266 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20268 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
20269 Undef to avoid warning.
20271 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20273 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
20274 * configure: Regenerate.
20276 2013-01-25 Jakub Jelinek <jakub@redhat.com>
20278 PR tree-optimization/56098
20279 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
20280 for stmts with volatile ops.
20281 (cond_store_replacement): Don't optimize if assign has volatile ops.
20282 (cond_if_else_store_replacement_1): Don't optimize if either
20283 then_assign or else_assign have volatile ops.
20284 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
20287 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20289 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
20291 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20293 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
20294 missing ':' in asm example.
20296 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
20298 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
20299 entries into lane and laneq entries.
20300 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
20301 Remove AdvSIMD scalar modes.
20302 (aarch64_sq<r>dmulh_laneq<mode>): New.
20303 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
20305 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
20306 builtin implementations to relfect changes in RTL in aarch64-simd.md.
20307 * config/aarch64/iterators.md (VCOND): New.
20310 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20313 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
20314 Add NULL LIBNAME argument to existing definitions.
20315 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
20316 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
20317 * config/avr/avr.c (DEF_BUILTIN): Same.
20318 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
20319 (avr_expand_builtin): Expand to a vanilla call if a libgcc
20320 implementation is available (DECL_ASSEMBLER_NAME is set).
20321 (avr_fold_absfx): New static function.
20322 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
20323 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
20324 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
20325 AVR_BUILTIN_ABSLLK.
20326 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
20327 (abshk, absk, abslk, absllk): Provide as static inline functions.
20329 2013-01-25 Marek Polacek <polacek@redhat.com>
20331 PR tree-optimization/56035
20332 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
20334 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
20336 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
20337 (*movtf_internal_rex64): Add (!o,C) alternative
20338 (*movxf_internal_rex64): Ditto.
20339 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
20341 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
20343 * doc/invoke.texi: fix typo.
20344 * doc/objc.texi: fix typo.
20346 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
20348 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
20349 for the first two alternatives.
20351 2013-01-24 Diego Novillo <dnovillo@google.com>
20353 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
20354 (ggc-zone.o): Remove.
20355 * configure.ac: Remove option --with-gc.
20356 * configure: Re-generate.
20357 * doc/install.texi: Remove documentation for --with-gc.
20358 * gengtype.c (write_enum_defn): Remove. Update all users.
20359 (write_Types_process_field): Remove generation of gt_e_* argument.
20360 (output_type_enum): Remove. Update all users.
20361 (write_enum_defn): Remove. Update all users.
20362 (enum alloc_zone): Remove. Update all users.
20363 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
20364 * ggc-common.c (ggc_splay_alloc): Remove first argument.
20365 Update all callers.
20366 (struct ptr_data): Remove field TYPE. Update all users.
20367 (gt_pch_note_object): Remove argument TYPE. Update all users.
20368 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
20370 * ggc-none.c (ggc_alloc_typed_stat): Remove.
20371 (struct alloc_zone): Remove.
20372 (ggc_internal_alloc_zone_stat): Remove.
20373 (ggc_internal_cleared_alloc_zone_stat): Remove.
20374 * ggc-page.c (ggc_alloc_typed_stat): Remove.
20375 (ggc_pch_count_object): Remove last argument. Update all users.
20376 (ggc_pch_alloc_object): Remove last argument. Update all users.
20377 (struct alloc_zone): Remove.
20378 * ggc-zone.c: Remove.
20379 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
20380 (struct alloc_zone): Remove.
20381 (ggc_alloc_typed_stat): Remove.
20382 (ggc_alloc_typed): Remove.
20383 (ggc_splay_alloc): Remove first argument.
20384 (rtl_zone): Remove. Update all users.
20385 (tree_zone): Remove. Update all users.
20386 (tree_id_zone): Remove. Update all users.
20387 (ggc_internal_zone_alloc_stat): Remove. Update all users.
20388 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
20389 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
20390 * tree-ssanames.c: Remove references to zone allocator in comments.
20392 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
20394 * config/avr/avr.c (avr_out_fract): Make register numbers that
20395 might be outside of source operand signed.
20397 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
20399 * config/i386/constraints.md (Yf): New constraint.
20400 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
20401 of f constraint to conditionaly disable x87 register preferences.
20402 (*movdf_internal): Ditto.
20403 (*movsf_internal): Ditto.
20405 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
20407 PR inline-asm/55934
20408 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
20409 that have operands with impossible constraints.
20410 Add a FIXME for a speed-up opportunity.
20411 * lra-constraints.c (process_alt_operands): Verify that a class
20412 selected from constraints on asms is valid for the operand mode.
20413 (curr_insn_transform): Remove incorrect comment.
20415 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
20417 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
20418 TOC operand is a valid symbol ref in the constant pool.
20420 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20422 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
20424 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
20427 * config/avr/stdfix.h: New file.
20428 * t-avr (stdfix-gcc.h): New rule to build it.
20429 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
20431 2013-01-23 Kostya Serebryany <kcc@google.com>
20433 * config/darwin.h: remove dependency on
20434 CoreFoundation (asan on Mac OS).
20436 2013-01-23 Jakub Jelinek <jakub@redhat.com>
20439 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
20440 instead of cmpdi_operand for first comparison operand.
20441 Don't assert that comparison operands aren't both constants.
20443 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20445 * doc/install.texi (Downloading the Source): Update references to
20446 downloading separate components.
20448 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20450 * doc/extend.texi (__int128): Improve grammar.
20452 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
20455 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
20456 alternative to (o,r).
20457 (*movdi_internal_rex64): Remove (!o,n) alternative.
20458 (DImode immediate->memory splitter): Remove.
20459 (DImode immediate->memory peephole2): Remove.
20460 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
20461 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
20462 alternative to (!o,*r).
20463 (*movtf_internal_sse): New pattern.
20464 (*movxf_internal_rex64): New pattern.
20465 (*movxf_internal): Disable for TARGET_64BIT.
20466 (*movdf_internal_rex64): Remove (!o,F) alternative.
20468 2013-01-22 Jakub Jelinek <jakub@redhat.com>
20470 PR middle-end/56074
20471 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
20472 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
20473 * tree-vect-loop-manip.c (find_loop_location): Also ignore
20474 stmt locations where LOCATION_LOCUS of the stmt location is
20475 UNKNOWN_LOCATION or BUILTINS_LOCATION.
20478 * config/i386/i386.md (UNSPEC_STOS): New.
20479 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20480 *strsetqi_1): Add UNSPEC_STOS.
20482 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20485 * doc/invoke.texi: Remove left over -Wsynth example.
20487 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20489 PR tree-optimization/56051
20490 * fold-const.c (fold_binary_loc): Don't fold
20491 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20492 a narrowing conversion, or widening conversion from signed
20495 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20497 PR rtl-optimization/56023
20498 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20499 dependent on debug instruction.
20501 2013-01-21 Martin Jambor <mjambor@suse.cz>
20503 PR middle-end/56022
20504 * function.c (allocate_struct_function): Call
20505 invoke_set_current_function_hook earlier.
20507 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20509 * reload1.c (init_reload): Only initialize reload_obstack
20510 during the first call.
20512 2013-01-21 Marek Polacek <polacek@redhat.com>
20514 * cfgloop.c (verify_loop_structure): Fix up grammar.
20516 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20518 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20519 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20521 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20524 * config/arm/marvell-pj4.md: Update copyright year.
20525 Fix up use of alu to alu_reg and simple_alu_imm.
20527 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20529 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20531 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20534 * lra-constraints.c (curr_insn_transform): Don't reuse original
20535 insn for secondary memory move when memory mode should be different.
20537 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20539 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20540 atomic_storedi_1): New patterns.
20542 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20544 btver2 pipeline descriptions.
20545 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20547 * config/i386/i386.md (btver2_decode): New type attributes.
20548 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
20550 * config/i386/btver2.md: New file describing btver2 pipelines.
20552 2013-01-19 Andrew Pinski <apinski@cavium.com>
20554 PR tree-optimization/52631
20555 * tree-ssa-sccvn (visit_use): Before looking up the original
20556 statement, try looking up the simplified expression.
20558 2013-01-19 Anthony Green <green@moxielogic.com>
20560 * config/moxie/moxie.c (moxie_expand_prologue): Set
20561 current_function_static_stack_size.
20563 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20565 PR tree-optimization/56029
20566 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20567 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20569 2013-01-18 Sharad Singhai <singhai@google.com>
20571 PR tree-optimization/55995
20572 * dumpfile.c (dump_loc): Print location only if available.
20573 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
20575 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
20578 * lra-constraints.c (curr_insn_transform): Reuse original insn for
20579 secondary memory move.
20580 (inherit_reload_reg): Use rclass instead of cl for
20581 check_secondary_memory_needed_p.
20583 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20585 PR middle-end/56015
20586 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
20587 the case where writing real complex part of target modifies op1.
20589 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
20591 * config/aarch64/aarch64-simd.md
20592 (aarch64_vcond_internal<mode>): Handle unordered cases.
20593 * config/aarch64/iterators.md (v_cmp_result): New.
20595 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
20596 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20598 * config/arm/marvell-pj4.md: New file.
20599 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
20600 * config/arm/arm.md (generic_sched): Add marvell_pj4.
20601 (generic_vfp): Likewise.
20602 * config/arm/arm-cores.def: Add marvell-pj4.
20603 * config/arm/arm-tune.md: Regenerate.
20604 * config/arm/arm-tables.opt: Regenerate.
20605 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
20606 * doc/invoke.texi: Document marvell-pj4.
20608 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
20610 * config/aarch64/arm_neon.h: Map scalar types to standard types.
20612 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20617 * var-tracking.c (negative_power_of_two_p): New.
20618 (global_get_addr_cache, local_get_addr_cache): New.
20619 (get_addr_from_global_cache, get_addr_from_local_cache): New.
20620 (vt_canonicalize_addr): Rewrite using the above. Adjust the
20622 (vt_stack_offset_p): Remove.
20623 (vt_canon_true_dep): Always canonicalize loc's address.
20624 (clobber_overlapping_mems): Make sure we have a MEM.
20625 (local_get_addr_clear_given_value): New.
20626 (val_reset): Clear local cached entries.
20627 (compute_bb_dataflow): Create and release the local cache.
20628 Disable duplicate MEMs clobbering.
20629 (emit_notes_in_bb): Clobber MEMs likewise.
20630 (vt_emit_notes): Create and release the local cache.
20631 (vt_initialize, vt_finalize): Create and release the global
20632 cache, respectively.
20633 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
20635 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20637 PR libmudflap/53359
20638 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20639 not found in the symtab.
20641 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20644 PR rtl-optimization/55547
20645 PR rtl-optimization/53827
20648 * alias.c (offset_overlap_p): New, factored out of...
20649 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20650 the conservative special case for symbolic constants. Don't
20651 adjust zero sizes on alignment.
20653 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20655 PR rtl-optimization/52573
20656 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20657 REG_UNUSED for the same register.
20659 2013-01-17 Richard Biener <rguenther@suse.de>
20660 Marek Polacek <polacek@redhat.com>
20662 PR rtl-optimization/55833
20663 * loop-unswitch.c (unswitch_loops): Move loop verification...
20664 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20665 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20666 Set it to true when we're removing a loop from hierarchy tree in
20667 an irreducible region.
20668 (fix_bb_placements): Adjust caller.
20669 (fix_loop_placements): Likewise.
20671 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
20673 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20674 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20675 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20676 Remove ID. Adjust comments.
20677 * config/avr/avr-c.c (avr_builtin_name): Remove.
20678 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20679 * config/avr/avr.c (avr_tolower): New static function.
20680 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20681 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20682 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20685 2013-01-17 Jan Hubicka <jh@suse.cz>
20687 PR tree-optimization/55273
20688 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20690 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
20693 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20694 store through atomic_store<mode>_1.
20695 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20697 2013-01-17 Martin Jambor <mjambor@suse.cz>
20699 PR tree-optimizations/55264
20700 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20701 for virtual methods.
20702 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20703 virtual methods before inlining is over.
20704 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20706 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20709 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20711 PR rtl-optimization/56005
20712 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20713 pending reads for prefetch.
20715 2013-01-16 Ian Bolton <ian.bolton@arm.com>
20717 * config/aarch64/aarch64.md
20718 (*cstoresi_neg_uxtw): New pattern.
20719 (*cmovsi_insn_uxtw): New pattern.
20720 (*<optab>si3_uxtw): New pattern.
20721 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20722 (*<optab>si3_insn_uxtw): New pattern.
20723 (*bswapsi2_uxtw): New pattern.
20725 2013-01-16 Richard Biener <rguenther@suse.de>
20727 * tree-inline.c (tree_function_versioning): Remove set but
20728 never used variable.
20730 2013-01-16 Richard Biener <rguenther@suse.de>
20732 PR tree-optimization/55964
20733 * tree-flow.h (rename_variables_in_loop): Remove.
20734 (rename_variables_in_bb): Likewise.
20735 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20736 (copy_loop_before): Adjust and delete update-ssa status.
20737 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
20738 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
20739 (rename_variables_in_loop): Remove.
20740 (slpeel_update_phis_for_duplicate_loop): Likewise.
20741 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20742 use available cfg machinery instead of duplicating it.
20743 Update PHI nodes and perform poor-mans SSA update here.
20744 (slpeel_tree_peel_loop_to_edge): Adjust.
20746 2013-01-16 Richard Biener <rguenther@suse.de>
20748 PR tree-optimization/54767
20749 PR tree-optimization/53465
20750 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20751 (vrp_visit_phi_node): For PHI arguments coming via backedges
20752 drop all symbolical range information.
20753 (execute_vrp): Compute backedges.
20755 2013-01-16 Richard Biener <rguenther@suse.de>
20757 * doc/install.texi: Update CLooG and ISL requirements to
20760 2013-01-16 Christian Bruel <christian.bruel@st.com>
20763 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20764 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20765 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20767 2013-01-16 DJ Delorie <dj@redhat.com>
20769 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20770 (UNSPECV_SP_SWITCH_E): New.
20771 (sp_switch_1): Change to an unspec.
20772 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20775 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
20777 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20778 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20779 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20780 (expand_mem_thread_fence): Ditto.
20781 (expand_mem_signal_fence): Ditto.
20782 (expand_atomic_load): Ditto.
20783 (expand_atomic_store): Ditto.
20785 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
20787 PR rtl-optimization/55547
20788 PR rtl-optimization/53827
20791 * alias.c (memrefs_conflict_p): Set sizes to negative after
20794 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20797 * function.c (thread_prologue_and_epilogue_insns): Always
20798 add crtl->drap_reg to set_up_by_prologue.set, even if
20799 stack_realign_drap is false.
20801 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20803 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20804 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20805 *call): Fix indention.
20807 2013-01-15 Tom de Vries <tom@codesourcery.com>
20810 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20813 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20815 PR rtl-optimization/55153
20816 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20818 2013-01-15 Martin Jambor <mjambor@suse.cz>
20820 PR tree-optimization/55920
20821 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20822 accesses as grp_to_be_debug_replaced.
20824 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20826 PR tree-optimization/55920
20827 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20828 there is non-useless type conversion needed from debug rhs to lhs,
20829 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20831 2013-01-15 Joseph Myers <joseph@codesourcery.com>
20832 Mikael Pettersson <mikpe@it.uu.se>
20835 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
20837 (ASM_OUTPUT_CASE_LABEL): Remove.
20838 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
20839 * final.c (shorten_branches): Update alignment of labels before
20840 jump tables if CASE_VECTOR_SHORTEN_MODE.
20842 2013-01-15 Richard Biener <rguenther@suse.de>
20845 * system.h: Do not include gmp.h for building host tools.
20847 2013-01-15 Richard Biener <rguenther@suse.de>
20849 PR middle-end/55882
20850 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
20851 account for bitpos when computing alignment.
20853 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
20855 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
20856 (ix86_target_macros_internal): Likewise.
20858 * config/i386/i386.c (m_CORE2I7): Removed.
20859 (m_CORE_HASWELL): New macro.
20860 (m_CORE_ALL): Likewise.
20861 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
20862 (initial_ix86_arch_features): Likewise.
20863 (processor_target_table): Initializations for Core avx2.
20864 (cpu_names): New names "core-avx2".
20865 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
20866 PROCESSOR_CORE_HASWELL.
20867 (ix86_issue_rate): New case.
20868 (ia32_multipass_dfa_lookahead): Likewise.
20869 (ix86_sched_init_global): Likewise.
20871 * config/i386/i386.h (TARGET_HASWELL): New macro.
20872 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
20873 (processor_type): New PROCESSOR_HASWELL.
20875 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20877 PR tree-optimization/55955
20878 * tree-vect-loop.c (vectorizable_reduction): Give up early on
20879 *SHIFT_EXPR and *ROTATE_EXPR codes.
20881 PR tree-optimization/48766
20882 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
20883 -ftrapv disable -fwrapv.
20885 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20888 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
20889 etc. to 1 and not to __flash.
20890 Use LL suffix for __INT24_MAX__ with -mint8.
20891 Use ULL suffix for __UINT24_MAX__ with -mint8.
20893 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20895 * config/avr/avr-arch.h
20896 (struct base_arch_s): Use typedef avr_arch_t instead.
20897 (struct arch_info_s): Use typedef avr_arch_info_t instead.
20898 (struct mcu_type_s): Use typedef avr_mcu_t instead.
20899 * config/avr/avr.c: Same.
20900 * config/avr/avr-devices.c: Same.
20901 * config/avr/driver-avr.c: Same.
20902 * config/avr/gen-avr-mmcu-texi.c: Same.
20903 * config/avr/avr-mcus.def: Adjust comment.
20905 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
20907 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
20908 * config/aarch64/iterators.md (VALLDI): New.
20910 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
20911 Andi Kleen <ak@linux.intel.com>
20914 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
20915 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
20918 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20920 * config/avr/avr-stdint.h: Remove trailing blanks.
20921 * config/avr/avr-log.h: Same.
20922 * config/avr/avr-arch.h: Same.
20923 * config/avr/avr-devices.c: Same.
20924 * config/avr/avr-dimode.md: Same.
20925 * config/avr/predicates.md: Same.
20926 * config/avr/avr-c.c: Same. And fix typo.
20928 * config/avr/avr-protos.h: Same. And:
20929 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
20930 (init_cumulative_args): Rename to avr_init_cumulative_args.
20931 (expand_prologue): Rename to avr_expand_prologue.
20932 (expand_epilogue): Rename to avr_expand_epilogue.
20933 (adjust_insn_length): Rename to avr_adjust_insn_length.
20934 (notice_update_cc): Rename to avr_notice_update_cc.
20935 (final_prescan_insn): Rename to avr_final_prescan_insn.
20936 * config/avr/avr.c: Same.
20937 * config/avr/avr.h: Same.
20938 * config/avr/avr.md: Remove trailing blanks.
20939 (prologue): Use avr_expand_prologue.
20940 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
20942 2013-01-14 Richard Biener <rguenther@suse.de>
20944 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
20945 verify_location, collect_subblocks): New functions.
20946 (verify_gimple_in_cfg): Verify that locations only reference
20947 BLOCKs in the functions BLOCK tree.
20949 2013-01-14 Richard Biener <rguenther@suse.de>
20951 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
20953 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
20955 (insert_out_of_ssa_copy_on_edge): Likewise.
20956 (rewrite_close_phi_out_of_ssa): Likewise.
20957 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
20959 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
20960 propagated constants.
20961 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
20964 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20966 * config/avr/avr-modes.def: Add GPL copyright notice.
20968 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
20970 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
20971 MEMMODEL_MASK to determine memory model.
20972 (atomic_store<mode>): Ditto from operands[2].
20973 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
20975 2013-01-13 Jakub Jelinek <jakub@redhat.com>
20978 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
20979 (fold_gimple_assign): Don't call unshare_expr here.
20980 (fold_ctor_reference): Call unshare_expr.
20982 2013-01-13 Terry Guo <terry.guo@arm.com>
20984 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
20985 * doc/fragments.texi: Document MULTILIB_REUSE.
20986 * gcc.c (multilib_reuse): New internal spec.
20987 (set_multilib_dir): Also search multilib from multilib_reuse.
20988 * genmultilib (tmpmultilib3): Refactor code.
20989 (tmpmultilib4): Ditto.
20990 (multilib_reuse): New multilib argument.
20992 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
20994 * Makefile.in: Update copyright.
20996 2013-01-12 Tom de Vries <tom@codesourcery.com>
20998 PR middle-end/55890
20999 * calls.c (expand_call): Check if arg_nr is valid.
21001 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
21003 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
21004 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
21005 documentation. Add missing '__' in front of
21006 __builtin_ia32_packssdw256.
21008 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21011 * config/s390/s390.c (s390_preferred_reload_class): Do not return
21012 NO_REGS for larl operands.
21013 (s390_reload_larl_operand): Use s390_load_address instead of
21016 2013-01-11 Richard Biener <rguenther@suse.de>
21018 * tree-cfg.c (verify_node_sharing_1): Split out from ...
21019 (verify_node_sharing): ... here.
21020 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
21022 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
21024 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
21025 Substitute TREECHECKING.
21026 * configure: Regenerate.
21027 * Makefile.in (TREECHECKING): New.
21029 2013-01-11 Richard Guenther <rguenther@suse.de>
21031 PR tree-optimization/44061
21032 * tree-vrp.c (extract_range_basic): Compute zero as
21033 value-range for __builtin_constant_p of function parameters.
21035 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
21037 Update copyright years.
21039 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
21041 PR rtl-optimization/55672
21042 * lra-eliminations.c (mark_not_eliminable): Permit addition with
21043 const to be eliminable.
21045 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
21047 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
21048 * configure: Regenerate.
21050 2013-01-10 Richard Biener <rguenther@suse.de>
21052 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
21054 2013-01-10 Richard Biener <rguenther@suse.de>
21057 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
21058 locations for virtual PHI arguments.
21059 (rewrite_update_phi_arguments): Likewise.
21061 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
21063 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
21066 2013-01-10 Jakub Jelinek <jakub@redhat.com>
21068 PR tree-optimization/55921
21069 * tree-complex.c (expand_complex_asm): New function.
21070 (expand_complex_operations_1): Call it for GIMPLE_ASM.
21072 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21075 * config/s390/s390.c (s390_symref_operand_p)
21076 (s390_loadrelative_operand_p): Merge the two functions.
21077 (s390_check_qrst_address, print_operand_address): Add parameters
21078 to s390_loadrelative_operand_p invokation.
21079 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
21080 (s390_reload_larl_operand, s390_secondary_reload): Use
21081 s390_loadrelative_operand_p instead of s390_symref_operand_p.
21082 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
21084 2013-01-09 Mike Stump <mikestump@comcast.net>
21086 * dse.c (record_store): Remove unnecessary assert.
21088 2013-01-09 Jan Hubicka <jh@suse.cz>
21090 PR tree-optimization/55569
21091 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
21092 * cfgloop.h (scale_loop_profile): Likewise.
21094 2013-01-09 Jan Hubicka <jh@suse.cz>
21097 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
21099 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
21101 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
21103 PR middle-end/55114
21104 * expr.h (maybe_emit_group_store): Declare.
21105 * expr.c (maybe_emit_group_store): New function.
21106 * builtins.c (expand_builtin_int_roundingfn): Call it.
21107 (expand_builtin_int_roundingfn_2): Likewise.
21109 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
21111 PR rtl-optimization/55829
21112 * lra-constraints.c (match_reload): Add code for absent output.
21113 (curr_insn_transform): Add code for reloads of matched inputs
21116 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
21118 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
21119 attribute of movddup insn to DF.
21120 (*vec_interleave_lowv2df): Ditto.
21121 (vec_dupv2df): Ditto.
21123 2013-01-09 Jan Hubicka <jh@suse.cz>
21125 PR tree-optimiation/55875
21126 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
21127 EVERY_ITERATION parameter.
21128 (number_of_iterations_exit): Check if exit is executed every iteration.
21129 (idx_infer_loop_bounds): Similarly here.
21130 (n_of_executions_at_most): Simplify
21131 to only test for cases where statement is dominated by the
21132 particular bound; handle correctly the "postdominance" test.
21133 (scev_probably_wraps_p): Use max loop iterations info
21134 as a global bound first.
21136 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
21137 Nick Clifton <nickc@redhat.com>
21139 * config/v850/v850.md (cbranchsf4): New pattern.
21140 (cstoresf4): New pattern.
21141 (cbranchdf4): New pattern.
21142 (cstoredf4): New pattern.
21143 (movsicc): Disallow floating point comparisons.
21144 (cmpsf_le_insn): Fix order of operators.
21145 (cmpsf_lt_insn): Likewise.
21146 (cmpsf_eq_insn): Likewise.
21147 (cmpdf_le_insn): Likewise.
21148 (cmpdf_lt_insn): Likewise.
21149 (cmpdf_eq_insn): Likewise.
21150 (cmpsf_ge_insn): Use LE comparison.
21151 (cmpdf_ge_insn): Likewise.
21152 (cmpsf_gt_insn): Use LT comparison.
21153 (cmpdf_gt_insn): Likewise.
21154 (cmpsf_ne_insn): Delete pattern.
21155 (cmpdf_ne_insn): Delete pattern.
21156 * config/v850/v850.c (v850_gen_float_compare): Use
21157 gen_cmpdf_eq_insn for NE comparison.
21158 (v850_float_z_comparison_operator)
21159 (v850_float_nz_comparison_operator): Move from here ...
21160 * config/v850/predicates.md: ... to here. Move GT and GE
21161 comparisons into v850_float_z_comparison_operator.
21162 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
21164 (v850_float_nz_comparison_operator): Likewise.
21166 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21168 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
21169 with calls to gen_insvsi/gen_insvdi.
21171 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21173 * config/i386/i386.c (initial_ix86_tune_features): Set up
21174 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
21176 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
21177 Jakub Jelinek <jakub@redhat.com>
21179 PR tree-optimization/48189
21180 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
21181 If nitercst is 0, don't predict the exit edge.
21183 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21185 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
21186 in asm_fprintf with reg_names.
21187 (aarch64_print_operand_address): Likewise.
21188 (aarch64_return_addr): Likewise.
21189 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
21191 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21193 * config/pa/pa.h (VAL_U6_BITS_P): Define.
21194 (INT_U6_BITS): Likewise.
21195 * config/pa/predicates.md (uint6_operand): New predicate.
21196 (shift5_operand, shift6_operand): Likewise.
21197 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
21199 (lshrdi3): Use shift6_operand.
21200 (shrpsi4, shrpdi4): New insn patterns.
21201 (extzv): Delete expander.
21202 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
21203 predicates in unamed zero extract patterns. Tighten common constraint.
21204 (extv): Delete expander.
21205 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
21206 predicates in unamed sign extract patterns. Tighten common constraint.
21207 (insv): Delete expander.
21208 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
21209 predicates in unamed insert patterns. Tighten common constraint.
21210 Change uint32_operand predicate to uint6_operand predicate in unamed
21211 DImode pattern to insert constant values of type 1...1xxxx.
21213 2013-01-04 Jan Hubicka <jh@suse.cz>
21215 PR tree-optimization/55823
21216 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
21219 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21220 Uros Bizjak <ubizjak@gmail.com>
21222 PR rtl-optimization/55845
21223 * df-problems.c (can_move_insns_across): Stop scanning at
21224 volatile_insn_p source instruction or give up if
21225 across_from .. across_to range contains any volatile_insn_p
21228 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21230 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
21231 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
21233 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
21234 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
21236 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21239 * asan.c (asan_clear_shadow): New function.
21240 (asan_emit_stack_protection): Use it.
21242 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21244 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
21245 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
21246 with tab instead of space.
21248 2013-01-08 Nick Clifton <nickc@redhat.com>
21250 * config/rl78/rl78.c (rl78_expand_prologue): Always select
21251 register bank 0 at the start of an interrupt handler.
21252 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
21255 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21257 * config/aarch64/aarch64-simd.md
21258 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
21259 (aarch64_simd_bsl): Likewise.
21260 (aarch64_vcond_internal<mode>): Likewise.
21261 (vcond<mode><mode>): Likewise.
21262 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
21263 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
21265 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21267 * config/aarch64/aarch64-builtins.c
21268 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
21270 2013-01-08 Martin Jambor <mjambor@suse.cz>
21273 * tree-sra.c (analyze_access_subtree): Return true also after
21274 potentially creating a debug-only replacement.
21276 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21278 PR middle-end/55890
21279 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
21281 PR tree-optimization/54120
21282 * tree-vrp.c (range_fits_type_p): Don't allow
21283 src_precision < precision from signed vr to unsigned_p
21284 if vr->min or vr->max is negative.
21285 (simplify_float_conversion_using_ranges): Test can_float_p
21286 against CODE_FOR_nothing.
21288 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21289 Richard Biener <rguenther@suse.de>
21291 PR middle-end/55851
21292 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
21293 types instead of just INTEGER_TYPE types.
21295 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
21297 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
21300 2013-01-07 Steve Ellcey <sellcey@mips.com>
21303 * config/mips/mips.opt: Change mad to mmad to match documentation.
21305 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21308 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
21309 .progmemx.data now.
21311 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21314 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
21315 (avr_addrspace_t): Add .section_name field.
21316 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
21318 (avr_addrspace): Same. Initialize .section_name. Remove last
21319 NULL entry. Put __memx into .progmemx.data.
21320 (progmem_section_prefix): Remove.
21321 (avr_asm_init_sections): No need to initialize progmem_section.
21322 (avr_asm_named_section): Use avr_addrspace[].section_name to get
21323 section name prefix.
21324 (avr_asm_select_section): Ditto. And use get_unnamed_section to
21325 retrieve the progmem section.
21326 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
21327 boundary to run over avr_addrspace[].
21328 (avr_register_target_pragmas): Ditto.
21330 2013-01-06 Jakub Jelinek <jakub@redhat.com>
21332 * varasm.c (output_constant_def_contents): For asan_protect_global
21333 protected strings, adjust DECL_ALIGN if needed, before testing for
21335 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
21336 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
21338 (output_object_block): For asan protected decls, emit asan padding
21339 after their contents.
21340 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
21341 (asan_finish_file): Test it here instead.
21343 2013-01-07 Nick Clifton <nickc@redhat.com>
21344 Matthias Klose <doko@debian.org>
21345 Doug Kwan <dougkwan@google.com>
21346 H.J. Lu <hongjiu.lu@intel.com>
21349 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
21351 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
21353 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
21355 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
21357 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
21359 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21362 * doc/install.texi (Cross-Compiler-Specific Options): Document
21365 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
21367 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
21368 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
21369 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
21370 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
21371 vqmovun_high_s64): Fix source operand number and update copyright.
21373 2013-01-07 Richard Biener <rguenther@suse.de>
21375 PR middle-end/55890
21376 * gimple.h (gimple_call_builtin_p): New overload.
21377 * gimple.c (validate_call): New function.
21378 (gimple_call_builtin_p): Likewise.
21379 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21380 Use gimple_call_builtin_p.
21381 (find_func_clobbers): Likewise.
21382 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
21383 (strlen_optimize_stmt): Likewise.
21385 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21387 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
21388 (vld1q_dup_*): Likewise.
21389 (vld1_*): Likewise.
21390 (vld1q_*): Likewise.
21391 (vld1_lane_*): Likewise.
21392 (vld1q_lane_*): Likewise.
21394 2013-01-07 Richard Biener <rguenther@suse.de>
21396 * lto-streamer.h (LTO_minor_version): Bump to 2.
21398 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21400 * config/aarch64/aarch64-protos.h
21401 (aarch64_const_double_zero_rtx_p): Rename to...
21402 (aarch64_float_const_zero_rtx_p): ...this.
21403 (aarch64_float_const_representable_p): New.
21404 (aarch64_output_simd_mov_immediate): Likewise.
21405 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
21406 move immediate case.
21407 * config/aarch64/aarch64.c
21408 (aarch64_const_double_zero_rtx_p): Rename to...
21409 (aarch64_float_const_zero_rtx_p): ...this.
21410 (aarch64_print_operand): Allow printing of new constants.
21411 (aarch64_valid_floating_const): New.
21412 (aarch64_legitimate_constant_p): Check for valid floating-point
21414 (aarch64_simd_valid_immediate): Likewise.
21415 (aarch64_vect_float_const_representable_p): New.
21416 (aarch64_float_const_representable_p): Likewise.
21417 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
21418 (aarch64_output_simd_mov_immediate): New.
21419 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
21420 (*movdf_aarch64): Likewise.
21421 * config/aarch64/constraints.md (Ufc): New.
21422 (Y): call aarch64_float_const_zero_rtx.
21423 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
21425 2013-01-07 Richard Biener <rguenther@suse.de>
21427 PR tree-optimization/55888
21428 PR tree-optimization/55862
21429 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
21430 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
21431 not if it is contained therein.
21433 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21435 * config/avr/t-avr: Typo.
21437 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21440 * config/avr/t-avr: Don't automatically rebuild
21441 $(srcdir)/config/avr/t-multilib
21442 $(srcdir)/config/avr/avr-tables.opt
21443 $(srcdir)/doc/avr-mmcu.texi
21444 (avr-mcus): New phony target to build them on request.
21445 (s-avr-mlib, s-avr-mmcu-texi): Remove.
21446 * avr/avr-mcus.def: Adjust comments.
21448 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
21450 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
21452 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21454 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
21456 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21458 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
21460 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
21462 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
21463 to generate profiling.
21464 * config/rs6000/aix64.h (LIB_SPEC): Same.
21466 2013-01-04 Andrew Pinski <apinski@cavium.com>
21468 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
21470 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
21472 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
21474 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
21476 (ix86_expand_move): Ditto.
21477 (ix86_zero_extend_to_Pmode): Ditto.
21478 (ix86_expand_call): Ditto.
21479 (ix86_expand_special_args_builtin): Ditto.
21480 (ix86_expand_builtin): Ditto.
21482 2013-01-04 Richard Biener <rguenther@suse.de>
21484 PR tree-optimization/55862
21485 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21486 translating them through PHI nodes.
21488 2013-01-04 Martin Jambor <mjambor@suse.cz>
21490 PR tree-optimization/55755
21491 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21492 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21494 2013-01-04 Marek Polacek <polacek@redhat.com>
21496 PR middle-end/55859
21497 * opts.c (default_options_optimization): Clarify error message.
21499 2013-01-04 Richard Biener <rguenther@suse.de>
21501 PR middle-end/55863
21502 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21505 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21508 * config/pa/pa.md (movsi): Revert previous change.
21509 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21512 2013-01-03 Richard Henderson <rth@redhat.com>
21514 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21515 after eliminating TLS symbols.
21517 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
21520 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21521 * graphite-poly.c (debug_gmp_value): Likewise.
21523 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
21526 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21527 selected code model, define __code_mode_small__, __code_model_medium__,
21528 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21529 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21530 xchg temporary register with %k. Declare temporary register as
21532 [__x86_64__]: For medium and large code models, preserve %rbx register.
21534 2013-01-03 Richard Biener <rguenther@suse.de>
21536 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
21537 (dump_subscript): Adjust.
21538 (finalize_ddr_dependent): Do not dump redundant info.
21539 (analyze_siv_subscript): Adjust.
21540 (subscript_dependence_tester): Likewise.
21541 (compute_affine_dependence): Likewise.
21543 2013-01-03 Richard Biener <rguenther@suse.de>
21546 2013-01-03 Richard Biener <rguenther@suse.de>
21548 PR tree-optimization/55857
21549 * tree-vect-stmts.c (vectorizable_load): Do not setup
21550 re-alignment for invariant loads.
21552 2013-01-02 Richard Biener <rguenther@suse.de>
21554 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21555 invariant load do not generate a vector load from the scalar location.
21557 2013-01-03 Richard Biener <rguenther@suse.de>
21559 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21560 for not vectorizing.
21561 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21562 not build INDIRECT_REFs, call get_name once only.
21563 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21564 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21566 2013-01-03 Richard Biener <rguenther@suse.de>
21568 PR tree-optimization/55857
21569 * tree-vect-stmts.c (vectorizable_load): Do not setup
21570 re-alignment for invariant loads.
21572 2013-01-03 Richard Biener <rguenther@suse.de>
21575 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
21576 prefer a built-in decl.
21578 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21580 * gcc.c (process_command): Update copyright notice dates.
21581 * gcov.c (print_version): Likewise.
21582 * gcov-dump.c (print_version): Likewise.
21584 PR rtl-optimization/55838
21585 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
21586 iv0.step, iv1.step and step.
21588 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21589 Marc Glisse <marc.glisse@inria.fr>
21591 PR tree-optimization/55832
21592 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
21593 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
21594 integer_{one,zero}_node.
21596 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21599 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
21600 * var-tracking.c (reverse_op): Don't add reverse ops to
21601 VALUEs that have already
21602 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
21604 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
21606 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
21608 2013-01-02 Teresa Johnson <tejohnson@google.com>
21610 * dumpfile.c (dump_loc): Print filename with location.
21611 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
21612 new location_t parameter to emit complete unroll message with
21613 new dump framework.
21614 (canonicalize_loop_induction_variables): Compute loops location
21615 and pass to try_unroll_loop_completely.
21616 * loop-unroll.c (report_unroll_peel): New function.
21617 (peel_loops_completely): Use new dump format with location
21618 for main dumpfile message, and invoke report_unroll_peel on success.
21619 (decide_unrolling_and_peeling): Ditto.
21620 (decide_peel_once_rolling): Remove old dumpfile message subsumed
21621 by report_unroll_peel.
21622 (decide_peel_completely): Ditto.
21623 (decide_unroll_constant_iterations): Ditto.
21624 (decide_unroll_runtime_iterations): Ditto.
21625 (decide_peel_simple): Ditto.
21626 (decide_unroll_stupid): Ditto.
21627 * cfgloop.c (get_loop_location): New function.
21628 * cfgloop.h (get_loop_location): Declare.
21630 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21632 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21635 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21637 PR middle-end/55198
21638 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21639 BLKmode objects when EXPAND_MEMORY is specified.
21641 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21643 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21645 (fold_builtin_cpu): Do not share cpu model decls across statements.
21647 2013-01-02 Jason Merrill <jason@redhat.com>
21650 * tree.c (build_array_type_1): Revert earlier change.
21652 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21654 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21656 * config/aarch64/aarch64-tune.md: Re-generate.
21658 2013-01-02 Richard Biener <rguenther@suse.de>
21660 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21661 invariant load do not generate a vector load from the scalar location.
21663 2013-01-02 Richard Biener <rguenther@suse.de>
21666 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21667 * configure: Regenerate.
21669 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21671 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21672 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21673 (expand_builtin_int_roundingfn_2): Keep the original target around
21674 for the fallback case.
21676 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21678 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21679 to be clear for sign changes.
21681 2013-01-01 Jan Hubicka <jh@suse.cz>
21683 * ipa-inline-analysis.c: Fix formatting.
21685 2013-01-01 Jakub Jelinek <jakub@redhat.com>
21687 PR tree-optimization/55831
21688 * tree-vect-loop.c (get_initial_def_for_induction): Use
21689 gsi_after_labels instead of gsi_start_bb.
21691 Copyright (C) 2013 Free Software Foundation, Inc.
21693 Copying and distribution of this file, with or without modification,
21694 are permitted in any medium without royalty provided the copyright
21695 notice and this notice are preserved.